Study On Extraction Method Of Water Level In Three Rivers Headwaters Lake Based On Multi-source Remote Sensing Data

The Three-River Headwaters region is a sensitive area and a significant area affected by the global climate change response.The changes of lake water situation are an important carrier of the response to climate change.Studying the changes of lake water situation is helpful to evaluate the response relationship of hydrological process to climate change.The climate and geographical conditions in the Three-River Headwaters Region are complex,with poor traffic and communication,and the ground monitoring data in this region are scarce.With the strengthening of earth observation technology,the multi-temporal resolution remote sensing technology has brought a new opportunity for the continuous and rapid monitoring of lake water level in a large range.This paper investigated and analyzed the effects of different algorithms and data to retrieve the water level of lakes in the Three-River Headwaters Region.Based on Cryo Sat-2,Sentienl-3A/B and ICESat-2 satellite altimeter data and Landsat images,waveform retracking algorithm was used to construct the water level series of 28 lakes in the Three-River Headwaters Region from 2010 to 2020.The specific research contents and results are as follows:(1)Comparison and analysis of waveform re-tracking algorithms for satellite altimeter with different measurement modes.Based on the Cryo Sat-2 LRM and Sentinel-3 SAR L1B waveforms,the lake water level was extracted by using the OCOG,TR,ICE-1,NPPR-COG and the NPPR-TR(30%,50%,80%)method,and the accuracy was evaluated by using the measured water level of the station.The results showed that the Sentinel-3 were optimized by NPPR-TR(80%in the inner lake and 30%in the outer lake),Cryosat-2 LRM were optimized by NPPR-30%TR.(2)Quality comparison of multi-source satellite altimeter data.The six water levels calculated by Cryo Sat-2,Sentinel-3 and ICESat-2 in four altimetry modes of LRM,SARIN,SAR and laser were compared in Qinghai Lake,and the quality of altimeter data was evaluated by combining waveform analysis.The results show that the Sentinel-3 data is of higher quality and has a wide application prospect.The Cryo Sat-2 SARIn data and ICESat-2 ATL13 data products can be directly used to extract the water level of the three rivers source lakes.(3)Constructing the time series of lake water level in the source area of the three rivers.Based on the ground trajectory of satellite altimeter and the time span of data,28 lakes were selected.Sentinel-3 used the NPPR-TR method(80%in the inner lake and 30%in the outer lake),Cryo Sat-2 LRM used the NPPR-30%TR method,Cryo Sat-2 SARIn and ICESat-2 ATL13 used the secondary data to construct the time series of lake water level at the Three-River Headwaters region.The correlation coefficients between Qinghai Lake,Eling Lake and Zhaling Lake and the measured data are all greater than 0.9,and RMSE is controlled within 0.2m.The R~2 of the lakes cross-verified by the datasets were all above 0.84,and the maximum value was 0.999,indicating that the water level sequences extracted in this paper were highly reliable.(4)Analysis of water level change and its causes in the Three-River Headwaters region.To average rate of change of fixed on water level,the peak water level change trend,years and years for the variation range of water level evaluation indexes,combining with rainfall and temperature change rule of three lakes water level rise on the whole,the Yangtze-river Headwaters basin and precipitation variations in temperature was increased more than the Yellow-River Headwaters basin and the Lancang-river Headwaters basin is slightly larger,the whole year large variation range of water level.