| With the intensive study of global environmental change，since the nineties of thetwentieth century, land use/cover change (LUCC) has caused the extensive concernof the international community, gradually becoming an important part and the corecontent of global environmental change. In the course of LUCC studies, land usespatial change models are an important technical means of analysis and research onland use change trends and its environmental effects in a certain area. Land usechange is a coefficient result of physic-geographical environment system andsocio-economic system. Using spatial simulation model to analyze the causes andprocess of land use change, and predict future patterns of land use spatial distributionhas become a hot area of today’s global environmental change research.In this paper, Shandong province Qixia city in JiaoDong Mountain Area wastaken an example. Based on remote sensing data in1992,2003and2010, this studyperformed analysis and division of landscape pattern according to the results ofcluster analysis of landscape indices in catchments. This paper chose19kinds ofdriving factors in natural and socio-economic aspects. CLUE-S model is a model ofspace module integrated development to predict land use pattern evolution based onLogistic regression analysis. This study explored the applicability of this model inQiXia city.Firstly,1992and2003land-use remote sensing data as the data source, takingdriving factors in natural and socio-economic aspects in these two years, and thatconfirming final driving factors with the model of multi-collinearity diagnosticanalysis. With the help of SPSS software, this paper analyzed all land use types anddriving factors using the Logistic regression model based on different scales, and analyzed driven mechanism of all land use types from1992to2010in Qixia cityaccording to the regression results. Then the simulating optimal-scale of land usechange was determined by ROC curve checkout. Secondly, using land use data in1992and2003to simulate land use spatial pattern in2010, and conducting precisionvalidation between simulation result maps and real land use map. In the end, on thebasic of1992simulating2010and2003simulating2010, Comprehensiveconsidering “Qixia city Land Use Planning (2006-2020)” and Qixia city otherdevelopment plans, this paper built three different simulating scenarios of foodsecurity, ecological protection and rapid urbanization. And then simulating land usespatial pattern in those simulating scenarios.The main research conclusions were as follows:(1) Analysis of driven mechanism of all land use types from1992to2010inQixia city were following.①Cultivatied land: Its spatial distribution was greatlyaffected by natural and location factors, and its area was mainly affected by humanactivities.②Woodland: Natural factors and socio-economic factors synergisticallyaffected woodland, whose area was affected by socio-economic factors totally.③Garden land: Socio-economic factors played a decisive role in addition to naturaland location factors for garden land.④Grass land: Natural factors affected itsspatial distribution of geographical position, socio-economic factors and locationfactors affected its area and the fragmentation degree of landscape patches.⑤Construction land: Socio-economic factors promoted expansion demand ofconstruction land; location factors determined the Expansion direction ofconstruction land.⑥Waters: Terrain factor affected spatial distribution of waters;climate factors affected the area of waters.⑦Other land: Terrain factor affectedspatial distribution of other land mainly; whose main feature was bare rock in theupper part of mountain.(2)This study used land use data of the study area in different years to simulateland use spatial pattern of the study area in the same year. That is to say, usingremote sensing data of Qixia city in1992and2003to simulate land use spatialpattern of Qixia city in2010, and then carrying out to verify accuracy between two simulating final maps with true land use map in2010. Study results were good.Based on multi-collinearity diagnostic analysis this paper chose eventually19driving factors: elevation, slope, aspect, annual accumulated temperature above10℃, annual average sunshine (h), soil thickness (mm), soil organic matter content(g/kg), distance to (reservoirs and lakes, rivers, railways, highways, provincialroads, national highways, urban center, economic development zone, ruralresidential area and town center), population density(the number of person persquare kilometers) and urbanization level. The simulating optimal-scale of land usechange was100m×100m determined by ROC curve. Upon checking, the overallaccuracy of1992simulated2010and2003simulated2010were97.83%,97.88%severally, and the Kappa indexes of the two periods were greater than0.83. Theresearch results showed out that CLUE-S model had a better ability to simulate landuse in Qixia city.(3)Three simulation scenarios predicted results show that there was a big spatialdifference in land use landscape pattern under different simulation scenarios.Farmland and forest land were different under different simulation scenarios. Gardenland and construction land were increased, while garss land, waters and other landwere reduced. Under food security scenario, duing to strict control of arable land intoother types and land development and land consolidation measures, farmlandincreased significantly; forest land and grass land were reduced; garden land andconstruction land increased. Under ecological protection scenario, farmlanddecreased; forest land and garden land increased largely, and waters did not change.Under rapid urbanization scenario, farmland decreased significantly, garden land andconstruction land were increased significantly, and waters decreased smally.Specific performance:①Cultivated land: Invariant area of cultivated landconcentrated in the flat area of western, southern and northeastern, abandonedcultivated land mainly located in the periphery of stable arable land, part of thesloping cultivated land in the east, the lower region of central hills and mountain;part of the old garden land was converted to farmland in northeast and southwest. ②Forest land: Invariant area of forest land concentrated in eastern and northwesternregions; increased forest land was mainly in the central of east region near YaMountain and Ai Mountain; partly grassland of suitable woodland cultivation wasconverted to forest land.③The area of garden land increased significantly. The studyarea has an apple production base, which was mainly located in the area northeastand southwest, also front and center inclined plains region of mountains, and therewere also scattered central region in central region. Farmland and forest land innortheast and southwest converted to garden land, forest land in hilly area, lower partof the ridge slope terrace and the outer edge of floodplain converted to garden land.④The area of grass land decreased. Invariant area of grass land concentrated inhigh-altitude areas of Takaoka mountainous region, and higher elevations of themountain regions of the northwest and southeast; grassland in lower elevations of theeast region of the center, and interaction region of forestry and farming converted toforest land.⑤A large area of construction land expansion was mainly occupied landsurrounding farmland and grassland, and developed other land. The main regionswere focus in the center of Qixia district, economic development zone and towns.⑥W aters and other landdecreased marginally.(4) The results showed that CLUE-S model had a good adaptability to the Qixiacity where was located in a typical JiaoDong mountain area. These researchconclusions can provide some reference value and significances to carry out similarstudies in the area of small and medium-scale complex mountainous terrain in thefuture. |
PDF Full Text Request