190
Global goals, local actions: A framework for integrating indigenous knowledge and ecological methods for rangeland assessment and monitoring in northern Kenya Hassan G. Roba Doctoral Thesis Department of International Environment and Development Studies, Noragric Norwegian University of Life Sciences (UMB) Ås, June 2008 Thesis no.: 2008:25 ISBN: 978-82-575-0828-9 ISSN: 1503-1667

Methodological Approaches for Integrating Indigenous ... · Global goals, local actions: A framework for integrating indigenous knowledge and ecological methods for rangeland assessment

Embed Size (px)

Citation preview

  • Global goals, local actions: A framework for integrating indigenous knowledge and ecological methods for rangeland assessment and monitoring in northern Kenya

    Hassan G. Roba Doctoral Thesis

    Department of International Environment and Development Studies, Noragric Norwegian University of Life Sciences (UMB)

    s, June 2008

    Thesis no.: 2008:25 ISBN: 978-82-575-0828-9 ISSN: 1503-1667

  • To Halima and Dirram

    ii

  • Table of Contents

    List of Acronyms iv

    List of Figures v

    List of Papers vi

    Acknowledgement vii

    Abstract viii

    1. Introduction 1 1.1 Background to the study 2 1.2. Organization of the thesis 6 1.3. Study sites 7

    2. Integration of global goals and local actions 11 2.1. Definition of terms and concepts 14

    3. Theoretical perspectives 19

    4. Methodological perspectives 23

    5. Framework for integrating local knowledge in the implementation of GECs 27 5.1 Indigenous knowledge 28 5.2 Ecological methods 30 5.3 The selection of indicators 32

    6. The implementation mechanisms of the framework 34 6.1. Step I: Interviews 34 6.2. Step II: Joint field assessments 35 6.3. Step III: The workshop 37

    7. Results and discussions 37 7.1. Impact of pastoral sedentarization on vegetation 37 7.2. Herder landscape classification, vegetation assessment and monitoring 40 7.3. Herders perceptions of land degradation 41 7.4 Long term environmental dynamics 43

    8. Implications for the global goals 45

    9. Conclusions and recommendations 46

    References 48

    Part B 60

    iii

  • List of Acronyms CBD Convention on Biological Diversity

    CBO Community Based Organizations

    CCD Convention on Combating Desertification

    EMC Environmental Management Committees

    GEC Global Environmental Conventions

    GLM General Linear Model

    IEK Indigenous Ecological Knowledge

    IPAL Integrated Project on Arid Lands

    ITK Indigenous Technical Knowledge

    LGP Landscape Grazing Potential

    LGS Landscape Grazing Suitability

    NAPs National Action Program

    NEMA National Environmental Management Authority

    NGOs Non Governmental Organizations

    PCA Principal Components Analysis

    PRA Participatory Rural Appraisal

    RRA Rapid Rural Appraisal

    SAS Statistical Analysis System

    SEEM Socio-Economic and Ecological Model

    SPSS Statistical Package for the Social Sciences

    TEK Traditional Ecological Knowledge

    UNCCD United Nations Convention to Combat Desertification

    UNCED United Nations Conference on Environment and Development

    UNCOD United Nations Conference on Desertification

    UNEP Unite Nations Environmental Programme

    iv

  • List of Figures

    Figure 1. Location of study sites on a map of the Marsabit District 8

    Figure 2a. Annual rainfall for Marsabit town 9

    Figure 2b. Mean monthly rainfall for Marsabit town between 1935 and 2004 9

    Figure 2c. Total annual rainfall for the Kargi station 10

    Figure 3. Framework for the implementation of Global Environmental Conventions, showing integration of

    indigenous knowledge and ecological methods for the assessment and monitoring of environmental change 28

    Figure 4. Framework for integrating herder knowledge and ecological methods for assessing and monitoring

    rangelands in northern Kenya 35

    v

  • List of Papers Paper I. Roba, H.G. and Oba, G. 2008. Integration of herder knowledge and ecological methods for land

    degradation assessment around sedentary settlements in a sub-humid zone in northern Kenya.

    International Journal of Sustainable Development & World Ecology 15 (2008) 251264 DOI

    10.3843/SusDev.15.3:8

    Paper II Roba, H. G and Oba, G. 2008. Community participatory landscape classification and biodiversity

    assessment and monitoring of grazing lands in northern Kenya.

    Journal of Environmental Management (In press). doi:10.1016/j.jenvman.2007.12.017

    Paper III Roba, H. G. and Oba, G. Efficacy of integrating herder knowledge and ecological methods for

    monitoring rangeland degradation in Northern Kenya (Submitted to Human Ecology)

    Paper IV

    Roba, H. G. and Oba, G. Pulling away the Last Tree or Reversing Desertification? Re-assessments

    of land cover changes around pastoral settlements in Northern Kenya (Revised, Global

    Environmental Change)

    vi

  • Acknowledgement The successful completion of this thesis is as a result of enormous support and encouragement I

    received from many people. I sincerely thank my supervisor Professor Gufu Oba for tirelessly

    supervising my work from the initial stage of developing research proposal, to the fieldwork and

    finally, the write up of this thesis. I benefited a lot from his guidance and scholarly skills for

    conducting research. During the fieldwork, I received valuable support from experienced field

    technicians of Kenya Agricultural Research Institutes (KARI) in Marsabit, namely, Hussein

    Wallaga, Peter Geykuku, and Diiba Guyo. I also thank Mr. Jacob Kimani of the Department of

    Resource Survey and Remote Sensing (DRSRS) of Kenya for his support in the interpretation of

    satellite imagery.

    I am grateful to the staff and colleagues at the Noragric Department for providing a pleasant and

    academically stimulating environment. Many thanks to the Librarians Liv Ellingsen and Ingeborg

    Brandtzg for being very helpful in availing relevant literature whenever I needed them throughout

    my study period. I acknowledge administrative and IT assistance I got from Josie Teurling, Frode

    Sundnes, Aslaug Gotehus, Lars imoen, Ann Marte, Anders Dysvik and Joanna Boddens-Hosang.

    I thank my colleagues Boku Tache, Zeinabu Khalif and Mohamed Guyo who accorded me a highly

    valued company and useful discussions related to my work. I also thank my other colleagues,

    Hussein Tadicha, Wario Tadicha, Bule Hallo, Safo Roba, Tari Doti, Philip Ebei, Steve Lesoron,

    Hussein Jemma, Charllote Nakakaawa, Ayele Tessema, Bed Mani Dahal, David Mwesigye

    Tumusiime for their encouragements. I thank Ali Ibrahim and Adan M. Boru for taking care of my

    family in my absence.

    I am grateful the Norwegian Research Council who funded the study through the project no.

    16139/S30 and to the National Museums of Kenya and in particular the Director Dr. Idle Farah for

    giving me paid leave to pursue Phd program.

    My special gratitude is due to the Ariaal and Rendille herders of Marsabit who accepted me as their

    friend and sacrificed their time to participate in the joint field work and for their willingness to share

    with me their knowledge on the environment.

    Finally my heartfelt gratitude goes to my wife Halima Ibrahim, for enduring the loneliness and also

    for taking care of our daughter Dirram while I was away. Her love and support provided a source of

    my inspiration.

    vii

  • Abstract

    This thesis is about testing a methodological framework for integrating indigenous knowledge

    and ecological methods for promoting local communities participation in the implementation of

    Global Environmental Conventions (GECs) such as the Convention on Combating Desertification

    (CCD) and the Convention on Biological Diversity (CBD) at local community levels. The thesis

    (divided into Part A and Part B) tackles the integration of indigenous knowledge and ecological

    methods for assessing and monitoring human impacts on the environments of northern Kenya around

    sedentarized settlements that was associated with land degradation and desertification as well as the

    loss of biodiversity. Development of the framework is based on in-depth analysis of theoretical and

    methodological analysis of environmental indicator selection. The implementation of the framework

    is approached at three levels: the global level related to the articles of the GECs, the national level

    concerned with the prioritizing of the action programs and finally, the local community levels where

    the actions of implementations of the conventions take place. In implementing the framework, the

    thesis focused on the third part which involved local communities in participatory research.

    Implementation at the local level involves the consideration of a diversity of ecological, production

    and social-cultural factors, and the use of local knowledge for resource assessment and monitoring.

    The implementation of the framework was conducted in Marsabit District, Northern Kenya, in two

    contrasting environments where two pastoral communities were used as partners. In the sub-humid

    (Ariaal) zone, the study sites were represented by the Karare and Lkijiji settlements. This zone

    included a National Forest Reserve and Game Reserve that is protected from livestock grazing,

    which served as a benchmark for monitoring vegetation change. The arid lowlands (Rendille), were

    represented by the Kargi and Korr settlements. In order to implement the framework the research

    was aimed at (a) understanding herder knowledge and ecological methods for assessing and

    monitoring the impact of pastoral sedentarization on land degradation; (b) applying herders

    knowledge of landscape classification by asking participatory questions for monitoring changes in

    biodiversity; (c) using herder and ecological indicators for assessing and monitoring land

    degradation; and (d) understanding long-term changes in vegetation cover using herder monitoring

    and ecological monitoring methods.

    For addressing the goals in implementing the framework, herders knowledge and ecological

    methods were integrated in three sequential steps. Semi-structured interviews and group discussions

    viii

  • with key informants were used to generate information on livestock management, changes in

    vegetation indicators and historical changes in land use patterns. Joint transect walks were conducted

    with knowledgeable herders to assess environmental change using ecological indicators (vegetation

    and soil) and herder anthropogenic indicators (i.e. landscape grazing potential and landscape grazing

    suitability). Monitoring of marked transects, satellite images taken at different times, and herders

    knowledge were used to evaluate long-term changes in vegetation cover around permanent

    settlements in the arid lowlands that were previously mapped as desertified sites. Finally, a workshop

    was organized with herders and Environmental Management Committees (EMCs) in which they

    participated in informal discussions on issues addressed by the joint research project.

    The analyses showed that the rangelands both the sub-humid (Ariaal) and the arid lowlands

    (Rendille) had shrunk during the previous three to four decades. Mobility had however remained as

    the main strategies for land use for grazing through splitting of the herds between the mobile (fora)

    managed in the remote rangelands, and home-based rangelands. Although it had been earlier

    hypothesized that pastoral sedentarization contributed to degradation of vegetation around

    settlements, the present research found no evidence of permanent degradation. In the sub-humid zone

    an increase in bush cover, a decline in herbaceous species, or an increase in unpalatable as opposed

    to palatable plant species was found according to herder assessments. For making comparisons, the

    herders separated the biodiversity into those species that are desirable for livestock grazing and those

    that are undesirable. For the herders, the qualitative changes in vegetation indicators were associated

    with continuous grazing, banning the use of fire for range management, and episodic rainfall.

    Ecologists on the other hand, considered the total species pool, which showed no variability across

    land use gradients from the settlements compared to the benchmark.

    In the arid lowlands, vegetation recovery around the Kargi and Korr settlements was observed.

    Around the Korr settlement that was mapped some 25 years ago as being desertified, key fodder

    species now dominate the areas around the settlement, which also has higher herbaceous and shrub

    cover. In these arid lowlands herders have been actively involved in environmental management.

    The home herds are located in pastoral camps outside the settlements, and due to protection, the

    rangelands within a 4 km radius of the settlements show recovery of both herbaceous and woody

    vegetation. Monitoring of the communal grazing areas did not show significant variations in species

    composition over the 24 year period, although the herders reported changes in land use patterns from

    seasonal to year round grazing. The changes in species composition were confirmed by ecological

    ix

  • methods, which linked the changes to seasonal variability, as opposed to permanent loss in species

    composition. Around the permanent settlements of Kargi and Korr, satellite imagery showed an

    increase in vegetation cover between1986 and 2000. The evidence shows that conservation methods

    practiced by local communities have reversed land degradation reported earlier in the 1970s. The

    evidence further shows that the arid ecosystems of northern Kenya exhibited resilience with a

    capacity for recovery, when appropriate management measures were taken. The research found no

    evidence of permanent degradation leading to desertification. Herders monitored land degradation

    using multiple indicators including livestock productivity, landscape grazing potential, landscape

    grazing suitability and changes in vegetation and soil characteristics. Herders perceptions of land

    degradation were influenced by livestock production performance, from which they inferred other

    qualitative indicators, in addition to biophysical and anthropogenic indicators. From the observations

    made, there is no evidence of permanent land degradation around the settlements in either the sub-

    humid or the arid lowlands, which could be attributed to pastoral sedentarization. On the contrary,

    even the areas reported as desertified some 25 years ago have shown recovery. The findings also

    support the recent greening of the sahel which has been marked by vegetation recovery after the

    return of normal rains. In this study, both stochastic rainfall and management practices contributed to

    the reversal of desertification. From the results we may conclude that local community participation

    in assessment and monitoring of environment change in the grazing lands of northern Kenya would

    contribute to the successful implementation of GECs at community levels. The thesis showed that

    integrated methods would improve local communities participation in the implementation of GECs.

    The most important contribution of this thesis is the evidence about the ability of herders to assess

    and monitor environmental change, and the use of herder knowledge for selecting sensitive

    indicators that meet the criteria defined in the articles of GECs, particularly those related to the CCD

    and the CBD. The thesis makes specific recommendations for achieving the global goals through

    local actions that are linked to traditional pastoral production in northern Kenya.

    Key words: Biodiversity; global environmental conventions; herder indicators; ecological indicators; land degradation; local participation

    x

  • 1. Introduction

    This thesis aims to develop a methodological framework for achieving local participation in the

    implementation of global environmental conventions such as the Convention on Combating

    Desertification (CCD) and the Convention on Biological Diversity (CBD). The study involved

    assessment and monitoring of land degradation, desertification and biodiversity loss in the grazing

    lands of northern Kenya. The rationale of the thesis is that participatory environmental assessment

    and monitoring in arid zones in Sub-Saharan Africa in general, and northern Kenya in particular,

    have been constrained by the lack of involvement of local communities. Instead, assessment and

    monitoring of environmental problems have relied exclusively on conventional ecological methods

    and indicators selected by scientists. This happens in spite of the global environmental conventions

    (GECs) such as CCD and CBD placing strong emphasis on the need for participation by local

    communities (UNCED, 1992). Among the factors that have impeded local participation in the

    implementation of the GECs is a lack of integration of indigenous knowledge and ecological

    methods (Seely and Moser, 2004; Oba et al., 2008). An integrated methodological approach is

    therefore needed to address environmental degradation, particularly that associated with changes in

    land use intensification and desertification around sedentary nomadic settlements (Lusigi, 1981).

    Sedentarization of pastoral populations is either voluntary, as an adaptation to changes in

    economic and environmental conditions, or through forceful settlements by states for purposes of

    development or economic rehabilitation (Salzman, 1980). Whatever the cause, pastoral

    sedentarization has environmental consequences. In Sahelian Africa, over-exploitation of vegetation

    resources around settlements is reported to be responsible for inducing desertification (Swift 1975;

    Mabbutt 1984; Mabbutt 1985; Thomas et al., 2000). Pastures are said to be depleted, showing

    dramatic declines along gradients of land use pressure by livestock grazing. Depletion of woody

    species around settlements is reported to be associated with over-exploitation of woody plants for

    the construction of livestock night enclosures (Lamprey and Yussuf 1981), the collection of wood

    for fuel (Benjaminsen, 1993), cultivation (Lamprey, 1976), and over-browsing by livestock (Oba,

    1998; Oba et al., 2000a). The changes are pertinent to the principal goals of the GECs, which are

    concerned with reversing such problems through better ecological understanding and improved land

    management.

    In northern Kenya, pastoral sedentarization has been associated with land degradation and

    desertification (Lamprey and Yussuf, 1981; Lusigi 1984). The problem has been captured within the

    1

  • global debate of environmental degradation and desertification. Many of the scholarly research

    conducted during the 1970s emphasized environmental degradation, both in the sub-humid zones

    and the arid lowlands, where land use associated with pastoralists sedentarization was blamed.

    However, previous attempts to assess and monitor land degradation induced by pastoralists

    sedentarization in northern Kenya used scientific methods alone (Lusigi et al., 1986, Hary et al.,

    1996, Keya, 1997; Oba et al., 2003). In the earlier attempts, local community knowledge on

    environmental change was ignored and, no effort was made to integrate local peoples knowledge

    with ecological assessments. However, the local communities were usually blamed for contributing

    to the process of environmental degradation around sedentary settlements (Field, 1981). Due to the

    exclusion of local people, the findings from environmental assessments, including those of northern

    Kenya, have remained contradictory and the implementation of the GECs has been poorly addressed

    (see section 2.0 of this thesis). This thesis aims to close the gaps by proposing an effective method

    for integrating the indigenous knowledge of pastoral herders with ecological methods used by

    trained scientists. The thesis attempts this task by developing participatory methods, testing the

    efficacy of the methods and complementing indigenous knowledge with conventional ecological

    methods, as a process that will finally be aimed at meeting the global goals using local actions. The

    research is conceptualized within the wider global goals related to the implementation of the GECs

    at local community levels.

    1.1 Background to the study

    During the previous four decades, the Ariaal and Rendille pastoralists systems of livestock

    management in northern Kenya have been transformed from mobile to sedentary systems (Oba,

    1994). The process of settlement by nomads was accelerated as a result of drought disasters that

    impoverished many herders in the 1970s, forcing them into famine relief camps. In the sub-humid

    zone, the settlements in Karare, in the Marsabit District, were initially developed to rehabilitate

    destitute nomads by means of crop cultivation, while in the arid lowlands, the Rendille settled

    around the main towns of Kargi and Korr to receive food relief (Fig. 1). Later, the government and

    missionaries initiated development programs that encouraged pastoralists to settle around these

    towns, which had grown into administrative, educational and security centers that provided watering

    facilities for humans and livestock. Despite the changed patterns of land use, the Ariaal and Rendille

    2

  • pastoralists maintained mobility of their herds, while the greater proportion of the human population

    remained in the sedentary settlements. The main concerns for the scholarly discussions in the 1970s

    were the impact of sedentary land use on the scarce vegetation resources. The scientists of the

    UNESCO-Integrated Project on Arid Lands (IPAL), who conducted much of the earlier work, were

    concerned about the accelerating land degradation and loss of vegetation cover, due to the over-

    exploitation of woody vegetation for the construction of night enclosures (boma), as well as

    overgrazing associated with high livestock stocking densities around the settlements. One estimate

    suggests that within the settlement rangelands, livestock stocking densities exceeded 25 Tropical

    Livestock Units per km, whereas in the remote rangelands the stocking densities were much lower.

    The heavy stocking was linked to the lack of woody plant regeneration (Lamprey, 1976).

    The research goals of the UNESCO-IPAL project were to understand the state of land

    degradation and livestock production within 18,000 km2 of the home range of the Rendille and the

    Ariaal pastoralists. Using vegetation maps, the grazing lands were delineated according to

    vegetation units (hereafter referred to as range units) that corresponded with grazing landscapes

    used for seasonal grazing by the multi-species livestock comprising sheep, goats, cattle and camels.

    During the initial surveys conducted throughout the grazing home range of the Ariaal and the

    Rendille pastoralists, the conclusion was that the conditions of the rangelands varied from fair to

    poor, except for a few vegetation types in the sub-humid zone of Marsabit mountain that were

    rated as good (Lusigi, 1984). Around the settlements, increased extraction of woody vegetation

    for building human shelters and livestock enclosures resulted in reduced woody cover, while

    livestock grazing pressure was reported to have had negative impacts on the herbaceous vegetation

    and woody species regeneration (Walther and Herlocker, 1980; Lamprey and Yussuf, 1981). In the

    lowlands, vegetation around settlements was over-utilized and the vegetation communities were

    mapped as man-made deserts, which in the arid lowlands, extended up to a distance of 8 km from

    the settlements (Lusigi, 1984). The general scientific perception at the time was that the man-made

    deserts around the settlements were expanding and threatening the surrounding grazing lands. The

    perception was influenced by discourses at the time on the topic of Sahelian desertification

    (Lamprey, 1983).

    Although an interdisciplinary team of researchers conducted the investigations aimed at

    placing the people central to the problem, there was little evidence that the ecological research

    3

  • included the local herders in conducting environmental assessments.1 Despite this lack, the

    anthropological research uncovered explicitly the wealth of knowledge that the pastoralists had in

    managing their arid environments (OLeary, 1985). However, mostly the scientific empirical

    ecological research findings were used to influence the recommendations proposed by the

    management plan (Lusigi, 1984).

    The management plan considered policy recommendations for the development of the Ariaal

    and Rendille rangelands, emphasizing the need for considering development according to local

    socio-economic needs and priorities set by the local people. However, it gave little weight to the

    role of indigenous knowledge in resource management, by recommending that range areas should

    be developed, conserved, and managed in accordance with the ecological principles of proper land

    use (Lusigi, 1984:484). At the time, the thinking of the IPAL research team was influenced by the

    equilibrium ecological theory that readily blamed the pastoralists for causing desertification due to

    high livestock stocking densities around settlements resulting in over-exploitation of vegetation

    resources (Field, 1981). Interestingly enough, whereas the ecological component of the project

    concluded that the environmental changes were permanent, leading to desertification, the social

    scientists in the team placed the problem within the wider socio-economic and political discourses

    (e.g. OLeary, 1984).

    By comparison, the assessment and monitoring of the status of the rangelands in northern

    Kenya conducted by UNEP and the Government of Kenya a few years after the IPAL project

    (Ottichilo, 1990), was more conservative concerning environmental degradation. The UNEP study

    used remote sensing and ground surveys of indicators of degradation, including physical (climate

    and soil), biological (vegetation) and socio-economic indicators, to model different processes

    influencing environmental change. It was concluded that the rangelands of northern Kenya,

    including those of the Marsabit District, had not suffered adverse changes over the monitoring

    periods. Another study by Herlocker and Walther (1991) reported that less than 1% of the district

    was in poor environmental condition. The discrepancies between the IPAL results of the 1970s

    and 1980s, and those of 1990s imply that there was inconsistency in the evaluations of land

    degradation. The present study therefore used the IPAL work as a baseline for monitoring and

    assessing vegetation changes around four main settlements, using multiplicities of methods that

    1 Much of the environmental awareness was, however, conducted by means of training local herders (see Oba, 1985a, 1985b; Lusigi, 1984).

    4

  • fully integrated herder indigenous knowledge with ecological methods, in order to understand the

    dynamics of land degradation.

    This thesis is aimed at improving the understanding of environmental changes in the sub-

    humid and the arid lowland systems of northern Kenya since the UNESCO-IPAL research was

    conducted in the early 1980s. The sub-humid zone and the arid lowlands were selected to represent

    the two dominant production systems used by the Ariaal (cattle) pastoralists and the Rendille

    (camel) pastoralists, respectively. According to the assessments of UNESCO-IPAL, the sub-humid

    zone suffered less environmental degradation, while the arid lowlands experienced localized

    desertification. This thesis explores the state of degradation processes across the two systems in the

    time period since the previous studies.

    The individual studies in the thesis were guided by the general hypothesis that, due to their

    ecological dispositions, the two systems have experienced different environmental changes under

    the influence of local land use in terms of livestock management. Throughout the thesis, the two

    systems were, however, not directly compared for the following reasons. Firstly, the impact of

    stocking density in both systems was unknown due to a lack of reliable livestock census data, and

    frequent livestock movements between the two systems. Secondly, the most important variable in

    measuring environmental change was vegetation characteristics. The two systems, due to

    differences in the levels of rainfall, altitude and topography, display differences in vegetation

    characteristics in terms of species composition and cover, regardless of levels of use. Thirdly, the

    study used local knowledge of the Ariaal and Rendille pastoralists. Since the two groups use

    different ecosystems, it was assumed that their knowledge of environmental change would differ;

    thus preference was given to using the knowledge of the two pastoral groups to understand

    environmental changes in their local land use contexts. Nonetheless, the selection of the sub-humid

    (equilibrium) and the arid zones (non-equilibrium) was useful in understanding the characteristics of

    vegetation changes that might be explained in terms of existing ecological theories. Furthermore,

    the use of different ecosystems was useful in investigating whether concepts such as land

    degradation and desertification might be relative to (a) use consistent with indigenous knowledge

    and (b) who is assessing and monitoring the herders or the ecologists.

    5

  • 1.2. Organization of the thesis

    This thesis is divided into two parts. Part A is the synthesis of a framework for the integration of

    indigenous knowledge and ecological methods for assessment and monitoring of environmental

    change at the local level (see section 2.0). The nine sections in this part of the thesis present the

    theoretical framework necessary for linking the implementation of the GECs and local

    communities actions to ecological principles. In section two of Part A, a brief description of the

    participatory research is presented and the frequently used terms in the thesis are defined in the

    context of their use. In section three, the theoretical perspectives for integrating indigenous

    knowledge and scientific methods are discussed. In section four, the methodological perspectives

    for achieving the integration of indigenous knowledge and ecological methods are described. In

    section five, the mechanisms for developing integration are given, using a schematic framework that

    links global and local goals. In section six, the steps followed in the implementation of the

    framework at local community levels in northern Kenya are described. In the seventh section, the

    synthesis of the key findings of the four case studies is provided. The eighth section provides

    discussions on the implications for the global goals. The final section of Part A provides conclusions

    and recommendations. In Part B of the thesis, four papers from the individual studies are presented.

    The general objectives of the thesis were as follows:

    (1) To develop a methodological framework for integrating indigenous knowledge and ecological

    methods for achieving the implementation of GECs at local community levels. The theoretical and

    methodological considerations for integrating indigenous knowledge and ecological methods were

    reviewed.

    (2) To understand the impact of sedentarization on the vegetation in the sub-humid and the arid

    lowlands. Key questions for addressing this objective included: How did herders livestock

    management strategies affect livestock distribution in the rangelands around settlements? What

    were the impacts of sedentarization on the vegetation around settlements?

    6

  • (3) To understand herders knowledge of biodiversity assessment and monitoring at landscape level.

    The key questions were: What were the herders landscape classification criteria; What indicators

    did the herders use for floral biodiversity assessment and monitoring; and what roles did the

    herders biodiversity assessment and monitoring play in the decision making process for livestock

    management?

    (4) To understand herders perceptions of land degradation and the influence it had on herd

    management strategies. The questions posed were: What were herders perceptions of good or

    bad environments? What indicators did the herders use for assessment and monitoring change in

    environments from good to bad, for purposes of livestock management?

    (5) To evaluate the efficacy of herders and ecological assessment and monitoring of long-term

    environmental changes in arid rangelands. The research questions were: What were the changes in

    vegetation characteristics in the communal grazing area over a period of 24 years (1982/1983-

    2005/2006)? What were the changes in vegetation cover around the Kargi and Korr settlements over

    a period of 14 years (1986-2000)? What factors contributed to vegetation cover changes in the

    communal rangelands and around the settlements of Kargi and Korr?

    1.3. Study sites

    The study sites were located in the sub-humid zone (Karare and Lkijiji) and the arid low lands (Korr

    and Kargi) of the Marsabit District (Fig. 1).

    7

  • Figure 1. Location of study sites on a map of the Marsabit District

    8

  • The Karare and Lkijiji sites receive a mean annual rainfall of about 600 mm per year (Fig. 2a). The

    rainfall is bimodal, divided into the long rains (March to June) and the short rains (October and

    November) (Fig. 2b). The vegetation of the sub-humid zone is classified as

    Pennisetum/Bothriochloa (Perennial grassland) (Awere-Gyekye, 1984). The Ariaal herders, who

    manage cattle and small stock, also conduct limited cultivation (the later system of land use is not

    discussed in this thesis). For about 30 years, the area has experienced continuous livestock grazing,

    mainly by cattle and small stock around the permanent settlements clusters of Karare and Lkijiji.

    0200400600800

    100012001400160018002000

    19

    35

    19

    40

    19

    45

    19

    50

    19

    55

    19

    60

    19

    65

    19

    70

    19

    75

    19

    80

    19

    85

    19

    90

    19

    95

    20

    00

    Rai

    nfal

    l (m

    m)

    Year

    Figure 2a. Annual rainfall for Marsabit town

    0

    50

    100

    150

    200

    250

    300

    Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

    Mea

    n ra

    inaf

    all (

    mm

    )

    Months

    Figure 2b. Mean monthly rainfall for Marsabit town between 1935 and 2004

    9

  • 050

    100150200250300350400

    1980

    1982

    1984

    1986

    1988

    1990

    2000

    2002

    2004

    2006

    Years

    Rai

    nfal

    l (m

    m)

    Figure 2c. Total annual rainfall for the Kargi station The arid lower lands of Korr and Kargi (Fig.1) receive less than 200 mm of annual rainfall, which is

    highly variable both temporally (Fig. 2c), and spatially (Pratt and Gwynne, 1977). Droughts are

    frequent in the area (Bake, 1983). The vegetation is mainly dwarf shrubs dominated by Indigofera

    spinosa and I. cliffodiana, and bushland dominated by Acacia sp. (Lusigi et al., 1986). The major

    land use type is grazing by multi-species livestock managed from pastoral camps (gob).

    In the sub-humid system, cattle and small stock herds are split into the home and the fora

    herds (Fratkin, 1986; Fratkin, 1987). During the wet season, the fora cattle are moved to the

    lowlands to exploit available fresh grasses and surface water. During the dry season, the mobile fora

    camps return to the home rangelands and the livestock are watered at the wells in the Marsabit

    Forest Reserve. The small stock remains close to the permanent settlements in the sub-humid zone.

    At the settlements, fewer animals (cattle and small stock) are managed for milking. But generally

    there is a circulation of livestock (milk vs. dry herds) between the settlements and the fora systems.

    The milk animals are brought back to the settlements and the dry animals are sent to the fora.

    Besides the sub-humid zones around Mt. Marsabit, the Ariaal pastoralists are also found

    around Logo Logo (approximately 40 km to the south of Marsabit), Merrile (approximately 100 km

    to the south of Marsabit) and around the Ndoto and Mathews mountain ranges bordering the

    Samburu District in the south. The Ariaal residing around the Ndoto Mountains and Mathews Range

    use the surrounding lowlands, in a similar way to their counterparts on the slopes of the Marsabit

    Mountain. According to Fratkin (1987), the Ariaal have remained nomadic, with some sub-clans

    occupying highland areas subsisting on cattle, while others living in the lowland areas manage

    10

  • camels. The livestock species managed by the Ariaal clans have different grazing requirements and

    are therefore herded as a separate group, resulting in high labour demands (Fratkin, 1987). The

    Ariaal living in the lowlands have adopted the Rendille cultural lifestyle, while the Rendille in the

    sub-humid zones have become bilingual in Maa and Rendille. The practice of intermarriage has

    created shifting identities between the two communities.

    In the arid lowlands, the Rendille pastoralists also practice herd splitting into mobile satellite

    camps (fora) and home herds (gob) (Fratkin and Smith, 1994). The camel mobile camps are located

    far from settlements and water points. The home settlements are positioned near permanent water

    points, as cattle and small stock need to be watered more frequently (Fratkin, 1986). Each Rendille

    clan resides in separate pastoral camps where the elders meet in the evening at a gathering in the

    centre of the camp called naabo, to discuss matters such as the state of grazing, livestock

    management and other matters of importance to the community. Aerial surveys of the camps done

    by IPAL in 1982/1983 show that the camps had 17-23 houses on average and 47% were located

    around the major settlements centers of Kargi, Korr, Logo Logo and Laisamis (OLeary, 1985).

    Unlike the Ariaal settlements, the Rendille pastoral camps are frequently moved and each time

    the moves are made, sometimes just a short distance away, new thorny trees are cut to build night

    livestock enclosures. Once a year the Rendille bring all their livestock and people to the main camps

    for the almado and sorio ritual ceremonies (OLeary, 1984; Fratkin 1986, 1987; Schlee 1991). The

    greatest threat to the environment is therefore over-harvesting of woody plants for the construction

    of night livestock enclosures (boma) and the greater frequencies of camp movements that require

    fresh tree materials each time. Consequently, the areas around the camps have been over-exploited

    and supplies of fencing materials have become scarce. In the settlements, where people previously

    transported fencing materials, the Rendille now use camels to ferry in the same materials from long

    distances. The environmental impact has important implications for the implementation of GECs at

    local community levels.

    2. Integration of global goals and local actions Participatory research and development is rooted in the shift in theories from modernization theory

    associated with top-down technological transfer, to neo-populist theory that advocates for local

    people participation which uses bottom-up approaches in research and development (Sillitoe,

    1998). The proponents of bottom-up or local participation approaches, present convincing

    11

  • arguments that local people have accumulated a wealth of knowledge over time, based on long-term

    experiences, that can complement scientific knowledge in environmental conservation (Richards,

    1980; Knight, 1980). Major emphasis is placed on the roles indigenous knowledge and local

    management play in conservation (Warren, 1992; Berkes et al., 2000), and protection of the land

    from degradation (see below for definitions). Additionally, there is growing interest on how

    indigenous knowledge and management practices can be used in collaboration with standard

    scientific methods for improving understanding of the environment (Dahlberg, 2000; Reed et al.,

    2007). Global environmental problems and the need for local participation were discussed during

    international conferences related to UNCOD and UNCED, where participating nations agreed on

    joint action plans. Since environmental changes are attributed to a multitude of factors (Geist,

    2005), appropriate methods and sensitive indicators are needed for assessment and monitoring.

    For the above reasons, Agenda 21 of the Rio Conference recognizes the role local

    communities play in environmental assessment and management (UNCED, 1992). According to

    Agenda 21, partnerships with local communities aim to achieve the global goals for the sustainable

    use of natural resources. For example, the role of indigenous knowledge in combating

    desertification and droughts is contained in Agenda 21, Chapter 12 part 18 (d), which states that the

    United Nations seeks to promote participatory management of natural resources, including

    rangeland, to meet both the needs of rural populations and conservation purposes, based on

    innovative or adapted indigenous technologies. Further, Agenda 21, Chapter 12, subsection 23 (a),

    states that government should integrate indigenous knowledge related to forests, forest lands,

    rangeland and natural vegetation into research activities on desertification and drought. In relation

    to the conservation of biological diversity, Agenda 21, Chapter 15, part 4 (g) states that global

    partners should recognize and foster the traditional methods and the knowledge of indigenous

    people ...relevant to the conservation of biological diversity and the sustainable use of biological

    resources.

    Various nations are signatories to the global conventions through the ratification of the

    different articles. The national goals are to implement the conventions by means of National Action

    Programs (NAPs), which in turn, comprise strategies and methods for implementing the global

    goals. Considering the broad geographical, and ecological variability, and the socio-economic

    factors linked to the process of environmental change, there is no single indicator that may be used

    for the assessment and monitoring thereof. The goals at the national level therefore include linking

    12

  • scientific assessment and monitoring activities with the environmental management practices of

    local communities. For example, according to Article 10 (f) of UNCCD (1994), parties to the

    convention should: provide for effective participation of local, national and regional levels of non-

    governmental organization and local population, ..particularly resource users including farmers

    and pastoralists and their respective organization in policy planning, decision making and review of

    national programs.

    Additionally, according to the Convention on Biological Diversity (CBD), each member state party

    to the convention is obliged to promote local participation in the management of biological

    diversity. Article 8j of the CDB states:

    Subject to its national legislation, respect, preserve and maintain knowledge, innovations and practices of indigenous and local communities embodying traditional lifestyles relevant for the conservation and

    sustainable use of biological diversity and promote their wider application with the approval and involvement

    of the holders of such knowledge, innovations and practices and encourage the equitable sharing of the benefits

    arising from the utilization of such knowledge, innovations and practices.

    The NAPs can benefit from global mechanisms (such as the Global Environmental Facility) that

    make technology and funds available for initiating implementation activities by means of local

    participation. By tapping into local knowledge systems, environmental assessment and monitoring

    in response to anthropogenic and natural ecosystem drivers at local levels can be achieved

    (Krugman, 1996). Previously, the use of indigenous knowledge for promoting local participatory

    assessment and monitoring of environmental change was constrained by the lack of integration of

    local knowledge systems with scientific methods. Conducting evaluations and monitoring human

    impacts on the environment are pre-requisites for accomplishing the implementation of the GECs.

    The purpose of this study is therefore to shift the approach to enable local communities, such

    as herders, and ecologists to be partners in the assessment and monitoring of the implementation of

    the GECs at community levels, using traditional systems of land use. Rather than discussing the

    actual implementation of the GECs, the main purpose of the thesis is to discuss the potential

    application of integrated methods for implementing the goals of NAPs at local community levels.

    The success of the integration of local knowledge in environmental assessment and monitoring is

    influenced by shifts in theoretical viewpoints on environmental discourses. Such discourses in turn

    13

  • determine the type of research methods that can be used to link management objectives of local

    communities and conservation goals. Participatory methods face challenges in terms of questions

    relating to how local knowledge can be used how to collaborate with conventional scientific

    methods and understand the perspectives of indigenous knowledge, particularly how it functions in

    relation to environmental change. Whereas the need for integrating participatory knowledge with

    conventional scientific methods is often demanded as a prerequisite for achieving global

    participation, any frameworks for achieving such integration are poorly documented (for a recent

    attempt, see Dougill and Reed 2006; Oba et al., 2008).

    A schematic framework for understanding how the objectives of the GECs could be used to

    guide national and local actions to achieve global and local goals is proposed in this study.

    Integrated methods (see Fig. 3, section 5) based on local environmental knowledge and ecological

    methods were tested to show how the broader global objectives for the implementations of GECs

    such as the CCD and CBD could be tackled at local community levels. Before discussing the

    approaches used for achieving these goals at community levels, it is necessary for the reader to be

    familiar with how some of the terminologies and concepts were applied in this thesis.

    2.1. Definition of terms and concepts

    In this thesis different terms are used in relation to the assessment and monitoring of land

    degradation and biodiversity loss. The terms are explained within the context of their use, and

    therefore universal agreement is not presumed. For example, the knowledge held by the local

    communities has been described using different terms including: indigenous knowledge, local

    knowledge, folk knowledge, indigenous technical knowledge (ITK), traditional ecological

    knowledge (TEK) and indigenous ecological knowledge (IEK), among others. Each of the terms

    has a different connotation in terms of importance and the application of knowledge by different

    communities. In this thesis, some of the terms are used interchangeably to describe the knowledge

    held by local communities. These knowledge systems are used by individuals but represent the sum

    total of knowledge used by particular local communities (Roba and Oba, 2008). However, reference

    is made to the most popular terms, such as indigenous knowledge/local knowledge and

    traditional ecological knowledge to describe the roles local communities play in environmental

    management. According to Warren (1991), the term indigenous knowledge describes the

    14

  • knowledge developed by a given community, which is different from scientific knowledge systems

    generated through universities or at government research stations. Indigenous ecological knowledge

    refers to experiences acquired over a lifetime through observations and in relation to social norms

    and institutions that shape human interaction with the environment (Berkes et al., 2000; Fernandez-

    Gimenez, 2000). Such knowledge is useful, for example, in describing concepts such as land

    degradation.

    Land degradation is a composite term and the definition depends on the context used by

    both scientists (Stocking and Murnaghan, 2001; Warren 2002) and local communities (Oba and

    Kotile, 2001). In general, land degradation is defined as the loss of utility or potential in relation to

    biological organisms as well as changes in the physical environment that would alter the functions

    of natural systems (Abel and Blaikie, 1989; Barrow, 1991). According to Dodd (1994), degradation

    may refer to a decrease in plant productivity or unfavorable changes in species composition, but

    does not imply that changes are permanent. In the rangelands, Abel and Blaikie (1989), defined land

    degradation as a permanent decline in land for the yielding of livestock products under a given

    system of production. This means that in terms of pastoral production, where milk and meat are

    major products, land degradation leads to a downward spiral in livestock productivity. In pastoral

    systems where multiple livestock species are managed, better insight about the processes of

    environmental change can be gained by considering herders perceptions. Herders define land

    degradation in relation to livestock productivity. A degraded environment, according to herders,

    does not support livestock productivity at optimal levels due to the loss of important fodder species.

    Accordingly, a landscape that is degraded for one species of livestock e.g. grazers, may not be so for

    browsers. Throughout this thesis, a broader approach to understanding land degradation is adopted,

    rather than narrow ecological definitions alone. In all cases, land degradation will be considered to

    be reversible under improved systems of management.

    Ecologists also use other terms such as desertification when referring to extreme levels of

    land degradation. The concept of desertification implies both temporal and spatial perspectives.

    Internationally, desertification is defined as land degradation in arid, semi-arid and dry sub-humid

    areas resulting from various factors, including climatic variations and human activities (UNCCD,

    1994). The nature, extent and reversibility of the changes associated with desertification have been

    at the center of scholarly debates (e.g. Stiles, 1995; Thomas and Middleton, 1994; Helldn, 1988).

    In this thesis, the terms desertification and land degradation are used concurrently, but often

    15

  • desertification implies permanent and irreversible changes in vegetation and soil conditions,

    whereas land degradation is used to describe changes which are reversible with management or

    when the anthropogenic pressures are removed. Given that the classification of both degradation and

    desertification might rely on similar indicators (see section 5), the issue is at what levels changes

    would be described as simple degradation or as desertification. One of the important indicators for

    evaluating land degradation and desertification is changes in floral biodiversity.

    Biodiversity, or biological diversity, is a broad concept that describes the variety of all

    living life forms encompassing species, genetics and ecosystems. In this thesis, the focus of

    biodiversity is mainly on the diversity of plant species. In the assessment and monitoring of changes

    in plant species diversity, conservation interests are based on the assumption that loss of species

    diversity (number of different species) and species richness (number of species per unit area) are

    important criteria for assessing ecosystem degradation. In the grazing lands, herders are concerned

    with satisfactory livestock production. Therefore the concept of the total species pool, which is

    important from the conservation viewpoint, does not adequately capture herders requirements.

    According to herders, changes in biodiversity in the grazing lands refers to changes in plant species

    composition in relation to livestock fodder requirements (Mapinduzi et al., 2003; Roba and Oba,

    2008). By focusing on key forage species, the herders use utilitarian definitions of biodiversity. The

    concept of invasive species is well known to ecologists, while herders refer to bad or good

    biodiversity in terms of the extent of unpalatable plant species, such as those that might be

    associated with bush encroachment. In this thesis, the use of the term biodiversity refers to both

    the conservation and utilitarian values in terms of how it is assessed and monitored.

    Assessment and monitoring of vegetation are relative terms used for understanding

    environmental dynamics. Assessment refers to observation of the status of various indicators that

    influence environmental health. It involves evaluation at one point in time to generate baseline data

    on vegetation and soil physical characteristics. For herders, assessment is done more frequently

    across the grazing landscape to ensure an acceptable quality and quantity of fodder for multiple

    livestock species. Ecological assessment of grazing ecosystems is less frequent and on limited

    spatial scales. Observations are usually made at a series of sampling transects and plots in order to

    generate data that are used to generalize the status of grazing resources. By comparison,

    monitoring is an evaluation process conducted several times over long periods to determine

    responses to management and other environmental factors such as rainfall (Holecheck et al., 1995).

    16

  • Monitoring of important environmental variables, such as change in species composition, is

    conducted using similar methods, for the same objectives and over a long period of time. The

    observed trends in the variables monitored can then be related to different drivers. Herders

    assessment and monitoring of environmental change is not limited to environmental indicators, but

    includes livestock production performances such as volume of milk and animal body condition. In

    this thesis, assessment and monitoring describe activities conducted by herders and ecologists to

    determine the suitability of the grazing environment on short-term and long-term basis. Assessment

    and monitoring must be conducted by range managers to understand how management has

    influenced the condition of the rangelands (short term) and trends (long term changes) in

    environmental indicators.

    Vegetation condition and trends are terms commonly used to describe the state of

    range health (NRC, 1994). Range condition is scaled in terms of the climax vegetation. The use

    of the concept is limited due to difficulties in establishing a climax vegetation composition for

    determining change in arid zones. The idea of using climax vegetation for determining healthy

    range condition shows that the term is closely related to the equilibrium ecological theory. In this

    thesis the concept is used in relation to the range status which the herders consider as being

    optimal for livestock production, while ecologists use the term to mean a departure from the

    assumed climax vegetation status. Thus in this thesis, the term condition is used in relation to

    utilization. Trend refers to the direction of change of range condition, which can be rated as

    upward, downward or stable (Holecheck et al., 1995). To describe trend, an observation of change

    in species composition is required. In pastoral systems, herders have accumulated knowledge of the

    direction of vegetation change more so than ecologists.

    Another term that appears often in the thesis is integration which, as used here, implies the

    combined use of local knowledge and scientific methods to understand environmental change.

    Integration of assessment and monitoring of environmental change is achieved by the simultaneous

    use of multiple indicators used by local people and ecologists. The local communities use

    indigenous knowledge and composite indicators (hereafter also referred to as anthropogenic

    indicators), while ecologists use ecological methods and ecological indicators. The integration of

    local knowledge and ecological methods improves understanding of environmental change, as the

    two systems complement each other. Thus, could integration be achieved by asking the local

    informants questions about the environment and using ecological methods to measure required

    17

  • variables? Or does integration involve a whole range of processes of environmental assessment and

    monitoring and decision-making? In this thesis integration is taken as a process that finally leads to

    rational decisions. Creating a situation for sharing information and understanding different

    viewpoints by local communities as well as by scientifically trained technicians, resulted in a

    common forum for discussing the problems of the environment. This approach has been the goal of

    this thesis. By jointly using the different methods it was possible to understand the perceptions

    about the reversibility of land degradation, which is related to the concept of resilience.

    The concept of ecological resilience was described by Holling (1973) to describe changes in

    ecosystems when subjected to perturbation. Resilience is defined as the capacity of a system to

    buffer or resist change in response to a given magnitude of disturbance, before losing the capacity to

    respond (Gunderson, 2000; Perrings and Walker 1995). The level of resilience can be defined in

    terms of systems potential. In grazing lands, for example, different landscapes disclose varied

    potential in relation to soils and vegetation. The potential of the resource system describes the

    capacity of the system to resist degradation. In this thesis, the concept of ecological resilience is

    used to understand change in species composition and vegetation structure. Long-term fluctuations

    in vegetation variables were examined to see if they present a characteristic of resilience under

    unpredictable rainfall regimes, or if they portray a more linear change as presented by the

    equilibrium ecological model. In addition, variables such as species inventory, change species

    frequency, and species cover over long periods of time are important for understanding the

    resilience of an ecosystem after years of droughts and sustained grazing pressure. The capacity of

    the system to spring back is what allows land degradation in arid ecosystems to be reversible; the

    lack of resilience would cause desertification (Binns, 1990; Oba et al., 2008). Herders are aware of

    the resilient property of arid ecosystems, based on several years of observation. In arid ecosystems,

    the concept of resilience has helped herders to develop adaptive management strategies that enable

    them to modify their management strategy according to prevailing environmental conditions. Using

    their knowledge of individual landscapes and their capacity to cope with grazing pressures, herders

    regulate grazing movements allowing the land to regenerate, even after heavy use. This shows that

    management strategies have an influence on the resilience of an ecosystem (e.g. Perrings and

    Walker, 1997).

    Each of the concepts or terms discussed here elicits different discourses in terms of

    environmental change. It is important to mention that no single theoretical viewpoint can adequately

    18

  • address the different discourses, necessitating the application of an interdisciplinary approach for

    understanding the process of environmental change. The terms discussed above are used in this

    thesis for discussing different theoretical and methodological perspectives related to local

    community participation in the implementations of the GECs related to the CCD and the CBD. The

    concepts are important for understanding continuous shifts in paradigms of environmental change

    from deterministic cause-effect views to multi-directional approaches that acknowledge the

    influence of environmental variability and the importance of indigenous knowledge.

    3. Theoretical perspectives

    Major environmental discourses have played a central role in shaping how environmental problems

    are perceived. The dominant environmental discourses such as desertification and biodiversity loss

    have attempted to portray a crisis scenario, especially in marginal environments (Lamprey, 1983).

    Increases in human and livestock populations have been associated with adverse effects on the

    environment. The crisis narratives are rooted in different environmental theories, which predict the

    relationships between biotic and abiotic, social-ecological and economic components of an

    ecosystem. For example, different viewpoints of land degradation in arid ecosystems have been

    influenced by different scientific theories. The ecological theories reflect a priori environmental

    functions in terms of processes, explained in terms of deductive relationships between causes and

    effects of land degradation. Management is often not part of the theory, although the impact of

    management on the environment uses the theory to analyze the effects. This implies that local

    knowledge in environmental assessment and monitoring is usually not part of the theory description

    and verification. This was until recently, when ecologists re-evaluated existing ecological theories

    for guiding management decisions, particularly in arid lands (Behnke and Scoones, 1993), for the

    following reasons. Firstly, ecological theories do not explain all the outcomes of environmental

    change, particularly where management or human decisions in land use are involved. Secondly, for

    ecosystems such as arid lands, earlier ecological theories had assumed stability, while the system

    behavior is better described by variability. Empirical evidence in support of spatial and temporal

    variability questioned the value of using ecological theories that prescribed stability and

    predictability (Ellis and Swift 1988). The stable and predictable systems failed to acknowledge the

    management systems of local resource users, while the variable systems considered the rationale of

    19

  • local resource use. These developments therefore became the impetus for the integration of

    indigenous knowledge and ecological methods.

    Various viewpoints from the perspective of the dynamics of vegetation in arid ecosystems in

    response to land use by local pastoralists are reviewed here. Since the early days of range

    management as a science (e.g. Clements 1916), ecologists considered proper management in terms

    of the equilibrium between grazer populations and allowable forage utilization (Heady 1975). This

    theory holds that range managers should maintain the numbers of grazers on a given range

    commensurate with its potential carrying capacity which is defined as the maximum stock carried

    per unit of land in a given time (Pratt and Gwynne, 1977; Bartels et al., 1993). The theory presumes

    that it is the grazers that drive the changes in plant composition and therefore by regulating stocking

    levels, the dynamics of vegetation can be maintained at desirable levels of plant composition to

    promote environmental sustainability.

    The equilibrium theory, which is described above in a simplified version, is deterministic and

    uni-directional, since it does not take into consideration environmental drivers, such as rainfall

    variability, as the principal control agents for driving range production. Indeed, the theory, which

    was designed under humid conditions, might work adequately in ecosystems with predictable

    rainfall, but problems arose in arid environments (Ellis and Swift, 1988). Using the responses of

    vegetation parameters to herbivore populations, the theory would predict negative changes when the

    population exceeds the carrying capacity or the stocking potential (Lamprey and Yussuf, 1981;

    Lamprey, 1983; Sinclair and Fryxell, 1985). Therefore, based on the equilibrium theory, land

    degradation that occurs in arid lands, would be blamed on management systems that ignore the

    equilibrial relationships between natures functions and use.

    The equilibrium model influenced early pastoral development throughout sub-Saharan Africa

    in a significant manner, with adverse consequences for the environment and production systems

    (Sandford, 1983). One such adverse consequence was sedentarization of former nomads that

    resulted in the breakdown of traditional systems of land use, causing precisely those environmental

    problems which the theory was meant to guard against (Sinclair and Fryxel, 1985). The proponents

    of the theory excluded local indigenous knowledge and purposely focused on the use of ecological

    indicators (see below) for environmental assessment and monitoring. There are, however, critical

    similarities between the equilibrium theory and its explanation of vegetation changes in relation to

    grazing pressures, and local knowledge of herders on how vegetation indicators respond to sustained

    20

  • grazing. The equilibrium theory postulates that sustained grazing pressure induces shifts in plant

    species composition, where some species that are more sensitive decrease, while others, that are

    unpalatable or more resistant to grazing pressure, increase or remain stable. Careful analysis of

    indigenous knowledge shows that herders have comparable understanding about the species that

    would decrease, and others that would increase or remain stable. The qualifying difference is that

    according to the local herders, the shifts in plant species vary with the type of livestock. The

    implication is that land degradation cannot be a universal problem in grazing lands. The land

    degraded for grazing livestock would still be sustainable for browsing stock and vice versa.

    Ecologists rarely took such views into account.

    Shifts in ecological thinking (e.g. Ellis and Swift 1988) propose that vegetation changes in

    highly variable environments, such as in arid lands, are more sensitive to environmental drivers

    including rainfall variability, than they are to grazer populations. The non-equilibrium theory

    proposes that in environments with high coefficients of variation, the spatial-temporal dynamics of

    vegetation resources cannot be accounted for by grazing alone (Fernandez-Gimenez and Allen-Diaz

    1999; Oba et al., 2000b; Sullivan and Rhode, 2002). Rather, the rhythms of range production are

    closely related to spatial and temporal rainfall variability. This means that the same environment

    might experience a boom at one time, and burst production at another, in response to varied

    rainfall regimes. Rainfall varies from season to season, resulting in substantial differences in range

    production according to seasons (temporally) and spatially by sites or geographical distribution of

    plant production. The non-equilibrium model of rangeland dynamics therefore puts emphasis on the

    unpredictable nature of ecosystems and the inability on the part of management to develop practical

    plans based on prior knowledge for manipulating stocking rates. Pastoral production, which

    involves livestock mobility, is adapted to variability (Fernandez-Gimenez and Allen-Diaz 1999;

    Oba et al., 2000b). Pastoral land use systems track the variable resources opportunistically (Behnke

    and Scoones, 1993).

    Herders are aware that plant species composition and cover change with variation in rainfall

    and across heterogeneous landscapes. Herders use this knowledge of seasonal and spatial variability

    of grazing resources to promote mobility. Local knowledge of resource management could therefore

    potentially contribute to the assessment and monitoring of the vegetation dynamics aimed at

    understanding the mechanisms described by different ecological models. The role of local

    21

  • knowledge is deemed to be even more relevant when the dynamics of arid ecosystems are described

    using social-ecological resilience viewpoints.

    According to the social-ecological resilience viewpoint, ecological processes are closely

    linked to social activities, including decisions on land use and livestock management (Berkes et al.,

    1998; Oba et al., 2008). In the social-ecological resilience model, adaptive management is central to

    pursuing livestock production goals. Pastoral adaptive management involves making adjustments to

    production variability by means of mobility. Thus, the explanation offered by the non-equilibrium

    theory about the dynamics of arid lands has closer parallels with the socio-ecological resilience

    theory. By means of the latter, the functions of indigenous knowledge in the management of

    variable environments can be justified. Based on their detailed knowledge of seasonal variations in

    species composition, herders might suggest that a particular plant species is not present at the time

    of assessment (e.g. in dry season), but will be seen again in the wet season. Thus, herders believe

    that arid ecosystems are highly resilient and in their view degradation occurs only when livestock

    mobility is curtailed and heavy grazing is sustained over a long period that would result in loss of

    key forage species.

    The importance of local knowledge in addressing complex environmental problems has been

    supported further by interdisciplinary studies that used the principles of political ecology. Political

    ecology underscores the importance of environmental narratives and discourses for addressing

    desertification and biodiversity loss (Leach and Mearns, 1996; Batterbury et al., 1997; Laris, 2004).

    Studies in human-environment relationships have shown that many of the environmental crises

    debated at global level are exaggerated (Bassett and Zuli, 2000). The narratives advanced at global

    levels on the status of environment should be related to the counter narratives of local people (e.g.

    Bassett and Crummey, 2003). Political ecology and environmental history therefore provide

    important links for analyzing human-environmental interactions (Benjaminsen and Lund, 2001).

    Local communities narratives on causes and trajectories of environmental change are important for

    the implementation of the GECs, as well as in mitigating processes such as desertification and

    biodiversity loss.

    Based on the above discussion, it is evident that processes of environmental change involve

    more than ecological changes and are influenced by social factors, including decision making by

    local land users (Blaikie and Brookfield, 1987; Oba et al., 2008). In the light of the recognition of

    the close interaction between social and ecological systems, there is a need to adopt a more holistic

    22

  • approach that goes beyond the disciplinary divides, and in addition co-opts local knowledge and

    practices using assessment and monitoring of environmental changes at local levels. To achieve this

    objective, indigenous knowledge and conventional scientific methods should not be portrayed as

    competing binary opposites, but rather, as complementary sets of research and management tools

    for addressing environmental problems (Agrawal, 1995; Nygren, 1999).

    Advances in interdisciplinary methods have transformed environmental research from purely

    ecological perspectives to incorporate wider social, political and economic aspects (Blaikie and

    Brookfield, 1987). This knowledge interplay underscores the role of local communities in studies on

    environmental change. A conceptual socio-economic and ecological model (SEEM) suggests the

    practicalities of such an approach (Oba et al., 2008). The SEEM model proposes that analysis of

    environmental change should consider local socio-economic and ecological drivers that may inform

    decision makers on the relationship between land use and environmental change. Land use decisions

    are influenced by socio-economic as well as ecological drivers. In the SEEM model, appropriate

    anthropogenic and ecological indicators are used to describe environmental changes. Details on the

    process of indicator selection and decision-making have been described in SEEM model by Oba et

    al. (2008). This thesis utilizes the SEEM approach for integrating indigenous and ecological

    methods by conducting environmental assessment and monitoring jointly with herders in the arid

    region of northern Kenya. The integration process demands the development of efficient methods to

    accommodate indigenous knowledge variables and ecological methods for environmental

    assessment and monitoring.

    4. Methodological perspectives

    The integration of indigenous knowledge in research and development faces some methodological

    challenges (Scoones and Thompson, 1993). An important methodological issue is the selection of

    knowledgeable members of the community. It has been argued that indigenous knowledge is not

    homogenous among the local people due to stratification into gender, age, social classes and other

    disparities in power relations (Sillitoe, 2002). Therefore, researchers are obliged to find a balance

    between participation by highly heterogeneous groups, while at the same time maintaining rigor

    demanded by standard scientific methods, such as random selection of informants. Another problem

    of integration arises because of epistemological differences between indigenous and standard

    scientific methods (Purcell and Onjoro, 2002). In scientific research, the focus is on cause and effect

    23

  • relationships and developing predictive models about future outcomes in relation to propositions

    made by different scientific theories (Freeman, 1992). The scientific method is deductive as well as

    reductive and may be deterministic in predicting processes of environmental change. By

    comparison, by its very nature, indigenous knowledge is rooted in local contexts, practices and

    beliefs (Woodley, 2005). It is better understood using cognitive anthropology, focusing on the

    relationships between human culture and human thoughts and beliefs. Unlike the traditional

    ethnographic approach, where an outsider uses his/her worldviews to study certain cultures,

    cognitive anthropology stresses how people make sense of reality according to their own indigenous

    cognitive categories. The cognitive anthropology approach aims at understanding the local peoples

    worldviews using the emic perspective, as opposed to traditional ethnographic outsiders etic views.

    The cognitive anthropological approach of eliciting local knowledge uses formal

    questionnaires. However, the questionnaires are developed with literate informants in mind and are

    often inappropriate in rural settings where research is conducted with local people (Antweiler,

    2004). Most researchers using questionnaires are not comfortable with the responses elicited,

    because they are void of capturing real experiences, and there is the added disadvantage of the time

    it takes (Kumar, 2002). Although questionnaires are prepared based on presumed problems and sets

    of objectives that may not address the goals of local people, they may be useful for soliciting

    generalized information, to be followed up by more participatory approaches. The time drawback

    has led to the adoption of more pragmatic, rapid participatory methods. Quick and popular methods

    such as Participatory Rural Appraisal (PRA) and Rapid Rural Appraisal (RRA) have been used to

    bring together local people and researchers/development agents to address local issues (Chambers,

    1999). These participatory approaches use visual learning such as drawings and maps so that non-

    literate and less articulate members of the local community can participate in exercise-oriented

    discussions (Kumar, 2002). RRA focuses on the use of observation and verbal interactions including

    semi-structured interviews, transect walks, and group discussions (Chambers, 1999). By

    comparison, PRA stresses the use of shared visual representation by local people. RRA and PRA

    methods allow local people to share knowledge and practice with external development and

    research agents to implement locally relevant development objectives (Chambers, 1994). Rapid

    participatory methods are, however, disadvantageous because of the lack of empirical and

    theoretical grounding (Antweiler, 2004).

    24

  • The data collected through rapid participatory methods represent mostly the local contexts in

    which they are generated, with limited wider application. The most appropriate forum for using

    participatory methods is organized workshops involving both local and technical groups, for one-to-

    one discussions concerning pertinent environmental or social issues. For example, in herding

    communities, information on the local environment is held in both the public domain and also by

    some key knowledgeable members of the society, including the elders and herder scouts. Due to

    their long-term experience, the elders have more detailed information on historical environmental

    changes including land use patterns, levels of livestock productivity and the extent and composition

    of vegetation in the grazing rangelands. The elders knowledge is continuously updated during daily

    assessment and monitoring of the status of grazing resources. Individuals rely on accumulated

    knowledge within the community, suggesting that there is uniformity of information, regardless of

    whether individuals or groups were involved in environmental assessments. The indigenous

    knowledge has two sides the knowledge related to the production systems, such as livestock, that

    is often used as a barometer of environmental change, and the knowledge of the environment itself

    in terms of grazing landscapes, soils, vegetation types and the history of land use. Thus, the herders

    have accumulated knowledge of the interaction of livestock production systems with biophysical

    resources. This knowledge is put into everyday practical use to regulate livestock-environment

    interactions.

    When selecting research methods, it is therefore important that collaborating researchers

    adequately understand the knowledge and practice aspects of indigenous knowledge systems. The

    basic resource management units are landscapes. Assessment and monitoring of environmental

    change is based on the knowledge of a landscape that has been accumulated over the long term and

    enriched by continuous observations using different environmental indicators. For the herders, the

    process has utilitarian value, which is aimed at maximizing livestock productivity through

    appropriate and timely management decisions (Oba et al., 2008). The knowledge of the local

    environment is therefore part of daily practice embedded in the local cultures and influenced by the

    requirements of the production systems.

    The scale of assessment and monitoring and the indicators used all have relevance to local

    production systems. Previously, different approaches have been used to elicit participation,

    including questionnaires, semi-structured interviews, workshops and collaborative fieldwork

    (Dougill et al., 2006; Huntington, 2000; Stringer et al., 2006). These methods have different levels

    25

  • of flexibility in accessing indigenous knowledge in different local contexts. Structured

    questionnaires are useful when quantification is desired and the interviewer has a clear

    understanding of the information required. For example, in pastoral environments, data on the

    number of livestock managed by different households in different management systems (e.g. fora

    and home herds) can be addressed adequately using a structured questionnaire. Semi-structured

    interviews are less restrictive and may use sketch guidelines, while the participants and the

    interviewer engage in open-ended discussions. Semi-structured discussions with individuals or

    groups help the interviewer to strike a rapport with principal participants in more informal settings,

    thereby allowing a free exchange of information. Similarly, workshops on specific environmental

    issues involving researchers and participants can result in fruitful discussions. In workshops,

    different issues raised earlier during interviews could be deliberated further to generate new insights

    on the topics under discussion. Interview methods are important for eliciting knowledge engrained

    in local land use practices. The approach requires developing personal relationships to promote

    dialogue and the exchange of ideas as part of participatory research.

    Scientific methods (hereafter referred to as ecological methods) are standardized and

    strongly empirical. In vegetation assessments, ecological methods include the use of transects to

    measure the density and composition of vegetation occurring along a line or belt transect, or a road

    may be used as a transect. Additionally quadrants or plots are used to sample vegetation in

    circumscribed areas (Sorrells and Glenn, 1991). There are, however, some important requirements

    in ecological sampling, including having representative samples by randomizing sampling units,

    including transects and quadrants/plots. The sampling units must also be of the right size in order to

    capture the diversity of plant life forms such as herbs, shrubs and trees. The selection of sampling

    strategies needs to consider variability along ecological or anthropogenic gradients. Another factor

    central to ecological monitoring is the use of the right scales for measuring the vegetation variables

    from landscapes to larger regional levels. Ecological methods linked to different scales of vegetation

    change are useful for testing hypotheses. Sampling designs must incorporate spatial and temporal

    factors that may influence vegetation changes. Spatial limitations of quadrant and transect methods

    in the assessment and monitoring of vegetation have been solved using remote sensing technology

    which has a wider spatial coverage. Satellite imagery can be used to compare vegetation covers at

    different times in order to analyze the temporal dynamics.

    26

  • The integration of indigenous knowledge and ecological methods for the implementation of

    GECs should harmonize the various methodological orientations. Before discussing the actual

    integration of the two methods for addressing local environmental problems, it is imperative to

    situate participatory actions at the local level within the context of the broader plans for the

    implementations of GECs.

    5. Framework for integrating local knowledge in the implementation of GECs

    The implementation of GECs is linked to the activities at global, national and local levels. At the

    global level, conventions are negotiated as part of collective global responsibilities as mentioned

    earlier. At the national level, each country implements the conventions according to the articles, by

    setting priorities according to national goals and implementation strategies. Implementation at the

    local level necessitates consideration of the diversity of ecological, production and social-cultural

    systems, and the use of local knowledge for resource assessment and monitoring. Local

    participation as recommended at the global level, and implementation plans through national and

    local levels are described using a conceptual framework (Fig. 3). The framework has three

    interrelated components (hereafter referred to as stages). In Stage I, global environmental

    problems such as desertification, land degradation and biodiversity loss are described. Stage II of

    the conceptual framework is concerned with the implementations of NAPs (Fig. 3).

    The activities described in Stages I and II are beyond the scope of this thesis. The thesis

    addresses the activities described under Stage III, which forms the core of local participation in the

    implementation of GECs (Fig. 3). In this thesis it is assumed that national action programs (NAPs)

    may be concerned with reversing land degradation and desertification, while at the same time

    conserving biodiversity using both conventional methods and local knowledge (Stage III). This may

    be done by working through technical departments that are represented at the local level. Thus t