Residue Management Strategies for the Rainfed N-Deprived Maize-legume Cropping
Systems of Central Mozambique
Nascimento Salomão Nhantumbo
(BSc in Agronomy and MSc in Erosion, Soil and Water Conservation)
A thesis submitted for the degree of Doctor of Philosophy at
The University of Queensland in 2016
Queensland Alliance for Agriculture and Food Innovation Institute (QAAFI)
Page | i
ABSTRACT
Identifying best-fit residue management strategies for Central Mozambique rain fed N-
deprived maize-legume cropping (LEINA) systems managed under conservation agriculture
is critical to improve resource use efficiency and yields. Therefore, understanding under what
circumstances positive and negative responses from carbon rich residue mulches are likely to
occur and what drives farmers’ decision on the adoption of technologies is critical to design
relevant and feasible recommendations that fit farmers complex farming circumstances. Here
I used household survey data, field experimentation and cropping systems modelling to: 1)
characterize farm diversity and understand how it affects smallholder farmers resource use
attitudes, management decisions and the likelihood to engage in sustainable intensification
(SI) practices; 2) identify best-fit residue management strategies that are more likely to
improve resource productivity and yields; and 3) develop simple rules of thumb to match
residue management solutions to N-deprived smallholder farming systems.
Household characterization data showed that, rather than farm size and labour availability,
farmers resource use attitudes and household capacity to meet its annual food security and
income generation targets where key farm differentiation factors, therefore critical to map
farm typology feasible intensification options. Using a heuristic decision tree to group
smallholder farmer into a new set of mutually exclusive farm typologies that highlight
existing intragroup diversity among Mozambican small-scale farmers, three key typologies
where identified. The resource endowed and innovative farmers (Type B); resource endowed
but change resistant farmers (Type - A2); and resource constrained (Type - A1) farmers, who
can also be maize self-sufficiency (Type-A1a) and the food insecure (Type-A1b). For the
already maize self-sufficient Type-B and Type-A2 farms, the challenge is to improve
agronomy to generate enough surplus and income from maize. Mechanization and labour
qualification are critical for these groups. For resource constrained Type-A1 farmers, the
primary challenge towards intensification is to generate enough income to improve the
household maize self-sufficiency and income in the system. Considering its high market
value and further contribution to soil N-availability, legume production is a feasible
investment alternative. Nevertheless, two key intensification traps were identified: the
household maize self-sufficiency and soil fertility perception, and land availability perception
trap (extensification). The traps are perception based mind-sets that reflect farmers
understanding of their circumstances and the strategies they use to materialize their annual
food security and income generation targets.
Page | ii
Field trials and model-assisted analyses (64 years) of carbon rich residues effects on a cereal
(maize) and a legume (cowpea) crop, at different levels of N supply, and three soils of
contrasting water holding capacity (WHC) showed that for Central Mozambique rain fed N-
deprived systems, residues are more likely to (i) improve the yield, WUE and NUE of the
unfertilized legume crop across all tested environments; and (ii) that the application of carbon
rich residues reduces maize yields, NUE and WUE in a high WHC sandy clay loam (cSaCL)
soil, and on the wet lowland sandy loam textured soil (SaL). In high WHC soils, only a 5%
chance of positive yield response from carbon rich mulched soils was simulated at 0 (ND)
and 23 kg N/ha (LNA). At 92 kg N/ha (HNA) benefits from carbon rich residues are expected
only in 20% of the seasons in high WHC soils. However, such benefits are only attained
during the driest years. In the dry and low WHC fine sandy soil (fSa), positive responses to
residue application were simulated in 85-90% of the seasons in both cowpea and maize. In
terms of resource productivity, maize proved to be more responsive to N-application,
especially on the wet and moisture rich SaL soil. Here, maize response to N-application was
attributed to better nitrogen uptake (NUpE) and translocation efficiency (NUtE) due to high
in crop moisture regimes. In the drier fSa, poor soil water availability led to poor NUpE and
consequently lower maize NUE and yields. For cowpea, yields at HNA were hindered by
poor N-translocation into grain. Despite maize providing the best responses to N-fertilizer
application in mulch free systems, yield penalties from residue applications in continuous
maize cropping systems.
The findings from this study showed that, in the rain fed N-deprived systems of central
Mozambique, the overall performance of maize-legume cropping systems under residues is
governed by two critical interactions: 1) a crop and soil water induced residue response trade-
off; 2) a residue modified nitrogen driven WUE and NUE trade-off. Understanding these two
responses is key to validate locally feasible resource management strategies crucial to the
effective validation of CA systems. However, because resource access and management
differs across households, an interactive and personalized socio-technical intervention that
takes into account each group biophysical and socioeconomic circumstances is required to
successfully involve farmers in co-designing group specific technological interventions
centred on the improved use of locally available resources and those that farmers can afford
to mobilize and use.
Page | iii
Declaration by Author
This thesis is composed of my original work, and contains no material previously published
or written by another person except where due reference has been made in the text. I have
clearly stated the contribution by others to jointly-authored works that I have included in my
thesis.
I have clearly stated the contribution of others to my thesis as a whole, including statistical
assistance, survey design, data analysis, significant technical procedures, professional
editorial advice, and any other original research work used or reported in my thesis. The
content of my thesis is the result of work I have carried out since the commencement of my
research higher degree candidature and does not include a substantial part of work that has
been submitted to qualify for the award of any other degree or diploma in any university or
other tertiary institution. I have clearly stated which parts of my thesis, if any, have been
submitted to qualify for another award.
I acknowledge that an electronic copy of my thesis must be lodged with the University
Library and, subject to the policy and procedures of The University of Queensland, the thesis
be made available for research and study in accordance with the Copyright Act 1968 unless a
period of embargo has been approved by the Dean of the Graduate School.
I acknowledge that copyright of all material contained in my thesis resides with the copyright
holder(s) of that material. Where appropriate I have obtained copyright permission from the
copyright holder to reproduce material in this thesis.
Page | iv
Publication during candidature
Peer-reviewed papers
Nhantumbo, N., Mortlock, M., Nyagumbo, I., Dimes, J and Rodriguez (nd) Residue and
fertilizer management strategies for the rain fed, N-deprived maize-legume cropping systems
in Central Mozambique, submitted to the Field Crop Research Journal
Conference abstracts
Nhantumbo, N. Dias, J., Mortlock, M., Nyagumbo, I., Dimes, J and Rodriguez, D (2014)
Best-fit residue allocation: A gate for legume Intensification in nitrogen constrained systems
of Central Mozambique. In “Proceedings from First African Congress on Conservation
Agriculture" (African Conservation Tillage Network, ed.), pp. 33-35. ACT, 18-21 March
Lusaka, Zambia.
Nhantumbo, N., Mortlock, M., Nyagumbo, I., Dimes, J and Rodriguez, D (2015) Improving
resource allocation in nitrogen constrained systems: acknowledging within and cross-farm
variability effect on yield and NUE. In “Proceedings from the 5th International Symposium
for Farming Systems Design: Multi-functional farming systems in a changing world”. Edited
by ESA and Agropolis International, pp 491-492, 7-10 September, Montpellier, France.
Nhantumbo, N., Mortlock, M., Nyagumbo, I., Dimes, J and Rodriguez, D (2015) Improving
farming systems design and management: the role of model assisted participatory crop season
planning in Sub-Saharan Africa. Presented at the TropAg2015 conference in Brisbane
Publications included in this thesis
None
Page | v
Contributions by others to the thesis
Key Mathews, a part-time statistical advisor for QAAFI RHD students, provided early advice
on experimental design and analysis techniques using GenStat during proposal
development and preparations for the first field trip.
For chapter 4, preliminary soil characterization and routine plant analysis were conducted at
the Cropnuts Soil and Plant Laboratory at ICRAF in Nairobi, with further analysis –
nitrate N (NO3-N) determined at the Soil and Plant Analysis Lab of the Faculty of
Agriculture of the Manica Higher Polytechnic Institute with Support from Eng.
Jacinto M. Andreque and Eng. Clemente Zivale.
Miranda Mortlock, have helped revise most of chapter 3, offering key advise on data
presentation and interpretation. Miranda also provided further statistical advice for
chapter 3 and 4 of this thesis.
John Dimes and Peter DeVoil provided technical backstop on APSIM model. This consisted
on direct training and further assistance with model parameterization for chapter 5.
Language editing: Dr John Schiller
Statement of parts of the thesis submitted to qualify for the award of another degree
None
Page | vi
Acknowledgements
I have learned a lot throughout this PhD journey. I am truly grateful to the SIMLESA Project
and also to the Australian Centre for International Agricultural Research (ACIAR) which,
through the John Allwright Fellowship Program, gave me this unique opportunity. My special
thanks go to my supervisor, Dr. Daniel Rodriguez, for keeping his word after our first casual
meeting in August 2010, during the SIMLESA launching meeting, when he gave me his
business card and immediately shared with me scholarship application information. I will be
always grateful for this gesture. I also acknowledge the SIMLESA Africa team, Dr Mullugeta
(SIMELSA Africa Coordinator), Dr Isaiah Nyagumbo (Malawi and Mozambique, Objective
2 Coordinator and my Co-supervisor) and Eng. Domingos Dias (SIMLESA Country
Coordinator) for bringing me into the SIMLESA family. I also want to thank Dr John Dimes
for the great discussions we had on residue management in Nairobi during the breaks of the
1st SIMLESA annual planning meeting back in March 2011. Thanks also for the endless
discussions and support relating to APSIM modelling throughout my PhD. My extended
thanks also go to the Instituto Superior Politecnico de Manica Director General, Dr Rafael
Massinga, for introducing me to the SIMLESA team. Without the direct or indirect
contribution of those people I wouldn’t have made it to UQ. For that reason, I will always be
grateful to you for all for this opportunity.
I am exceedingly grateful to my team of advisors for their support throughout my PhD
studies, and especially for keeping up with me throughout the process. Managing my field
work back in Mozambique, the re-allocation to Australia far from my family, and
administrating the tight schedule between my arrival in Australia after field work and the
milestone attainment deadlines was really challenging. I couldn’t have made it without the
support of my advisory team, especially Dr. Daniel Rodriguez, who had to directly manage
all these issues. To Miranda, for the valuable inputs in the survey design and analysis, and
precious time spent talking about it. Thanks also for bringing the human element into my
research and time I spent in Australia. All these experiences were rewarding for me and made
me grow both as a researcher and human being.
Page | vii
I would also like to send my special thanks to Ms Sharon Harvey in ACIAR for her support
throughout the scholarship process. Special thanks, also goes to my fellow PhD colleagues,
Abeya, Mesfin, Solomon Admassu and his family and Solomon Hassem, for the time we
spent in Australia together. To Caspar for the talks in Mozambique during our field work. I
wish you all the best, hoping that this is not the last time we do something together.
My special thanks to my beloved wife Ana, without whom I couldn’t have made it so far. Te
amo muito e ser-te-ei eternamente grata pelo amor, apoio incondicional, amizade,
companheirismo e principalmente por me aturares e teres sido o alicerce da família durante
este processo todo. To our kids, Makhida, Aione and Thulani, o pai pede imensas desculpas
por ter perdido estes 4 anos da vossa vida. Espero que um dia vos possa compensar e me
possam perdoar pela ausência. Espero também que este sacrifício tenha valido a pena e
possa ajudar a mostrar-vos que a gente pode chegar onde quiser desde que estude e corra
atras dos nossos objetivos.
Special thanks also to my family - mum, dad, brothers Betinho, Pique, Nina, and Salo, and
their families, for the endless support and words of encouragement throughout my studies.
Nothing that I do will ever payback the support you always gave me. Betinho obrigado por
olhares pelos meus matewes na minha ausência.
I would also like to thank the Instituto Superior Politecnico de Manica management board for
allowing me to pursue my studies. To my colleagues at the Faculty of Agriculture (DivAG),
without forgetting the Faculty of Economy, the Technical and Administrative Support Team
(CTA’s), students, my field team and friends in general, for their words of encouragement
throughout this journey. I hope my journey can help inspire you to fallow the same path so
that in the near future we can have an institution and country with people highly qualified and
actively conducting research that can help make a difference in the lives of those most in
need.
Page | viii
Keywords:
residue-fertilizer management, resource allocation, farm typologies, water use efficiency
(WUE), nitrogen use efficiency (NUE), maize-legume cropping systems, conservation
agriculture, APSIM, sustainable intensification, technology adoption
Australian and New Zealand Standard Research Classifications (ANZSRC)
ANZSRC code: 070105 Agricultural Systems Analysis and Simulation, 40 %
ANZSRC code: 070107 Farming Systems Research, 30 %
ANZSRC code: 070302 Agronomy, 30%
Fields of Research (FoR) Classification
FoR code: 0701 Agriculture, Land and Farm Management, 60 %
FoR code: 0703 Crop and Pasture Production, 30%
FoR code: 0799 Other Agricultural and Veterinary sciences, 10%
Page | ix
Table of Contents
ABSTRACT i
List of Tables xiii
List of Figures xv
Abbreviations xix
Thesis outline xx
CHAPTER 1 23
1. General Introduction 23
1.1 General Research Background 23
1.2. Research Questions, Objectives and Hypothesis 25
1.3 Study area description 27
1.3.1 Target Agroecological zone 28
1.4 General Methodology 29
1.4.2 Trial site characterization 34
CHAPER 2: 39
2. Literature Review 39
2.1 Fertilizer use and demand promotion in Sub-Saharan Africa 39
2.2 Tailoring fertilizer recommendations to SSA farming environments 40
2.2.1 Optimized fertilizer recommendations and Micro-dosing 41
2.2.2 Integrated Nutrient Management (INM) 42
2.2 Fertilizer, Residues and Legumes interaction in SSA Conservation Agriculture 45
2.2.1 Crop residue contribution to yield, water and N-availability 45
2.2.2 Yield penalties from residue retention in continuous cereal systems 46
2.2.3 Competing claims on residue use: from mixed crop-livestock systems to non-
cattle dominated systems 47
2.3 Legumes integration in SSA smallholder systems managened under CA 48
2.3.1 Legume performance in Africa 48
2.3.2 Potential contribution from legume crops in improving N-fixation 52
2.4 Using agricultural simulation models as decision support tools in SSA 54
2.4.1 Using the Agricultural Production Systems Simulation Model (APSIM) 55
2.4.2 Gaps in modelling crop residue decomposition and N-cycling with APSIM 57
References 59
CHAPTER 3: 74
Page | x
Characterising Farmers Diversity and Opportunities for Sustainable Intensification 74
Abstract 75
3.2 Materials and Methods 78
3.2.1 Geographic setting and context of our study 78
3.2.2 Sampling and data analysis 79
3.3 Results 83
3.3.1 Farm household characteristics - Household demographic and land use 83
3.3.2 Farming systems design and management 84
3.3.3 Resource use and allocation strategies 87
3.3.4 Farm typologies formulation 90
3.4.5 Farm typologies 94
3.5 Discussion 98
3.5.1 Improving agricultural advisory services through typology tailored intervention 98
3.5.2 Handling the key intensification traps across contrasting farm typologies 100
3.5.3 Building from current farming systems design 103
3.6 Conclusions 104
References 105
CHAPTER 4: 111
Best-fit Residue Management in Rain Fed N-Deprived Systems 111
Abstract 112
4.1 Introduction 113
4.2. Materials and Methods 114
4.2.1 Field Experimentation 114
4.2.2 Measurements 116
4.2.3 Statistical analysis of field experiment data sets 117
4.3. Results 117
4.3.1 Crop growing conditions – rainfall distribution 117
4.3.2 Maize and cowpea yield response to residue and fertilizer application across
sites 118
4.3.3 Resource Utilization 122
4.3.4. Short-term effects from residues and N-fertilizer allocation startegies 130
4.4. Discussion 134
4.6. Conclusions 138
Page | xi
Acknowledgements 139
References 139
CHAPTER 5: 146
Pathways for Sustainable Residue Management 146
Abstract 147
5.1 Introduction 148
5.1.1 Conceptual framework 149
5.2. Materials and Methods 151
5.2.1 Experiment for model parameterization 151
5.2.2 Long term simulations: residue induced yield, NUE and WUE responses 154
5.2.3 Model testing and Evaluation 155
5.3. Results 155
5.3.1 Model evaluation 155
5.3.2 Interactions between residue and fertilizer and their impact on yield 158
5.3 Resource productivity 161
5.4. Seasonal rainfall variability and their effects on residue responses 163
5.5 Discussion 166
5.6 Conclusion 170
References 170
CHAPTER 6 177
General Discussion 177
6.1 Summary of key findings 178
6.2 Typology tailored intervention: key entry points to downscale technologies 180
6.3 Optimizing responses through smart and pragmatic resource allocation strategies 182
6.3.1 Residue and fertilizer responses: linking empirical and model assisted analysis 182
6.3.2 Model assisted analysis and decision support framework 183
6.3.4 Demonstrating short term effects from alternative resource allocation strategies 185
6.4 Conclusions and way forward 186
6.4.1 Scaling out and scaling up innovations 187
References 191
Appendix 3.1 Generalized crop management timelines across regions and farmer
groups 199
Page | xii
Appendix 3.2 Crop distribution across farming environments and soils in both AEZ-
R4 and AEZ-R10. 200
Appendix 3.4: Farm household clusters by their fertilizer and manure allocation
preferences 201
Appendix 3.5: Households clustered by their reasoning for not using fertilizers or
manure 201
Appendix 3.6: Fertilizer and manure allocation preferences among residue users 202
Appendix 3.7: Correlation between household size and number of household farm
active people disaggregated by age and sex 204
Page | xiii
List of Tables
Table 1-1 Treatment key with indication of tested factors disaggregated by site ................... 33
Table 1-2 Soil starting condition in the first season (2013-14) of trials ................................. 35
Table 2-1 Main Lessons from Fertilizer Research and Promotion in SSA in Recent
Years .......................................................................................................................... 44
Table 2-2 Current findings on legume contributions to nitrogen fixation in African
cropping systems ....................................................................................................... 51
Table 2-3 The potential of APSIM in aiding agricultural research in SSA ............................ 56
Table 3-1 Household and farming systems characterization variables used for the
categorical principal components analysis (CATPCA) ............................................. 80
Table 3-2 Household characterization based on demographic data and land use ................... 84
Table 3-3 Loading values for the correlation matrix for 21 farm household
characteristics in Matsinho (Vanduzi) and Marera (Macate). ................................... 91
Table 3-4 Households distribution and characterization across farm types and sites ............. 95
Table 4-1 Significance levels for the main factors (Crop, R-levels, N-levels, Site)
and their interactions effects on yield, stover, crop water use (CWU), water
use efficiency (WUE), nitrogen use efficiency (NUE), nitrogen uptake
efficiency (NUpE), nitrogen utilization efficiency (NUtE) and N-
partitioning into shoot N (Nsh) and Grain N (Ng), at physiologic maturity ........... 121
Table 4-2 Significance levels for the residual effects from residue-fertilizer
allocation strategies on yield and resource productivity of cowpea-maize
and maize-maize sequences measured in the 2014-15 season ................................. 131
Table 4-3 Short term yield, N-uptake, NUpE, NUtE, NUE and WUE advantages
from residue and nitrogen fertilizer application into cowpea attained in a
subsequent bioassay maize crop. Results were obtained after one-year
comparison between unfertilized maize grown in plots previously planted to
cowpea and maize at different residue and N-application levels. ........................... 133
Page | xiv
Table 5-1 Initial soil organic carbon (SOC), nitrate-nitrogen (NO3-N) and soil water
description parameters from soil samples collected in the 2013-14 seasons
and were used for model parameterization .............................................................. 153
Table 5-2 Overall model evaluation for selected parameters ................................................ 156
Table 5-3 Inter-seasonal performance evaluated through variations in long term
simulated maize and cowpea water productivity (WUE) and nitrogen use
efficiency (NUE) response to residue application across N-levels for a low
N and low water holding capacity environment. ..................................................... 165
Page | xv
List of Figures
Figure 1-1 Trial layout. For each site a layout was generated through the GenStat 16
trial design tool. At site level, crop types was the whole plot, residue levels
where the sub-plots and N-level the sub-subplots ..................................................... 33
Figure 1-2 Topsoil volumetric soil water content (VWC) measured between
flowering and physiological maturity across the three tested soils in the
2013-14 and 2014-15 cropping seasons. ................................................................... 37
Figure 1-3 Post-harvest soil nitrate (NO3-N) distribution down the profile on maize
(ABC) and cowpea (DEF) treatments and sites. Shapes identification: ∆ -
Sandy Clay loam (cSaCL); O Fine sandy (fSa) and ◊ Sandy loam soil
(SaL). For each soil and N-level (ND: 0 kg N/ha; LNA: 23 kg N/ha and
HNA: 92 kg N/ha), R represents residue treatments, i.e., with (R+) and
without (R-) residue application ................................................................................ 38
Figure 3-1 Cropping systems description – crops and sequences for the AEZ-R4
gathered through the model assisted focus group discussions and resource
allocation maps (RAM’s) designed by local farmers ................................................ 86
Figure 3-2 Preferred fertilizer, manure and residue allocation across crops over the
last 10 years in sampled districts. Interpretation: 1-Mz is maize, 2-Leg is
legume and 3-Veg is vegetables and 4 is shared across crops, excluding
vegetables. ................................................................................................................. 89
Figure 3-3 Framework for farmer categorization and personalize agricultural
intervention. Key grouping factors were attitudes towards resource use
(fertilizer, manure) and the household maize self-sufficiency level .......................... 93
Figure 3-4 Hypothetical model for resource endowment based farm typologies.
Typology specific develop pathways are indicated by dashed arrows.
Continuous arrow indicates a stepping out situation (Dorward et al., 2009),
i.e., a level of income diversification where the farmer is able to generate
enough off the farm income to reduce dependence from farming. ............................ 94
Figure 4-1 Rainfall distribution, daily maximum, minimum temperature and
radiation from 1st September 2013 to 4th April 2014 in the 2013-14 cropping
Page | xvi
season (left) and 2014-15 cropping season (right). S is sowing date, (F)
flowering and (H) harvest ........................................................................................ 118
Figure 4-2 Crop response to residue application across different N-fertilization
levels on three contrasting soil types from Central Mozambique. ND:
Unfertilized N-deficient system (0 kg N/ha), LNA: low nitrogen application
system (23 kg N/ha), HNA: high N-application (92 kg N/ha). S – significant
differences between residue application levels. Different letters across N-
levels for same R-level indicated significant differences at p<0.05
significance. Caps lock – separate treatments without residues. Small letter
separate situations with residues .............................................................................. 120
Figure 4-3 Maize and cowpea grain yield as a function of crop water use from
emergence to maturity measured across sites. The continuous line
represents the attainable water use efficiency of 20 kg grain/ha.mm (a) and
10 kg grain/ha.mm (b), for maize and cowpea respectively. The dotted lines
represent the maximum actual WUE for each crop across the tested
environments. Theoretical WUE, based on APSIM model simulations for
this agroecological zone are of 17.9 kg grain/ha.mm for maize and 7.74 kg
grain/ha.mm for cowpea, respectively. .................................................................... 123
Figure 4-4 Relationship between crop WUEyield and N-fertilizer application across
sites (N-level x Site interaction). N-levels: ND: N-deficient system (0 kg
N/ha), LNA: low N-application system (23 kg N/ha) and HNA: Non-
limiting N (92 kg N/ha). Soil textures: fSa – fine sand soil, SaL – sandy
loam soil and cSaCL - coarse sandy clay loam soil. The vertical line in each
plot represents the LSD value for WUEyield comparison between N-levels
for each site at p<0.05. N-levels spaced higher than the LSD bar for the
same site are significantly different. ........................................................................ 125
Figure 4-5 Relationship between crops NUE and N-fertilizer application across
sites. N-levels: ND: N-deficient system (0 kg N/ha), LNA: low N-
application system (23 kg N/ha) and HNA: Non-limiting N (92 kg N/ha).
Soil textures: fSa – fine sand soil, SaL – sandy loam soil and cSaCL -
coarse sandy clay loam soil. Continuous line in each plot represent the LSD
value for NUE comparison between N-levels for each site at p<0.05. N-
Page | xvii
levels spaced higher than the LSD bar for the same site are significantly
different. ................................................................................................................... 127
Figure 4-6 Relationship between NUE and its components observed across sites of
contrasting moisture regime. In maize as NUpE was the primary driver of
NUE (a.1), while in cowpea (b.1, b.2) a, NUtE appeared to be major driver
of NUE changes. Shape identification: ∆ - Sandy Clay loam (cSaCL); O
Fine sandy (fSa) and ◊ Sandy loam soil (SaL). Empty and filled shape
indicate residue application levels ........................................................................... 128
Figure 4-7 Nitrogen driven trade-offs between NUE-WUE. NUE and WUE were
averaged across sites and residue application levels for maize (a) and
cowpea (b). Bellow: Maize (c) and cowpea (d) trade-offs as affected by
residue application levels. NUE and WUE for both crops were averaged
across sites. Residue application levels: R- and R+ indicate residue free and
residue applied systems. .......................................................................................... 129
Figure 4-8 Short term yield, NUpE and NUE return from a maize crop grown in
plots previously planted to cowpea (Cwp-Mz) and maize (Mz-Mz) under
contrasting residue and nitrogen application levels Treatment description:
ND, LNA and HNA represent unfertilized, low and high N-application
respectively. Previous residue status: with (+) and without (-) residue. (s)
represents significant differences between residual treatment effects at
p<0.05. ..................................................................................................................... 132
Figure 5-1 Above - Comparison between observed and APSIM model simulated
maize and cowpea yield distributed across three sites. Shapes identification:
∆ - Sandy Clay loam (cSaCL); O Fine sandy (fSa) and ◊ Sandy loam soil
(SaL). Filled shapes for each soil type represent residue treatments while
empty shapes did not receive any residue. Middle and bellow (A, B and C):
Overall comparison between observed and simulated maize and cowpea
yield (a), NUE (b) and WUE(c) distributed across tested sites for.
Continuous lines indicate a 1:1 fit and dotted lines indicate a best fit line
between observed and simulated. ............................................................................ 157
Figure 5-2 long term simulated maize and cowpea yield responses to residue
application and N-application levels (0, 23, 92 kg N/ha representing ND,
Page | xviii
LNA and HNA systems) in a high WHC (cSaCL) and prolonged wetness
(SaL) soils. No residue (R-); Residue applied (R+) ................................................ 159
Figure 5-3 Long term simulated maize and cowpea yield response to residue and N-
application levels (0, 23, 92 kg N/ha representing ND, LNA and HNA
systems) in low WHC (fSa) soil. No residues (R-), Residue applied (R+) ............. 161
Figure 5-4 Residue modified nitrogen driven NUE and WUE trade-off simulated
across maize and cowpea grown in contrasting farming environments of
central Mozambique. Overall means are presented for each level of N-
application namely: unfertilized (ND, 0 kg N/ha), Limited N-applications
(LNA, 23 kg N/ha) and high N-application (HNA, 92 kg N/ha). Residue
application levels are indicated as R- and R+ for no residue and residue
applied plots respectively. ....................................................................................... 162
Figure 5-5 Above, in the rainfall distribution graphs (A, B and C) for the period
1951-2014, seasons are categorized according to simulated maize yield
variations under high C/N residue application at three different soils. The
horizontal line indicates annual average rainfall over the simulated period.
Dark bars – indicate seasons were yield losses from residue application
were simulated; Red bars indicate seasons in which positive responses to
residue application are expected across N-application levels. In graphs A1
to C3, relative maize yield variations to residue application disaggregated
by soil and N-application levels over the simulated period. .................................... 164
Figure 5-6 Nitrogen and water stress levels in maize across residue application
levels for a high WHC environment in a normal and poor rainfall season. A
stress factor of 1 indicates full stress and 0 absence of stress. Stress indexes:
npho is the N-stress and wpho is water stress. ......................................................... 168
Page | xix
Abbreviations
APSIM Agricultural Production System Simulator
BNF Biological nitrogen fixation
CATPCA Categorical Principal Component Analysis
C/N Carbon to nitrogen ratio
CA Conservation agriculture
CMYP Continuous maize yield penalties
cSaCL Coarse sandy clay loam soil
CWU Crop water use
DAP Draft animal power
fSa Fine sandy soil
HNA High nitrogen application system
HTP Hired tractor power
INM Integrated Nutrient Management
ISFM Integrated Soil Fertility Management
LENI(S)A Low external nitrogen input sustainable agricultural systems
LENIA Low external nitrogen input agricultural systems
LNA Low nitrogen application system
LLR Land to labour productivity ratio
LSD Least significant difference
ND Unfertilized nitrogen deficient system
NNA Non-limiting nitrogen application
NO3-N Nitrate nitrogen
NUE Nitrogen use efficiency
NUpE Nitrogen uptake efficiency
NUtE Nitrogen utilization efficiency
PAWC Plant available water content
ppm Parts per million
R+ and R- With and without residue application
SaL Sandy loam soil
SIMLESA Sustainable Intensification of Maize-legume systems in Eastern and Southern Africa
SOC Soil organic carbon
WHC Water holding capacity
WUE Water use efficiency
YLR Yield to labour productivity ratio
Page | xx
Thesis outline
This study is about improving agricultural production and household incomes in nitrogen
constrained smallholders farming systems in Mozambique. This thesis document comprises
the following components (Figure 1):
Chapter I: Introduction
In this chapter, the general research background is set with a focus on resource use and
resource allocation for the rain fed N-deprived systems of Central Mozambique. The general
research background, the main question, objectives, hypothesis and general methodology are
presented. A short characterization of the environments where the trials were conducted, with
a focus on the biophysical aspects, is also presented.
Chapter 2: Literature Review
This chapter reviews the literature in reference to the key issues of: (1) fertilizer use in SSA;
(2) the effect of crop residues on yield, water and nutrient cycling in maize-legume cropping
systems; (3) legumes performance and potential contribution to improve N-availability
through biological N-fixation in Africa; and (4) how agricultural simulation models, such as
the Agricultural Production System Simulator (APSIM), can be used as analytical and
decision support tools to generate and disseminate relevant agricultural information in
Africa.
Chapter 3: Farmers diversity and opportunities to sustainable intensification
This chapter, reports on the findings of a household survey aimed at assessing farmers’
attitudes towards residues, fertilizers, manure and legume use, in their cropping systems.
Functional farmer’s typologies based on the combination of overall household livelihood
strategies, resource use behaviour and their perceived impact on local farming systems design
and the household ability to meet seasonal food security needs and income generation targets,
were built. These typologies were then used as the bases to formulate entry points for
sustainable intensification of maize-legume cropping systems. In the process, three main
technology adoption traps were identified, and their potential roles are also discussed.
Page | xxi
Chapter 4: Residue management in rain fed N-deprived systems
This chapter builds on the findings from chapter 3. The results from a multiple-site trial to
assess the short term responses from crop residue-fertilizer interactions and their impact on
yields and resource productivity (WUE, NUE, NUpE, NUtE) across crops and contrasting
farming environments are presented. Sites were selected based on data relating to plot
ownership collated during the survey. The focus of this chapter is in demonstration of
alternative allocation strategies of high C/N ratio residues and N-fertilizers that are more
likely to improve yields in the short term, and resource productivity for the rain fed N-
deprived systems.
Chapter 5: Pathways for sustainable residue management
The main focus of this chapter is on matching technological solutions to farmers’ needs in
terms of residue, legume and fertilizer management strategies in rain fed N-deprived
environments. A model based decision support framework, is presented and discussed as an
alternative toolkit to generate generalizable information that can be used to tailor residue
management strategies in the region. The framework presented in this chapter focus the
analysis of long term seasonal responses to residue application across crops, levels of N-
application and sites of contrasting soil water holding capacity. In this framework, the
understanding of seasonal variability effects on crop responses to residues is used as the key
to map the likelihood of positive or negative responses to residues for specific crops, sites and
levels of N-application.
Chapter 6: General discussion and conclusions
This chapter outlines and discusses the key findings from the study, the lessons learned, the
applicability of the results and the way forward in scaling out the results.
Page | xxii
Figure 1 Thesis outline layout, which includes chapters’ sequence and methodology used
Page | 23
CHAPTER 1
1. General Introduction
1.1 General Research Background
Over recent decades, the need to sustainably increase crop and land productivity in Sub-Saharan
Africa (SSA) has led to the widespread promotion of several agricultural technologies in Africa.
Conservation agriculture (CA) is one of the technologies that has received wide attention and has
been widely promoted throughout Africa under the assumption that it could help revert soil
degradation - erosion, mineral and organic matter decline, chemical in SSA agriculture (Giller et
al., 2011a; Ngwira et al., 2013; Thierfelder et al., 2013; Wall, 2007). However, the achievements
with CA in Africa are still far below those obtained in America, Asia, Europe and Australia
(Brouder and Gomez-Macpherson, 2014). In these regions, the adoption of CA has reached 70-
90% of the cropped land, and proved effective in slowing soil degradation while increasing crop
yields at the same time (D'Emden et al., 2006; Derpesch, 2008). However, attributing gains in
productivity solely to the adoption of CA principles is difficult, given the multiplicity of
associated investments made in those regions e.g. irrigation systems, mechanization, improved
seed and crop varieties, high use of inorganic fertilizers and chemical weed control.
It has been estimated that more than 50% of registered increases in agricultural production
achieved during the green revolution is attributable to high input uses, mechanization, irrigation
and enhanced crop response to inorganic fertilizers (Kelly, 2006; Roy et al., 2006). In SSA,
however, the practice of extensive rain fed agriculture (Alene and Coulibaly, 2009; Giller et al.,
2006), the recurrent poor access to inputs (Bationo et al., 2012a; Chianu et al., 2012), lower
mechanization levels and weak access to relevant agricultural information, are among the factors
that have hindered the realization of the continents agricultural potential. Despite these
constraints, in 2006 the Abuja declaration pledged to help increase fertilizer use in Africa from
8.3 to 50 kg/ha by 2015 (Kintché et al., 2015)1. However, to date there remain a number of
constraints to increased fertilizer use in Africa.
1 The declaration refers to gross nutrient amounts.
Page | 24
Despite the acknowledgement that, in the short term, “inorganic fertilizers are the best way to
feed Africa” (Gilbert, 2012), it is also true that smallholder farmers in Africa have been reluctant
to invest in inorganic fertilizers. Major constraints to fertilizer use in Africa are high fertilizer
costs mainly exacerbated by poor fertilizer markets, lack of institutional support and
inappropriate fertilizer packaging (Ade Freeman and Omiti, 2003; Chianu et al., 2012). Poor
perception about fertilizer benefits exacerbated by unreliable responses negatively affects
farmers decision to use inorganic fertilizers (Chianu and Tsujii, 2005). As a result, fertilizer
application rates in SSA are extremely low ca. 9-13 kg/ha (Tenkorang and Lowenberg-DeBoer,
2008). Therefore, within the context of poorly resourced smallholder farmers and the nitrogen
deprived farming systems of SSA, a more integrated and pragmatic resource use approach that
takes into account the socio-economic and biophysical diversity of smallholder farmers and the
resource pool at their disposal, appears to be a more feasible option to increase yields and
farmers’ food security prospects. In Africa, ‘integrated soil fertility management’ (ISFM) has
been presented as a viable and locally feasible alternative to improve nutrient availability
(Bationo et al., 2012b; Bationo and Waswa, 2011). The focus on a combined use of locally
available amendments and legume to increase nutrient availability links well with CA strategy in
Africa. Nevertheless, tailoring both ISFM and CA still poses a major challenge in Africa.
The CA strategy for Africa is built under the assumption that, minimum soil disturbance, crop
residue retention and legume integration in maize dominated systems offer a more affordable
option to sustainably increase soil fertility and yields in SSA. However, small amounts of plant
residues are left as ground cover due to low biomass production, high competition for the
residues mainly in livestock dominated areas (Erenstein, 2003; Erenstein et al., 2015;
Hauggaard-Nielsen et al., 2003; Valbuena et al., 2012). Poor legume performance in Africa
(Giller et al., 2009), is also a recurrent problem in Africa. Furthermore, the most widely available
and used residues in most smallholder farming systems are of high C/N ratio, and the
decomposition of such material is known to interact strongly with soil mineral N supply (Ros et
al., 2011). However, this relationship has been poorly studied, and integrated datasets reporting
on the combined effect of all CA principles are still scarce, particularly in SSA.
From 2011-2014 (SIMLESA Phase 1) to 2014-2019 (SIMLESA Phase 2), Mozambique has been
involved in an Australian Centre for International Agriculture Research (ACIAR) funded project
Page | 25
– The Sustainable Intensification of Maize-legume Cropping Systems in Eastern and Southern
Africa (SIMLESA). The SIMLESA project is being implemented across 5 countries, Ethiopia,
Kenya, Tanzania, Malawi and Mozambique. The intensification of maize-legume cropping
systems through the CA-ISFM platform is the focus of the SIMLESA intervention. However,
there is still a need to identify feasible options to sustainably intensify these maize-legume
cropping systems among resource poor smallholder farmers, considering that their farming
systems are diverse, complex and highly context bounded. Therefore, understanding the
interactions between farmers’ circumstances, their resource use behaviour and farming systems
design, is a critical step towards tailoring new technological packages to fit the local reality –
farmer groups and geographic context. In the meantime, is also true that little research has been
done on achieving an understanding of the SSA N-deprived smallholder farming systems those
dynamics in Central Mozambique.
This research, which is an integral part of the SIMLESA project in Mozambique is focused on:
1) identifying the best fit allocation strategies for limited resources such as high C/N ratio crop
residues and inorganic N-fertilizers, across crops and environments; 2) generating information
for a better understanding of the interactions between the widely used high C/N ratio residues
with N-availability in rain fed N-deprived systems; 3) understanding to what extent farmers
circumstances and their resource use behaviour affects local farming systems design and the
ability of those farmers to engage in new practices; and finally 4) because no change can be
achieved without a reliable and time effective decision support system, this study also explored
the capabilities of the Agricultural Production Systems sIMulator (APSIM) to simulate
alternative residue-fertilizer allocation strategies across crops and environments, in the hope that
the model can be used as part of a model assisted decision support framework on residue
management.
1.2. Research Questions, Objectives and Hypothesis
The overall research question of this study is: what are the benefits and trade-offs from
alternative crop residue management in the rain fed low external N-input maize-legume systems
in central Mozambique. The general hypothesis is that: ‘Best fit technological interventions need
to be found from an integrated systems approach that evaluates the benefits and trade-offs from
Page | 26
alternative resource management strategies and sustainable intensification pathways, in terms of
food production, risks and environmental outputs’.
General aim and specific objectives:
The aim of this research is to identify more productive and locally feasible sustainable residue
management practices capable of helping improve rainfall capture and nitrogen use efficiency,
and which are more likely to be adopted by resource poor smallholder farmers in rain fed low
external N-input agricultural (LENIA) systems. To achieve this goal, the following specific
research objectives were proposed:
Socio-economic and bio-physical characterization: To assess the performance of
existing practices of residue and nitrogen management, and to identify problems and
opportunities for improvements that are likely to be adopted by the poorly resourced
farmers of Central Mozambique;
Cropping systems research: To identify and test opportunities for better allocation of
crop residue and fertilizers that are more likely to provide short term benefits in terms of
improving crop yield, soil nitrogen, and water availability, in nitrogen constrained
environments;
Pathway to impact: To develop simple recommendations for better matching
technological solutions to smallholder farmers’ needs in terms of residue, legume and
fertilizer management strategies.
Hypothesis Tested
The objective of the research trials was to test the following specific hypotheses:
1) For the nutrient constrained smallholder farming systems of central Mozambique there
are household and site-specific management combinations of technologies e.g. residues,
and fertilizers that best fit farmers’ bio-physical and socioeconomic circumstances.
2) In wetter nitrogen constrained environments, investing the limited availability of high
C/N ratio stubble in improving the performance of the sole legume crop is more
beneficial to the system than using the residues for the maize crop.
Page | 27
3) In drier and variable climates, investing the limited availability of maize stubble in
improving the rainfall capture for the maize crop is likely to provide more benefits than
applying the residues to the legume crop.
4) For wetter soils, the beneficial effects of applying crop residues onto the legume crop are
likely to be observed on the following cereal crop.
5) The balance between the positive and negative effects of residue retention in maize-
legume systems (i.e. water balance, and the temporary nitrogen immobilization), can be
managed with nitrogen fertilisers.
1.3 Study area description
Central Mozambique is home for 58.6% of the country’s agricultural households distributed
across four provinces, Sofala, Manica, Tete and Zambezia. The region has one of the most
important agricultural breadbaskets, the Beira Agricultural Growth Corridor (BAGC) with ca.,
227.000 km2 covering Sofala, Manica and Tete provinces. The BAGC, has a total population of
4.8 million (19 persons/km2) and 10 million hectares of arable land, however, only 16.5% are
cultivated according to estimates from the last agricultural census (INE, 2011). The average
yields are around 1.2 t/ha for maize and less than 0.5 t/ha for most legume crops (FAOSTAT,
2014).
In 2011, with the approval of the countries Strategic Plan for the Development of the
Agricultural Sector (PEDSA)2 which points at the need to increase agricultural productivity
across the country’s agricultural breadbaskets as critical to improve the performance of the
agricultural sector, the BAGC have attracting several investments from the government and its
developmental partners, among them the World Bank, the Alliance for Green Revolution in
Africa (AGRA), the Food and Agriculture Organization (FAO), the International Fund for
Agricultural development (IFAD), the United States Agency for International Development
(USAID) and most recently the Australian Agency for International Development (AusAID)
through ACIAR agencies. This attention is in part due to: (1) the region’s rich agricultural
2 PEDSA – Plano Estratégico de Desenvolvimento do Sector Agrário. PESDA was approved in May 2011
and its main pillars are: (1) Increase agricultural productivity (ii) Facilitate access to market through the
improvement of infrastructure and development of value chain; (iii) Promote a sustainable use of natural
resource (iv) Strengthen local institutions – research and collaborators;
Page | 28
potential, has it hosts most of the highly productive agroecological regions of the country with an
expected production potential of more than 8.0 t/ha for maize and 3.0 t/ha for soybeans; (2) it
concentrates the highest extension of arable and cultivated land and (3) the existence of a
favourable support infra-structure mainly access to water, roads and transport to facilitate trade
within the regions and its surrounding provinces and neighbouring countries.
The vast mosaic of agroecological zones (AEZ) in central Mozambique, makes agriculture a
highly complex and diverse activity in the area. The agroecological diversity heavily influences
farming systems design and management strategies within and across regions. For example,
within the four provinces that constitute the central Mozambique agricultural development
region, at least 3 agro-ecologies can be identified in each central Mozambique province –
Manica (AEZ-R4, AEZ-R5, AEZ-R6, and AEZ-R10), Sofala (AEZ-R5, AEZ-R4, and AEZ-R6),
Tete (AEZ-R6, AEZ-R7, and AEZ-R10) and Zambezia (AEZ-R5, AEZ-R7, AEZ-R8, and AEZ-
R10) making farming highly context bounded. The intra and cross-agroecological variability that
characterizes the region can be seen in its climatic resource map as described in (Reddy, 1984).
1.3.1 Target Agroecological Zone (AEZ)
The study was conducted in Manica province in central eastern Mozambique in the medium
altitude agroecological zone (AEZ-R4). The AEZ-R4, falls within the plateau zone and are
characterized by two well-defined rainfall seasons: a wet season from November to late April
which receives 78-99% of the total annual rainfall, and a dry season that runs from May to
October. AEZ-R4 is characterized by a tropical dry climate and covers 80% of the Manica
province and part of the Sofala. The region is located in a medium planaltic zone with altitudes
between 200-800m. The annual rainfall is of approximately 834 mm, calculated over 63 years
between 1951 to 2014. The average minimum temperature in the region is 16.6°C, with an
average maximum of 32.5°C. The soils are predominately red ferralsols and luvisols (Gouveia
and Azevedo, 1954).
Page | 29
1.4 General Methodology
A combined methodology integrating socio-economic surveys, empirical research trials and bio-
physical modelling, were used to test the proposed hypotheses, as outlined below.
Hypothesis 1: For the nutrient constrained smallholder farming systems of central Mozambique
there are site and management specific combinations of technologies e.g. residues, legumes and
fertilizers, that best fit farmers’ bio-physical, market and socio-economic circumstances, which
will be tested using data generated from the following activities:
Socio-economic survey: The objective of this survey was to understand how farmers use
and manage crop residues, legumes and fertilizers, in their current cropping systems, and
how their attitudes towards resource use affects their ability to meet the household food
security and income generation targets. Survey data was also used to identify key
variables for the construction of new framework for farmer characterization that can help
aid the identification of scalable intensification pathways. To achieve this objective, a
random sample of 131 farming households were interviewed across two villages of
contrasting resource endowment – Matsinho in Vanduzi and Marera in Macate, both
located in the AEZ-R4. In addition, two focus group discussion were held in Marera with
29 farmers representing 7 farmer organizations from Macate, Zembe, Matsinho, Boavista
and Marera.
Participatory Crop Modelling (PCM) Workshops: These were used to complement
the household survey. The PCM workshops helped gather in-depth information on local
farming systems design, i.e., crops, sowing windows, cropping sequences, management
practices, yields, most common farming environments and their relative advantage. In
each meeting, resource allocation maps (RAM) were developed by farmers and used to
parameterize agent based typologies and also APSIM. APSIM was used to run technical
analysis of current cropping systems with an emphasis on plant densities and N-fertilizer
application. The simulations helped understand current cropping practices and identify
opportunities for improvements. The PCM’s also served as a platform to test APSIM
suitability to serve as a decision support tool for local research and agricultural extension
networks.
Page | 30
Hypotheses 2, 3 and 4: In wetter, nitrogen constrained environments, investing the limited
availability of high C/N residues on improving the performance of the sole legume crop is more
productive and beneficial to the system than applying those residues on the maize crop. In drier
and more variable climates, using the limited availability of maize stubble for improving rainfall
capture for the maize crop is likely to provide more benefits than applying these residues on the
legume crop. For wetter environments, the beneficial effects of applying crop residues on the
legume crop are likely to be observed on the following cereal crop. These hypotheses were tested
using the data generated from the following activities:
Cropping Systems Modelling: Modelling was used ex-ante to fine tune the treatments
for empirical research, and ex-post to: (1) explain the results from the empirical
experimentation; and (2) scale out the results from the study across seasonal conditions.
The Agricultural Production Systems Modelling (APSIM) version 7.7 through its
nitrogen (SOILN), RESIDUE and water (SOILWAT) modules, were used as the main
modelling platform.
Field Trials: a researcher managed trials were conducted in Matsinho, central-eastern
Manica province (AEZ-R4) over two seasons to generate GxExM information to identify
best fit residue management strategies for the rain fed N-deprived systems managed
under conservation agriculture. Responses where tested across two crop types, i.e., sole
maize and sole cowpea. Three soils (environments) of contrasting texture and water
holding capacity (WHC) commonly used in Central Mozambique were used. A red
coarse sandy clay loam (cSaCL) textured Ferralsol which is a typical upland soil and
most dominant in the region; a dark sandy loam soil (SaL) Gleysol, typical lowland soil
with prolonged wetness periods; and a dry fine sandy textured (fSa) Arenosol of low poor
water holding capacity. Nitrogen was applied in three levels: 0 kg N/ha representing an
unfertilized N-deficient system (ND), 23 kg N/ha representing a low N-application (LNA
system and 92 kg N/ha representing a high N-application system (HNA). Finally, high
C/N crop residue mulches were applied in two levels, 0 and 6 t/ha. (Table 1.1)
The N-application levels were selected based on two criteria. The 23 kg N/ha assumed for the
LNA systems is an approximation of fertilizer package in FAO input voucher program in
Mozambique which provides approximately 29 kg N/ha through a 50 kg bag of NPK 12:24:12
Page | 31
and a 50 kg bag urea (46 N) to be applied as basal and top-dress fertilizer. To facilitate
recommendations and align the figure with regional fertilizer which range between 18.3 and 24.1
kg N/ha among resource endowed smallholder farmers in our study area an average N-amount of
23.8 kg N/ha averaged across this three values was kept to 50 kg bag of urea assuming only a
top-dress application. The HNA was based on reported N-application amount needed to satisfy
N-immobilization demands during high carbon residue decomposition in SSA LENIA systems
which ranges between 80-100 kg N/ha (Homann-Kee Tui et al., 2015; Vanlauwe et al., 2014).
The amount was kept to 92 kg N/ha, i.e., 4 bags of urea (46 N). The 6 t/ha residue application
rate, was selected to mimic an environment with a high N-immobilization risk due to continuous
incorporation of carbon rich maize stubble and grasses characteristic of central Mozambique
LENIA systems. This value is close to the average amount of maize stubble produced in
Southern Africa if we assume a 3.2 t/ha at a 0.45 harvest index (Kintché et al., 2015). Ex-ante
simulations for Central Mozambique showed an 80-75% chance of achieving a 3 and 4.5 t/ha
yield in ND and LNA systems without residue application meaning that 4±5.5 t/ha maize stubble
are likely to be attained in the region using a maize hybrid. These figures are in line with data
presented in (Baudron et al., 2015c) who estimated an approximately 2428 ± 5890 kg maize
biomass year-1 farm-1 in central Mozambique.
The experiment was conducted as completely randomized block design in a 3x3x2x2 factorial
arrangement, laid in a 4x3 layout (Figure 1-1). All the plots received the same recommended
amount of phosphorous (P) and potassium (K) as NPK 12:24:12 at sowing and nitrogen dozes
were applied in a 50:50 ratios as basal and top-dress. Cowpea, was inoculated at sowing. Plots
were 12m x 6m in the first season and were split into two 6x4m plots in the second season to
accommodate a bioassay maize crop. The bioassay was used to measure the short term residual
effect of the first year treatments while in the other half of the plot the first season trial was
continued to measure cross seasonal effects. No fertilizer neither residues were applied to the
bioassay maize crop in season two and residues from the previous season where retained in the
plot.
Page | 32
1.4.1 Trial data analysis
Before analysis, trial data was checked for homogeneity of variance using the Bartlett’s test.
Because variances where homogeneous, analysis of variance was performed using the following
statistical models for yield and resource productivity response to residue management strategies in year
one and year two bioassay maize crop:
Crop response to residue management strategies (Main Factors: Crop types, N-application levels,
and sites of contrasting soil water holding capacity). Three way interactions where tested as
model bellow shows:
i. (YEAR 1) Response variable = Constant + Crop + R_level + N_level + Site + Crop*R_level
+ Crop*N_level + R_level*N_level + Crop*Site + R_level*Site + N_level*Site
+ Crop*R_level*N_level + Crop*R_level*Site + Crop*N_level*Site +
R_level*N_level*Site + Crop*R_level*N_level*Site
ii. (YEAR 2 - Bioassay crop) Residual responses = Constant + P_Crop + R_level + N_level +
Site + P_Crop *R_level + P_Crop*N_level + R_level*N_level + P_Crop *Site
+ R_level*Site + N_level*Site + P_Crop*R_level*N_level +
P_Crop*R_level*Site + P_Crop *N_level*Site + R_level*N_level*Site +
P_Crop *R_level*N_level*Site
iii. Random model: blocks + blocks*wplots + blocks*wplots*subplots
iv. Number of units: 264
NB: For year 1 statistical model, Crop is maize and cowpea grown in this specific season while in year 2
model, P_Crop represents the previous crop sown, i.e., the sequences maize-maize or cowpea-maize
which were used to measure residual treatment effect.
For each individual crop, cross-site responses to residues across N-fertilizer application levels
were tested for the variables where sites proved to have a significant effect. The model below was
used:
i. Response variable: Constant + R_level + N_level + Site + R_level*N_level +
R_level*Site + N_level*Site + R_level*N_level*Site
ii. Random model: blocks + blocks*wplots + blocks*wplots*subplots
iii. Number of units:132
Page | 33
Table 1-1 Treatment key with indication of tested factors disaggregated by site
Factors Treatment3
Sites
(Whole plots)
Crop
(Subplot)
R-level
(Sub-subplot 1)
N-level
(Sub-subplot 2)
SaL fSa
0 t/ha
0 kg N/ha Cwp0R0N
Cowpea
23 kg N/ha Cwp0R23N
cSaCL
92 kg N/ha Cwp0R92N
6 t/ha
0 kg N/ha Cwp6R0N
23 kg N/ha Cwp6R23N
92 kg N/ha Cwp6R92N
Maize
0 t/ha
0 kg N/ha Mz0R0N
23 kg N/ha Mz0R23N
92 kg N/ha Mz0R92N
6 t/ha
0 kg N/ha Mz6R0N
23 kg N/ha Mz6R23N
92 kg N/ha Mz6R92N
3 Treatment key: Crop * R-level * N-level * Site
Figure 1-1 Trial layout. For each site a layout was generated through the GenStat 16 trial
design tool. At site level, crop types was the whole plot, residue levels where the sub-plots and
N-level the sub-subplots
Page | 34
Hypothesis 5: The balance between the positive and negative effects of residue retention on
maize yield, i.e. grain water productivity (WUE), and the temporary nitrogen immobilization,
respectively, can be managed with the use of nitrogen fertilisers. The experimental results were
used to test the performance of APSIM to simulate the observed crop responses to residue and N
inputs across the farming environments. The calibrated APSIM model and quality historical
climate records were used to extend the analysis across seasonal conditions as a means to
generate relevant and generalizable information.
1.4.2 Trial site characterization
Three commonly used soil types of Central Mozambique were used. These three soils are found
across two distinct farming environments, namely: 1) the upland farming environments, mostly
dominated by the red ferralsols (cSaCL) and Arenosols (fSa); and 2) the lowland wet farming
environments locally called Baixas, which are characterized by prolonged wet periods and are
dominated by Gleysol. Samples for initial soil characterization where collected at the start of the
2013-14 cropping season to 1.20 m depth (Table 1-2) at intervals of 0.15 m for the first two
layers and 0.30 m for the following five layers.
Sampling for starting soil water content and plant water use determination was done at the same
depth intervals, at sowing and physiological maturity. Additional soil moisture measurements
were conducted fortnightly across both seasons, 2013-14 and 2014-15 between maize flowering
and physiological maturity. Measurements were made at 0-15 cm depth in three points per
treatment with the help of a MPKit-406 Soil Moisture Instant Reading Kit from ICT
International. Results showed that the sandy loam soil (SaL), registered the highest topsoil (0-
15cm) volumetric soil water content (VWC) in all sampled dates. The fine sandy soil (fSa)
registering the lowest VWC (Figure 1-2). VWC measurements, as presented in figure 1-1 were
also critical to show the current site variability on in crop soil moisture regimes, an indication of
water holding capacity.
Page | 35
Permanent wilting point (CL15) and field capacity (DUL) were determined opportunistically by
combining the APSIM soil characterization protocol described by Dalgliesh and Foale (1998)
and also Pala et al. (2007). Here, a pond experiment was established to measure soil water
holding capacity and strengthen the texture based method.
Pre-sowing and post-harvest soil nitrate (NO3-N) content down the soil profile (0-1.20 m) was
measured across treatments in both seasons. The colorimetric method based on Merckoquant
nitrate strips - Merck KGaA from Germany was used (Cossani et al., 2012). Results from post-
harvest NO3-N measurements (0-1.20 m) at the end of the first season, showed higher residual
NO3-N in the cSaCL and fSa (Figure 1-3). These figures are in line with data on starting NO3-N
starting which was higher on both soils. Nevertheless, a slight reduction in the overall NO3-N
was measured compared to the starting condition in all soils. In the N-rich cSaCL, residue
application reduced the amount of profile residual NO3-N in ND and HNA systems meaning that
residue application favoured N-extraction especially at higher N-application (HNA). In the drier
fSa, the highest residual NO3-N was measured in residue free maize treatments, meaning that the
crop was unable to extract most of the applied and mineralized nitrogen. Contrarily to maize,
post-harvest NO3-N in cowpea was lower, meaning that the crop managed to extract most of the
N-applied at except at HNA under residue in the SaL.
Table 1-2 Soil starting condition in the first season (2013-14) of trials
Red ferralsol (Red sandy clay loam – cSaCL, local name: Kupswuka)
Soil layer
pH
P(O) K SOC NO3-N SON BD CLL15 PAWC
(ppm) (ppm) (%) (ppm) (%) g/cm3 (%) (mm)
0-15 6.50 25.3 296 1.93 7.79 0.19 1.50 12.1 0.36
15-30 6.37 19.6 206 1.67 6.92 0.19 1.45 13.9 4.47
30-60 6.62 49.2 190 0.90 4.12 0.15 1.52 10.3 19.4
60-90 6.45 27.4 121 0.62 2.29 0.17 1.39 17.9 9.94
90-120 6.47 22 78.5 0.59 1.36 0.16 1.39 17.2 7.81
Page | 36
Gleysol (Dark sandy loam – SaL, local name: N’dongo wa Kupfipa)
Soil layer
pH
P(O) K SOC NO3-N SON BD CLL15 PAWC
(ppm) (ppm) (%) (ppm) (%) g/cm3 (%) (mm)
0-15 4.93 19.8 14.5 2.04 5.10 0.17 1.50 11.4 4.32
15-30 4.89 18.2 13.5 2.37 0.10 0.17 1.52 10.9 8.20
30-60 4.89 18.7 15.2 1.14 0.10 0.19 1.52 10.9 19.1
60-90 4.89 19.3 10.7 0.51 0.01 0.09 1.45 14.0 32.5
90-120 5.01 21.3 13.4 0.31 0.01 0.06 1.43 15.2 27.1
Arenosol (Fine sandy soil – fSa, local name: Djetcha)
Soil layer
pH
P(O) K SOC NO3-N SON BD CLL15 PAWC
(ppm) (ppm) (%) (ppm) (%) g/cm3 (%) (mm)
0-15 672 35.4 97.7 0.59 5.72 0.15 1.64 7.34 0.24
15-30 6.55 42.9 105 0.56 3.27 0.15 1.65 7.14 0.26
30-60 6.61 28.3 100 0.13 0.86 0.12 1.65 7.13 2.74
60-90 6.41 14.3 87.9 0.07 0.87 0.12 1.52 11.4 3.76
90-120 6.36 51.7 60 0.02 0.67 0.09 1.52 11.4 4.67
Page | 37
Figure 1-2 Topsoil volumetric soil water content (VWC) measured between flowering and physiological maturity across the three
tested soils in the 2013-14 and 2014-15 cropping seasons.
Page | 38
Figure 1-3 Post-harvest soil nitrate (NO3-N) distribution down the profile on maize (ABC)
and cowpea (DEF) treatments and sites. Shapes identification: ∆ - Sandy Clay loam (cSaCL);
O Fine sandy (fSa) and ◊ Sandy loam soil (SaL). For each soil and N-level (ND: 0 kg N/ha;
LNA: 23 kg N/ha and HNA: 92 kg N/ha), R represents residue treatments, i.e., with (R+) and
without (R-) residue application
Page | 39
CHAPER 2:
2. Literature Review
2.1 Fertilizer use and demand promotion in Sub-Saharan Africa
In Africa, optimizing resource use and productivity among resource poor smallholder farmers is
paramount to feed the continent’s and growing world population. This is an important step
towards improving yields and food security prospects in the low external input smallholder
farming systems where extensive and poorly managed agriculture has prevailed for years.
Increased fertilizer use is regarded as being responsible for almost 50% of yields increases
during the Green Revolution (Kelly, 2006; Roy et al., 2006). However, despite being recognised
as critical to improve crop yields specially in Africa (Gilbert, 2012) fertilizer use is still the
lowest in the world. In the developing world, especially Sub-Saharan Africa, access to affordable
fertilizer is still a major problem (Darko et al., 2014; Poulton et al., 2006). For example, in the
period 1992-2002, Africa accounted for only 2.87% of the global fertilizer use (FAOSTAT,
2012). Over the same period, Asia, America and Europe, accounted for 56.2%, 21.1% and 17.4%
of global fertilizer consumption, respectively.
Despite a predicted 2.6% annual increase in global fertilizer consumption since the economic
crisis in 2008, Africa still accounts for less than 3.0% of global fertilizer consumption (FAO,
2010; Heffer and Prud'homme, 2011). This is a marginal contribution when compared to Asia
and America that are expected to account for 69% and 19% of the predicted annual growth
(FAO, 2010). The high transport cost in many landlocked countries and regions, poor conditions
of roads, and low population density (people/km2) in rural areas create barriers to the
establishment of viable fertilizer markets (FAO, 2008). The price of agricultural commodities,
farmers risk perception of fertilizer profitability, farmers’ liquidity and the lack of appropriate
fertilizer policies to regulate distribution, mainly among smallholder farmers in remote areas
(Kelly, 2006; Morris et al., 2007), are factors limiting fertilizer access and use. Moreover, lower
fertilizer returns especially in the rain fed smallholder farming systems of SSA where climate
variability has undermined fertilizer response is also a critical adoption factor (Darko et al.,
2014). As a result, only 9 kg/ha of fertilizer are used, a figure 82% less that the 50 kg inorganic
fertilizer use target set for 2015 in Arusha declaration in 2006 (Lambrecht et al., 2014).
Page | 40
In recent years, several attempts to increase inorganic fertilizer use in Africa have been put in
place. Fertilizer promotion initiatives, e.g., the Agro-dealers Development and Support Program
(ADP/ASP) are being implemented in Malawi, Kenya, Tanzania, Rwanda, Ghana and
Mozambique (IFDC, 2012; Minot and Benson, 2009; Poulton et al., 2006). In these countries,
subsidized input vouchers were distributed to smallholders farmers, allowing them to purchase
fertilizer at almost half the market price (IFDC, 2012; Minot and Benson, 2009). However, these
initiatives have come under criticism mainly because of its failure to stimulate independent
demand from smallholder farmers by creating dependence on government handouts (Jayne et al.,
2013). Nevertheless, this cannot be generalized since there is little data to back the claim. In
Mozambican the program has succeeded in helping establish a network of agro-dealers. It is true
however, that in many cases, the vouchers failed to reach the designated beneficiaries, i.e., the
poorest farmer – creating the “elite capture” which is when farmers who have the means to pay
are handed the vouchers in detriment of the one that really need them due to their poor purchase
power (Crawford et al., 2006; Jayne et al., 2013; Minot and Benson, 2009). Moreover, the
voucher programs are not accompanied by proper research and extension programs aimed at
educating farmers on associated practices that could help enhance fertilizer productivity
(Crawford et al., 2006).
2.1.1 Tailoring fertilizer recommendations to SSA farming environments
Despite the low fertilizer use in Africa it is well agreed that, in the short term, inorganic
fertilizers are the best way to improve yields in Africa (Gilbert, 2012). Therefore, extensive
research to optimize fertilizer use and response in SSA farming systems has been conducted in
recent years (Ade Freeman and Omiti, 2003; Kintché et al., 2015; Tabo et al., 2011). Here,
evaluating crop responses has been the main focus.
Several approaches to promote fertilizer use in SSA agro-environments have been hypothesized
in results of ongoing fertilizer use research. Among them, there is the development and
dissemination of broad and site specific fertilizer recommendations (Bationo et al., 2012a;
Bationo et al., 2012b), micro-dosing (Ncube et al., 2007; Tabo et al., 2007; Twomlow et al.,
2010) and most importantly integrated nutrient management (INM) strategies. In the INM
Page | 41
platform, integrated soil fertility management (ISFM) in systems managed under conservation
agriculture has been the major flagship in Africa (Bationo and Waswa, 2011; Mafongoya et al.,
2006).
2.1.2 Optimized fertilizer recommendations and Micro-dosing
The development and dissemination of broad and site specific fertilizer recommendations has
been pointed as a critical step towards improving fertilizer allocation and productivity in Africa.
The underlining principle behind this approach is that knowing African soils is key to improve
fertilizer use and allocation strategies (Bationo et al., 2012a). This goal has been pursued through
initiatives such as the African Soil Information Service (AfSIS) led by the Tropical Soil Biology
and Fertility Institute (TSBF) (Hengl et al., 2015; Towett et al., 2015) and the OFRA4 project led
by CABI International and the Nebraska Lincoln University and covers Kenya, Uganda,
Tanzania, Rwanda, Zambia, Malawi, Mozambique, Ghana, Mali, Burkina-Faso, Niger, Ethiopia
and Nigeria. Both interventions have invested considerable efforts in improving soil
characterization across SSA. Improved soil characterization is key for the development of better
fertilizer recommendation. Despite all efforts, current fertilizer recommendation in Africa are
still unattractive to smallholder farmers who cannot pay for fertilizers. This has been one of the
major criticism of optimal fertilizer recommendations in Africa, since fertilizer use levels among
African smallholder farmers are low. In response to SSA lower fertilizer use levels, alternative
solutions have been developed to overturn smallholder farmers’ inability to invest in the
recommended fertilizer amounts. Micro-dosing is one of those alternatives.
However, the successful implementation of micro-dosing is strongly linked to the existence of
reliable site specific fertilizer recommendations, which is not the case in Africa. Until now,
broad fertilizer recommendations based on the supply of optimal crop requirement ranges under
the assumption that farmers have the capacity to supply such amounts, are promoted. As for
micro-dosing, positive responses were reported across Africa. In Zimbabwe, 30-50% increases in
maize grain yield were measured from the application of 17 kg N/ha (Twomlow et al., 2010). A
42% and 55% grain yield increase in the Sahel was reported after the application of 0.3g N per
4 Optimizing Fertilizer Recommendations in Africa (OFRA)
Page | 42
plant hill in the form of NPK 16:16:16, i.e., 3 and 10 kg N/ha for sorghum and pearl millet after
three years (Aune and Bationo, 2008). Despite proved benefits results from these studies like
many nitrogen nutrition trials conducted to date in Africa and across the world, are highly
context bounded, therefore not generalizable (Sadras and Lemaire, 2014).
Despite the acknowledgement that site-specific fertilizer recommendations need to account for
site variability, little information is yet available on this aspect, as intra-agroecological zone
variability has been poorly studied. As presented in Table 2-1, fertilizer use and its relationship
with crop productivity in Africa has been widely studied. Factors hindering adoption and ways to
improve both fertilizer use and soil fertility management among SSA smallholder farmers, have
been widely theorized. However, there is still an inability to translate these findings into
pragmatic measures at farmer level.
Despite it being true that there has been considerable increase in the average use of fertilizers in
Africa it is also true that the rates are likely to vary across locations and farming environments,
within and across agro-ecologies. Therefore, optimizing inorganic fertilizer allocation in SSA
smallholder farming systems is as critical as access. This involves selecting the right fertilizer
allocation between crops and sites, choosing the right source of nutrients, the right timing of
application and the right amount – the 4R’s Principle to optimize fertilizer response (Ciampitti
and Vyn, 2014; Kaiser and Rubin, 2013; Ross R. Bender, 2013). Optimizing fertilizer allocations
and returns by advocating ‘best fit’ allocations of available fertilizer amounts in Africa is
important to improve nutrient productivity and yields in SSA diversified smallholders cropping
systems, where the main issue for farmers is finding the best ways to produce enough yield and
income with a reasonable amount of work and resources (Vanclay, 2004b) and resources.
2.1.3 Integrated Nutrient Management (INM)
In recent years, the concept of Integrated Nutrient Management (INM) practices have been
widely promoted in Africa (Bationo and Waswa, 2011; Mafongoya et al., 2006). Under INM,
locally tailored soil fertility management strategies focused in the combined use of inorganic
fertilizers with locally available amendments (manure, crop residues) and improved agronomical
Page | 43
practices such as, residue retention, legume integration and adequate soil and water conservation
measures. Integrated soil fertility management (ISFM) and conservation agriculture (CA) have
been the core of INM.
INM has been presented as a more farmer friendly technology since inorganic fertilizer use in the
practice can be kept at minimum and complemented by local amendments. However, like other
technologies promoted to date, INM is also knowledge intensive. Moreover, INM needs to be
tailored to meet contrasting agroecological and socioeconomic contexts. The fact that INM is
context bounded and knowledge intensive is among the factors that undermine its adoption.
Nevertheless, INM is an affordable and environmentally friendly practice with high potential to
improve soil fertility and raise crop yields in Africa where average maize yields are around 1.7
t/ha (1.2 t/ha in Mozambique), against 8.6 t/ha in developed countries (FAOSTAT, 2014).
Despite its potential to improve soil fertility and increase crop yield, INM benefits will not come
as easy as it is thought. INM needs to be part of an integrated farming systems designs where
optimizing resource allocation and responses for improved crop yield are the core of crop
husbandry. Here, alternative resource management strategies such as improved crop residue and
fertilizer allocation across crops and farming environments through GxExM analysis.
Nevertheless, a better understanding of residue-soil nutrient dynamics (Erenstein, 2003) within
existing cropping systems are yet to be tested.
Page | 44
Table 2-1 Main Lessons from Fertilizer Research and Promotion in SSA in Recent Years
Source Key message (Findings)
Ade Freeman and
Omiti (2003)
Understanding smallholder farmers (SHF) behaviour regarding fertilizer is key to fostering adoption
Bationo et al.,
(2012a)
Improving soil classification and availability of soil information can help improve fertilizer recommendations and nutrient
allocation
Bationo et al.
(2012b)
Tailoring fertilizer recommendations to the social and biophysical contexts of smallholder farmers are important if
fertilizer efficiency is to be achieved
Bationo and Waswa
(2011)
Crop productivity can be improved in Africa if widespread adoption of ISFM is achieved. However socio-economic and
biophysical factors still hinder ISFM adoption in Africa
Bekunda et al.
(2010)
Nutrient restoration in depleted SSA soils can be achieved through several cheap technologies, including ISFM and
biological nitrogen fixation (BNF). However, a proper policy environment is required
Chianu et al. (2012) Balanced fertilizer applications, the development of improved input-output markets, improved crop management and use
of nutrient budgets to demonstrate fertilizer profitability, is required to guarantee that SHF adopt fertilizers
Chianu and Tsujii
(2005)
Farmers decisions to adopt or not adopt the use of inorganic fertilizer is determined by their education level and knowledge
about fertilizers, and their perceptions of fertilizer impact on crop yield and profitability.
Enyong et al. (1999) Farmers’ implementation of soil fertility enhancing technologies (SFETs) is influenced by access to information and the
existence of a favourable policy environment.
Gilbert (2012) The key for tackling hunger in Africa is through enriching its soils. However, this cannot be achieved by just investing in
fertilizers. More locally tailored technologies, including the use of appropriate legumes for BNF are also important.
Mafongoya et al.
(2006)
ISFM through the combined use of inorganic fertilizers, legumes, animal manure and agroforestry, are the way to increase
and maintain soil fertility in SSA.
Page | 45
2.2 Fertilizer, Residues and Legumes interaction in SSA Conservation Agriculture
2.2.1 Crop residue contribution to yield, water and N-availability
In recent years, extensive research has been conducted to assess the complex interactions
between crop residues, fertilizers and tillage on yield, soil water availability and soil nutrient
turnover (Gangwar et al., 2006; Hoyle and Murphy, 2011; Kumar and Goh, 2002; Malhi and
Lemke, 2007; Mazzoncini et al., 2011; Sidhu and Beri, 1989). Nevertheless, contrasting results
have been found.
In India, results from a 9 year experiment designed to assess the impacts of incorporating 4 t/ha
of chopped wheat and rice residues on the yield of maize, wheat and rice crops grown at 0, 40,
80 and 120 kg N/ha (Sidhu and Beri, 1989), showed no significant effects on maize yield in the
first year. Nevertheless, significant yield changes were measured after the second year.
Significant increases in Total-N, 460 – 672 mg/kg were measured when compared to the initial
soil status. Data from the same trial also showed that, the incorporation of wheat residues
without fertilizer application in wheat plots, resulted in a significant yield decline in the first four
seasons compared to the fertilized plots. In this systems, no enough nitrogen was available to
satisfy the N-immobilization demands from the decomposition of incorporated residues.
Therefore, negative yields responses where measured. Negative returns from carbon rich
residues incorporation were also reported by Baudron et al. (2015a) with N-applications below
100 kg N/ha after 3 t/ha of maize stubble were applied to cotton.
The negative yield responses were also measured in unfertilized fields with high C/N residues
mulches (Cheshire et al., 1999; Turmel et al., 2015). These negative yield responses were
attributed to poor N-availability to plants due to N-immobilization. However, yield loss due to N
immobilization in residue applied fields can be minimized through increased high N-application.
Crop sequences and residue management were also found to significantly affect crop response to
residue application. Bakht et al. (2009), found that residue incorporation immediately after
harvest led to a 1.31 times increase in wheat yield. Legume-cereal sequencing (mung bean-wheat
rotations) led to the production of 2.09 and 2.16 times more wheat stover and grain yield when
compared to maize-wheat sequences. These results corroborate the findings of earlier studies
(Aggarwal et al., 1997; Gangwar et al., 2006; Samra et al., 2003; Sidhu and Beri, 1989; Wang et
Page | 46
al., 2010b), where positive impacts from high C/N residue application on yield, soil water
availability, soil biota and N-uptake were measured under adequate fertilizer-tillage-crop
sequencing management. It is important however, to take into consideration that most of these
studies were conducted in the humid tropics, and that the levels of N-fertilizer applied in the
residue-fertilizer interaction were rather high, i.e., between 80 to 300 kg N/ha in some studies.
This figures, differ from most of the environments where residue retention is promoted in Africa.
2.2.2 Yield penalties from residue retention in continuous cereal systems
Most positive responses to high C/N ratio residue use in conservation agriculture systems
reported to date comes from research dataset generated under considerable high inorganic N-
application (Kushwaha et al., 2000; Thierfelder et al., 2015; Vanlauwe et al., 2014; Wang et al.,
2011), which contrast SSA smallholder farming systems. Nevertheless, in high biomass
production systems like the corn-belt in the USA, a growing interest in understanding the trade-
offs between crop residue retention and removal have been reported in an attempt to minimize
yield penalties from residue incorporation in continuous maize systems (Luce et al., 2015;
Morrison, 2013).
An alternative to improve nutrient availability in low N systems managed under conservation
agriculture in Africa, is through the addition of N-rich legume residues (Giller et al., 2009). The
promotion of legume-cereal cropping sequences is another alternative (Luce et al., 2015).
Mazzoncini et al. (2011), found that the incorporation of N-rich legume residues grown as cover
crops, is able to contribute up to 430 kg/ha of extra soil organic carbon (SOC) and 20 kg N/ha a
year. Kumar and Goh (2002), measured significant improvements in yield and nitrogen uptake in
wheat grown after the incorporation of white clover and pea residues, compared to wheat grown
under wheat and ryegrass residues. These results strengthen the fact that responses from residue
application are highly dependent on residue type, amount, N-application levels, crop sequence
and the overall residue management strategy.
A recent simulation study based on data from the DayCent model, showed that up to 70% of
maize stover could be removed for biofuel production without significant losses in SOC (Liska et
al., 2014). Nevertheless, the 70% removal cup was contested (Robertson et al., 2014). More
recently, Wu et al. (2015) showed that 50% residue removal could be a reasonable cup for
Page | 47
maintaining SOC and soil fertility levels during the first years in these high biomass production
systems.
In the rain fed and N-deprived cropping systems of SSA managed under conservation
agriculture, smarter residue management strategies still need to be tested. Here, the emerging
question is on the best allocation of the widely available high C/N ratio residues, in both
livestock dominated and non-livestock dominated areas. Cross-site and crop-specific allocations
strategies which are key to optimize residue use have been little studied in Africa. Residue
allocation strategies and its interactive relation with crop type and inorganic fertilizer levels in
the SSA smallholder farming systems is an important area of attention, especially considering the
fact that crop residues can either help offset the nutrient cycling benefits in the system or
improve it if proper fertilizer compensation is made.
2.2.3 Competing claims on residue use: from mixed crop-livestock systems to non-cattle
dominated systems
Crop residues are an integral component of the conservation farming platform in Africa (Wall,
2007). If well managed, crop residue have the potential to improve soil microbial pool, thereby
influencing the nutrient cycling rate (Malhi and Lemke, 2007; Yadvinder et al., 2005). In the
semi-arid tropics where the soils are characteristically low in soil organic matter, crop residues
retention is crucial to compensate for physical biological degradation. Nevertheless, more has to
be done to identify under what circumstances positive and negative responses to residue
application are likely to occur.
In most mixed crop-livestock systems of SSA, residue use as feedstock is understood to limit
residue retention in the field (Erenstein, 2002). In this context, finding the best balance between
crop residue retention and removal for animal feedstock is critical to aid farmer’s decision
making process on best-fit residue use (Erenstein et al., 2015; Homann-Kee Tui et al., 2015).
Nevertheless, it also true that, in non-livestock based systems, residue retention is not an issue.
Here, analysing the potential trade-offs e.g., economic and environmental benefits associated
with the best-bet options for residue use – across contrasting crops, soil types, agro-environments
and rainfall season is crucial but stills lacking in Africa.
Page | 48
2.3 Legumes integration in SSA smallholder systems managed under CA
2.3.1 Legume performance in Africa
Almost 27 million hectares of land are planted to legumes in SSA every year (Tsedeke et al.,
2012), of which 11 million ha are concentrated in Western Africa, with 5 million being
distributed between Eastern and Southern Africa (Lupwayi, et al., 2011). Cowpea, Bambara
groundnut and common beans in particular, have received much attention due to their wide
adaptation in the diverse SSA farming environments (Dakora and Keya, 1997; Danso, 1984;
Lupwayi, 2011). Over recent decades, the need to improve productivity in Africa has led to
increased interest in the potential contribution of legume crops to the soil nitrogen pool (Asiwe,
2009; Jeranyama, 2007; Siame et al., 1998) through biological nitrogen fixation (BNF). Due to
this interest, an increasing range of grain legumes and green manure cover crops have attracted
the interest of SSA researchers and smallholder farmers. However, legume production potential
in African farming systems has not yet been fully realized (Sprent et al., 2010). This is in
contrast to the success of legume adoption in Latin America, Asia and Australia where legumes
are reported to perform well (Peoples et al., 2009) in terms of yield and nitrogen fixation.
Average yields between 0.45 and 0.6 t/ha for cowpea and common bean have been reported in
SSA (Chianu et al., 2011; Lupwayi, 2011; Tsedeke et al., 2012). Poor natural soil fertility,
characterized by low soil nitrogen and phosphorous content (Giller and Cadisch, 1995; Sanginga,
2003) in the dried semi-arid regions (Dakora and Keya, 1997) is undermining legume
performance. Drought in particular is of much importance. Drought and high temperature in the
semi-arid areas of SSA are threats to legume productivity, as they significantly reduce legumes
yields, biomass production and most of all, their N-fixation (Leport et al., 2006; Sangakkara et
al., 2002).
Despite the importance of legume-cereal rotations for the success of CA in SSA, most of the
legumes are grown as intercrops (Lupwayi, 2011), with a relatively small area being dedicated to
sole legume cropping. In West Africa. for example, approximately 90% of cowpea is grown as
an intercrop to sorghum and pearl millet. Despite the focus on intercropping, monoculture crops
have been proven to produce up to 38% higher yield and 35% more biomass compared to
intercrops (Dakora and Keya, 1997). Not less important is relay cropping that has been practiced
Page | 49
in the sub-humid and humid tropics of Africa (Awonaike et al., 1990). Here bimodal rainfall
patterns, e.g. in Guinean Savannah, offer the possibility of double cropping legumes. However,
little information is available on legume intensification systems in Africa.
Nonetheless, despite the on-going discussion over intercrops and sole cropping, increasing
research into the identification of adequate cropping sequences in legume-cereal cropping
systems has been gaining a momentum in Africa (Adjei-Nsiah, 2012; Bado et al., 2006; J. M.
Vesterager, 2007; Jeranyama, 2007; Sun et al., 2010). Maize yields of up to 7.0 t/ha was
measured in a pigeon pea-maize sequence, compared to 2.0 t/ha in a continuous maize systems in
Ghana (Adjei-Nsiah, 2012). Umeh and Mbah (2010) reported a significant yield increase in a
cassava crop harvested 8 months after soybean.
Legume-cereal and legume-tuber sequences are of much importance for SSA N-deprived
systems. In these systems the production and retention of N-rich residues provides a quick
alternative to improve N-availability. Therefore, considering the fact that most of the residual N
will be available for uptake by the second crop considerable investment needs to be made to
improve legume productivity (Derpesch, 2008). The current challenge in most SSA farming
environments is to establish which farming system design will provide farmers with better
legume yields and biomass production. Until now, intercrops and one cycle legume rotations
have not managed to produce enough N-rich biomass and/or fix the amounts of N achieved in
other parts of the world.
In West Africa, specifically, legume trees and shrubs have been alternatively used as N-sources
in legume-cereal and mixed crop-livestock systems. Alley cropping, planting hedgerows and
relay green cover crops using Acacia, Sesbania, Albizia, Leucaena and Gliricidia genus, have
been reported to fix between 36-581 kg N/ha year in trials conducted in Senegal, Nigeria and
Tanzania (Dakora and Keya, 1997; Gathumbi et al., 2002; Giller and Cadisch, 1995; Oyun,
2007). In these systems, leaf pruning from woody species and residues retained from relay green
cover crops have been successfully used as N-sources. Tree legumes have a double advantage,
especially in mixed crop-livestock systems where they can be used as N-sources and also as
stock feed, thereby reducing the pressure on field crop residues. Another important use of tree
legumes is for the control of erosion.
Page | 50
With regards to legume shrubs, they have been widely used as relay green manure cover crops,
which can later be incorporated as in-situ mulching. Crotalaria, Mucuna and Tephrosia have
been of great importance in these systems, due to their ability to fix nitrogen while providing
extra soil cover and weed suppressing effects (Odhiambo et al., 2010). However, N2-fixation
from green manure cover crops requires adequate in-field management, as their contribution to
N2-fixation is highly influenced by the cropping system and cultivation practices (Sanginga,
2003). Sanginga (2003) found that N2-fixation in Mucuna pruriens and Lablab purpureus
responded differently when used as live mulches when compared to in-situ mulching. Higher N-
fixation was achieved with in-situ mulching. In addition, Mucuna has been reported to fix
between 133 to 188 kg N/ha in N-fertilized and rhizobia inoculated plots.
Despite all the measured impacts from legume incorporation into SSA N-deprived farming
systems, a lot still need to be done, mainly in designing legume favourable cropping systems,
management and resource allocation strategies. This will help increase the potential contribution
from legumes (Table 2-2). However, independent of the achievements to date, it is agreed that,
(1) legumes in Africa are still an underestimated crop and have not yet achieved their full
production and nitrogen fixation potential. This is in part due to the continent’s focus on maize
which is the main dominant staple crop in Africa (Giller et al., 2011b); (2) improving legume
performance, especially in relation to biomass production in SSA smallholder cropping systems,
is of considerable importance if higher contributions to BNF are to be achieved; and (3)
improving management conditions and cropping systems design are key tools for improving
legume performance. Only good performing legumes will fix enough biological N and add
substantial N-rich residues into the systems.
Page | 51
Table 2-2 Current findings on legume contributions to nitrogen fixation in African cropping systems
Citation
Agro-ecological
Condition
Legume
used
N2-Fixation
(Kg/ha year)
Systems Comments
Adjei-Nsiah et al.
(2008)
Forest/savannah
transitional zone
(1271 mm)
Cowpea 32-44 Farmers evaluation of productivity
in cowpea-maize crop sequences
Indeterminate varieties produced 2.6
t/ha stover while determinate varieties
produced, on average, 1.2-1.3 t/ha;
Adjei-Nsiah
(2012)
Semi-deciduous
forest zone
(1433 mm)
Cowpea,
Pigeon pea
Groundnuts
64-160
N-return and yield in legume-
maize-cassava cropping sequences
Total N-returned to the system through
residues after harvest in one season in a
crop sequence that included maize and
cassava.
Yusuf et al. (2009) Guineas Savannah
(1038 mm)
Cowpea and
Soybean
61-83 Rotational effect of legumes and N-
fertilization on maize yield and
nitrogen fixation
Maize after legume yielded better than
continuous maize. High N-uptake in the
soybean-maize sequence was measured.
N agronomic efficiency (AEN) was
affected by N rate and rotation
Somado et al.
(2006)
Humid forest
(green house)
Crotalaria
micans
29-116 P-response effect on yield and N-
fixation
Legume N2-fixation increased fourfold
compared with the unfertilized legume.
Legumes produced less than 1- ton
biomass without P application
Siame et al. (1998) Kasama, Zambia
1286.5mm
Common
beans
- Intercropping and N-response At higher N-levels increased vegetative
growth of the legume decreased maize
yield by 30%, with some competition
noticed at lower N-levels
Jeranyama (2007) Sub-humid
(800-900mm)
Groundnuts 16-40 Rotation effect Maize yields increased by 0.7 t/ha
compared to continuous maize
Sanginga (2003) Northern Guinea
Savannah
Soybean 38-126 BNF in legume based cropping
systems
Soil accumulated N ranged between 8-
47 kg/ha, depending on the soybean
cultivar, with 10-24 kg/ha residual N in
the soybean-maize rotation
Page | 52
2.3.2 Potential contribution from legume crops in improving N-fixation
On a global scale, it has been estimated that chemical fertilizers only account for 25% of
total-N2 in the soil, with atmospheric biological nitrogen fixation (BNF) from legumes
contributing with almost 60% (Zahran, 1999). This represents a total annual input of
approximately 139 to 170 million tons of N2-biologically fixed, of which 25-30% (34-45
million tons of N) being in cropped land and 30% (45 million tons of N) being in permanent
grazing land (Peoples et al., 1995). Therefore, enhancing N-fixation is important, especially
for fertilizer constrained systems in the developing world, where soil quality depletion due to
continuous nutrient mining and poor crop management practices, are common. As in the
developing world smallholder farmers are less likely to invest in inorganic fertilizers for
staple food crop production (Serraj, 2005), cropping systems that advocate proper legume
incorporation would provide a long term alternative for improving soil fertility and yields.
However, for legumes to realize their nitrogen fixing potential, proper cropping
arrangements, sequences and varieties, need to be in place.
In recent decades, nitrogen fixation through symbiotic association between leguminous plants
and rhizobia has been widely studied (Di Ciocco et al., 2008; Giller and Cadisch, 1995;
Herridge et al., 2008; Peoples et al., 2009; Salvagiotti et al., 2008). Determining the
partitioning between plant parts of nitrogen fixed from the atmosphere (Ndfa) through 15N-
labelling and natural abundance methods (Peoples et al., 2009), the identification of the best
association between rhizobia species and legume crops (Awonaike et al., 1990), and the
performance of inoculants (Laditi et al., 2012), are just some of the key areas of progress in
legume research that can be used for the benefit of N-deficient systems. However, most of the
studies conducted to date have been in highly productive environments in the humid tropics
and temperate regions. Less information is available on BNF, especially in the SSA cropping
systems where underperforming legumes are grown in multi-nutrient depleted conditions,
e.g., N. P and K have yet to deliver the advantages shown in other regions, e.g., Latin
America, Asia and Australia.
Despite the success experienced in Latin America, Asia and Australia, BNF does not provide
straightforward benefits. Negative and positive N-balances have been reported in legume-
cereal cropping systems. N-balances have ranged from -132 to + 135 kg N/ha in soybean,
groundnut, chickpea, lentil, pea and lupin crops (Peoples et al., 1995). Several factors have
Page | 53
been cited as contributing to the negative N balance. The fact that most of the N2-biologically
fixed by legumes is stored in the seeds which are primary exports from the systems, is one of
the factors. Di Ciocco et al., (2002), found that the negative N-balances have been mainly
associated with high yielding legume varieties, e.g., soybean with high harvest index. With
such varieties, the N-harvested from seed is higher than the N2-fixed in vegetative parts,
leaving less N to be returned to the soil. In addition, the removal of crop residues from the
system (Peoples et al., 1995; Stern, 1993) is also a contributing factor for the negative N
balance. Adding to the discussion, Austin et al., (2006) argued that the highly fertile soils and
high input cropping systems where most of the legumes are grown in the humid tropics, were
acting against nitrogen fixation by legumes. This is because the high fertility of the soils and
the use of fertilisers, has a negative impact on the nodulation capacity of legumes (Austin et
al., 2006), thereby reducing N2-fixation.
The relationship between nitrogen fertilization and BNF is an important point of attention
from a management perspective, especially in SSA where most of the soils are known to be
N-deficient and require a certain level of nitrogen application to stimulate BNF. However,
rather than soil starting nitrogen levels, there are other potential constraints to nitrogen
fixation. In Africa, for instance, crop varieties and cropping systems are among the most
important factors affecting legume nodulation and N fixation capacity.
The selection of appropriate legume species, varieties and management options is also
important. Growing environments also play a major issue on BNF. In high rainfall and
bimodal environments for example, legume intensification systems and better resource
allocation strategies between cereals and legumes, can help improve N-fixation and
availability in the system. The same would be less likely to happen in unimodal regions
except if short cycled varieties are used to allow double sowing. Nevertheless, the complexity
of SSA smallholder cropping systems and the poor investment in agricultural research and
development makes the discussion about legume intensification systems in unimodal rainfall
regions such as central Mozambique rather difficult and mundane. Therefore, investing in
new approaches to test and convey relevant agricultural information to improve decision
making processes at farm and regional level is crucial in Africa. In that quest, agricultural
simulation models, can be an important part of the solution in SSA.
Page | 54
2.4 Using agricultural simulation models as decision support tools in SSA
In recent decades, with the advent of farming systems research, (Bouma and Jones, 2001;
Davidson, 1987; Jordan et al., 1997; Maxwell, 1986), the modelling of agronomic processes
has played an important role in helping researchers generate new virtual information that is
allowing them to integrate and understand the interactions between crops, soils, climate
processes and management practices. According to Mathews and Stephens (2002), the
integration of simulation modelling in research programs has helped to: (1) identify gaps in
existing knowledge; (2) generate and test hypotheses, and design of experiments; (3) select
the most influential parameters in a system (Seligman, 1990); and (4) bring researchers,
experiments and farmers together to solve problems (Carberry et al., 2009; Whitbread et al.,
2010).
In developed countries where simulation models are widely used, they have helped improve
the understanding of crop eco-physiological responses and allowed researchers to answer
basic questions, such as identifying the most suitable and remunerative cropping systems and
patterns for different climatic and socioeconomic conditions (Cox et al., 2010; Meinke et al.,
2001; Rodriguez et al., 2011). The use of crop and biophysical models capable of generating
ex-ante information through the simulation of real world responses to factors such as the
climate, soil and management (Bouma and Jones, 2001), has allowed researchers to provide
farmers with timely and relevant advice that has helped shape local farming practices. This
advice has significantly improved farmers’ ability to manage risk.
Despite the attention that agricultural simulation models experienced in developed countries,
they had been poorly integrated in SSA agricultural research. Several factors contribute to
this. The reliance on reductionist processes, such as the conventional agricultural research
model that views fields, plants and animals as separate and non-related factors within a farm
unit or the household (Ikerd, 1993); the lack of training and technical capacity to gather
quality data to test and validate crop models (Bouma and Jones, 2001); and above all, the lack
of faith in models’ predictive capacity (Mathews and Stephens, 2002), have hindered the use
of models in SSA. However, in SSA where expenditure on agricultural research and
development is low (Beintema and Stads, 2008), the use of simulation models that emphasize
process based research (Matthews and Stephens, 2002a) could potentially help increase the
efficiency of local research programs. Models would help define attainable research priorities
Page | 55
and suitable technologies to be tested with farmers, which would help make research more
effective. However, the selection and use of appropriate models is also challenging and
would depend on the questions needed to be answered.
For the specific case of SSA, models, need to be able to answer the practical problems faced
by farmers. In order to achieve this, models must be able to capture the complex routines of
the decision making process at farm level, while avoiding the mismatch between model
enquiry and end user needs (Matthews and Stephens, 2002a), a common burden with some
models. The collection and availability of quality data, e.g., soil and climate data, for initial
parameterization are key issues that need to be addressed in Africa. However, once these
issues are resolved, the time and resources invested in conventional research programs can be
improved, as models can be parameterized to a vast range of farming environments and
systems from where ex ante simulation can be developed in order to assist local research
initiatives.
The fact that current models are based on a sustainable systems research platform rooted in a
holistic approach which views the different components of a farming system as complex
interrelated sub-organisms, with distinct biophysical and socioeconomic boundaries
(Bawden, 1991; Ikerd, 1993) is an advantage. The ability of models such as the Agricultural
Production Systems Simulator (APSIM) (Holzworth et al., 2014; Keating et al., 2003;
McCown et al., 1996) to predict potential environmental and economic impacts of
management decisions on the systems short and long-term responses is a major potential
benefit for agricultural research. Therefore, in the following section, APSIM capabilities as a
decision support tool will be discussed, with specific interest in its potential to simulate the
nitrogen and crop residue interactions in N-deficient cropping CA systems of SSA.
2.4.1 Using the Agricultural Production Systems Simulator (APSIM)
In recent years, APSIM has been used in studies covering economic, biophysical and
management interfaces of the model. In these studies, APSIM has been successfully used in
simulating a wide range of interactions. Among them, tillage responses, soil water balances,
nutrient and residue responses have received much attention (Table 2-3). With the growing
interest in conservation agriculture and integrated soil fertility management, which have been
largely promoted as the flagship for improving soil and crop productivity in Africa, several
technologies, including crop residues, conservation tillage, fertilizers, farmyard manure and
Page | 56
legumes, have been increasingly promoted to smallholder farmers (Chianu et al., 2011;
Giller et al., 2011a; Machido, 2011). However, information on the adaptability, best-fit
resource allocation and response of these practices, is still scarce and not yet tailored to the
diverse farming environments of Africa. Under these circumstances, modelling can be of
great use in: (1) synthesising current knowledge; (2) exploring alternative agronomic
practices; (3) measuring the effects of management options; and (4) bringing theory and
practice together (Stern, 1993). Therefore, calibrating and validating models such as APSIM,
have the potential to provide new information, especially in SSA.
APSIM’s major advantage is its ability to simulate biophysical processes in systems where
there is also interest in the economic and ecological impacts of management practices
(Keating et al., 2003). APSIM also includes most of the major tropical crops such as maize,
cotton and legumes e.g., cowpea, mucuna, pigeon pea, soybean and peanut (Keating et al.,
2003), which are of great importance in low input farming systems. This makes APSIM an
appealing analytical and decision support tool for application in SSA. In such places APSIM
could help generate timely information to support experimental design and discussions with
farmers and agribusinesses in participatory modelling research activities.
Table 2-3 The potential of APSIM in aiding agricultural research in SSA
Author Location Modelled parameters
Power et al. (2011) Australia Resource allocation in a farm business
Rodriguez et al. (2011) Australia Whole-Farm business design and analysis
Sadras and Rodriguez
(2010)
Australia Trade-off in Nitrogen and Water Use Efficiency
Balwinder et al. (2011) India Crop residue management and water use in wheat
crops
Mupangwa et al. (2011) Zimbabwe Tillage and residue effects on yield
MacCarthy et al. (2009) Ghana Nutrient and residue management
Lisson and Cotching
(2011)
Australia Fate of water and nitrogen in multiple cropping
systems
Ncube et al. (2009) Zimbabwe Residual effect of grain legumes,
Cox et al. (2010) Australia Profitability of cropping patterns,
Mupangwa and Jewitt
(2011)
Zimbabwe No-tillage effect on water fluxes and crop
productivity
Akponikpè et al. (2010) Niger Long term simulation of the effect of nitrogen
management
Page | 57
2.4.2 Gaps in modelling crop residue decomposition and N-cycling with APSIM
Residue decomposition and N-mineralization have been widely simulated with APSIM in
varying cropping environments, from drought affected semi-arid to humid tropics (Balwinder
et al., 2011; Gaydon et al., 2012; Mohanty et al., 2012; Mupangwa et al., 2007; Nascimento
et al., 2011). A mixture of results has been observed. APSIM simulated the decomposition of
Stizolobium aterrimum and Calopogonium mucunoides with a 37.6% and 48.2% error,
respectively, compared with the observed data in the absence of calibration, highlighting that
the reason for this error is the model’s inability to accurately capture initial residue
decomposition in the first 30 days. Separate studies have successfully predicted crop
responses to residue application (Balwinder et al., 2011; Gaydon et al., 2012) after assuming
a constant moisture factor of 0.5. The APSIM-residue module uses a combination of
algorithms where residue decomposition is a function of moisture, temperature and residue
C/N ratio (Probert et al., 1998), with adjustments made to the latter(Thorburn et al., 2001).
The algorithms have been basically developed to represent residue decomposition in large
residue mass per unit area, i.e., higher than 10 t/ha. A later modification was made for residue
amounts below 5-7 t/ha (Thorburn et al., 2001), which is a closer representation of SSA
smallholder farming systems where less residue is retained.
A positive aspect of the modified algorithms for residue decomposition is that they were
based on field experiments different from the mineralization studies. However, despite the
representativeness that the modification brings, further adjustments need to be made in the
algorithm of the RESIDUE module, to take account of the diversity of residues incorporated
through tillage in most conditions where the model is used. For some areas of Mozambique
and other semi-arid regions for instance, there is a need to make a distinction between termite
infested and non-infested soils, as the residue break-down and decomposition rate will differ
between these environments. However, while there is scope within the APSIM framework to
deal with termite loss, until now, this aspect has not yet been developed and tested.
As per N-dynamics, the process of N-mineralization and immobilization in the soil is the
main driver of nitrogen supply in cropping lands, thus the centre of research on N-dynamics
in agricultural land. In recent years, apart from model simulations such as in APSIM, N-
mineralization and immobilization have been studied through several incubation experiments
across the world (Coppens et al., 2006; Mary, 1996; Mohanty et al., 2010; Ros, 2012).
Page | 58
Initially, there was common agreement that the rate of N-mineralization and immobilisation
during crop residue decomposition depends largely on the type of residue applied and on the
existence of adequate soil moisture and favourable soil temperature to start the process. As a
result, several incubation studies have been conducted to analyse the decomposition rate of
different crop residue types under different management conditions and environments.
Despite several mineralization studies being conducted in soils with high amounts of mineral
N, it was possible to demonstrate that in N-constrained soils, the decomposition of soil
organic matter was not totally stopped but decreased; therefore, the mineralization intensity
was reduced and N-mineralization was delayed (Mary, 1996). The findings from incubation
trials for contrasting soil N-levels, (i.e., N-rich and N-poor soils), have proven that the
immobilization and mineralization process can actually be controlled through a feedback
effect by managing the soil N availability, plus the amount and type of residue applied,
thereby providing options for an improvement in the design of better crop residue
management strategies. However, there still remain questions to be answered regarding
mineralisation-immobilization turnovers in soils, especially regarding the role of the different
soil organic matter pools involved in N-cycling in the soil, and to what extent the laboratory
based mineralization results can be extrapolated to field conditions (Mary, 1996).
In a recent review that re-analysed 59 published datasets on the predictions of N-
mineralization, it was found that the amount of N mineralized from applied surface residues
was primarily linked to the size of the extractable organic matter (EOM) fraction, whereas
variables reflecting soil texture and organic matter quality were less important (Ros, 2012).
These findings add to the discussion that more mechanistic research is needed to explore the
biochemical basis of the relationship between total organic matter, EOM and N-
mineralization, rather than focusing on the C/N ratio of the material.
While using multivariate statistical modelling to analyse the role of organic matter and soil
proprieties in determining the variations of mineralizable N in 98 agricultural soils under
different management, it was found that the influence of variables reflecting soil texture or
organic matter quality (e.g., C/N ratio and extractable C and N), had little influence on the
amount of mineralizable N, compared to the size of the total organic matter pool. This means
that OM quality variables such as C:N ratio actually have less potential to improve the
accuracy of mineralizable N predictions (Ros et al., 2011). Another issue raised in this study
was the relevance of C-P and N-P ratios in explaining the magnitude of mineralizable N
Page | 59
variation, especially in N-constrained soils. In the experiment, these two ratios significantly
helped explain the variations in mineralizable N; however, the soil to which the study related
was not P-deficient, opening the question of whether these processes should be also studied
on P-deficient soils.
Based on these findings it can be concluded that key areas of research in terms of the
assumptions in the calculation of N and residue dynamics in the APSIM model are: (1)
modelling the contribution of the different organic matter and nitrogen pools to N-
mineralization and immobilization under different residue and nitrogen application levels,
especially in N-deficient soils; (2) to understand how management, mainly tillage, affects the
availability of the different organic matter and nitrogen pools, and its linkage with the
amounts of N-mineralization and immobilization under different residue-fertilizer
combinations in N-deficient soils; and (3) measuring N-mineralization and immobilization
during fallow is an area for exploration, as most of the residues from each season, is, in some
systems, incorporated right after crop harvest and left to decompose during the fallow period.
References
Ade Freeman, H., and Omiti, J. (2003). Fertilizer use in semi-arid areas of Kenya: Analysis of
smallholder farmers' adoption behaviour under liberalized markets. Nutrient Cycling
in Agroecosystems 66, 23-31.
Adjei-Nsiah, S. (2012). Evaluating sustainable cropping sequences with cassava and three
grain legume crops: Effects on soil fertility and maize yields in the semi-deciduous
forest zone of Ghana. Journal of Soil Science and Environmental Management 3, 49-
55.
Adjei-Nsiah, S., Kuyper, T., Leeuwis, C., Abekoe, M., Cobbinah, J., Sakyi-Dawson, O., and
Giller, K. (2008). Farmers’ agronomic and social evaluation of productivity, yield and
N-fixation in different cowpea varieties and their subsequent residual N effects on a
succeeding maize crop. Nutrient Cycling in Agroecosystems 80, 199-209.
Aggarwal, R. K., Praveen, K., and Power, J. F. (1997). Use of crop residue and manure to
conserve water and enhance nutrient availability and pearl millet yields in an arid
tropical region. Soil and Tillage Research 41, 43-51.
Akponikpè, P. B. I., Gérard, B., Michels, K., and Bielders, C. (2010). Use of the APSIM
model in long term simulation to support decision making regarding nitrogen
Page | 60
management for pearl millet in the Sahel. European Journal of Agronomy 32, 144-
154.
Alene, A. D., and Coulibaly, O. (2009). The impact of agricultural research on productivity
and poverty in sub-Saharan Africa. Food Policy 34, 198-209.
Asiwe, J. A. N. (2009). Needs assessment of cowpea production practices, constraints and
utilization in South Africa. African Journal of Biotechnology 8, 5383-5388
Aune, J. B., and Bationo, A. (2008). Agricultural intensification in the Sahel – The ladder
approach. Agricultural Systems 98, 119-125.
Austin, A. T., Piñeiro, G., and Gonzalez-Polo, M. (2006). More is less: agricultural impacts
on the N cycle in Argentina. Biogeochemistry 79, 45-60.
Awonaike, K. O., Kumarasinghe, K. S., and Danso, S. K. A. (1990). Nitrogen fixation and
yield of cowpea (Vigna unguiculata) as influenced by cultivar and Bradyrhizobium
strain. Field Crops Research 24, 163-171.
Bado, B., Bationo, A., and Cescas, M. (2006). Assessment of cowpea and groundnut
contributions to soil fertility and succeeding sorghum yields in the Guinean savannah
zone of Burkina Faso (West Africa). Biology and Fertility of Soils 43, 171-176.
Bakht, J., Shafi, M., Jan, M. T., and Shah, Z. (2009). Influence of crop residue management,
cropping system and N fertilizer on soil N and C dynamics and sustainable wheat
(Triticum aestivum L.) production. Soil and Tillage Research 104, 233-240.
Balwinder, S., Gaydon, D. S., Humphreys, E., and Eberbach, P. L. (2011). The effects of
mulch and irrigation management on wheat in Punjab, India—Evaluation of the
APSIM model. Field Crops Research 124, 1-13.
Bationo, A., Hartemink, A., Lungu, O., Naimi, M., Okoth, P., Smaling, E., Thiombiano, L.,
and Waswa, B. (2012a). Knowing the African Soils to Improve Fertilizer
Recommendations. In "Improving Soil Fertility Recommendations in Africa using the
Decision Support System for Agrotechnology Transfer (DSSAT)." (J. Kihara, D.
Fatondji, J. W. Jones, G. Hoogenboom, R. Tabo and A. Bationo, eds.), pp. 19-42.
Springer Science and Business Media, Netherlands.
Bationo, A., Kihara, J., and Adesina, A. (2012b). Beyond Biophysical Recommendations:
Towards a New Paradigm. In "Improving Soil Fertility Recommendations in Africa
using the Decision Support System for Agrotechnology Transfer (DSSAT)." (J.
Kihara, D. Fatondji, J. W. Jones, G. Hoogenboom, R. Tabo and A. Bationo, eds.), pp.
169-184. Springer Science and Business Media, Netherlands.
Page | 61
Bationo, A., and Waswa, B. S. (2011). New Challenges and Opportunities for Integrated Soil
Fertility Management in Africa. In "Innovations as Key to the Green Revolution in
Africa" (A. Bationo, B. Waswa, J. M. M. Okeyo, F. Maina and J. M. Kihara, eds.), pp.
3-17. Springer Science and Business Media, Netherlands.
Baudron, F., Delmotte, S., Corbeels, M., Herrera, J. M., and Tittonell, P. (2015). Multi-scale
trade-off analysis of cereal residue use for livestock feeding vs. soil mulching in the
Mid-Zambezi Valley, Zimbabwe. Agricultural Systems 134, 97-106.
Bawden, R. J. (1991). Systems Thinking and Practice in Agriculture. Journal of Dairy
Science 74, 2362-2373.
Beintema, N. M., and G. J. Stads. (2008). Agricultural R&D in Sub-Saharan Africa: An era of
stagnation. ASTI Background Paper. International Food Policy Research Institute
Washington, USA
Bekunda, M., Sanginga, N., and Woomer, P. L. (2010). Chapter Four - Restoring Soil
Fertility in Sub-Sahara Africa. In "Advances in Agronomy" (L. S. Donald, ed.), Vol.
Volume 108, pp. 183-236. Academic Press, USA.
Bouma, J., and Jones, J. W. (2001). An international collaborative network for agricultural
systems applications (ICASA). Agricultural Systems 70, 355-368.
Brouder, S. M., and Gomez-Macpherson, H. (2014). The impact of conservation agriculture
on smallholder agricultural yields: A scoping review of the evidence. Agriculture,
Ecosystems & Environment 187, 11-32.
Carberry, P. S., Hochman, Z., Hunt, J. R., Dalgliesh, N. P., McCown, R. L., Whish, J. P. M.,
Robertson, M. J., Foale, M. A., Poulton, P. L., and van Rees, H. (2009). Re-inventing
model-based decision support with Australian dryland farmers. 3. Relevance of
APSIM to commercial crops. Crop and Pasture Science 60, 1044-1056.
Cheshire, M. V., Bedrock, C. N., Williams, B. L., Chapman, S. J., Solntseva, I., and
Thomsen, I. (1999). The immobilization of nitrogen by straw decomposing in soil.
European Journal of Soil Science 50, 329-341.
Chianu, J., Chianu, J., and Mairura, F. (2012). Mineral fertilizers in the farming systems of
sub-Saharan Africa. A review. Agronomy for Sustainable Development 32, 545-566.
Chianu, J., Nkonya, E., Mairura, F., Chianu, J., and Akinnifesi, F. (2011). Biological nitrogen
fixation and socioeconomic factors for legume production in sub-Saharan Africa: a
review. Agronomy for Sustainable Development 31, 139-154.
Page | 62
Chianu, J. N., and Tsujii, H. (2005). Determinants of farmers' decision to adopt or not adopt
inorganic fertilizer in the savannas of northern Nigeria. Nutrient Cycling in
Agroecosystems 70, 293-301.
Ciampitti, I. A., and Vyn, T. J. (2014). Nutrient Sufficiency Concepts for Modern Corn
Hybrids: Impacts of Management Practices and Yield Levels. Crop Management 47,
38-44.
Coppens, F., Garnier, P., De Gryze, S., Merckx, R., and Recous, S. (2006). Soil moisture,
carbon and nitrogen dynamics following incorporation and surface application of
labelled crop residues in soil columns. European Journal of Soil Science 57, 894-905.
Cox, H. W., Kelly, R. M., and Strong, W. M. (2010). Pulse crops in rotation with cereals can
be a profitable alternative to nitrogen fertiliser in central Queensland. Crop & Pasture
Science 61, 752-762.
Crawford, E. W., Jayne, T. S., and Kelly, V. A. (2006). Alternative Approaches for
Promoting Fertilizer Use in Africa. pp. 88. World Bank, Washington.
D'Emden, F. H., Llewellyn, R. S., and Burton, M. P. (2006). Adoption of conservation tillage
in Australian cropping regions - An application of duration analysis. Technological
Forecasting and Social Change 73, 630-47.
Dakora, F. D., and Keya, S. O. (1997). Contribution of legume nitrogen fixation to
sustainable agriculture in Sub-Saharan Africa. Soil Biology and Biochemistry 29, 809-
817.
Danso, S. K. and Eskew D.L (1984). Enhancing biological nitrogen fixation. IAEA Bulletin
Vol 26 Nr. 2 p.1-6.
Dalgliesh, N., and Foale, M. (1998). "Soil matters: Monitoring soil water and nutrients in
dryland farming," 1/Ed. CSIRO, Australia, Toowoomba, Australia.
Davidson, A. P. (1987). Does farming systems research have a future? Agricultural
Administration and Extension 24, 69-77.
Derpesch, R. (2008). Critical Steps to No-Till Adoption. In "No-till Farming Systems" (T.
Goddard and M. A. Zoebisch, Gan, Y., Ellis, W., Watson, A. and Sombatpanit, S.,
eds.), World Association of Soil and Water Conservation. pp. 479 – 95.
Di Ciocco, C., Coviella, C., Penon, E., Diaz-Zorita, M., and Lopez, S. (2008). Short
communication. Biological fixation of nitrogen and N balance in soybean crops in the
pampas region. Spanish Journal of Agricultural Research 6, 114-119.
Page | 63
Enyong, L. A., Debrah, S. K., and Bationo, A. (1999). Farmers' perceptions and attitudes
towards introduced soil-fertility enhancing technologies in western Africa. Nutrient
Cycling in Agroecosystems 53, 177-187.
Erenstein, O. (2002). Crop residue mulching in tropical and semi-tropical countries: An
evaluation of residue availability and other technological implications. Soil and
Tillage Research 67, 115-133.
Erenstein, O. (2003a). Smallholder conservation farming in the tropics and sub-tropics: a
guide to the development and dissemination of mulching with crop residues and cover
crops. Agriculture, Ecosystems and Environment 100, 17-37.
Erenstein, O. (2011). Cropping systems and crop residue management in the Trans-Gangetic
Plains: Issues and challenges for conservation agriculture from village surveys.
Agricultural Systems 104, 54-62.
Erenstein, O., Gérard, B., and Tittonell, P. (2015). Biomass use trade-offs in cereal cropping
systems in the developing world: Overview. Agricultural Systems 134, 1-5.
FAO (2008). Current world fertilizer trends and outlook to 2011-12. pp. 57. FAO, Rome,
Italy.
FAO (2010). Current world fertilizer trend and outlook to 2014. The Fertilizer Organization
Working Group, FAO. Italy, Rome.
Fischer, R. A., Santiveri, F., and Vidal, I. R. (2002). Crop rotation, tillage and crop residue
management for wheat and maize in the sub-humid tropical highlands: II. Maize and
system performance. Field Crops Research 79, 123-137.
Gangwar, K. S., Singh, K. K., Sharma, S. K., and Tomar, O. K. (2006). Alternative tillage
and crop residue management in wheat after rice in sandy loam soils of Indo-Gangetic
plains. Soil and Tillage Research 88, 242-252.
Gathumbi, S. M., Cadisch, G., and Giller, K. E. (2002). 15N natural abundance as a tool for
assessing N2-fixation of herbaceous, shrub and tree legumes in improved fallows. Soil
Biology and Biochemistry 34, 1059-1071.
Gaydon, D. S., Probert, M. E., Buresh, R. J., Meinke, H., Suriadi, A., Dobermann, A.,
Bouman, B., and Timsina, J. (2012). Rice in cropping systems—Modelling transitions
between flooded and non-flooded soil environments. European Journal of Agronomy
39, 9-24.
Gilbert, N. (2012). Dirt Poor: The key to tackle hunger in Africa is enriching its Soils. Nature
483, 525-227.
Page | 64
Giller, K. E., and Cadisch, G. (1995). Future benefits from biological nitrogen fixation: An
ecological approach to agriculture. Plant and Soil 174, 255-277.
Giller, K. E., Corbeels, M., Nyamangara, J., Triomphe, B., Affholder, F., Scopel, E., and
Tittonell, P. (2011). A research agenda to explore the role of conservation agriculture
in African smallholder farming systems. Field Crops Research 124, 468-472.
Giller, K. E., Rowe, E. C., de Ridder, N., and van Keulen, H. (2006). Resource use dynamics
and interactions in the tropics: Scaling up in space and time. Agricultural Systems 88,
8-27.
Gregory, D. I., and Bumb, B. L. (2006). Factors Affecting Supply of Fertilizer in Sub-
Saharan Africa. In "Agriculture and Rural Development Discussion Papers", Vol. 24,
pp. 81. The World Bank, Washington, USA.
Hauggaard-Nielsen, H., Ambus, P., and Jensen, E. S. (2003). The comparison of nitrogen
uses and leaching in sole cropped versus intercropped pea and barley. Nutrient
Cycling in Agroecosystems 65, 289-300.
Heffer, P., and Prud'homme, M. (2011). "Fertilizer Outlook 2011-2015." IFA, Paris, France.
Herridge, D. F., Peoples, M. B., and Boddey, R. M. (2008). Global inputs of biological
nitrogen fixation in agricultural systems. Plant and Soil 311, 1-18.
Holzworth, D. P., Huth, N. I., deVoil, P. G., Zurcher, E. J., Herrmann, N. I., McLean, G.,
Chenu, K., van Oosterom, E. J., Snow, V., Murphy, C., Moore, A. D., Brown, H.,
Whish, J. P. M., Verrall, S., Fainges, J., Bell, L. W., Peake, A. S., Poulton, P. L.,
Hochman, Z., Thorburn, P. J., Gaydon, D. S., Dalgliesh, N. P., Rodriguez, D., Cox,
H., Chapman, S., Doherty, A., Teixeira, E., Sharp, J., Cichota, R., Vogeler, I., Li, F.
Y., Wang, E., Hammer, G. L., Robertson, M. J., Dimes, J. P., Whitbread, A. M., Hunt,
J., van Rees, H., McClelland, T., Carberry, P. S., Hargreaves, J. N. G., MacLeod, N.,
McDonald, C., Harsdorf, J., Wedgwood, S., and Keating, B. A. (2014). APSIM –
Evolution towards a new generation of agricultural systems simulation.
Environmental Modelling and Software 62, 327-350.
Homann-Kee Tui, S., Valbuena, D., Masikati, P., Descheemaeker, K., Nyamangara, J.,
Claessens, L., Erenstein, O., van Rooyen, A., and Nkomboni, D. (2015). Economic
trade-offs of biomass use in crop-livestock systems: Exploring more sustainable
options in semi-arid Zimbabwe. Agricultural Systems 134, 48-60.
Page | 65
Hoyle, F. C., and Murphy, D. V. (2011). Influence of organic residues and soil incorporation
on temporal measures of microbial biomass and plant available nitrogen. Plant and
Soil 347, 53-64.
IFDC (2012). "IFDC report: An update on the work and progress of IFDC." IFDC, Alabama.
USA
Ikerd, J. E. (1993). The need for a system approach to sustainable agriculture. Agriculture,
Ecosystems and Environment 46, 147-160.
Vesterager, J. M., Nielsen N. E. and Høgh-Jensen, H (2007). Nitrogen budgets in crop
sequences with or without phosphorus-fertilised cowpea in the maize-based cropping
systems of semi-arid eastern Africa. African Journal of Agricultural Research 2, 261-
268.
Jayne, T. S., Mather, D., Mason, N., and Ricker-Gilbert, J. (2013). How do fertilizer subsidy
programs affect total fertilizer use in sub-Saharan Africa? Crowding out, diversion,
and benefit/cost assessments. Agricultural Economics 44, 687-703.
Jeranyama, P., Waddington, S. R., Hesterman, O. B. and Harwood, R. R. (2007). Nitrogen
effects on maize yield following groundnut in rotation on smallholder farms in sub-
humid Zimbabwe. African Journal of Biotechnology 6, 1503-1508
Jordan, V. W. L., Hutcheon, J. A., Donaldson, G. V., and Farmer, D. P. (1997). Research into
and development of integrated farming systems for less-intensive arable crop
production: experimental progress (1989–1994) and commercial implementation.
Agriculture, Ecosystems and Environment 64, 141-148.
Kaiser, D. E., and Rubin, J. C. (2013). Corn Nutrient Uptake as Affected by In-Furrow Starter
Fertilizer for Three Soils. Agronomy Journal 105, 1199-1210.
Keating, B. A., P.S. Carberry, G.L. Hammer, M.E. Probert, M.J. Robertson, D. H., N.I. Huth,
J.N.G. Hargreaves, H. Meinke, Z. H., G. McLean, K. Verburg, V. Snow, J.P. Dimes,
M. S., E. Wang, S. Brown, K.L. Bristow, and S. Asseng, S. C., R.L. McCown, D.M.
Freebairn, C.J. Smith (2003). An overview of APSIM, a model designed for farming
systems simulation. European Journal of Agronomy, 267-288.
Kelly, V. A. (2006). Factors Affecting Demand for Fertilizer in Sub-Saharan Africa. In
"Agricultural and Rural Development Discussion Papers", pp. 89. The World Bank,
Washington, USA.
Kintché, K., Guibert, H., Sogbedji, J. M., Levêque, J., Bonfoh, B., and Tittonell, P. Long-
term mineral fertiliser use and maize residue incorporation do not compensate for
Page | 66
carbon and nutrient losses from a Ferralsol under continuous maize–cotton cropping.
Field Crops Research. 184, 192-200
Kumar, K., and Goh, K. M. (2002). Management practices of antecedent leguminous and
non-leguminous crop residues in relation to winter wheat yields, nitrogen uptake, soil
nitrogen mineralization and simple nitrogen balance. European Journal of Agronomy
16, 295-308.
Laditi, M. A., Nwoke, O. C., Jemo, M., Abaidoo, R. C., and Ogunjobi, A. A. (2012).
Evaluation of microbial inoculants as biofertilizers for the improvement of growth and
yield of soybean and maize crops in savanna soils. African Journal of Agricultural
Research 7, 405-413.
Lafolie, F., Bruckler, L., de Cockborne, A. M., and Laboucarié, C. (1997). Modeling the
water transport and nitrogen dynamics in irrigated salad crops. Irrigation Science 17,
95-104.
Leport, L., Turner, N. C., Davies, S. L., and Siddique, K. H. M. (2006). Variation in pod
production and abortion among chickpea cultivars under terminal drought. European
Journal of Agronomy 24, 236-246.
Liska, A. J., Yang, H., Milner, M., Goddard, S., Blanco-Canqui, H., Pelton, M. P., Fang, X.
X., Zhu, H., and Suyker, A. E. (2014). Biofuels from crop residue can reduce soil
carbon and increase CO2 emissions. Nature Climate Change 4, 398-401.
Lisson, S. N., and Cotching, W. E. (2011). Modelling the fate of water and nitrogen in the
mixed vegetable farming systems of northern Tasmania, Australia. Agricultural
Systems 104, 600-608.
Littleboy, M., D.M. Freebairn, Silburn, D. M., Woodruff, D. R., and Hammer, G. L. (1999).
PERFECT Version 3.0 - A computer simulation model of Productivity, Erosion,
Runoff Functions to Evaluate Conservation Techniques.
Lupwayi, N. Z., Kennedy, A. C and Chirwa, R. M. (2011). Grain legume impacts on soil
biological processes in sub-Saharan Africa. African Journal of Plant Science 51, 1-7.
MacCarthy, D. S., Sommer, R., and Vlek, P. L. G. (2009). Modeling the impacts of
contrasting nutrient and residue management practices on grain yield of sorghum
(Sorghum bicolor (L.) Moench) in a semi-arid region of Ghana using APSIM. Field
Crops Research 113, 105-115.
Page | 67
Machido, D. A., Olufajo, O. O, Yakubu, S. E., and Yusufu, S. S. (2011). Enhancing the
contribution of the legumes to the N-fertility of soils of the semi-arid zone of Nigeria.
African Journal of Biotechnology 10, 1848-1853.
Mafongoya, P., Bationo, A., Kihara, J., and Waswa, B. (2006). Appropriate technologies to
replenish soil fertility in southern Africa. Nutrient Cycling in Agroecosystems 76, 137-
151.
Malhi, S. S., and Lemke, R. (2007). Tillage, crop residue and N fertilizer effects on crop
yield, nutrient uptake, soil quality and nitrous oxide gas emissions in a second 4-yr
rotation cycle. Soil and Tillage Research 96, 269-283.
Mary, S. R., D. Darwis and D. Robin (1996). Interactions between decomposition of plant
residues and nitrogen cycling in soil. Plant and soil, 71-82.
Mathews, R. B., and Stephens, W. (2002). "Crop-soil simulation models applications in
developing countries," CABI- Publishing, UK.
Matthews, R. B., and Stephens, W. (2002). "CROP-SOIL SIMMULATION MODELS:
Aplications in Developing Countries," CABI Publishing, Wallingford, UK.
Maxwell, S. (1986). Farming systems research: Hitting a moving target. World Development
14, 65-77.
Mazzoncini, M., Sapkota, T. B., Barberi, P., Antichi, D., and Risaliti, R. (2011). Long-term
effect of tillage, nitrogen fertilization and cover crops on soil organic carbon and total
nitrogen content. Soil and Tillage Research 114, 165-174.
McCown, R. L., Hammer, G. L., Hargreaves, J. N. G., Holzworth, D. P., and Freebairn, D.
M. (1996). APSIM: a novel software system for model development, model testing
and simulation in agricultural systems research. Agricultural Systems 50, 255-271.
Meinke, H., Baethgen, W. E., Carberry, P. S., Donatelli, M., Hammer, G. L., Selvaraju, R.,
and Stöckle, C. O. (2001). Increasing profits and reducing risks in crop production
using participatory systems simulation approaches. Agricultural Systems 70, 493-513.
Minot, N., and Benson, T. (2009). Fertilizer Subsidies in Africa: Are Vouchers the Answer?
IFPRI, Issue Brief 60, Washington, USA
Mohanty, M., Probert, M. E., Reddy, K. S., Dalal, R. C., Mishra, A. K., Subba Rao, A.,
Singh, M., and Menzies, N. W. (2012). Simulating soybean–wheat cropping system:
APSIM model parameterization and validation. Agriculture, Ecosystems and
Environment 152, 68-78.
Page | 68
Mohanty, M., Probert, M. E., Reddy, K. S., Dalal, R. C., Rao, A. S., and Menzies, N. W.
(2010). Modelling N mineralization from high C: N rice and wheat crop residues.
Proceedings of the 19th World Congress of Soil Science: Soil solutions for a changing
world, Brisbane, Australia, 1-6 August 2010. Symposium 1.5.2 Modelling critical
processes in changing soil, 28-31.
Morris, M., Kelly, V. A., Kopicki, R. J., and Byerlee, D. (2007). Fertilizer Use in African
Agriculture: Lessons Learned and Good Practice Guidelines. pp. 162. World Bank,
Wanshigton, USA.
Mupangwa, W., Dimes, J., Walker, S., and Twomlow, S. (2011). Measuring and simulating
maize (Zea mays L.) yield responses to reduced tillage and mulching under semi-arid
conditions. Agricultural Sciences 2, 167-174.
Mupangwa, W., and Jewitt, G. P. W. (2011). Simulating the impact of no-till systems on field
water fluxes and maize productivity under semi-arid conditions. Physics and
Chemistry of the Earth, Parts A/B/C 36, 1004-1011.
Mupangwa, W., Twomlow, S., Walker, S., and Hove, L. (2007). Effect of minimum tillage
and mulching on maize (Zea mays L.) yield and water content of clayey and sandy
soils. Physics and Chemistry of the Earth, Parts A/B/C 32, 1127-1134.
Naab, J., Chimphango, S., and Dakora, F. (2009). N2 fixation in cowpea plants grown in
farmers’ fields in the Upper West Region of Ghana, measured using 15N natural
abundance. Symbiosis 48, 37-46.
Nascimento, A. F. d., Mendonça, E. d. S., Leite, L. F. C., and Neves, J. C. L. (2011).
Calibration of the century, apsim and ndicea models of decomposition and n
mineralization of plant residues in the humid tropics. Revista Brasileira de Ciência do
Solo 35, 917-928.
Ncube, B., Dimes, J., Twomlow, S., Mupangwa, W., and Giller, K. (2007). Raising the
productivity of smallholder farms under semi-arid conditions by use of small doses of
manure and nitrogen: a case of participatory research. Nutrient Cycling in
Agroecosystems 77, 53-67.
Ncube, B., Dimes, J. P., van Wijk, M. T., Twomlow, S. J., and Giller, K. E. (2009).
Productivity and residual benefits of grain legumes to sorghum under semi-arid
conditions in south-western Zimbabwe: Unravelling the effects of water and nitrogen
using a simulation model. Field Crops Research 110, 173-184.
Page | 69
Ndakidemi, P. A. (2006). Manipulating legume/cereal mixtures to optimize the above and
below ground interactions in the traditional African cropping systems. African
Journal of Biotechnology 5 2526-2533.
Ngwira, A. R., Thierfelder, C., and Lambert, D. M. (2013). Conservation agriculture systems
for Malawian smallholder farmers: long-term effects on crop productivity,
profitability and soil quality. Renewable Agriculture and Food Systems 28, 350-363.
Odhiambo, J. J. O., Ogola, J. B. O., and Madzivhandila, T. (2010). Effect of green manure
legume - maize rotation on maize grain yield and weed infestation levels. African
Journal of Agricultural Research 5, 618-625.
Oyun, M. B. (2007). Nodulation and Rhizobium Population in Root Nodules of Selected Tree
Legumes. Nigerian Journal of Soil Science 17, 65-70
Pala, M., Ryan, J., Zhang, H., Singh, M., and Harris, H. C. (2007). Water-use efficiency of
wheat-based rotation systems in a Mediterranean environment. Agricultural Water
Management 93, 136-144.
Peoples, M., Brockwell, J., Herridge, D., Rochester, I., Alves, B., Urquiaga, S., Boddey, R.,
Dakora, F., Bhattarai, S., Maskey, S., Sampet, C., Rerkasem, B., Khan, D.,
Hauggaard-Nielsen, H., and Jensen, E. (2009). The contributions of nitrogen-fixing
crop legumes to the productivity of agricultural systems. Symbiosis 48, 1-17.
Peoples, M. B., Herridge, D. F., and Ladha, J. K. (1995). Biological nitrogen fixation: An
efficient source of nitrogen for sustainable agricultural production. In "Management
of Biological Nitrogen Fixation for the Development of More Productive and
Sustainable Agricultural Systems." (J. K. Ladha and M. B. Peoples, eds.), Vol. 65, pp.
3-28. Springer Science and Business Media, Netherlands.
Poulton, C., Kydd, J., and Dorward, A. (2006). Increasing Fertilizer Use in Africa: What
Have We Learned? In "Agriculture and Rural Development Discussion Paper 25",
Vol. 25, pp. 52. The World Bank, Washington, USA.
Power, B., Rodriguez, D., deVoil, P., Harris, G., and Payero, J. (2011). A multi-field bio-
economic model of irrigated grain–cotton farming systems. Field Crops Research
124, 171-179.
Probert, M. E., Dimes, J. P., Keating, B. A., Dalal, R. C., and Strong, W. M. (1998). APSIM's
water and nitrogen modules and simulation of the dynamics of water and nitrogen in
fallow systems. Agricultural Systems 56, 1-28.
Page | 70
Robertson, G. P., Grace, P. R., Izaurralde, R. C., Parton, W. P., and Zhang, X. (2014). CO2
emissions from crop residue-derived biofuels. Nature Climate Change 4, 933-934.
Rodriguez, D., deVoil, P., Power, B., Cox, H., Crimp, S., and Meinke, H. (2011). The
intrinsic plasticity of farm businesses and their resilience to change. An Australian
example. Field Crops Research 124, 157-170.
Ros, G. H. (2012). Predicting soil N mineralization using organic matter fractions and soil
properties: A re-analysis of literature data. Soil Biology and Biochemistry 45, 132-
135.
Ros, G. H., Hanegraaf, M. C., Hoffland, E., and van Riemsdijk, W. H. (2011). Predicting soil
N mineralization: Relevance of organic matter fractions and soil properties. Soil
Biology and Biochemistry 43, 1714-1722.
Ros, G. H., Hoffland, E., van Kessel, C., and Temminghoff, E. J. M. (2009). Extractable and
dissolved soil organic nitrogen – A quantitative assessment. Soil Biology and
Biochemistry 41, 1029-1039.
Ross R. Bender, J. W. H., Matias L. Ruffo, and Fred E. Below (2013). Modern Corn Hybrids’
Nutrient Uptake Patterns. Better Crops 97, 7-10.
Roy, R. N., Finck, A., Blair, G. J., and Tandon, H. L. S. (2006). Plant nutrition for food
security - A guide for integrated nutrient management. (FAO, ed.), pp. 368, Rome.
Sadras, V. O., and Rodriguez, D. (2010). Modelling the nitrogen-driven trade-off between
nitrogen utilisation efficiency and water use efficiency of wheat in eastern Australia.
Field Crops Research 118, 297-305.
Salvagiotti, F., Cassman, K. G., Specht, J. E., Walters, D. T., Weiss, A., and Dobermann, A.
(2008). Nitrogen uptake, fixation and response to fertilizer N in soybeans: A review.
Field Crops Research 108, 1-13.
Samra, J. S., Bijay, S., and Kumar, K. (2003). "Managing crop residues in the rice-wheat
system of the Indo-Gangetic Plain.p.173-195. In "Ladha J.K. (Ed.) Improving the
productivity and sustainability of rice–wheat systems, Issues and impact." American
Society of Agronomy Special Publication, 65 Wisconsin, USA
Sangakkara, U. R., Ratnayake, A. D. A., and Kumarasinghe, K. S. (2002). "Impact of soil
moisture on nodulation, biological nitrogen fixation and yields of selected tropical
legumes," International Atomic Energy Agency (IAEA), Vienna, Austria.
Sanginga, N. (2003). Role of biological nitrogen fixation in legume based cropping systems;
a case study of West Africa farming systems. Plant and Soil 252, 25-39.
Page | 71
Shah, Z., Shah, S. H., Peoples, M. B., Schwenke, G. D., and Herridge, D. F. (2003). Crop
residue and fertiliser N effects on nitrogen fixation and yields of legume-cereal
rotations and soil organic fertility. Field Crops Research 83, 1-11.
Shivakumar, B. G., Mshhra, B. N., and Gautam, R. C. (2004). Nutrient economy through land
configuration and residue management in greengram (Phaseolus radiatus L.)-Wheat
(Triticum aestivum L.) cropping sequence with limited water supllies. Acta
Agronomica Hungarica 52, 369-379.
Siame, J., Willey, R. W., and Morse, S. (1998). The response of maize/Phaseolus
intercropping to applied nitrogen on Oxisols in northern Zambia. Field Crops
Research 55, 73-81.
Sidhu, B. S., and Beri, V. (1989). Effect of Crop Residue Management on the Yields of
Different Crops and on Soil Properties. Biological Wastes 27, 15-27.
Somado, E. A., Sahrawat, K. L., and Kuehne, R. F. (2006). Rock Phosphate-P enhances
biomass and nitrogen accumulation by legumes in upland crop production systems in
humid West Africa. Biology and Fertility of Soils 43, 124-130.
Sprent, J. I., Odee, D. W., and Dakora, F. D. (2010). African legumes: a vital but under-
utilized resource. Journal of Experimental Botany 61, 1257-65.
Stern, W. R. (1993). Nitrogen fixation and transfer in intercrop systems. Field Crops
Research 34, 335-356.
Tabo, R., Bationo, A., Gerard, B., Ndjeunga, J., Marchal, D., Amadou, B., Garba Annou, M.,
Sogodogo, D., Taonda, J.-B., Hassane, O., Diallo, M., and Koala, S. (2007).
Improving cereal productivity and farmers’ income using a strategic application of
fertilizers in West Africa Advances. In "in Integrated Soil Fertility Management in
sub-Saharan Africa: Challenges and Opportunities." (A. Bationo, B. Waswa, J.
Kihara and J. Kimetu, eds.), pp. 201-208. Springer Science and Business Media,
Netherlands.
Tenkorang, F., and Lowenberg-DeBoer, J. (2008). Forescasting Long-ter Global Fertilizet
Demand. FAO, Rome, Itally.
Thierfelder, C., Cheesman, S., and Rusinamhodzi, L. (2012). A comparative analysis of
conservation agriculture systems: Benefits and challenges of rotations and
intercropping in Zimbabwe. Field Crops Research 137, 237-250.
Thierfelder, C., Mombeyarara, T., Mango, N., and Rusinamhodzi, L. (2013). Integration of
conservation agriculture in smallholder farming systems of southern Africa:
Page | 72
identification of key entry points. International Journal of Agricultural Sustainability
11, 317-330.
Thorburn, P. J., Probert, M. E., and Robertson, F. A. (2001). Modelling decomposition of
sugar cane surface residues with APSIM-residue. Field Crops Research 70, 223-232.
Tsedeke, A., Alene, A. D., Bergvinson, D., Shiferaw, B., Silim, S., Orr, A., and Asfaw, S.
(2012). "Tropical Grain Legumes in Africa and South Asia: Knowledge and
Opportunities," International Crops Research Institute for the Semi-Arid Tropics,
Nairobi, Kenya.
Turmel, M.-S., Speratti, A., Baudron, F., Verhulst, N., and Govaerts, B. (2015). Crop residue
management and soil health: A systems analysis. Agricultural Systems 134, 6-16.
Twomlow, S., Rohrbach, D., Dimes, J., Rusike, J., Mupangwa, W., Ncube, B., Hove, L.,
Moyo, M., Mashingaidze, N., and Mahposa, P. (2010). Micro-dosing as a pathway to
Africa’s Green Revolution: evidence from broad-scale on-farm trials. Nutrient
Cycling in Agroecosystems 88, 3-15.
Umeh, S. I., and Mbah, B. N. (2010). Measuring the benefits of biological nitrogen fixation
of soybean (Glycine max (l.) Merrill) in cassava (Manihot esculenta crantz) and
soybean intercrop. African Journal of Agricultural Research 5, 3354-3359.
Unger, P. W., Stewart, B. A., Parr, J. F., and Singh, R. P. (1991). Crop residue management
and tillage methods for conserving soil and water in semi-arid regions. Soil and
Tillage Research 20, 219-240.
Valbuena, D., Erenstein, O., Homann-Kee Tui, S., Abdoulaye, T., Claessens, L., Duncan, A.
J., Gérard, B., Rufino, M. C., Teufel, N., van Rooyen, A., and van Wijk, M. T. (2012).
Conservation Agriculture in mixed crop–livestock systems: Scoping crop residue
trade-offs in Sub-Saharan Africa and South Asia. Field Crops Research 132, 175-184.
Valbuena, D., Tui, S. H.-K., Erenstein, O., Teufel, N., Duncan, A., Abdoulaye, T., Swain, B.,
Mekonnen, K., Germaine, I., and Gérard, B. (2015). Identifying determinants,
pressures and trade-offs of crop residue use in mixed smallholder farms in Sub-
Saharan Africa and South Asia. Agricultural Systems 134, 107-118.
Vanclay, F. (2004). Social principles of agricultural extension to assist in the promotion of
natural resource management. Australian Journal of Experimental Agriculture 44,
213–222.
Page | 73
Verburg, K., Bond, W. J., and Hunt, J. R. (2012). Fallow management in dryland agriculture:
Explaining soil water accumulation using a pulse paradigm. Field Crops Research
130, 68-79.
Wall, P. C. (2007). Tailoring Conservation Agriculture to the Needs of Small Farmers in
Developing Countries. Journal of Crop Improvement 19, 137-155.
Wang, X., Wu, H., Dai, K., Zhang, D., Feng, Z., Zhao, Q., Wu, X., Jin, K., Cai, D., Oenema,
O., and Hoogmoed, W. B. (2011). Tillage and crop residue effects on rain fed wheat
and maize production in northern China. Field Crops Research. 132, 106-116
Wang, Y., Xu, J., Shen, J., Luo, Y., Scheu, S., and Ke, X. (2010). Tillage, residue burning
and crop rotation alter soil fungal community and water-stable aggregation in arable
fields. Soil and Tillage Research 107, 71-79.
Whitbread, A. M., Robertson, M. J., Carberry, P. S., and Dimes, J. P. (2010). How farming
systems simulation can aid the development of more sustainable smallholder farming
systems in southern Africa. European Journal of Agronomy 32, 51-58.
Wu, Y., Liu, S., Young, C. J., Dahal, D., Sohl, T. L., and Davis, B. (2015). Projection of corn
production and stover-harvesting impacts on soil organic carbon dynamics in the U.S.
Temperate Prairies. Scientific Reports 5: 10830.
Yadvinder, S., Bijay, S., and Timsina, J. (2005). Crop Residue Management for Nutrient
Cycling and Improving Soil Productivity in Rice-Based Cropping Systems in the
Tropics. In "Advances in Agronomy", Vol. Volume 85, pp. 269-407. Academic Press.
Yusuf, A. A., Iwuafor, E. N. O., Abaidoo, R. C., Olufajo, O. O., and Sangin, N. (2009).
Effect of crop rotation and nitrogen fertilization on yield and nitrogen efficiency in
maize in the northern Guinea savanna of Nigeria. African Journal of Agricultural
Research 4, 913-921.
Zahran, H. H. (1999). Rhizobium-legume symbiosis and nitrogen fixation under severe
conditions and in an arid climate. Microbiology and Molecular Biology Reviews 63,
968-89
Page | 74
CHAPTER 3:
Characterising Farmers Diversity and Opportunities for Sustainable Intensification
No smallholder farmer chooses to remain a subsistence farmer. Smallholders are always
adapting and looking for a way out! Our responsibility is to provide farmers with the right
set of tools for them to make this transformation happen in a more productive, profitable, risk
minded, and planet friendly way, through a more conscious exploitation of the resources at
their disposal.
Page | 75
Abstract
Understanding farm diversity and its effects on farming system design and the household
ability to meet its food security and income generation targets is critical to identify scalable
intensification options and constraints. In this study, a new set of simple and mutually
exclusive farm typologies were generated and used to identify group best-fit intensification
options. A mixed dataset of qualitative and quantitative household survey data and resource
allocation maps collected across two villages in Central Mozambique was used to identify
key farm differentiation variables using Categorical Principal Component Analysis
(CATPCA). Selected variables were then used to define mutually exclusive agents using a
heuristic decision tree. In the classification tree Boolean statements based on the two key
household differentiation thresholds namely resource endowment (access and usage) and
household subsistence (food security and income generation capacity) level were developed.
Five different farm typologies were developed, resource endowed and semi-intensive farmers
(Type B); resource endowed but change resistant farmers (Type - A2); and finally resource
constrained farmers (Type - A1) who can be either maize self-sufficient (Type-A1a), or food
constrained (Type-A1b). For maize self-sufficient Type-B and Type-A2 farms, the challenge
is to improve agronomy to generate enough surplus and income. For that to happen,
mechanization and labour qualification are critical. For resource constrained Type-A1
farmers, the primary challenge towards intensification is to generate enough income to
improve the household maize self-sufficiency. Due to their high market value and further
contribution to soil N-availability, improving legume productivity appears to be a feasible
alternative for Type-A1 farmers. Two intensification constraints, were identified - the soil
fertility and maize self-sufficiency relation, and the land availability for extensive farming.
Based on current farm diversity, farm typology targeted agricultural interventions are
required to actively involve farmers in the design and implementation of locally feasible
intensification options centred on a smarter exploitation of available resources.
Keywords: Farm typologies, personalized agricultural extension, adoption traps
Page | 76
3.1 Introduction
In the last decades, the need to improve yields and food security prospects among Sub-
Saharan Africa (SSA) smallholder farmers, led to the promotion of several agricultural
technologies in a clear attempt to reproduce the 1960’s green revolution. Technological
packages such as conservation agriculture, (Giller et al., 2011a; Thierfelder et al., 2013; Wall,
2007), and more recently, sustainable intensification (Robinson et al., 2015; Tittonell and
Giller, 2013; Zimmerer et al., 2015) which has recently been presented as the “new paradigm
for Agriculture in Africa” in the 2013 Montpellier Panel Report under the proposition that
food production targets should be attained with the lowest possible environmental impact.
CA and SI are mere examples of the vast array of technologies and developmental models
that are being promoted across Africa. However, despite the efforts, technology adoption
rates in SSA are still low (Baudron et al., 2015c) compared to the rest of the world.
The low technology adoption rates among SSA farmers are in part due to the existing
inability to downscale complex and knowledge intensive technological packages into
practical onsite measures that fit farmer’s circumstances. This is continued due to the
prevalence of a supply driven agricultural intervention model, (Bembridge, 1987; Binns et al.,
1997), centred on single sized technological packages. Supply driven agricultural intervention
have failed to capture the complex and strong social dimension of technology adoption
(Vanclay, 2004a). Studying and incorporating these social dynamics when planning
agricultural intervention has been a major challenge in many developing countries. This is
exacerbated by the departmentalized interventions which lead to poor linkages between the
key players, i.e., researchers across disciplines, research-extension-farmers and policy makers
(Alene and Coulibaly, 2009). In SSA complex and diverse farming environments, bringing
farmers to the negotiation table is critical to understand their circumstances and how their
perception of those same circumstances shape their livelihood strategies, farming systems
design and the likelihood to engage in new practices (Giller et al., 2011b).
Clustering farmers in functional typologies have helped shed light into how local systems
work and how understanding farm households and regional dynamics can help trigger a
conscious change of practice. Nevertheless, despite farm typologies being pinpointed as a
critical tool to improve agricultural planning back in the 1970’s (Kostrowicki, 1976;
Page | 77
Kostrowicki, 1977), rigid typologies based on farm size and on the idea of large
homogeneous groups have dominated agricultural intervention in SSA. In recent years,
however, mapping intra and intergroup diversity became useful to understand farmers’
adoption profiles and livelihood strategies (Nainggolan et al., 2013; Valbuena et al., 2008).
Since then, several approaches to group farmers in functional typologies have been
hypothesized. Perception-based farmer typologies (Guillem et al., 2012); resource
endowment (Cortez-Arriola et al., 2015) and livelihood aspirations (Dorward et al., 2009)
have been used to group farmers across the globe. Daloğlu et al. (2014) proposed a
framework to link farmer typologies with an agent based model (ABM) to aid soil and water
assessment planning in the USA Corn Belt. In Malawi, Franke et al. (2014) grouped farmers
in typologies to map their likelihood to benefit from legume intensification systems. Despite
the late developments in farm categorization, agricultural interventions in SSA still fails to
incorporate the complex social dynamics that are characteristic of smallholder farm
enterprises into the design of locally feasible technological packages.
In this paper data from a farm household characterization survey conducted in Manica
province, central Mozambique is used to map farm diversity and identify the key farm
differentiation factors that can be used to develop an alternative and easy to use farm
categorization framework. The framework would be an entry point to explore the within
group diversity and help tailor sustainable agricultural intensification practices. This study
builds on two key prepositions:
1) Understanding smallholder farmers’ biophysical and socioeconomic
circumstances, is key to effectively involve them in the co-design of resource
efficient and profitable farming systems capable of helping them meet their food
and income generation goals at the minimum possible effort.
2) Smallholder farmers’ livelihood strategies shape their farming practices.
Therefore, understanding how individual farmers and groups pursue their seasonal
food security and income generation goals is critical towards mapping farmer’s
adaptability and likelihood to engage in new practices.
Page | 78
3.2 Materials and Methods
3.2.1 Geographic setting and context of the study
The study was conducted in Manica province in central Mozambique. Manica falls within the
plateau zone and is characterized by two well-defined seasons: a wet season with 78-99% of
the total annual rainfall, i.e., from November to late April; and a dry season that runs from
May to October. Data was collected across agroecological zone R4 (AEZ-R4). AEZ-R4 has a
tropical dry climate with average annual rainfall of 834.7 mm (1951-2014). The soils are
predominately red ferralsols and luvisols (Gouveia and Azevedo, 1954).
Two villages of contrasting were selected for the farm household characterization survey
namely Matsinho in Vanduzi and Marera in Macate districts. Matsinho is a resource limited
area dominated by maize, sorghum, sunflower and commercial cotton production. In contrast,
Marera is the most productive of the two districts with well-established commercial vegetable
and fruit production based systems.
In both communities, farmers are grouped in three major categories: small-scale (0-5 ha),
semi-commercial (5-20 ha) and large scale commercial farmers (>20ha)(INE, 2011a). These
rigid typologies emerged in late 1970’s right after independence and are based on farm size,
access to production means, farm labour and the share of the production destined for
household consumption and market sales.
According to the 2009-2010 Mozambican Agricultural Census (CAP-2009), small-scale
farmers represent 96.4% of the agricultural explorations in the country (Cunguara and
Darnhofer, 2011; INE, 2011a). The Mozambican farm categorization system like many in
SSA are somehow a rigid version of Dorward et al. (2009) classification where small-scale
family farmers would fall within the “hanging in” subsistence farming households. However,
several structural changes have occurred since the 1970’s e.g., the bankruptcy of most state-
owned agricultural enterprises in the 1980’s and the markets liberalisation in the 1990s,
forced the household farm enterprise to evolve in order to adapt to the new context. Farmers
became more diverse and the boundaries between groups became less distinct. Differences
Page | 79
have also emerged within groups, mainly based on farmer’s individual approach to their farm
enterprise management in order to meet the household food security and income generation
goals (Cunguara and Darnhofer, 2011). All these structural changes call for a renewed look
into how the “farm enterprise” operates in the increasingly dynamic and diverse farming
circumstances. Therefore, understanding what factors differentiate each farm from another is
critical to design interventions that are more likely to be adopted.
3.2.2 Sampling and data analysis
3.2.2.1 Farm household and farming systems characterization
Semi-structured interviews where used to collect data on 1) household composition –
household size and farm labour availability; 2) resource use behaviour - land, fertilizer,
manure and residue use; 3) farming systems design and management, and 4) farm
productivity. A mixed dataset composed of quantitative and categorical variables was
generated from the questionnaire (Table 3-1). These variables where then screened for
relevance to mapping farm diversity. In total a random sample of 131 households distributed
across two villages in the AEZ-R4, namely 65 of 5874 in Matsinho-Vanduzi and 66 of 3804
in Marera-Macate (INE, 2011b). In total 10 communities were sampled across both districts.
Additionally, 58 people met across two focus group discussions in Marera and Rotanda were
used to cross-validate the survey data.
Two model assisted focus group discussion were held to collect additional information that
would help glean a better light into local cropping systems and sequences, resource use and
allocation strategies with focus on farm inputs, risk management strategies, yields and future
aspirations. The focus group discussion also helped parameterize the pre-defined farm
typologies. Therefore, During the focus group discussion, resource allocation maps (RAMs)
were built to gather in depth information on the whole farm household enterprise functioning
and its effect on field level management decisions and farm productivity. Two model assisted
focus group discussion were held. The first one in Marera (Macate) and was attended by 29
farmers not involved in the survey and representing 11 farmer organisations from Matsinho,
Zembe, Boavista and Marera-Sede. Two agronomists and three extension officers from the
Gondola Agricultural Directorate (SDAE-Gondola) also attended the meeting. In Rotanda, 22
Page | 80
farmers representing 7 FO’s from Messambuzi, Tsetsera and Rotanda-Sede, attended the
meeting. Three extension officers, from the SDAE-Sussundenga, also attended the Rotanda
meeting. The survey data and PCM for the household characterization was explored using
simple descriptive statistics.
3.2.2.2. Farm level productivity
Maize (Zea mays L.) self-sufficiency index, was calculated as the ratio between total maize
production and annual household maize consumption requirements. Total household maize
consumption requirements were generated by correcting Mozambique maize daily intake of
128 g maize/person which supplies 20 % daily energy intake (Nuss and Tanumihardjo, 2011;
Ranum et al., 2014) to a fully maize based diet capable of supplying 80 % daily energy intake
required to meet 2280 Kcal daily caloric requirement. Adult equivalent scales (AE) were used
to narrow the food consumption gaps between members of the same household with distinct
energy requirements. The number AE was calculated using Deaton and Zaidi (2002)
guidelines to generate consumption aggregates. For this study, adult well all farm active
people older than 15. Household consumption needs were corrected to a 17% post-harvest
loss during storage. Land to labour ratio (LLR) and yield to labour ratio (YLR), i.e., share of
land management and yield attributed to each farm active member as determined in
(Leonardo et al., 2015; Lukanu et al., 2007) were also analysed. This analysis excluded hired
labour under the assumption that each household farm active member managed his share of
labour in his portion of managed land. Maize surplus was determined as the difference
between food requirements and total maize production. The income from maize selling was
the product between maize surplus and maize market prize.
Table 3- 1 Household and farming systems characterization variables used for the categorical
principal components analysis (CATPCA)
Input variable Unit Abbreviation
District dummy Dist
Household head age # Hh_age
Household head gender dummy Hh_gender1F2M
Household head literacy level dummy Hhlireacy_lvl
Page | 81
Household size # Hhld_size
Household farm active people # Hhld_farmactpeople
Household farm active elder people # Hhld_farmactEldery
Household farm active youth # Hhld_farmactYouth
Household farm active females # Hhld_farmactFemale
Household farm active males # Hhld_farmactMale
Total number of fields cultivated in the sampled season # TotlplotOwnd
Total farm size # Total_FarmSIZE
Fertilizer use dummy Fert_Use
Fertilizer application years # FertyearsApplied
Preferred fertilizer allocation (1-maize, 2-legumes, 3-
vegetables)
dummy FertCropApplied
Best fertilizer returns crops (1-maize, 2-legumes, 3-
vegetables)
dummy fertBestrespo
Reasons not to use fertilizers dummy fertReasNOT2use
Conditions to use fertilizers dummy fertCond2Use
Fertilizer investment source dummy Fert_InvstSOURCE
Manure Use dummy Manre_Use
Crop where manure is applied (1-maize, 2-legumes, 3-
vegetables)
dummy mnreAlloction
Best manure returns crops (1-maize, 2-legumes, 3-
vegetables)
dummy mBestresponse
Reasons not to use manure dummy mReasNOT2USE
Conditions to use manure dummy mCond2Use
Cattle ownership dummy Cattle ownership
Crop residue retention dummy rsd_Use
Highest recalled yield # TtplotRecalledyield
Maize self-sufficiency index # Maize_SFIndex
labour to yield productivity index # LP_index
Labour to land ratio # LLR
Income from maize grain selling # Income_Grainselling
Maize surplus # Maize_surplus
Page | 82
3.2.2.3. Farm categorization
Variable selection
Survey data was explored by categorical principal component analysis (CATPCA), using
SPSS 23. CATPCA is a non-linear Principal Component Analysis (PCA) method used in
social and behavioural science research where mixed datasets composed of binary and
quantitative variables are common and standard PCA or Multiple Correspondence Analysis
(MCA) cannot be used (Linting et al., 2007). In agricultural research and related fields,
CATPCA allows the analysis of farmer’s behavioural attitudes that are key to understanding
decision-making processes and tailor agricultural intervention (Rouabhi et al., 2014).
CATPCA has been successful used in agriculture and related fields (Browne et al., 2014;
Comoé and Siegrist, 2013; Dossa et al., 2011a; Keshavarz and Karami, 2014; Rouabhi et al.,
2014). In this study, CATPCA was used for two purposes: first as dimension reduction
method to reduce the original set of variables, into a smaller one and to understand how much
of the variability found in the data could be explained by each selected variable; second, to
select the most influential set of variables for farm categorization that could be used to aid the
design of sustainable agricultural intensification pathways that are reflective of farmers’
needs. Therefore, the variables should be reflective of farmer’s socioeconomic circumstances
(Valbuena et al., 2008) and highlight existing farm diversity based on the associations
between farmers’ resource use behaviour and farm level productivity. Proper variable
selection for farm categorization has been pointed as more important than the classification
technique itself (Daloğlu et al., 2014; Kostrowicki, 1977) especially in highly diverse and
complex farming environments such as SSA and central Mozambique in particular.
Therefore, CATPCA is used in this study to explore a mixed dataset in search for the most
relevant categorical and quantitative variables to aid the design of simple and mutually
exclusive farm typologies. Power scores in the general dataset and loadings across the first 4
components for each district were used to rank the individual contribution of tested variables.
At the end, only variables with loadings higher than ± 0.5 were retained for analysis (Dossa et
al., 2011b).
Page | 83
Typology formulation - agent definition and parameterization
Because a mixed dataset of qualitative and quantitative variables was used as base for farm
categorization, a “classification tree” was chosen as the most appropriate method for
typology formulation (Rounsevell et al., 2012; Valbuena et al., 2008). To effectively capture
farmer’s contrasting socioeconomic circumstances, attitudes towards resource use and their
impact of the household subsistence level and field level productivity, the classification tree
is based on Boolean statements that were key to set the household operational boundaries for
the selected variables. Two thresholds were fixed. For resource endowment level, the
threshold was fixed on the household capacity to mobilize enough funds to pay for an
unsubsidized FAO Group B input package with a market cost value of 7000 MT (205 USD at
34 MT/USD). For the household subsistence level, the threshold was fixed based on farm
productivity level measured on maize self-sufficiency basis. Once the typologies were
defined, they were parameterized based on additional variables, namely household
demographic composition (household size, number of farm active people youth and adults),
farming system design and management (crops, size of cultivated land, number of fields).
Analysis of variance (ANOVA) was used to define whether the difference between
typologies were significant based on selected quantitative variables. Grouping variables for
the analysis were selected based on the CATPCA results.
3.3 Results
3.3.1 Farm household characteristics - Household demographic and land use
The average household size was c.a., 8 people, of which half were farm active (Table 3-2). Of
the household farm active members, 43.1% were female and 47.1% were men, which shows a
gender balance in farm household labour. However, of the total sampled households, 33.6%
were female-headed and the remaining 66.4% were headed by men. In Marera (Macate), 33.7
% of the households were female-headed and 66.3% were male-headed. In Matsinho
(Vanduzi), a similar trend was observed, with male-headed households (66.7%), appearing in
a higher number than female-headed ones. In terms of literacy levels, 52.7% of the household
heads had primary education, 21.4% had secondary education and less than 1% had a
university degree. The overall illiteracy level was 25.2 %. Marera, had the lowest percentage
of illiterate household heads, with 24.2% compared with the 26.2% registered in Matsinho.
Household head literacy levels, where highly affected by gender. A 45.5% illiteracy level was
found in female-headed households but only 18.3% in male headed households.
Page | 84
Table 3-2 Household characterization based on demographic data and land use
* Significant at 0.05 probability
** Significant at 0.01 probability
3.3.2 Farming systems design and management
Both, the survey and focus group discussions highlighted that rain fed maize based systems
are the most dominant systems in both villages. Nevertheless, cropping systems and
sequences were found to be highly influenced by the farm position in the landscape catena -
upland and lowland farming environments (Figure 3-1). Soil type, water availability and in
crop soil moisture regimes were the key differentiating factors between farming
environments. To take advantage of such biophysical variability, 74.1% of the farmers in the
region cultivated more than two separate fields per season. Of those households, 62.8% had
two fields and a third and fourth field was owned by 29.9% and 7.21% households
respectively. Only 25.9% of the households cultivated one single field per season.
Most of the cultivated fields are located in the upland farming environment where more land
is available. Here, apart from maize, patches of early intercropped cowpea (Vigna
unguiculata (L.) Walp) are grown. Sunflower (Helianthus annuus) and sorghum (Sorghum
bicolor (L.) Moench) are also cultivated mostly in relay intercrops with maize. The lowland
landscapes (Baixas), are wet environments dominated by mixed vegetable-cereal-legume
systems. Production in this environment is intensive due to high water availability and incrop
Household demographic
Number of cultivated fields
1 (N=34) 2 (N=61) 3 (N=29) 4 (N=7) MEAN P-value
Household size 7.03 6.89 8.10 8.43 7.61
Farm active people 3.59 3.85 4.38 5.43 4.31
Farm active adults 1.72 1.67 1.48 2 1.66
Farm active youth 2.62 2.71 3.65 3.71 2.95 *
Total cultivated land (ha) 2.03 3.09 6.20 6.14 3.67 **
Field size (ha) 2.03 1.54 2.69 1.535 1.95
Page | 85
moisture regimes which allows for season-long commercial vegetable production and sowing
a second opportunistic maize crop (Figure 3-1). Farmers in these environment normally
invest in fertilizer, manure and improved seed.
Legumes are a significant component of local cropping systems. Nearly 72.5% of the
households grow legumes. Of planted legumes, cowpea (72.6%) and common beans
(Phaseolus vulgaris L.) (62.3%) are the most planted legumes. Additionally, groundnuts
(Arachis hypogaea L.) and Bambara groundnuts (Vigna subterranean (L) Verdc) are grown
by 24.7 and 23.2% of the farmers. Legumes are mostly planted in monoculture (52.3%) and
62.1% legume growers believe the best legume yields are achieved in monocultures. In
25.3% of the farms early sowing legumes (Nov-Dec window) was identified to be where the
best yield responses, were achieved mainly with determinate erect cowpea varieties.
However, the vast majority sow cowpea in the usual January to February window. Yield
(30.1%), local adaptability (15.0%) and marketability (12.4%) were the major factors for
legume selection. Seed availability (7.84%) is also a significant determinant for legume
selection.
Page | 86
End fallow Rain Season Cool season AEZ-R4
(Matsinho and Marera) Aug Sep Oct Nov Dec Jan Feb March April May June July
Fertilized commercial vegetable production for season long producers
Mixed
Vegetable-Cereal systems
(Lowland fields – Matoros)
Veg-Maize opportunistic sequence
Fertilized vegetable Regular Maize sowing
(Veg-Maize sequence)
Vegetable production
Maize harvest
Patches of early sown
legume with Maize
Fresh beans Legume harvest
Regular legume sowing
(Sole + relay Intercrop)
Dry beans
Harvest Maize Normal Maize sowing Maize harvest Opportunistic Maize Maize based systems
(low land fields – N’dongo) Fallow period Fallow period - Unmanaged
Fallow period Normal Maize sowing Maize harvest Fallow period
Mixed cropping systems
Upland fields
(Kupswuka/Tchica)
Patches of early sown
Legume with Maize
Regular legume sowing
(Sole + relay Intercrop)
Legume harvest
1st Sorghum sowing
(Sole/Intercrop-Mz)
Sorghum harvest
Sunflower and Sorghum
(Sole or relay with Maize)
Sunflower harvest
Figure 3-1 Cropping systems description – crops and sequences for the AEZ-R4 gathered through the model assisted focus group discussions
and resource allocation maps (RAM’s) designed by local farmers
Page | 87
3.3.3 Resource use and allocation strategies
Fertilizer allocation and perceived yield advantages
Only 15.3% of the surveyed households reportedly applied fertilizers in the last 10 years. Of
these fertilizer users, 45% are located in the Marera (Macate) and the remaining 55% in
Matsinho (Vanduzi). In Matsinho, an average amount of 270 kg/season was used over the last
4.27 years. This represents approximately 63.7 kg of fertilizer per year. Since most farmers
reportedly use the compost NPK 12-24-12 as basal and Urea (46% N) in top-dress, around
18.3 kg N/year were applied by each farmer, assuming a 50/50 share between fertilizer types.
Despite an apparently lower number of people reportedly using fertiliser in Marera, annual
fertiliser application rates were 86.4 kg/year, which represents 24.1 kg N/year per farmer.
In total, 65% of the farmers applied fertilizer into their commercial vegetable production
which is where the best economic returns to investments were expected. Only 20% have
applied fertiliser directly to maize, and just 15% to legume crops (Figure 3-2) - common
beans (Phaseolus vulgaris) and soybeans (Glycine max) in particular. As a result, 65% of the
farmers reportedly got the funds to purchase fertiliser from the income generated from selling
their vegetables. Money from household-owned businesses was also used to purchase
fertilizers in 30% of the cases. Among non-fertilizer users, 58.9% cited the high cost of
fertiliser as the main reason for not using fertilizers, while 16.7% identified knowledge as the
main factor. Furthermore, willingness to use fertiliser was identified in 27.1% of the farmers
under the condition that a local provider was available and they were instructed about how to
efficiently use fertilisers. However, 17.6% of sampled farmers did not see a need to invest
into fertilizers and pointed at decreasing yields or soil fertility has the main condition to use
fertilizers, since they believed their soils are fertile enough to keep up with their production
requirements.
Manure use and perceived yield advantages
Manure is currently used by 22.3% of the households in the area. The vast majority of
manure users, were located in Marera (58.6%). In Matsinho, only 16.7% of manure users got
it directly from their own kraal and the remaining users either purchased (50%) or got it free
Page | 88
(33.3%). In Macate, 29.1% used manure from their own kraal and only 17.6% purchased
from animal owners. Free access from neighbouring animal owner was high in Macate
(52.9%). In terms of allocation, 58.8% of the farmers applied manure on their commercial
vegetables, production gardens where the best economic returns are expected (Figure 3-2).
Only 23.5% and 14.7% of manure users reportedly applying it into maize and legumes,
respectively. Among non-users, the major determinants to start using manure were cattle
ownership, purchase capacity, knowledge and transport means - 30.6 %, 24.7%, 17.6% and
7.06% respectively. Furthermore, 16.5 % of the surveyed households sow no benefits from
manure use.
Residue use and perceived yield advantage
Residues are reportedly retained by 74.1% of the sampled farmers across both communities.
Macate is more residue friendly with 60.8% of the households reportedly retaining crop
residues in the field. In Vanduzi, only 39.2% are retaining crop residues. Residues are
incorporated through ploughing in 85.6% of the farms. In 5.15% of the farms residues were
piled off the field and used as feedstock. Post-harvest grazing was reported by 11.3% of the
households. Piling residues to be used as feedstock is a common practice among small scale
dairy farmers in both Matsinho and Macate. Despite the existence of such farmers in the
region, no-one is reportedly selling maize stubble. Of all farmers retaining residues, only,
26.3% of them use fertilizers. Returns from residue application were perceived to be high in
maize by 58.8% of the households. Application into legumes and vegetables was identified as
beneficial only by 6.18% and 3.09% of households, respectively.
Page | 89
Figure 3- 2 Preferred fertilizer, manure and residue allocation across crops over the last 10 years in sampled districts. Interpretation: 1-Mz is
maize, 2-Leg is legume and 3-Veg is vegetables and 4 is shared across crops, excluding vegetables.
Page | 90
3.3.4 Farm typologies formulation
3.3.4.1 Farm diversity – key farm differentiation factors
The CATPCA revealed that resource use related variables – manure and fertilizer use,
contributed the most for the differences between households in both villages. Among
resource use related variables, farmers reasoning for not using manure was the most
influential variable with a 0.69 and 0.82 contribution to the observed variation in the first
principal component (PC1). The first two principal components together captured 57.6% of
the variability in Matsinho (Vanduzi) and 59.1% in Marera (Macate). In table 3-3, loadings
for the correlation matrix for 21 household characteristics in Matsinho and Marera are
presented. The variable loading values suggest that among the categorical variables, resource
use related variables i.e., manure and fertilizer use, associated with farmers reasoning for not
using neither of them could be used as key variables to categorize farm households in the
region. Among the quantitative variables, the household food security and income generation
status measured through maize self-sufficiency index and income from maize surplus selling
proved relevant for farm categorization. The CATPCA also showed that for both villages,
household composition related variables, i.e., the household size, the age and literacy level of
the household head were less influential. Despite being a key determinant of farmer’s risk
management strategy and significantly affecting the overall size of the household cultivated
land, the number of cultivated fields had a poor contribution to the observed variability as
CATPCA results showed. Nevertheless, similarly to the gender of the household head,
despite not explaining much of the variability observed these two variables could be used as
important grouping factors for farm categorization. Based on CATPCA analysis results which
indicated farmer’s attitudes towards resource use (access capacity and usage level) and farm
level productivity (maize self-sufficiency and income) as the key farm differentiation factors,
a heuristic decision tree (Figure 3-3) is proposed to group farmers into simplistic as mutually
exclusive typologies that are representative of existing intragroup diversity.
Page | 91
Table 3-3 Loading values for the correlation matrix for 21 farm household characteristics in
Matsinho (Vanduzi) and Marera (Macate).
Variables Categorical Principal Components
Component 1 Component 2 Component 3 Component 4
(a) Matsinho
(Vanduzi)
Manre_Use 0.853 -0.362 -0.293 0.064
mReasNOT2use 0.697 -0.409 -0.164 -0.043
Fert_Use 0.687 0.276 0.620 -0.088
mCond2Use 0.573 -0.369 0.082 -0.033
fertReasNOT2use 0.321 0.208 0.450 0.663
fertCond2Use 0.006 0.244 0.551 0.604
Income ($) _Maize -0.214 -0.799 0.470 0.032
Maize_SFIndex -0.255 -0.737 0.469 -0.102
Maize_surplus -0.258 -0.822 0.449 -0.060
Total Prod -0.311 -0.776 0.419 -0.021
LP_index -0.326 -0.691 0.495 -0.009
LLR -0.491 -0.323 -0.079 0.335
fertAmountoverTiMe -0.522 -0.295 -0.614 0.149
FertyearsAppld -0.549 -0.329 -0.647 0.192
fertCropAppld -0.667 -0.262 -0.623 0.126
fertBestrespo -0.667 -0.262 -0.623 0.126
mnreAlloction -0.806 0.354 0.331 0.005
mBestresponse -0.806 0.354 0.331 0.005
manre_Type -0.809 0.258 0.226 -0.243
manre_Source -0.832 0.295 0.298 -0.042
manre_yrsofUse -0.848 0.362 0.249 -0.101
(b) Marera (Macate)
Manre_Use 0.908 -0.118 -0.350 -0.060
mReasNOT2use 0.824 0.048 -0.342 0.034
mCond2Use 0.598 -0.237 -0.387 0.175
Fert_Use 0.528 0.744 0.359 -0.141
fertReasNOT2use 0.291 0.507 -0.094 0.575
Page | 92
fertCond2Use 0.122 0.539 -0.099 0.576
LLR 0.055 0.232 0.088 0.541
fertAmountoverTiMe -0.415 -0.629 -0.441 0.088
LP_index -0.416 0.526 -0.616 -0.040
fertCropAppld -0.467 -0.774 -0.313 0.171
Maize_surplus -0.478 0.520 -0.656 -0.114
Income ($) Maize -0.481 0.538 -0.640 -0.110
FertyearsAppld -0.486 -0.760 -0.354 0.159
fertBestrespo -0.528 -0.744 -0.359 0.141
Total Prod -0.540 0.446 -0.676 -0.051
Maize_SFIndex -0.545 0.526 -0.626 -0.087
manre_Source -0.735 0.142 0.467 0.158
manre_Type -0.775 0.315 0.359 0.001
mBestresponse -0.802 0.144 0.435 0.043
mnreAlloction -0.841 0.120 0.416 0.030
manre_yrsofUse -0.851 -0.030 0.327 0.026
The variables are ranked in importance to the first principal component
Page | 93
How resources are
allocated?
Cropij ... CropN
Type B (a, b, ... N) Type A2 (a, b, …N)
(Type R ij * Cropij ... Type RN * CropN)
Type A1(a, b, ... N)
(Type R ij * Cropij ... Type RN * CropN)
RAM’s: Farm productivity and livelihood strategy: Are subsistence requirements met, i.e., maize self-sufficient and maize income? (YES/NO).
Secondary data: household demographics (size, farm active people), farming systems design and management (crops, and size of cultivated land)
Household Income and assets (Resource Endowment and usage)
(Can the household generate enough income to pay for 1 Group-B FAO input package: 50 kg NPK 12:24:12 + 50 kg Urea + 12.5 kg Maize seed)
(Yes) (No)
Uses? YES!
Resource Endowed and Uses either
fertilizer or Manure
Uses? NO!
Resource Endowed but Uses neither
Fertilizer nor Manure
Resource Constrained and Uses neither of them
Reasons not to Use
(Rij … RN)
Conditions to Use
(Condij … CondN)
Reasons not to Use
(Rij … RN)
Conditions to Use
(Condij … CondN)
Figure 3-3 Framework for farmer categorization and personalize agricultural intervention. Key grouping factors were attitudes towards resource
use (fertilizer, manure) and the household maize self-sufficiency level
Page | 94
3.4.5 Farm typologies
Three key structural typologies (Figure 3-4) were identified based on resource endowment
level and usage: (1) the Type-B farming households, which are resource-endowed innovative
farmers; (2) Type-A2 farmers who are resource endowed but technology averse farmers and
finally the (3) Type-A1 farmers who are resource constrained smallholder farmers. Type-A1
farmers can be either maize self-sufficient (Type-A1b) or not (Type-A1a). In both villages,
Type-B households had the largest household size. However, the number of farm active
people only differed across typologies in Marera where the number of young farm active
people also differed. The household subsistence level differed across typologies. Here, farm
productivity indexes, i.e., maize self-sufficiency, maize income, the labour to land ration
(LLR) and crop production per farm active member (YLR) differed significantly across
typologies in both villages. Maize self-sufficiency in particular, proved to be affected by the
gender of the household head specially in Matsinho.
Figure 3-4 Hypothetical model for resource endowment based farm typologies. Typology
specific develop pathways are indicated by dashed arrows. Continuous arrow indicates a
stepping out situation (Dorward et al., 2009), i.e., a level of income diversification where the
farmer is able to generate enough off the farm income to reduce dependence from farming.
Page | 95
Table 3-4 Households distribution and characterization across farm types and sites
Sites variables Hhold size Farm active people Youth Adults Cultivated land Maize Income LLR YLR_index SSFIndex
Farm types (FT) # p/Hh # p/Hh # p/Hh # p/Hh (ha) (USD) ha/p kg/p
Matsinho
(Vanduzi)
A1a 7.22 4.04 2.96 1.09 2.04 - 0.613 66 0.16
A1b 5.67 3.62 2.67 0.88 2.46 265.1 0.787 683 1.55
A2 4.5 2 1.5 1 5.12 652.2 2.562 2125 6.01
B 8.12 4.94 3.31 1.62 7.27 130.1 1.447 420 2.75
FT - - - 0.044 0.01 <.001 <.001 <.001 <.001
FT x #_fields - - - 0.036 - - - - -
FT x Hhhead gender - - - - - - - - 0.032
Marera
(Macate)
A1a 7.22 3.24 1.89 0.73 3.68 - 1.28 62 0.13
A1b 7.25 4 2.75 1.25 4.34 440 1.01 1089 0.95
A2 6.33 4.67 3.67 1 4.92 508.7 1.09 1105 3.77
B 9.93 5.27 3.13 1.2 3.23 100.7 0.99 267 3.07
FT 0.041 0.026 0.013 - - <.001
<.001 <.001
FT x #_fields - - - - - - - - -
FT x Hhhead gender - - - - - - - - -
Legend: p/Hh is people per household; ha/p is hectares per person; kg/p is kg of maize produced per farm active person
(a) Resource endowed innovative farmers – Type B (23.7%)
Farmers in this group normally use fertilisers and manure. Applied manure is from their own
kraal or acquired locally. Type – B farmers cultivate at least two fields per season, of which one
is the wet lowland farming environment where they run a commercial vegetable production farm.
This lowland commercial vegetable production unit is where most of the fertilizer and manure
are allocated. Type-B farmers normally sell part of their maize. However, despite cultivating
more land than other groups (Matsinho) and also being maize self-sufficient (Table 3-4), Type-B
had lower yield to labour productivity ratio (YLR) and income from maize selling. This is
explained by the fact that most of their income is being generated from their commercial
vegetable production and legume crops rather than maize grain selling. Poor maize agronomy is
also an issue. Type-B farmer’s lowland mixed vegetable-maize systems are semi-intensive and
specialized in a reduced number of crops per season. Normally marketability is the key crop
selection factor. Here, an opportunistic maize crop is normally grown right after the end of the
cool season’s in June-August and runs up until the start of the regular maize sowing season in
October-December (Figure 3-1). In this sequence, maize mostly benefits from the residual N
from vegetable production, which also includes N-fixing fresh and dry beans. To take advantage
of the residual N and prolonged wetness of these lowland fields, farmers normally purchase high
yielding hybrid varieties. In terms of labour, due to external cash availability and maize surplus,
Type B farmers are able to hire draft animal power (DAP), tractor (HTP) for land preparation.
Nevertheless, sowing is not mechanized. The household head is mostly well connected in the
community. These farmers are also market oriented legume growers.
(b) Non-resource users: Type - A (76.3%)
This group hosts approximately 76 % of surveyed households. The vast majority of the farmers
in this group run up to three or more fields per year. These are extensive and pro-diversification
farm enterprises. Within this group, two distinct categories of non-resource users can be
distinguished, based on their assets and reasons for not using either fertilizers or manure. The
resource endowed but resistant change farmer (Type-A2) and the resource constrained farmer
(Type-A1).
Page | 97
Resource endowed but change resistant farmers: (Type - A2; 6.11%)
These are multipurpose and labour intensive farms. These farmers are engaged in several on
farm and off-farm income generating activities that allow the household to hire DAP/HTP and
labour mainly for land preparation and weeding respectively. The majority of these farmers are
cattle owners, but do not use manure. Type -A2 farmers, generate enough income (Table 3-4) to
invest in fertilizers in both villages but don’t. The group’s maize self-sufficiency levels, 6.07 in
Matsinho and 3.77 in Marera contribute to the situation, since farmers manage to meet their food
security targets and generate enough surplus for income generation. Apart from assumed high
fertilizer price, Type – A2 farmer’s poor investment in fertilizers is associated with their maize
self-sufficient which gives the farmers the perception that their land is productive enough for
them to meet their production targets. However, their food security is attained through a labour
intensive farming approach. In Macate, labour competition between fruit crops – citrus and
banana with cereal and vegetables are common.
Two subgroups can be distinguished among type – A2 farmers. The traditional investment
resistant farmers. The household head is older (>45 years old) and points at decreasing soil
fertility and poor yields as pre-conditions to invest in fertilisers. However, this is less likely to
happen while the farmer has the means to run his normal farming cycle. Moreover, their
tightened budget and farm management approach is also a hindering factor, since most of the
farm income, is reinvested back into keeping the laborious farm cycle running. In Macate,
maintaining the citrus and banana orchards or the year-long commercial vegetable production
consumes most of the household income. There is also the resource endowed but input aware
and willing farmers. These are increasingly younger farmers, who like the previous subgroup,
have the asset base to invest in external inputs and practices that would likely reduce their work
load and improve the whole system performance. However, they are poorly connected to
information networks, therefore, do not have enough technical support to identify and test new
practices. Labour qualification and technical assistance is critical for this group.
Page | 98
Resource constrained farmers (Type - A1)
Most of the surveyed smallholder farmers fall within this group. They do not use fertilizers or
manure. Despite a vast majority being aware that fertilizers and manure would most likely
improve their yields, they are resource limited. Limited access to knowledge networks and a poor
asset base are major constraints to engage in new technologies especially among the food-
insecure ones, i.e., Type –A1a farmers (45.8%). Farmers in this group had the lowest YLR, 66
and 62 kg/Hh farm active household member in Matsinho and Marera respectively. Maize self-
sufficient was the lowest among Type-A1a farmers (Table 3-4). For that reason, most of the
farmers in this group engage in part-time on-farm and off-farm jobs to cover their expenses
while waiting for the next harvest. There is however, a growing number of resource constrained
but maize self-sufficient (Type –A1b; 24.4%) young male-led households where the household
head normally works several part-time jobs during the season or has an off-farm job. Therefore,
he is on the farm only during pick times, such as land preparation and weeding. Normally his
off-farm income is invested in small amounts of OPV seeds and is also used to pay for additional
labour for land preparation and weeding and cover household expenses during the low food
periods, i.e., December to March in bad years as they are self-sufficient most of the years. These
farmers are also typical traditional legume growers (L2-Type), planting a wide range of legumes
across the season from where they get most of their on-farm income. Fresh and dry groundnuts,
cowpeas sold as fresh pods and dry grain, Bambara groundnuts and also tuber and root crops are
key income sources. However, the amounts sold are considerable lower compared to other
groups.
3.5 Discussion
3.5.1 Improving agricultural advisory services through typology tailored intervention
Clustering farmers into functional typologies is crucial for central Mozambique and most SSA
countries, where farming across groups is a highly context bounded activity in both space and
time as contrasting household livelihood strategies, biophysical and socioeconomic
circumstances shape farming system design and management. Therefore, clustering farmers into
small homogeneous typologies that are more reflective of farmers biophysical and
socioeconomic circumstances is critical to glean a better light in farmers’ decision making
Page | 99
process. The advantage of clustering farmers as a mean to aid agricultural intervention has
already been proved valid in East Africa (Pannell and Vanclay, 2011; Tittonell et al., 2010)
where it helped understand how farmers’ livelihood strategies influence soil fertility
management practices. In Malawi, southern Africa, Franke et al. (2014), built typologies to
understand how different farmer groups would benefit from sustainable intensification. In both
situations, clustering farmers in functional typologies helped identify key points of intervention.
However, finding better ways to use the data from this type of exercises as baseline to
interactively design personalised agricultural support programs that suits each group
circumstances has been the main challenge in the pipeline to trigger effective change of practice
in SSA.
Typology targeted agricultural intervention, i.e., focusing on each group unique circumstances
and developmental needs, is the closest approach to 1:1 peer tutoring defined in Maertens and
Barrett (2013) as the most effective learning approach. Personalizing agricultural intervention
and allowing farmers to experiment and learn by doing in their own fields is a more effective
way to trigger adoption (Cameron, 1999; Munshi, 2004) compared to the conventional
intervention methods which treats farmers as equal. However, although, splitting farmers into
small functional typologies is a pivotal tool to make learning and interaction more effective, it
also true that typologies are dynamics and smallholder farmers are always evolving. Therefore,
exploring the similarities and transition factores between groups is as important as having the
contrasts clearly set.
To validate typology tailored agricultural intervention, it is important to design and validate new
and easy to use farm categorization frameworks that are likely to be used by field staff when
performing simple appraisals of agricultural intervention needs at community level. Based on
that, a new framework to map intragroup diversity (figure 3-3) is proposed in this study. The
framework is flexible in time and can accommodate the seasonal changes in farmer’s livelihood
strategies and socioeconomic circumstances. This is because the typologies generated through
the framework clearly explore the intra-group diversity based on the premise that the
differentiation is crucial to understand how each groups socioeconomic status and livelihood
strategy are critical to build the foundations to an effective design of group-specific agricultural
Page | 100
interventions (Daloğlu et al., 2014). The ability of the framework in exploring farmers reasoning
to use or not a certain resource and the preconditions to engage in its use is critical to the effort
of validating farm typology tailored interventions. This approach is important in the sense that,
rigid characterisation of farmer groups as it currently happens tells us little about each
household’s productive capabilities or their innovative approaches to farming which is critical to
re-negotiate and effectively co-design group specific interventions.
3.5.2 Handling the key intensification constraints across contrasting farm typologies
Soil fertility and maize self-sufficiency
Despite poor soil fertility being vastly identified as one of the major reason behind lower yields
in SSA agriculture (Bekunda et al., 2010; Mwangi, 1996), farmers identified decreasing yields
and soil fertility loss as pre-conditions for investing in fertilizers and manure. This attitude is
based on the premise that their soils are “fertile enough” to produce what they require to be food
safe - the soil fertility and maize self-sufficiency perception. This trend was observed across both
districts, mainly among maize self-sufficient Type-A2 and resource constrained Type-A1b
farmers, who are typical extensive farmers. However, this is a counterproductive approach since
the household food security targets are achieved through a laborious extensive farming approach
(Lotter, 2015). Nevertheless, misconceptions about soil fertility among smallholder farmers in
SSA are common especially among maize self-sufficient farmers (Scoones, 2001) who tend to
relate poor yield with low rainfall pattern rather than soil fertility.
The challenge in triggering a change of practice among Type-A2 farmers is that, the overall farm
performance is assessed through seasonal maize self-sufficiency and income levels rather than
technical efficiencies. This is in line with findings from Vanclay (2004a) who pointed that
“sustainability” for the farmer is about being in the farm and keep up with it.. An analogy can be
made to Type-A2 farmers in central Mozambique, to whom everything revolves around being
able to produce enough grain to keep up with the household’s basic needs – food and income,
throughout the year independently of how technically inefficient and laborious is the process: the
maize self-sufficiency trap. Therefore, introducing technological changes into these group needs
to take into account that they are already self-sufficient and that only interventions that can prove
Page | 101
capable to lift their economic and social status without an abrupt disruption of their own
reasoning to do the things the way they do are more likely to be put to the test. However, as in all
adoption processes which are basically ignited by a change of mind-set, it will not happen
immediately (Maertens and Barrett, 2013). However, continuous and personalized agricultural
support services aiming at demonstrating short term returns from investment and help farmers
visualize long term benefits of each agreed management change will likely challenge farmers to
move out of their comfort zone. These proposed changes need to be imbibed in what farmers
already do and fit into their own development plan. A similar approach was proposed by Vanclay
(2004a) when studying farmers’ attitudes and the entry points for successful extension programs
in Australia. The author pointed that for most farmers, farming is a socio-cultural practice rather
than just a technical activity, and most farmers do the things they do because it has become a
way of living for them. Therefore, understanding and taking into account these ways of living
when designing agricultural interventions is the first step towards stimulating a conscious change
of practice, since no single sized technological package is likely to fit all farms.
Land access and labour availability perception trap
Independently of the household size, cultivating multiple fields in a season is a common risk
management practice implemented by most households. This is done in an attempt to take
advantage of existing biophysical and climatic gradients – perceived soil fertility, rainfall, soil
moisture regimes and land productivity (Table 3-4). Despite being a traditional risk management
practice management capabilities differ across groups, consequently the viability of the practice
as well. Nevertheless, extensification rather than “(Sustainable) intensification” as a way of
improving production and on farm income among smallholder farmers in environments where
access to land is not a limiting factor was reported before (Cameron, 1999; Leonardo et al.,
2015). In Central Mozambique, access to additional man power, draft animal power (DAP) and
capacity to hire “tractorized” power (HTP) among resource endowed Type-B, Type-A2 and
some resource constrained self-sufficient farmers (Type-A1a) is a key factor in determining the
household likelihood to cultivate more than one fields. The number of households with access to
HTP and DAP has increased significantly from 7% in 1996 to 39.9 % in 2010 (INE, 2011a). This
figure is expected to have increased to approximately 51.7% by 2015 assuming a 2.35% annual
increase since 2010. In the last 5-10 years, facilities to HTP and DAP hiring have been created
Page | 102
through the local agricultural directorates (SDAE) and private providers from where farmers can,
access those services at rates ranging from 11 and 22 USD for an hour of tractor harrowing and
ploughing. Approximately 18 USD for animal draft power to cover an area of 0.25 ha. Most
farmers are willing to pay those prices mainly because cultivating multiple fields is also a
measure of social status (Place, 2009) and a mean to achieve surplus (Leonardo et al., 2015).
However, this come at a price mainly for resource constrained Type-A1 farmers who are unable
to mobilize enough man power to weed.
Mechanization of farm activities as a driver to optimize land usage and productivity is a valid
option to the already maize self-sufficient and non-labour constrained Type-B and Type-A2
farmers. In these farms, improving farm income from maize based system is the main driver for
mechanization. However, mechanization needs to accommodate the whole spectrum of farm
activities (Twomlow et al., 2002), i.e., land preparation, using improved varieties, fertilizer use
and most importantly weeding so that the benefits from sowing enough land to achieve surplus
are not offset by farmer’s inability to weed their fields as it is happening. The use of small
multipurpose machines like the two-wheel tractors (2WTs), (Baudron et al., 2015b), for sowing
and weeding would help improve crop arrangement, increase yields prospects and reduce time
invested in those activities. For resource constrained Type-A1 farmers, the primary challenge is
to generate enough income to supplement their low maize self-sufficiency and start investing in
external inputs - fertilizer, manure and improved seeds to boost crop land productivity.
Because Type-A1 farmers are capital constrained to invest in inorganic fertilisers (Liverpool-
Tasie and Takeshima, 2013), legume intensification systems are a feasible alternative to increase
farm income while improving soil N availability. However, more legume friendly resource
allocations strategies and cropping systems need to be considered. To validate these systems,
special attention needs to be given to legume selection in order to find a suitable balance between
grain yield and biomass production since high harvest index legumes do less for soil fertility
(Vanlauwe and Giller, 2006). For the laborious Type-A2 farms and labour constrained Type-A1b
farmers, improving labour quality, i.e., labour-land ratio (LLR) and yield to labour productivity
ratio (YLR) through the specialization of farm labour, is critical before considering field size
Page | 103
increments (Table 3-4). However, this needs to happen in parallel with investment in external
inputs and production means – fertilizers, improved seeds.
3.5.3 Building from current farming systems design
Resource access and smart allocation strategies
As already found in several other studies across SSA, capital constraints are a key limiting factor
for investing in inorganic fertilizers among resource poor Type-A1 farmers (Ade Freeman and
Omiti, 2003; Chianu and Tsujii, 2005; Lambrecht et al., 2014). Nevertheless, our results showed
that, despite lower use there is preferential resource allocation as well. These results are in line
with findings from Vanlauwe and Giller (2006) who also argued that fertilizers are used in SSA
often where they offer a favourable value-to-cost ratio. In central Mozambique, fertilizers and
manure are applied into vegetables were they are believed to provide better returns.
Nevertheless, more still needs to be done mainly regarding fertilizer promotion strategies so that
they are attractive to farmers. The fact that responses from low fertilizer rates are being reported
across Africa (Kgonyane et al., 2013; Vanlauwe et al., 2011) is a good starting point. However,
the low rate response data contradicts current blanket fertilizer recommendations (Chianu et al.,
2012; Rware et al., 2014), meaning that fertilizer recommendations still need to be tailored to fit
crops, farm variability and environments, purchase capacity and field sizes in some cases.
Farming environments proved also to play a pivotal role in shaping local cropping systems’
design (Figura 3-1). Understanding the biophysical differences across farming environments has
allowed smallholder farmers to innovate in order to take advantage of existing gradients. The
crops, sequences and resource allocation preferences across farming environments is a proof, of
how farmers are using the existing variability to their advantage. Better maize arrangements,
were observed among farmers who used improved OPV’s and hybrid maize seed varieties, once
germination of this material is not questioned (Biemond et al., 2013). As a result, yields and
management practices were better, especially when DAP was used to sow and weed in Type B
and Type A2 farms. In these systems, yields as high as 1.4 t/ha were measured with unfertilized
cross pollinated (OPV) maize in the upland maize systems in Macate. In the wet lowland mixed
maize-vegetable systems, a 3.78 t/ha yield was measured in Matsinho, with the hybrid PAN67,
Page | 104
planted at 3.3 pls/m2 density in a field previously under fertilized fresh beans. Taking local
management tactics into account when planning agricultural intervention is important as it will
help generate information that is likely to be used by farmers in making informed decisions
regarding best fit resource allocation strategies in their system. For that to happen, a whole farm
household analysis that looks at the overall livelihood strategies, resource allocation across
activities, crop management responses and income returns between management options across
typologies is required since there are clear trade-offs to be considered at household and field
level. Therefore, facilitating access to information is critical to help farmers smartly decide what
management option better suits his farm and will most likely help him attain his seasonal and
long term development goals more efficiently.
3.6 Conclusions
Overall, the findings from this study point that, understanding farmer’s socioeconomic
circumstances, their resource use behaviour and the reasoning behind their management
decisions, is critical to involve them in the co-design of locally feasible and sustainable farming
systems. In order to achieve that, a socio-technical approach that captures the dynamics within
the “whole farm enterprise” and builds from the existing set up, i.e., the whole household
livelihood strategies and own development needs is required. This approach is favoured over a
linear and technical intervention approach focused sole on field level agronomic management
since the farm enterprise is far more complex than the field. Our analysis also suggests that, at
field level, there are circumstantially induced trade-offs, i.e., plant densities, N-fertilizers,
varieties and residues in the system that need to be considered to see how each one fits into each
farm enterprise management model and possible development pathways. Adding to that, the
existing diversity of farms, livelihood strategies and management decisions, reveals that there are
farm-specific management combinations of technologies that best fit each farmers’ bio-physical
and socioeconomic circumstances. Therefore, no generalized single-sized solution is likely to fit
all farms. This because all modifications to be introduced on the way the “farm household
enterprise” is managed need to be rooted on farmers’ circumstances and elected paths to pursue
the household food and income generation targets. Assuring this key principle is critical to
actively involve farmers in the identification and implementation of locally feasible
intensification pathways.
Page | 105
Acknowledgement
The authors would like to thank the SIMLESA project that through the Australian Centre for
International Agriculture Research (ACIAR) awarded me the scholarship to pursue my studies.
The Soil Health Program of the Alliance for Green Revolution in Africa (AGRA) through the
2011SHP007 Project and 201-Inv-FNI for helping fund this work. Extended thanks also goes to
Eng.º Domingos Dias (IIAM), Clemente Zivale and Ricardo Maquina for the help with the
surveys. Thanks also to Eng.º Edwin Pinho, Eng.º Chiwiro Filipe, Tec., Rosario and Tec. David
from the Gondola and Sussundenga Agricultural Directorate for their help in organizing and
running the model assisted participatory focus group discussions in Marera and Rotanda.
References
Ade Freeman, H., and Omiti, J. (2003). Fertilizer use in semi-arid areas of Kenya: analysis of
smallholder farmers' adoption behavior under liberalized markets. Nutrient Cycling in
Agroecosystems 66, 23-31.
Alene, A. D., and Coulibaly, O. (2009). The impact of agricultural research on productivity and
poverty in sub-Saharan Africa. Food Policy 34, 198-209.
Baudron, F., Sims, B., Justice, S., Kahan, D., Rose, R., Mkomwa, S., Kaumbutho, P., Sariah, J.,
Nazare, R., Moges, G., and Gérard, B. (2015a). Re-examining appropriate mechanization
in Eastern and Southern Africa: two-wheel tractors, conservation agriculture, and private
sector involvement. Food Security 7, 889-904.
Baudron, F., Thierfelder, C., Nyagumbo, I., and Gérard, B. (2015b). Where to Target
Conservation Agriculture for African Smallholders? How to Overcome Challenges
Associated with its Implementation? Experience from Eastern and Southern Africa.
Environments 2, 338-357.
Bekunda, M., Sanginga, N., and Woomer, P. L. (2010). Chapter Four - Restoring Soil Fertility in
Sub-Sahara Africa. In "Advances in Agronomy" (L. S. Donald, ed.), Vol. Volume 108,
pp. 183-236. Academic Press, Netherlands.
Bembridge, T. J. (1987). Agricultural extension in the less developed areas of Southern Africa.
Agricultural Administration and Extension 27, 245-265.
Page | 106
Biemond, P. C., Oguntade, O., Stomph, T.-J., Kumar, P. L., Termorshuizen, A., and Struik, P.
(2013). Health of farmer-saved maize seed in north-east Nigeria. European Journal of
Plant Pathology 137, 563-572.
Binns, T., Hill, T., and Nel, E. (1997). Learning from the people: Participatory rural appraisal,
geography and rural development in the ‘new’ South Africa. Applied Geography 17, 1-9.
Browne, M., Ortmann, G. F., and Hendriks, S. L. (2014). Household food security monitoring
and evaluation using a resilience indicator: an application of categorical principal
component analysis and simple sum of assets in five African countries. Agricultural
Economics Research, Policy and Practice in Southern Africa 53, 25-46.
Cameron, L. A. (1999). The Importance of Learning in the Adoption of High-Yielding Variety
Seeds. American Journal of Agricultural Economics 81, 83-94.
Chianu, J., Chianu, J., and Mairura, F. (2012). Mineral fertilizers in the farming systems of sub-
Saharan Africa. A review. Agronomy for Sustainable Development 32, 545-566.
Chianu, J. N., and Tsujii, H. (2005). Determinants of farmers' decision to adopt or not adopt
inorganic fertilizer in the savannas of northern Nigeria. Nutrient Cycling in
Agroecosystems 70, 293-301.
Comoé, H., and Siegrist, M. (2013). Relevant drivers of farmers' decision behavior regarding
their adaptation to climate change: a case study of two regions in Côte d'Ivoire.
Mitigation and Adaptation Strategies for Global Change, 1-21.
Cortez-Arriola, J., Rossing, W. A. H., Massiotti, R. D. A., Scholberg, J. M. S., Groot, J. C. J.,
and Tittonell, P. (2015). Leverages for on-farm innovation from farm typologies? An
illustration for family-based dairy farms in north-west Michoacán, Mexico. Agricultural
Systems 135, 66-76.
Cunguara, B., and Darnhofer, I. (2011). Assessing the impact of improved agricultural
technologies on household income in rural Mozambique. Food Policy 36, 378-390.
Daloğlu, I., Nassauer, J. I., Riolo, R. L., and Scavia, D. (2014). Development of a farmer
typology of agricultural conservation behavior in the American Corn Belt. Agricultural
Systems 129, 93-102.
Deaton, A., and Zaidi, S. (2002). Guidelines for constructing consumption aggregates for welfare
analysis. (T. W. Bank, ed.). The World Bank, Washington, USA.
Page | 107
Dorward, A., Anderson, S., Bernal, Y. N., Vera, E. S., Rushton, J., Pattison, J., and Paz, R.
(2009). Hanging in, stepping up and stepping out: livelihood aspirations and strategies of
the poor. Development in Practice 19, 240-247.
Dossa, L. H., Abdulkadir, A., Amadou, H., Sangare, S., and Schlecht, E. (2011a). Exploring the
diversity of urban and peri-urban agricultural systems in Sudano-Sahelian West Africa:
An attempt towards a regional typology. Landscape and Urban Planning 102, 197-206.
Franke, A. C., van den Brand, G. J., and Giller, K. E. (2014). Which farmers benefit most from
sustainable intensification? An ex-ante impact assessment of expanding grain legume
production in Malawi. European Journal of Agronomy 58, 28-38.
Giller, K. E., Corbeels, M., Nyamangara, J., Triomphe, B., Affholder, F., Scopel, E., and
Tittonell, P. (2011a). A research agenda to explore the role of conservation agriculture in
African smallholder farming systems. Field Crops Research 124, 468-472.
Giller, K. E., Tittonell, P., Rufino, M. C., van Wijk, M. T., Zingore, S., Mapfumo, P., Adjei-
Nsiah, S., Herrero, M., Chikowo, R., Corbeels, M., Rowe, E. C., Baijukya, F., Mwijage,
A., Smith, J., Yeboah, E., van der Burg, W. J., Sanogo, O. M., Misiko, M., de Ridder, N.,
Karanja, S., Kaizzi, C., K’ungu, J., Mwale, M., Nwaga, D., Pacini, C., and Vanlauwe, B.
(2011b). Communicating complexity: Integrated assessment of trade-offs concerning soil
fertility management within African farming systems to support innovation and
development. Agricultural Systems 104, 191-203.
Gouveia, D. H. G., and Azevedo, A. L. (1954). The Provisional Soil Map of Mozambique. In
"Second Inter-African Soil Conference", Vol. 1459-1468. ISRIC, Leopoldville.
Guillem, E. E., Barnes, A. P., Rounsevell, M. D. A., and Renwick, A. (2012). Refining
perception-based farmer typologies with the analysis of past census data. Journal of
Environmental Management 110, 226-235.
INE (2011). "Censo Agro-Pecuario 2009-2010: Resultados Preliminares." Instituto Nacional de
Estatistica(INE), Ministerios da Agricultura (MINAG), Maputo, Mozambique.
Keshavarz, M., and Karami, E. (2014). Farmers' decision-making process under drought. Journal
of Arid Environments 108, 43-56.
Kgonyane, M. C., Mariga, I. K., and Dimes, J. P. (2013). Low rates of nitrogen and phosphorus
as fertilizer options for maize production by smallholding farmers in drier regions of
South Africa. Research on Crops 14, 444-454.
Page | 108
Kostrowicki, J. (1976). Agricultural typology as a tool in planning spatial organization of
agriculture—IGU Commission on Agricultural Typology. Geoforum 7, 241-250.
Kostrowicki, J. (1977). Agricultural typology concept and method. Agricultural Systems 2, 33-
45.
Lambrecht, I., Vanlauwe, B., Merckx, R., and Maertens, M. (2014). Understanding the Process
of Agricultural Technology Adoption: Mineral Fertilizer in Eastern DR Congo. World
Development 59, 132-146.
Leonardo, W., van de Ven, G. J., Udo, H., Kanellopoulos, A., Sitoe, A., and Giller, K. (2015).
Labour not land constrains agricultural production and food self-sufficiency in maize-
based smallholder farming systems in Mozambique. Food Security, 1-18.
Linting, M., Meulman, J. J., Groenen, P. J. F., and Van der Kooij, A. J. (2007). Nonlinear
Principal Components Analysis: Introduction and Application. Psychological Methods
12, 336-358.
Liverpool-Tasie, L. S. O., and Takeshima, H. (2013). Input promotion within a complex
subsector: fertilizer in Nigeria. Agricultural Economics 44, 581-594.
Lotter, D. (2015). Facing food insecurity in Africa: Why, after 30 years of work in organic
agriculture, I am promoting the use of synthetic fertilizers and herbicides in small-scale
staple crop production. Agriculture and Human Values 32, 111-118.
Lukanu, G., Green, M., and Worth, S. (2007). The influence of smallholder labour demand on
cultivation of cash crops in northern Mozambique. Development Southern Africa 24, 553-
573.
Maertens, A., and Barrett, C. B. (2013). Measuring Social Networks' Effects on Agricultural
Technology Adoption. American Journal of Agricultural Economics 95, 353-359.
Munshi, K. (2004). Social learning in a heterogeneous population: technology diffusion in the
Indian Green Revolution. Journal of Development Economics 73, 185-213.
Mwangi, W. M. (1996). Low use of fertilizers and low productivity in sub-Saharan Africa.
Nutrient Cycling in Agroecosystems 47, 135-147.
Nainggolan, D., Termansen, M., Reed, M., Cebollero, E., and Hubacek, K. (2013). Farmer
typology, future scenarios and the implications for ecosystem service provision: a case
study from south-eastern Spain. Regional Environmental Change 13, 601-614.
Page | 109
Nuss, E. T., and Tanumihardjo, S. A. (2011). Quality Protein Maize for Africa: Closing the
Protein Inadequacy Gap in Vulnerable Populations. Advances in Nutrition: An
International Review Journal 2, 217-224.
Pannell, D., and Vanclay, F. (2011). Changing Land Management: Adoption of New Practices
by Rural Landholders. CSIRO Publishing, Melbourne, Australia.
Place, F. (2009). Land Tenure and Agricultural Productivity in Africa: A Comparative Analysis
of the Economics Literature and Recent Policy Strategies and Reforms. World
Development 37, 1326-1336.
Ranum, P., Peña-Rosas, J. P., and Garcia-Casal, M. N. (2014). Global maize production,
utilization, and consumption. Annals of the New York Academy of Sciences 1312, 105-
112.
Robinson, L. W., Ericksen, P. J., Chesterman, S., and Worden, J. S. (2015). Sustainable
intensification in drylands: What resilience and vulnerability can tell us. Agricultural
Systems 135, 133-140.
Rouabhi, A., Hafsi, M., and Kebiche, M. (2014). Assessment of the farming transformation in a
rural region of Sétif province in Algeria. New Mediterranean 13, 38-46.
Rounsevell, M. D. A., Robinson, D. T., and Murray-Rust, D. (2012). From actors to agents in
socio-ecological systems models. Philosophical Transactions of the Royal Society of
Biological Sciences 367, 259-269.
Rware, H., Wairegi, L., Oduor, G., Macharia, M., Romney, D., Tarfa, B. D., de Maria, R., Ley,
G., Tetteh, F., Makumba, W., Dicko, M., Nouri, M., Leon, N. N., Gondwe, B., Kaizzi,
K., Negash, D., Kibunja, C., Alawy, A., Wartmann, C., and Rutsimba, E. (2014).
Assessing the Potential to Change Stakeholders Knowledge and Practices on Fertilizer
Recommendations in Africa. Agricultural Sciences 5, 1384-1391.
Scoones, I. (2001). "Dynamics and diversity: soil fertility and farming livelihoods in Africa: case
studies from Ethiopia, Mali, and Zimbabwe," Earthscan Publications Ltd, London.
Thierfelder, C., Mombeyarara, T., Mango, N., and Rusinamhodzi, L. (2013). Integration of
conservation agriculture in smallholder farming systems of southern Africa: identification
of key entry points. International Journal of Agricultural Sustainability 11, 317-330.
Page | 110
Tittonell, P., and Giller, K. E. (2013). When yield gaps are poverty traps: The paradigm of
ecological intensification in African smallholder agriculture. Field Crops Research 143,
76-90.
Tittonell, P., Muriuki, A., Shepherd, K. D., Mugendi, D., Kaizzi, K. C., Okeyo, J., Verchot, L.,
Coe, R., and Vanlauwe, B. (2010). The diversity of rural livelihoods and their influence
on soil fertility in agricultural systems of East Africa – A typology of smallholder farms.
Agricultural Systems 103, 83-97.
Twomlow, S., O'Neill, D., Sims, B., Ellis-Jones, J., and Jafry, T. (2002). RD—Rural
Development: An Engineering Perspective on Sustainable Smallholder Farming in
Developing Countries. Biosystems Engineering 81, 355-362.
Valbuena, D., Verburg, P. H., and Bregt, A. K. (2008). A method to define a typology for agent-
based analysis in regional land-use research. Agriculture, Ecosystems and Environment
128, 27-36.
Vanclay, F. (2004). Social principles for agricultural extension to assist in the promotion of
natural resource management. Australian Journal of Experimental Agriculture 44, 213–
222.
Vanlauwe, B., and Giller, K. E. (2006). Popular myths around soil fertility management in sub-
Saharan Africa. Agriculture, Ecosystems and Environment 116, 34-46.
Vanlauwe, B., Kihara, J., Chivenge, P., Pypers, P., Coe, R., and Six, J. (2011). Agronomic use
efficiency of N fertilizer in maize-based systems in sub-Saharan Africa within the context
of integrated soil fertility management. Plant and Soil 339, 35-50.
Wall, P. C. (2007). Tailoring Conservation Agriculture to the Needs of Small Farmers in
Developing Countries. Journal of Crop Improvement 19, 137-155.
Zimmerer, K. S., Carney, J. A., and Vanek, S. J. (2015). Sustainable smallholder intensification
in global change? Pivotal spatial interactions, gendered livelihoods, and agrobiodiversity.
Current Opinion in Environmental Sustainability 14, 49-60.
Page | 111
CHAPTER 4:
Best-fit Residue Management in Rain Fed N-Deprived Systems
This chapter was submitted for publication in the Field Crop Research Journal and the
original paper format was kept in the chapter:
Nhantumbo, N., Mortlock, M., Nyagumbo, I., Dimes, J and Rodriguez (nd) Residue and fertilizer
management strategies for the rain fed, N-deprived maize-legume cropping systems in
Central Mozambique.
Nhantumbo, N., Mortlock, M., Nyagumbo, I., Dimes, J and Rodriguez, D (2015) Improving
resource allocation in nitrogen constrained systems: acknowledging within and cross-
farm variability effect on yield and NUE. “In proceedings from the 5th International
Symposium for Farming Systems Design: Multi-functional farming systems in a
changing world”. Edited by ESA and Agropolis International, pp 491-492, 7-10
September, Montpellier, France
Page | 112
Residue and fertilizer management strategies for the rain fed, N-deprived maize-legume
cropping systems in Central Mozambique
Nascimento Nhantumbo1,2, Miranda Mortlock3, Isaiah Nyagumbo4, Daniel Rodriguez1
1 Queensland Alliance for Agriculture and Food Innovation (QAAFI), The University of Queensland PO
Box 102 Toowoomba Qld (4350) Australia; 2 Divisão de Agricultura (DivAG), Instituto Superior
Politecnico de Manica; Corresponding author: [email protected]; 3School of Agriculture and Food Science (SAFS), the University of Queensland; 4CIMMYT-Harare;
Abstract
Improving crop residues and nitrogen fertiliser management strategies, is critical to increase resource
productivity and yields in Sub-Saharan Africa’s N-deprived maize-legume systems. In this study we
quantified the immediate benefits and returns from the allocation of carbon rich residues and nitrogen
fertilisers between maize and cowpea crops, grown across soils of contrasting level of fertility and
moisture regime in Mozambique. Three contrasting soils: a fine sandy soil (fSa), a wetter sandy loam soil
(SaL), and a sandy clay loam soil (cSaCL) were used. The experiment was carried over two seasons. In
the first season maize and cowpea were sown under two levels of crop residue cover, i.e. 0 and 6t/ha;
three levels of N supply i.e. 0, 23 and 92 kg N/ha, representing unfertilized (ND), limited N-applications
(LNA), and high N-application (HNA) systems. Giant thatching grass was used as mulching. In the
second season a bio-assay maize crop was planted to evaluate the short term residual effects from
previous season treatments. Changes in yield, water use efficiency (WUE), nitrogen responses, nitrogen
use efficiency (NUE), and short term residual effects were analysed. Results showed that the application
of high C/N residues as mulches (i) improved cowpea yield and biomass production at all three sites; and
(ii) reduced the yield of unfertilized maize at all sites but effects were only significant on the fine sandy
soil, and sandy clay loam soil. Residue application in maize only proved beneficial under the HNA.
Residue application significantly increased WUE, NUE and nitrogen uptake efficiency (NUpE) across
residue levels in cowpea; though these efficiency indices were reduced on the maize crop. A higher soil
moisture regime in the wetter SaL soil led to higher NUE, while increased N supply reduced NUE and
increased WUE across all the sites. Results from the bioassay maize crop showed a 23.2%, 27.7% and
56.7% lower maize yield, NUpE and NUE, from plots previously under maize when residues were
applied. Results from this study suggest that there are crop and site specific responses and trade-offs that
need to be considered when promoting high C/N residues and fertilizers among resource poor smallholder
farmers of Mozambique.
Key words: Crop residues, Water use efficiency (WUE), Nitrogen use efficiency (NUE), maize-legume
systems, conservation agriculture
Page | 113
4.1 Introduction
Benefits from the adoption of conservation agriculture principles in Africa (Ngwira et al., 2013;
Thierfelder et al., 2012; Thierfelder et al., 2015) are contestable, reason current adoption levels
are low compared to the rest of the world (Baudron et al., 2015c; Giller et al., 2009; Govaerts et
al., 2009). In particular, the use of carbon rich mulches that are known to interact negatively with
plant N availability when used in N-deprived cereal cropping systems (Baudron et al., 2015c;
Giller et al., 2011a). Furthermore, data sets showing benefits from the use of C rich mulches are
usually from trials where considerable amounts of nitrogen fertilizer have been used (Ngwira et
al., 2013; Thierfelder et al., 2012; Vanlauwe et al., 2014); which does not compare to common
farmer practice. High prices, limited cash, and unreliable fertilizer responses resultant from poor
recommendations (Crawford et al., 2006; Enyong et al., 1999), have kept fertilizer application
rates in Sub Saharan Africa (SSA) extremely low, ca. 9-13 kg fertiliser/ha (Chianu et al., 2012).
This is much lower than the 50 kg fertiliser/ha target expected to be achieved by 2015 in the
Abuja declaration (Kintché et al., 2015).
An alternative to inorganic N-fertilizers in Africa, are rotations with legumes. However, legumes
are known to underperform in most SSA farming environments, producing lower yields, lower
shoot biomass and low N fixation (Giller et al., 2009). The availability of maize stubble is also
limited in regions were livestock freely roam open fields (Erenstein et al., 2015). In central
Mozambique the scarcity of residues is worsened by the presence of termites that feed on any
available source of carbon (Nyagumbo et al., 2015). As a result, incorporating locally available
grasses as an alternative to the in demand maize stubble maize is a common practice mainly in
cattle and termite dominated areas of Central Mozambique, thought poorly studied. African
grasses like Hyparrhenia rufa, Andropogon gayanus, Panicum maximum, Eragrostis ciliaris
(L.), and Penissetum glaucum are widely available in Central Mozambique and across southern
Africa where they colonize maize and sorghum fields (Vibrans et al., 2014) during undamaged
fallows.
The incorporation of alternative grasses and shrubs in CA based systems in Africa to build soil
organic matter (SOM) have been recommended in west African agroecosystems (Lahmar et al.,
Page | 114
2012). This is an important step towards tailoring CA in SSA since grasses and shrubs are
responsible for the increase in fire-mediated nitrogen losses in the tropics (Kugbe et al., 2015;
Rossiter-Rachor et al., 2008). Nevertheless, the incorporation of these carbon rich grasses in CA
based systems have detrimental effect on N-availability that still needs to be understood specially
in SSA N-deprived agroecosystems. Therefore, identifying optimum strategies to use and
allocate the available amounts of these C rich mulches, is of pivotal importance for the
sustainable intensification of agricultural production in smallholder farmers’ managed under
conservation agriculture (Tittonell et al., 2015).
The objective of this study was therefore, to quantify the short term returns and trade-offs from
the alternative allocation of carbon rich residues and the limited availability of mineral nitrogen
fertilisers, between maize and cowpea crops, grown on soils having contrasting level of moisture
regime. Responses were analysed in terms of water use (WU), nitrogen responses (NU) and
resource productivity – water use efficiency (WUE) and nitrogen use efficiency (NUE). This
study adds value to the current efforts to improve resource management in smallholder farming
systems in the SSA. Here, we hypothesize that, the best response from high C/N ratio mulches
and residues should be expected from the legumes phase of the rotation; and that limited
availabilities of nitrogen fertilisers should be primarily used to improve yields on the residue free
maize crop.
4.2. Materials and Methods
4.2.1 Field Experimentation
Field experiments were conducted at three soils of contrasting soil texture and moisture regime.
The soils were a red coarse sandy clay loam (cSaCL) Ferralsol i.e. a dry upland soil most
dominant in the region; a dark sandy loam soil (SaL) Gleysol, i.e. a wetter lowland soil, medium
to high fertility, located across the valleys and river streams; and a fine sandy soil (fSa)
Arenosol, mostly identified as dry and of low fertility soil. Plant available water content
(PAWCs) at sowing was 93 mm, 41.9 mm and 11.7 mm for the SaL, cSaCL and fSa respectively.
Initial nitrogen content was 23.6, 33.4 and 66.6 kg NO3-N/ha, in the SaL, fSa, and cSaCL soils,
Page | 115
to 1.20 m depth respectively (Table 1-1). Initial N (NO3-N and NH4-N) was determined using
NIR spectroscopy at ICRAF - Cropnuts Laboratory in Nairobi. Seasonal N-mineralization and
immobilization was monitored as extractable NO3-N at harvest of season 1 and 2 (Figure 1-3).
The colorimetric method based on Merckoquant nitrate strips - Merck KGaA from Germany was
used (Cossani et al., 2012). The SaL and fSa sites were established on-farm, while the cSaCL
was located at the ISPM research station. At each site, a fully randomized split plot factorial
(2x3x3) experiment was established in a 4x3 arrangement layout. Treatments included three
factors: crop type i.e., sole maize and sole cowpea; residue application at two levels (0 and 6
t/ha); and N-fertilization at three levels, an unfertilized N-deficient system (ND); a low N-
application rate (LNA, i.e. 23 Kg N/ha); and a high N application rate (HNA, 92 Kg N/ha). The
combination of these three factors resulted in 12 treatment replicated four times in each site. Plot
sizes were 12x4m in the first season (2013-14), and each plot was divided into two 6x4m in
2014-15 season to accommodate a bioassay maize crop planted without fertilizer application nor
residue to determine 1st season’s treatment effects. The other half was kept as in the first season.
Residues from the previous seasons where kept in the field and not incorporated in the bioassay.
The trial was dry sown between the 19th and 21st of November, in both seasons. Weeds were
brown killed using 6 l/ha glyphosate prior to sowing. Three in-crop weeding’s were conducted in
the maize, while one weeding, just before flowering, was required for the cowpea. All the plots
received 60 kg P/ha at sowing as single superphosphate, and the legume was inoculated prior to
sowing. Thatching grass straw having a 54:1 C/N ratio, was applied as mulch immediately after
sowing. The N-fertilizer was applied as Urea in a 50:50 rate at sowing and four weeks after
sowing. A short duration and self-pollinated erect cowpea variety IT16 was planted at 22.2 pl/m2
(0.45x0.20 m). A medium duration maize hybrid from PANNAR (PAN 67) was planted at 0.75
m between rows and at 0.25 m between plants, i.e. 5.3 pl/m2. An automatic weather station was
used to record daily values of rainfall, maximum and minimum temperatures, and solar radiation.
All trials were within a 2 km distance from the weather station.
Page | 116
4.2.2 Measurements
Soil water content (SWC) was determined gravimetrically in each plot at sowing and
physiological maturity at 0-15, 15-30, 30-60, 60-90 and 90-120 cm depth intervals using a soil
corer. Crop water use (WU) was estimated as the difference between soil water content at sowing
and physiological maturity, plus in crop rainfall (Aggarwal and Sinha, 1983; Cossani et al.,
2012). Losses by runoff were minimal, as the site was flat, though the method ignores water
losses by deep drainage. Plant available water content (PAWC) at each site was calculated by
correcting the SWC by a crop lower limit (CLL) (Dalgliesh and Foale, 1998; Pala et al., 2007).
Yield and dry matter production: dry matter production was determined at physiological maturity
by sampling two consecutive rows of 3 m length from the centre of each plot, i.e. 2.7 and 4.5 m2
in cowpea and maize, respectively. The same procedure was used to determine grain yield.
Resource productivity: the mass difference method was used to measure nitrogen use (NU)
(Montemurro et al., 2006). For the grain and straw, samples were collected at physiological
maturity, and total nitrogen content determined using Kjeldhal method. Nitrogen use efficiency
(NUE) was determined as the ratio between grain yield (Gw, kg/ha) and nitrogen supply (Ns, kg
N/ha), where Ns was the sum of applied inorganic N-fertilizer, aboveground plant N from the
control plots where no N fertilisers was added (Nt, kg N/ha), and the residual postharvest
mineral-N in the soil (Nsph, kg N/ha) (Huggins and Pan, 1993; Moll et al., 1982). Two NUE
components were also calculated. Nitrogen uptake efficiency (NUpE), which was determined as
Nt/Ns; and nitrogen utilization efficiency (NUtE), calculated as Gw/Nt. (Giambalvo et al., 2010).
Soil nitrogen availability (Nav) was estimated as the sum of aboveground N and postharvest
mineral N calculated as an approximation from soil nitrate levels, (Aggarwal et al., 1997; Bakht
et al., 2009). Aboveground nitrogen partitioning between stover and grain was calculated by
multiplying the percentage of nitrogen in the stover or grain, by the respective amount of
biomass of stover or grain produced. Water use efficiency (WUE, kg/ha mm-1), was calculated as
the ratio between grain yield and WU assuming 100 mm losses in soil evapotranspiration during
early crop establishment (Dalgliesh and Foale, 1998; French and Schultz, 1984).
Residual responses from residue and fertilizer treatments were calculated as the relative
difference between the yield of the maize crop in the bioassay maize crop grown in plots
previously under maize and the bioassay yield from plots previously under cowpea as described
Page | 117
in Gentry et al. (2013). The difference between those yields was defined as yield penalty from
growing a continuous maize (CMYP). Similar calculations were made for WUE, NUE, NUpE
and NUtE. Positive values indicate benefits from the cowpea-maize sequences, while negative
values indicate benefits from the continuous maize systems.
4.2.3 Statistical analysis of field experiment data sets
Analyse of variance (ANOVA) were performed using a generalized fixed linear model (RELM)
in GenStat 16 which was also used for trial design. ANOVA was conducted to test main factors
effect (residues, fertilizer, site) and their interactions on tested crops. Response variables where,
yield, water and N productivity (CWU, WUE, NUE, NUpE, NUtE). ANOVA was performed in
two stages. In the first season to test whether cowpea and maize responded differently to residue
and fertilizer application across tested sites. In the cases where different crop responses where
observed, further analysis where performed separately for each crop. Means were separated by
least significant differences (LSD) where significant differences where observed. Furthermore,
correlation analysis where conducted to explore the strength and significance of the relation
between the parameters of interest. In the second season, results from the bioassay maize crop
where analysed to measure treatment effects and identify the best fit resource residue and N-
fertilizer application across tested crops and sites. In the second season, because only 5 out of 12
treatments, i.e., 3 cowpea-maize and 2 maize-maize sequences were harvested in the drier fSa
soil residual treatment effects are only reported for the cSaCL and SaL sites where all treatments
were harvested.
4.3. Results
4.3.1 Crop growing conditions – rainfall distribution
Rainfall distribution throughout the growing season varied largely between months and was
better distributed in the first season, during the first season the trials experienced almost no
rainfall during the first 21 days after sowing (Figure 4-1). The amount of pre-sowing rainfall
Page | 118
differed between the two seasons though, during the 2013-14 season in-crop rainfall was 484 and
559 mm for the cowpea and maize, respectively. During the 2014-15 season in-crop rainfall on
the maize bioassay was 594 mm. During the first season crops experienced a short periods of
water stress at grain filling in cowpea and during early flowering of the maize. During the second
season rainfall distribution was much poorer, with 72.8% (432.9mm) falling between 9th
December to 9th January. During this season the sites experienced dry spells around flowering
and during grain filling of the maize
Figure 4-1 Rainfall distribution, daily maximum, minimum temperature and radiation from
1st September 2013 to 4th April 2014 in the 2013-14 cropping season (left) and 2014-15
cropping season (right). S is sowing date, (F) flowering and (H) harvest
Page | 119
4.3.2 Maize and cowpea yield response to residue and fertilizer application across sites
Maize and cowpea yield responses to the addition of the high C/N ratio residues differed across
sites and levels of N-applied. For cowpea, the application of high C/N ratio residues significantly
increased yields by 45.4% and 41.7% in the unfertilized (ND) cowpea of the high starting N
cSaCL and wetter SaL sites. On the drier fine sandy soil (fSa), positive responses to residue
application were also measured though the differences were not significant.
In maize, the response to residue application differed across levels of N-application (R-level x N-
level, p<0.05), and site (site, p<0.05, table 4-1). The application of high C/N ratio residues led to
significant yield losses in unfertilized (ND) maize plots, except on the wetter SaL site (Figure 4-
2). In the high starting N cSaCL site, residue application significantly reduced the yield of
unfertilized (ND) maize (33.8%), and significantly increased maize yield (44.7%) under HNA.
No significant yield increases were measured in the LNA plots (Figure 4-2). At the drier fSa site,
residue application significantly reduced the yield of the unfertilized (ND) maize (43.5%). In
contrast to the measured yield loss after the application of high C/N residues in unfertilized
maize, maize responded positively to the addition of the high C/N ratio residue for the at HNA
treatment, except at the drier sandy soil (fSa).
Cowpea and maize responded differently to nitrogen fertilisation (N-levels, p<0.05). Maize yield
responded positively to inorganic N-applications, while the yield of cowpea was reduced by N-
applications above 23 kg N/ha. The response to N-application also differed between sites (N-
level x Site, p<0.05). The largest responses were observed at the wetter SaL site. Individual
crops responses to N-applications (Figure 4-2), showed that maize’s responded to inorganic N-
applications at the high starting N but water limited cSaCL site, particularly when residues were
applied.
Page | 120
Figure 4-2 Crop response to residue application across different N-fertilization levels on
three contrasting soil types from Central Mozambique. ND: Unfertilized N-deficient system
(0 kg N/ha), LNA: low nitrogen application system (23 kg N/ha), HNA: high N-application
(92 kg N/ha). S – significant differences between residue application levels. Different letters
across N-levels for same R-level indicated significant differences at p<0.05 significance.
Caps lock – separate treatments without residues. Small letter separate situations with
residues
Table 4-1 Significance levels for the main factors (Crop, R-levels, N-levels, Site) and their interactions effects on yield, stover, crop
water use (CWU), water use efficiency (WUE), nitrogen use efficiency (NUE), nitrogen uptake efficiency (NUpE), nitrogen
utilization efficiency (NUtE) and N-partitioning into shoot N (Nsh) and Grain N (Ng), at physiologic maturity
Fixed term Yield Stover CWU WUEyld Shoot N Grain N Total N-uptake NUpE NUtE AEN
Crop ***
*** *** *** ** *** *** ***
R_level
**
**
N_level *** *** *** *** ** *** ***
Site *** *** *** *** *** *** *** *** ***
Crop x R_level
**
**
Crop x N_level ***
* ** * **
R_level x N_level *
*
*
Crop x Site *** *** *** *** *** *** *** ***
R_level x Site
*
N_level x Site ***
** *** *** * ***
Crop x R_level x N_level ** * **
** *
Crop x R_level x Site
*
*
Crop x N_level x Site ** **
**
** ***
R_level x N_level x Site
*
Crop x R_level x N_level x Site * *
*
* Significant at 0.05 probability level
** Significant at 0.01 probability level
*** Significant at 0.001 probability level
4.3.3 Resource Utilization
Crop water use (CWU)
Estimated crop water use (CWU) varied from a minimum of 384.8 mm to a maximum of 572.3
mm across crops and sites. Differences in WU across crops were due to different crop durations
and rainfall distribution, which affected the amount of in crop rainfall. However, site differences
effect on CWU were related to contrasting starting soil water contents (SWC) (R2 = 0.70,
p<0.05) and starting plant available water content (PAWC) (R2 = 0.59, p<0.05). In maize,
residue application significantly increased CWU.
Significant differences in CWU across sites were measured at both flowering and physiological
maturity. The cSaCL site, showed the highest value of CWU at flowering for both crops, and
differed significantly from the other sites. For both crops, CWU was different across levels of N-
application at all sites (N-level x Site, p<0.05). Figure 4.3 shows that the variation in CWU was
relative small compared to the variation in the yields of maize and cowpea. Figure 4-3 also
shows that in maize the maximum WUE was 13.8 kg grain/ha.mm, corresponding to the highest
levels of N supply under the use of residues at the SaL site. This value represented 77% of the
achievable WUE simulated by the APSIM model (17.9 kg grain/ha.mm) (Figure 4-3a). Figure 4b
shows that in cowpea the maximum WUE was 4.9 kg grain/ha.mm, corresponding to the SaL site
and low levels of N supply, with and without the use of residues. This value represented ca. 63%
of the achievable WUE calculated by the APSIM model (Figure 4-3b).
Figure 4-3 Maize and cowpea grain yield as a function of crop water use from emergence to maturity measured across sites. The
continuous line represents the attainable water use efficiency of 20 kg grain/ha.mm (a) and 10 kg grain/ha.mm (b), for maize and cowpea
respectively. The dotted lines represent the maximum actual WUE for each crop across the tested environments. Theoretical WUE, based
on APSIM model simulations for this agroecological zone are of 17.9 kg grain/ha.mm for maize and 7.74 kg grain/ha.mm for cowpea,
respectively.
Page | 124
Water use efficiency
Residue application significantly increased cowpea WUEyield, (Table 4-1). For maize, WUEyield
responses to residue application differed across levels of N-application (R-level * N-levels, p<
0.05). Nitrogen application significantly increased WUEyield (Table 10). For maize, the highest
WUEyield was measured under HNA and differed significantly from the other N-levels. For
cowpea, the highest WUEyield was achieved under LNA at all sites. Individually (Sites, p<0.05)
and through its interaction with N-levels (N-levels x sites, p<0.05), had the largest effect on
WUEyield. Crop WUEyield at cSaCL (9.16 kg grain/ha.mm) and SaL (8.67 kg grain/ha.mm) were
significantly higher than the values at the drier fSa site, which registered the lowest WUEyield
(Table 4-1). Analysis of individual treatment responses showed that the highest WUEyield for
maize was measured at the SaL site, in residue applied HNA (13.8 kg grain/ha.mm) plots.
However, this was not significantly different from the measured WUEyield in the cSaCL for the
same treatment. In cowpea the highest WUEyield, was observed on the sandy loam soil, i.e. 4.93
kg grain/ha.mm and differed significantly from the other two sites. Different effects from N-
application across sites were measured on WUEyield (N-level x Sites, p<0.05). In maize (Figure
4-4a), N-application increased WUEyield by an overall 76% across N-levels in the wetter SaL site.
However, no significant differences in WUEyield across N-application levels were observed for
the cSaCL and dry fSa sites. No intra-site effects were observed in cowpea (Figure 4-4b), but
clear inter-site differences in WUEyield were observed across N-levels.
Page | 125
Figure 4-4 Relationship between crop WUEyield and N-fertilizer application across sites (N-level x Site
interaction). N-levels: ND: N-deficient system (0 kg N/ha), LNA: low N-application system (23 kg N/ha)
and HNA: Non-limiting N (92 kg N/ha). Soil textures: fSa – fine sand soil, SaL – sandy loam soil and
cSaCL - coarse sandy clay loam soil. The vertical line in each plot represents the LSD value for WUEyield
comparison between N-levels for each site at p<0.05. N-levels spaced higher than the LSD bar for the
same site are significantly different.
Page | 126
Nitrogen Use Efficiency (NUE)
Residue application and N-levels differently affected NUE across crops and sites. No significant
differences were observed in maize. In cowpea, an overall 14.7% increase in NUE was
calculated for the treatments were high C/N residues were applied (Table 1). Across all sites, the
response of cowpea NUE to the use of residues was highest at the SaL site, where an overall
27.1% NUE increase was observed (R-level x Site, P<0.05). N and sites had the largest impact
on crop NUE. NUE was significantly reduced by N-applications above LNA, for both crops and
sites. For maize, 21.3 kg grain/kg N-supplied were produced under LNA; this was significantly
higher than the value under the HNA (16.2 kg grain/kg N-supplied). For cowpea the highest
NUE was 7.35 kg grain/kg N-supplied, which was observed at the LNA treatment. However,
NUE responses to N-applications differed across sites (N-level x Site, P<0.05). For maize, the
largest NUE loss from increased N-application was observed at the cSaCL site (Figure 4-5a). For
cowpea, variations in NUE as a function of the level of N-application were highest at the SaL
site. Moisture and starting N driven changes in NUE were observed. As we move from the drier
fSa to the high WHC cSaCL and wetter SaL sites NUE increased. Nevertheless, at high N-levels,
the starting high soil N-status negatively affected the values of NUE. As a result, no significant
differences between maize NUE were observed between cSaCL and SaL sites at HNA for maize
(Figure 4-5a). For cowpea (Figure 4-5b), the high starting N in cSaCL, led to poor NUE as well.
Page | 127
Relationship between NUE, NUpE and NUtE
NUE was highly related to its components in both crops. However, the magnitude the
contribution of each component to explain the variation in NUE differed between crops and sites.
In maize changes in NUE (Figure 4-6a) were primarily driven by changes in NUpE (R2 = 0.97, p
< 0.05) rather than NUtE (R2 = 0.25, p < 0.05). However, because NUE, NUpE and NUtE
differed significantly across sites (Table 4-1) the relationships between NUE and its components
also differed across sites. A split site analysis showed that for the cSaCL, NUE was sensitive to
variations in both NUpE (R2=0.72) and NUtE (R2=0.71), while at the wetter SaL, variations in
NUE were better explained by variations in NUtE, (R2=0.51). At the drier and N limited fSa site,
NUE was primarily driven by NUpE (R2=0.64), this compares with NUtE (R2=0.47). In cowpea,
overall NUE (Figure 4-6) shared a strong relation with both NUpE (R2 = 0.79, p < 0.05) and
Figure 4-5 Relationship between crops NUE and N-fertilizer application across sites. N-levels: ND: N-
deficient system (0 kg N/ha), LNA: low N-application system (23 kg N/ha) and HNA: Non-limiting N (92
kg N/ha). Soil textures: fSa – fine sand soil, SaL – sandy loam soil and cSaCL - coarse sandy clay loam
soil. Continuous line in each plot represent the LSD value for NUE comparison between N-levels for each
site at p<0.05. N-levels spaced higher than the LSD bar for the same site are significantly different.
Page | 128
NUtE (R2 = 0.82, p < 0.05). However, in cowpea NUE was more sensitive to the N-translocation
into grain (NUtE), particularly at high levels of N supply (HNA). A cross site analysis showed
that, at the SaL site, NUE was better related to NUpE (R2 = 0.58) than NUtE (R2 = 0.23). While
at the cSaCL site, NUE, was better related to NUtE (R2 = 0.94), and had almost no relation with
NUpE. At the drier, fSa, NUE was strongly related to both NUpE (R2 = 0.67) and NUtE (R2 =
0.83).
Figure 4-6 Relationship between NUE and its components observed across sites of contrasting moisture
regime. In maize as NUpE was the primary driver of NUE (a.1), while in cowpea (b.1, b.2) a, NUtE
appeared to be major driver of NUE changes. Shape identification: ∆ - Sandy Clay loam (cSaCL); O Fine
sandy (fSa) and ◊ Sandy loam soil (SaL). Empty and filled shape indicate residue application levels
Page | 129
Resource allocation trade-offs: yield and resource productivity
A nitrogen driven trade-off between NUE and WUE was identified (Figure 4-7). Increases in N-
supply tended to increase WUE in detriment of NUE. However, the magnitude of this trade-off -
positive or negative - depended on the crops differential response to N-application (Figure 4-7a,
maize, Figure 4-7b cowpea) and residue treatments. As of residues, a crop driven residue
response modified trade-off between NUE and WUE was observed in both maize and cowpea.
Results also showed the existence of a cross-site interactive relationship between NUE and
WUE, which explained most of the variation in NUE. Here, improved soil water, i.e., a move
from the drier fSa to the relatively wetter SaL site, increased NUE.
Figure 4-7 Nitrogen driven trade-offs between NUE-WUE. NUE and WUE were averaged across sites
and residue application levels for maize (a) and cowpea (b). Bellow: Maize (c) and cowpea (d) trade-
offs as affected by residue application levels. NUE and WUE for both crops were averaged across sites.
Residue application levels: R- and R+ indicate residue free and residue applied systems.
Page | 130
4.3.4. Short-term effects from residues and N-fertilizer allocation strategies
Results from the bioassay crop, demonstrated that yield and resource productivity of the
subsequent maize crop were found to differ between plots previously under cowpea and maize
(Table 4-2). Of all tested factors, site differences had the largest effects on the subsequent crop
performance. In general, higher yields and resource productivity where measured from plots
previously planted to cowpea under both residue and fertilizer application. Resource allocation
into cowpea, led to an overall, 31.4%, 17.4% and 56.8 % higher yield, NUpE and NUE, in the
subsequent maize crop compared to the applications into maize. Moreover, an overall “penalty”
of 23.2%, 27.7% and 56.7% on yield, NUpE and NUE, was calculated from residue application
into maize (Table 4-3). Nevertheless, the largest yield, NUpE and NUE differences between
sequences was attained in the SaL site. Despite the highest N-uptake levels being observed at the
cSaCL site, the wetter SaL site, showed the largest overall NUE, particularly in plots previously
under cowpea (Table 4-3). However, differences were only significant at specific treatments
(Figure 4-8). Results also show that, at specific circumstances, residue allocation into the
continuous maize sequences are likely to bring benefits to the systems.
Page | 131
Table 4- 2 Significance levels for the residual effects from residue-fertilizer allocation strategies
on yield and resource productivity of cowpea-maize and maize-maize sequences measured in the
2014-15 season
Fixed term Yield Grain N N-uptake WUEyld NUE NUpE NUtE
Previous crop ** * *** *** **
R_level
N_level *
Site *** *** *** *** *** *** ***
Crop x R-level * *
Crop x N-level
R_level x N_level
Crop x Site
R_level x Site
N_level x Site
Crop x R-level x N-level
Crop x R-level x Site ***
Crop x N-level x Site
R-level x N-level x Site
Crop x R-level x N-level x Site
* Significant at 0.05 probability level,
** Significant at 0.01 probability level,
*** Significant at 0.001 probability level
Page | 132
Figure 4-8 Short term yield, NUpE and NUE return from a maize crop grown in plots previously
planted to cowpea (Cwp-Mz) and maize (Mz-Mz) under contrasting residue and nitrogen
application levels Treatment description: ND, LNA and HNA represent unfertilized, low and
high N-application respectively. Previous residue status: with (+) and without (-) residue. (s)
represents significant differences between residual treatment effects at p<0.05.
Page | 133
Table 4- 3 Short term yield, N-uptake, NUpE, NUtE, NUE and WUE advantages from residue
and nitrogen fertilizer application into cowpea attained in a subsequent bioassay maize crop.
Results were obtained after one-year comparison between unfertilized maize grown in plots
previously planted to cowpea and maize at different residue and N-application levels.
Fixed
term
Factor
level
∆yield
(%)
∆N-uptake
(%)
∆NUpE
(%)
∆NUtE
(%)
∆NUE
(%)
∆WUE
(%)
R-levels R- 39.7 17.1 7.06 12.9 56.9 25.6
R+ 23.2 10.7 27.7 11.0 56.7 25.8
N-level ND 23.6 16.6 36.4 7.61 41.9 6.43
LNA 37.9 13.9 0.94 14.4 71.9 34.0
HNA 32.8 11.1 14.8 13.8 56.5 36.8
Sites cSaCL 34.9 3.07 4.89 24.4 81.1 29.8
SaL 28.0 24.7 29.9 -0.54 32.4 21.6
cSaCL
ND - 24.0 11.8 42.4 13.6 49.4 22.1
LNA - 131 42.1 13.3 54.7 237.5 85.2
HLA - 13.1 -9.18 -12.5 13.3 44.2 1.85
ND + -7.48 -24.4 16.5 22.2 27.4 -21.9
LNA + 6.91 -13.2 -40.1 11.5 31.6 37.1
HNA + 41.7 11.3 9.77 31.1 96.7 54.7
SaL
ND - 15.3 29.0 0.26 -13.3 -12.5 12.1
LNA - 15.6 22.8 4.09 -2.93 -0.23 7.49
HNA - 39.2 6.20 -5.22 11.9 22.8 24.9
ND + 62.7 50.2 86.4 7.93 103.1 13.4
LNA + -1.98 3.97 26.5 -5.62 18.8 6.08
HNA + 37.1 36.2 67.3 -1.19 62.4 65.7
Interpretation: Residue levels: R- is without residues and R+ with residues applied. N-levels: ND
is 0 kg N/ha, LNA is 23 kg N/ha, HNA is 92 kg N/ha; Sites: Sandy clay loam soil (cSaCL) and
Sandy loam soil (SaL). For each specific circumstance, positive variations (∆), indicate gains
from resource allocation into cowpea in the previous season, while negative variations indicate
gains from resource allocation into maize
Page | 134
4.4. Discussion
Residue responses
Negative responses to high C/N ratio residues in unfertilized maize crops were reported before
and were mostly attributed to reduced nitrogen availability due to N immobilization stress in
mulched soils (Huang et al. 2008). This is in line with results from Andraski and Bundy (2008)
who found that mulching with high C/N ratio residues reduced NO3-N early in the season,
consequently reducing maize yield in unfertilized No-till plots. In a different study, Hoyle and
Murphy (2011), observed significant reductions in plant N concentration and yield in wheat after
the incorporation of high C/N ratio (40/1) oat residues. Similar results were also reported by
(Ambus and Jensen, 2001; Wells et al., 2013), who observed a reduction in soil inorganic N and
yield of cereals in mulched soils.
Despite the fact that the N-immobilized in mulched soils is retained in the soil and is expected to
become available to plants with time (Kumar and Goh, 2002), unless enough inorganic N-
fertilizer is supplied to satisfy N-immobilization demands. However, time is a limited
commodity for resource poor farmers, to whom the use of mineral fertilisers is unaffordable.
Therefore, demonstrating potential short-term technologies to improve soil N-availability and
yields in systems where high C/N residues are used as mulches, primarily to reduce erosion or
reduce soil evaporation losses (Bescansa et al., 2006; Ward et al., 2012), is of critical
importance. Based on our results, shifting residue application from cereals into legume crops
were significant yield performance improvement was measured, is a much better resource
allocation strategy for the N-deprived maize-legume cropping systems of Central Mozambique.
The high N turnover from the legume biomass (Nsh) to the system under cowpea, is more
beneficial to the system, especially if a cereal rotation is followed. In the USA corn belt,
interrupting continuous maize sequences with a legume cycle proved effective in helping reduce
maize yield penalties from the yearly incorporation of high C/N ratio maize residues in
continuous maize systems (Morrison, 2013).
Positive responses to high C/N ratio residue application observed in cowpea (Figure 4-2), for all
tested site. Rebafka et al. (1993), measured increased nitrogen fixation and dry matter production
in groundnut grown under 4 t/ha of millet straw in Niger per year. These response can be
Page | 135
explained by the legume ability to biologically fix their own N (Sprent et al., 2010), which
protects the crop from the severe N-immobilization induced plant N-stress. Positive responses to
residue application in legume crops were measured in other studies (Herrmann et al., 2014;
Krishna, 2001; Shah et al., 2003a) The positive responses to residue application in cowpea,
especially in unfertilized (ND) and low N-application (LNA) systems are an important result to
improve N-availability in the N-deprived maize-legume cropping systems managed under
conservation agriculture in Central Mozambique and SSA. This shift in resource allocation
already proved to provide better short term returns to subsequent cereal crops compared to
continuous maize systems (Homann-Kee Tui et al., 2015). In our trial, applying residues into
cowpea led to an overall 23.2% more yield in cowpea-maize sequences compared to continuous
maize under residue application.
Nitrogen fertilizer responses
Sites differences in starting mineral N and in-crop soil moisture regimes were detrimental to
observed differences N-fertilizer responses (N-level x Site, p<0.05). In the specific case of
maize, higher and more significant yield increments from increased N-application were
measured in the wetter SaL site, (Figure 4-2), compared to the high starting N cSaCL and drier
fSa were not enough soil water was available for the crop to use all available N. Water deficit
conditioned crop N responses as in our trial were also reported for several crops including maize
(Lemaire and Gastal, 2009; Pandey et al., 2000). Furthermore, cross-site variability effect on
resource allocation responses were also reported across SSA (Giller et al., 2006; Tittonell et al.,
2007b), strengthening the idea that there are site-specific responses that need to be considered to
improve resource allocation strategies in the N-deprived maize-legume systems of central
Mozambique.
The low N responses in the high starting N cSaCL for this AEZ-R4 is a critical result to raise
awareness on fertilizer recommendations in such unreliable rainfall environments (Dimes, 2011).
The measured poor response can be attributed to poor N-translocation into grain (NUtE)
resulting from lower water availability to match the nitrogen availability (Sadras, 2004). The
negative response to high N-application measured in the cSaCL and fSa especially in the absence
of residues, is also in line with findings from Cossani et al. (2010). They proposed that for rain
Page | 136
fed conditions N-availability needs to match water supply for the crop to maximize its grain
yield for a given environment. For our trail condition, a better match between N-supply and soil
water availability was only achieved in the wet SaL site which had the highest topsoil in-crop
moisture regime (Figure 1-2), consequently the highest NUpE and NUtE (Figure 4-6). In the
drier and loose fSa, poor N-responses can be partly explained by poor NUpE (Figure 4-6a). The
strong interactive relationship between water availability and NUE as reported in (Albrizio et al.,
2010; Teixeira et al., 2014), also explains how moisture regimes conditioned nitrogen responses
across our tested sites. Similar result was reported in Sun et al. (2012), while studying the effects
of different water and nitrogen management practices on yield and NUE of rice in China. They
found that, increased water availability improved NUE and rice yield compared to dry land rice,
were water deficit led to poor NUE. Despite being conducted under supplementary irrigation,
similarities can be drawn with central Mozambique were there are clear moisture driven cross-
site variability within and across agro-ecologies, which conditions resource responses and need
to be considered when recommending fertilizer management among resource poor smallholder
farmers of central Mozambique.
Resource productivity and nitrogen driven WUE - NUE trade-offs
Inorganic N-fertilizer applications above 23 kg N/ha significantly reduced NUE and its
components NUpE, NUtE. Decreases in NUE and its components with increased N-application
have been reported before (Giambalvo et al., 2010; Sadras and Rodriguez, 2010). NUE decreases
above such low levels of N-application clearly demonstrates how risky and unprofitable (Dimes,
2011) applying high amounts of inorganic N-fertilizer can be in unreliable rainfall environments.
Measured crop WUE was considerably lower than the levels proposed by Shah et al. (2003b) for
rain fed conditions i.e., 20 kg grain/ha.mm and 10 kg grain/ha.mm for maize and legume crops
grown respectively (Figure 4-3). It can be speculated that, poor rainfall distribution and the
occurrence of dry spells during early flowering and grain filling especially in the second season
negatively affected crop WUE. Rainfall distribution was reported to affect WUE in China (Sun
et al., 2010). Wang et al. (2010a) calculated lower maize WUE in dryland environments with
regular dry spells during tasselling. Despite, the observed low WUE, increased N-application
proved to increase WUE, as also reported by (Sun et al., 2012).
Page | 137
A nitrogen driven trade-off between NUE and WUE was also observed as in (Gong et al., 2011;
Sadras and Rodriguez, 2010). For both maize and cowpea, increased N-application reduced NUE
in detriment of WUE, whereas increased water availability reduced WUE in detriment of NUE
(Figure 4-7). Nevertheless, the trade-off was modified by residue application in both crops. The
direction and magnitude of these changes, depended on the individual crop and site response to
both nitrogen and residue application. Understanding the underlying factor behind the
modification of such important trade-offs and how these variations can be used for the benefit of
the system is fundamental to guide the design of improved and sustainable site and crop-specific
resource allocation recommendations for the N-deprived maize-legume systems of central
Mozambique.
Demonstrating immediate returns from alternative resource allocation strategies
Results from the bioassay crop, showed that for the N-deprived maize-legume cropping systems
of central Mozambique, the highest net returns from both high C/N ratio residues and N-fertilizer
application were achieved with applications into cowpea rather than maize. This is in line with
findings from Bakht et al. (2009), who reported a strong carry-over effect in legume-cereal
sequences as a key factor to improve N-availability and reduce N-losses in residue applied fields
(Ambus and Jensen, 2001). The high return measured in the cowpea-sequence, especially in ND
and LNA systems where losses are expected from residue applied maize, shows how allocation
the limited available resources into improving the performance of a sole legume crop is a better
and more pragmatic strategy (Kirkegaard et al., 2014), to improve N-deprived yields of the
following cereal crop in the N-deprived systems of Central Mozambique. This shift in resource
allocation is in part sustained on the premise that, with current low fertilizer use in SSA (Gilbert,
2012; Heffer and Prud'homme, 2011), resource poor farmers are less likely to afford the amounts
of N-fertilizers required to limit N-immobilization from yearly incorporation of high C/N
residues in continuous maize cropping systems (Morrison, 2013). In the USA corn belt, legumes
have also proved successful in reducing the economic optimum N-rates and improve yields in
wheat and maize systems under residue retention (Luce et al., 2015; Morrison, 2013). These
results strengthen our proposition that, in N-deprived maize-legume cropping system of central
Mozambique, investing the limited resources, i.e., residues and N-fertilizers to improve crop
Page | 138
performance during the legume phase of the rotation is likely to yield better returns to the
system.
The high N returns from previous crop and NUpE in the subsequent maize crop measured in the
cowpea-maize sequence, are in part resultant from the significant difference in N turnover from
biomass registered across sequences. In plots previously grown to cowpea, an average 122.7 kg
N/ha were returned to the soil, against 30.2 kg N/ha under maize. Because, in the continuous
maize sequence, N is still trapped into its high C/N ratio biomass, whose decomposition requires
further N to satisfy N-immobilization demands (Huang et al., 2008), yields penalties from
retained maize residue were measured in the subsequent season. Reversely, better residual
returns were attained in the cowpea-maize sequence where N from decomposing legume residues
were easily available through mineralized N (Kumar and Goh, 2002). For that reason, NUpE,
NUE and yields in the following season, were considerably higher in the cowpea-maize.
4.6. Conclusions
Results from this study, indicate that there are crop and site specific responses and trade-offs to
residue and N-fertilizer allocation that need to be considered in order to improve resource
productivity and yields in the rain fed N-deprived maize-legume cropping systems of central
Mozambique. Trial results clearly demonstrated that, N-fertilizers are better off, if used to
improve maize yields in residue free systems while legumes are a better match to residues. Better
returns from a combined application of available high C/N ratio residues and limited amounts of
inorganic N-fertilizers into the legume phase of the rotation were also measured. This shift in
practice proved beneficial in improving N-availability and yields of the following maize crop in
cowpea-maize sequences, contrary to the continuous maize systems yield losses from residue
application were registered. A nitrogen driven trade-off between NUE-WUE, modified by
residue and N-responses across crops and sites, was also demonstrated in our results. Therefore,
understanding how these trade-offs occurs in key to the design of feasible and locally tailored
resource management recommendations that are likely to help amplify desired responses or
alleviate unwanted ones in order to improve resource productivity and yields. To conclude, our
study results are of critical importance to SSA N-deprived maize-legume systems managed under
conservation agriculture, where demonstrating potential short term benefits of widely promoted
Page | 139
technologies such as residue retention, N-application and legume rotations through the
promotion of smart and pragmatic resource allocation strategies across crops and farming
environments is critical to easy a more conscious and effective adoption process. Nevertheless,
more research is required to create generalizable and robust information that can help understand
under what specific circumstances, i.e., farming environments, management and rain fall
patterns, benefits and penalties from residues are likely to be observed for a specific agro-
ecologies.
Acknowledgements
We would like to thank the Sustainable Intensification of Maize and Legume farming systems
for Eastern and Southern Africa (SIMLESA) program funded by the Australian Centre for
International Agriculture Research (ACIAR). ACIAR also awarded the PhD scholarship under
which this research was conducted. The AGRA Soil Health Program through the 2011SHP007
Project partially funded the fieldwork. We also thank Jacinto Andreque, Clemente Zivale,
Manuel Chirindzane, Cristovão D. Maquina, Eliseu D. Maquina and Ricardo Maquina for their
help with the fieldwork. We particularly acknowledge the support from SIMLESA’s
Mozambique team and country coordinator Mr Domingos Dias.
References
Aggarwal, P. K., and Sinha, S. K. (1983). Water stress and water-use efficiency in field grown
wheat: A comparison of its efficiency with that of C4 plants. Agricultural Meteorology
29, 159-167.
Aggarwal, R. K., Praveen, K., and Power, J. F. (1997). Use of crop residue and manure to
conserve water and enhance nutrient availability and pearl millet yields in an arid tropical
region. Soil and Tillage Research 41, 43-51.
Albrizio, R., Todorovic, M., Matic, T., and Stellacci, A. M. (2010). Comparing the interactive
effects of water and nitrogen on durum wheat and barley grown in a Mediterranean
environment. Field Crops Research 115, 179-190.
Page | 140
Ambus, P., and Jensen, E. S. (2001). Crop residue management strategies to reduce N-losses-
interaction with crop N-supply. Communications in Soil Science and Plant Analysis 32,
981-996.
Andraski, T. W., and Bundy, L. G. (2008). Corn Residue and Nitrogen Source Effects on
Nitrogen Availability in No-Till Corn All rights reserved. No part of this periodical may
be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system,
without permission in writing from the publisher. Agronomy Journal 100, 1274-1279.
Bakht, J., Shafi, M., Jan, M. T., and Shah, Z. (2009). Influence of crop residue management,
cropping system and N fertilizer on soil N and C dynamics and sustainable wheat
(Triticum aestivum L.) production. Soil and Tillage Research 104, 233-240.
Baudron, F., Thierfelder, C., Nyagumbo, I., and Gérard, B. (2015). Where to Target
Conservation Agriculture for African Smallholders? How to Overcome Challenges
Associated with its Implementation? Experience from Eastern and Southern Africa.
Environments 2, 338.
Bescansa, P., Imaz, M. J., Virto, I., Enrique, A., and Hoogmoed, W. B. (2006). Soil water
retention as affected by tillage and residue management in semiarid Spain. Soil and
Tillage Research 87, 19-27.
Chianu, J., Chianu, J., and Mairura, F. (2012). Mineral fertilizers in the farming systems of sub-
Saharan Africa. A review. Agronomy for Sustainable Development 32, 545-566.
Cossani, C. M., Slafer, G. A., and Savin, R. (2010). Co-limitation of nitrogen and water, and
yield and resource-use efficiencies of wheat and barley. Crop and Pasture Science 61,
844-851.
Cossani, C. M., Slafer, G. A., and Savin, R. (2012). Nitrogen and water use efficiencies of wheat
and barley under a Mediterranean environment in Catalonia. Field Crops Research 128,
109-118.
Crawford, E. W., Jayne, T. S., and Kelly, V. A. (2006). Alternative Approaches for Promoting
Fertilizer Use in Africa. pp. 88. World Bank, Washington, USA.
Dalgliesh, N., and Foale, M. (1998). "Soil matters: Monitoring soil water and nutrients in
dryland farming," 1/Ed. CSIRO, Australia, Toowoomba, Australia.
Page | 141
Dimes, J. (2011). Formulating Crop Management Options for Africa’s Drought-Prone Regions:
Taking Account of Rainfall Risk Using Modeling. In “Innovations as Key to the Green
Revolution in Africa.” (A. Bationo, B. Waswa, J. M. M. Okeyo, F. Maina and J. M.
Kihara, eds.), pp. 785-794. Springer Science and Business Media, Netherlands.
Enyong, L. A., Debrah, S. K., and Bationo, A. (1999). Farmers' perceptions and attitudes towards
introduced soil-fertility enhancing technologies in western Africa. Nutrient Cycling in
Agroecosystems 53, 177-187.
Erenstein, O., Gérard, B., and Tittonell, P. (2015). Biomass use trade-offs in cereal cropping
systems in the developing world: Overview. Agricultural Systems 134, 1-5.
French, R., and Schultz, J. (1984). Water use efficiency of wheat in a Mediterranean-type
environment. II. some limitations to efficiency. Australian Journal of Agricultural
Research 35, 765-775.
Gentry, L. F., Ruffo, M. L., and Below, F. E. (2013). Identifying Factors Controlling the
Continuous Corn Yield Penalty. Agronomy Journal 105, 295-303.
Giambalvo, D., Ruisi, P., Di Miceli, G., Frenda, A. S., and Amato, G. (2010). Nitrogen Use
Efficiency and Nitrogen Fertilizer Recovery of Durum Wheat Genotypes as Affected by
Interspecific Competition All rights reserved. No part of this periodical may be
reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system,
without permission in writing from the publisher. Agronomy Journal 102, 707-715.
Gilbert, N. (2012). Dirt Poor: The key to tackle hunger in Africa is enriching its Soils. Nature
483, 525-227.
Giller, K. E., Corbeels, M., Nyamangara, J., Triomphe, B., Affholder, F., Scopel, E., and
Tittonell, P. (2011). A research agenda to explore the role of conservation agriculture in
African smallholder farming systems. Field Crops Research 124, 468-472.
Giller, K. E., Rowe, E. C., de Ridder, N., and van Keulen, H. (2006). Resource use dynamics and
interactions in the tropics: Scaling up in space and time. Agricultural Systems 88, 8-27.
Giller, K. E., Witter, E., Corbeels, M., and Tittonell, P. (2009). Conservation agriculture and
smallholder farming in Africa: The heretics’ view. Field Crops Research 114, 23-34.
Page | 142
Gong, X. Y., Chen, Q., Lin, S., and Brueck, H. (2011). Tradeoffs between nitrogen- and water-
use efficiency in dominant species of the semiarid steppe of Inner Mongolia. Plant and
soil 340, 227-238.
Govaerts, B., Verhulst, N., Castellanos-Navarrete, A., Sayre, K. D., Dixon, J., and Dendooven,
L. (2009). Conservation Agriculture and Soil Carbon Sequestration: Between Myth and
Farmer Reality. Critical Reviews in Plant Sciences 28, 97-122.
Heffer, P., and Prud'homme, M. (2011). "Fertilizer Outlook 2011-2015." IFA, Paris, France.
Herrmann, L., Chotte, J. L., Thuita, M., and Lesueur, D. (2014). Effects of cropping systems,
maize residues application and N fertilization on promiscuous soybean yields and
diversity of native rhizobia in Central Kenya. Pedobiologia 57, 75-85.
Homann-Kee Tui, S., Valbuena, D., Masikati, P., Descheemaeker, K., Nyamangara, J.,
Claessens, L., Erenstein, O., van Rooyen, A., and Nkomboni, D. (2015). Economic trade-
offs of biomass use in crop-livestock systems: Exploring more sustainable options in
semi-arid Zimbabwe. Agricultural Systems 134, 48-60.
Hoyle, F. C., and Murphy, D. V. (2011). Influence of organic residues and soil incorporation on
temporal measures of microbial biomass and plant available nitrogen. Plant and Soil 347,
53-64.
Huang, Z., Xu, Z., Blumfield, T., Chen, C., and Bubb, K. (2008). Soil nitrogen mineralization
and fate of (15NH4)2SO4 in field-incubated soil in a hardwood plantation of subtropical
Australia: the effect of mulching. Journal of Soils and Sediments 8, 389-397.
Huggins, D. R., and Pan, W. L. (1993). Nitrogen Efficiency Component Analysis: An Evaluation
of Cropping System Differences in Productivity. Agronomy Journal 85, 898-905.
Kintché, K., Guibert, H., Sogbedji, J. M., Levêque, J., Bonfoh, B., and Tittonell, P. (2015).
Long-term mineral fertiliser use and maize residue incorporation do not compensate for
carbon and nutrient losses from a Ferralsol under continuous maize–cotton cropping.
Field Crops Research 184, 192-200.
Kirkegaard, J. A., Conyers, M. K., Hunt, J. R., Kirkby, C. A., Watt, M., and Rebetzke, G. J.
(2014). Sense and nonsense in conservation agriculture: Principles, pragmatism and
productivity in Australian mixed farming systems. Agriculture, Ecosystems and
Environment 187, 133-145.
Page | 143
Krishna, N. R. (2001). Effects of Cereal and Legume Cover Crop Residues on Weeds, Yield, and
Net Return in Soybean (Glycine max). Weed Technology 15, 660-668.
Kugbe, J., Fosu, M., and Vlek, P. L. G. (2015). Impact of season, fuel load and vegetation cover
on fire mediated nutrient losses across savanna agro-ecosystems: the case of northern
Ghana. Nutrient Cycling in Agroecosystems 102, 113-136.
Kumar, K., and Goh, K. M. (2002). Management practices of antecedent leguminous and non-
leguminous crop residues in relation to winter wheat yields, nitrogen uptake, soil nitrogen
mineralization and simple nitrogen balance. European Journal of Agronomy 16, 295-308.
Lahmar, R., Bationo, B. A., Dan Lamso, N., Guéro, Y., and Tittonell, P. (2012). Tailoring
conservation agriculture technologies to West Africa semi-arid zones: Building on
traditional local practices for soil restoration. Field Crops Research 132, 158-167.
Lemaire, G., and Gastal, F. (2009). Chapter 8 - Quantifying Crop Responses to Nitrogen
Deficiency and Avenues to Improve Nitrogen Use Efficiency. In "Crop Physiology" (V.
Sadras and D. Calderini, eds.), pp. 171-211. Academic Press, San Diego, USA.
Moll, R. H., Kamprath, E. J., and Jackson, W. A. (1982). Analysis and Interpretation of Factors
Which Contribute to Efficiency of Nitrogen Utilization1. Agronomy Journal 74, 562-564.
Montemurro, F., Maiorana, M., Ferri, D., and Convertini, G. (2006). Nitrogen indicators, uptake
and utilization efficiency in a maize and barley rotation cropped at different levels and
sources of N fertilization. Field Crops Research 99, 114-124.
Morrison, L. (2013). Residue management can reduce continuous corn yield penalty. Corn and
Soybean Digest. Vol. 75 Issue 9, p 40
Ngwira, A. R., Thierfelder, C., and Lambert, D. M. (2013). Conservation agriculture systems for
Malawian smallholder farmers: long-term effects on crop productivity, profitability and
soil quality. Renewable Agriculture and Food Systems 28, 350-363.
Pala, M., Ryan, J., Zhang, H., Singh, M., and Harris, H. C. (2007). Water-use efficiency of
wheat-based rotation systems in a Mediterranean environment. Agricultural Water
Management 93, 136-144.
Pandey, R. K., Maranville, J. W., and Admou, A. (2000). Deficit irrigation and nitrogen effects
on maize in a Sahelian environment: I. Grain yield and yield components. Agricultural
Water Management 46, 1-13.
Page | 144
Rebafka, F. P., Ndunguru, B. J., and Marschner, H. (1993). Crop residue application increases
nitrogen fixation and dry matter production in groundnut (Arachis hypogaea L.) grown
on an acid sandy soil in Niger, West Africa. Plant and Soil 150, 213-222.
Rossiter-Rachor, N. A., Setterfield, S. A., Douglas, M. M., Hutley, L. B., and Cook, G. D.
(2008). Andropogon gayanus (Gamba Grass) Invasion Increases Fire-mediated Nitrogen
Losses in the Tropical Savannas of Northern Australia. Ecosystems 11, 77-88.
Sadras, V. O. (2004). Yield and water-use efficiency of water- and nitrogen-stressed wheat crops
increase with degree of co-limitation. European Journal of Agronomy 21, 455-464.
Sadras, V. O., and Angus, J. F. (2006). Benchmarking water-use efficiency of rain fed wheat in
dry environments. Australian Journal of Agricultural Research 57, 847-856.
Sadras, V. O., and Rodriguez, D. (2010). Modelling the nitrogen-driven trade-off between
nitrogen utilisation efficiency and water use efficiency of wheat in eastern Australia.
Field Crops Research 118, 297-305.
Shah, Z., Shah, S. H., Peoples, M. B., Schwenke, G. D., and Herridge, D. F. (2003). Crop residue
and fertiliser N effects on nitrogen fixation and yields of legume-cereal rotations and soil
organic fertility. Field Crops Research 83, 1-11.
Sprent, J. I., Odee, D. W., and Dakora, F. D. (2010). African legumes: a vital but under-utilized
resource. Journal of Experimental Botany 61, 1257-65.
St. Luce, M., Grant, C. A., Zebarth, B. J., Ziadi, N., O’Donovan, J. T., Blackshaw, R. E., Harker,
K. N., Johnson, E. N., Gan, Y., Lafond, G. P., May, W. E., Khakbazan, M., and Smith, E.
G. (2015). Legumes can reduce economic optimum nitrogen rates and increase yields in a
wheat–canola cropping sequence in western Canada. Field Crops Research 179, 12-25.
Sun, H., Shen, Y., Yu, Q., Flerchinger, G. N., Zhang, Y., Liu, C., and Zhang, X. (2010). Effect of
precipitation change on water balance and WUE of the winter wheat–summer maize
rotation in the North China Plain. Agricultural Water Management 97, 1139-1145.
Sun, Y., Ma, J., Sun, Y., Xu, H., Yang, Z., Liu, S., Jia, X., and Zheng, H. (2012). The effects of
different water and nitrogen managements on yield and nitrogen use efficiency in hybrid
rice of China. Field Crops Research 127, 85-98.
Teixeira, E. I., George, M., Herreman, T., Brown, H., Fletcher, A., Chakwizira, E., de Ruiter, J.,
Maley, S., and Noble, A. (2014). The impact of water and nitrogen limitation on maize
Page | 145
biomass and resource-use efficiencies for radiation, water and nitrogen. Field Crops
Research 168, 109-118.
Thierfelder, C., Cheesman, S., and Rusinamhodzi, L. (2012). A comparative analysis of
conservation agriculture systems: Benefits and challenges of rotations and intercropping
in Zimbabwe. Field Crops Research 137, 237-250.
Thierfelder, C., Matemba-Mutasa, R., and Rusinamhodzi, L. (2015). Yield response of maize
(Zea mays L.) to conservation agriculture cropping system in Southern Africa. Soil and
Tillage Research 146, Part B, 230-242.
Tittonell, P., Gérard, B., and Erenstein, O. (2015). Tradeoffs around crop residue biomass in
smallholder crop-livestock systems – What’s next? Agricultural Systems 134, 119-128.
Tittonell, P., Zingore, S., van Wijk, M. T., Corbeels, M., and Giller, K. E. (2007). Nutrient use
efficiencies and crop responses to N, P and manure applications in Zimbabwean soils:
Exploring management strategies across soil fertility gradients. Field Crops Research
100, 348-368.
Vanlauwe, B., Wendt, J., Giller, K. E., Corbeels, M., Gerard, B., and Nolte, C. (2014). A fourth
principle is required to define Conservation Agriculture in sub-Saharan Africa: The
appropriate use of fertilizer to enhance crop productivity. Field Crops Research 155, 10-
13.
Vibrans, H., García-Moya, E., Clayton, D., and Sánchez-Ken, J. G. (2014). Hyparrhenia
variabilis and Hyparrhenia cymbaria (Poaceae): New for the Americas, Successful in
Mexico. Invasive Plant Science and Management 7, 222-228.
Wang, X., Dai, K., Wang, Y., Zhang, X., Zhao, Q., Wu, X., Cai, D., Hoogmoed, W. B., and
Oenema, O. (2010). Nutrient management adaptation for dryland maize yields and water
use efficiency to long-term rainfall variability in China. Agricultural Water Management
97, 1344-1350.
Ward, P. R., Flower, K. C., Cordingley, N., Weeks, C., and Micin, S. F. (2012). Soil water
balance with cover crops and conservation agriculture in a Mediterranean climate. Field
Crops Research 132, 33-39.
Wells, M. S., Reberg-Horton, S. C., Smith, A. N., and Grossman, J. M. (2013). The Reduction of
Plant-Available Nitrogen by Cover Crop Mulches and Subsequent Effects on Soybean
Performance and Weed Interference. Agronomy Journal 105, 539-545.
Page | 146
CHAPTER 5:
Pathways for Sustainable Residue Management
Page | 147
Model assisted decision support frameworks for residue management in N-deprived maize-
legume cropping systems
Nhantumbo, Nascimento12., Mortlock, M 3, Nyagumbo, I4., and Rodriguez, D1 1 Queensland Alliance for Agriculture and Food Innovation (QAAFI), the University of Queensland PO
Box 102 Toowoomba Qld (4350) Australia; 2 Divisão de Agricultura, Instituto Superior Politecnico de
Manica; Corresponding author: [email protected]; 3School of Agriculture and Food Science (SAFS), the University of Queensland; 4CIMMYT-Harare
Abstract
Quality information is needed to support decision making on the use of limited resources in low
external N-input agricultural systems (LENIA) to improve yields and farmers’ profits,
particularly when high C/N residues are being promoted as mulches in conservation agriculture
(CA) systems. In this study, field trials and the Agricultural Production Systems Simulator
(APSIM) were used to research under what circumstances, i.e., crop type, farming environments,
and N-fertilizer regimes, the addition of high C/N ratio crop residues in CA systems is likely to
be beneficial in terms of crop yields and resource productivity. Results showed that the use of
high C/N ratio residues in LENIA systems managed under CA produced two responses 1) a crop
and soil type mediated residue effect on yield, water use efficiency (WUE) and nitrogen use
efficiency (NUE); 2) a residue driven nitrogen trade-off that was different between N fixing
crops (cowpea) and maize. In soils of high water holding capacity (WHC), gains in crop WUE
are likely to be outperformed by yield losses due to poor NUE caused by N-immobilization
during residue decomposition. Simulations showed a 90% chance of maize yield losses when
high C/N ratio residues are applied. However, cowpea responded positively to this type of
mulches. For dry and low WHC soils, moisture benefits from the application of high C/N residue
mulches were higher than any yield losses due to N-immobilization, and positive yield responses
from residue application were simulated. Results from this study shows the value of using
simulation modelling in combination with field experimentation to assess the outcome of
complex and dynamic processes to generate residue management recommendations.
Key words: APSIM, Conservation agriculture, residues, maize, Mozambique,
Page | 148
5.1 Introduction
Across most Sub-Saharan Africa (SSA) smallholder farming systems, improving the allocation
of locally available resources such as high C/N ratio residues, manure and limited amounts of
inorganic fertilizers, is critical to improve resource productivity and overall system performance.
However, in SSA low external N-input agricultural systems (LENIA) managed under
conservation agriculture, the role of high C/N ratio residue mulches has generated highly
contrasting views (Giller et al., 2009; Sommer et al., 2014; Vanlauwe et al., 2014). In part, most
of the discussion surrounding residues and fertilizer use in smallholder conservation agriculture
(CA) farming systems in Africa is related to the perceived need to produce generalizable
information easy to scale across regions, and highly contrasting farming situations.
To date, benefits from CA principles have been reported for specific situations (Baudron et al.,
2015c). Therefore, understanding the interactive relationships between the components of CA
system mainly crop residue mulches, fertilizers, crops and their management effect on yields and
resource productivity across farming environments (Giller et al., 2011a), is of extreme
importance to tailor CA principles to SSA diverse and complex farming circumstances.
Here we propose that understanding under what circumstances the use of high C/N ratio residues
as mulches is likely to increase or reduce yields, can help design more productive CA systems.
The use of dynamic cropping systems models can help quantify and untangle some of the
complexities in SSA smallholder CA system (Bouma and Jones, 2001; Jordan et al., 1997;
Maxwell, 1986). In Australia, for instance, the Agricultural Production Systems Simulator
(APSIM) (Holzworth et al., 2014) helped researchers identify the most suitable and remunerative
cropping systems for different climatic and socio-economic contexts (Cox et al., 2010; Meinke et
al., 2001; Rodriguez et al., 2011). Moreover, the integration of simulation modelling in research
programs across the world have helped to (1) identify gaps in existing knowledge; (2) generate
and test hypothesis to improve the design of experiments; (3) determine the most influential
parameters of a system (Matthews and Stephens, 2002b); and (4) bring researchers, experiments
and farmers together to discuss problems and identify solutions in participatory modelling
exercises (Whitbread et al., 2010).
Page | 149
Despite the attention that agricultural simulation models experienced in developed countries,
there has been little use in the developing world, were public expenditure in agricultural research
and development is the lowest (Beintema and Stads, 2008). Nevertheless, cost effective
approaches like agronomic models could help define attainable research outputs, and facilitate
the selection of suitable technologies to be tested in farmers’ fields. Among several models,
APSIM has proven valuable in Africa (Chikowo et al., 2008; Masikati et al., 2014; Mupangwa et
al., 2011; Ncube et al., 2009). The success of APSIM in modelling farming systems in Africa,
comes mainly as a result of integrating a systems approach (Whitbread et al., 2010) that allows it
to capture and analyse GxExM interactions (Holzworth et al., 2014), relevant to improving
research design and interpretation of African smallholder farming systems.
In this study, we used APSIM v7.7 to model the observed yield responses to crop residue mulch
management strategies and its linkages with water use efficiency (WUE) and nitrogen use
efficiencies (NUE) in maize-legume systems. The main objective of this study was to develop
simple rules of thumb to guide the allocation of high C/N ratio mulches between legumes and
cereals in LENIA systems managed under CA.
5.1.1 Conceptual framework
In the last decades, all alone in SSA, vast research have been conducted to demonstrate the
potential benefits of the conservation agriculture principles, of which residues retention are an
integral part (Ngwira et al., 2013; Thierfelder et al., 2012; Thierfelder et al., 2015; Thierfelder et
al., 2013). Fertilizer research has also deserved its share of attention (Gilbert, 2012; Ncube et al.,
2007; Twomlow et al., 2010). Nevertheless, most of these research was conducted under specific
trial conditions and focused in demonstrating the contribution of individual CA principals rather
than their interaction. Contrary to models, field trials by themselves have been so far unable to:
1) capture and explain the complex GxExM interactions and 2) explore the dynamics of WUE
and NUE which are of critical importance to generate relevant recommendation on the use of
high C/N residues and fertilizes that are relevant to mostly N-deprived SSA farming systems.
Therefore, results generated to date are of local interest with less generalizable value (Sadras and
Lemaire, 2014). To close this information gap, an APSIM based decision support framework for
residue management is proposed. The framework makes use of pre-established knowledge on
Page | 150
residue-fertilizer interactions to underpin under which circumstances benefits and losses from
alternative residues management strategies are likely to occur. To validate the framework, the
following operational principles were assumed:
1. In rain fed N-deprived systems, maize yield increases under residue application, are
likely to be attained at specific circumstances, i.e., crops, soils, N-availability levels, and
season types. Benefits from residues application are attained only when residue induced
crop WUE increments surpass yield losses from residue induced NUE decrease due to N-
immobilization.
2. There will always be a nitrogen driven trade-off between NUE and WUE (Sadras and
Rodriguez, 2010) whose magnitude can be alleviated or worsened by the amount and
type of mulch used for each crop type and fertilizer application level.
3. The balance between N-mineralization and immobilization during residue decomposition
is affected by the soils moisture regimes and N-supply to the system (McCown et al.,
1996; Probert et al., 1998). Therefore, wet and low N environments are expected to
provide a quick residue decomposition, and consequently higher chances of N-
immobilization induced plant N-stress than water limited environments.
Four propositions (rules of thumb) are used as background for the framework proposed in this
study:
1. In wet and high WHC soils, the use of high C/N ratio residues to improve the
performance of the sole legume crop is a better strategy than applying those residues on
the maize crop.
2. In dryer and low WHC soils, maize yields are likely to be increased by residue
application as a direct result of improved rainfall capture.
3. For wet and high WHC soils, the beneficial effects of applying crop residues on the
legume crop are likely to be observed on the following cereal crop.
4. The balance between the positive and negative effects of residue application on yield,
grain water productivity, and the temporary nitrogen immobilization in the rain fed
LENIA systems, can be managed with the use of nitrogen fertilisers and improved
legume-cereal sequences.
Page | 151
5.2. Materials and Methods
5.2.1 Experiment for model parameterization
The APSIM model was parameterised using data from field experiments conducted in central
Mozambique designed to evaluate the effects of different residue and fertilizer allocation
strategies in maize legume cropping systems. APSIM was parameterised for three contrasting
environments, namely: 1) a red ferralsol with sandy clay loam (cSaCL) texture representing an
upland farming environment with high starting soil N and high WHC; 2) a fine sandy (fSa)
textured Arenosol representing a dry and low WHC upland farming environment; 3) a dark
sandy loam (SaL) textured Gleysol representing a lowland farming environment with low
starting N and prolonged wetness, i.e., high in crop soil moisture regime. Soils were classified
based on the FAO-Unesco soil classification system.
In each environment twelve treatments were simulated, resulting from the combination of three
factors. Factors included crop type i.e., sole maize and sole cowpea, two levels of residue
application (0 and 6 t/ha); and three levels of N-fertilizer application - 0, 23 and 92 kg N/ha
representing an unfertilized N-deficient system (ND), a low N-application (LNA) and a high N
application system (HNA), respectively. The 6 t/ha residue application rate, was selected to
mimic an environment with a high N-immobilization risk due to continuous incorporation of
carbon rich maize stubble and grasses characteristic of central Mozambique LENIA systems.
This value is proposed on ex-ante model simulations for Central Mozambique which showed an
80-75% chance of achieving a 3 and 4.5 t/ha yield in ND and LNA systems without residue
application meaning that a 4±5.5 t/ha maize stubble amount is likely to be attained in the region
using a maize hybrid. Maize stubble amount is calculated based on Southern Africa average
maize yield of 3.2 t/ha assuming a 0.45 harvest index (Kintché et al., 2015). These figure are in
line with data presented in (Baudron et al., 2015c) who estimated an approximately 2428 ± 5890
kg maize biomass per year in central Mozambique farms.
The N-application levels were selected based on two criteria. The 23 kg N/ha assumed for the
LNA systems is an approximation of the fertilizer package in FAO input voucher in
Mozambique which provides approximately 29 kg N/ha through a 50 kg bag of NPK 12:24:12
and a 50 kg bag urea (46 N) to be applied as basal and top-dress fertilizer. To facilitate
Page | 152
recommendations and align the figure with sampled fertilizer applicant amounts in in our study
area (18.3 ± 24.1 kg N/ha) an average N-amount of 23.8 kg N/ha averaged across this three
values was kept to 50 kg bag of urea assuming only a top-dress application. The HNA was based
on reported N-application amount needed to satisfy N-immobilization demands during high
carbon residue decomposition in SSA LENIA systems which ranges between 80-100 kg N/ha
(Homann-Kee Tui et al., 2015; Vanlauwe et al., 2014). The amount was kept to 92 kg N/ha, i.e.,
4 bags of urea (46 N).
Soil parameterization in APSIM
Actual soil data collected at the field trials was used for the model parameterization. This
included bulk density, soil water, soil organic carbon (SOC) and mineral N (NO3-N) measured at
the start of the cropping (Table 5-1). In APSIM the nitrogen module was initialized with
measured soil nitrate N (NO3-N) in each site. Crop lower limit (CLL) and drained upper limit
(DUL) were determined opportunistically across the seasons (Pala et al., 2007). The lowest
measured soil water content at harvest in each soil was used as CLL. The DUL was averaged
between the highest measured soil water in each soil and the results from a pond experiment as
described in the soil matter procedural manual from Dalgliesh and Foale (1998). Soil profile (0-
1.20 m) plant available water content at sowing (PAWCs) of 91 mm, 41 mm and 11 mm for the
sandy loam, sandy clay loam and fine sand respectively were used to initialize the model. The
soil profile was considered to be filled from the top and the soil C/N ratio was set at 14.5 based
on soil analysis data.
Crop parameterization
For maize, a mid-duration hybrid PAN67 from PANNAR Seeds, was used. An existing cultivar
description in APSIM (SC501, a hybrid) was modified to represent PAN 67 yield potential (c.a.
8t/ha) and duration to physiological maturity (120 days). The main adjustments for this
calibration were; maximum grain number = 550 (from 520), grain growth rate was kept at 9
mg/grain/day, and thermal time from flowering to maturity = 980 oC days (from 730). The
genetic coefficients for maize were estimated using a vet fit method (Balwinder et al., 2011).
Here, variation on the genetic coefficients in a pre-existing variety were made in order to
produce a close match between measured and simulated phenology and grain yield. For cowpea,
Page | 153
an existing short duration variety, Banjo was adjusted to represent the IT16. Adjustments were
made on thermal degrees’ day from emergence to floral initiation (552 to 565), cumulative
vernal day from 100 to 85, estimated days from emergence to floral initiation from 20 to 25 days
to reproduce the yield of the unfertilized and no residue applied cowpea treatment. These
adjustments were based on averaged data from two cowpea crops, sown in November (early
sowing), and February of the 2012-13 and 2013-14.
Table 5-1 Initial soil organic carbon (SOC), nitrate-nitrogen (NO3-N) and soil water description
parameters from soil samples collected in the 2013-14 seasons and were used for model parameterization
Farming Environment Parameter
Soil layer (cm)
0-15 15-30 30-60 60-90 90-120
Arenosol
Low N and low WHC (fSa)
SOC (%) 0.594 0.564 0.133 0.070 0.020
NO3-N (ppm) 5.720 3.270 0.860 0.870 0.670
Air dry (mm/mm) 0.030 0.050 0.087 0.127 0.137
CLL (mm/mm) 0.074 0.074 0.087 0.127 0.134
DUL (mm/mm) 0.140 0.140 0.140 0.140 0.140
SAT (mm/mm) 0.331 0.327 0.327 0.376 0.376
Bulk density (g/cm3) 1.640 1.647 1.648 1.521 1.521
Gleysol
Low N and high WHC (SaL)
SOC (%) 2.040 2.370 1.140 0.520 0.370
NO3-N (ppm) 5.100 0.100 0.100 0.010 0.010
Air dry (mm/mm) 0.040 0.060 0.180 0.239 0.252
CLL (mm/mm) 0.100 0.110 0.180 0.239 0.252
DUL (mm/mm) 0.200 0.200 0.283 0.290 0.300
SAT (mm/mm) 0.384 0.376 0.376 0.403 0.410
Bulk density (g/cm3) 1.503 1.518 1.522 1.455 1.432
Ferralsol
High N and high WHC (cSaCL)
SOC (%) 1.930 1.670 0.900 0.620 0.590
NO3-N (ppm) 7.790 6.920 4.120 2.290 1.360
Air dry (mm/mm) 0.050 0.070 0.183 0.193 0.199
CLL (mm/mm) 0.121 0.121 0.183 0.193 0.199
DUL (mm/mm) 0.258 0.286 0.296 0.298 0.299
SAT (mm/mm) 0.384 0.403 0.420 0.425 0.422
Bulk density (g/cm3) 1.500 1.450 1.520 1.390 1.390
Page | 154
5.2.2 Long term simulations: residue induced yield, NUE and WUE responses
APSIM v7.7, was used to evaluate the expected variation in maize and cowpea response to a
range of fertiliser x mulch inputs in Chimoio. As cowpea is a major legume for farmers in this
region, the analysis was extended to assess whether farmers would obtain a higher return from
high C/N maize residues if applied as mulch to cowpea crops. Once the model was
parameterized, it was used to:
a) Test whether the model was able to reproduce the observed treatment and site responses;
b) Identify under which circumstances, i.e., crops, sites, N-levels and seasons, responses to
high C/N residues would be likely. Responses were evaluated in term of yield (kg/ha),
water use efficiency (WUE, kg grain/ha.mm), and NUE (kg grain/kg N-supplied).
Multi-year simulations (64 years) were conducted to assess the potential response of maize and
cowpea to residues. The C/N ratio of applied residues was kept at 54:1. Daily climate records for
rainfall, temperature and radiation at Chimoio were available for the period January 1951 to
April 2015, were used. Simulations were initiated on 01 September (end of dry season) and
ended on 15th April. Weed competition was simulated using a summer grass cultivar with a
population of 3 pl/m2, based on weed counts taken along the maize crop. The soil module was
initialised at 27 and 64% plant available water content evenly distributed in the soil profile to
simulate two contrasting environments. Initial NO3-N and NH4-N in the soil profile was set at 10
and 5 kg N/ha to 1.20 m, respectively.
Calculations
Water use efficiency (WUE) was calculated for both maize and cowpea as the ratio between
yield and evapotranspiration. Evapotranspiration was determined as the difference between in-
crop rainfall and soil water losses by the system which a sum of surface soil runoff and water
losses through deep drainage outside the root zone. Nitrogen use efficiency (NUE) was
calculated based on (Moll et al., 1982; Sadras and Lemaire, 2014).
Page | 155
5.2.3 Model testing and Evaluation
Simulated model outputs were compared with observed values for crop yield, WUE and NUE for
all treatments and sites for the 2013-14 cropping season. The fit between observed and simulated
results was determined using the root mean squared error (RMSE), model efficiency (ME) and
the coefficient of determination (r2). The coefficient of determination r2 was calculated from the
regression of observed against simulated data. RMSE and ME were calculated through formulas
(1) and (2) respectively. A ME and r2 vary between -1 and +1. A ME as closes as possible to 1
indicates a good fit, while a negative ME suggests that the observed data are a better
representation of the situation in cause. ME close to zero represents as poor fit.
RMSE = (1)
And
ME = 1 - (2)
5.3. Results
5.3.1 Model evaluation
The model was able to reproduce the experimental results i.e. overall model efficiency (ME) of
0.64 and RMSE of 957.3 kg/ha; with a r2 of 0.75 between observed and predicted yields (Table
5-2). A separate analysis of model performance for each individual crop showed an overall yield
RMSE of 1131.8 kg/ha and 0.60 ME for maize. However, the model over predicted maize yield
for specific treatments (Figure 5-1), especially HNA in fSa and cSaCL. For cowpea model
performance was less impressive, i.e. with an overall ME of 0.37 and RMSE of 232.2 kg/ha.
However, over predictions of cowpea yield were observed for the cSaCL site. The model
represented better crop NUE than WUE (Table 5-2).
Page | 156
Table 5- 2 Overall model evaluation for selected parameters
Model attribute Number of data
points
r2 RMSE ME
Maize yield (kg/ha) 18 0.63 1131.8 0.60
Cowpea yield (kg/ha) 18 0.42 233.2 0.37
Overall Grain yield (kg/ha) 36 0.78 831.5 0.73
Water use efficiency (WUE, kg grain/ha.mm) 36 0.45 - -
Nitrogen Use Efficiency (NUE, kg grain/kg N-
applied)
36 0.81 4.912 0.80
Page | 157
Figure 5-1 Above - Comparison between observed and APSIM model simulated maize and cowpea
yield distributed across three sites. Shapes identification: ∆ - Sandy Clay loam (cSaCL); O Fine sandy
(fSa) and ◊ Sandy loam soil (SaL). Filled shapes for each soil type represent residue treatments while
empty shapes did not receive any residue. Middle and bellow (A, B and C): Overall comparison
between observed and simulated maize and cowpea yield (a), NUE (b) and WUE(c) distributed across
tested sites for. Continuous lines indicate a 1:1 fit and dotted lines indicate a best fit line between
observed and simulated.
Page | 158
5.3.2 Interactions between residue and fertilizer and their impact on yield
A crop and soil driven yield response to a high C/N ratio mulch and inorganic N fertilizer
application was simulated. In general, cowpea responded positively to residue application in all
environments, except in the drier and low water holding capacity sandy soil (fSa), were mixed
responses to residues were simulated across levels of N-application. For maize, responses to
residue application were affected by the starting soil N status and N-fertilizer application levels.
Soil moisture regimes also led to contrasting responses to residue and N-application.
Residue responses in high WHC and wetter soils
Maize yield losses were simulated in almost 85% to 80% of the seasons in ND and LNA systems
under residue application in the of the N-rich cSaCL (Figure 5-2). In HNA systems, yield losses
from residue application were simulated in 75% of the seasons with positive responses to
residues being expected in approximately 20 % of the seasons. Nevertheless, yield losses due to
high C/N residue application were reduced with increased N-application in this soil. Overall
maize yield losses of 11.2% were calculated for HNA systems, against 31.2% and 38.4% in LNA
and ND systems after high C/N residue application. Simulations results also showed that a 3 t/ha
yield is only likely to be attained in 10% and 40 % of the seasons in ND and LNA systems under
residue application but in almost 80% of the seasons in residue free ND and LNA systems. In
contrast, a 95% chance of achieving yields higher than 3 t/ha was simulated in HNA systems
under residue application. This was 5% higher than residue free systems under HNA. For the wet
sandy loam soil (SaL), were enough soil moisture to trigger a faster residue decomposition and
consequently N-immobilization, high C/N residue application led to an overall yield loss of
46.3%, 40.7% and 18.6% in ND, LNA and HNA maize systems over the simulation period. In
this soil, benefits from residue application were only attained in 5 % of the seasons in ND and
LNA systems. Nevertheless, maize yields in those seasons were lower than 1 t/ha. In HNA
systems, positive yield responses were simulated in 20% of the seasons under high C/N residue
application. Simulation results also showed that the likelihood of producing yields higher than 3
t/ha, was of approximately 80 % at all levels of N-application in residue free systems compared
to less than 1% and 20% in ND and LNA systems under residue application. In HNA systems, an
85% chance to produce yields higher than 3 t/ha yield plateau was simulated.
Page | 159
In both soils, positive responses to high C/N residue application were simulated in more than
90% of the seasons in ND, LNA and HNA systems for cowpea. Nevertheless, increased N-
applications have reduced cowpea yields. Cowpea yields higher than 1t/ha were attained in
almost 85% of the season in ND and LNA systems but in less than 80% in HNA systems in the
N-rich cSaCL. In the wet SaL, increased N-application have reduced cowpea yields to less than 1
t/ha in 85% in ND and LNA systems. In HNA systems, yields higher than 1 t/ha were only
attained in almost 75% of the seasons under residue application.
Figure 5-2 long term simulated maize and cowpea yield responses to residue application and N-
application levels (0, 23, 92 kg N/ha representing ND, LNA and HNA systems) in a high WHC
(cSaCL) and prolonged wetness (SaL) soils. No residue (R-); Residue applied (R+)
Page | 160
Residue responses in dry and low WHC soils
In dry and low WHC (fSa), yield increments from the application of high C/N ratio residues
were simulated for maize in almost 90% of the seasons at all N-application levels. Maize yields
were lowest at the fSa site (Figure 5-3). Nevertheless, residue application in this environment,
increased the chances of achieving yields higher than 1 t/ha, from less than 40%, 30% and 25%
of the seasons in residue free systems under ND, LNA and HNA, to more than 75% of the
seasons in all systems once residues were added. Overall maize yield increments of 43.3%,
22.2% and 16% in ND, LNA and HNA were attained under residue application in the fSa.
Despite the measured positive response to residue application, a 3 t/ha yield is likely to be
surpassed in less than 30% and 35% of the seasons in ND and LNA systems. A 50% chance of
surpassing the 3 t/ha mark was simulated for HNA systems under residue application in the drier
fSa.
For cowpea, simulated results showed almost 90 and 80% probability of surpassing 0.5t/ha
(SSA, average) and 1t/ha, respectively, were simulated under residue application in all sites.
Despite an overall positive response to residue application in cowpea, mixed responses to residue
were simulated in this drier and low water holding capacity soil. In unfertilized ND cowpea
systems, advantages from residue application, were only observed in 50% of the simulated
seasons. However, advantages from residues application were increased with levels of N-
application. A pre-assessment of crop performance based on residue application, would suggest
that for this site, when fertilizers are not available, investing residues to improve moisture
retention in the maize crop, is a more viable option that applying then into the legume crop
whose response to residues proved to increase with levels of N-application.
Page | 161
5.3 Resource productivity
Residue modified N-driven trade-off between NUE and WUE
A nitrogen driven trade-off between NUE and WUE was simulated (Figure 5-4). For both crops,
increased N-applications above 23 kg N/ha, reduced NUE and improved WUE. Increased water
availability across environments increased crop NUE especially for maize. This phenomenon can
be seen in the fSa-SaL and fSa-cSaCL transitions. The magnitude of the nitrogen driven NUE
and WUE trade-offs was dependent on individual sites, i.e. soil moisture and initial soil N-status.
Results also show that in LENIA systems, the N-driven NUE and WUE trade-offs is affected by
site and crop individual responses to the application of high C/N residues. Across all sites,
increased N-fertilizer applications reduced NUE but improved WUE (Figure 5-4).
Figure 5-3 Long term simulated maize and cowpea yield response to residue and N-application
levels (0, 23, 92 kg N/ha representing ND, LNA and HNA systems) in low WHC (fSa) soil. No
residues (R-), Residue applied (R+)
Page | 162
Figure 5-4 Residue modified nitrogen driven NUE and WUE trade-off simulated across maize and cowpea grown in contrasting farming
environments of central Mozambique. Overall means are presented for each level of N-application namely: unfertilized (ND, 0 kg N/ha),
Limited N-applications (LNA, 23 kg N/ha) and high N-application (HNA, 92 kg N/ha). Residue application levels are indicated as R- and R+
for no residue and residue applied plots respectively.
Page | 163
5.4. Seasonal rainfall variability and their effects on residue responses
Responses to residue application across crops and farming environments were found to be
largely affected by in-crop rainfall distribution patterns (Figure 5-5). For each environment,
positive and negative yield variations across N-application levels were simulated at specific
seasons. In the low WHC and drier fSa, positive contributions from residue application into
maize yield were simulated in most of the seasons, except in the wettest years, i.e., 1952, 1956,
1960, 1977, 1989 and 2001 which registered annual rainfall higher than 1000 mm. In the high
WHC cSaCL and wet SaL, were soil moisture regimes are considerable good enough to sustain
crop growth over the season, benefits from residue application were only simulated in the driest
(e.g.,1973, 1968, 1992 and 2003) and poorly distributed rainfall seasons (e.g., 1954, 1983 and
2010).
Page | 164
Figure 5-5 Above, in the rainfall distribution graphs (A, B and C) for the period 1951-2014, seasons are
categorized according to simulated maize yield variations under high C/N residue application at three
different soils. The horizontal line indicates annual average rainfall over the simulated period. Dark bars –
indicate seasons were yield losses from residue application were simulated; Red bars indicate seasons in
which positive responses to residue application are expected across N-application levels. In graphs A1 to
C3, relative maize yield variations to residue application disaggregated by soil and N-application levels
over the simulated period.
Page | 165
Table 5-3 Inter-seasonal performance evaluated through variations in long term simulated maize and cowpea water productivity (WUE) and
nitrogen use efficiency (NUE) response to residue application across N-levels for a low N and low water holding capacity environment.
Farming Environment N-level Systems
Performance Maize WUE Cowpea WUE Maize NUE Cowpea NUE
∆WUE (%) Seasons ∆WUE (%) Seasons ∆NUE (%) Seasons ∆NUE (%) Seasons
Arenosol
Low N and low WHC
(fSa)
0N Benefits 135 0.79 32.5 0.95 77.4 0.25 22.0 0.41
Losses -34.6 0.21 -12.0 0.05 -29.8 0.75 -3.38 0.59
23N Benefits 72.1 0.78 27.4 0.98 52.9 0.24 24.3 0.40
Losses -27.4 0.22 -19.9 0.02 -29.9 0.76 -3.30 0.60
92N Benefits 54.7 0.83 31.8 0.95 54.8 0.40 25.5 0.65
Losses -29.8 0.17 -9.49 0.05 -17.8 0.60 -2.68 0.35
Gleysol
Low N and high WHC
(SaL)
0N Benefits 70.4 0.18 64.6 1 51.9 0.37 33.2 0.94
Losses -37.8 0.82 - - -17.6 0.63 -2.94 0.06
23N Benefits 71.3 0.20 51.2 0.98 92.1 0.20 35.8 0.94
Losses -36.1 0.80 -0.95 0.02 -25.4 0.80 -4.01 0.06
92N Benefits 88.2 0.30 44.4 1 87.0 0.30 42.0 0.97
Losses -14.6 0.70 - - -16.6 0.70 -5.66 0.03
Ferralsol
High N and high WHC
(cSaCL)
0N Benefits 64.9 0.22 33.1 0.56 49.0 0.49 61.0 0.95
Losses -32.9 0.78 -19.1 0.44 -6.28 0.51 -2.38 0.05
23N Benefits 52.4 0.19 34.0 0.44 83.3 0.30 40.2 0.84
Losses -33.4 0.81 -13.6 0.56 -20.2 0.70 -0.90 0.16
92N Benefits 35.4 0.33 29.1 0.59 35.7 0.76 45.0 0.87
Losses -12.3 0.67 -11.1 0.41 -21.8 0.24 -2.26 0.13
Legend: Systems performance: Mean is the overall relative variation (%) of yield, WUE and NUE across residue application levels; Benefits and
losses, are averaged positive and negative variations in yield, WUE and NUE over the simulated period. Seasonal responses: presented as a ratio
between the number of seasons with calculated benefits over the all spectrum of simulated seasons (N/64), Benefits and losses are averaged across
seasons were positive and negative variations in simulated yield, WUE and NUE were calculated for each specific environment
Page | 166
5.5 Discussion
APSIM suitability to simulate residue-fertilizer interactions in N-deprived systems
The overall performance of APSIM in reproducing observed maize and cowpea yield
responses to residue mulch and fertilizer application across farming environments was good.
However, in certain circumstances, the model either over predicted or under predicted
individual treatments responses. Example, for the specific case of the cSaCL, a response to
N-application was simulated in the absence of residues in both years contradicting the poor
response to N-application attained in the field during the first season. However, long-term
simulations showed that in low and poorly distributed rainfall seasons a similar result is likely
to occur in this environment. In the wet but low N environment (SaL), a good agreement
between model and observed data was achieved. In this environment, positive responses to
N-application in both situations, with and without residues were also simulated. Nevertheless,
a lower response was simulated in the transition between ND and LNA systems in both
residue levels.
In the low N and drier environment (fSa), the model has done well at lower N-levels but did
not quite well mimicked trial results at higher N-level. Poor APSIM performance in water
limited environments was reported before (Balwinder et al., 2011). For this specific case,
there is a need to mention that, a single crop soil parameterization, i.e., maize was used,
letting all alone the model to capture cowpea responses for the same conditions. This
somehow would likely affect model predictions for cowpea. However, it was not the case,
since a better model agreement was simulated in cowpea. Despite the discrepancy between
observed and simulated yield in some treatments, there was an overall good agreement
between observed and simulated responses in all tested environments, which makes APSIM a
candidate tool for ex-ante and post-ante analysis that are residue management and maize-
legume cropping systems design in Central Mozambique and SSA as shown before
(Akponikpè et al., 2010; Mohanty et al., 2010; Ncube et al., 2009). Nevertheless, model
accuracy can still be improved, in order to make the model an integral part of local research
systems. However, improved crop phenology descriptors for locally grown cereals and
especially legume varieties need to be developed.
Page | 167
Understanding residue induced responses through seasonal yield, WUE, NUE variations
Simulation results showed that there is almost 85 % chance that residue application will
reduce maize yield in rain fed LENIA systems of central Mozambique, especially in ND and
LNA maize crops cultivated in considerably wet and high WHC soils. In the AEZ-R4, which
has an average annual rainfall of 834 mm (1951-2014), appears to provide enough soil
moisture to trigger residue decomposition and consequently a N-immobilization induced
plant N-stress which leads to maize yield losses especially in wet and high WHC soils
(Figure 5-5).
In a normal rainfall season no significant water stress is likely to occur throughout the crop
cycle, especially in the predominantly cultivated soil - red cSaCL ferralsol in AEZ-R4.
Therefore, maize yield losses are derived from poor NUE resulting from N-immobilization
induced plant N-stress under residue application (Table 5-3). However, in poor rainfall
seasons, i.e., less than 20% over the simulated period, both water and N-stress are likely to
occur. Nevertheless, residue application is likely to improve soil water availability and
consequently yields only in poor rainfall seasons – low rainfall and badly distributed in-crop
rainfall years.
Poor responses from C rich residues are likely to occur across most central Mozambique
agro-ecologies, namely the medium altitude – AEZ-R5, AEZ-R7 and AEZ-R8 and in the
central Mozambique highlands (AEZ-R10) which have rainfall regimes between 800-1000
mm/year and above 1200 mm/year respectively (Reddy, 1984). Here, more pragmatic residue
application strategies need to be considered to take advantage of residues in the system. In the
semi-arid regions of northern Manica province and south Tete provinces where lays the R6
agro-ecology with annual rainfall between 400-600 mm, benefits from residue application are
likely to occur as simulated for the drier and low WHC fSa soils. In drier and low WHC soils,
where there is not enough soil moisture to trigger a faster residue decomposition, residues are
likely to help increase maize WUE at all levels of N-application, and consequently increase
yields (Proposition 2).
Page | 168
Figure 5-6 Nitrogen and water stress levels in maize across residue application levels for a
high WHC environment in a normal and poor rainfall season. A stress factor of 1 indicates
full stress and 0 absence of stress. Stress indexes: npho is the N-stress and wpho is water
stress.
Maize yield losses, under residue application in considerably wet environments, are
associated to the high moisture induced residue decomposition, which leads to N-
immobilization. In high WHC and clay rich soils, gross N-immobilization lasting between
14-545 days were reported (Gentile et al., 2008; Recous et al., 1995; Sakala et al., 2000).
High moisture and microbial activity in those soils trigger a fast residue decomposition
leading to N-immobilization at low N-levels (Guntiñas et al., 2012; Tian et al., 1992). Chen et
al. (2012), reported high N-mineralization and N-immobilization during residue
decomposition in soils at 75% WHC compared to soils at 35% WHC.
Contrarily to high WHC soils, poor microbial activity in low WHC soils, i.e., dry
environments like AEZ-R6, delays residue decomposition allowing them to stay longer in the
soil providing further surface cover and soil moisture benefits to the crops. As a result,
positive yield responses mainly due to increased WUE are expected in such circumstances as
the simulation results also showed for the fSa. Nevertheless, a risk of N-immobilization is
also present and residue application are likely to reduce maize NUE. However, losses from
NUE do not surpass the gains from WUE in this environment. Benefits from residue
application, in drier and coarse textured low WHC soil such as the drier fSa were also
Page | 169
reported by Gentile et al. (2008). Improved maize yields under residue application were also
reported under drought conditions in the Central Mexico Highlands (Romero-Perezgrovas et
al., 2014), strengthening our finding that high C/N ratio residues can be used to improve
water capture and increase yields in considerably dry soils and years.
Experimental data has also showed that yield losses from N-immobilization during residue
decomposition in N-deprived systems might be alleviated by improved residue management
strategies. Increased inorganic N-application especially in wet and high WHC soils, was
reported to reduce N-immobilization and improve N-release to the crops (Esther et al., 2014;
Recous et al., 1995; Ribeiro et al., 2002). Nevertheless, results also showed that the
magnitude of the negative effects from high C/N ratio residues application on maize yields,
WUE and NUE was largely dependent on the initial soil N status and could be alleviated by
increased levels of N-application in wet environments (Proposition 4). In most field trials,
however positive, yield and resource productivity attained under high C/N ratio residue
application where reported at considerably high N-application levels, i.e., applications above
80 kg N/ha or more in N-rich environments where inorganic N-supply levels are high enough
to partially or fully satisfy N-immobilization demands during residue decomposition (Bakht
et al., 2009; Baudron et al., 2015a; Tao et al., 2015; Verhulst et al., 2011; Xiukang et al.,
2015). However, this is not the case of most resource poor farmers in central Mozambique
and SSA.
An alternative to inorganic fertilizer whose use is considerable low among resource poor
smallholder farmers of Central Mozambique, is shifting high C/N ratio residue mulch
application into legumes crops. This presents as more affordable option to improve soil N-
availability in the system once positive yield, NUE and WUE responses to high C/N ratio
residues were simulated in almost 50-90 % of the seasons with an early sown cowpea crop
across tested soils in the AEZ-R4. For example, a 13% grain yield benefit in mungbean under
wheat residue application were reported by Shah et al. (2003b). In the same trial, the
percentage of mungbean N-derived from N2-fixation was increased by 55.6% and a +64 kg
N/ha average N-balance was achieved under residue application. Positive yield responses and
high N-turnover from N-rich residues returned to the system from legume crops grown under
high C/N ratio residue applied plots, were reported in separate studies (Luce et al., 2015;
Monzon et al., 2014; Neugschwandtner and Kaul, 2015). These results, prove that using
Page | 170
residues to improve legume performance in maize dominated systems, might be a viable
option to improve N-availability in smallholder farming systems. This is a critical result for
central Mozambique maize dominated systems, where legume incorporation is key to help
reduce yield penalties from continuous maize systems under conservation agriculture.
5.6 Conclusion
As field trials can only capture a limited part of the vast and complex interactions between
crops, farming environments and their response to several management conditions. Model
assisted research and decision support is critical especially in countries where research is
scantly funded. Therefore, owning to its simplicity, time and cost effectiveness, the model
assisted decision support framework proposed here can be useful to capture not only crop and
site-specific responses to residue and fertilizer allocation strategies, but also the trade-offs
characteristic of central Mozambique complex farming environment mosaic. Nevertheless,
despite APSIM’s acceptable performance in simulating the complex residue and fertilizer
interaction and their effects on crop yields, WUE and NUE, model accuracy can still be
improved to make it a reliable and integral component of local research and extension
systems.
Acknowledgements
Would like to thank the SIMLESA project and the Australian Centre for International
Agriculture Research (ACIAR) for the scholarship under which this research was conducted.
The Soil Health Program of the Alliance for Green Revolution in Africa (AGRA) through the
2011SHP007 Project for helping fund the fieldwork. Thanks also to Peter De Voil (APSRU-
DAF-QAAFI) for support with APSIM modelling.
References
Akponikpè, P. B. I., Gérard, B., Michels, K., and Bielders, C. (2010). Use of the APSIM
model in long term simulation to support decision making regarding nitrogen
management for pearl millet in the Sahel. European Journal of Agronomy 32, 144-
154.
Page | 171
Bakht, J., Shafi, M., Jan, M. T., and Shah, Z. (2009). Influence of crop residue management,
cropping system and N fertilizer on soil N and C dynamics and sustainable wheat
(Triticum aestivum L.) production. Soil and Tillage Research 104, 233-240.
Balwinder, S., Gaydon, D. S., Humphreys, E., and Eberbach, P. L. (2011). The effects of
mulch and irrigation management on wheat in Punjab, India—Evaluation of the
APSIM model. Field Crops Research 124, 1-13.
Baudron, F., Delmotte, S., Corbeels, M., Herrera, J. M., and Tittonell, P. (2015a). Multi-scale
trade-off analysis of cereal residue use for livestock feeding vs. soil mulching in the
Mid-Zambezi Valley, Zimbabwe. Agricultural Systems 134, 97-106.
Baudron, F., Thierfelder, C., Nyagumbo, I., and Gérard, B. (2015b). Where to Target
Conservation Agriculture for African Smallholders? How to Overcome Challenges
Associated with its Implementation? Experience from Eastern and Southern Africa.
Environments 2, 338.
Bouma, J., and Jones, J. W. (2001). An international collaborative network for agricultural
systems applications (ICASA). Agricultural Systems 70, 355-368.
Chen, Q.-H., Feng, Y., Zhang, Y.-P., Zhang, Q.-C., Shamsi, I. H., Zhang, Y.-S., and Lin, X.-
Y. (2012). Short-Term Responses of Nitrogen Mineralization and Microbial
Community to Moisture Regimes in Greenhouse Vegetable Soils. Pedosphere 22,
263-272.
Chikowo, R., Corbeels, M., Tittonell, P., Vanlauwe, B., Whitbread, A., and Giller, K. E.
(2008). Aggregating field-scale knowledge into farm-scale models of African
smallholder systems: Summary functions to simulate crop production using APSIM.
Agricultural Systems 97, 151-166.
Cox, H. W., Kelly, R. M., and Strong, W. M. (2010). Pulse crops in rotation with cereals can
be a profitable alternative to nitrogen fertiliser in central Queensland. Crop and
Pasture Science 61, 752-762.
Dalgliesh, N., and Foale, M. (1998). "Soil matters: Monitoring soil water and nutrients in
dryland farming," 1/Ed. CSIRO, Australia, Toowoomba, Australia.
Esther, O. J., Guo, C.-h., Tian, X.-h., Li, H.-y., and Zhou, Y.-x. (2014). The Effects of Three
Mineral Nitrogen Sources and Zinc on Maize and Wheat Straw Decomposition and
Soil Organic Carbon. Journal of Integrative Agriculture 13, 2768-2777.
Page | 172
Gentile, R., Vanlauwe, B., Chivenge, P., and Six, J. (2008). Interactive effects from
combining fertilizer and organic residue inputs on nitrogen transformations. Soil
Biology and Biochemistry 40, 2375-2384.
Gilbert, N. (2012). Dirt Poor: The key to tackle hunger in Africa is enriching its Soils. Nature
483, 525-227.
Giller, K. E., Corbeels, M., Nyamangara, J., Triomphe, B., Affholder, F., Scopel, E., and
Tittonell, P. (2011). A research agenda to explore the role of conservation agriculture
in African smallholder farming systems. Field Crops Research 124, 468-472.
Giller, K. E., Witter, E., Corbeels, M., and Tittonell, P. (2009). Conservation agriculture and
smallholder farming in Africa: The heretics’ view. Field Crops Research 114, 23-34.
Guntiñas, M. E., Leirós, M. C., Trasar-Cepeda, C., and Gil-Sotres, F. (2012). Effects of
moisture and temperature on net soil nitrogen mineralization: A laboratory study.
European Journal of Soil Biology 48, 73-80.
Holzworth, D. P., Huth, N. I., deVoil, P. G., Zurcher, E. J., Herrmann, N. I., McLean, G.,
Chenu, K., van Oosterom, E. J., Snow, V., Murphy, C., Moore, A. D., Brown, H.,
Whish, J. P. M., Verrall, S., Fainges, J., Bell, L. W., Peake, A. S., Poulton, P. L.,
Hochman, Z., Thorburn, P. J., Gaydon, D. S., Dalgliesh, N. P., Rodriguez, D., Cox,
H., Chapman, S., Doherty, A., Teixeira, E., Sharp, J., Cichota, R., Vogeler, I., Li, F.
Y., Wang, E., Hammer, G. L., Robertson, M. J., Dimes, J. P., Whitbread, A. M., Hunt,
J., van Rees, H., McClelland, T., Carberry, P. S., Hargreaves, J. N. G., MacLeod, N.,
McDonald, C., Harsdorf, J., Wedgwood, S., and Keating, B. A. (2014). APSIM –
Evolution towards a new generation of agricultural systems simulation.
Environmental Modelling and Software 62, 327-350.
Homann-Kee Tui, S., Valbuena, D., Masikati, P., Descheemaeker, K., Nyamangara, J.,
Claessens, L., Erenstein, O., van Rooyen, A., and Nkomboni, D. (2015). Economic
trade-offs of biomass use in crop-livestock systems: Exploring more sustainable
options in semi-arid Zimbabwe. Agricultural Systems 134, 48-60.
Jordan, V. W. L., Hutcheon, J. A., Donaldson, G. V., and Farmer, D. P. (1997). Research into
and development of integrated farming systems for less-intensive arable crop
production: experimental progress (1989–1994) and commercial implementation.
Agriculture, Ecosystems and Environment 64, 141-148.
Kintché, K., Guibert, H., Sogbedji, J. M., Levêque, J., Bonfoh, B., and Tittonell, P. (2015).
Long-term mineral fertiliser use and maize residue incorporation do not compensate
Page | 173
for carbon and nutrient losses from a Ferralsol under continuous maize–cotton
cropping. Field Crops Research 184, 192-200.
Luce, M., Grant, C. A., Zebarth, B. J., Ziadi, N., O’Donovan, J. T., Blackshaw, R. E., Harker,
K. N., Johnson, E. N., Gan, Y., Lafond, G. P., May, W. E., Khakbazan, M., and
Smith, E. G. (2015). Legumes can reduce economic optimum nitrogen rates and
increase yields in a wheat–canola cropping sequence in western Canada. Field Crops
Research 179, 12-25.
Masikati, P., Manschadi, A., van Rooyen, A., and Hargreaves, J. (2014). Maize–mucuna
rotation: An alternative technology to improve water productivity in smallholder
farming systems. Agricultural Systems 123, 62-70.
Matthews, R. B., and Stephens, W. P. D. (2002). "Crop-soil simulation models: applications
in developing countries," CABI Pub, Wallingford, England.
Maxwell, S. (1986). Farming systems research: Hitting a moving target. World Development
14, 65-77.
McCown, R. L., Hammer, G. L., Hargreaves, J. N. G., Holzworth, D. P., and Freebairn, D.
M. (1996). APSIM: a novel software system for model development, model testing
and simulation in agricultural systems research. Agricultural Systems 50, 255-271.
Meinke, H., Baethgen, W. E., Carberry, P. S., Donatelli, M., Hammer, G. L., Selvaraju, R.,
and Stöckle, C. O. (2001). Increasing profits and reducing risks in crop production
using participatory systems simulation approaches. Agricultural Systems 70, 493-513.
Mohanty, M., Probert, M. E., Reddy, K. S., Dalal, R. C., Rao, A. S., and Menzies, N. W.
(2010). Modelling N mineralization from high C/N rice and wheat crop residues.
Proceedings of the 19th World Congress of Soil Science: Soil solutions for a changing
world, Brisbane, Australia, 1-6 August 2010. Symposium 1.5.2 Modelling critical
processes in changing soil, 28-31.
Moll, R. H., Kamprath, E. J., and Jackson, W. A. (1982). Analysis and Interpretation of
Factors Which Contribute to Efficiency of Nitrogen Utilization1. Agron. J. 74, 562-
564.
Monzon, J. P., Mercau, J. L., Andrade, J. F., Caviglia, O. P., Cerrudo, A. G., Cirilo, A. G.,
Vega, C. R. C., Andrade, F. H., and Calviño, P. A. (2014). Maize-soybean
intensification alternatives for the Pampas. Field Crops Research 162, 48-59.
Page | 174
Mupangwa, W., Dimes, J., Walker, S., and Twomlow, S. (2011). Measuring and simulating
maize (Zea mays L.) yield responses to reduced tillage and mulching under semi-arid
conditions. Agricultural Sciences 2, 167-174.
Ncube, B., Dimes, J., Twomlow, S., Mupangwa, W., and Giller, K. (2007). Raising the
productivity of smallholder farms under semi-arid conditions by use of small doses of
manure and nitrogen: a case of participatory research. Nutrient Cycling in
Agroecosystems 77, 53-67.
Ncube, B., Dimes, J. P., van Wijk, M. T., Twomlow, S. J., and Giller, K. E. (2009).
Productivity and residual benefits of grain legumes to sorghum under semi-arid
conditions in south-western Zimbabwe: Unravelling the effects of water and nitrogen
using a simulation model. Field Crops Research 110, 173-184.
Neugschwandtner, R. W., and Kaul, H.-P. (2015). Nitrogen uptake, use and utilization
efficiency by oat–pea intercrops. Field Crops Research 179, 113-119.
Ngwira, A. R., Thierfelder, C., and Lambert, D. M. (2013). Conservation agriculture systems
for Malawian smallholder farmers: long-term effects on crop productivity,
profitability and soil quality. Renewable Agriculture and Food Systems 28, 350-363.
Pala, M., Ryan, J., Zhang, H., Singh, M., and Harris, H. C. (2007). Water-use efficiency of
wheat-based rotation systems in a Mediterranean environment. Agricultural Water
Management 93, 136-144.
Probert, M. E., Dimes, J. P., Keating, B. A., Dalal, R. C., and Strong, W. M. (1998). APSIM's
water and nitrogen modules and simulation of the dynamics of water and nitrogen in
fallow systems. Agricultural Systems 56, 1-28.
Recous, S., Robin, D., Darwis, D., and Mary, B. (1995). Soil inorganic N availability: Effect
on maize residue decomposition. Soil Biology and Biochemistry 27, 1529-1538.
Reddy, S. J. (1984). Clima Geral de Mocambique In "Serie Terra e Água" (INIA, ed.), pp. 41.
Instituto de Investigação Agronómica de Moçambique (INIA), Departamento de Terra
e Agua (DTA), Maputo, Mozambique.
Ribeiro, C., Madeira, M., and Araújo, M. C. (2002). Decomposition and nutrient release from
leaf litter of Eucalyptus globulus grown under different water and nutrient regimes.
Forest Ecology and Management 171, 31-41.
Rodriguez, D., deVoil, P., Power, B., Cox, H., Crimp, S., and Meinke, H. (2011). The
intrinsic plasticity of farm businesses and their resilience to change. An Australian
example. Field Crops Research 124, 157-170.
Page | 175
Romero-Perezgrovas, R., Verhulst, N., De La Rosa, D., HernÁNdez, V., Maertens, M.,
Deckers, J., and Govaerts, B. (2014). Effects of Tillage and Crop Residue
Management on Maize Yields and Net Returns in the Central Mexican Highlands
Under Drought Conditions. Pedosphere 24, 476-486.
Sadras, V. O., and Lemaire, G. (2014). Quantifying crop nitrogen status for comparisons of
agronomic practices and genotypes. Field Crops Research 164, 54-64.
Sadras, V. O., and Rodriguez, D. (2010). Modelling the nitrogen-driven trade-off between
nitrogen utilisation efficiency and water use efficiency of wheat in eastern Australia.
Field Crops Research 118, 297-305.
Sakala, W. D., Cadisch, G., and Giller, K. E. (2000). Interactions between residues of maize
and pigeon pea and mineral N fertilizers during decomposition and N mineralization.
Soil Biology and Biochemistry 32, 679-688.
Shah, Z., Shah, S. H., Peoples, M. B., Schwenke, G. D., and Herridge, D. F. (2003). Crop
residue and fertiliser N effects on nitrogen fixation and yields of legume–cereal
rotations and soil organic fertility. Field Crops Research 83, 1-11.
Sommer, R., Thierfelder, C., Tittonell, P., Hove, L., Mureithi, J., and Mkomwa, S. (2014).
Fertilizer use should not be a fourth principle to define conservation agriculture:
Response to the opinion paper of Vanlauwe et al. (2014) ‘A fourth principle is
required to define conservation agriculture in sub-Saharan Africa: The appropriate use
of fertilizer to enhance crop productivity’. Field Crops Research 169, 145-148.
Tao, Z., Li, C., Li, J., Ding, Z., Xu, J., Sun, X., Zhou, P., and Zhao, M. (2015). Tillage and
straw mulching impacts on grain yield and water use efficiency of spring maize in
Northern Huang–Huai–Hai Valley. The Crop Journal 3, 445-450.
Thierfelder, C., Cheesman, S., and Rusinamhodzi, L. (2012). A comparative analysis of
conservation agriculture systems: Benefits and challenges of rotations and
intercropping in Zimbabwe. Field Crops Research 137, 237-250.
Thierfelder, C., Matemba-Mutasa, R., and Rusinamhodzi, L. (2015). Yield response of maize
(Zea mays L.) to conservation agriculture cropping system in Southern Africa. Soil
and Tillage Research 146, Part B, 230-242.
Thierfelder, C., Mombeyarara, T., Mango, N., and Rusinamhodzi, L. (2013). Integration of
conservation agriculture in smallholder farming systems of southern Africa:
identification of key entry points. International Journal of Agricultural Sustainability
11, 317-330.
Page | 176
Tian, G., Kang, B. T., and Brussaard, L. (1992). Biological effects of plant residues with
contrasting chemical compositions under humid tropical conditions—Decomposition
and nutrient release. Soil Biology and Biochemistry 24, 1051-1060.
Twomlow, S., Rohrbach, D., Dimes, J., Rusike, J., Mupangwa, W., Ncube, B., Hove, L.,
Moyo, M., Mashingaidze, N., and Mahposa, P. (2010). Micro-dosing as a pathway to
Africa’s Green Revolution: evidence from broad-scale on-farm trials. Nutrient
Cycling in Agroecosystems 88, 3-15.
Vanlauwe, B., Wendt, J., Giller, K. E., Corbeels, M., Gerard, B., and Nolte, C. (2014). A
fourth principle is required to define Conservation Agriculture in sub-Saharan Africa:
The appropriate use of fertilizer to enhance crop productivity. Field Crops Research
155, 10-13.
Verhulst, N., Govaerts, B., Nelissen, V., Sayre, K. D., Crossa, J., Raes, D., and Deckers, J.
(2011). The effect of tillage, crop rotation and residue management on maize and
wheat growth and development evaluated with an optical sensor. Field Crops
Research 120, 58-67.
Whitbread, A. M., Robertson, M. J., Carberry, P. S., and Dimes, J. P. (2010). How farming
systems simulation can aid the development of more sustainable smallholder farming
systems in southern Africa. European Journal of Agronomy 32, 51-58.
Xiukang, W., Zhanbin, L., and Yingying, X. (2015). Effects of mulching and nitrogen on soil
temperature, water content, Nitrate-N content and maize yield in the Loess Plateau of
China. Agricultural Water Management 161, 53-64.
Page | 177
CHAPTER 6
General Discussion
The most important thing in triggering any kind of change is by first finding people’s own
passion to grow. Once we find that passion our responsibility is to provide them with the
right kind of knowledge and technical support to materialize their passion! For that to
happen we need to listen and not force our will into people. (Ernesto Sirolli, in Ted Talks
2015)
(Agricultural intervention is not an exception to that rule)
Page | 178
6.1 Summary of key findings
Over the last decades, residue retention and fertilizer application have been widely promoted
across Africa. Crop residues are used as part of a wider conservation agriculture (CA)
platform that also includes rotation with legumes and minimum soil disturbance (Wall, 2007).
In the short term fertilizers appear to be the best way to improve yields in Africa (Chianu et
al., 2012; Gilbert, 2012); though there are many situations where the degree of soil
degradation is such that fertilisers alone will not be enough to lift crop productivity(Tittonell
et al., 2007a). CA appears like an optimum solution then in the view of researchers (Ngwira
et al., 2013; Thierfelder et al., 2012; Thierfelder et al., 2015) while adoption of these
practices continues to be a low priority for resource poor smallholder farmers in Africa. So,
compared to the rest of the world, adoption of CA principles in Sub Saharan Africa (Baudron
et al., 2015c), as that of fertilizers (Heffer and Prud'homme, 2011) remains negligible. This
raises questions over the suitability of these technologies and the way they are promoted for
the complex and diverse African situations.
Single-sized technological packages have been widely promoted (Giller et al., 2010), based
on results from researcher-managed trials. Nevertheless, these trials have failed to
demonstrate the how, where, when and what technology would fit what situation. This
because field trials alone have made it harder to scrutinize the real source of reported CA
benefits making most results reported to date not generalizable since they fail to capture the
complex of the system and situation where CA is promoted. The complexity of the situation
involves not only the need to understand numerous components of the system i.e., farmer’s
endowment levels, access to inputs and markets, availability and quality of mulches, crop and
sites responses across seasons; but also risk perceptions, values, and aspirations of poorly
resourced farmers. In addition to that, the use of inorganic N-fertilizers a fundamental input
to validate CA and increase yields is SSA is prohibitive to cash constrained farmers (Chianu
et al., 2012; Rware et al., 2014), and is making fertilizer use less attractive to most farmers.
Despite resource access being essential, recent research has demonstrated that rather than
access itself, optimizing resource management, i.e., allocation across farms, crops and
locations (Bucagu et al., 2014; Giller et al., 2006; Rusinamhodzi et al., 2016) may play a key
role in improving the performance of the highly diverse and complex SSA farming systems
(Tittonell et al., 2010; Tittonell et al., 2007a; Vanlauwe et al., 2006). Therefore, this study
Page | 179
investigated: 1) how farmer’s circumstances and perception of their biophysical and
socioeconomic circumstances affect their resource use decisions and farming systems design
and management; 2) how can carbon rich residues and available limited amounts of N-
fertilizers be optimally used in rain fed N-deprived maize-legume cropping systems of central
Mozambique. This study was conducted around four key propositions:
1) There are household and site-specific managements e.g. fertilizers, manure, residues
and legumes, that best fit farmers’ bio-physical and socioeconomic circumstances.
2) The balance between the positive and negative effects from the use of carbon rich
residues in maize-legume systems can only be managed with nitrogen fertilisers.
3) In wetter nitrogen constrained environments, investing the limited availability of high
C/N ratio residues on improving the performance of the sole legume crop is more
beneficial than applying those residues on the maize crop. The benefits from this
resource allocation strategy are likely to be observed in the following cereal crop.
4) In dryer nitrogen constrained agricultural systems, investing the limited availability of
maize stubble on improving rainfall capture for the maize crop is likely to provide
more benefits than applying those residues on the legume crop.
Each of the results chapters, i.e., 3, 4 and 5, address one or a set of the propositions presented
above under the assumption that they are the key to optimize resource allocation strategies
and their responses. The main hypothesis in the thesis is that the likelihood to sustainably
improving resource productivity, yields, and the household food and income situation in the
N-deprived systems of central Mozambique, can be increased through a better allocation of
available resources. However, this can only be done through interactive co-design of new
practices with active farmer participation from the onset of the project. Furthermore, it is
hoped that the results from this research will be widely applicable to a range of agro-
ecologies of central Mozambique and across SSA since there is little generalizable
information that can help design sustainable intensification intervention that are more likely
to be adopted by poorly resourced farmers.
Page | 180
6.2 Typology tailored intervention: key entry points to downscale technologies
The results in chapter 3, show that, understanding the drivers for smallholder farmers’
decisions in the quest to meet seasonal food security and income generation targets is
paramount to learn how farmers can be actively involved in the co-design of more relevant
and actionable management options. As African smallholder farms are highly diverse in
space and time, grouping farmers in small homogeneous functional groups or typologies can
help target interventions so that adoption is more likely to take place (Daloğlu et al., 2014;
Dorward et al., 2009; Guillem et al., 2012). For the specific case of Mozambique, exploring
the intra-group diversity among small scale farmers in the existing rigid categorization (INE,
2011b) is key to develop recommendations that are more representative of farmers
circumstances. Here we proposed a framework to categorize farmers in small mutually
exclusive functional typologies.
The typologies proposed in this study are flexible in time and are in line with findings
showing how contrasting households access knowledge (Kraaijenbrink and Wijnhoven,
2008); and understanding of their levels of resources endowment to achieve food self-
sufficiency (Bucagu et al., 2014; Guillem et al., 2012) affected the way farmers designed and
managed their farms. Nevertheless, in central Mozambique, the relation between farmer’s
perception of their biophysical and socioeconomic circumstances and the way it affects local
farming systems design and management (Zingore et al., 2007) translated into technology
adoption traps. The adoption traps, are perception based mind-sets that reflects farmers
understanding of their circumstances and the farm management strategies they choose to
implement and believed are more likely to help them achieve their seasonal food security and
income generation targets considered their current circumstances. These management choices
are made in detriment of existing technological recommendation – fertilizers. Three key
adoption traps were identified across sampled typologies in central Mozambique, i.e.,
farmers’ perception about the level of soil fertility in their farm, the degree of maize self-
sufficiency, and availability of land i.e. farm size.
The results from chapter 3 also validate the proposition 1, that, there might be household
(typology) and site-specific management combinations of technologies e.g. fertilizers,
manure, residues, legumes and resources usage that best fit farmers’ bio-physical and
socioeconomic circumstances. For example, farmers’ investment on improved hybrid seed
Page | 181
planted mainly in the lowland mixed vegetable-maize farms that are normally fertilized since
farmers believe to have better returns from investments, is in line with the findings from
(Vanlauwe and Giller, 2006), who showed that farmers invest when they believe there is a
clear value-to-cost ratio. In the meanwhile, resource poor farmers, choose to diversify their
crops and cultivate more plots a season or engage in on and off the farm jobs in order to
improve their food and income prospects (Cunguara et al., 2011; Sichoongwe et al., 2014).
Therefore, understanding and taking these combinations of factors and management
circumstances into account is key to actively involve farmers in scaling out knowledge
intensive technologies such as the conservation agriculture principles and sustainable
intensification into locally feasible practices that are more likely to be tested and adopted by
farmers. However, it is also clear that given the diversity in households’ sources of
livelihoods there is no single-sized technology that will fit all households. Recent studies
revealed that factors such as the adopter characteristics (Meijer et al., 2015; Muzari et al.,
2012) are critical determinants of adoption.
Results from the household survey also showed that despite labour constraints (Leonardo et
al., 2015), farmers opt for extensive farming as a mean to minimize risk and achieve a
surpluses. Nevertheless, the average cultivated land area was usually not larger than 3ha/year
in our sampled population. Plot size is an issue to be further studied across typologies. Here,
identifying economic plot sizes for each typology is a critical step towards making farms
manageable and profitable (Ali et al., 2015). For example, a rough estimate indicates that
each household needs to produce a surplus of at least 7.8 ton/maize/year to generate enough
income to pay for 12 months of basic salary in the agricultural sector, i.e., ca. 960 USD/year
at a rate of 80 USD/month to the household head. Unfortunately, this cannot be achieved
from 0.5-1.0 ha of a maize monoculture with current varieties and levels of resource use.
Diversification of crops, income sources and cultivated land are the only ways this can be
achieved. The likelihood of producing surplus or generate the income from other farm and
off-farm activities will differ across groups. Therefore, an exhaustive analysis of household
dynamics and the trade-offs associated to different management strategy at both plot and
household level, needs to be considered in order to design resource efficient and profitable
systems capable of providing farmers with what they need.
Page | 182
6.3 Optimizing responses through smart and pragmatic resource allocation strategies
6.3.1 Residue and fertilizer responses: linking empirical and model assisted analysis
In chapter 4, the results clearly demonstrate that in the rain fed N-deprived systems of Central
Mozambique where there is enough soil moisture to trigger residue decomposition, a more
pragmatic and cautious approach needs to be taken when recommending residue application.
Field trial results showed that high C/N residues are better used to improve the performance
of the legume crop during the rotation phase rather than applied directly into an unfertilized
or poorly fertilized maize crop, where yield losses due to residue application were registered.
Nevertheless, positive responses to crop residue application into maize were attained with
increased N-applications. HNA, i.e., 92 kg N/ha, an amount far from the reach of most
resource poor farmers proved effective in offsetting the N-immobilization demand, allowing
enough N to be available to plant growth in tested seasons.
Previous research has also demonstrated that yield losses in CA cropping systems can be
alleviated by using inorganic N-application, which reduces N-immobilization demands and
improves N-release to the crop (Esther et al., 2014; Recous et al., 1995; Ribeiro et al., 2002).
Inorganic N-applications of around 80-100 kg N/ha in soils with considerable soil N content
were considered good enough to partially or fully off-set N-immobilization (Bakht et al.,
2009; Baudron et al., 2015a; Tao et al., 2015; Verhulst et al., 2011; Xiukang et al., 2015).
However, this is not the case for most resource poor households from central Mozambican.
Therefore, applying high C residues to legumes and using the limited amounts available of N
(ca. 23 kg N/ha) on the maize crop, appears to be the best resource allocation option for the
N-deprived maize-legume systems managed under conservation agriculture. This is because
better returns from residue and N-fertilizer application were measured in the legume-maize
sequence after just one season. This shift in practice is supported by the cowpeas positive
response to residues, as also reported in other studies before (Herrmann et al., 2014; Krishna,
2001; Shah et al., 2003a). In maize, fertilizer application at lower levels should only be
recommended in residue free systems, where better responses are attained.
Maize proven to be more responsive to increased N-application than cowpea. Cowpea yield
was actually reduced by N-applications above 23 kg N/ha. Despite, positive responses to N-
application in maize, responses differed across sites, with the wet sandy loam soil (SaL)
Page | 183
showing the highest overall agronomic response. The high returns from N-fertilizer
application in the wetter environments (SaL) rather than water limited fSa and cSaCL sites, is
in line with results from other studies (Cossani et al., 2010, 2012; Shangguan et al., 2000;
Sun et al., 2012; Teixeira et al., 2014). In these studies, strong interactive and co-limitation
effects between nitrogen use and soil water availability were observed. This is an important
finding, considering the high intra and cross-agroecological zone soil diversity in Central
Mozambique that implies that smarter fertilizer allocation strategies are required. Farmers
preferential fertilizer allocation into the wet lowland farming environment (Chapter 3), where
fertilizers are believed to offer better economic returns is in line with the interactive relation
between N-responses and water availability.
Fertilizer responses and productivity analysis demonstrated that the highest return from each
kg of N applied (NUE) was achieved under limited N-application (LNA), i.e., 23 kg N/ha
compared to 92 kg N/ha, in all the tested environments. This result is important in regards to
raising awareness about the importance of understanding the value-to-cost ratio of existing
blanket fertilizer recommendations in such unreliable rainfall environments (Dimes, 2011).
Nitrogen driven trade-offs between NUE and WUE were also observed (Sadras and
Rodriguez, 2010). In maize, increased N-application increased WUE and reduced WUE,
while increased water availability increased NUE to the detriment of NUE. A strong
relationship between NUE and its components was also observed with site moisture regime
playing a detrimental role in NUpE and its contribution to NUE. NUE in maize was highly
affected by variation in NUpE, while for cowpea NUtE had the highest impact on NUE,
especially at higher levels of N-application. Understanding all this interactions, resource
productivity trade-offs and the responses to several management conditions (e.g., residue
application and sites variability) is key to generate relevant and generalizable information
(Sadras and Lemaire, 2014) that could be used to optimize resource allocation.
6.3.2 Model assisted analysis and decision support framework
The results from long-term simulations with APSIM presented in Chapter 5, reproduced the
field trial results of chapter 4, and allowed further extrapolation. To facilitate the analysis in
an area with a vast array of soils, soil textures (soil WHC), was used to group residue
responses across soil types and agro-ecologies. The results demonstrated that there are site
Page | 184
and crop-specific responses to residues and inorganic N-fertilizer application that need to be
considered to design effective management recommendations in the N-deprived systems of
Central Mozambique. A 90% chance that a high C/N ratio residue application as mulch will
likely reduce maize yield, was simulated in considerably wet and high WHC soils – the
lowland sandy loam (SaL) and the upland sandy clay loam soils (cSaCL). For example,
simulated results showed an 80% chance to achieve 3 t/ha in residue free maize at 23 kg N/ha
in both cSaCL and SaL. This contrasted with only 40% at the cSaCL, and 10 % at the SaL
sites. In these soils, high soil moisture content leads to high microbial activity which triggers
faster residue decomposition resulting in N-immobilization, particularly in soils where not
enough N is available to satisfy N-immobilization demands and increase N-availability to
crops (Guntiñas et al., 2012; Probert et al., 1998; Tian et al., 1992). High N-immobilization
during residue decomposition in soils at 75% WHC compared to soils at 35% WHC was also
reported by Chen et al. (2012). In contrast, positive responses to residue application into
cowpea were simulated in almost 95% of the seasons in the same environments. This
validates our third proposition that in wetter nitrogen constrained environments, investing the
limited availability of high C/N ratio stubble on improving the performance of the sole
legume crop is more beneficial to the system than applying those residues on the maize crop.
The returns from this resource allocation strategy are likely to be observed in the following
cereal crop.
Results also showed that, the negative responses to high C/N ratio residue application on
maize yields, WUE and NUE will largely depend on the levels of N-application and type of
season. In both SaL and cSaCL soils, benefits from residue application were only observed
in the driest seasons, i.e., 5-10% seasons over the simulated period. Similar results were
reported in the Central Mexico Highlands where improved maize yields under residue
application were attained in dry years (Romero-Perezgrovas et al., 2014). In contrast to the
wet and high WHC soils, in drier and low WHC coarse textured soils (fSa), residue
application increased maize yields through improved water capture (WU) at all levels of N-
application. In these dry and low WHC soils, the poor microbial activity delays
decomposition (Chen et al., 2012), allowing residues to remain longer and provide further
soil cover. Benefits from residue application in drier and coarse textured soils were also
reported by Gentile et al. (2008). These are the only soils were residue applications into
maize proved to be more beneficial to the system compared to applications into legumes.
Page | 185
6.3.4 Demonstrating short term effects from alternative resource allocation strategies
Demonstrating short term effects from investments in residues and fertilizer application, in
the N-deprived smallholder farming systems is a critical challenge defined in the CA research
agenda of most research and development institutions across SSA (Giller et al., 2011a).
Rising awareness about short-term effects can help promote a more conscious and informed
adoption process. In this study higher returns from residue and limited N-fertilizer application
were obtained for the legume phase of the cowpea-maize sequence. High returns from
investment in legume-cereal sequences have been reported in several other studies before
(Luce et al., 2015; Monzon et al., 2014; Neugschwandtner and Kaul, 2015), mainly resulting
from high N-turnover and carry-over effect from the retained N-rich legume biomass. In the
continuous maize systems, penalties to yield from residue application and retention (NUpE
and NUE), were observed in the subsequent bio-assay maize crop. Penalties from yearly
residue incorporation in continuous maize systems have also been reported in the United
States corn belt (Gentry et al., 2013; Sindelar et al., 2013), though those are high residue and
N-input systems.
For Central Mozambique and SSA N-deprived smallholder farming systems managed under
conservation agriculture, where residue retention has been widely recommended as a mean to
reduce soil erosion, yield penalties from residue retention in continuous maize systems need
to be further studied in order to find effective ways to alleviate them. In the meanwhile,
promoting regular legume rotations (Morrison, 2013; Stute and Posner, 1995) appears to be a
cost-effective option to minimize such penalties among resource poor farmers who are unable
to invest in fertilizers. In these systems, effective legume phases can also help reduce the
economic optimum of N-fertilizer application (Luce et al., 2015). In our experiments
(Chapter 4), differences between the yield of fertilized continuous maize systems and the
unfertilized maize yield in the cowpea-maize sequence were only found for in specific
treatments. This means that more pragmatic and smarter resource management strategies
need to be considered for the N-deprived smallholder farming systems. However, this needs
further research. For the specific case of the N-deprived systems of Central Mozambique,
bold measures need to be considered in order to increase soil N-availability and reduce
penalties from yearly incorporation of high C/N ratio residues in continuous maize systems.
Climatic analyses for Central Mozambique indicate the existence of regions where double
growing periods are possible (Reddy, 1984). This means that an opportunistic legume crop,
either early or as a relay crop, could be grown. In the course of this research, an early cowpea
Page | 186
crop was successfully grown (Nhantumbo et al., 2014), leaving enough time to grow a
second crop starting in February. This proves that two legume crops can be grown in the
region, even if only one is for grain and the other one as a green manure cover crop.
6.4 Conclusions and way forward
From this study it can be concluded that, apart from the protracted problem of limited access
to production means, optimizing resource allocation across activities within the household is
critical to improve the performance of the whole farm enterprise and consequently their food
and income generation prospects. However, because the farm enterprise, its design and
management approach differs in space, time and across typologies, there is no single-sized
technological package capable of solving all farmers’ problems at once. Therefore, an
integrated socio-technical approach that looks at the whole household “farm enterprise”
dynamics (i.e., from the homestead to the plot level management) is required. This is because
the overall household livelihood strategy and the farmers’ ability to meet their annual food
and income generation targets, affects their resource use behaviour – investment and
relocation across activities and plot level management. Personalized typology specific
agricultural interventions are required to effectively involve farmers in the negotiation and
co-design of more farmer suitable and sustainable farming systems and management
practices.
At the filed level, the results have demonstrated that there is crop, site and management
specific responses and trade-offs that need to be understood and explored in a smarter and
more pragmatic way. This, in order to improve resource productivity and yields in N-deprives
smallholders farming systems managed under conservation agriculture in Central
Mozambique. However, because field trials can only capture a limited part of the vast and
complex GxExM interactions and their effect on resource productivity and yields, model
assisted research and decision support systems are critical to improve decision making
processes, especially in contexts where research for development is poorly funded, like
Mozambique. Nevertheless, model integration into local research and extension platforms
needs to be part of a wider intervention platform that not only looks at plot level responses,
but which can also capture and exploit the inter linkages between components of the system –
a whole household farm enterprise. It’s critical, however, to combine agronomic simulation
Page | 187
models such as APSIM with agent based models (ABM) (e.g., multiple goal programming or
the trade-off analysis model (TOA-MD Model)) to set clear optimization goals and co-design
development roadmaps to each farm enterprise and region. This approach is critical in the
sense that it would help generate information that can help farmers visualize all available
alternatives on what they can achieve for their farms and what approach better suits their
circumstances.
6.4.1 Scaling out and scaling up innovations
Central Mozambique, presents a vast mosaic of agro-ecologies with highly diverse edaphic
and climatic characteristics. For example, in the four provinces that constitute the Central
Mozambique agricultural growth corridor, at least 3 agro-ecologies can be identified in each
one of them – Manica (AEZ-R4, AEZ-R6, and AEZ-R10), Sofala (AEZ-R5, AEZ-R4, and
AEZ-R6), Tete (AEZ-R6, AEZ-R7, and AEZ-R10) and Zambezia (AEZ-R7, AEZ-R5, AEZ-
R8, and AEZ-R10). Moreover, clear intra and cross-agroecological zone differences with
regards to soil type, soil fertility and soil texture induced in crop moisture regimes are also
characteristic of the region, and dictate how farmers manage their systems within the context
of existing and perceived variability. Farmer diversity is also an issue that characterizes the
agricultural sector in the region, as our results demonstrated. All these factors make farming a
highly complex activity in the region. Because no single-sized technological package will
help solve all farmers’ problems at once, understanding how each group of farmers perceive
their circumstances and manage their farm enterprise is critical to actively involve them in the
co-generation of relevant agricultural information that can help make farming a smarter and
more effective activity, is critical. However, finding timely and cost-effective methods to co-
generate relevant and pragmatic agricultural information that can provide feedback to farmers
within a time frame that can allow them to reflect and use it to improve their farming systems
design and management still constitutes a major challenge.
In this thesis, three new approaches that need to be further explored at both national and local
level by research institutions and also local agricultural authorities, since they have the
potential to help minimize the existing information gap and its relevance burden by making
agricultural interventions more reflective of farmers’ circumstances and the complexity of
their farming systems design, management and their agro-environment are proposed:
Page | 188
1) Personalized agricultural interventions: a framework to cluster and characterize
farmers in new simple and mutually exclusive functional typologies that reflect
the diversity among smallholder farmers, based on their resource use behaviour
and openness to engage in new practices. Contrary to the current farmer
classification that assumes a wide homogeneity among farmers, the approach
proposed in this thesis is based on the proposition that farmers are different, and
their livelihood strategy and perception of their own circumstances shapes how
they manage their whole household farm enterprise.
2) Creating more generalizable information by conducting research that explores the
interactions within the farm household enterprise and finding ways to amplify or
minimize management associated trade-offs: Model assisted decision support
frameworks for residue and fertilizer management in N-deprived systems. The
frameworks will allow researchers to extend the scope of their analysis in order to
produce information that has the potential for more general application, this being
a problem to date. This framework used a simplistic approach that allows the
grouping of several soils into the three textural classes – heavy, medium and light
textured, as a key to further explore the responses and resource productivity trade-
offs associated with residues and N-fertilizer allocation strategies across farms and
agro-ecologies.
3) Bringing researchers and farmers closer: The use of model assisted participatory
crop season planning has also been tested and proved to be a valid option for
integration into local research and extension platforms. This is an interactive
method that allows a closer farmer-researcher interaction.
Taking into account the fact that the farm enterprise is managed differently across typologies,
depending on each household biophysical and socioeconomic circumstances, which
influences current farming system design and management, agricultural interventions needs
to match household’s circumstances. Therefore, the following management challenges were
identified as key to sustainably improve current farming systems performance and boost each
farm typology and the household ability to meet their annual food and income generation
targets:
Page | 189
1) For resource endowed Type-B and Type-A2 farms who are already maize self-
sufficient and have the necessary investment capacity to purchase the basic inputs
required, the challenge is to improve resource productivity (land, labour and
inputs) and yields in order to generate enough surplus and on farm income. This
can be achieved through smarter resource mechanization of farm activities and
relocation strategies.
2) For resource constrained Type-A1 farmers, the primary challenge is to generate
enough income either on-farm or off-farm to supplement their low maize self-
sufficiency. For these farms, improving legume productivity appears to be the best
option considered legumes high market price and further contribution to soil N-
availability. Diversification of farm activities is also a better option for this group
since it will reduce the risk.
3) Maximizing land use and productivity through better crop arrangement and
improved plot level management is critical to improve returns from land and
labour. Here, mechanizing farm activities such as sowing and weeding is
fundamental for all groups especially considering the fact that labour availability
and quality are limiting and each farm will need to smartly and efficiently
cultivate all owned land to produce enough surplus to generate on farm income to
cover household daily expenses. This is critical in Central Mozambique, were land
is not yet a limiting factor, but surplus will hardly be achieved in 0.5 ha plots of
land in a region with less than 50 days of rainfall a season.
At field level improving resource allocation and productivity is a key area to be addressed
across all farm types. Here understanding field level responses and resource allocation trade-
offs is critical to take advantage of existing cross-site variability and potential short term
benefits. Therefore, the following rules of thumb are proposed in order to guide resource
allocation – residues and inorganic N fertilizers, across farms, in the region:
1) Although residues from previous maize crops and weeds germinated during
unmanaged fallows, are incorporated across all farm types, residues application
into maize is only viable if enough fertilizer (≥ 92 kg N/ha), can be mobilized to
Page | 190
satisfy N-immobilization demands which is not the case of most resource poor
farmers. Therefore, shifting residue application into legumes is the best short term
option.
2) For the ones using inorganic N-fertilizers, investing the limited amounts of N-
fertilizers into improving the performance of the legume crop (biomass and yield)
or increasing the yield of residue free maize is the best N-fertilizer allocation
strategy compared to applying this N into residue applied maize;
3) Cattle owners and those who can afford to purchase enough manure should also
consider incorporating manure into maize fields rather than just using it in
vegetable production fields were it is likely to help improve soil biophysical
proprieties.
Because central Mozambique presents a vast mosaic of soils and farming environment, it is
critical that resource allocation strategies match existing cross-site variability. To fine tune
residue and fertilizer allocation and field level, a set of rules based on existing soil WHC
gradients in the region are proposed:
1) In high water holding capacity (WHC) soils where moisture is not a limiting
factor, high C/N ratio residues should be used to improve the performance of the
sole legume crop during the legume phase of the rotation rather than applied into
an unfertilized cereal crop where the soil moisture advantage provided by residues
(WUE increments) do not compensate for the yield losses due to N-
immobilization;
2) High C/N ratio residues only offer a significant yield advantage to the cereal crops
in dryer and low WHC soils. In these soils, residues significantly improve incrop
rainfall capture generating a moisture benefits that surpasses the negative impacts
from N-immobilization on maize yield;
3) Best responses from inorganic N-fertilizer (NUE, NUpE, NUtE and NHI) are
likely to be attained in wet and high WHC environments where there is enough
moisture for the crop to efficiently use the supplied inorganic N fertilizer. This
because in dry and low WHC soils, poor soil moisture regimes reduce crop
nitrogen uptake leading to poor responses to inorganic N-application;
4) For wet and high WHC soils, the beneficial effects of applying crop residues on
the legume crop are likely to be observed on the following cereal crop.
Page | 191
References
Ali, D., Daniel Ayalew, A., and Klaus, D. (2015). Is There a Farm Size-Productivity
Relationship in African Agriculture? Evidence from Rwanda. Land economics 91,
479-500
Bakht, J., Shafi, M., Jan, M. T., and Shah, Z. (2009). Influence of crop residue management,
cropping system and N fertilizer on soil N and C dynamics and sustainable wheat
(Triticum aestivum L.) production. Soil and Tillage Research 104, 233-240.
Baudron, F., Delmotte, S., Corbeels, M., Herrera, J. M., and Tittonell, P. (2015a). Multi-scale
trade-off analysis of cereal residue use for livestock feeding vs. soil mulching in the
Mid-Zambezi Valley, Zimbabwe. Agricultural Systems 134, 97-106.
Baudron, F., Thierfelder, C., Nyagumbo, I., and Gérard, B. (2015b). Where to Target
Conservation Agriculture for African Smallholders? How to Overcome Challenges
Associated with its Implementation? Experience from Eastern and Southern Africa.
Environments 2, 338-357.
Bucagu, C., Vanlauwe, B., Van Wijk, M., and Giller, K. (2014). Resource use and food self-
sufficiency at farm scale within two agro-ecological zones of Rwanda. Food Security
6, 609-628.
Chen, Q.-H., Feng, Y., Zhang, Y.-P., Zhang, Q.-C., Shamsi, I. H., Zhang, Y.-S., and Lin, X.-
Y. (2012). Short-Term Responses of Nitrogen Mineralization and Microbial
Community to Moisture Regimes in Greenhouse Vegetable Soils. Pedosphere 22,
263-272.
Chianu, J., Chianu, J., and Mairura, F. (2012). Mineral fertilizers in the farming systems of
sub-Saharan Africa. A review. Agronomy for Sustainable Development 32, 545-566.
Cossani, C. M., Slafer, G. A., and Savin, R. (2010). Co-limitation of nitrogen and water, and
yield and resource-use efficiencies of wheat and barley. Crop and Pasture Science 61,
844-851.
Cossani, C. M., Slafer, G. A., and Savin, R. (2012). Nitrogen and water use efficiencies of
wheat and barley under a Mediterranean environment in Catalonia. Field Crops
Research 128, 109-118.
Page | 192
Cunguara, B., Langyintuo, A., and Darnhofer, I. (2011). The role of nonfarm income in
coping with the effects of drought in southern Mozambique. Agricultural Economics
42, 701-713.
Daloğlu, I., Nassauer, J. I., Riolo, R. L., and Scavia, D. (2014). Development of a farmer
typology of agricultural conservation behavior in the American Corn Belt.
Agricultural Systems 129, 93-102.
Dimes, J. (2011). Formulating Crop Management Options for Africa’s Drought-Prone
Regions: Taking Account of Rainfall Risk Using Modeling. In “Innovations as Key to
the Green Revolution in Africa.” (A. Bationo, B. Waswa, J. M. M. Okeyo, F. Maina
and J. M. Kihara, eds.), pp. 785-794. Springer Science and Business Media,
Netherlands.
Dorward, A., Anderson, S., Bernal, Y. N., Vera, E. S., Rushton, J., Pattison, J., and Paz, R.
(2009). Hanging in, stepping up and stepping out: livelihood aspirations and strategies
of the poor. Development in Practice 19, 240-247.
Esther, O. J., Guo, C.-h., Tian, X.-h., Li, H.-y., and Zhou, Y.-x. (2014). The Effects of Three
Mineral Nitrogen Sources and Zinc on Maize and Wheat Straw Decomposition and
Soil Organic Carbon. Journal of Integrative Agriculture 13, 2768-2777.
Gentile, R., Vanlauwe, B., Chivenge, P., and Six, J. (2008). Interactive effects from
combining fertilizer and organic residue inputs on nitrogen transformations. Soil
Biology and Biochemistry 40, 2375-2384.
Gentry, L. F., Ruffo, M. L., and Below, F. E. (2013). Identifying Factors Controlling the
Continuous Corn Yield Penalty. Agronomy Journal 105, 295-303.
Gilbert, N. (2012). Dirt Poor: The key to tackle hunger in Africa is enriching its Soils. Nature
483, 525-227.
Giller, K. E., Corbeels, M., Nyamangara, J., Triomphe, B., Affholder, F., Scopel, E., and
Tittonell, P. (2011). A research agenda to explore the role of conservation agriculture
in African smallholder farming systems. Field Crops Research 124, 468-472.
Giller, K. E., Rowe, E. C., de Ridder, N., and van Keulen, H. (2006). Resource use dynamics
and interactions in the tropics: Scaling up in space and time. Agricultural Systems 88,
8-27.
Page | 193
Guillem, E. E., Barnes, A. P., Rounsevell, M. D. A., and Renwick, A. (2012). Refining
perception-based farmer typologies with the analysis of past census data. Journal of
Environmental Management 110, 226-235.
Guntiñas, M. E., Leirós, M. C., Trasar-Cepeda, C., and Gil-Sotres, F. (2012). Effects of
moisture and temperature on net soil nitrogen mineralization: A laboratory study.
European Journal of Soil Biology 48, 73-80.
Heffer, P., and Prud'homme, M. (2011). "Fertilizer Outlook 2011-2015." IFA, Paris, France.
Herrmann, L., Chotte, J. L., Thuita, M., and Lesueur, D. (2014). Effects of cropping systems,
maize residues application and N fertilization on promiscuous soybean yields and
diversity of native rhizobia in Central Kenya. Pedobiologia 57, 75-85.
INE (2011). Censo Agropecuário 2009-2010: Resultados Preliminares. pp. 105. Instituto
Nacional de Estatistica(INE), Ministerios da Agricultura (MINAG), Mapuo,
Mozambique.
Kraaijenbrink, J., and Wijnhoven, F. (2008). Managing heterogeneous knowledge: a theory
of external knowledge integration. Knowledge Management Research and Practice 6,
274-286.
Krishna, N. R. (2001). Effects of Cereal and Legume Cover Crop Residues on Weeds, Yield,
and Net Return in Soybean (Glycine max). Weed Technology 15, 660-668.
Leonardo, W., van de Ven, G. J., Udo, H., Kanellopoulos, A., Sitoe, A., and Giller, K.
(2015). Labour not land constrains agricultural production and food self-sufficiency in
maize-based smallholder farming systems in Mozambique. Food Security, 1-18.
Meijer, S. S., Catacutan, D., Ajayi, O. C., Sileshi, G. W., and Nieuwenhuis, M. (2015). The
role of knowledge, attitudes and perceptions in the uptake of agricultural and
agroforestry innovations among smallholder farmers in sub-Saharan Africa.
International Journal of Agricultural Sustainability 13, 40-54.
Monzon, J. P., Mercau, J. L., Andrade, J. F., Caviglia, O. P., Cerrudo, A. G., Cirilo, A. G.,
Vega, C. R. C., Andrade, F. H., and Calviño, P. A. (2014). Maize-soybean
intensification alternatives for the Pampas. Field Crops Research 162, 48-59.
Morrison, L. (2013). Residue management can reduce continuous corn yield penalty. Corn
and Soybean Digest. Vol. 75 Issue 9, p 40
Page | 194
Muzari, W., Gatsi, W., and Muvhunzi, S. (2012). The Impacts of Technology Adoption on
Smallholder Agricultural Productivity in Sub-Saharan Africa: A Review. Journal of
Sustainable Development 5, 69-77.
Neugschwandtner, R. W., and Kaul, H.-P. (2015). Nitrogen uptake, use and utilization
efficiency by oat–pea intercrops. Field Crops Research 179, 113-119.
Ngwira, A. R., Thierfelder, C., and Lambert, D. M. (2013). Conservation agriculture systems
for Malawian smallholder farmers: long-term effects on crop productivity,
profitability and soil quality. Renewable Agriculture and Food Systems 28, 350-363.
Nhantumbo, N., Dias, J., Mortlock, M., Nyagumbo, I., Dimes, J., and Rodriguez, D. (2014).
Best-fit residue allocation: A gate for legume intensification in nitrogen constrained
cropping systems of Central Mozambique. In "First African Congress on
Conservation Agriculture" (A. C. T. Network, ed.), pp. 33-35. ACT, 18-21 March
Lusaka, Zambia.
Probert, M. E., Dimes, J. P., Keating, B. A., Dalal, R. C., and Strong, W. M. (1998). APSIM's
water and nitrogen modules and simulation of the dynamics of water and nitrogen in
fallow systems. Agricultural Systems 56, 1-28.
Recous, S., Robin, D., Darwis, D., and Mary, B. (1995). Soil inorganic N availability: Effect
on maize residue decomposition. Soil Biology and Biochemistry 27, 1529-1538.
Reddy, S. J. (1984). Clima Geral de Mocambique in "Serie Terra e Água" (INIA, ed.), pp. 41.
Instituto de Investigação Agronómica de Moçambique (INIA), Departamento de Terra
e Agua (DTA), Maputo, Mozambique.
Ribeiro, C., Madeira, M., and Araújo, M. C. (2002). Decomposition and nutrient release from
leaf litter of Eucalyptus globulus grown under different water and nutrient regimes.
Forest Ecology and Management 171, 31-41.
Romero-Perezgrovas, R., Verhulst, N., De La Rosa, D., Hernandez, V., Maertens, M.,
Deckers, J., and Govaerts, B. (2014). Effects of Tillage and Crop Residue
Management on Maize Yields and Net Returns in the Central Mexican Highlands
Under Drought Conditions. Pedosphere 24, 476-486.
Rusinamhodzi, L., Dahlin, S., and Corbeels, M. (2016). Living within their means:
Reallocation of farm resources can help smallholder farmers improve crop yields and
soil fertility. Agriculture, Ecosystems and Environment 216, 125-136.
Page | 195
Rware, H., Wairegi, L., Oduor, G., Macharia, M., Romney, D., Tarfa, B. D., de Maria, R.,
Ley, G., Tetteh, F., Makumba, W., Dicko, M., Nouri, M., Leon, N. N., Gondwe, B.,
Kaizzi, K., Negash, D., Kibunja, C., Alawy, A., Wartmann, C., and Rutsimba, E.
(2014). Assessing the Potential to Change Stakeholders Knowledge and Practices on
Fertilizer Recommendations in Africa. Agricultural Sciences 5, 1384-1391.
Sadras, V. O., and Lemaire, G. (2014). Quantifying crop nitrogen status for comparisons of
agronomic practices and genotypes. Field Crops Research 164, 54-64.
Sadras, V. O., and Rodriguez, D. (2010). Modelling the nitrogen-driven trade-off between
nitrogen utilisation efficiency and water use efficiency of wheat in eastern Australia.
Field Crops Research 118, 297-305.
Shah, Z., Shah, S. H., Peoples, M. B., Schwenke, G. D., and Herridge, D. F. (2003). Crop
residue and fertiliser N effects on nitrogen fixation and yields of legume-cereal
rotations and soil organic fertility. Field Crops Research 83, 1-11.
Shangguan, Z. P., Shao, M. A., and Dyckmans, J. (2000). Nitrogen nutrition and water stress
effects on leaf photosynthetic gas exchange and water use efficiency in winter wheat.
Environmental and Experimental Botany 44, 141-149.
Sichoongwe, K., Mapemba, L., Tembo, G., and Ng'ong'ola, D. (2014). The Determinants and
Extent of Crop Diversification Among Smallholder Farmers: A Case Study of
Southern Province Zambia. Journal of Agricultural Science 6, 150-159.
Sindelar, A. J., Coulter, J. A., Lamb, J. A., and Vetsch, J. A. (2013). Agronomic Responses of
Continuous Corn to Stover, Tillage, and Nitrogen Management. Agronomy Journal
105, 1498-1506.
St. Luce, M., Grant, C. A., Zebarth, B. J., Ziadi, N., O’Donovan, J. T., Blackshaw, R. E.,
Harker, K. N., Johnson, E. N., Gan, Y., Lafond, G. P., May, W. E., Khakbazan, M.,
and Smith, E. G. (2015). Legumes can reduce economic optimum nitrogen rates and
increase yields in a wheat–canola cropping sequence in western Canada. Field Crops
Research 179, 12-25.
Stute, J. K., and Posner, J. L. (1995). Legume Cover Crops as a Nitrogen Source for Corn in
an Oat-Corn Rotation. Journal of Production Agriculture 8, 385-390.
Page | 196
Sun, Y., Ma, J., Sun, Y., Xu, H., Yang, Z., Liu, S., Jia, X., and Zheng, H. (2012). The effects
of different water and nitrogen managements on yield and nitrogen use efficiency in
hybrid rice of China. Field Crops Research 127, 85-98.
Tao, Z., Li, C., Li, J., Ding, Z., Xu, J., Sun, X., Zhou, P., and Zhao, M. (2015). Tillage and
straw mulching impacts on grain yield and water use efficiency of spring maize in
Northern Huang–Huai–Hai Valley. The Crop Journal 3, 445-450.
Teixeira, E. I., George, M., Herreman, T., Brown, H., Fletcher, A., Chakwizira, E., de Ruiter,
J., Maley, S., and Noble, A. (2014). The impact of water and nitrogen limitation on
maize biomass and resource-use efficiencies for radiation, water and nitrogen. Field
Crops Research 168, 109-118.
Thierfelder, C., Cheesman, S., and Rusinamhodzi, L. (2012). A comparative analysis of
conservation agriculture systems: Benefits and challenges of rotations and
intercropping in Zimbabwe. Field Crops Research 137, 237-250.
Thierfelder, C., Matemba-Mutasa, R., and Rusinamhodzi, L. (2015). Yield response of maize
(Zea mays L.) to conservation agriculture cropping system in Southern Africa. Soil
and Tillage Research 146, Part B, 230-242.
Tian, G., Kang, B. T., and Brussaard, L. (1992). Biological effects of plant residues with
contrasting chemical compositions under humid tropical conditions—Decomposition
and nutrient release. Soil Biology and Biochemistry 24, 1051-1060.
Tittonell, P., Muriuki, A., Shepherd, K. D., Mugendi, D., Kaizzi, K. C., Okeyo, J., Verchot,
L., Coe, R., and Vanlauwe, B. (2010). The diversity of rural livelihoods and their
influence on soil fertility in agricultural systems of East Africa – A typology of
smallholder farms. Agricultural Systems 103, 83-97.
Tittonell, P., Vanlauwe, B., de Ridder, N., and Giller, K. E. (2007). Heterogeneity of crop
productivity and resource use efficiency within smallholder Kenyan farms: Soil
fertility gradients or management intensity gradients? Agricultural Systems 94, 376-
390.
Vanlauwe, B., and Giller, K. E. (2006). Popular myths around soil fertility management in
sub-Saharan Africa. Agriculture, Ecosystems and Environment 116, 34-46.
Page | 197
Vanlauwe, B., Tittonell, P., and Mukalama, J. (2006). Within-farm soil fertility gradients
affect response of maize to fertiliser application in western Kenya. Nutrient Cycling in
Agroecosystems 76, 171-182.
Verhulst, N., Govaerts, B., Nelissen, V., Sayre, K. D., Crossa, J., Raes, D., and Deckers, J.
(2011). The effect of tillage, crop rotation and residue management on maize and
wheat growth and development evaluated with an optical sensor. Field Crops
Research 120, 58-67.
Wall, P. C. (2007). Tailoring Conservation Agriculture to the Needs of Small Farmers in
Developing Countries. Journal of Crop Improvement 19, 137-155.
Xiukang, W., Zhanbin, L., and Yingying, X. (2015). Effects of mulching and nitrogen on soil
temperature, water content, Nitrate-N content and maize yield in the Loess Plateau of
China. Agricultural Water Management 161, 53-64.
Zingore, S., Murwira, H. K., Delve, R. J., and Giller, K. E. (2007). Soil type, management
history and current resource allocation: Three dimensions regulating variability in
crop productivity on African smallholder farms. Field Crops Research 101, 296-305.
Page | 198
Appendix 1.1 Mozambique agroecological regions with their main farming systems
Altitude Description Agro-
ecological
zone (AEZ)
Farming Systems
Low altitude
Semi-arid and arid
regions of Southern
Mozambique
In lad semi-arid zone in Maputo province R1 Dry land maize and pastoralism
Semi-arid littoral in southern Mozambique,
Inhambane and Maputo Province
R2 Dry land maize-legume based systems. Coconut
based systems in the cost intercropped with
groundnuts and cassava
Arid and semi-arid parts of in land Gaza, north
Maputo and Inhambane province
R3 Dry land maize, lowland irrigated rice and maize
systems (Chókwè) with Commercial vegetables;
Cotton based systems; crop-livestock systems
Medium altitude
Covers almost 80% of Manica province and part of
Sofala in central Mozambique
R4 Maize-legume with Banana based systems in the
highlands of Manica
Coastal area of central Mozambique covering
Sofala and Zambezia province
R5 Maize-legume, rice based systems and
commercial sugar cane
Semi-arid region of Southern Tete and northern
Manica and Zambezia
R6 Livestock dominated systems with sorghum and
dryland maize
The largest agroecological region, covering 5
provinces in Central and Northern Mozambique,
namely Zambezia, Nampula, Tete, Niassa and
Cabo Delgado
R7 Maize-legumes and cassava-legume systems;
Cotton and tobacco based systems. Groundnuts,
beans, pegeonpea are important legume crops
High altitude
Northern Mozambique littoral covers Zambezia,
Nampula and Cabo Delgado
R8 Coconuts based systems
Maize/cassava-legume systems
Highlands of Cabo Delgado – planalto de Mueda R9 -
Manica, Tete, Zambezia and Niassa highlands R10 Maize-legume systems
Small patches of wheat in Rotanda
Commercial vegetable gardens
Compiled from: (Kajisa and Payongayong, 2011; Lukanu et al., 2009; Reddy, 1984)
Page | 199
Appendix 3.1 Generalized crop management timelines across regions and farmer groups
Activity Timing
Land preparation Land preparation mostly takes place from September-December before maize sowing.
In Rotanda, ploughing right after harvest in March/April and August/September is a
common practice among cattle owners, to incorporate residues and give them time to
decompose.
Maize sowing
window
Normally split across 15 October to 15 December: Early sowing in mid-October to
late November after the first rainfall, depending on the region. This window is mostly
used by resourceful farmers who prepare their land on time.
Normal sowing in early December using early to late maturity varieties. In irrigated
areas: mid-September and irrigated until the start of the rainy season.
Legumes sowing
window
Small patches of cowpea at the same time as maize for leaf and fresh pods.
The main window is in January/February after the rains, extending to March in the
wetlands. Harvest is normally in April/May. Sole legumes–common beans are sown
in March in irrigated areas and in the wetlands.
Varieties used Maize: local varieties, Matuba (OPV) and local hybrids PAN 67 and PAN 53 from
Pannar Seeds; Beans: Diacol Calima (IIAM).
Cowpea: IT16 and IT18 (IIAM) but mostly local varieties with indeterminate growth.
Spacing Random sowing with 3-6 pls/station for farmers without animal draft power. A better
in-line sowing pattern is seen with farmers with animals. Plant spacing recommended
by extension: 90x50 cm for Matuba, 90x40 cm for PAN67 and 90x60 cm for
Rasposta, with 2 pls/station, as recommended in Farmer Field Schools.
Fertilizer
application
(when available)5
Basal: at sowing with 3.4g/hill of NPK 12:24:12
Top dressing: 30-40 DAS with 3.4 g/hill Urea (46% N). This application coincides
with the second weeding.
Weeding 1st weeding at 21-30 DAS and 2nd weeding in January around tasselling. The third
weeding takes place in early February after the first rains and is also intended to
prepare land for beans that will planted as a relay crop.
5 8 farmers of the 20 that attended the PCM workshop in Rotanda, reported using fertilizer in the last season. The
amounts reported were, 75 kg NPK 12:24:12 (1.5-50 kg bags in a 1ha), 100 kg Urea (2-50 kg bag in 1ha), 100 kg of
Urea (2-50kg bags in 0.5ha), 100 kg of urea (2-50 kg bags in 0.5ha), 100 kg of urea (2-50 kg bags in 1ha), 100 kg of
urea (2-50 kg bags in 0.5ha), 50 kg of urea (1-50 kg bags in 0.5ha) and 100 kg of urea (2-50 kg bags in 1ha)
Page | 200
Harvest Harvesting from late March to early May for maize. Beans are harvested around May-
June.
Crop residues Left in the field in Macate. In Rotanda, the residues are normally incorporated after
harvest for those with access to draft animal power, and sometimes burned during
land preparation by those without access to draft animals. Residues are not sold and
animal graze freely after harvest.
Manure All present animal owners reportedly use manure but most of it is applied into
commercial vegetables production.
Appendix 3.2 Crop distribution across farming environments and soils in both AEZ-R4
and AEZ-R10.
Soil Type (local name) Fertility Yield6 Most common crop allocation
Red sandy clay loam
(Kupswuka)
Average Average Maize, sunflower, sorghum, legumes, cotton and
sesame
Sandy soil (Djetcha) Poor Low Maize-sorghum-sunflower and legumes
Dark sandy loam
(N’dongo & Matoro)
High Higher Vegetables, maize, legumes, tubers and root crops
6 Yield capacity is defined on a maize base: In an average season, without applying any fertilizer or
manure, farmers in Rotanda (R10) reportedly harvest between 27 bags of 50kg/bag maize in the lowland
wet environments and only 16 bags in upland environments using recycled local seed. During the focus
group discussion, a farmer reported having harvested 32 bags of 50 kg/bag in the lowland environment in
the 2012-13 season after planting 0.5 ha with PAN67, a high yielding hybrid with a yield potential of 6-8
t/ha
Page | 201
Appendix 3.4: Farm household clusters by their fertilizer and manure allocation
preferences
Figure A-0-1 Farm household clustered by fertilizer and manure allocation across
crops. Codding: X-Y-Z(v): number X is gender of the household head (1-M 2-F), Y-Z
household size and farm active people; (v) and (M), where (v) is Matsinho in Vanduzi
district and (M) is Marera in Macate district. Resource use: fert+manure: household uses
fertilisers and also apply manure; Nofert_Usemanure: household doesn’t use fertilizer but
applies manure; Allocation preference presented according to the resource in use: VEG -
resource applied into vegetable, Mz - applied into maize, Leg – applied into a legume,
shared – resource shared across all crops.
Page | 202
Appendix 3.5: Households clustered by their reasoning for not using fertilizers or manure
Codding: Household characteristics and location indicated through X-Y-Z(v): number X is gender of
the household head (1-Male, 2-Female); Y-Z are household size and farm active people respectively;
(v) and (M) stand for sampled communities, where (v) is Matsinho in Vanduzi district and (M) is
Marera in Macate district. Resource use: fert+manure: household uses fertilisers and also apply
manure; Nofert_Usemanure: doesn’t use fertilizer but applies manure; Nofert_NoManure: doesn’t use
both. Reasons for not using fertilizer and manure: f(X-Y) reasons for not using fertilizers. 1- No
interest, 2 - See no benefit, 3 - high cost, 4 – No local supplier, 5 – don’t know how to use them
properly; m(X-Y) reasons for not using manure: 1- No interest, 2 – Don’t see a benefit, 3 - high cost,
4 – No local supplier, 5 – don’t know how to use them properly, 6 – other
Page | 203
Appendix 3.6: Fertilizer and manure allocation preferences among residue users
Codding: Household characteristics and location indicated through X-Y-Z(v): number X is gender of
the household head (1-Male, 2-Female), Y-Z are household size and farm active people, respectively;
(v) and (M) stand for sampled communities where (v) is Matsinho in Vanduzi district and (M) is
Marera in Macate district. Resource use: fert+manure: household uses fertilisers and also applies
manure; Nofert_Usemanure: household doesn’t use fertilizer but applies manure; Allocation
preference presented according to the resource in use: VEG - resource applied into vegetable, Mz -
applied into maize, Leg – applied into a legume, shared – resource shared across crops in the farm
enterprise
Page | 204
Appendix 3.7: Correlation between household size and number of household farm active
people disaggregated by age and sex
Legend: 1-household size, 2-household farm active people (age>16), 3-household farm active
older people (age>45), 4-household farm active youth (age: 16-35), 5-household farm active
female members, 6-household farm active male members.