Spatio-temporal Evolution Pattern Of Land Use And Assessment Of Environmental Carrying Capacity In Typical Port Areas Along The Belt And Road

In 2015,China formally proposed the Belt and Road Initiative.As an important core area of international trade and economic cooperation,Port areas along the Belt and Road have developed rapidly and attracted more attention.Thus,it is necessary to explore the process,driving factors and environmental effects of land use transformation on port areas along the Belt and Road in recent years.In this study,Landsat remote sensing image data and socio-economic data from 2005 to 2020,were used to analyze the spatio-temporal evolution pattern of land use,assess environmental carrying capacity,investigate characteristics of construction land expansion and its socio-economic driving forces in typical port areas along the Belt and Road,in order to fully understand the law of development and effectively provide reliable reference for future development of port areas.This study is helpful to transform the port area research from point to surface and from qualitative to quantitative.The main contents and conclusions include:（1）To construct more accurate model for land use classification,Landsat satellite imagery of port areas in 2020,three indexes,and four texture features extracted by Principal component analysis were used in four models including maximum likelihood model,support vector machine model,BP neural network model,and random forests model.Producer’s accuracy,overall accuracy and Kappa coefficient were used to assess the accuracy of models.The results show that the overall accuracy and Kappa coefficient of random forest models are higher than those of support vector machine models in Asian port areas,and slightly higher than those of BP neural network models in European,African,Oceanian port areas.Meanwhile,the accuracy of maximum likelihood models is lowest.In summary,compared with other methods,random forest is better for use with land use classification on these port areas.（2）Based on the Landsat satellite imagery from 2005 to 2020,the land use information of 10 port areas was extracted by random forest method,and analyzed the land use structure characteristics and spatial distribution.In addition,the spatiotemporal evolution pattern of land use in each port area were compared and discussed from three perspectives of change rate,change direction and structure evolution.The results suggest that forest land and urban construction land are the main land types in these port areas.On the whole,land use changed rapidly from 2005 to 2020.During this period,urban construction land area increased continually,while forest land area decreased first and then increased in Asian port areas,and decreased continually in European,African and Oceanian port areas.The land use structure of most Asian port areas returned to a relatively orderly state from a relatively disorderly state,except West Asian port areas,which remained in a disorderly state,and the land use structure of port areas in Europe,Africa and Oceania maintained a stable state from 2005 to 2020.（3）To analyze spatio-temporal distribution of habitat quality,habitat degradation degree and its cold/hot spots,which can reflect the environmental carrying capacity of each port area,the InVEST model and Gi^* coefficient were used to calculate the habitat quality,habitat degradation degree and cold/hot spots.The results indicate that the habitat quality of 10 port areas was on the decline in the past 15 years,and the habitat quality of Asian port areas decreased slowly after 2015.The spatial distribution of habitat quality showed a gradual decrease from the sea area far from land to the coastal sea area near construction land,especially port land,and area of low habitat quality was expanding from 2005 to 2020.These results suggest that the expansion of port land and urban construction land is the main reason of habitat quality decline.The hot spots of habitat degradation degree were mainly distributed in the sea area bordering the land and the coastal land area,and the cold spots were mainly distributed in the marginal sea area and the land area far from the construction land.The number of hot spots were increasing from 2005 to 2015,but it remained unchanged or even gradually decreased after 2015.（4）To further explore the law of the construction land expansion and its driving factors,the annual increase and the annual growth rate were calculated as indexes of the construction land expansion size and rate,and some social and economic variables were used to analyze correlation and build the individual fixed effects regression model.The results suggest that construction land expansion size and rate in Asian port areas increased significantly after 2015,while they showed a downward trend from 2005 to2020 in European,African and Oceanian port areas,because of their earlier development.Urbanization rate and the proportion of import and export trade with China are the main social and economic factors influencing the construction land expansion,which indicates that urbanization development and rational allocation of industrial structure are conducive to construction land expansion,and the degree of trade cooperation with China has also become an important factor in promoting the construction land expansion,especially after proposing the Belt and Road Initiative.