Agricultural System

8/13/2019 Agricultural System

1/12

Modelling agricultural expansion in Kenyas Eastern Arc Mountains biodiversityhotspot

Eduardo Eiji Maeda * , Barnaby J.F. Clark, Petri Pellikka, Mika SiljanderUniversity of Helsinki, Department of Geosciences and Geography, Gustaf Hllstrmin katu 2, 00014 Helsinki, Finland

a r t i c l e i n f o

Article history:Received 18 January 2010Received in revised form 7 July 2010Accepted 16 July 2010Available online 19 August 2010

Keywords:Land use changeSimulation modelEastern Arc MountainsTaita Hills

a b s t r a c t

The Taita Hills are the northernmost part of the Eastern Arc Mountains of Kenya and Tanzania, which isone of the most important regions for biological conservation in the world. The indigenous cloud forestsin this area have suffered substantial degradation for several centuries due to agricultural expansion. Inthe Taita Hills, currently only 1% of the original forested area remains preserved. In order to create effec-tive policies to preserve the natural resources and biodiversity of the Eastern Arc Mountains it is crucial tounderstand the causes and interactions involved in the landscape changes in the most degraded areas.The research presented here aimed to understand the role of landscape attributes and infrastructurecomponents as driving forces of agricultural expansion in the Taita Hills. Geospatial technology toolsand a landscape dynamic simulation model were integrated to identify and evaluate the driving forcesof agricultural expansion and simulate future landscape scenarios. The results indicate that, if currenttrends persist, agricultural areas will occupy roughly 60% of the study area by 2030. Agricultural expan-sion will likely take place predominantly in lowlands and foothills throughout the next 20 years, increas-ing the spatial dependence on distance to rivers and other water bodies. The main factors driving thespatial distribution of new croplands were the distance to markets, proximity to already established agri-cultural areas and distance to roads. Other driving forces of the agricultural expansion, as well as theirimplications for natural resources conservation, are discussed. Further studies are necessary to integratethe effects of population pressure and climate change on the sustainability and characteristics of localagricultural systems.

2010 Elsevier Ltd. All rights reserved.

1. Introduction

The Eastern Arc Mountains (EAM) of Kenya and Tanzania main-tain some of the richest concentrations of endemic animals andplants on Earth, and are thus considered one of the worlds top25 biodiversity hotspots ( Myers et al., 2000 ). The EAM comprisea chain of mountains located in southern Kenya and eastern Tanza-nia and are home for at least 96 endemic vertebrate species and800 endemic vascular plant species ( Burgess et al., 2007 ). Althoughthis region is among the most important areas for biological con-servation in the world, it has already lost approximately 80% of its original forest area ( Hall et al., 2009 ).

Forest loss in sub-Saharan Africa is proceeding at an alarmingrate of 2.8 million ha per year; particularly in Afromontane areasthe decrease is estimated to be 3.8% annually ( Eva et al., 2006 ).Most of these losses are caused by agricultural expansion. Be-tween the years 1975 and 2000 the agricultural areas increased57% in sub-Saharan Africa ( Brink and Eva, 2009 ). Although the

development of the agricultural sector is essential to improvefood security in this region, the expansion of croplands withoutlogistic and technological planning is a severe threat to the envi-ronment. Besides the imminent risk to biodiversity, indiscrimi-nate agricultural activities may pose serious obstacles to waterresources and soil conservation.

One of the EAM sections most affected by agricultural expan-sion is the Taita Hills, which is the northernmost part of theEAM. Between 1955 and 2004, the indigenous forest areas in theTaita Hills decreased by 50% ( Pellikka et al., 2009 ). Today, only1% of the original forested area remains. Although only a smallfraction of the indigenous cloud forests is preserved, the Taita Hillscontinue to have an outstanding diversity of ora and fauna and ahigh level of endemism. It is home for six endemic vertebrate spe-cies, three endemic bird species and at least 13 endemic plant spe-cies ( Burgess et al., 2007; Brooks et al., 1998 ). Hence, a detailedstudy of this specic region is essential to preserve the remainingbiodiversity and, most importantly, to expand the understandingof the interactions between human activities and landscapechanges in the EAM. Such work can contribute to improving envi-ronmental protection policy in the regions of the EAM that are stillintact but currently threatened by agricultural expansion.

0308-521X/$ - see front matter 2010 Elsevier Ltd. All rights reserved.doi: 10.1016/j.agsy.2010.07.004

* Corresponding author. Tel.: +358 44 2082876.E-mail address: eduardo.maeda@helsinki. (E.E. Maeda).

Agricultural Systems 103 (2010) 609620

Contents lists available at ScienceDirect

Agricultural Systems

j o u r na l h om e pa ge : www.e l s e v i e r. c om / l oc a t e / a gs y
http://dx.doi.org/10.1016/j.agsy.2010.07.004mailto:[email protected]://dx.doi.org/10.1016/j.agsy.2010.07.004http://www.sciencedirect.com/science/journal/0308521Xhttp://www.elsevier.com/locate/agsyhttp://www.elsevier.com/locate/agsyhttp://www.sciencedirect.com/science/journal/0308521Xhttp://dx.doi.org/10.1016/j.agsy.2010.07.004mailto:[email protected]://dx.doi.org/10.1016/j.agsy.2010.07.004


2/12

The improvement in models and computer capacity during thepast decades has allowed an increasing number of studies aimingat the sustainable use of natural resources and land use planning.For instance, land use and land cover change (LUCC) simulationmodels provide robust frameworks to cope with the complexityof land use systems ( Veldkamp and Lambin, 2001; Washingtonet al., 2010 ). Such models are considered efcient tools to project

alternative scenarios into the future and to test the stability of interrelated ecological systems ( Koomen et al., 2008 ). Understand-ing the circumstances and driving forces of change is an essentialstep for elaborating public policies that can effectively lead tothe conservation of natural resources.

Among dynamic spatial models, frameworks operating on a cel-lular automata (CA) basis have arisen as a feasible alternative forthe analysis of land use dynamics and in the exploration of futurelandscape scenarios. CA models consist of a simulation frameworkin which space is represented as a grid of cells, and a set of transi-tion rules determine the attribute of each given cell considering theattributes of its neighboring cells ( Almeida et al., 2003 ). For in-stance, CLUE-S is a spatially explicit, multi-scale model that canbe applied to describe land change dynamics through the determi-nation and quantication of the bio-geophysical and human driv-ers of land use ( Verburg et al., 2002 ). Another widely used CAsimulation model is Dinamica-EGO ( Soares-Filho et al., 2002,2009 ). Dinamica-EGO is an environmental modelling platformdeveloped by the Centre for Remote Sensing of the Federal Univer-sity of Minas Gerais, Brazil (CSR-UFMG). This platform allows thedesign of static or dynamic simulations involving nested iterations,dynamic feedbacks and multi-scale approaches.

Moreover, geospatial technologies, such as remote sensing andgeographical information systems (GIS), have made available anunprecedented opportunity for new studies in terms of data collec-tion, availability and processing capacity. Nevertheless, scientistscurrently face the challenge of integrating these technologies tobetter understand the coupled relations between human activitiesand environmental changes. In this context, land change sciencehas emerged as an interdisciplinary eld that aims to understandthe dynamics of LUCC as a coupled human environment system(Turner et al., 2007 ).

This study addresses this exact issue, aiming to understand therole of landscape attributes and infrastructure components as driv-ing forces of agricultural expansion in the Taita Hills. In order toachieve this objective, remote sensing, GIS techniques and a LUCCsimulation model were integrated to identify and evaluate thedriving forces of LUCC and simulate future landscape scenarios.Thus, it is hoped that the results of this study may represent animportant instrument for households, researchers and policy mak-ers to better cope with future changes in local agricultural systems.

2. Study area

Taita Hills is the northernmost part of the EAM biodiversity hot-spot, situated in the middle of the Tsavo plains of the Taita-TavetaDistrict in the Coastal Province, Kenya ( Fig. 1). Taita Hills cover anarea of 1000 km 2 . The population of the whole Taita-Taveta districthas grown from 90,146 (1962) persons to over 300,000 (Republicof Kenya, 2001). The indigenous cloud forests have suffered sub-stantial loss and degradation for several centuries as abundantrainfall (annual 1100 mm) and rich soils (cambisols and humicnitosols) have created good conditions for agriculture. The agricul-ture in the hills is intensive small-scale subsistence farming. In thelower highland zone and upper midland zone, the typical crops aremaize, beans, peas, potatoes, cabbages, tomatoes, cassava and ba-

nana ( Soini, 2005 ). In the slopes and lower parts of the hills withaverage annual rainfall between 600 and 900 mm, early maturing

maize species and sorghum and millet species are cultivated. Inthe lower midland zones with average rainfall between 500 and700 mm, dryland maize types and onions are cultivated, amongothers. Moreover, the area is considered to have high scienticinterest and there is a high potential for succeeding in connectingeconomic development and community-based natural resourcemanagement ( Himberg et al., 2009 ).

3. Material and methods

This research integrated remote sensing, GIS techniques and aspatially explicit simulation model of landscape dynamics, Dinam-ica-EGO (Soares-Filho et al., 2007 ), to assess the driving forces of agricultural expansion in the study area and simulate future sce-narios of land use. A general description of the applied method isillustrated in Fig. 2.

The model receives as inputs land use transition rates, land-scape variables and landscape parameters. The landscape parame-ters are intrinsic spatially distributed features, such as soil typeand slope, which are kept constant during the simulation process.The landscape variables are spatialtemporal dynamic featuresthat are subjected to changes by decision makers, for instanceroads and protected areas. The land use transition rates were alsoconsidered to be decision variables, given that this modelling exer-cise was based on the assumption that agricultural expansion ratescan be modied by public policies or other external forces.

The model is driven by land use and land cover maps (LULCM)from two selected dates: 1987 (initial landscape) and 2003 (nallandscape), which are used as inputs to represent the historicalland use transitions in the study area. The dates of the LULCM werechosen based on two criteria. The rst criterion was that the land-scape changes between the initial and nal landscape should accu-rately represent the ongoing land change activities in the studyarea. That is to say, the agricultural expansion rates between1987 and 2003 were assumed to be representative of currenttrends. The second criterion relied on the availability of cloud freesatellite images to assemble the LULCM.

3.1. Landscape variables and parameters

In total, ten landscape attributes (variables/parameters) wereused as inputs for the model, nine of which were static and oneof which was dynamic. Static inputs are those that are kept con-stant throughout the model run, while dynamic inputs refer tothose that undergo changes during the model run. All landscapeattributes were represented by raster images with a 20 m spatialresolution. The description of each layer is as follows:

Distance to roads (DRo) : Euclidian distance in meters to main

and secondary roads.Distance to Markets (DM) : The markets were represented by

main villages in the region; the distance to markets raster was cre-ated by calculating the Euclidian distance in kilometers to centre of each village.

Digital Elevation Model (DEM) : The 20 m spatial resolution DEMwas interpolated from 50-feet interval contours captured from1:50,000 scale topographic maps, deriving an estimated altimetricaccuracy of 8 m and an estimated planimetric accuracy of 50 m.

Distance to Rivers (DRi) : Represented by the Euclidian distancein meters to main rivers. Two sources were used to extract the riv-er network in the study area. Firstly, GIS tools were used to auto-matically identify the rivers based on a ow accumulation gridobtained using the DEM. Subsequently, eventual errors in the auto-

matic classication were corrected using a 1:50,000 scale topo-graphic map.

610 E.E. Maeda et al. / Agricultural Systems 103 (2010) 609620


3/12

Protected Areas (PA) : Characterized by the national parks andconservation areas close to the Taita Hills. A segment of Tsavo EastNational Park is located in the northeastern part of the study areaand a small section of Taita Hills Game Sanctuary in thesouthwestern.

Soil Type (ST): The soil map was obtained from the Soil and Ter-rain Database for Kenya (KENSOTER), at scale 1:1 M, compiled bythe Kenya Soil Survey ( Batjes and Gicheru, 2004 ).

Slope (S) : The slope (%) was extracted from the DEM.Insolation (I) : Annual average solar radiation in watt hours per

square meter (W h/m 2) for the whole year was created from theDEM using ArcGIS 9.3.

Mean annual precipitation (P) was obtained by the compilationof long term mean precipitation grids interpolated from available

meteorological data in surrounding areas using ANUSPLINE soft-ware ( Hutchinson 1995; Maeda et al., 2010a ).

Distance to croplands (DC) : Represented by the Euclidian dis-tance to already established croplands. This layer was the only dy-namic landscape attribute used as an input for the model, whichmeans that this variable undergoes changes during the modelrun as new cropland patches are created.

3.2. Global transition rates and local transition probabilities

Global transition rates refer to the total amount of changes foreach type of land cover given in the simulation period, without tak-

ing into account the spatial distribution of such changes. The tran-sition rates were calculated by cross-tabulation, which produced as

Fig. 1. Geographic location of the study area.

Fig. 2. General description of the method, in which landscape attributes obtained using remote sensing and GIS techniques are used as inputs for a LUCC model. The modelevaluates the role of each attribute in the land changes and simulates future landscape scenarios.

E.E. Maeda et al./ Agricultural Systems 103 (2010) 609620 611


4/12

output a transition matrix between the LULCMs from 1987 and2003. The transition matrix describes a system that changes overdiscrete time increments, in which the value of any variable in a gi-ven time period is the sum of xed percentages of the value of thevariables in the previous time period ( Soares-Filho et al., 2009 ).

The land use/land cover for the Taita Hills was mapped fromSPOT 4 HRVIR satellite images (path & row 143357), with 20 m

spatial resolution and green, red and Near-infrared (NIR) spectralbands. Because of the rugged terrain in the area, the images wereorthorectied using a 20 m planimetric resolution DEM. Atmo-spheric correction was implemented utilizing the historical empir-ical line method (HELM) ( Clark and Pellikka, 2009 ). The SPOTimages were classied according to a nomenclature derived usingthe land cover classication system (LCCS) of the Food and Agricul-ture Organization of the United Nations (FAO) and the United Na-tions Environment Programme (UNEP) ( Di Gregorio, 2005 ).

The land cover classication methodology itself utilized a mul-ti-scale segmentation/object relationship modelling (MSS/ORM)approach implemented with the Deniens software tool ( Burnettand Blaschke, 2003 ). Central to this methodology is the generationof meaningful image objects, relating to homogeneous land coverpatches, by multi-scale segmentation based on both spectral andtextural characteristics of the imagery. More detailed descriptionof the method is presented in Clark and Pellikka (2009) . An accu-racy assessment for the 2003 classication was undertaken basedon ground reference test data, independent of the training data,collected during eld visits in 2005/2006 using stratied randomroad sampling. Additional reference points for the test data werecollected from 0.5 m resolution airborne true-colour digital cameraimagery acquired in January 2004.

The local transition probabilities, different from the global tran-sition rates, are calculated for each grid cell considering the naturaland anthropogenic characteristics of the site. Different methodscan be used to estimate local probabilities in Dinamica-EGO, suchas logistic regression and neural networks ( Soares-Filho et al.,2002; Almeida et al., 2008 ). In the particular case of this work,the transition probability of each cell was calculated in Dinami-ca-EGO using the weights of evidence (WoE) method. The WoE isa Bayesian method, in which the effect of each landscape variableon a transition is calculated independently of a combined solution(Soares-Filho et al., 2002 ). The spatial probability of a transition isgiven by the following equation ( Bonham-Carter, 1994 ):

P x; yf T =V 1 \ V 2 \ \ V ng Of T g eP

n

i1W x; y

1 Of T g P t j1 ePn

i1W x; y

; 1

where P x,y is the probability of transition in a cell with coordinates x, y; T is land use/land cover transition; V i is all possible landscapevariables i selected to explain transition T ; and O{T } is the odds of a transition, represented by the ratio between a determined transi-

tion probability and the complementary probability of non-occur-rence, described by following equation:

Of T g P f T gP f T g

; 2

where P{T } is the probability of transition T occurring, given by thenumber of cells where the relevant land use/land cover transitionoccurred divided by the total number of cells in the study area;P {T } is probability of transition T not occurring, given by the num-ber of cells where the relevant land use/land cover transition is ab-sent divided by the total number of cells in the study area; and W x; yis the weight of evidence for a determined landscape variable range,dened by the following equation:

W log eP f Vi=T gP f Vi=T g ; 3

where P{V i/T } is the probability of variable V i occurring in the pres-ence of transition T , given by the number of cells where both V i andT are found divided by the total number of cells where T is found;and P {V i/T } is the probability of variable V i occurring in the absenceof transition T , given by the number of cells where both V i and T arefound divided by the total number of cells where T is not found.

Hence, the W + values represent the attraction between a deter-mined landscape transition and a certain variable. The higher theW + value is, the greater is the probability that a certain transitiontakes place. On the other hand, negative W + values indicate lowerprobability of a determined transition occurring in the presence of the respective variable range. Based on the W + values of each rangefor every variable considered, Dinamica-EGO generates a spatiallyexplicit probability map, in which each cell is assigned the proba-bility for a determined transition.

3.3. Calibration of Dinamica-EGO internal parameters

The Dinamica-EGO platform has the advantage of using sto-chastic algorithms for land use change allocation. Two transitionalgorithms are responsible for the allocation of the land use/landcover changes: expander and patcher . The expander function per-forms the expansion of previously existing patches of a certainclass. The patcher function, in turn, is designed to generate newpatches through a seed formation mechanism ( Soares-Filho et al.,2002 ). The applied algorithms consist of scanning the initial landcover/land use map to sort out the cells with the highest probabil-ities and then arrange them in a data array. Following this proce-dure, cells are selected randomly from top to bottom of the dataarray (the internal stochastic selection mechanism can be loosenedor tightened depending on the degree of randomisation desired). Ina nal step, the land cover/land use map is again scanned to per-form the selected transitions. Hence, the rst parameter to be cal-ibrated in the simulation is the percentage of changes that will beaddressed by each of these two algorithms. For instance, in regionswhere landscape changes happen exclusively by the expansion of existing patches, all changes should be arranged by the expander function.

The next parameters to be adjusted are the mean and varianceof new patches sizes. These parameters can be independently adjusted for the expander and patcher functions. The model also in-cludes another heuristic parameter denoted the patch isometryindex. A high isometry index results in compact patches, whilelow values are reected in more fragmented formations.

The calibration of all parameters was conducted simulta-neously. Firstly, initial values are set based on empirical knowledgeobtained in previous eld work in the region. Next, the model isrun to simulate the landscape scenario in 2003 using the map from1987 as the initial landscape. The simulated landscape for 2003 iscompared with the reference LULCM by visual analysis, and the

parameters adjusted through heuristic procedures in order to im-prove the results. The model is iteratively executed and evaluated,and the results compared with the previous run. These proceduresare repeated until an optimal result is found in the modelevaluation.

The model tting was evaluated using an adaptation of themethod proposed by Hagen (2003) , in which multiple resolutionwindows are used to compare the simulated and the referencemaps within a neighbourhood context. Each type of change is ana-lysed separately using pairwise comparisons involving maps of dif-ferences: (i) between the initial land use/land cover map and asimulated one, and (ii) between the same initial land use/land cov-er map and the reference one. This modication is able to tackletwo matters. First, as it deals with only one type of change at a

time, the overall two-way similarity measure can be applied tothe entire map, regardless of the different number of cells per



5/12

category. Second, the inherited similarity between the initial andsimulated maps can be eliminated from this comparison by simplyignoring the null cells from the overall count ( Almeida et al., 2008 ).

Approaches considering neighbourhood contexts are useful incomparing maps that do not exactly match on a cell-by-cell basis,but still present similar spatial patterns within certain cell vicinity(Soares-Filho et al., 2002 ). The method retrieves a fuzzy similarity

index dened inside a window that is gradually expanded, allow-ing the assessment of model performance at multiple resolutions.This fuzzy similarity index is based on the concept of fuzzinessof location, in which a representation of a cell is inuenced bythe cell itself and by the cells in its vicinity ( Hagen, 2003 ). Thecomparison results in a map that species for each pixel the degreeof similarity on a scale of 01, where zero represents total dis-agreement and one represents identical maps.

3.4. Simulation of future scenarios

After calibration, the model is executed using a recent LULCM asan initial landscape, producing simulated landscape scenarios asoutput. In this study, the LULCM from the year 2003 was consid-

ered to be the initial landscape, and the model was applied to sim-ulate two different scenarios up to 2030. One exploratory and oneprescriptive scenario were simulated. A more detailed descriptionof each scenario is presented below:

Business as Usual (BAU) : This scenario was simulated using anexploratory approach. An exploratory scenario is a sequence of emerging events ( Alcamo, 2001 ). Here, the average agriculturalexpansion rates observed from 1987 to 2003 in the study area wereused to build an exploratory scenario with stationary behaviour forthe year 2030. Hence, in this case the agricultural expansion ratesare considered a static input for the model and are not affected bypolicy decisions.

Governance (GOV) : In this case, a prescriptive scenario was sim-ulated. Prescriptive scenarios are established a priori by the model-

ler in accordance with a targeted future ( Alcamo, 2001 ). Thus, theagricultural expansion rates were associated with ctitious gover-nance policies, in which the LUCCs were constrained according tothe availability of Renewable Freshwater Resources (RWR). RWR is dened as the water that is continuously recharged in the hydro-logical cycle. Here, it was represented by the annual average rain-fall volume.

The assumption made was that annual Irrigation WaterRequirements (IWR) could not exceed 70% of the total RWR, leav-ing the remaining 30% to be used for residential or commercialpurposes. Consequently, the agricultural expansion rates are grad-ually decreased through simulated time as the IWR would ap-proach the RWR limit. The 70% threshold was based on theglobal average distribution of water resources withdraws ( FAO,2005 ) and used as a virtual limit of water consumption. This ap-proach, however, does not represent any real policy or water man-agement strategies. The equation used to calculate the transitionrates is written as follows:

Ryab Riab IWR y 0; 7 RWR Riab

IWR i 0; 7 RWR ; 4

where Ryab is the transition rate from a to b in year y; Riab is transi-tion rates for the same land use types in the beginning of the sim-ulation; IWR y is annual average irrigation water requirementsduring year y; IWR i is annual average irrigation water requirementsin the beginning of the simulation; and RWR is the annual averagerenewable freshwater resources.

Some important concepts are central to a better understanding

of this approach. For instance, Crop Water Requirement (CWR) isdened as the amount of water required to compensate the

evapotranspiration loss from a cropped eld ( Allen et al., 1998 ),and is represented here by the Crop Evapotranspiration (ETc).Evapotranspiration (ET) is dened as the combination of two sep-arate processes, in which water is lost on the one hand from thesoil surface by evaporation and on the other hand from the cropby transpiration ( Allen et al., 1998 ).

In cases where all the water needed for optimal growth of the

crop is provided by rainfall, irrigation is not required and the Irri-gation Water Requirement (IWR) is equal to zero. In cases whereall water has to be supplied by irrigation the IWR is equal to theCWR (ETc). However, when part of the CWR is supplied by rainfalland the remaining part by irrigation, the IWR is equal to the differ-ence between the ETc and the Effective Precipitation (Peff). In suchcases, the IWR was computed using the following equation ( FAO,1997 ):

IWR m ETcm 30 Peff m ; 5

where IWR m is the monthly average crop water requirement inmonth m, [mm]; ETc m is mean daily crop evapotranspiration inmonth m , [mm day 1]; Peff m is the average effective precipitationin month m , [mm].

The Peff is dened as the fraction of rainfall retained in the rootzone, which can be effectively used by plants: that is, the portion of precipitation that is not lost by runoff, evaporation or deep perco-lation ( Brouwer and Heibloem, 1986 ). The ET was calculated usingthe Hargreaves model ( Hargreaves and Samani, 1985 ). A detailedillustration of this method, as well as the procedures used toparameterize and calibrate the ET model, are beyond the scope of this paper, and can be found in Maeda et al. (2010b) .

4. Results

4.1. Global transition rates

The overall accuracy of the 2003 land cover map was 89%, with

a Kappa index for agreement of 0.87. The class specic producersand users accuracy assessment results are shown in Table 1 . Theproducers accuracy indicates the probability of a reference pixel

Table 1

2003 Land use and land cover map classication accuracy assessment.

Land cover class # Reference testdata sites

Produceraccuracy (%)

Useraccuracy

Cropland 122 95.9 81.8%Shrubland 67 64.2 82.7%Woodland 47 91.5 91.5%Plantation forest 72 97.2 94.6%Broadleaved forest 33 97.0 100.0%Grassland 31 71.0 95.7%Bare soil 59 84.7 90.9%Built-up area 25 96.0 92.3%Water 20 100.0 100.0%

Overall accuracy 89.0%

Overall kappa index of agreement

0.87

Table 2

Annual average agricultural expansion rates (baseline 19872003).

Original vegetation Annual conversion rate (%) (baseline 19872003)

Shrubland 1.305Woodland 2.013Plantation forest 1.161Broadleaved forest 0.289Grassland 0.310



6/12

for a particular category on the map being correctly classied, andis a measure of the omission error. The users accuracy indicatesthe probability that a pixel classied on the map actually repre-sents that category on the ground, and is a measure of the commis-sion error. It can be seen that, for the most part, accuracies weregood with the croplands class, for example, having a producersaccuracy of 96% and a users accuracy of 82%. The exception was

the lower producer accuracies of the shrubland and grassland clas-ses, due to misclassication errors with certain areas of croplandwhere either shrub-like or grass-like crops gave very similar spec-tral and textural characteristics in the SPOT imagery. Because of alack of timely ground reference test data or aerial photography, theaccuracy of the 1987 classication could not be assessed directly.However, as the same classication methodology was applied to

Fig. 3. Graphics showing the W + values attributed for each range of six landscape attributes most related to the shrublands to croplands transition.



7/12

both scenes, the 1987 map accuracy was assumed to be similar tothat of the 2003 map.

The annual average agricultural expansion rates observed from1987 to 2003 are shown in Table 2 . The highest conversion rateswere observed in the transition from woodlands to agriculture.However, considering absolute numbers, shrubland areas are themost affected, given that currently it represents the predominant

vegetation type in the region. The small regions covered withbroadleaved forests were nearly untouched, presenting low con-version rates; the total area decreased from 7.7 to 6.9 km 2 duringthe observed period.

4.2. Local transitional probabilities

The most relevant W + values obtained during the model cali-bration are shown in Figs. 35 . This information represents theattraction between a determined landscape transition and a certainlandscape attribute. Several attributes were important in the con-version from shrublands to croplands: distance to rivers, insola-tion, distance to croplands, DEM, distance to roads and distant tomarkets were particularly associated with this transition ( Fig. 3).The distance to croplands is an important driving factor for all tran-sitions ( Figs. 35 ) indicating that the proximity to previouslyestablished croplands is a key factor for agricultural expansion in

this region. Areas with low insolation were indicated as havinglower probability of being converted to croplands.

Although areas close to rivers did not exhibit high positive W +values, the importance of water bodies for croplands is clearly re-ected in distance to rivers, where high negative W + values are ob-served. Hence, the results indicate that patches further than 1 kmfrom water bodies have lower probability of being converted to

cropland ( Figs. 3 and 4 ). Distance to roads also demonstrated aclear pattern in inuencing the transition from shrublands to crop-lands ( Fig. 3). That is to say, the probability of a shrubland areabeing converted to cropland linearly decreases with distance awayfrom the roads. Nevertheless, DRo did not show very high W + val-ues, possibly due to the fact that the Taita Hills includes a relativelydense road network, diminishing the contrasts between areas nearto and away from roads.

The distance to markets, here represented by the Euclidean dis-tance to the main villages, was the most representative drivingforce for shrublands to croplands transition ( Fig. 3). Shrublandspatches within up to 100 m from markets have particularly highprobability of being converted to croplands, while patches fartherthan 10 km from markets are unlikely to be used as agriculturallands.

Particular to the woodlands to croplands transition is the inu-ence of the slope parameter on the transition probability. On slopes

Fig. 4. Graphics showing the W + values attributed for each range of four landscape attributes most related to the woodland to croplands transition.



8/12

higher than 30% woodland patches are unlikely to be converted tocroplands. This can be explained by the fact that woodlands mayact as a soil conservation component in steep slopes areas, and atthe same time sloping terrains are less appropriate for agriculture.

For the transition from plantation forests to croplands, the re-sults show that areas with low precipitation (i.e. lower than550 mm/year) have higher probability of being converted ( Fig. 5).Some factors may explain this result. Firstly, the low availabilityof areas suitable for agricultures in the highlands, where the pre-cipitation rates are higher, drives the agricultural expansion tothe remaining areas with lower rainfall. Secondly, although crop-lands can, to some extent, migrate to lower precipitation areas,drier lands are not likely to be suitable for plantation forests. More-

over, most plantation forests in the Taita Hills are governmentowned, and for this reason cannot be converted to croplands.Once the driving forces of LUCC are dened, the model is itera-

tively executed and calibrated, using as an initial landscape theLULCM from the year 1987. The objective of this approach was toexecute the model to simulate the landscape observed in the year2003 and compare with the reference LULCM obtained from the sa-tellite image classication. An important parameter to be cali-brated during this process is the size and characteristics of newcropland patches. It was found that areas converted to agriculturehave an average size of 3 ha, with a standard deviation of 3.2 ha.From the new cropland patches stochastically allocated in the sim-ulations, 90% were handled by the patcher algorithm and 10% bythe expander algorithm. In actual fact this means that, on average,90% of new agricultural areas are created to be separate from exist-ing crop elds, and the remaining 10% are a result of the expansionof current elds.

The evaluation of the LUCC model performance is illustrated inFig. 6, which shows the fuzzy similarity indices achieved using dif-ferent window sizes. The maximum fuzzy similarity indices rangedfrom around 55% at a spatial resolution of 20 m to 90% at a spatialresolution of 380 m.

4.3. Scenarios simulation

After the model is calibrated and the role of each landscape var-iable dened, transition probability maps are created for each sim-ulated year. In Fig. 7, an example of a transition probability map for

the year 2003 is illustrated. In the map, the light colors representareas with higher probabilities of being converted to croplands.

Fig. 5. Graphics showing the W + values attributed for each range of two landscape attributes most related to the plantation forest to croplands transition.

Fig. 6. Fuzzy similarity indices based on multiple size windows obtained in themodel tting evaluation.

Fig. 7. Map illustrating the probability and area being converted to cropland duringthe years 2003/2004.



9/12

Once spatial probabilities are dened, the new agricultural patchesare stochastically allocated by the expander and patcher algorithms.

In Fig. 8, the historical land use maps for 1987 and 2003 are dis-played (upper left and upper right) together with the land usemaps for 2030 resulting from the BAU and GOV scenarios simula-tions (lower left and lower right). It is observed that, in 1987, crop-

lands were already clearly established in the highlands (centralarea in the maps). This is explained by the favorable climatic andedaphic conditions for agricultural activities (e.g. high precipita-tion rates), which resulted in the clearance of large areas of forestduring the last century.

The protected areas, situated in the northeast (grassland area)and southwest parts of the study area ( Fig. 8) were effective in con-taining the agricultural expansion between 1987 and 2003. Theinuence of this variable was reected in both simulated scenarios,where a clear boundary constraining the expansion of croplands isobserved in the limits of the protected areas. Hence, although ahigh environmental pressure is present along the fringes of theconservation areas, such as Tsavo East National Park, the areas in-side the park are likely to be preserved.

Between 1987 and 2003, croplands started to be implementedwith higher intensity in the lowlands, given that suitable areasfor agriculture activities in the highlands were already almost en-tirely taken. This trend is clearly reected in the LUCC simulationresults. Although each simulated scenario was created indepen-dently, using different transitions rates, the spatial distribution of new cropland patches followed the same patterns in all simula-

tions. As suitable agricultural areas in highland disappeared, theexpansion of new patches was distributed in the foothills. Amongthe main driving forces of such distribution were the distance tomarkets (here represented by villages or towns), distance to roadsand distance to rivers.

Distance to markets and roads played an interesting role incroplands distribution, in the sense that the effects of these twovariables in the landscape dynamic were closely related. Townsand villages acted as core points, which are interconnected byroads creating axes in which new cropland patches were settled.Such patterns were observed mainly in the southern and south-western parts of the study area.

It is also important to notice the enhanced importance attrib-uted to rivers in the land use dynamic. Given that in this region

Fig. 8. Historical land use maps for 1987 and 2003 (upper left and upper right) and simulated scenarios for 2030 (lower left and lower right).



10/12

the foothills typically have higher average temperatures and lowerprecipitation volumes, the proximity to water bodies is essential tothe establishment of agricultural activities.

The numerical results of the simulations are presented in Fig. 9aand b, which display the percentage of the main land use/land cov-er classes during the years analyzed, and total cropland areas in thestudy site from 2010 to 2030, respectively. In the BAU scenario, thecropland areas expanded to around 515 km 2 in 2030, correspond-ing to about 60% of the study area ( Fig. 9a). This represents an in-crease of 40% in comparison to the year 2003, when croplandsoccupied around 365 km 2. Although the effects of the governancescenario cannot be easily identied in the map showed in Fig. 8,the simulated land use policies resulted in a signicant reductionin agricultural expansion ( Fig. 9b). The total area used for agricul-ture in 2030 for the GOV scenario was approximately 485 km 2 .

5. Discussion

In contrast to studies aiming to assess the social and economicalcomponent of LUCC, this study focused on the inuence of physicallandscape attributes and infrastructure components on the spatialdistribution patterns of agricultural expansion. Although socio-economic studies are helpful to better understand the uctuationsand tendencies in land change rates ( Rudel, 2009 ), they may beinadequate to numerically analyse the spatial probabilities of changes and the consequent development of plausible landscapescenarios.

The results obtained in the LUCC driving forces analysis closelyagree with previous studies carried out in other locations in Kenya.

For instance, studying the Narok District in Kenya, Serneels andLambin (2001) found that the expansion of small holder agricul-ture is mainly controlled by proximity to permanent water, landsuitability and vicinity to villages. Moreover, the authors found dif-ferences in the factors driving mechanized agriculture, where con-versions are mainly driven by distance to the markets and agro-climatic potential.

In another similar work, Mertens and Lambin (2000) designed aspatially explicit model driven by a spectrum of socio-economicand infrastructure variables to simulate deforestation in SouthernCameroon. According to the authors, in this study case the resultssuggested that roads mostly increased the accessibility of the for-est for migrants rather than providing incentives for the establish-ment of market oriented farming systems.

Therefore, the comparison with previous studies highlights thefact that, although many similarities can be found at large scales,

some characteristics of the landscape dynamic are intrinsic, andcan only be assessed locally. According to Lambin et al. (2003) , landchanges are in general caused by multiple interacting factors,which vary in time and space, according to specic human-envi-ronment conditions. In this context, similarities and disparitiesare also observed when comparing the driving forces of LUCC inthe Taita Hills with agricultural expansion areas in other tropicalforests. For instance, in the rain forests of South America, distanceto roads and distance to markets arise as common factors affectingcroplands expansion ( Aguiar et al., 2007 ). In agreement with thepresent study, Alves (2002) indicates that new agricultural patchesin the Brazilian Amazon spring up predominantly in the surround-ings of pioneer areas. On the other hand, in contrast to the indus-trial agricultural activities in the Brazilian Amazon, new croplandpatches in the Taita Hills are predominantly small, targeting sub-sistence or local production. New cropland patches in the TaitaHills were found to have an average size of 3 ha, with a varianceof 10 ha, while in the agricultural expansion areas in So Feliz doXingu, Brazilian Amazon, new agricultural patches have an averagesize of 300 ha, with a standard deviation of 22.3 ha ( Ximenes et al.,2008 ).

The performance achieved in the LUCC model calibration is con-sidered satisfactory. In the multiple resolution tting procedure,the simulations achieved spatial ttings from 75%, at a spatial res-olution of 100 m, up to 90% at a spatial resolution of 380 m. Suchresults are consistent with previous studies carried out with Din-amica-EGO. For instance, aiming to simulate the landscape dynam-ics in an Amazonian colonization frontier, Soares-Filho et al. (2002)calibrated the model for two study sites, reaching average spatialtting values between 63.3% and 82.4% at a spatial resolution of

100 m. In another similar study, the CLUE-S model was appliedto simulate deforestation in the Klang-Langat watershed, Malaysia(Verburg et al., 2002 ). The study evaluated the deforestationprocess as a function of socio-economic and biophysical drivingfactors and also considered driving forces related to the urbaniza-tion process, such as the accessibility of the area by different roadtypes. In this study case, the model t ranged from approximately65% at a window size of 150 m, to 84% with a window of 10.5 km.

Although the remaining biodiversity hotspot forests in the TaitaHills are likely to be preserved, the simulated scenarios clearlyindicate a tendency towards the continuity of habitat fragmenta-tion caused by agricultural expansion. It is known that habitat frag-mentation can signicantly reduce species richness and changeecosystem functioning ( Flynn et al., 2009 ). Likewise, the destruc-

tion of natural habitats may signicantly increase the dependenceof species on protected areas ( Jackson and Gaston, 2008 ). Studying

Fig. 9. (a) Percentage of the main land use/land cover classes during the years 1987, 2003 and in both scenarios simulated for 2030; (b) LUCC simulation results from 2010 to2030.



11/12

the effects of habitat fragmentation on seed dispersal of an interiortree in the Taita Hills, Lehouck et al. (2009) show that forest dete-rioration reduced avian visitation and seed removal rates, creatinglong-term effects on tree communities. In this context, Hall et al.(2009) state that in order to ensure the efcacy of conservation ef-forts in the EAM hotspots, one needs to consider the extent of hab-itat changes within and across the entire elevational ranges found

in these ecosystems.The agricultural expansion simulated in the presented studywill have direct impacts on soil and water conservation issues bythe year 2030. The natural vegetation protects the soil againstthe impacts of rainfall and provides organic matter to the soil.These factors improve water inltration and recharging of ground-water reservoirs. When this vegetation cover is displaced the soilbecomes vulnerable to compaction and the water that wouldotherwise have inltrated the soil is turned into surface runoff,which can potentially carry out sediments and nutrients to rivers(Maeda et al., 2009 ). Such changes in the landscape dynamic canpotentially cause environmental problems such as erosion, siltingof rivers, eutrophication, and water contamination, among others(Maeda et al., 2008, 2010a ).

The benets of increasing agricultural areas for the local econ-omy need to be carefully analyzed. Although the local economyis mainly based on agricultural activities, the results of this studyshow that croplands are currently expanding throughout the foot-hills, that is to say, in the direction of areas with lower precipita-tion and higher temperatures. This dynamic will likely lead to ahigher water resources demand for irrigation ( Maeda, 2009 ), hav-ing the potential to generate water conicts. In addition, Thorntonet al. (2010) warn that, due to continuing population increases, to-gether with climate change, agricultural systems may have to un-dergo substantial intensication during the next decades.Therefore, alternative solutions, such as the implementation of drought resistant crops, should be explored in order to meet agri-cultural production requirements. Such approaches can increasecrop yield and potentially spare land for wild nature ( Ewerset al., 2009 ).

Hence, agricultural expansion will likely increase land pressure,having direct and indirect consequences on the characteristics of farming systems. For instance, Fermont et al. (2008) show that,during the last few decades, increasing populations and loweravailability of arable land contributed to signicant changes inthe crops cultivated in parts of mid-altitude zones of East-Africa.Traditional farming systems based on millet, cotton, sugarcaneand/or banana have evolved into continuously cultivated cassavaor cassava/maize-based systems. The authors state that the declin-ing soil fertility and food shortage is likely the primary trigger forthis transformation, given that cassava can recycle nutrients andtolerate poor soils.

In this sense, the scenarios simulated in this study can offer a

valuable tool to guide informed policy decisions in the directionof the improvement of agricultural systems, land use allocationand environmental protection. An immediate and straightforwardapplication of the results of this study will be in supporting localauthorities and researchers in implementing new water qualitycontrol and stream ow measuring stations. Water quality mea-surement stations aiming to assess the impacts of agriculturalexpansion are expected to be implemented during the comingyears. The results presented here can elucidate the priority regionswhere new croplands are likely to occur and, consequently, theapproximate location where measurements need to be taken in or-der to achieve consistent results.

The use of scenarios to improve agricultural systems strategieshas been extensively reported in the literature. Such studies dem-

onstrate the importance of this approach in simulating majordirections of agricultural development, alternative strategies for

soil conservation, irrigation and land use allocation. The presentstudy represents an important contribution in this context. It canbe considered an additional step in the direction of lling an evi-dent gap in the understanding of environmental changes at localscales in Eastern Africa.

6. Conclusions

The research presented here was able to identify the main land-scape attributes driving the agricultural expansion in the TaitaHills. A connected relationship between villages and roads is evi-dent in the denition of new cropland patches. Proximity to al-ready established crop elds is also one of the key factors drivingagricultural expansion.

If current trends persist, it is expected that agricultural areaswill occupy 60% of the study area by 2030. LUCC simulations indi-cate that agricultural expansion will likely take place predomi-nantly in lowlands and foothills throughout the next 20 years.Such dynamics will increase the spatial dependence on distanceto rivers and other water bodies due to the higher potential evapo-transpiration in these areas. On the other hand, given that the mostfavorable areas for agriculture are already taken, agro-climatic fac-tors (e.g. precipitation) will have a decreased inuence on the spa-tial distribution of croplands.

Current trends indicate that the small residual areas of tropicalcloud forest, home for a great part of the biodiversity in the TaitaHills, will likely remain intact throughout the coming years. Never-theless, the impact of increasing habitat fragmentation on suchbiodiversity, caused by the agricultural expansion discussed in thispaper, is a relevant issue that must be addressed in furtherresearch.

The results described in this study have good potential to beused by policy makers in improving the identication of priorityregions from the point of view of land use allocation and environ-mental risks. Moreover, the integrated modelling presented repre-

sents an important tool for researchers to understand the human-environment relations in this region.

Acknowledgements

Part of this research was carried out during the Young ScientistsSummer Program (YSSP) at the International Institute for AppliedSystem Analysis (IIASA), Austria. The authors kindly thank the IIA-SAs researchers Dr. Marek Makowski, Dr. David Wiberg and Dr.Tatiana Ermolieva for their comments. Thanks are also given toDr. Cludia Maria de Almeida from the National Institute for SpaceResearch (INPE), Brazil, for her contribution in improving part of the methodological descriptions presented in this paper. The re-search was funded by the Centre of International Mobility (CIMO),University of Helsinki and Academy of Finland for TAITATOOproject.

References

Aguiar, A.P.D., Cmara, G., Escada, M.I.S., 2007. Spatial statistical analysis of land-usedeterminants in the Brazilian Amazonia: exploring intra-regionalheterogeneity. Ecological Modelling 209 (24), 169188.

Alcamo, J., 2001. Scenarios as tools for international environmental assessments. In:Experts Corner Reports Prospects and Scenarios No 5. European EnvironmentAgency, Copenhagen.

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration.Guidelines for Computing Crop Water Requirements. FAO Irrigation andDrainage Paper 56. FAO, Rome.

Almeida, C.M., Batty, M., Monteiro, A.M.V., Camara, G., Soares-Filho, B.S., Cerqueira,G.C., Pennachin, C.L., 2003. Stochastic cellular automata modeling of urban landuse dynamics: empirical development and estimation. Computers,Environment and Urban Systems 27 (5), 481509.



12/12

Almeida, C.M., Gleriani, J.M., Castejon, E.F., Soares-Filho, B.S., 2008. Using neuralnetworks and cellular automata for modeling intra-urban land use dynamics.International Journal of Geographical Information Science 22 (9), 943963.

Alves, D.S., 2002. Spacetime dynamics of deforestation in Brazilian Amazonia.International Journal of Remote Sensing 23 (14), 29032908.

Batjes, N.H., Gicheru, P., 2004. Soil data derived from SOTER for studies of carbonstocks and change in Kenya (GEF-SOC Project; Version 1.0). Technical Report2004/01. ISRIC World Soil Information, Wagningen.

Bonham-Carter, G., 1994. Geographic Information Systems for Geoscientists:Modelling with GIS. Elsevier Science, Pergamon, Kidlington.

Brink, A.B., Eva, H.D., 2009. Monitoring 25 years of land cover change dynamics inAfrica: a sample based remote sensing approach. Applied Geography 29 (4),501512.

Brooks, T., Lens, L., De Meyer, M., Waiyaki, E., Wilder, C., 1998. Avian biogeographyof the Taita Hills, Kenya. Journal of East African Natural History 87, 189194.

Brouwer, C., Heibloem, M., 1986. Irrigation Water Management. Training Manual(FAO), vol. 3. Prov. ed. FAO, Rome, Italy (60p).

Burgess, N.D., Butynski, T.M., Cordeiro, N.J., et al., 2007. The biological importance of the Eastern Arc Mountains of Tanzania and Kenya. Biological Conservation 134(2), 209231.

Burnett, C., Blaschke, T., 2003. A multi-scale segmentation/object relationshipmodelling methodology for landscape analysis. Ecological Modelling 168 (3),233249.

Clark, B.J.F., Pellikka, P.K.E., 2009. Landscape analysis using multiscale segmentationand object orientated classication. In: Rder, A., Hill, J. (Eds.), Recent Advancesin Remote Sensing and Geoinformation Processing for Land DegradationAssessment. ISPRS Book Series in Photogrammetry, Remote Sensing andSpatial Information Sciences, vol. 8. Taylor & Francis Group, pp. 323342.

Di Gregorio, A., 2005. Land Cover Classication System (LCCS), Version 2:Classication Concepts and User Manual. FAO Environmental and NaturalResources Series 8, FAO, Rome, 208pp.

Eva, H.D., Brink, A., Simonetti, D., 2006. Monitoring land cover dynamics in sub-Saharan Africa. In: A Pilot Study Using Earth Observing Satellite Data from 1975and 2000. Institute for Environment and Sustainability, Joint Research Centre of the European Commission, p. 44.

Ewers, R.M., Scharlemann, J.P.W., Balmford, A., Green, R.E., 2009. Do increases inagricultural yield spare land for nature? Global Change Biology 15 (7), 17161726.

FAO Food and Agriculture Organization of the United Nations, 1997. Land andWater Development Division. Irrigation potential in Africa: A basin approach,FAO, 177p.

FAO Food and Agriculture Organization of the United Nations, 2005. Land andWater Development Division. AQUASTAT Information System on Water andAgriculture: Online database. FAO, Rome.

Fermont, A.M., van Asten, P.J.A., Giller, K.E., 2008. Increasing land pressure in EastAfrica: the changing role of cassava and consequences for sustainability of farming systems. Agriculture, Ecosystems and Environment 128 (4), 239250.

Flynn, D.F., Gogol-Prokurat, M., Nogeire, T., Molinari, N., Richers, B.T., Lin, B.B.,Simpson, N., Mayeld, M.M., DeClerck, F., 2009. Loss of functional diversityunder land use intensication across multiple taxa. Ecology Letters 12 (1), 2233.

Hagen, A., 2003. Fuzzy set approach to assessing similarity of categorical maps.International Journal of Geographical Information Science 17 (3), 235249.

Hall, J., Burgess, N., Lovett, J., Mbilinyi, B., Gereau, R.E., 2009. Conservationimplications of deforestation across an elevational gradient in the Eastern ArcMountains, Tanzania. Biological Conservation 142, 25102521.

Hargreaves, G.H., Samani, Z.A., 1985. Reference crop evapotranspiration fromtemperature. Applied Engineering in Agriculture 1 (2), 9699.

Himberg, N.L.M.A.Omoro., Pellikka, P., Luukkanen, O., 2009. The Benets andConstraints of Participation in Forest Management, vol. 187. The case of TaitaHills, Kenya. Fennia, pp. 6176.

Hutchinson, M.F., 1995. Interpolating mean rainfall using thin plate smoothingsplines. International Journal of Geographical Information Science 9, 305403.

Jackson, S.F., Gaston, K.J., 2008. Land use change and the dependence of national priority species on protected areas. Global Change Biology 14 (9),21322138.

Koomen, E., Rietveld, P., Nijs, T.de., 2008. Modelling land-use change for spatialplanning support. The Annals of Regional Science 42 (1), 110.

Lambin, E.F., Geist, H.J., Lepers, E., 2003. Dynamics of land-use and land-coverchange in tropical regions. Annual Review of Environment and Resources 28,205241.

Lehouck, V., Spanhove, T., Colson, L., Adringa-Davis, A., Cordeiro, N.J., Lens, L., 2009.Habitat disturbance reduces seed dispersal of a forest interior tree in afragmented African cloud Forest. Oikos 118 (7), 10231034.

Maeda, E.E., 2009. Impacts of agricultural expansion on irrigation waterrequirements in Taita hills, Kenya. Interim Report (IR-09-056) InternationalInstitute for Applied Systems Analysis, Laxenburg, Austria.

Maeda, E.E., Formaggio, A.R., Shimabukuro, Y.E., 2008. Impacts of land use and landcover changes on the sediment yield in Brazilian Legal Amazon drainage basin.Geoscience & Remote Sensing 45, 443453.

Maeda, E.E., Formaggio, A.R., Shimabukuro, Y.E., Kaleita, A.L., 2009. Impacts of agricultural expansion on surface runoff: a case study of a river basin in theBrazilian Legal Amazon. International Journal of Geoinformatics 5 (3), 3341.

Maeda, E.E., Pellikka, P.K.E., Siljander, M., Clark, B.J.F., 2010a. Potential impacts of agricultural expansion and climate change on soil erosion in the Eastern ArcMountains of Kenya. Geomorphology, doi:10.1016/j.geomorph.2010.07.019 .

Maeda, E.E., Wiberg, D.A., Pellikka, P.K.E., 2010b. Estimating referenceevapotranspiration using remote sensing and empirical models in a regionwith limited ground data availability in Kenya. Applied Geography. doi:10.1016/j.apgeog.2010.05.011.

Mertens, B., Lambin, E.F., 2000. Land-cover-change trajectories in SouthernCameroon. Annals of the Association of American Geographers 90 (3), 467494.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A.B., Kent, J., 2000.Biodiversity hotspots for conservation priorities. Nature 403 (6772), 853858.

Pellikka, P.K.E., Ltjnen, M., Siljander, M., Lens, L., 2009. Airborne remote sensing of spatiotemporal change (19552004) in indigenous and exotic forest cover inthe Taita Hills, Kenya. International Journal of Applied Earth Observations andGeoinformation 11 (4), 221232.

Republic of Kenya, 2001. The 1999 Population & Housing Census. Central Bureau of Statistics. Ministry of Planning and National Development, Kenya.

Rudel, T.K., 2009. How do people transform landscapes? A sociological perspectiveon suburban sprawl and tropical deforestation. American Journal of Sociology115 (1), 129154.

Serneels, S., Lambin, E.F., 2001. Proximate causes of land-use change in Narokdistrict, Kenya: a spatial statistical model. Agriculture, Ecosystems andEnvironment 85 (13), 6581.

Soares-Filho, B.S., Pennachin, C.L., Cerqueira, G., 2002. DINAMICA a stochasticcellular automata model designed to simulate the landscape dynamics in anAmazonian colonization frontier. Ecological Modelling 154 (3), 217235.

Soares-Filho, B.S., Garcia, R.A., Rodrigues, H.O., et al., 2007. Coupling socioeconomicand demographic dimensions to a spatial simulation model of deforestation forthe Brazilian Amazon. Paper presented at the LBA-ECO 11th Science TeamMeeting.

Soares-Filho, B.S., Rodrigues, H.,Costa, W., 2009. Modeling Environmental Dynamicswith Dinamica EGO. ISBN: 978-85-910119-0-2 < www.csr.ufmg.br/dinamica >.

Soini, E., 2005. Livelihood capital, strategies and outcomes in the Taita Hills of Kenya. In: ICRAF Working Paper No. 8. World Agroforestry Centre, Nairobi,Kenya, p. 48.

Thornton, P.K., Jones, P.G., Alagarswamy, G., Andresen, J., Herrero, M., 2010.Adapting to climate change: agricultural system and household impacts inEast Africa. Agricultural Systems 103 (2), 7382.

Turner, B.L., Lambin, E.F., Reenberg, A., 2007. The emergence of land change sciencefor global environmental change and sustainability. Proceedings of the NationalAcademy of Sciences 104, 2066620671.

Veldkamp, T., Lambin, E.F., 2001. Predicting land-use change. Agriculture,Ecosystems & Environment 85 (13), 16 (Editorial, Special issue).

Verburg, P.H., Soepboer, W., Limpiada, R., Espaldon, M.V.O., Sharifa, M.A., Veldkamp,A., 2002. Modelling the spatial dynamics of regional land use: the CLUE-Smodel. Environmental Management 30, 391405.

Washington, C., Pijanowski, B., Campbell, D., Olson, J., Maitima, J., Musili, A., Kibaki,T., Kaburu, H., Hayombe, P., Owango, E., Irigia, B., Gichere, S., Mwangi, A., 2010.Using a role-playing game to inform the development of land-use models forthe study of a complex socio-ecological system. Agricultural Systems 103 (3),117126.

Ximenes, A.C., Almeida, C.M., Amaral, S., et al., 2008. Dynamic deforestationmodeling in the Amazon. Boletim de Cincias Geodsicas 14, 370391 (inPortuguese).

http://dx.doi.org/10.1016/j.geomorph.2010.07.019http://www.csr.ufmg.br/dinamicahttp://www.csr.ufmg.br/dinamicahttp://dx.doi.org/10.1016/j.geomorph.2010.07.019

Date post:	04-Jun-2018
Category:	Documents
Upload:	jaramillo-rojas-isaac
View:	221 times
Download:	0 times

Agricultural System

Documents