A Framework for the Land Use Change Dynamics Model ... · 2 AdvancesinMeteorology human-landscapesystemandfeedsbacktothesubsequent developmentoftheseinteractions[14].Mostlandusechange

Hindawi Publishing CorporationAdvances in MeteorologyVolume 2013 Article ID 658941 7 pageshttpdxdoiorg1011552013658941

Research ArticleA Framework for the Land Use Change Dynamics ModelCompatible with RCMs

Xiangzheng Deng12 Jiyuan Liu1 Yingzhi Lin3 and Chenchen Shi14

1 Institute of Geographic and Natural Resources Research Chinese Academy of Sciences Beijing 100101 China2 Center for Chinese Agricultural Policy Chinese Academy of Sciences Beijing 100101 China3 School of Mathematics and Physics China University of Geosciences (Wuhan) Wuhan 430074 China4 State Key Laboratory of Water Environment Simulation School of Environment Beijing Normal University Beijing 100875 China

Correspondence should be addressed to Xiangzheng Deng dengxzccapgmailcom

Received 22 August 2013 Accepted 12 November 2013

Academic Editor Burak Guneralp

Copyright copy 2013 Xiangzheng Deng et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

A framework of land use change dynamics (LUCD) model compatible with regional climate models (RCMs) is introduced in thispaper The LUCD model can be subdivided into three modules namely economic module vegetation change module and agent-based module The economic module is capable of estimating the demand of land use changes in economic activities maximizingeconomic utility A computable general equilibrium (CGE) modeling framework is introduced and an approach to introduce landas a production factor into the economic module is proposedThe vegetation change module provides the probability of vegetationchange driven by climate change The agroecological zone (AEZ) model is supposed to be the optimal option for constructing thevegetation change module The agent-based module identifies whether the land use change demand and vegetation change can berealized and provides the land use change simulation results which are the underlying surfaces needed by RCM By importing theRCMsrsquo simulation results of climate change and providing the simulation results of land use change for RCMs the LUCD modelwould be compatible with RCMsThe coupled simulation system composed of LUCD and RCMs can be very effective in simulatingthe land surface processes and their changing patterns

1 Introduction

There are two primary factors that contribute to climatechange land use change and greenhouse gas emission [1 2]Land use change which has been found to affect climatechange in both biogeochemical and biogeophysical ways isfundamentally important for the researches of regional cli-mate change [3] In regional climate modeling land use dataare applied as underlying surfaces and definitively determinethe simulation results of regional climate [4] Many simula-tion experiments have proven that the simulation results ofRCMs are sensitive to underlying land use and land coverchanges (LUCC) [5 6] While the interaction between landuse change and climate change has been fully realized mostRCMs introduce LUCC data exogenously [7 8] Always theyapply the LUCC data of one year of history as underlying sur-faces and keep them constant ignoring the interaction

between LUCC and climate variations This paper provides aframework of land use change dynamics (LUCD)model com-patible with RCMs to introduce parameterized LUCC intoregional climate changemodeling endogenously Several sug-gested models are introduced and some specific parameterprocessing approaches are explained in detail This modelingframework helps to enhance the understanding of the cou-pling mechanism of land use system and climatic system andstrengthen the simulation capability of land system

Land system is geographically complex which is com-posed of natural factors human land-use activities and otherimpact factors [9ndash11] Land use change simulation is a predic-tion of when where why and how land use pattern changes[12 13] However studies on land use change processes areoften challenged by the complex and unexpected human act-ivities and natural constraints Land use change emerges fromthe interactions among various components of the coupled

2 Advances in Meteorology

human-landscape system and feeds back to the subsequentdevelopment of these interactions [14] Most land use changesimulation models simulate successional pattern change ofland use under the macrobackground of the regional popula-tion growth economic development social progress changesin the natural environment and other facts [15 16] On thewhole the land use change simulationmodels can be broadlydivided into three major categories empirical statisticalmodel agent-based model (ABM) and raster neighborhoodrelationship based model [17]

There are abundant empirical statisticalmodels applied toland use change simulation A typical example is the Conver-sion of land use and its effect at small regional extent (CLUE-S) model whose application in land use change simulation iscurrently in the ascendant [18ndash20]TheCLUE-Smodel is con-structed to simulate land use change and its effects on enviro-nment at mesomicroscale It has the capability of synchro-nously simulating the changes of multiple types and intro-duces the dynamic driving factors (such as population andeconomic growth) to improve the simulation accuracy Sincethe 1990s along with the rapid development of complexity science ABM began to be applied to land use change research[21] The agent-based models of land use and land cover(ABMLUCC) were specially discussed by LUCC reportnumber 6 inwhich the development prospect ofABM in landuse change simulation is highly valued [22] An ABM modelmainly identifies the linkage between agents and environ-ment by describing the interaction and affiliation of indepen-dent agents [23] By combining ABM and cellular automata(CA) model the simulation of land use change is character-ized by multiscale and becomes more effective in multiobjec-tive decisionmaking Semboloni et al [24] established amul-tiagent system with the residents industry practitioners ser-vice practitioners and developers as themain agents to simu-late the land use change in urban areas Zhang et al [25] builtan ABM to simulate the impacts of the climate changes on theregional land use changes and agentsrsquo economic benefits intheThree-River Headwaters Region of China It is found thatthe agents would get more wealth under the scenario withoutclimate changes in the long term even though the totalincome is lower than that under the scenario with climatechanges As a representative of raster neighborhood relation-ship basedmodel CAmodel iswidely used in landuse changesimulation especially urban expansion [26 27] Syphard et al[28] analyzed the distinction of LUCC caused by urbanexpansion in areas with different slope with the CA modelOne of the superiority of the CA model in land use changesimulation is that it supports visualization of the simulationprocess The structure of the CA model makes it difficultconsidering the impacts of land use policies

In this study we developed an LUCD model com-patible with RCMs to describe the interdependencies andfeedback mechanisms between social economics ecosystemenvironment and irrational decision-making process TheLUCD model describes a combined and complex systemcomposed of social economic ecosystem components anddecision-making process and consequences It provides aconsistent and comprehensive framework of land use changemodeling and emphasis on how themodels work together By

introducing agroecological zone (AEZ) based on the sim-ulation results of RCMs the LUCD model is compatiblewith RCMs and constitutes an iterative simulation system ofLUCC and climate changes

2 Land Use Change Dynamics (LUCD) Model

21 Model Structure The LUCD model can be subdividedinto three modules namely economic module vegetationchange module and agent-based module The economicmodulcalculates the area demand for all land use types ineconomic activitiesmaximizing economic utility of land usesThe vegetation change module provides the probability ofvegetation change driven by climate changes And the agent-based module identifies whether the land use demand andvegetation change can be realized and provides the land usechange simulation results which are the underlying surfacesneeded by RCM To feed the LUCDmodel results into RCMsthe land use system applied to the LUCD model should beconsistent with the underlying surfaces used in RCMs(Figure 1) By iteratively using the output of one model as theinput of another the LUCDmodel is compatible with RCMsIn Figure 1 the dotted lines show the data transmission be-tween the LUCDmodel andRCMswhile the solid lines standfor the flow of information in the LUCD model The LUCDmodel provides the simulation results of land use change forRCMs as underlying surface data then RCMs can simulatethe climate change resulted from the land use change Theresults of climate change simulated by RCMs are furtherimported into the LUCD model and affect land use change

The economic module estimates the land use change de-mand driven by human activityThe current condition of landuses is introduced in this module as one of the limitations ofeconomic activity as well as land use decisions The equilib-rium of markets determines the commodity supply and inturn influences the land use demand Combining the land usedemand the limited amount of land and the current land usestatus the land use change demand is obtained The vegeta-tion change module describes the possible vegetation changedriven by climate change The AEZ is the key concept thatlinks the climate change and vegetation change and helps tocouple human activity with climate change The climatechange leads to change of AEZs which determines the growthof vegetation [29 30] Consequently the climate changeaffects not only the evolution of natural vegetation but alsothe human activities including planting and breeding Byoverlying the AEZs on the current vegetation pattern thesuitability of vegetation change can be evaluated The agent-based module describes the procedure of land use decisioncoupled with the land use change demand and vegetationchange suitability using the agent-based simulation technol-ogy This module identifies whether or not the theoreticalland use change demand and the possible vegetation changeestimated by the economicmodule and the vegetation changemodule can be realized The output of this module land usechange is the underlying surfaces that needed by RCMs Byembedding the LUCD in RCMs an iterative simulation sys-tem of land use change and climate change is constructed(Figure 1)

Advances in Meteorology 3

Agent-based land use decision

Agroecologicalzone change

Macroeconomic equilibrium

Human activity

Land use change demand

Climate change

Vegetationcharacter

Vegetation change probability

Vegetation suitability

Land use change

Land use status

Regional climate model (RCM)

Land

use

clas

sifica

tion

syste

m

Land

use

clas

sifica

tion

syste

mFigure 1 Framework of feeding LUCC model results into RCMssimulation

22 EconomicModule Theeconomicmodule should providea comprehensive macroeconomic framework to describemarket-oriented economies Computable general equilib-rium (CGE) models are suggested to be appropriate for sucha macroeconomic framework [31] For convenient applica-tion the way that induces land into the economic moduleunder a CGEmodeling framework is proposed as well in thisstudy (Figure 2) Land is one of the three primary factorsinput in commodity production And there are five compo-nents producers households government trade and mar-kets in CGE model [32] Producers decide demand of inputsincluding primary factors of land labor and capital and sup-ply of outputs (commodities) to maximize their profitsHouseholds decide demand of commodities and supply oftheir endowments of labor and capital to maximize their eco-nomic utility Government imposes taxes and expends themin public consumption and savings The savings of govern-ment and households transform into investment according toreserve requirements which is also an important componentin demand And we employ the small-country assumptionthat the study area is too small to affect prices in internationalmarkets Thus import and export prices which this countryfaces are given for it in foreign currency terms The demandand supply of commodities and primary factors are equilibr-ated in markets by price adjustment With this module wecan compute land uses in various equilibria to simulate whatwill happen in the future

ThoughCGEmodels are good at describing the quantitiesand prices variation as others we do not introduce land pricesbut land area in the economic module This is because theland prices vary along with not only time but also locationproductivity and so forth And as a macroeconomic modelCGE model does not support a diverse prices modelingframework Thus we summarize land uses in economicdevelopment as follows

119884119894119890= 119887119894119890prod

ℎ

119865120573ℎ119894119890

ℎ119894119890prod

119897

119860119890119888120589119897119894119890

119897119894119890 (1)

Total

supplyland

Other factorsFactor market

Incomes

Domestic output

Commodity supply

Commoditydemand

Import

Export

Commodity market

Total utility

zone changeAgroecological

Figure 2 Overview flow chart of economicmodule applying a CGEmodeling framework

119860119890119888119897119894119890=120577119897119894119890119884119894119890

119887119894119890

(2)

where 119894 is the index of commodities 119890 is the index of AEZs 119897is the index of land use type ℎ is the index of primary factors(labor and capital)119884

119894119890is the value added of the 119894th firm in the

119890th AEZ 119860119890119888119897119894119890

is the input area of the 119897th land use type for119894th commodity production in the 119890th AEZ119865

ℎ119894119890is the input of

the ℎth factor by the 119894th firm in the 119890th AEZ 119887119894119890is the scaling

parameter in production function also called total factor pro-ductivity (TFP) 120577

119897119894119890is the share parameter in production

functions and 120573ℎ119894119890

is the share parameter in productionfunctions Considering that the value added is proportionalto the input land area under the certain technique conditionand primary factors input the input area of each land use typeis calculated by (2)

The input land area of each type of land use per unit ofeach commodity output is inversely correlated with primaryfactors input besides TFP Consequently the share parameter120577119897119894119890

is determined by the input of the ℎth factor by the 119894thfirm in the 119890th AEZ 119865

ℎ119894119890

120577119897119894119890= 119891 (119865labor119894119890 119865capital119894119890) (3)

The area demand of lands input in commodity produc-tion is determined by the economic system because the eco-nomic links in the comprehensive macroeconomic frame-work provided byCGEmodel are tightly connectedwith eachother Each shock to economic system will influence the areademand of lands input in commodity production For exam-ple the growth in the rate of direct tax will lead to an increasein government revenues and a decrease in household incomeThen the structure differences of investments and consump-tions between government and household determine thechange of commodity demand structure Under the market-clearing condition the commodity production and supplystructure should be altered And finally the area demand oflands input in commodity production will change

As most of the economic models the economic moduleassumes that the ultimate purpose of economic developmentis to increase the economic utility of household Householdrsquos


economic utility is dependent on the amount of consumptionof commodity which are purchased from producers

119880119890119888 = prod

119894

119883119901120594119894

119894 (4)

where 119880119890119888 is the economic utility 119883119901119894is the amount of con-

sumption of the 119894th commodity and 120594119894is the share parameter

in the economic utility functionThe economic utility is indirectly restrained by the area of

land used for economic development In reality the earningsfrom endowments of land are the component householdrsquosincome which is the constraint of household consumptionAnd the input of land in production by producer determinesthe output and supply of commodity Nevertheless we onlyconsider the constraint function of land in production in thismodule because the earnings from the endowment of land arenot included when accounting income constraints The eco-nomic development can be summarized as the following opti-mization problem

maximize119883119901119894

119880119890119888 = prod

119894

119883119901120594119894

119894

subject to 119879119897119886119899119889 = sum

119890

sum

119897

sum

119894

119860119890119888119897119894119890

(5)

where 119879119897119886119899119889 is the total land area which is exogenouslydefined Equation (4) shows the objective function of eco-nomic utility to bemaximized and (5) is a total land area con-straint equation meaning that total land areas used for com-modity production must be equal to the total land area usedin economic activity on the left-hand side of the equationThesimulated land use change demand at regional scale can beallocated to grids by using DLS (dynamics of land systems)model CLUE-S model or CA models [33ndash36] and so forth

23 Vegetation Change Module The vegetation change mod-ule assesses the growth suitability of specific vegetation andprovides the possibility of vegetation changeThere are manymodels including dynamic global vegetation model [37 38]Holdridge life zone model [39] and AEZmodel [40] that canbe used to describe the vegetation change driven by climatechange In this study we propose AEZ model as the optimaloption because it is naturally correlatedwithAEZs facilitatingthe coupling of economic module and vegetation changemodule We also illustrate how to estimate the possibility ofvegetation change usingAEZmodelTheAEZmodel is devel-oped by Food and Agriculture Organization (FAO) of theUnitedNations with the collaboration of the International In-stitute for Applied Systems Analysis (IIASA) [41] Climatetopography and soil characteristics are three key inputs of theAEZ model The model can estimate the climate limited veg-etation productivity Assuming that the estimated climatelimited productivity of the Vth type of vegetation in the pixel119901in the 119905th year is 119884V119901119905 the possibility of vegetation change ofthe Vth type of vegetation in the pixel 119901 in the (119905 + 1)th year is

119875V119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884Vmax (6)

where 119884Vmax is the maximum climate limited productivity ofthe Vth type of vegetation and 119875V119901119905+1 is the possibility of veg-etation change of the Vth type of vegetation in the pixel 119901 inthe (119905 + 1)th year

A positive119875V119901119905+1 implies that the Vth type of vegetation inthe pixel 119901 will expand or be more thickly forested (119905 + 1)thyear while a negative 119875V119901119905+1 means that the Vth type of veg-etation in the pixel 119901 will be inclined to degrade in the in the(119905+1)th yearThepossibility of vegetation change provides thecomparison criterion of specific vegetation change of differ-ent pixels in different timesWhen comparing the superiorityof different vegetation in the specific pixel and time a super-iority index 119878V119906119901119905 is proposed

119878V119906119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884119906119901119905+1

minus 119884119906119901119905

(7)

where 119878V119906119901119905+1 is the superiority index of the Vth type of veg-etation compared with the 119906th type of vegetation in the pixel119901 in the (119905 + 1)th year

The superiority index cannot depict the dominance rela-tions between two types of vegetation by itself The applica-tion of this index should be combined with the possibility ofvegetation change For instance when 119875V119901119905+1 is positive and119878V119906119901119905+1 is larger than 1 the Vth type of vegetation is moresuperior than the 119906th type of vegetation in the pixel 119901 in the(119905+1)th year Amore exact mathematical formula for judgingthe dominance relations ofmultiple types of vegetation is pro-posed in the agent-based module

24 Agent-Based Module Determination of land use changeis partly characterized by nonrationality such as tradition andcustom The agent-based module identifies whether the landuse change demand simulated by economic module and thepossible vegetation change assessed by vegetation changemodule can be realized under the background of irrationaldecisions Agent-based modeling is able to simulate land usechange by measuring the individual behavior and results ofland use over time [42] Take the decision of land use changeof a given household for instanceThe dissimilarities betweena given household ℎ and all defined household groups in thepopulation can be measured

119863ℎ119892=

119878

sum

119904=1

119908119904[

[

(119881ℎ119904minus 119881119892119904)2

10038161003816100381610038161003816119881ℎ119904+ 119881119892119904

10038161003816100381610038161003816

]

]

(8)

where119863ℎ119892

is the distance from household ℎ (ℎ = 1 2 119867)to the household group 119892 (119892 = 1 2 119866)119881

ℎ119904is the value of

variable 119904 (119904 = 1 2 119878) representing the character of house-hold ℎ119881

119892119904is the average value of variable 119904 of households in

household group 119892119908119904is the weight coefficient of the variable

119904 in explaining the character of household and householdgroup

The household ℎ is assigned into the most similar house-hold group andmakes the same land use change decisionwiththe household group

1198921015840= arg min 119863

ℎ1 119863ℎ2 119863

ℎ119892 119863

ℎ119866 (9)


where 1198921015840 is the most similar household group to householdℎ By establishing a case database of land use change decisionwe can assign each household into one similar enough house-hold group and deduce the land use decision It helps correctthe land use change results simulated of economic modulebased on ideas of optimization

For the assessment result of vegetation change modulethe agent-based module also provides a criterion to judgewhich kind of vegetation change will happen in a specificpixel

119871V119901

=

1

if forforall119906 = V 119875V119901119905+1 gt 0 119878V119906119901119905+1 gt 1 or le 0or119875V119901119905+1 le 0 119878V119906119901119905+1 gt 0 or le 1

0if forforall119906 = V 119875V119901119905+1 gt 0 119878V119906119901119905+1 gt 0 or le 1or119875V119901119905+1 le 0 119878V119906119901119905+1 gt 1 or le 0

(10)

where 119871V119901 = 1 denotes that the Vth type of vegetation isthe dominant vegetation in the pixel 119901 and 119871V119901 = 0 denotesthat the Vth type of vegetation is not the dominant vegetationin the pixel 119901 This criterion defines that for any other veg-etation type 119906 when 119875V119901119905+1 is positive and 119878V119906119901119905+1 is largerthan 1 or no larger than 0 or 119875V119901119905+1 is not positive and119878V119906119901119905+1 is smaller than 0 or no larger than 1 the Vth type ofvegetation is the dominant vegetation in the pixel 119901 in the(119905 + 1)th year when 119875V119901119905+1 is positive and 119878V119906119901119905+1 is largerthan 0 or no larger than 1 or 119875V119901119905+1 is not positive and119878V119906119901119905+1 is larger than 1 or no larger than 0 the Vth type of veg-etation is not the dominant vegetation in the pixel 119901 in the(119905 + 1)th year

For a specific pixel vegetation change will happen as longas the productivity of the new dominant vegetation exceedsthat of the original dominant vegetation

119871119881119901119905+1

= V if for forall119906 = V 119877119884V119901119905+1 gt 119877119884119906119901119905+1

119877119884V119901119905+1 = 119877119884V119901119905 +119877119884V119901119905

119877119884119901119905

119884V119901119905+1

1198771198841199010= sum

V

119860V1199010

119860119901

119884V1199010

(11)

where 119871119881119901119905+1

denotes the new vegetation type that charact-erized the pixel 119901 in the (119905 + 1)th year 119877119884V119901119905+1 is the pro-ductivity of the Vth type of vegetation in the pixel 119901 in the (119905+1)th year119877119884

119901119905is the total productivity of all the vegetation in

the pixel 119901 in the 119905th year 1198771198841199010

is the total productivity of allthe vegetation in the pixel 119901 in the base year 119860

119901is area of

pixel 119860V1199010 is area the Vth type of vegetation in the pixel 119901 inthe base year and 119884V1199010 is the productivity of the Vth type ofvegetation in the pixel 119901 in the base year

3 Discussion and Conclusions

In this paper we introduced the LUCDmodel which is com-patible with RCMs to provide endogenous underlying surface

for climate modeling This model is constituted by economicmodule vegetation changemodule and agent-basedmoduleThe economicmodule calculates the land use change demanddriven by economic activities aiming at maximizing eco-nomic utility The vegetation change module evaluates theprobability of vegetation change driven by climate changeThese two modules depict the land surface process under thecondition of rational decision making and ideal circumst-ances To couple the economicmodule and vegetation changemodule the AEZ was introduced in the LUCD model Theagent-based module identifies whether the land use changedemand and vegetation change can be realized under the con-dition of irrational decision making and multiple vegetationcompetition By introducing the simulation results of theLUCD model in RCM and applying the simulation results ofRCM in the LUCD model a coupled simulation system ofland surface system simulation can be established

In addition to themodeling framework several suggestedmodels were introduced and some specific parameter pro-cessing approaches were explained in detail for the constitu-tion of the LUCD model For the economic module a CGEmodeling framework and the difference between land andother production factors in CGEmodel were introducedTheeffects of climate change on human activities were also takeninto consideration by establishing production function foreach AEZ The AEZ model was suggested for the vegetationchange module and two indexes (possibility of vegetationchange and superiority index) were supposed to determinethe climate-induced vegetation change For the agent-basedmodule an example of land use change decision making andthe criterion of vegetation change was provided

The LUCD model offers a framework integrating humanactivities and climate change rational and irrational decisionmakings and macro- and microdynamic models into theland use changeThemodel is spatially explicit and has a goodempirical applicability by integrating natural vegetationchange and land use change By embedding the LUCDmodelresults into RCMsrsquo climate simulation regional land usechange and climate change can be iteratively simulated whichwill undoubtedly enhance the understanding of land surfacesystem dynamics

To ensure that the output on LUCCof LUCDmodel easilyfeeds into RCMsrsquo simulation the classification system ofLUCC in the LUCDmodel should be comparable with that ofunderlying surface data needed by RCMs The classificationsystem determines the choice of driving factors that affectland use change vegetation change and decision makingprocesses in the LUCD model In other words the modelingapproaches of three modules of the LUCD model should beaccordant with specific RCM which is one of the major rea-sons by which we keep the LUCC classification flexible in theLUCD model The proposed LUCC classification system bydefault in the LUCDmodel is compatible for most of RCMsAnd the specific parameter processing approaches providedin this study can also serve as valuable examples even if a newmodeling approach is used in module construction in theLUCD model

Both the LUCD and RCMs are grid based while the gridscale RCMs used are always much more rough for land use


change modeling Though there are more and more re-searches applying RCM based on high resolution grid datathe difference in grid scale is still kept as one of themajor chal-lenges in coupling the LUCD model and RCMs [43ndash46] Wesuggest introduction of subgrid to solve the grid scale issueThe LUCC simulation in the LUCDmodel can be based on ahigh resolution grid data and the simulation of climate changein RCMs can be implemented at a low resolution grid scale

Besides the grid scale issue the temporal scale differenceis the problem that hinders the seamlessly embedding of theLUCD model into RCM The simulation of RCM should behourly while that of the LUCDmight be yearly based at leastThe alteration of underlying surface data in RCMs will cer-tainly result in a sudden variation of simulation results Con-sequently the analysis of hourly climate data exported byRCMs will make little sense Therefore we suggest the sim-ulation results of RCM be reconciled to monthly or yearly foranalysis And it is apparent that the simulation results ofRCMs should be reconciled to yearly for the purpose of inputinto the LUCD model

Acknowledgments

This research was supported by National Key Programme forDeveloping Basic Science in China (no 2010CB950900) theKey Project funded by the Chinese Academy of Sciences (noKZZD-EW-08) and the National Department Public BenefitResearch Foundation of the Ministry of Land and Resourcesof China (no 201311001-5)

References

[1] E Kalnay and M Cai ldquoImpact of urbanization and land-usechange on climaterdquo Nature vol 423 pp 528ndash531 2003

[2] C D Thomas A Cameron R E Green et al ldquoExtinction riskfrom climate changerdquo Nature vol 427 pp 145ndash148 2004

[3] R A Pielke Sr ldquoLand use and climate changerdquo Science vol 310no 5754 pp 1625ndash1626 2005

[4] A Grell J Dudhia andD R Stauffer ldquoA description of the fifth-generation Penn StateNCARmesoscalemodel (MM5)rdquoNCARTechnical Note Boulder Colo USA 1994

[5] JM ShepherdMCarterMManyinDMessen and S BurianldquoThe impact of urbanization on current and future coastal pre-cipitation a case study for houstonrdquo Environment and PlanningB vol 37 no 2 pp 284ndash304 2010

[6] C-Y Lin W-C Chen P-L Chang and Y-F Sheng ldquoImpact ofthe urban heat island effect on precipitation over a complex geo-graphic environment in Northern Taiwanrdquo Journal of AppliedMeteorology and Climatology vol 52 no 3 pp 570ndash587 2011

[7] Y Z Lin A P Liu E J Ma X Li and Q L Shi ldquoImpacts offuture urban expansion on regional climate in the NortheastMegalopolis USArdquo Advances in Meteorology vol 2013 ArticleID 362925 10 pages 2013

[8] Y J Cai L C Tan H Cheng et al ldquoThe variation of summermonsoon precipitation in central China since the last deglacia-tionrdquo Earth and Planetary Science Letters vol 291 no 1ndash4 pp21ndash31 2010

[9] B L Turner IIWC Clark RWKates J F Richards J TMath-ews and W B Meyer The Earth as Transformed by Human

Action Global and Regional Changes in the Biosphere over thePast 300 Years Cambridge University Press Cambridge UK1993

[10] J Y Liu M L Liu D F Zhuang Z X Zhang and X Z DengldquoStudy on spatial pattern of land-use change in China during1995ndash2000rdquo Science inChinaD vol 46 no 4 pp 373ndash384 2003

[11] GLP ldquoGLP science plan andimplementation strategyrdquo IGBPReport no 53 IHDP Report no 19 IGBP Secretariat Stock-holm Sweden 2005

[12] X Z Deng Q O Jiang H B Su and F Wu ldquoTrace forest con-versions in Northeast China with a 1-km area percentage datamodelrdquo Journal of Applied Remote Sensing vol 4 no 1 ArticleID 041893 pp 1ndash13 2010

[13] C H Zhao X Z Deng Y W Yuan H M Yan and H LiangldquoPrediction of drought risk based on theWRFmodel in YunnanProvince of Chinardquo Advances in Meteorology vol 2013 ArticleID 295856 9 pages 2013

[14] Q B Le S J Park P L G Vlek and A B Cremers ldquoLand-usedynamic simulator (LUDAS) a multi-agent system model forsimulating spatio-temporal dynamics of coupled human-land-scape system I structure and theoretical specificationrdquo Ecolog-ical Informatics vol 3 no 2 pp 135ndash153 2008

[15] J Y Liu and X Z Deng ldquoProgress of the research methodolo-gies on the temporal and spatial process of LUCCrdquo Chinese Sci-ence Bulletin vol 55 no 14 pp 1354ndash1362 2010

[16] J Y Liu and X Z Deng ldquoImpacts and mitigation of climatechange on Chinese citiesrdquo Current Opinion in EnvironmentalSustainability vol 3 no 3 pp 188ndash192 2011

[17] J Y Liu H Q Tian M L Liu D F Zhuang J M Melillo andZ X Zhang ldquoChinarsquos changing landscape during the 1990slarge-scale land transformations estimated with satellite datardquoGeophysical Research Letters vol 32 no 2 pp 1ndash5 2005

[18] A Veldkamp and L O Fresco ldquoCLUE a conceptual model tostudy the conversion of land use and its effectsrdquo EcologicalModelling vol 85 no 2-3 pp 253ndash270 1996

[19] Z Q Duan F R Zhang and LMMiao ldquoNeighborhood-basedmethod for land-use spatial pattern analysis and its applicationrdquoTransactions of the Chinese Society of Agricultural Engineeringvol 22 no 6 pp 71ndash76 2006

[20] X Z Deng Q O Jiang J Y Zhan S J He and Y Z Lin ldquoSim-ulation on the dynamics of forest area changes in NortheastChinardquo Journal of Geographical Sciences vol 20 no 4 pp 495ndash509 2010

[21] T P Evans and H Kelley ldquoMulti-scale analysis of a householdlevel agent-basedmodel of landcover changerdquo Journal of Enviro-nmental Management vol 72 no 1-2 pp 57ndash72 2004

[22] W J McConnell ldquoAgent-based models of land-use and land-cover changerdquo LUCC Report no 6 2001

[23] S Manson ldquoLand use in the Southern Yucatan peninsularregion of Mexico scenarios of population and institutionalchangerdquo Computers Environment and Urban Systems vol 30no 3 pp 230ndash253 2006

[24] F Semboloni J Assfalg S Armeni R Gianassi and F MarsonildquoCityDev an interactive multi-agents urban model on the webrdquoComputers Environment and Urban Systems vol 28 no 1-2 pp45ndash64 2004

[25] T Zhang J Y Zhan J Huang R Yu and C C Shi ldquoAn agent-based reasoning of impacts of regional climate changes on landuse changes in the Three-River Headwaters region of ChinardquoAdvances in Meteorology vol 2013 Article ID 248194 9 pages2013


[26] M Batty Y C Xie and Z L Sun ldquoModeling urban dynamicsthroughGIS-based cellular automatardquoComputers Environmentand Urban Systems vol 23 no 3 pp 205ndash233 1999

[27] J I BarredoMKasankoNMcCormick andC Lavalle ldquoMod-elling dynamic spatial processes simulation of urban future sce-narios through cellular automatardquo Landscape and Urban Plan-ning vol 64 no 3 pp 145ndash160 2003

[28] A D Syphard K C Clarke and J Franklin ldquoUsing a cellularautomaton model to forecast the effects of urban growth onhabitat pattern in Southern Californiardquo Ecological Complexityvol 2 no 2 pp 185ndash203 2005

[29] FAO Agro-Ecological Zoning Guidelines vol 73 of FAO SoilsBulletin Food and Agriculture Organization of the UnitedNations Rome Italy 1996

[30] E Stehfest M Heistermann J A Priess D S Ojima and JAlcamo ldquoSimulation of global crop production with the ecosys-temmodel DayCentrdquo EcologicalModelling vol 209 no 2ndash4 pp203ndash219 2007

[31] X ZDengH B Su and J Y Zhan ldquoIntegration ofmultiple datasources to simulate the dynamics of land systemsrdquo Sensors vol8 no 2 pp 620ndash634 2008

[32] X Z Deng F Yin Y Z Lin Q Jin and R J Qu ldquoEquilibriumanalyses on structural changes of land uses in Jiangxi ProvincerdquoJournal of Food Agriculture and Environment vol 10 no 1 pp846ndash852 2012


[34] P H Verburg W Soepboer A Veldkamp R Limpiada VEspaldon and S S A Mastura ldquoModeling the spatial dynamicsof regional land use the CLUE-S modelrdquo Environmental Man-agement vol 30 no 3 pp 391ndash405 2002

[35] K H Lau and B H Kam ldquoA cellular automata model for urbanland-use simulationrdquo Environment and Planning B vol 32 no2 pp 247ndash263 2005

[36] D Li X Li X P Liu et al ldquoGPU-CAmodel for large-scale land-use change simulationrdquo Chinese Science Bulletin vol 57 no 19pp 2442ndash2452 2012

[37] S Scheiter L Langan and S I Higgins ldquoNext-generationdynamic global vegetation models learning from communityecologyrdquoThe New Phytologist vol 198 no 3 pp 957ndash969 2013

[38] F I Woodward andM R Lomas ldquoVegetation dynamicsmdashSim-ulating responses to climatic changerdquo Biological Reviews of theCambridge Philosophical Society vol 79 no 3 pp 643ndash6702004

[39] T X Yue J Y Liu S E Joslashrgensen ZQGao SH Zhang andXZ Deng ldquoChanges of Holdridge life zone diversity in all ofChina over half a centuryrdquoEcologicalModelling vol 144 no 2-3pp 153ndash162 2001

[40] Q L Shi Y Z Lin E P Zhang H M Yan and J Y ZhanldquoImpacts of cultivated land reclamation on the climate and grainproduction inNortheast China in the future 30Years rdquoAdvancesin Meteorology vol 2013 Article ID 853098 8 pages 2013

[41] J Schmidhuber and F N Tubiello ldquoGlobal food security underclimate changerdquo Proceedings of the National Academy of Sciencesof the United States of America vol 104 no 50 pp 19703ndash197082007

[42] P Deadman D Robinson E Moran and E Brondizio ldquoCol-onist household decisionmaking and land-use change in theAmazon Rainforest an agent-based simulationrdquo Environmentand Planning B vol 31 no 5 pp 693ndash709 2004

[43] S RGaffinC Rosenzweig R Khanbilvardi et al ldquoVariations inNew York cityrsquos urban heat island strength over time and spacerdquoTheoretical and Applied Climatology vol 94 no 1-2 pp 1ndash112008

[44] J M Liu and X Z Deng ldquoInfluence of different land use onurban microenvironment in Beijing City Chinardquo Journal ofFood Agriculture and Environment vol 9 no 3-4 pp 1005ndash10112011

[45] L Tian J Q Chen and S X Yu ldquoHow has Shenzhen beenheated up during the rapid urban build-up processrdquo Landscapeand Urban Planning vol 115 pp 18ndash29 2013

[46] J Y Liu Z X Zhang X L Xu et al ldquoSpatial patterns and drivingforces of land use change inChina during the early 21st centuryrdquoJournal of Geographical Sciences vol 20 no 4 pp 483ndash4942010

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

ClimatologyJournal of

EcologyInternational Journal of


EarthquakesJournal of


Hindawi Publishing Corporationhttpwwwhindawicom

Applied ampEnvironmentalSoil Science

Volume 2014

Mining


Journal of

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014

International Journal of

Geophysics

OceanographyInternational Journal of


Journal of Computational Environmental SciencesHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal ofPetroleum Engineering


GeochemistryHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Atmospheric SciencesInternational Journal of


OceanographyHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Advances in


MineralogyInternational Journal of


MeteorologyAdvances in

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Paleontology JournalHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

ScientificaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Geological ResearchJournal of


Geology Advances in


human-landscape system and feeds back to the subsequentdevelopment of these interactions [14] Most land use changesimulation models simulate successional pattern change ofland use under the macrobackground of the regional popula-tion growth economic development social progress changesin the natural environment and other facts [15 16] On thewhole the land use change simulationmodels can be broadlydivided into three major categories empirical statisticalmodel agent-based model (ABM) and raster neighborhoodrelationship based model [17]

There are abundant empirical statisticalmodels applied toland use change simulation A typical example is the Conver-sion of land use and its effect at small regional extent (CLUE-S) model whose application in land use change simulation iscurrently in the ascendant [18ndash20]TheCLUE-Smodel is con-structed to simulate land use change and its effects on enviro-nment at mesomicroscale It has the capability of synchro-nously simulating the changes of multiple types and intro-duces the dynamic driving factors (such as population andeconomic growth) to improve the simulation accuracy Sincethe 1990s along with the rapid development of complexity science ABM began to be applied to land use change research[21] The agent-based models of land use and land cover(ABMLUCC) were specially discussed by LUCC reportnumber 6 inwhich the development prospect ofABM in landuse change simulation is highly valued [22] An ABM modelmainly identifies the linkage between agents and environ-ment by describing the interaction and affiliation of indepen-dent agents [23] By combining ABM and cellular automata(CA) model the simulation of land use change is character-ized by multiscale and becomes more effective in multiobjec-tive decisionmaking Semboloni et al [24] established amul-tiagent system with the residents industry practitioners ser-vice practitioners and developers as themain agents to simu-late the land use change in urban areas Zhang et al [25] builtan ABM to simulate the impacts of the climate changes on theregional land use changes and agentsrsquo economic benefits intheThree-River Headwaters Region of China It is found thatthe agents would get more wealth under the scenario withoutclimate changes in the long term even though the totalincome is lower than that under the scenario with climatechanges As a representative of raster neighborhood relation-ship basedmodel CAmodel iswidely used in landuse changesimulation especially urban expansion [26 27] Syphard et al[28] analyzed the distinction of LUCC caused by urbanexpansion in areas with different slope with the CA modelOne of the superiority of the CA model in land use changesimulation is that it supports visualization of the simulationprocess The structure of the CA model makes it difficultconsidering the impacts of land use policies

In this study we developed an LUCD model com-patible with RCMs to describe the interdependencies andfeedback mechanisms between social economics ecosystemenvironment and irrational decision-making process TheLUCD model describes a combined and complex systemcomposed of social economic ecosystem components anddecision-making process and consequences It provides aconsistent and comprehensive framework of land use changemodeling and emphasis on how themodels work together By

introducing agroecological zone (AEZ) based on the sim-ulation results of RCMs the LUCD model is compatiblewith RCMs and constitutes an iterative simulation system ofLUCC and climate changes

2 Land Use Change Dynamics (LUCD) Model

21 Model Structure The LUCD model can be subdividedinto three modules namely economic module vegetationchange module and agent-based module The economicmodulcalculates the area demand for all land use types ineconomic activitiesmaximizing economic utility of land usesThe vegetation change module provides the probability ofvegetation change driven by climate changes And the agent-based module identifies whether the land use demand andvegetation change can be realized and provides the land usechange simulation results which are the underlying surfacesneeded by RCM To feed the LUCDmodel results into RCMsthe land use system applied to the LUCD model should beconsistent with the underlying surfaces used in RCMs(Figure 1) By iteratively using the output of one model as theinput of another the LUCDmodel is compatible with RCMsIn Figure 1 the dotted lines show the data transmission be-tween the LUCDmodel andRCMswhile the solid lines standfor the flow of information in the LUCD model The LUCDmodel provides the simulation results of land use change forRCMs as underlying surface data then RCMs can simulatethe climate change resulted from the land use change Theresults of climate change simulated by RCMs are furtherimported into the LUCD model and affect land use change

The economic module estimates the land use change de-mand driven by human activityThe current condition of landuses is introduced in this module as one of the limitations ofeconomic activity as well as land use decisions The equilib-rium of markets determines the commodity supply and inturn influences the land use demand Combining the land usedemand the limited amount of land and the current land usestatus the land use change demand is obtained The vegeta-tion change module describes the possible vegetation changedriven by climate change The AEZ is the key concept thatlinks the climate change and vegetation change and helps tocouple human activity with climate change The climatechange leads to change of AEZs which determines the growthof vegetation [29 30] Consequently the climate changeaffects not only the evolution of natural vegetation but alsothe human activities including planting and breeding Byoverlying the AEZs on the current vegetation pattern thesuitability of vegetation change can be evaluated The agent-based module describes the procedure of land use decisioncoupled with the land use change demand and vegetationchange suitability using the agent-based simulation technol-ogy This module identifies whether or not the theoreticalland use change demand and the possible vegetation changeestimated by the economicmodule and the vegetation changemodule can be realized The output of this module land usechange is the underlying surfaces that needed by RCMs Byembedding the LUCD in RCMs an iterative simulation sys-tem of land use change and climate change is constructed(Figure 1)





Human activity


Climate change

Vegetationcharacter



Land use change

Land use status


Land

use

clas

sifica

tion

syste

m

Land

use

clas

sifica

tion

syste




119884119894119890= 119887119894119890prod

ℎ

119865120573ℎ119894119890

ℎ119894119890prod

119897

119860119890119888120589119897119894119890

119897119894119890 (1)

Total

supplyland


Incomes

Domestic output

Commodity supply

Commoditydemand

Import

Export

Commodity market

Total utility



119860119890119888119897119894119890=120577119897119894119890119884119894119890

119887119894119890

(2)



119890th AEZ 119860119890119888119897119894119890






functions and 120573ℎ119894119890




ℎ119894119890

120577119897119894119890= 119891 (119865labor119894119890 119865capital119894119890) (3)





119880119890119888 = prod

119894

119883119901120594119894

119894 (4)





maximize119883119901119894

119880119890119888 = prod

119894

119883119901120594119894

119894

subject to 119879119897119886119899119889 = sum

119890

sum

119897

sum

119894

119860119890119888119897119894119890

(5)



119875V119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884Vmax (6)



119878V119906119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884119906119901119905+1

minus 119884119906119901119905

(7)




119863ℎ119892=

119878

sum

119904=1

119908119904[

[

(119881ℎ119904minus 119881119892119904)2

10038161003816100381610038161003816119881ℎ119904+ 119881119892119904

10038161003816100381610038161003816

]

]

(8)

where119863ℎ119892








1198921015840= arg min 119863

ℎ1 119863ℎ2 119863

ℎ119892 119863

ℎ119866 (9)




119871V119901

=

1



(10)



119871119881119901119905+1


119877119884V119901119905+1 = 119877119884V119901119905 +119877119884V119901119905

119877119884119901119905

119884V119901119905+1

1198771198841199010= sum

V

119860V1199010

119860119901

119884V1199010

(11)

where 119871119881119901119905+1





119901is area of












Acknowledgments


References


























































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in





Human activity


Climate change

Vegetationcharacter



Land use change

Land use status


Land

use

clas

sifica

tion

syste

m

Land

use

clas

sifica

tion

syste




119884119894119890= 119887119894119890prod

ℎ

119865120573ℎ119894119890

ℎ119894119890prod

119897

119860119890119888120589119897119894119890

119897119894119890 (1)

Total

supplyland


Incomes

Domestic output

Commodity supply

Commoditydemand

Import

Export

Commodity market

Total utility



119860119890119888119897119894119890=120577119897119894119890119884119894119890

119887119894119890

(2)



119890th AEZ 119860119890119888119897119894119890






functions and 120573ℎ119894119890




ℎ119894119890

120577119897119894119890= 119891 (119865labor119894119890 119865capital119894119890) (3)





119880119890119888 = prod

119894

119883119901120594119894

119894 (4)





maximize119883119901119894

119880119890119888 = prod

119894

119883119901120594119894

119894

subject to 119879119897119886119899119889 = sum

119890

sum

119897

sum

119894

119860119890119888119897119894119890

(5)



119875V119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884Vmax (6)



119878V119906119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884119906119901119905+1

minus 119884119906119901119905

(7)




119863ℎ119892=

119878

sum

119904=1

119908119904[

[

(119881ℎ119904minus 119881119892119904)2

10038161003816100381610038161003816119881ℎ119904+ 119881119892119904

10038161003816100381610038161003816

]

]

(8)

where119863ℎ119892








1198921015840= arg min 119863

ℎ1 119863ℎ2 119863

ℎ119892 119863

ℎ119866 (9)




119871V119901

=

1



(10)



119871119881119901119905+1


119877119884V119901119905+1 = 119877119884V119901119905 +119877119884V119901119905

119877119884119901119905

119884V119901119905+1

1198771198841199010= sum

V

119860V1199010

119860119901

119884V1199010

(11)

where 119871119881119901119905+1





119901is area of












Acknowledgments


References


























































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in



119880119890119888 = prod

119894

119883119901120594119894

119894 (4)





maximize119883119901119894

119880119890119888 = prod

119894

119883119901120594119894

119894

subject to 119879119897119886119899119889 = sum

119890

sum

119897

sum

119894

119860119890119888119897119894119890

(5)



119875V119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884Vmax (6)



119878V119906119901119905+1 =119884V119901119905+1 minus 119884V119901119905

119884119906119901119905+1

minus 119884119906119901119905

(7)




119863ℎ119892=

119878

sum

119904=1

119908119904[

[

(119881ℎ119904minus 119881119892119904)2

10038161003816100381610038161003816119881ℎ119904+ 119881119892119904

10038161003816100381610038161003816

]

]

(8)

where119863ℎ119892








1198921015840= arg min 119863

ℎ1 119863ℎ2 119863

ℎ119892 119863

ℎ119866 (9)




119871V119901

=

1



(10)



119871119881119901119905+1


119877119884V119901119905+1 = 119877119884V119901119905 +119877119884V119901119905

119877119884119901119905

119884V119901119905+1

1198771198841199010= sum

V

119860V1199010

119860119901

119884V1199010

(11)

where 119871119881119901119905+1





119901is area of












Acknowledgments


References


























































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in




119871V119901

=

1



(10)



119871119881119901119905+1


119877119884V119901119905+1 = 119877119884V119901119905 +119877119884V119901119905

119877119884119901119905

119884V119901119905+1

1198771198841199010= sum

V

119860V1199010

119860119901

119884V1199010

(11)

where 119871119881119901119905+1





119901is area of












Acknowledgments


References


























































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in




Acknowledgments


References


























































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in
































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in










Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in

Documents

A Framework for the Land Use Change Dynamics Model ... · 2 AdvancesinMeteorology human-landscapesystemandfeedsbacktothesubsequent developmentoftheseinteractions[14].Mostlandusechange