Study On Identification Method Of Dominant Factors Of Lake Area Disturbances In Yunnan Province Based On Time Series Remote Sensing Data

Lakes are important carriers of freshwater resources and maintain the natural ecological balance.Inland lake variations are considered sensitive indicators of global climate change.For the lakes with frequent human activities in the river basin,whether the influence of human activities on the changes of the lake is dominant or whether the natural factors are the main factors that cause the changes of the lake area is crucial to the study of lake ecological security,lake evolution and river basin water ecological security.Based on this,this paper uses remote sensing technology to establish a time series based on the lake area changes in Yunnan Province in the past 31 years,and proposes a time series mutation recognition method based on Douglas-Peucker algorithm and Bend simplification algorithm to realize the recognition of lake disturbance.The accuracy of the method was evaluated with the lake disturbance events recorded in the official records.The results show that:(1)From 1987 to 2017,the total area of the 23 lakes in the study area showed an overall trend of fluctuation reduction.In 2013,the total area of the lake decreased to a minimum value(1034.21km2).Compared with 1987,the total area of the lake decreased by 73.99 km2(7.15%).After 2013,the total area of the lake has increased year by year;Among the 23 lakes in the study area,the area of Qilu Lake and Yilong Lake had the most serious reductions.In 2016 and 2015,the lake area was reduced to the lowest value.Compared with 1987,the lake area decreased by 21.48km2(46.25%)And 30.71km2(71.08%).The spatial changes of Qilu Lake are concentrated in the south and southwest of the lake,and Yilong Lake changes around the lake,with the most dramatic changes in the west,southwest,and northwest of the lake;In addition,the area of the Sanjiaohai Lake,Chaheihai Lake,and Wenhai Lake with relatively small lake areas fluctuates greatly,and the lake change rate is as high as 500% in some years;Among the lakes selected in the study,only Bitahai Lake has not been subject to large human disturbances,and the lake area has a small fluctuation.Yuxian Lake was severely disturbed by nature and man-made disturbances.The area of the lake was completely dried up between 2012 and 2014.(2)The disturbance event recognition method based on the lake area time series curve proposed in this study can accurately locate the main lake changing events,and accurately classify the disturbances identified as human or natural disturbance events.The accuracy of this model for segmenting the time series of lake surface area in our study area was 94.73%.Our proposed disturbance classification method achieved an overall accuracy of 87.75%,with an F-score of 85.71 for anthropogenic disturbances and an F-score of 88.89 for natural disturbances.(3)Lakes in Yunnan Province were strongly disturbed from 1987 to 2017.A total of 42 disturbance events occurred in the lakes in the nine selected research areas.The frequency of man-made disturbances is almost twice that of natural disturbances,indicating that the disturbance of human activities on lakes has increased,and the disturbance of human activities on lake areas is more intense and lasting than natural disturbances.For example,in 1993-1995,Qilu Lake 2010-2015,Yilong Lake 2010-2016,Lashihai Lake 1992-1993 and Yuxian Lake 2006-2011,the lake area affected by human activities increased(or decreased)1.4,1.5,2.2,2.03 and 2.1 times of the original area before and after the change,respectively.The lake disturbance dominated by natural factors,such as Shudu Lake in 1998-1999,Bitahai Lake in 1990-1992,Lashihai Lake in 2010-2013 and Haixihai Lake in 2010-2013,increased(or decreased)to 1.2,1.1,1.1,1.2 times of the original area before and after the change,and recovered to the same level in a short time(within 2 years in this study).Lake disturbance event identification method is of great significance to reveal whether inland lake disturbances are affected by human activities or natural events,and can monitor whether lake disturbances in a certain area have intensified.