Study On Spatial Reconstruction Of Historical Land Use In Zhenlai County, Jiiin Province, China

Understanding long-term human-environment interactions is essential tounderstanding changes in terrestrial ecosystems and this requires historicalreconstruction of past land cover changes. Historical reconstruction of land use/landcover (LULC) aims to reproduce information concerning past land use, not only thequantity of land use/cover in a historical period, but also the spatial distribution.Recently, improved remote sensing technology has made feasible the continuousobservation of land cover. However, remotely-sensed data have only existed for the lastfour decades at most, following the advent of the first land satellite, LandSat-1,launched in1972. Prior to that, other data sources must be relied on, which may coverthe global scale but often inconsistently. In this context, increasing numbers ofresearchers have made efforts to reconstruct historical LULC based on prime datasources and research approaches. And significant progress in gathering historical landchange data has been made both at global and regional scales. However, most of theexisting historical LULC reconstructions do not sufficiently meet the requirements ofclimate assessments due to insufficient spatial and thematic detail and the lack ofconsideration of various land change types. Most current studies do not thematicallyrepresent100%of the land area, and ignore the consideration of competing landcategories and land conversion types. Current research mainly focuses on arable land,wetland and forestland and it does not provide information on land categories such as settlement, water, and other land types. It is a research direction to build historicalspatial land use and land cover datasets with high resolution.Based on CAS “Strategic Priority Research Program—Dynamic Changes ofFarming-pastoral Ecotone Pattern in Northern China Response to Climate Warmingover the Past Century” and National Natural Science Foundation of China “Study ondigital reconstruction of land use change and vulnerability in the past century infarming-forestry ecotone of Northeast China”, this dissertation tries to explore thepossibility of using historical documents together with a Historical Land UseReconstruction Model (HLURM) to rebuild land use in the1930s in Zhenlai County.And then the simulation result is also validated using a three-map comparison method.The main conclusions of this dissertation are as follows.(1) Historical topographic maps are useful for retrospective analysis of land coverpatterns and their change over time. Historical maps containing land-cover informationwith temporal layer, can serve as a basis for the reconstruction of past land use and landcover, especially when they are used in a geographic information system (GIS). It isfeasible to extract historical land use data from topographic maps. The resolution oftopographic maps at the scale of1:100,000could meet the current study of land use andland cover change.(2) The farming radius is less accurate calculated by the regular equal method thanthat based on the ratio of farm land and rural settlement. It is feasible to reconstructarable land based on the relationship between settlement distribution and farming radius.Yet it is just an ideal reconstruction model for arable land.(3) Results of bi-temporal change detection showed that land use/cover changedsignificantly in Zhenlai County from1954to2005. Arable land expanded at theexpense of grassland and wetland loss. Meanwhile, plenty of grassland converted toother unused land, indicating seriously environmental degradation in Zhenlai Countyduring the past decades. Trajectory analysis of land use and land cover changedemonstrated that settlement, arable land, water bodies were relatively stable in termsof coverage and spatial distribution, while grassland, wetland and forest land had weak stability. Many transitions from wetland to other unused land indicated that landdegradation had difficult reversibility. Besides, natural forces were still dominating theenvironmental processes of the study area while human-induced changes also playedan important role in environmental change. Lorez curve-Gini coefficient could beapplied effectively to analyze the land use structure changes. Arable land was the mostdecentralized, whereas forest land was the most concentrated.(4) The Historical Land Use Reconstruction Model (HLURM) contains fourmodules: land quantity module, land conversion module, land suitability’s probabilitymodule, and spatial allocation module. Taking the grid as the research unit, this modeldistributes the total quantity of each land category into grids according to thedescending order of the probability and some spatial allocation rules. The validation ofHLURM model requires a three-map comparison and then the result shows that thesimulation model of HLURM has the high precision compared with the null model andsome other simulation models. The major differences among the simulated1932map,the reference1932map and the reference1954map, have less to do with the simulationmodel and more to do with the inconsistencies among the land categories during thestudy period.