| According to the “National New-type Urbanization Plan(2014-2020)”,urban agglomerations will be the main form of urban development in China in the future.Compared with a single city,the unreasonable spatial layout of urban agglomerations will produce superimposed effects of various ecological and environmental problems.Currently,urban agglomerations in most developed countries have already matured and face fewer problems;while China is still in the exploration stage of urban agglomeration development.Thus,studying the spatiotemporal evolution of urban agglomerations in China will have significant theoretical and practical implications for the healthy development of urban agglomerations in the future.Many studies have been conducted on the expansion of urban construction land in terms of characteristics,mechanism,prediction and impacts,but there are still several shortcomings in general:(1)In terms of research data,most of the existing data sources have either coarse spatial or temporal resolution.While the cost of the collection and processing of datasets with medium-high spatial and temporal resolution on large scale is high,which hinders the continuous and accurate observation of urban expansion process.(2)In terms of evolution characteristics and mechanism analysis,most of the studies,on the one hand,divide the expansion stages mechanically based on data collected from several time points,which cannot grasp the exact stages of urban expansion;on the other hand,focus more on megacities or were conducted at a single scale.However,the comprehensive cognition and analysis of the spatial evolution of the less developed small and medium cities,which will be the main body of future urbanization,is still lacking.(3)In terms of urban growth simulation and prediction,SLEUTH has been widely used all over the world due to its good scalability and flexible self-adaptive modification mechanism.However,most of the SLEUTH-based studies ignore the spatial nonstationarity of the driving factors,resulting in unreliable forecasting results.To address the above issues,this study took one of the five national urban agglomerations with the highest proportion of small and medium-sized cities(80.65%at the end of 2019),the middle Yangtze River basin(MYRB)urban agglomeration,as the study area.Based on remote sensing,GIS and statistic data,the spatial and temporal evolution of urban land in MYRB urban agglomeration in the past three decades was explored by the route of information extraction,phenomenon discovery,mechanism exploration as well as simulation and prediction by integrating remote sensing and GIS techniques,mathematical and statistical methods as well as econometric and cellular automata models,aiming to provide scientific reference for the study of less developed urban agglomerations in China.The specific work includes:(1)Semi-automatic large-scale and long time series urban land mapping for MYRB urban agglomeration by integrating the crowdsourced OpenStreetMap(OSM)data with free Landsat images.Random training samples were generated based on the updated OSM data and used for image classification for the latest year.For each historical year,training samples were obtained with a proposed transferring schema by which only the samples whose classes were unchanged were selected through a change detection analysis.By taking advantages of the high-performance computing capability of Google Earth Engine(GEE)as well as the large volume images archived on it,we realized a semi-automatic mapping framework to map urban land in MYRB urban agglomeration from 1987 to 2017.According to the accuracy assessment results,the overall accuracies of all years were above 96% and the average Kappa coefficient was0.75.Our product,when compared with other current 30 m land-cover products,showed similar accuracies but more spatial details.(2)Spatiotemporal analysis of the evolution characteristics of urban land in MYRB urban agglomeration from 1987-2017.Based on the extracted time-series urban maps,the overall temporal characteristics and spatial divergence of urban expansion process(growth type,speed and intensity,direction and form)and pattern(fragmentation,complexity and aggregation)in MYRB urban agglomeration were investigated at urban agglomeration,prefecture and grid scale,respectively,by using mathematical and statistical indices,GIS spatial analysis methods as well as landscape metrics.Results demonstrated that: in terms of scale,1)the construction land of MYRB has grown exponentially in the past 30 years with no obvious spatial differentiation among different cities.2)The urban expansion in MYRB experienced four stages in terms of speed and intensity: low-low speed,low-high speed,medium-high speed,and high-high speed.The time nodes were 1998,2006 and 2009,respectively;the fastest expansion area in the four stages shifted from Jianghan Plain and the west of Dongting Lake Plain to Changsha-Zhuzhou-Xiangtan(CZT)Metropolitan Area and its surrounding area,the east of Wuhan Metropolitan Area(WMA)and the whole urban agglomeration,respectively;the expansion speed of small and medium-sized cities gradually exceeded that of the large cities;the trajectory evolution of the hotspots of expansion speed and intensity at grid scale indicated that WMA and CZT have formed a single core layer and planar development mode,respectively,while Poyang Lake Metropolitan Area has not yet formed.In terms of spatial characteristics: 3)the urban expansion pattern in the past three decades was dominated by edge expansion and showed a significant upward trend,supplemented by infilling and outlying expansion,but showed no significant trend;no spatial differentiation in the dominant expansion mode was found among different cities.4)The new construction land in MYRB mainly distributed along the “northwest-southeast” axis,and the expansion trajectory gradually moved from the northwest to the southeast,but there were obvious dominant expansion wings among different cities.In terms of landscape pattern evolution: 5)the degree of fragmentation of the construction land in MYRB decreased continuously by 33.01%,while the degree of complexity and aggregation increased by 32.50% and 28.25%,respectively,and there existed spatial divergence in the distribution of urban landscape pattern change,among which the fragmentation degree in western Jianghan Plain and Dongting Lake Plain showed an upward increase,and the increase of landscape complexity in the right region bounded by Jiujiang and Changsha was higher than that in the left region,while the regions with greater increase of landscape aggregation and connectivity were mainly concentrated in the northeast and eastern regions.(3)The theoretical driving mechanism of urban expansion was analyzed based on a supply-demand schema.By integrating the econometric model,logistic regression(LR)model,generalized additive model and geographically weighted logistic regression(GWLR)model,the main driving forces influencing the area growth and expansion direction of urban land and the spatiotemporal evolution of their influences were explored from the perspectives of macroeconomics and micro-geographic location,respectively,by using statistical and geospatial data.Results showed that the main cause of the area increase of construction land in MYRB urban agglomeration was the capital construction investment,and the government promoted capital construction investment through fiscal budget expenditure.Population migration had significant negative effect on the area increase,which is different from the other three coastal urban agglomerations;capital construction investment had a continuous influence,while real estate development investment gradually started to have an impact on the megacities;there existed heterogeneity of the influences of the driving factors among cities of different levels.The global factors affecting the expansion direction of urban land were mainly social location and supplemented by natural location,which did not change with time.Jianghan Plain was mainly affected by topographic conditions,and Poyang Lake Plain and Dongting Lake Plain were mainly affected by water,while the north,central,and eastern regions were all mostly affected by transportation infrastructures.As time goes on,the constraint intensity of natural environment on urban expansion gradually decreased,and the influence range of road networks on urban expansion direction generally increased from 2000 to 2017,indicating that transportation infrastructures have a guiding role on the spatial form of urban land.(4)A new exclusion layer that combining the urban development suitability surface obtained from GWLR and the standard exclusion layer was generated and used as the new exclusion layer for SLEUTH calibration.Different scenarios were designed to make possible urban projections taking WMA as an example.Results showed that the coupled GWLR-SLEUTH model performed better than either the standard SLEUTH model or the LR-SLEUTH model according to both the calibration and the accuracy verification indices,proving that the employment of GWLR model can improve the prediction accuracy of SLEUTH.By comparing the prediction results under different scenarios,it was found that the land consumption of WMA will be reduced by 9% and 26% under the scenarios of “development under future master plan”and “ecologically sustainable growth”,respectively,indicating that future urban planning policy can mitigate the excessive growth of urban land if being well obeyed and the land urbanization of WMA in the future will still at the expense of the consumption of arable land according to the result of the third scenario.It will be of great practical significance to further explore the impact of urban expansion on the hydrological cycle in MYRB by integrating urban growth models with hydrological models in the future. |
PDF Full Text Request