Inversion Of The Impervious Surface Percentage And Analysis Its Correlation With Land Surface Temperature Of Important Node Cities In The Bangladesh-China-India-Myanmar Economic Corridor

The Bangladesh-China-India-Myanmar(BCIM)Economic Corridor is an important road section of the ancient Southern Silk Road and a significant strategic passage for the construction of the “Belt and Road” Initiative(BRI).The construction of the BRI will inevitably promote the urbanization process,which will lead to significant changes in the the underlying surface structure and type,like many vegetation,water and other natural surfaces are replacing by impervious surfaces such as buildings and roads.The increase of impervious surfaces will inevitably cause the deterioration of urban thermal environment,the rise of land surface temperature(LST),and the formation of urban heat island effect.Therefore,the remote sensing monitoring of the impervious surface percentage(ISP)and the analysis of the correlation between the ISP and the LST of the key cities in the economic corridor will be of great significance to the sustainable development planning and ecological effect cognition to the key cities in the BCIM Economic Corridor.Based on the present research state,this paper took the key cities in the BCIM economic corridor as the research area,developed a medium-spatial resolution ISP inversion model,and obtained the ISP dataset of key cities in the BCIM Economic Corridor.On this basis,this paper analyzed the spatial and temporal pattern differences of the ISP of different cities in the economic corridor.Then selecting the representative node city,Kolkata and analysing the seasonal correlation between ISP and LST.This can provide data and method support for the sustainable development of ecosystem of key cities in the BCIM Economic Corridor.The results showed that:(1)The accuracy of impervious surface percentage inversion model was generally highThe ISP reference value(training sample set and test sample set)and the corresponding model prediction value for each city were ploted and the linear regression model were fitted,and all cities' ISP inversion models showed strong fitting ability and predictive ability at the same time.Based on the 30% independent test data set,the accuracy of the ISP inversion model was verified.The results showed that the accuracy of ISP inversion model was generally high: the Mean Absolute Error(MAE)values were from 6.1% to 7.2%,and the Root Mean Squared Error(RMSE)values were from 13.85% to 15.04%.(2)The development of different size cities was slow and the pattern was different in economic corridorAccording to the spatial distribution of ISP of eight important node cities in Myanmar,Bangladesh and India,the backward transportation infrastructure hindered the development of cities and prevented them from making better use of their own geographical advantages and resources.The urban development of Yangon,Mandalay,Chattogram and Ruili were limited by natural factors,such as terrain and rivers,cities will develop better if they can break through the restrictions of natural factors.The urban development direction of Baoshan and Dhaka was the most obvious,Myitkyina and Ruili had the weakest direction of urban development,that of the other six cities was more obvious.The impervious surface landscape types of medium-sized cities were more heterogeneous and more balanced than those of large-sized and small-sized cities.The medium density and extremely high density impervious landscapes in small-sized cities had the highest Patch Density(PD)values and the highest degree of fragmentation.The high density impervious landscapes in medium-sized cities had the highest Perimeter-Area Fractal Dimension(PAFRAC)values and the highest complexity of shape.The low density,high density and extremely high density impervious landscapes in large-sized cities had the highest Patch Cohesion Index(COHESION)values and connectivity.(3)Correlation between LST and ISP in different seasonsIn the study area,the dense vegetation coverage area and water body area were low temperature or relatively low temperature areas,and the urban residential area was medium temperature and above area.In general,the distribution of low temperature and high temperature areas were relatively concentrated,and the high temperature area was concentrated in the built-up area.In the dry season,the temperature of the study area was from 26.54 ? to 48.99 ?,and the areas of low temperature and relatively low temperature were the least;In the rainy season,the temperature was from 20.93 ? to 41.78 ?,and the area of sub-medium temperature was the least;In the cool season,the temperature was from11.43 ? to 27.50 ?,and the areas of sub-high temperature,high temperature and extra high temperature were the least.The overall heat island effect of Kolkata had been decreasing from the dry season to the rainy season and then to the cool season,and the urban heat island effect was the strongest in the dry season and the weakest in the cool season.In each season,the LST was positively correlated with the ISP.When the ISP increased,the LST increased.In dry season,the LST increased fastest,and in cool season,the LST increased slowest.When the ISP increased by 0.1,the LST increased by 0.53 ? in dry season,0.35 ? in rainy season and 0.26 ? in cool season.