Research On Urban Expansion And Simulation In Xining City And Its Surrounding Areas

Xining City is the largest city in Qinghai Province and on the entire Qinghai-Tibet Plateau.Since ancient times,it has been an integral route,part of the"Tang-Tibetan Ancient Road"and the"Silk Road",an important northwest traffic route,and a military center.Because of its importance as the provincial capital of Qinghai and an important gateway to the Qinghai-Tibet Plateau,in 2015,the State Council approved it as an important central city in Northwest China.As of forty years ago,Xining’s economic foundation and urban-rural development were relatively backward.Since the reform and opening up,especially the implementation of the national"Western Development"strategy,and the ongoing"One Belt One Road"construction,Xining’s urbanization has accelerated.Improvement,especially urban expansion,has been significant.As with other typical mountain and valley cities,Xining City’s urban spatial expansion pattern is strictly restricted by geomorphology.Thirty years of urban expansion and construction have resulted in the occupation of a large amount of cropland,forestland and grassland in the valley plains,which has caused an increase in the negative effects of the urban ecological environment,including urban heat islands,deterioration of water quality,and other environmental issues such as air pollution.This study focuses on the study area of the river valley basin in Xining city in order to explore the temporal and spatial characteristics of urban land cover changes in the process of urbanization,the driving forces behind urban expansion,predict the expansion trend of urban construction land under future scenarios,and optimize the future urban land space and sustainable development of urban land.We use the GEE cloud platform together with long time series Landsat satellite image data(1987-2019),using the random forest algorithm,feature parameter optimization to explore land use dynamics and land cover spatiotemporal change.We further introduce the geography Detector(Geodetector)to carry out factor detection,interactive detection,and risk detection on the driving factors of urban construction land and cultivated land changes from 2000 to 2010 and 2010 to 2019.Based on the geographic simulation and optimization system(Geo SOS),these simulate the spatial pattern of land cover in the study area in 2009 and 2019,under the natural development scenario of slope restriction,and predict the spatial pattern of land cover in Xining City in 2025.The main research conclusions are as follows:(1)Supported by the GEE cloud platform,based on pixel stitching technology,we obtained a long-term sequence of the study area from 1987 to 2019,on June to October and November to March of the following year each year,and used the pixel-by-pixel fusion image data according to the median method.Using the random forest classification method,combined with data such as spectral characteristics,spectral index,texture characteristics,night light data,and climate factors,we gradually determined each of the 33 years of land cover in the study area through sample distribution optimization,texture feature optimization,climate feature optimization and optimal feature plan determination.Data were classified and the accuracy was evaluated.The final overall classification accuracy reached79.80-92.46%,and the Kappa coefficient reached 0.73-0.89.(2)On the basis of previous studies,taking the construction land in the land cover data of the study area as an example,an algorithm for checking the temporal and spatial consistency of urban construction land in the topographical area of river valleys was realized,and the construction land data for 33 years was subjected to temporal filtering and logical reasoning,and the final classification results were obtained.The overall classification accuracy of the final classification data reached83.33-94.03%,and the Kappa coefficient reached 0.75-0.90.The accuracy of developed land increased to 83.35-97.95%.(3)Analysis of the temporal and spatial changes in land cover shows that from1987 to 2019,land cover changes are characterized by a decrease in cultivated land and grassland area,an increase in developed land and forest,and smaller changes in water bodies and unused land.Arable land and grassland were reduced from 686.74km~2and 731.84 km~2in 1987 to 369.08 km~2and 626.5 km~2in 2019,corresponding to a net decrease of 317.80 km~2and 105.41 km~2,with an average annual change rate of1.44%and 0.45%,respectively.The area of the construction land and forestland increased from 54.67 km~2and 138.75 km~2in 1987 to 293.91 km~2and 334.15 km~2in2019,corresponding to a net increase of 279.76 km~2and 155.17 km~2,with an average annual change of 15.99%and 3.49%,respectively.(4)Based on Geodetector,the driving forces affecting the changes in developed and cultivated land were determined in the study area from 2000 to 2010 and from2010 to 2019.Distance from the main railway had the greatest impact on changes in developed and cultivated land during the period from 2000 to 2010,followed by the distance from the main river,and the distance from the main road.From 2010 to 2019,the distance to the main railway remained the primary driving force for changes to developed and arable land.Changes in total fixed asset investment and unit population were secondary driving forces for changes in developed land,while total investment and slope were the secondary driving forces for changes to cultivated land.Interaction factor detection showed that the changes in developed and cultivated land were driven by both economic and natural factor.Dual factors had a stronger effect on the reduction of cultivated land and the expansion of developed land.(5)Using Geo SOS,the three models of ANN-CA,DT-CA,and Logistic-CA were used to simulate the land cover changes in the study area from 1999 to 2009.The overall accuracy of the three models reached 70%.Among them,ANN-CA had the highest accuracy,at 71.49%.The ANN-CA model was used to simulate the land cover changes from 2009 to 2019,and conversion rules for urban construction land were established.With a limiting factor of slope<15°,this model was used to predict the expansion of urban construction from 2019 to 2025.The results indicated that by 2025,developed land area will expand from its 2019 extent of 334.6 km~2to405.5 km~2,a net increase of 70.9 km~2,with an average annual change rate of 3.53%and an average annual change in arable land of-2.55%.According to the spatial distribution map of land cover predicted in 2025,the river valley plain area in the study area will have essentially been converted into developed land,and the expansion will have reached a saturated state.After 2025,further city development may be eastward along the Huangshui River in the direction of Pingan,and along the Shatangchuan River Valley.