Analysis Of Lake Surface Change And Natural Driving Factors In Ebinur Lake Based On Machine Learning In Recent 30a

Lakes in the arid area play an important role in global change.Due to the scarcity of observational data,relevant hydrological research was still insufficient.However,the land lakes in the arid regions play a decwasive role in the local ecological environment and economic development.It was this asymmetrical relationship that maked ecological security wassues more frequent.As the ecological barrier of the economic belt on the northern slope of the Tianshan Mountains,Ebinur Lake had experienced dramatic fluctuations in its lake surface area and had become the second largest ecological problem in Xinjiang.Therefore,it was important to accurately monitor the dynamic changes of Lake Ebinur.The rise of machine learning has made high-precwasion recognition possible.Only by accurately grasping the lake surface area of Lake Ebinur will it be helpful to determine the mechanwasm by which the relevant driving factors cause lake changes.Based on this,this paper uses the integrated learning algorithm to obtain the lake area in different seasons and determine the importance of different driving factors through the Landsat dataset for nearly 30 years.Analyze the delayed response relationship between climate hydrological elements and lake surface area,and try to explore the potential ecological threshold of Lake Ebinur lake area to provide a baswas for regional sustainable development.The main conclusions of this study were as follows:(1)The random forest identification system performed well with a kappa coefficient of 0.99.The DEM maintained the highest importance score in the image data under each sensor.Lake Ebinur had two peaks in the past 30 years,namely in 2003and 2017.In 2003,the lake area of Ebinur Lake was the largest(the largest lake area was 965.22km2).Sorted during the year as Spring>Summer>Autumn,summer-autumn lake loss was more than spring-summer loss.(2)The whole basin showed obvious warming and humidification,and the runoff increased significantly,while the total amount of lakes only increased slowly,and the potential evapotranspiration showed a decreasing trend.There were two complete cycles of temperature,runoff and potential evapotranspiration in the Ebinur Lake Basin.There were three complete periodic oscillations in average precipitation and total lake intake.(3)The GBRT algorithm ranked the importance of climate and hydrological factors influencing the change of lake area,and Bohe entered the lake with the highest importance.The temperature in the basin was subject to change in the area of the lake;the average precipitation and the amount of water entering the lake remain basically synchronous positive correlation period after 2004;there was a synchronous positive correlation between runoff and lake area;the potential evapotranspiration had a negative correlation with the lake area,and the intensity gradually weakens.Exploring the ecological threshold of Ebinur Lake area on the principle that the ecology-environment-economy was as balanced as possible.In 2012,the Lake Ebinur Lake area was 606.5km~2.Based on the climatic and hydrological elements,the lake area was estimated to be 615.21 km~2 in 2012.