Batlle-Aguilar Etal2011_Modelling Soil Carbon and Nitrogen Cycles During Land Use Change. a Review

8/6/2019 Batlle-Aguilar Etal2011_Modelling Soil Carbon and Nitrogen Cycles During Land Use Change. a Review

1/24

c INRA, EDP Sciences, 2010DOI:10.1051 / agro / 2010007

Review article

Modelling soil carbon and nitrogen cycles during land use change.A review

J. Batlle -Aguilar 1*, A. Brovelli 1, A. Porporato 1,2, D.A. Barry 1

1 Institute for Environmental Engineering, cole Polytechnique Fdrale de Lausanne, Station 2, 1015, Lausanne, Switzerland2 Permanent address: Civil and Environmental Engineering Department, Duke University, Durham, NC 27708, USA

(Accepted 9 February 2010)

Abstract Forested soils are being increasingly transformed to agricultural elds in response to growing demands for food crop. ication of the land use is known to result in deterioration of soil properties, in particular its fertility. To reduce the impact of activities and mitigate their eff ects on the soil, it is important to understand the factors responsible for the modication of soil propethis paper we reviewed the principal processes aff ecting soil quality during land use changes, focusing in particular on the eff ect of soil mois-ture dynamics on soil carbon (C) and nitrogen (N) cycles. Both physical and biological processes, including degradation of litter and soil moisture evolution at the diurnal and seasonal time scales were considered, highlighting the impact of hydroclimatic vanutrient turnover along with the consequences of land use changes from forest to agricultural soil and vice-versa. In order to identextent diff erent models are suitable for long-term predictions of soil turnover, and to understand whether some simulators are morspecic environmental conditions or ecosystems, we enumerated the principal features of the most popular existing models dealingN turnover. Among these models, we considered in detail a mechanistic compartment-based model. To show the capabilities of theto demonstrate how it can be used as a predictive tool to forecast the eff ects of land use changes on C and N dynamics, four diff erent scenarioswere studied, intertwining two diff erent climate conditions (with and without seasonality) with two contrasting soils having physical pr

that are representative of forest and agricultural soils. The model incorporates synthetic time series of stochastic precipitation, ansoil moisture evolution through time. Our main ndings in simulating these scenarios are that (1) forest soils have higher concentand N than agricultural soils as a result of higher litter decomposition; (2) high frequency changes in water saturations under seasoscenarios are commensurate with C and N concentrations in agricultural soils; and (3) due to their diff erent physical properties, forest soilsattenuate the seasonal climate-induced frequency changes in water saturation, with accompanying changes in C and N concentrmodel was shown to be capable of simulating the long term eff ects of modied physical properties of agricultural soils, being thus a promtool to predict future consequences of practices aff ecting sustainable agriculture, such as tillage (leading to erosion), ploughing, harveirrigation and fertilization, leading to C and N turnover changes and in consequence, in terms of agriculture production.

soil organic matter / biogeochemical cycles / agricultural soil / forest soil / soil nutrients / soil moisture dynamics / soil restoration

Contents

1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.1 Land use change: forest versus agricultural soils. . . . . . . . . . . . . . . . 21.2 Modelling of soil C and N cycles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 C and N cycles in soil. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Moisture dynamics as a controlling factor of soil carbon and nitrogen

cycles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Existing models of soil carbon and nitrogen turnover. . . . . . . . . . . . . . . . . 65 Description of the C-N model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Modelling scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

6.1 Climate scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116.2 Soil scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2

7 Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 48 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 9

* Corresponding author: [email protected]

Agron. Sustain. Dev. (2011) 31: 251 274

(Published online: 2 Ju 2010)ne
http://dx.doi.org/10.1051/agro/2010007http://dx.doi.org/10.1051/agro/2010007http://dx.doi.org/10.1051/agro/2010007http://dx.doi.org/10.1051/agro/2010007http://dx.doi.org/10.1051/agro/2010007http://www.edpsciences.org/http://www.agronomy-journal.org/http://dx.doi.org/10.1051/agro/2010007


2/24

1. INTRODUCTION

Soils are complex systems sustaining life on Earth. Amongother functions, soils maintain plant and animal growth, recy-

cle nutrients and organic wastes, lter and purify water. Pre-cisely, soil quality refers to a combination of chemical, physi-cal, and biological processes that confers to the soil the abilityto carry out, among others, these particular ecological func-tions. Numerous human activities however utilise soil, mod-ify its physical and chemical properties and change the com-position of its ecosystems. As a result, in the last century awidespread decrease of soil quality has been observed, to-gether with a deterioration of its functioning (Brady and Weil,2004).

The main component of soils is organic matter (SOM),which shows a variable degree of decomposition, from freshlitter to highly decomposed humus. SOM stores three to

four times the amount of carbon (C) than found in all liv-ing vegetation. Other than C, soils also contain nearly allthe macro- (nitrogen, N, phosphorous, P, and potassium, K)and micro-nutrients required by living organisms. Among themacronutrients N plays a major role since it is essential for lifebut its bio-available forms are seldom abundant in the environ-ment. Therefore in many ecosystems the N cycle controls theoverall soil turnover and functioning. For these reasons andwithout neglecting the importance of other nutrients, in thispaper the focus is on soil C and N cycles.

1.1. Land use change: forest versus agricultural soils

The increasing demand on food crops, pasture, rewoodand timber is at the origin of worldwide changes of land-use in forested areas. This situation is worrying in some ar-eas of the planet, such as South America, where 12% and 7%of forestland was converted to pasture and croplands, respec-tively, between 1850 and 1985 (Houghton et al.,1991). Land-use changes, and especially cultivation of previously forestedland, reduce signicantly the soil quality (e.g., changes inSOM content and decomposition rates, changes in soil chem-ical and physical properties), leading to a permanent degra-dation of land productivity (Nye and Greenland,1964; Islamet al.,1999). Furthermore, it has been reported that deforesta-tion increases carbon dioxide (CO2) release to the atmosphere(Houghton,2002), which contributes to global warming.All studies that focused on the eff ects of land conversionfrom forest to cultivated land concluded that land-use changeinduces a reduction of the available soil C and a decrease in itsquality. The maximum rate of loss occurs during the rst 10 yof cultivation, with total C decrease up to 30% (Davidson andAckerman,1993; Lugo and Brown,1993; Murty et al.,2002)followed by reduced but still signicant reduction rate (Brams,1971; Martins et al.,1991; Bonde et al.,1992; Motavalli et al.,2000). Furthermore, it was reported that the loss rate is highlyvariable and inuenced by several factors such as the na-tive vegetation, climate, soil type and management practices(Mann,1986; Davidson and Ackerman,1993; Bruce et al.,

1999).

Contrasting with the conversion from forest to cultivateland, controversy exists when the change is from forest to pature lands. The overall change in soil C has been shown tbe either positive or negative. For instance, de Moraes et a

(1996) found an increase up to 20% in total soil C 20 y aftethe change in land use, while Veldkamp (1994) reported a netsoil organic C loss up to 18% after 25 y. Johnson(1992) alsoobserved that changes in soil C in both land-use cultivatioand pasture were associated with changes in soil N. Reiners al. (1994) found that the transformation of forest land to pasture led to important changes in the N cycling. For examplthe ammonium (NH+4) pool was larger in pasture lands whilethe nitrate (NO3) pool was less important in pasture than for-est lands. This is consistent with a low rate of plant uptakof NH+4 and slow nitrication rates (Vitousek,1984; Vitousekand Sanford,1986).

One of the important aspects that aff ect SOM cycling in

the transition from forest to cultivated soil is the removal most of the fresh organic C (litter) due to harvesting (Smi1999). However, harvesting is not the only factor responsiblfor the soil organic C loss. Some other processes that weralso recognized to contribute to change the amount of soC are the changes in litter chemical properties (Feigl et al1995; Ellert and Gregorich,1996; Scholes et al.,1997), soiltype (Feller and Beare,1997; Scholes et al.,1997; Garca-Oliva et al.,1999), microbial community (Prasad et al.,1995),changes in soil N cycling (DalalandMayeer,1986; Brown andLugo,1990; Desjardins et al.,1994) and management prac-tices (Feller and Beare,1997; Fernandes et al.,1997; Bruceet al.,1999). Soil tillage and ploughing promote redistribu-tion of residues and their decomposition. As a result, soil and N pools are depleted and soil fertility is lost. Soil C oxidized to CO2 and lost to the atmosphere contributing tothe increase of greenhouse gases in the atmosphere.Moreovetillage improves soil aeration, destroys macro-aggregates anchanges the hydrological cycle, with an increase of the respiration rates and ultimately an additional depletion of the pool (Juo and Lal,1979; Agboola,1981; Ellert and Gregorich,1996; Reicosky et al.,1997; Bruce et al.,1999).

In agricultural areas, the root zone (soil depth aff ected byplant roots) remains constant over time and is relatively shalow. Diff erent rootingpatterns have directeff ects on the C ux,since they aff ect soil porosity and soil aeration (Berger et al.,2002). Therefore, changes in land use resulting in a modierooting depth often have a direct inuence on soil respiratioand C mineralization rates, and thus on soil turnover (Howaand Howard,1993).

Ecological restoration is the process of assisting the recovery of an ecosystem that has been degraded, damaged or dstroyed as a consequence of human activities (Young et al2005), and typically involves a land use change. During threstoration, environmental conditions (e.g., type of vegetatioecosystem corridors or soil practices) are manipulated to crate ecological conditions suitable for the successful establishment of a target composition of species (Prober et al.,2005).The change from agricultural soil to the original forest istypical example of soil restoration, where natural soil prope

ties and vegetation are amended, resulting in an improveme

252 J. Batlle-Aguilar et al.


3/24

Modelling soil carbon and nitrogen cycles during land use change. A review

Figure 1. Simplied representation of the global carbon cycle.

of soil fertility and an ecosystem close to its former naturalcondition.

1.2. Modelling of soil C and N cycles

Numerical tools are becoming increasingly used to under-stand the modications induced in ecosystems as a result of changes in land use, and it has been found that understandingthe coupled N and C dynamics is of primary importance forpredictive models of SOM evolution, for example to changesin land use and responses to global changes (Rodriguez-Iturbeet al.,2001). Modellingof soil biogeochemical processesdatesback to the 1930s (Manzoni and Porporato,2009), and nowa-days an extended list of stochastic, empirical and mechanisticmodelsincorporatingsoil nutrient dynamics is available. Mod-els vary signicantly in terms of complexityand mathematicaldescription of the biological and geochemical processes in-volved. Manzoni andPorporato (2009) reviewedandclassiedabout 250 diff erent mathematical models developed over 80 y.Most of the models currently available evolved from early ef-forts to provide a concise mathematical description of the soilcycles, and have been adapted and improved to specic appli-cations. The aims of the diff erent models are numerous andinclude, for example: understanding and prediction of feed-backs between terrestrial ecosystems and global climate (e.g.,estimate and predict climatological and biological eff ects of human activities) (Agren et al.,1991; Melillo,1996; Mooreet al.,2005); inuence of climate changes on nutrient cyclingin soils (Pastor and Post,1986; Hunt et al.,1991; Moorheadet al.,1999; Eckersten et al.,2001; Ito,2007); prediction of changes in soil C and N cycles related to possible land usechanges (Eckersten and Beier,1998; Paul and Polglase,2004;

Christiansen et al.,2006; Findeling et al.,2007; Pansu et al.,

2007; Post et al.,2007; Kaonga and Coleman,2008); and fore-casts of crop productivity and system response under specphysical soil changes (Wolf et al.,1989; Wolf and Van Keulen,1989; Matus and Rodrguez,1994; Parton and Rassmussen,1994; Henriksen and Breland,1999; Nicolardot et al.,2001).

The aim of this manuscript is to provide an overview the main processes, mechanisms and parameters aff ecting theevolution of selected soil nutrient cycles (soil C and N) ato provide a modelling framework that incorporates the kmechanisms. Both physical and biological processes, incluing degradation of litter and humus, and soil moisture evlution on diurnal and seasonal time scales are considered.the rst part of the manuscript, soil C and N cycles are sumarized, followed by an overview of the most popular mels dealing with soil nutrient turnover. In the second partcompartment model based on Porporato et al. (2003) is de-scribed and applied to simulate soil C and N dynamics,well as degradation and transformation processes occurriunder diff erent precipitation and soil scenarios. Contrastinsoil types and precipitation regimes are considered, to ill

trate modelling capabilities and to show how numerical tocan be used to understand eff ects of land use changes oversoil C and N uxes and, thus, the feasibility and viabilityecological restoration regarding the modelled ecosystem asurroundings.

2. C AND N CYCLES IN SOIL

The global C cycle can be depicted as consisting of a ries of interconnected compartments (terrestrial, aquatic aatmospheric) where C is stored and transformed. Soils are pof the terrestrial C pool (Fig.1). The amount of C stored in

the (living and dead) organic matter in soils is three to fo

25


4/24


times higher than that in the atmosphere (Bruce et al.,1999).The circulation rates are also high. For these reasons, soil Cturnover is of primary importance to developing understand-ing and forecasting global changes in biogeochemical cycles

and climate change (Stevenson and Cole,1999; Rodriguez-Iturbe and Porporato,2004). The total global emission of CO2from soils is probably the largest ux in the global C cy-cle, and small changes in the magnitude of soil respiration, if they take place at large scale, could have a tremendous eff ecton the concentration of CO2 in the atmosphere (Schlesingerand Andrews,2000; Murty et al.,2002). By the same argu-ment, soils have also a great potential for long-term C storage.Whether a soil will act as a sink or source of CO2 depends on anumber of environmental factors, including climatic variabil-ity and anthropogenic changes in land use, which for examplemay result in a modied composition of the vegetation andtherefore of the quality and quantity of litter inputs (Gignoux

et al.,2001).The principal C exchange processes between soil and at-mosphere arephotosynthesis and respiration . PhotosyntheticC xation by plants often named primary producers con-verts atmospheric CO2 and is the main source of soil organicC. Briey, during photosynthesis CO2 is used as a C sourceto produce complex organic molecules, using sunlight as anenergy source (e.g., Killham and Foster,1994):

CO2 + H2O+ Energy CH2O + O2. (1)The complex organic molecules produced by plants enter thesoil C cycleas decayingorganic matter (litter) andareprogres-sively convertedto simpler molecules.A signicant fraction of the organic C introduced in the soil is directly used as an en-ergy source to sustain pedofauna metabolism, and is releasedagain to the atmosphere in form of CO2 through respiration:

CH2O + O2 CO2 + H2O + Energy. (2)Another part of the soil C is assimilated by vegetation and -nally transferred to the soil as plant litter, becoming part of SOM (Porporatoet al.,2003). Organic C is available in soils ina large variety of forms. Killham and Foster(1994) partitionedthe soil organic C into three main pools: insoluble, soluble andbiomass. Insoluble soil organic C includes plant residues andpartially decomposed material, which forms the litter and thehumus. Soluble C is a fraction of the humus further decom-posed and is rapidly assimilated as a substrate by the pedo-fauna. The fast consumption of soluble C explains its oftenlow concentration in the soil (1%) in comparison to insolubleorganic C (90%). Soil biomass (9%) consists of microbes andanimals (e.g., macroinvertebrates), the decomposition activityof which is mostly responsible for the C decomposition andrecycling (Killham and Foster,1994).

Within the soil, organic C is transferred between the dif-ferent pools (or compartments) by means of decompositionprocesses, which are regulated by environmental conditions(e.g., soil moisture) and the C / N ratio (Brady and Weil,2004).These factors will be discussed subsequently. Litter under-going decomposition is mainly composed of plant residues(fallen leaves, roots, etc.).Decompositionratesarehighlyvari-

able in time, and are mainly controlled by the environmental

Figure 2. Inuence of the soil water content on bacterial activity indiff erent processes of nitrogen transformations (after Fenchel et al1998).

conditions (e.g., soil moisture level, aeration, soil temperatureand the quality of the added litter. Complex organic moleculcan be decomposed under either aerobic or anaerobic condtions. Under normal conditions, soils are unsaturated and thuO2 is likely to be always available. However, even in the vadose zone saturated conditions can result from, for examplsignicant precipitation events. Wetlands are a particular caswhere saturated conditions are found permanently or seasonally. In general, microbialdecompositionrates are largerundeaerobic conditions (Brady and Weil,2004), where O2 acts asthe electron acceptor during oxidation of organic compound(Barry et al.,2002). On the other hand, slow decompositionrates under anaerobic conditions can result in accumulation considerable amounts of partially decomposed organic matt(Fig.2).

Soil N comes mainly from the atmosphere, which is thlargest N pool and contains almost 75% of the total N avaiable on Earth (Barbour et al.,1999). In brief, the soil N cycle isbased upon the uptake of the inorganic forms (NO3 , NH+4 ) byplants. N returns to the soil in organic form as plant residuewhich are decomposed by the soil pedofauna (e.g., invertebrates, microbes, fungi) and are made available to plants iinorganic form.

The total amount of organic N in soils varies greatly anis inuenced by the soil-forming factors likely climate, topography, vegetation, parent material and age. The N cyclis tightly coupled to the C cycle, since most of the microbiN transformations (e.g., nitrication) use energysupplied by C(Paul,1976). Although locally N is also incorporatedinto soilsthroughdry or wet directdeposition, the largest fractionof soorganic N xation is done biologically (conversion from N2gas to organic forms, mediated by specic microbial strainsN is found in soils mainly within the organic matter fractiofor example humiccompounds,plant roots, microbial biomas

and decomposing organic materials. The amount of organic

25


5/24


contained in soils far exceeds that which is present in plant-available inorganic forms.

The soil N, mainly present in organic form as previouslymentioned, is almost unavailable for plants. The vegetation

mainly uses inorganic forms of N, which are made availableby the SOM decomposition. Soil microorganisms convert theN contained in the organic matter to NH+4 in a process namedmineralization (Schinner et al.,1995), further subdivided intotwo processes. The organic N is initially transformed viaam-monication , and if O2 is available NH+4 is subsequentlyoxidized to nitrite (NO2) and NO3 , throughnitrication :

NH+4 + O2 + H+ + 2e NH2OH+ H2O NO2 + 5H+ + 4e , (3)

NO2 + H2O NO3 + 2H+ + 2e . (4)Although plants can use both forms of inorganic N, NO3 is

used in preference to NH+

4 because of its greater solubility inwater. In other words, nitrates quickly dissolve in the pore so-lution, which is taken up by plants. On the other hand, how-ever,thisalsomeans thatNO3 is easilyushed to groundwater.NH+4 is instead less mobile because it is strongly adsorbed onclay minerals due to its positive charge.

Denitrication is the anaerobic microbial reduction of N,and NO3 is used as an electron acceptor (i.e., source of en-ergy), resulting in a transfer of soil N to the atmosphere(Groff man et al.,2002):

2NO3 + 10e + 12H+ N2(g)+ 6H2O. (5) Immobilisation is a process involving microbial uptake of nu-trients, where inorganicN is converted into organic form, suchas amino acids and biological macro-molecules.

The carbon-to-nitrogen ratio (C / N) is an important factoraff ecting the overall turnover rates of SOM (Young andYoung,2001). Bacterial sensitivity to the C / N ratio is due to the factthat bacteria need a constant C / N ratio, while this ratio ishighly variable in substrate. For example, intense competi-tion among microorganisms for available N occurs when soilresidues have a high C / N ratio, i.e., the substrate is poor in Nmaking it the limiting factor (Brady and Weil,2004). Environ-mental conditions (e.g., soil moisture and temperature) havea direct inuence on bacterial activity and thus on this ratio(Koch et al.,2007). The C / N ratio of the substrate tends to de-crease as the SOM becomes more decomposed from freshlitter to highly transformed humus when microbes are solelyresponsible fordecomposition(Zhenget al.,1999) because themicrobial C / N ratio is lower than that of litter (Persson,1983).In other words, the humus is enriched in N compared to thelitter. For this reason the C / N ratio of the litter pool controlsthe rates of mineralization / immobilisation. Young and Young(2001) identied a thresholdof the C / N ratio which determinesthe bacterial activity. When C / N > 25, microbes respire com-pletely using the available C and thus assimilate the entire Nmineralized,andconsequentlyN is immobilised. In contrast, if C / N < 25, the SOM N content far exceeds the immobilisationcapacity of microbial populations and the result is a net miner-alization. Although this threshold seems to be directly related

to the C / N ratio needs of bacteria, White (1997) argued that

this threshold value is variable among diff erent ecosystems,for example because the C / N ratio of the vegetation changesdepending on the composition and relative frequency of easpecies. For example, pines produce litter with C / N ratio as

high as 90, while litter originating from cereal crops has C / Nratio of 80 and tropical forest trees produce litter with C / Nratios around 30 (Young and Young,2001).

3. MOISTURE DYNAMICS AS A CONTROLLINGFACTOR OF SOIL CARBON AND NITROGENCYCLES

Soil moisture results from the interactions between climasoil type (texture, granulometry, organic matter content) avegetation, and it is consequently variable both in space atime. Among the possible physical processes the dynamics

soil moisture exerts the greatest inuence over SOM turnovmineralization,decomposition, leaching and uptake, and its fects are complex and non-linear. As an example to illustrthis complexity, the production of plant residues the masource of litter and therefore of energy for the pedofaundepends on the growth rate of vegetation, which is controlby water availability. Accumulation of SOM can increase water retention capacity of the soil, with a positive feed-bon the vegetation. Moreover, the soil biota activity depenon the soil water content, and optimal decomposition rates only achieved within a relatively narrow soil moisture rang

Soil biota is sensitive to moisture level for several reasoIn order to preserve cell integrity, when the soil water cotent decreases bacteria increase the intracellular solute cocentration to compensate for the extracellular concentratiand counterbalance the increased osmotic pressure (Stark aFirestone,1995; Bell et al.,2008). Therefore, a high concen-tration of solutes results in an inhibition of the enzymatic tivity and therefore decreased cellular activity. Additionalas the soil becomes drier, water in soil pores becomes a thlayer covering soil grains and substrate availability becomdiff usion-limited. In consequence, microbial activity is furthreduced (Csonka,1989; Stark and Firestone,1995; Fenchelet al.,1998).

It is however difficult to identify a unique threshold moisture level under which soil respiration (or microbial actity) diminishes. Davidson et al.(1998) and Rey et al.(2002)estimated that 75% of the soil eld capacity correspondsthe soil moisture level below which soil respiration decreaswhile accordingto Xu et al. (2004) a more likely value is 42%.

A number of studies have shown that soil moisture eff ectson soil C and N turnover also depend on the time-scale of terest. Curiel Yuste et al. (2007) found that, at the seasonalscale, the eff ect of temperature and soil moisture on CO2 ef-ux (e.g., soil respiration) was very similar for ponderosa pand oak savannah ecosystems. For shorter time scales (edaily), decompositionof organicmatter was mainly controlby temperature during wet periods and a combination of teperature and soil moisture during dry periods. Soil bactergrowth (or soil respiration) a parameter often used as a m

sure of microbial activity shows a maximum at about 30

C

25


6/24


(Pietikinen et al.,2005). Nevertheless, the inuence of temperature on microbial activity is generally consideredmuch less important than soil moisture (Rodriguez-Iturbe andPorporato,2004) because, although important diff erences in

soil temperature are likely to occur at a daily and seasonalscale in the uppermost soil (e.g., rst few centimetres), yearlyaverage values at depth are much more constant than those of soil moisture.

Typically, summer drought decreases substantially decom-position rates (Curiel Yuste et al.,2007), but it has been ob-served that sporadic rains during these dry periods tends toincrease the decomposition efficiency of the bacterial commu-nities (Borken et al.,1999, 2002; Savage and Davidson,2001;Goulden et al.,2004; Xu et al.,2004; Misson et al.,2005;Scott-Denton et al.,2006). Kieft et al.(1998) and Moore et al.(2008) observed an increase of root density and soil micro-bial activity rate in response to isolated moisture pulses in arid

soils, although the response of root density occurred at longertime-scale. A fast rewetting of the soil prole is likely to havenegative consequences on microbial populations in that it cangenerate an osmotic shock and result in cell lyses (Kieft et al.,1987; Van Gestel et al.,1993). In contrast, Ryel et al.(2004),Schwinning and Sala (2004) and Bell et al. (2008) found that,in arid soils, while plantsusually do nottakeadvantage of brief pulses of moisture generated by short precipitation events,microbial mineralization is stimulated. Consequently, short-term increases in soil microbial activity triggered by mois-ture pulses will not typically correlate with an increase in pri-mary production at the same time scale, conrming that plantgrowth is not only dependent on soil microbial activity, butalso on other factors such as the precipitation event duration,amount of soil water inltrated and the overall change in soilmoisture. The magnitude and timing of intra-seasonal precipi-tation becomes therefore a key regulator for microbial activity(Bell et al.,2008). Since decomposition and consequent min-eralization can be stimulated by moisture pulses that are toobrief to benet primary producers (e.g., plants) (Cui and Cald-well,1997; Schwinning et al.,2003; Austin et al.,2004), inaridsoils there is potential for soil nutrientpools to accumulateover time and become available to plants as heavier precipita-tion occurs. The inuence of soil moisture over soil nutrientdynamics has been also studied in temperate (Davidson et al.,1998; Buchmann,2000; Reichstein et al.,2003) and tropicalforests (Conant et al.,2000; Davidson et al.,2000; Kiese andButterbach-Bahl,2002; Epron et al.,2004). It was concludedthat a strong inuence of the soil moisture over microbial ac-tivity exists, but that the degree of correlation varies stronglyamong diff erent ecosystems (Buchmann,2000; Rustad et al.,2000).

4. EXISTING MODELS OF SOIL CARBONAND NITROGEN TURNOVER

At least 250 models dealing with soil C and nutrientturnover exist (Manzoni and Porporato,2009). Classicationof all these simulators is difficult because they are based on

a wide range of physical and biogeochemical descriptions of

the processes and the underlying assumptions vary signicantly. Nevertheless, based on their internal structure modedescribing SOM dynamics can be divided into (1) processoriented, (multi)-compartment models; (2) organism-oriente

(food-web) models; (3) cohort models describing decomposition as a continuum; and (4) a combination of model typ(1) and (2) (Brussaard,1998; Smith et al.,1998; Post et al.,2007). Process-oriented or compartment models (each compartment or pool is a fraction of SOM with similar chemcal and physical characteristics) are built considering the prcesses involved in the migrationof SOM across the soil proand its transformations (Smith et al.,1998). Models belongingto this class can potentially have a variable degree of complexity, from the simplest case with no compartments (considering degradation as a continuum) to more rened, multcompartment models, with each compartment composed oorganic matter with similar chemical composition or degra

ability. Process-oriented models can be combined with GIsoftware, giving a modelling platform well suited for regionascale studies. Examples of successful coupling between soturnover and GIS software are CANDY (Franko,1996), CEN-TURY (Schimel et al.,1994) and RothC (Post et al.,1982;Jenkinson et al.,1991). On the other hand, the theoreticalcompartments that dene the structure of multi-compartmeprocess-oriented models are difficult to compare with the mea-surements of SOM fractions, and therefore the testing and vaidation is difficult and limited (Christensen,1996; Elliott et al.,1996). Among the most popular process-oriented models aralso DAISY (Hansen et al.,1991), NCSOIL (Molina et al.,1983) and SOILN (Johnsson et al.,1987).

In organism-based models the SOM ows from one organism pool to another, which in turn are classied dependinupon their taxonomy or metabolism. The main advantage organism-oriented models is that the main drivers of SOMuxes and transformations the pedofauna are explicitly acounted for. However, as noted in Post et al. (2007), to datethere is no general acceptance of the existence of a relatiobetween soil biota abundance and degradation rates. In cotrast, the relationship between degradation rate and amoun(or concentration) of substrate, as in process-based modelis well recognized. Simple rst-order kinetic rates are oftesuitable to model the transformations, and the reaction ratecan be easily estimated from laboratory experiments and drectly used in process-oriented models. Site-specic calibrtion of organism-oriented models involves the characterization of the soil microbial consortiaand thereforerequires morcomplex techniques, while process-oriented models are lesinuenced by the features of the microbial communities, anhave a larger range of application to diff erent environments.To summarize, process-oriented modelsareeasier to apply ancalibrate than organism-oriented, which explains their greatpopularity.Nevertheless, organism-oriented models have beeproposed by several authors, including Moore et al. (2004),Kuijper et al. (2005), Zelenev et al. (2006) and Cherif andLoreau (2009).

A cohort is a set of items sharing some particular characteristic. Cohort models divide SOM into cohorts, whic

are further divided into diff erent pools (e.g., C, N). Contrary

25


7/24


to process-based models where decay is usually treated as apurely physical or biochemical process, e.g., described by arst-order rate, cohort models consider explicitly microbialphysiology as the driving factor of decomposition. An exam-

ple of a cohort model was proposed by Furniss et al. (1982),where SOM was divided into three cohorts considering age,origin and size, with each cohort subdivided into a numberof chemical constituents. Gignoux et al. (2001) developedSOMCO (soil organic matter cohort), where SOM is dividedinto diff erent cohorts in a demographic sense, meaning that acohort is a set of items of the same age. At each time step anew cohort is dened and its fate is followed until its relativeamount to total SOM becomes negligible. Other examples of models belonging to this class are those of Pastor and Post(1986), Bosatta and gren (1991, 1994) and Frolking et al.(2001).

The last group of models consists of a combination of

process- and organism-oriented models, which are seldomused because their applicability is limited by the data requiredto dene the organism-oriented components (Smith et al.,1998). Some examples of combined models are proposed byOBrien(1984) and Paustian et al.(1990).

In order to identify to what extent diff erent models aresuitable for long-term predictions of soil turnover, and tounderstand whether some simulators are more suited tospecic environmental conditions or ecosystems, model com-parisons were conducted using long-term experiments andmulti-annual datasets. De Willigen(1991) tested 14 diff er-ent models comparing their ability to simulate soil N turnover(e.g., mineralization and plant uptake). It was concluded thataboveground processes (e.g.,plant growth) were easier to sim-ulate than belowground transformations (e.g., soil water andN content), and that the more complex, multi-compartmentmodels do not necessarily provide better results in terms of predictive capabilities. Rodrigo et al. (1997) compared the ef-fects of soil moisture and temperature variations on nine diff er-ent models (NCSOIL; SOILN; DAISY; Kersebaums model,Kersebaum and Richter,1991; MATHILD, Lafolie,1991;TRITSIM, Mirschel et al.,1991; NLEAP, Shaff er et al.,1991;SUNDIAL, Bradbury et al.,1993; CANTIS, Neel,1996) onpredictions of soil C and N turnover. Not surprisingly, theyobserved the highest C decomposition and N mineralisationrates close to eld capacity conditions and decreasing ratesduring soil drying. In this study good agreement between thediff erent models for low moisture conditions was observed,whereas poor agreement was found in wet soils, with wa-ter saturation equal or above eld capacity. A complete com-parison of nine process-oriented multi-compartment models(CANDY; NCSOIL; RothC; DAISY; CENTURY, Parton et al.,1987; Verberne model, Verberne et al.,1990; ITE, Thornley,1991; DNDC, Li et al.,1994; SOMM, Chertov and Komarov,1997) was presented by Smith et al. (1997). A qualitative andquantitative evaluation of the performance of the models wascarried out by comparing their ability to simulate observeddata from seven diff erent sites in temperate regions. A generalconclusionof all these comparisonswas that the errors derivedfrom the tested models are not signicantly diff erent, mean-

ing that the models provide consistent results except when a

model is used for an application for which it was not devoped. For example, the ITE and SOMM models were devoped for grasslands while in the study of Smith et al. (1997)they were applied to crops. Model calibration is an additio

source of uncertainties and makes the comparison of diff er-ent models difficult. Pansu et al. (2004) presented a qualitativecomparison of the predictive performance of a family of multi-compartment models, MOMOS-2 to -6, using14C- and15N-labelled species in eld experiments. These models uthe same conceptual approach but have diff erent complexity,in that the number of compartments varies from 3 to 5 and description of the biochemical transformation uses a diff erentlevel of detail and simplication. Pansu et al. (2004) concludedthat the simplications do not decrease signicantly model curacy, but that the use of additional compartments resultsimproved long-term predictions.

Most of the currently available models are updated versio

of earlier and original versions that have been modied to tend the applicability to specic ecosystems. TableI presentsa list of the ve most popular models, their main featurtogether with the key references. The popular CENTURmodel, originally devised for modellingsoil nutrient dynamin grassland systems, has been considerably modied sincerst version. Smith et al. (1997) and Parton and Rassmussen(1994) modied the CENTURY model for application to crand pasture systems, while Kelly et al. (1997), Peng et al.(1998) and Kirschbaum and Paul (2002) modied the modelto be applied to forest ecosystems. Despite the ad hoc mocations, contrasting results in terms of predictive capabiliwere obtained. The RothC model of Jenkinson et al.(1990), isan evolution of the model previously presented by Jenkinsand Rayner (1977), named Rothamsted. TOUGHREACT-N(Maggi et al.,2008) was developed to study the biogeochemical soil N cycle under diff erent conditions of fertilization andirrigation. It is based on the multi-phase, multi-component active transport model TOUGHREACT (Xu et al.,2006), inturn an evolved and improved version of TOUGH2 (Pruet al.,1999). SWIM (Krysanova et al.,1998), based on pre-viously developed tools (MATSALU, Krysanova et al.,1989;SWAT, Arnold et al.,1993) and originally devised for mod-elling soil N cycle in mesoscale watersheds (102 to 104 km2),has recently been extended to better describe groundwater namics and processes in riparian zones (Hattermann et a2004; Wattenbach et al.,2005). FullCAM (Richards,2001)accounts for full C turnover in forests, and is an integrasuite of sub-models: the empirical C tracking model CAMFor (Richards and Evans,2000), the tree growth model 3PG(Landsberg and Waring,1997), the litter decomposition modelGENDEC (Moorhead and Reynolds,1991) and the soil Cturnover model RothC (Jenkinson,1990). PASTIS (Lafolie,1991; Garnier et al.,2001, 2003) is a one-dimensional mecha-nistic model that simulates the transport of water, solutes aheat using Richards equation for water ow, the advectiodispersion equation for solute transportand the diff usion equa-tion for heat ow. Some variations to this model have beimplemented, such as PASTISmulch (Findeling et al.,2007),which extends the original capabilities by including the phy

cal eff ects of a surface residue mulch on rain interception a

25


8/24

8

Table I. Main characteristics of the ve most popular models.

Model Type Characteristics References to updated versionsCENTURYParton et al.(1987)

Process-orientedmulti-compartment

Long-term (decades to centuries) of SOM dynamics,plant growth and cycling of N, P and S, using monthlyinput data;

Three organic compartments: active, slow and passive; Soil texture regulates the C transfer from one compart-

ment to another.

Parton and Rassmussen(1994)Fernandes et al. (1997)Kelly et al.(1997)Smith et al.(1997)Bolker et al. (1998)Peng et al. (1998)Kirschbaum and Paul(2002)Easter et al.(2007)

DAISYHansen et al. (1991)


Simulates crop production, soil water and N dynamics; Four diff erent modules: hydrological, soil temperature,

soil nitrogen, and crop models; Decomposition rate constants depend on clay content; Vertical soil prole divided in layers considering phys-

ical and chemical characteristics.

Jensen et al.(1997)Smith et al.(1997)

NCSOILMolina et al. (1983)


Simulates C and N uxes through soil microbes andorganic components;

Four organic compartments: plant residues, microbialbiomass, humus and stable organic matter.

Smith et al.(1997)Hadas et al.(2004)

CANDY(CArbon and NitrogenDYnamics)Franko et al.(1995)


Composed of four diff erent submodels: soil temper-ature, hydrological, crop and organic matter turnovermodels;

Daily time step; Simulates soil N dynamics, temperature and moisture.

Franko et al. (1996)Franko et al. (1997)Smith et al.(1997)

RothC(Rothamsted model)Jenkinson and Rayner (1977)


N and C are not interconnected; C turnover sensitive to soil type, temperature, moisture

and plant cover; Monthly input data.

Smith et al.(1997)Easter et al.(2007)Kaonga and Coleman (2008)Jenkinson and Coleman(2008)

evaporation.Another example of model evolution is the familyof models MOMOS-2 to -6, which are modied versions fromthe initial MOMOS-C (Sallih and Pansu,1993) and MOMOS-N (Pansu et al.,1998) models. TRIPLEX (Peng et al.,2002) isa model of forest growth and C and N dynamics, and is a com-bination of three prior well-established models: 3PG (Lands-berg and Waring,1997), TREEDYN3.0 (Bossel,1996) andCENTURY4.0 (Parton et al.,1993). Easter et al.(2007) de-veloped a soil C modelling system, GEFSOC, aimed at mod-elling soil C stocks and exchange rates at regional or countryscales in response to land use changes. The developed tool isbased on three well-recognized models: the CENTURY gen-eral ecosystem model, the RothC soil C decomposition model

and the empirical IPCC method (IPCC,2003) for assessingsoil C stock changes at regional scales. The model can be cou-pled to a soil and terrain digital database to include the topog-raphy and spatial soil variability of the studied area.

5. DESCRIPTION OF THE C-N MODEL

The model we describe and use in this work was pre-sented by Porporato et al.(2003). It belongs to the group of process-based models, with the soil organic matter and nutri-ents divided into ve pools.Three pools consist of SOM (litter,

humusand microbial biomass), while the remainder are for in-

organic N. The model is applied to the root zone treated assingle unit, i.e., spatial variations are ignored.The framework with three organic pools is in good agree

ment with Jenkinson(1990), who proposed that process-basedmodels should have between two and four pools to obtain rliable results. These pools represent the main components othe system, and C and N concentrations correspond to aveage values over the rooting depth ( Z r ) (Rodriguez-Iturbe andPorporato,2004). This simplication is justied because soilshave often a uniform distribution of SOM and inorganic Nover the whole rooting depth (Porporato et al.,2003). This isnot true however for the uppermost soil layer, where organresidues tend to accumulate, and acts as a sourceof litter to thlayers beneath.Additionally, some other simplications were made during the development. First, SOM decomposition rates arknown to vary over orders of magnitude among the diff erentcomponents and, as already described, each functional grouof organisms has specic and highly variable decompositioand mineralization rates. In the model however no distinction is made between diff erent microbial populations. Rather,for each pool, a single, rst-order kinetic rate is used, whicrepresents an average transformation rate. This approach, athough approximate, reduces the number of model parameteand therefore simplies its calibration. Decomposition ratevary however among the diff erent pools: litter has faster de-

composition than the humus pool. The second approximatio



9/24


concerns the C / N ratio. As for the transformation rates, themodel considers a single C / N ratio for each pool, again rep-resenting an average value. In this case, the litter C / N ratiocan be computed, for example, as the weighted average of the

C / N ratios of the diff erent species, weighted by their relativeamount in the ecosystem. Other than this, vegetation charac-teristics (maximum evapotranspiration, wilting point, incipi-ent stress point, etc.) are assumed constant. This is an impor-tant simplication, since in previous sections it was pointedout that climatic conditions inuence vegetation growth anddeposition of fresh organic matter. The advantage is that wereduce and simplify the external factors inuencing C and Nturnover to soil type and moisture content dynamics.

Model inputs are precipitation and litter fall rates, while onoutput the extent of soil respiration, plant uptake, transpira-tion and leaching are recovered. The amount and frequencyof precipitation are the only climatic variables considered.Isothermal conditions are assumed, meaning that variations of the average daily temperature within the year are limited. Thisassumption is clearly not satised in many climatic regions(e.g., at high latitude). On the other hand, the model can stillbe applicable given that, during the unfavourable season (toohigh or low temperature), the moisture content becomes anadditional limiting factor, thus inhibiting soil respiration andtransformations.

The model of Porporato et al.(2003) is comprised of aset of coupled non-linear ordinary diff erential equations. Eachequation describes the mass balance of C and N in the vepools. An overview of the reaction network is given in Fig-ure3. Moreover, since the soil moisture is the key factor in thismodel, and inuences the decomposition and turnover rates asoutlined above, soil water variations are computed from thewater balance at one point. In order to facilitate model under-standing and comparisonwith previousworks, here we use thesame notation as in Porporato et al. (2003), DOdorico et al.(2004) and Rodriguez-Iturbe and Porporato (2004).

The evolution of C in the litter, humus and biomass pools isgiven by:

d Cldt

= ADD + BD DEC l , (6)

d Chdt = r h DEC l DEC h , (7)

d Cbdt

= (1 r h r r ) DEC l + (1 r r ) DEC h BD , (8)

where Cl, Ch and Cb are the C concentrations in the litter, hu-mus and biomass pools respectively [M L3], ADD is the exter-nal input of C to the system [M L 2 T 1], BD is the recyclingrate of decaying biomass in the litter pool [M L 3 T 1], DEC landDEC h are the C uxes leaving the litter and humus poolsdue to microbial decomposition [M L 3 T 1], whiler h and r r are non-dimensional coefficients representing the fractions of decomposed organic C that go into the humus pool and to res-

piration, respectively.

The combination of equations (6)(8) gives the overallbalance equation (Ctot ) in the system:

d Ctot

dt

= ADD r r DEC l r r DEC h . (9)

The ux of C between two pools is described by rst-ordkinetic equations ( DEC l, DEC h and BD), where the reactionrates (k l, k h andk d , respectively) are weighted averages of thdecomposition rates of the diff erent organic molecules. Therst-order kinetic equations of C decomposition and microbdeath for the litter, humus and biomass pool are:

DEC l = f d (s) Cb k l Cl , (10) DEC h = f d (s) Cb k h Ch , (11)

BD = Cb k d , (12)

where is a non-dimensional factor that accounts for a possble reduction of the decomposition rate when the litter is vpoor in N (high C / N ratio) and the immobilization is not sufcient to integrate the required N by the bacteria. This fachas an important inuence on the dynamics of the biomevolution, and details on how it is dened and computed cbe found in Porporato et al. (2003). The termf d (s) is a non-dimensional parameter that describes soil moisture eff ects ondecomposition:

f d (s) =

ss f c

, s s f c,

s f cs

, s > s f c,(13)

whereS f c is the soil eld capacity (water content held in so

after excess water drained away by gravity). The main moparameters are listed in TableII.The N balance in the litter, humus and biomass pools

computed from the C balance equations, scaled by the apppriate C / N ratio:

d Nldt

=ADD

(C/ N)add +

BD(C/ N)b

DEC l(C/ N)l

, (14)

d Nhdt

= r h DEC l(C/ N)h

DEC h(C/ N)h

, (15)

d Nbdt

= [1 r h(C/ N)l(C/ N)h

] DEC l(C/ N)l+

DEC h(C/ N)h

BD

(C/ N)b ,

(16)

where Nl, Nh and Nb are the N concentrations in the litter, humus and biomass pools, respectively [M L 3 T 1], (C / N)add ,(C / N)l, (C / N)h and (C / N)b are the C to N ratios of added or-ganic matter, litter, humus and biomass pools, respectiveand is a term that takes into account the contribution dto either the net mineralization or to the immobilization L 3 T 1]. This term relates the total mineralization and immbilization rates:

= MI N IMM , (17)whereMIN expresses the mineralization rate [M L 3 T 1] and IMM is the total rate of immobilization (sum of the N immbilization rate in the NH+4 IMM + and NO3 IMM pools, re-

spectively) [M L 3

T 1

]. WhenIMM is equal to zero,MIN is

25


10/24

Nitrogen plant uptake

Figure 3. Schematic representation of soil C and N cycles (after Rodriguez-Iturbe and Porporato, 2004).

equal to, while whenMIN is zeroIMM is equal to . The(C / N)b ratio is one of the most important parameters in themodel, since the switch between mineralization and immobi-lization is dened in order to maintain as constant the C / N ra-tio of the biomass pool. If the organic matter is rich in N (and(C / N)b is smaller than the value required to sustain growth of microbial biomass), decomposition results in surplus N. Thisis used by the microorganisms, and mineralization occurs. Incontrast, if decomposition produces an environment poor in N,microorganisms will increase the immobilization rate of NH+4and NO

3in order to meet their requirements. This process is

rather complex and very dynamic, as explained in Porporatoet al.(2003).

N transfer between the pools is described by the same rst-order kinetic transfer parameters used for C, with each termscaled by the corresponding C / N ratio (Fig.3). The balanceequations of inorganic N are:

d N+dt

= MI N + IMM + NIT LE + UP + , (18)

d Ndt

= NI T IMM LE UP , (19)

where N+ and N are the inorganic N concentrations in the

NH+4 and NO

3 pools, respectively [M L

3], NIT is the nitri-

cation rate [M L 3 T 1], UP + andUP are the N uptake byplants from the NH+4 and NO3 pools, respectively [M L 3 T 1],andLE + andLE are N uxes from the root zone towards thegroundwater [M L 3 T 1].

The combination of equations (1416, 18, 19) gives thoverall evolution of total N (Ntot ) in the system:

d Ntot dt

= ADD

(C/ N)add LE + UP + LE UP . (20)

Equations (9) and (20) represent the total C and N sinks ansources of the system depicted in Figure3.

Although we have described the main elements of thmodel here, further descriptions for example the rates omineralization ( MIN ), immobilization ( IMM + andIMM ), ni-trication ( NIT ), plant uptake (UP + and UP ) and leach-ing ( LE + and LE ) and their associated variables can befound in Porporato et al. (2003), DOdorico et al. (2004) andRodriguez-Iturbe and Porporato (2004), together with addi-tional discussion about the underlying assumptions and sim

plications introduced in this model.



11/24


12/24

Table III. Soil and climate parameters corresponding to the modelled scenarios.

Climatic parameters ClimateA(seasonality)

ClimateB(no seasonality)

Mean storm frequency () d 1 0.3, 0.1 0.2Mean storm [] depth () m 1.55 10 2, 4.8 10 3 1.08 10 2Soil parameters Soil I (agricultural) Soil II (forest)Saturated hydraulic conductivity (K ) m d 1 1 2Active root depth ( Z r ) m 0.3 0.7Pore size distribution index (b) - 4.05 5.39Porosity (n) - 0.35 0.45Tortuosity (d ) - 1.5 3Field capacity (s f c) - 0.35 0.55

Vegetation parameters

Maximum transpiration (T max ) m 3.6 10 3 4.63 10 3Canopy interception threshold m 5 10 4 2 10 3Maximum evapotranspiration ( E max ) m d 1 3.7 10 3 4.76 10 3Evapotranspiration wilting point ( E w) m d 1 10 4 1.3 10 4Hygroscopic point (sh) - 0.04 0.14Point of incipient stress (s) - 0.16 0.37Wilting point (sw) - 0.05 0.17

Modelled scenariosScenarioAI Agricultural soil with climatic seasonalityScenarioBI Agricultural soil without seasonalityScenarioAII Forest soil with climatic seasonalityScenarioBII Forest soil without seasonality

Table IV. Parameters related to carbon and nitrogen soil dynamicsused in all model scenarios. Shaded parameters correspond to vari-ables that are not focused upon in this manuscript. Their details canbe found in Porporato et al. (2003).

Initial soil moisture s 0.15C litter pool Cl g m 3 1200C humus pool Ch g m 3 8500C biomass pool Cb g m 3 50N ammonium pool N + g m 3 0N nitrate pool N g m 3 1Added litter Add g C m 3 d 1 1.5C / N ratio of added litter (C / N )add 58C / N ratio of litter (C / N )l 22C / N ratio of humus (C / N )h 22C / N ratio of microbial biomass (C / N )b 11.5C / N ratio of ammonium (C / N )+ 1C / N ration of nitrate (C / N ) 1Isohumic coefficient r h 0.25Respiration coefficient r r 0.6Litter decomposition rate k l m3 d 1 g C 1 6. 5 10 5

Factor of carbon return to litter pool k d d 1 8.5 10 3Humus decomposition rate k h m3 d 1 g C 1 2.5 10 6Rate of nitrication k n m3 d 1 g N 1 0.6Ammonium immobilization coefficient k amm m3 d 1 g N 1 1Nitrate immobilization coefficient k nit m3 d 1 g N 1 1Ammonium plant demand DEM + g N m 3 d 1 0.2Nitrate plant demand DEM g N m 3 d 1 0.5Fraction of dissolved ammonium a + 0.05Fraction of dissolved nitrate a 0.1

Parameters representing the two climates considered arepresented in TableIII. ClimateA is characterized by season-

ality represented by two wet and two dry seasons over a year.

This climate can be consideredcomparable to a Mediterraneaclimate, with two wet seasons, spring and fall (e.g., high and) and two dryseasons, summer and winter (e.g., low and).In contrast, climateB is characterized by a lack of seasonal-ity, with relatively low but homogeneous amount of precipit

tion randomly distributed over the year, using and betweenthose of the wet and dry seasons considered in climateA.Although in Section 3 was pointed out that temperature ex

erts a control over soil C and N cycles, in this study onlisothermal conditions were considered. This assumption wamade for two reasons. First, the eff ect of temperature in manyclimates is less important than that of soil moisture and, seond, because considering the eff ects of soil moisture alone thenumber of factors aff ecting soil nutrient cycles is reduced, andis therefore easier to understand the inuence and feedbackon soil changes and nutrient dynamics. Temperature variationare however closely related to climate conditionsand thereforthis factor should be considered in future analyses.

6.2. Soil scenarios

We seek to identify whether diff erent patterns of soil mois-ture evolve through time as a consequence of the combinatioof diff erent processes in the soil-plant-atmosphere system. Tothis end, equation (21) is used to calculate the soil moistubalance at a point (Laio et al.,2001):

nZ r ds(t )

dt = R(t ) I (t ) Q[s(t ); t ] E [s(t )] L[s(t )], (21)

where n is the porosity;Z r is the depth of active soil or

root depth [L];s(t ) is the relative soil moisture content

J. Batlle-Aguilar et al.262


13/24


(0 s(t ) 1); R(t ) is the rainfall rate [L T 1]; I (t ) is theamount of rainfall lost through interception by canopy cover[L T 1]; Q[s(t ); t ] is the rate of runoff [L T 1]; E [s(t )] is theevapotranspiration rate [L T 1]; andL[s(t )] is the leakage rate

[L T 1

].The soil was assumed as a horizontal and homogeneouslayer of depthZ r . This is an important assumption becausesoil depth depends in time and space on two main parameters,soil structure and vegetation. In the simulations we consideredthe same soil depth for water balance and nutrient cycles. Wa-ter inltration into the soil and runoff are entirely controlledby soil moisture dynamics, since water will inltrate into thesoil if there is available storage. Excess rainfall that cannot bestored in the soil is converted into runoff .

Although the vegetation type depends on both climate andsoil, here the vegetation parameters were xed for each soilin order to reduce the number of variables aff ecting changes

in C and N uxes (Tab.III). This is also justied by the factthat vegetation parameters mostly depends on soil moisture,which directly depends on soil characteristics. Evapotranspi-ration varies from a maximum valueE max when soil moistureranges between the maximum, unity, and the point of incipi-ent stress,s(soil moisture level at which the plants begin toclose stomata in response to water stress). The evapotranspi-ration rate decreases linearly fromE max to E w, the latter ratecorresponding to the soil moisture at the wilting pointsw (soilwater content at which plants wilt and can no longer recoveror, in terms of water potential, is dened as the suction headbeyond which the plant can no longer take up water). Belowthis value, only transpiration is active, and the water loss rateis linear fromE w to zero at the point of hygroscopic watersh(microscopic lm of water covering soil particles not availablefor plants). More details and assumptions concerning evapo-transpiration are given by Laio et al. (2001).

Verhoef and Brussaard (1990) dened a series of functionalgroupsof pedofaunabased on their contribution to nutrient de-composition and mineralization. Organisms belonging to thesame functional group play a similar role in decomposition-mineralization transformations. For example, there is a func-tional group that includes organisms that pulverize, mix andgranulate the soil. Such organisms are rather important be-cause (i) they contribute to incorporate the organic residuesavailable on the surface into the lower horizons; and (ii)they create large pores and channels that guarantee aerationof the soil prole and eliminate excess water. Other func-tional groups include pedofauna specialized in breaking downwoody recalcitrant materials, in degrading litter and digestingorganic residues, etc. (Brady and Weil,2004). Although thefunctional group concept is useful for modelling of soil nutri-ent cycles, here a more simplied approach considering a sin-gle value for nutrient decomposition and mineralization ratesis used, representing the contribution of the entire pedofaunato these processes.

In combination with climatesA and B, two soil types named I and II are considered, representative of agriculturaland forest soils respectively (Tab.III). Agricultural and forestsoils have contrasting physical properties mainly due to man-

agement practices and the type of vegetation supported (Lutz

and Chandler,1955; Carmean,1957; Garca-Oliva et al.,1994;de Moraes et al.,1996; Islam and Weil,2000). The main dif-ferences between these soils are:

Silt and clay content . Agricultural soils have loweramounts of silt and clay than natural forest soils, mosas a result of preferential removalof these particles by wter erosion (Islam and Weil,2000).

Soil aggregate stability . A higher input of litter fall com-bined with the absence of tillage and ploughing practicgivesrise to forest soils with greater soil aggregatestabili(Islam and Weil,2000). Furthermore, forest soils are natu-rally protected from the impact of raindrops by the canoand organic matter at the soil surface that absorbs raindrenergy (Carmean,1957). In practice, the eff ect of rain-drops is (i) removal of silt and clay particles and (ii) disrution of soil aggregates that subsequently can block larpores and reduce water percolation.

Bulk density and porosity . Agricultural soils have higherbulk density and lower porosity than forest soils, mainbecause of a greater residual sand content combined wpoorer soil aggregation (Garca-Oliva et al.,1994; deMoraes et al.,1996).

Soil structure . Agricultural soils often have a deterioratestructure in comparisonto forestsoils. This deteriorationapparent in pore modication, increased bulk density, icreased compaction, and less stable aggregates (Carmea1957). In addition, compacted, impermeable layersor panwithin the soil prole often develop as a consequenof repeated ploughing, mainly under wheel track patter(Roger-Estrade et al.,2004; Coquet et al.,2005).

Inltration rate . As a consequence of above mentionedproperties, which contribute to reduce the average posize and their connectivity, the rate of inltration is rduced in agricultural soils. Additionally, forest vegetatihas more extensive root networks, leading to large numbof interconnected channels leading to rapid water inlttion (Lutz and Chandler,1955).

Runo ff and soil erosion . Low inltration rates of agricul-tural soils contribute to increased runoff , which empha-sizes soil erosion and removal of silt-clay soil particl(Lutz and Chandler,1955).

Due to the above diff erences, the agricultural soil (soil I) ischaracterized by a relatively low saturated hydraulic condtivity (K ), as well as lower values of pore size distribution (b),porosity (n) and soil eld capacity (s f c) than soil II, representa-tive of a forest soil (Ndiaye et al.,2007). Furthermore, the soiltortuosity is likely to be aff ected by the loss of structure andby the less extendedroot network of agricultural soils, the lof connected porosity due to tillage processes and disturbaggregates clogging large pores. The soil tortuosity index the agricultural soil is also thus decreased in comparison wthe forest soil (Tab.III).

The rooting depth ( Z r ) considered is larger for forest soilsthan agricultural soils, since the root network is much mimportant for forest vegetation than agricultural. As preously mentioned, vegetation depends on both climate and s

However, we have dened the vegetation parameters only

26


14/24

J. Batlle-Aguilar et al.

Figure 4. Simulated rainfall and corresponding water saturation of climateA (with seasonal eff ects) and climateB (without seasonal eff ects)for idealized agricultural (a, b and c) and forest (d, e and f) soils.

function of the soil type. As a consequence of the low C / Nratio of agricultural vegetation and microbial decomposers as-sociated, cultivated soils typically have lower C / N ratio thanforest soils (Zheng et al.,1999). Nevertheless, to facilitate thecomparisonbetween the four scenarios we assumed equal C / Nratios for both soils, as well as initial C and N concentrations.

7. RESULTS

Figures4a and4d present the precipitation over 20 y forclimatesA (seasonality) andB (no seasonality), respectively,while the evolutionof water saturation for thesame period ascomputed with equation (21) for each climate and soil typeis depicted in Figures4b4f, respectively. There is a markeddiff erence in precipitation distribution between climatesA and B, with wet and dry seasons in climateA and random uni-formly distributed precipitation in climateB. Soil water sat-uration follows the dynamics imposed by precipitation, morenotably in agricultural soils while the trend in water satura-tion evolution is smoothed in forest soils. As expected, peaks

of soil saturation are lower for those soils under the inuence

of climateB than climateA, while water saturation in agricul-tural soils is lower than those of forest soils. The latter is due tdiff erent vegetation parameters associated with each soil typspecically to the wilting point (sw), xed at 0.05 and 0.17 foragricultural and forest soils, respectively. It is interesting tnote that forest soils attenuate changes of soil saturation mucmore than agricultural soils (e.g., compare Figs.4b to 4c oreven Figs.4e to4f ), and delays water saturation peaks, mostlydue to diff erence rooting depth.

The evolution of the ve diff erent nutrient pools, for agri-cultural and forest soils under the conditions of climateA, arepresented in Figure5. Figures corresponding to the same poolare depicted with the same vertical scale in order to facilitacomparison between them. From the comparison is evidenthat seasonal eff ects are much more visible in agricultural thanin forest soils. The lower rooting depth ( Z r ), soil hydraulicconductivity (K ) and soil porosity (n) in the agricultural soilmay be at the origin of these diff erences. Although forest soilspresent similar behaviour, small peaks observed in agricutural soils and corresponding to high frequency changes i

precipitation, are not evident. As previously mentioned, litt

264


15/24


500

1000

1500

2000

2500

3000a Litter pool

C ( g m -

3 )

Soil I (agricultural)

0

100

200

300

400 c Biomass pool

C ( g m

- 3 )

0

0.002

0.004

0.006

0.008

0.01d Ammonium pool

N ( g m

- 3 )

0 5 10 15 200

0.75

1.5

2.25

3e Nitrate pool

N ( g m

- 3 )

Time (years)

500

1000

1500

2000

2500

3000f Litter pool

Soil II (forest)

0

100

200

300

400 h Biomass pool

0

0.002

0.004

0.006

0.008

0.01i Ammonium pool

0 5 10 15 200

0.75

1.5

2.25

3 j Nitrate pool

Time (years)

8300

8800

9300

9800

10300g Humus pool

8300

8800

9300

9800

10300b Humus pool

C ( g m

- 3 )

C ( g m -

3 )

C ( g m

- 3 )

N ( g m

- 3 )

N ( g m

- 3 )

C ( g m

- 3 )

Figure 5. Simulated organic carbon and inorganic nitrogen concentrations for agricultural (left side) and forest (right side) soils undeAconditions.

26


16/24


0

5

10

15

20

25

30

35a Litter decomposition

C ( g m

- 3 d

- 1 )


0

0.2

0.4

0.6

0.8

1

b Net Mineralization

N ( g m

- 3 d

- 1 )

0

0.2

0.4

0.6

0.8

1c Nitrate uptake

N ( g m

- 3 d

- 1 )

0 5 10 15 200

0.1

0.2

0.3

0.4d Nitrate leaching

N ( g m

- 3 d

- 1 )

Time (years)

0

5

10

15

20

25

30

35e Litter decomposition

C ( g m

- 3 d

- 1 )

Soil II (forest)

0

0.2

0.4

0.6

0.8

1f Net Mineralization

N ( g m

- 3 d

- 1 )

0

0.2

0.4

0.6

0.8

1g Nitrate uptake

N ( g m

- 3 d

- 1 )

0 5 10 15 200

0.1

0.2

0.3

0.4h Nitrate leaching

N ( N m

- 3 d

- 1 )

Time (years)

Figure 6. Simulated rates of litter decomposition (a, e), net nitrogen mineralization (b, f), nitrate uptake (c, g) and nitrate leaching (both agricultural and forest soils under climateA conditions.

decomposition in agricultural soils is enhanced by tillage andplough practices, represented in the model by lower valuesof saturated hydraulic conductivity (K ), pore size distribution(b), porosity (n) and soil eld capacity (s f c). This explainsthe lower values of C litter pool in agricultural soils (com-pare Figs.5a and5f), although the high inuence of seasonalclimates over agricultural soils is likely to hide this fact. Fur-thermore, this trend is likely to inuence the evolution of sub-sequent pools (Figs.5g5 j). Peaks of litter C match well thebiomass pool decrease, while decreases of litter C concen-trations well correlate with the peaks of C concentration in

the biomass pool for both agricultural and forest soils. This

is not surprising since the augmentation of the biomass pois linked to degradation of the litter pool. C concentrationin the humus pool (Ch) for agricultural and forest soils arelower than in other pools, since this pool is an intermediapool between litter and biomass and there is almost no inteaction with other variables. Concentrations of NH+4 (Figs.5dand 5i) show a trend similar to that of C in the litter poolalthough NH+4 variations are almost negligible since its con-centration is very low. The NO3 pool preserves much of thehigh-frequency variability imposed by the random forcing oprecipitation, which is not surprising since NO3 dynamics are

the nal productof a numberof intertwined processes in which

266


17/24


0 5 10 15 200

2

4

6

8

10

12 j Nitrate pool

Time (years)

0

100

200

300

400h Biomass pool

9300

9600

9900

10200

10500g Humus pool

500

1000

1500

2000

2500

3000f Litter pool

Soil II (forest)

0

0.002

0.004

0.006

0.008

0.01i Ammonium pool

0 5 10 15 200

2

4

6

8

10

12e Nitrate pool

Time (years)

0

100

200

300

400c Biomass pool

9300

9600

9900

10200

10500b Humus pool

500

1000

1500

2000

2500

3000a Litter pool


0

0.002

0.004

0.006

0.008

0.01

d Ammonium pool

C ( g m

- 3 )

C ( g m

- 3 )

N ( g m

- 3 )

N ( g m

- 3 )

C ( g m

- 3 )

C ( g m

- 3 )

C ( g m

- 3 )

N ( g m

- 3 )

N ( g m

- 3 )

C ( g m

- 3 )

Figure 7. Simulated organic carbon and inorganic nitrogen concentrations for agricultural (left side) and forest (right side) soils undeBconditions.

26


18/24


0 5 10 15 200

0.1

0.2

0.3

0.4h Nitrate leaching

N ( N m

- 3 d

- 1 )

Time (years)

0

0.2

0.4

0.6

0.8

1g Nitrate uptake

N ( g m

- 3 d

- 1 )

0

5

10

15

20

25

30

35e Litter decomposition

C ( g m

- 3 d

- 1 )

Soil II (forest)

0

0.2

0.4

0.6

0.8

1f Net Mineralization

N ( g m

- 3 d

- 1 )

0

0.2

0.4

0.6

0.8

1c Nitrate uptake

N ( g m

- 3 d

- 1 )

0

5

10

15

20

25

30

35a Litter decomposition

C ( g m

- 3 d - 1 )


0

0.2

0.4

0.6

0.8

1b Net Mineralization

N ( g m

- 3 d

- 1 )

0 5 10 15 200

0.1

0.2

0.3

0.4d Nitrate leaching

N ( g m

- 3 d

- 1 )

Time (years)

Figure 8. Simulated rates of litter decomposition (a, e), net nitrogen mineralization (b, f), nitrate uptake (c, g) and nitrate leaching (both agricultural and forest soils under climateB conditions.

both high- and low-frequency components interact. It is inter-esting to note the low levels of NO3 obtained for agriculturalsoils (Fig.5e) compared with the relatively high levels in for-est soils (Fig.5 j). This is reected in Figure6d and6h, whereNO3 leaching is almost non-existent in forest soils while it isrelatively important for agricultural soils. The higher rates of mineralizationand NO3 uptakein agricultural soils than in for-est soils aid to explain this fact. Note that fertilizationpracticescommonly undertaken in agricultural soils are not consideredin the model. Thus, simulated results are in good agreementwith the need of agricultural soils for regular fertilizer input,

since high decomposition and mineralization rates (Figs.6a

and 6b), NO3 uptake by plants (Fig.6c) and NO3 leaching(Fig.6d) to lower layers results in low levels of NO3 availablefor plants.

The results obtained, especially in forest soils, providevidence of a structure with temporal amplitudes consideably larger than the one induced by the stochastic hydrologforcing. This behaviour was already observed by Thornleet al. (1995) and DOdorico et al. (2003). The presence of this behaviour in soil nutrient dynamics (see Figs.5f5j), apriori not related to climatic conditions, is a manifestation the nonlinearity and degree of complexity of the whole sy

tem. Thornley et al. (1995) pointed out that such behaviour

268


19/24


suggests that soil nutrients cycles could show cases of richer(perhaps chaotic) dynamics enmeshed with the variability thatis directly induced by the stochastic hydrologic uctuations.This issue is not fully assessed yet to date, although a partial

discussion is provided by Manzoni and Porporato (2007).The evolution through time of the C and N pools in agri-cultural and forest soils under climateB (no seasonality) ispresented in Figure7. A general trend for both agriculturaland forest soil is that simulated concentrations in all C and Npools closely reproducetheconcentrationsobserved in climate A (seasonal). This is supported by the similar decompositionrates between climateA (Fig.6) andB (Fig.8). Nevertheless,the absence of seasonality, which is reected in agriculturaland forest soils by a loss of the high frequency nutrient peaks,is nally translated in less nitrate leaching and less variabil-ity in the soil nutrients. While water saturation changes un-der seasonal climates contribute to hinder decomposition rates

from time to time, this does not occur in climates without sea-sonal eff ects, where decomposition peaks are more continu-ous through time. The loss of climate seasonality aff ects nei-ther the inverse trend between litter and biomass pools, nor thehigh frequency of NO3 changes in both agricultural and forestsoils.

As observed under seasonal conditions, the frequency of changes in C and N concentrations is higher in agriculturalthan in forest soils, although the general trend remains quitesimilar for both soils. Again, C concentrations in the humuspool are relatively constant, and NH+4 concentrations are low.Concentrations of NO3 are lower in agricultural soils than inforest soils, as observed under seasonal climate conditions.This fact supports what was mentioned in previous sectionsthat agricultural soils are relatively poor in nutrients, indepen-dently of climate conditions. Furthermore, this is consistentwith the higher mineralization rate (Fig.8b), higher NO3 up-take by plants (Fig.8c) and higher loss of NO3 by leaching(Fig.8d) in agricultural than in forest soils.

Finally, some general insights canbe gainedfromthemodelresults. Great variations of the average value of C and N con-centrations in forest soils are obtained both under seasonal anduniformlyvariableclimatic conditions.These results show thata point measurement of litter, for example, is almost meaning-less, and that time-series are neededto provideuseful informa-tion for evaluation of soil nutrient turnover. It is worth notingalso that steady state concentrations are not always reached,which means that longer simulations should be applied to bet-ter evaluate long term nutrient concentrations, mainly undernon-seasonal conditions (see for example Fig.5g in compari-son to 5b).

8. CONCLUSIONS

Land use changes aff ect soil properties and, thus, nutrientcycling dynamics. Changing soil properties means also chang-ing the type of vegetation, altering even more the concomi-tant nutrient cycles. The example most widely observed is thechange from forest to agricultural soils, due to increasing an-

thropogenic demands for food production. Agricultural prac-

tices generally cause changes in soil structure, compromisaggregation and porosity, leading to a soil structure decliTillage practices like mechanical mixing compact and reduthe size of aggregatesand lls porespaces with nes. The si

ulation of two diff erent climates applied to two contrastingsoitypes(wherespecic vegetationpropertieswere linkedto easoil), showed that decomposition rates in agricultural soils higher, which in addition to lower input of fresh litter ova year in comparison to a forested area, results in lowerand N concentrations. In consequence, less NO3 is availableto plants, since plant NO3 uptake is higher and leaching is en-hanced under these conditions. The leached NO3 is compen-sated for by the input of extra nutrients in agricultural soilsfertilization, a practice that, if poorly managed, contributesdiminish the quality of aquifers.

Model application under seasonal and non-seasonal cmatic conditions resulted in a higher attenuation of punct

growing concentrations of soil C and N under the eff

ect of aseasonal climate, as well as averageconcentrationshigher ththose under the eff ect of non-seasonal climates. Furthermorethe high frequency imposed by seasonal climates is attenuain forest soils, while this high frequency is reected in agcultural soil nutrient cycles. Nevertheless, the general trendthe temporal dynamic is similar under both seasonal and nseasonal conditions.

Land use changes should include long term practices avoid the loss of soil properties, contributing to the mainnance of optimal conditions for long term agricultural pduction. Crop rotation is an important management practto avoid soil C losses following conversion from forestagricultural land (Murty et al.,2002). Furthermore, Agboola(1981) and Bruce et al. (1999) proposed that a diminution oftillage processes minimizes soil erosion and decompositirates, and thus soil C losses. The results presented here uderscore that models dealing with soil nutrient turnover potentially promising tools to design new soil practices apredict long-term eff ects of these practices on soil fertility.

Acknowledgements: This research is part of the RECORD project of thCompetence Centre Environment and Sustainability (CCES,http: // www.cces.ethz.ch / projects / nature / Record). Funding has been provided by the Swiss National Science Foundation under grant 200021-113296.

REFERENCES

Agboola A.A. (1981) The eff ects of diff erent soil tillage and managementpractices on the physical and chemical properties of soil and mayields in a rainforest zone of Western Nigeria, Agron. J. 73, 24251.

Agren G.I., McMurtrie R.E., Parton W.J., Pastor J., Shugart H.H. (19State-of-the-art of models of production-decomposition linkagesconifer and grassland ecosystems, Ecol. Appl. 1, 118138.

Arnold J.G., Allen P.M., Bernhardt G. (1993) A comprehensive surfagroundwater ow model, J. Hydrol. 142, 4769.

Austin A.T., Yahdjian L., Stark J.M., Belnap J., Porporato A., NorU., Ravetta D.A., Schaeff er S.M. (2004) Water pulses and biogeo-chemical cycles in arid and semiarid ecosystems, Oecologia 1

221235.

62
http://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Recordhttp://www.cces.ethz.ch/projects/nature/Record


20/24

2 0 J. Batlle-Aguilar et al.

Barbour M.G., Burk J.H., PittsW.D.,Gilliam F.S., Schwartz M.W. (1999)Terrestrial plant ecology. Pearson Benjamin Cummings, MenloPark, Canada.

Barry D.A., Prommer H., Miller C.T., Engesgaard P., Brun A., ZhengC. (2002) Modelling the fate of oxidisable organic contaminants ingroundwater, Adv. Water Resour. 25, 945983.

Bell C., McIntyre N., Cox S., Tissue D., Zak J. (2008) Soil microbialresponses to temporal variations of moisture and temperature in aChihuahuan desert grassland, Microb. Ecol. 56, 153167.

Berger T.W., Neubauer C., Glatzel G. (2002) Factors controlling soil car-bon and nitrogen stores in pure stands of Norway spruce (Piceaabies) and mixed species stands in Austria, For. Ecol. Manage. 159,314.

Bolker B.J., Pacala S.W., Parton W.J. (1998) Linear analysis of soil de-composition: insights from the century model, Ecol. Appl. 8, 425-439.

Bonde T.A., Christensen B.T., Cerri C.C. (1992) Dynamics of soil or-ganic matter as reected by natural 13C abundance in particle sizefractions of forested and cultivated oxisols, Soil Biol. Biochem. 24,275277.

Borken W., Xu Y.-J., Davidson E.A., Beese F. (2002) Site and temporalvariation of soil respiration in European beech, Norway spruce, andScots pine forests, Glob. Change Biol. 8, 12051216.

Borken W., Xu Y.J., Brumme R., Lamersdorf N. (1999) A ClimateChange Scenario for Carbon Dioxide and Dissolved OrganicCarbon Fluxes from a Temperate Forest Soil: Drought andRewetting Eff ects, Soil Sci. Soc. Am. J. 63, 18481855.

Bosatta E., Agren G.I. (1991) Theoretical analysis of carbon and nutrientinteractions in soils under energy-limited conditions, Soil Sci. Soc.Am. J. 55, 728733.

Bosatta E., gren G.I. (1994) Theoretical analysis of microbial biomassdynamics in soils, Soil Biol. Biochem. 26, 143148.

Bossel H. (1996) TREEDYN3 forest simulation model, Ecol. Model. 90,187227.Bradbury N.J., Whitmore A.P., Hart P.B.S., Jenkinson D.S. (1993)

Modelling the fate of nitrogen in crop and soil in the years follow-ing application of 15N-labelled fertilizer to winter wheat, J. Agric.Sci. 121, 363379.

Brady N.C., Weil R.R. (2004) Elements of the nature and properties of soils. Pearson Prentice Hall, Upper Saddle River, New Jersey, USA.

Brams E. (1971) Continuous cultivation of West African soils: Organicmatter diminution and eff ects of applied lime and phosphorus, PlantSoil 35, 401414.

Brown S., Lugo A. (1990) Eff ects of forest clearing and succession on thecarbon and nitrogen content of soils in Puerto Rico and US VirginIslands, Plant Soil 124, 5364.

Bruce J.P., Frome M., Haites E., Janzen H., Lal R., Paustian K. (1999)Carbon sequestration in soils, J. Soil Water Conserv. 54, 382389.

Brussaard L. (1998) Soil fauna, guilds, functional groups and ecosystemprocesses, Appl. Soil Ecol. 9, 123135.

Buchmann N. (2000) Biotic and abiotic factors controlling soil respirationrates in Picea abies stands, Soil Biol. Biochem. 32, 16251635.

Carmean W.H. (1957) The structure of forest soils, Ohio J. Sci. 57, 165168.

Cherif M., Loreau M. (2009) When microbes and consumers determinethe limiting nutrient of autotrophs: a theoretical analysis, Proc. R.Soc. B 276, 487497.

Chertov O.G., Komarov A.S. (1997) SOMM: a model of soil organic mat-

ter dynamics, Ecol. Model. 94, 177189.

Christensen B.T. (1996) Matching measurable soil organic matter frations with conceptual pools in simulation models of carboturnover: revision of model structure, in: Powlson D.S., SmitP., Smith J.U. (Eds.), Evaluation of soil organic matter modelsSpringer Verlag, Berlin, pp. 143160.

Christiansen J.R., Elberling B., Jansson P.E. (2006) Modelling water baance and nitrate leaching in temperate Norway spruce and beecforests located on the samesoil typewiththe CoupModel, For. EcoManage. 237, 545556.

Conant R.T., Klopatek J.M., Klopatek C.C. (2000) Environmental factocontrolling soil respiration in three semiarid ecosystems, Soil ScSoc. Am. J. 64, 383390.

Coquet Y., Simunek J., Coutadeur C., van Genuchten M.T., Pot V., RogEstrade J. (2005) Water and solute transp

Documents

Batlle-Aguilar Etal2011_Modelling Soil Carbon and Nitrogen Cycles During Land Use Change. a Review