Estimation And Change Analysis Of Lake Surface Water Temperature Of Nine Lakes In Yunnan Province

Lake is not only an important ecological resource,but also a carrier of a series of social and economic processes.The quality of the lake's ecological environment determines the quality of the regional ecological environment and the sustainable development of the city.At present,an important physical parameter in the field of lake research-Lake Surface Water Temperature(Lake Surface Water Temperature)has become a hot spot of concern.Lake water hydrodynamics,physical properties and chemical reactions,substances required for the growth of aquatic organisms and energy metabolism are inseparable from the lake surface water temperature.^[1].Many studies in the field of lake research are carried out on larger spatial scales and longer time scales.The development of information technology promoted the application of artificial intelligence and gave birth to machine learning.Machine learning,an important branch in the field of artificial intelligence,is widely used.Yunnan Province is one of the provinces with the most natural lakes.There are 37 lakes with an area of less than 1square kilometer.Among them,there are 9 lakes with a lake area of more than 30 square kilometers.Dianchi Lake is in Kunming.Fuxian Lake.Qilu Lake and Xingyun Lake is in Yuxi.Erhai Lake is in Dali.Chenghai Lake is in Lijiang.Yilong Lake is in Honghe.Yangzonghai Lake is on the border of Kunming and Yuxi.Lugu Lake is on the border of Yunnan and Sichuan Provinces.Since the reform and opening up,rapid economic development and rapid urbanization have caused many irreversible effects on the lake's water environment.It is necessary to do further research on lake water resources to protect the lake water ecological environment from damage from its root causes.In this study,we selected nine typical plateau lakes in Yunnan as the study area.Three meteorological factors(air temperature,wind speed,solar radiation)in each basin from 1981 to 2018 and MODIS 11A2 surface temperature products of each lake from 2001 to 2018 were used as basic dataset.The artificial neural network(ANN)method constructed a estimation model to estimate the lake surface water temperature of each lake from 1981 to 2018.The least squares linear regression method,Mann-Kendall test and Pearson correlation coefficient were used to analyze the change characteristics of lake surface water temperature and air temperature from 1981 to 2018,mainly from four aspects:trend change,significance test,correlation and sudden change.Four evaluation indicators were used to measure the performance of the estimation model.The model had good performance with small errors and high accuracy,and could accurately estimate the lake surface water temperature of historical lakes.The results of the study showed that the water extraction methods applicable to each lake are different.AWElsh was used to extract Dianchi Lake and Lugu Lake;NDWI was used to extract Fuxian Lake,Xingyun Lake and Qilu Lake;EWI was used to extract Yangzonghai and Chenghai;NWI was used to extract Longhu Lake and Erhai Lake.The monthly average lake surface water temperature of Chenghai from 1981 to2018 showed an upward trend with an average increase rate of 0.1198°C/year.The monthly average lake surface water temperature of the remaining eight lakes showed different degrees of decline in different months.Except for the summer average lake surface water temperature of Lugu Lake and the spring average lake surface water temperature of Yangzonghai,which had an insignificant downward trend,the other seasonal average lake surface water temperature were warming.Except for Lugu Lake,the annual average lake surface water temperature of the nine plateau lakes all showed warming trends significantly.The monthly average air temperature of Lugu Lake and Xingyun Lake showed a downward trend in individual months,and the monthly average air temperature of the remaining lakes increased significantly in all months from 1981 to 2018.The seasonal average air temperature of all lake basins showed a significant warming trend.The annual average air temperature of the basin warmed significantly,which passed the significant test at the 0.1 level.The mutation test results showed that Dianchi Lake,Fuxian Lake,Qilu Lake,Xingyun Lake and Yangzonghai all had two uncertain mutation time points from 1997 to 2001,and the specific mutation time should be determined in conjunction with other testing methods.The abrupt characteristics of the annual average lake surface water temperature of Chenghai and Yilong Lake were consistent with those of the annual average air temperature.Dianchi Lake,Lugu Lake and Yangzonghai have the same temperature changes in the same basin,and there are two sudden changes in the significant interval.The sudden changes of air temperature in Dianchi Lake,Lugu Lake and Yangzonghai were the same,with two mutation time points in the significant interval.Correlation analysis results showed that the average annual lake surface water temperature and the annual average air temperature were all significantly correlated,and the seasonal average lake surface water temperature and the seasonal average air temperature were also significantly correlated except for summer.