Spatio-temporal Evolution And Simulation Of Land Use In Xi’an Based On Markov-PLUS Model

As a central city in the northwestern region,Xi’an has experienced rapid economic development,rapid population growth,increasing urbanisation and major changes in urban land use over the past two decades.The increasing contradiction between people and land has led to the unbalanced development of the land system and the gradual deterioration of the urban ecological environment,and has caused a series of social problems.Therefore,it is important to study the changes in land use patterns in Xi’an over the past two decades,explore the driving forces of land use changes,simulate and predict future land use patterns for ecological environmental protection and sustainable socio-economic development in Xi’an.Based on Google Earth Engine(GEE)cloud platform,this paper uses Landsat remote sensing images,DEM data,natural environment and socio-economic data of Xi’an city in 2000,2010 and 2020 as support,and constructs different combinations of classification features based on spectral,topographic and textural features,Random Forest(RF)and Support Vector Machine(SVM)were used to classify land use in Xi’an from 2000 to 2020,the spatial and temporal evolution characteristics of land use in Xi’an were analysed using the land usedynamics,the comprehensive index of land use degree and the transfer matrix.The main driving factors affecting the land use pattern of Xi’an were explored based on land use expansion strategies;on this basis,the PLUS model was used to forecast and simulate the land use pattern of Xi’an in2030 under four different scenarios.The main results of the research are as follows:(1)The GEE cloud-based platform can achieve fast and efficient classification of large area land use over long time series.The experimental results show that the multi-feature classification scheme has the highest accuracy,and the classification accuracy of the RF algorithm is better than the SVM machine algorithm.During the experiments,the highest accuracy was obtained for the RF combination of fused multiple features using spectral,index,texture and topographic features,with classification accuracies of 92.72%,91.76% and 92.12% for the years 2000,2010 and2020 respectively.Compared to the use of spectral features alone,the simultaneous addition of index,texture and topographic features improved the overall accuracy by8.93,7.87 and 7.33 percentage points for each year,and the kappa coefficients by 0.1,0.09 and 0.06 respectively.(2)In terms of temporal evolution,the proportion of the total area of land use types in Xi’an in the last 20 years was in the order of arable land,forest land,grassland,construction land,water and unused land.Among them,the area of arable land is the largest,while forest land and grassland also occupy a large proportion.Over the past 20 years,the area of arable land has continued to decline,decreasing by450.16 km2;the area of forest land has not changed much,increasing by 12.61 km2;from 3,017.3 km2 in 2000 to 3,604.94 km2 in 2010;the area of grassland has first decreased and then increased,decreasing by a total of 46.14 km2;the area of water has gradually increased,increasing by a total of 39.57 The area of water area gradually increased,with a total increase of 39.57 km 2;the area of built-up land gradually increased,with a cumulative increase of 443.03km2;the area of unused land accounted for a small proportion,with a slight increase of 1.42km2.Although the area of arable land,forest land and grassland is large and accounts for a large proportion of the total,the variation is relatively small,while the construction land has the largest variation,followed by the unused land,and the water area also has a large variation.2000-2010 and 2010-2020,Xi’an’s comprehensive land use dynamics are 3.55% and 3.14% respectively,indicating that land use changes in Xi’an were relatively active during 2000-2010,and the latter period The stability of land use has increased in the latter period.(3)In terms of Spatialevolution,arable land is mainly distributed in the counties around the urban area and in the areas north of the Qinling Mountains.Forests and grasslands are mainly located along the Qinling Mountains.Water areas are mainly located in the Xi’an section of the Wei River and near the eight rivers of the "Eight Waters that Run Xi’an".Construction land is concentrated in the urban areas of Xi’an.Unused land is mainly located in parts of Zhouzhi and Lantian counties.Over the past20 years,the area of arable land has been decreasing,with grassland being the main source of inflow to arable land and grassland being the main source of outflow to construction land.The area of construction land has been increasing,with cropland being the main source of inflow of construction land,while construction land is mainly outflow to cropland;the total amount of unused land has basically not increased or decreased.The degree of land use has been on the rise year by year,with Yanliang District,Weiyang District and Yanta District experiencing greater growth.(4)The analysis of the drivers of land use change in Xi’an shows that the main driving factors for the expansion of different land types are different.In the process of applying the strategy of land use expansion analysis,the magnitude of the contribution of these drivers to the expansion of each land category was analysed according to the principles of driver selection.The results show that elevation plays a very significant role in the expansion of land use types in each category.Based on the above,using the various modules provided by Markov-PLUS and setting reasonable neighbourhood weighting parameters based on the probability of land expansion after several experiments,the land use of Xi’an in 2030 is predicted under urban development,farmland protection,ecological priority and natural development scenarios.For different scenario simulations,parameters such as the transfer probability matrix and the initial land use distribution of each type of land can be adjusted to obtain the simulation results of land use changes under different scenarios.The results show that the decrease of arable land and the increase of building land are the main trends of future land use,especially under the urban expansion scenario.Under the arable land conservation scenario,the conversion of forest and grassland to arable land increases,leading to a rapid increase in the amount of arable land.