Variations Of Ice-on Date And Its Key Controlling Factors In Typical Lakes Of The Qinghai-Tibet Plateau Revealed By A Lake Temperature Model

In northern China and the Qinghai-Tibet Plateau region,seasonal ice formation processes exist in lakes,and the timing of ice-on and ice-off dates,the duration of ice cover,and changes in ice thickness all have a significant impact on lake-atmosphere interactions,lake environment,and ecology.Under climate warming,the phenology of lake ice is highly sensitive,which inevitably drives the evolution of the ecological environment of cold-region lakes.The ice-on date is a critical parameter in the characterization of lake ice phenology,as it impacts the initial state of the winter lake ecosystem.The lakes in the Qinghai-Tibet Plateau region,which is one of the areas most affected by global climate change,lack sufficient historical observation data to establish long ice-on date sequences due to various limitations.Therefore,there is a need to further investigate changes in ice-on dates and identify the primary driving factors.This article focuses on three lakes located on the Qinghai-Tibet Plateau,Lake Ngoring,Lake Gyaring,and Lake Qinghai.To reconstruct the historical ice-on dates of these lakes from1979 to 2021,the FLake and SELF one-dimensional water temperature models were coupled,with meteorological reanalysis data serving as the driving factor for the models.The reconstructed ice-on dates were validated using MODIS water surface temperature and lake ice remote sensing products.By sensitivity experiments,the study identified the primary factors that controlled the historical evolution of ice-on dates.Furthermore,the article predicted future ice-on dates for the three lakes based on CMIP6 climate prediction data for different scenarios.The main findings and conclusions are as follows:(1)Based on remote sensing data on lake ice phenology,during the period from 2002 to2015,for the average ice-on date,Lake Qinghai was latest(occurring on December 15th),and Lake Ngoring was earlier(occurring on November 30th),while Lake Gyaring was at the earliest(occurring on November 15th).All three lakes exhibited a trend of delayed ice-on dates over time.Additionally,the average date of complete ice-off was the earliest for Lake Qinghai(occurring on April 4th)and the latest for Lake Ngoring(occurring on April 30th).Lake Qinghai had the shortest and most stable ice cover duration,whereas Lake Ngoring and Lake Gyaring had similar durations,both of which exhibited a trend of shortening.It’s also shown that,under the same climate conditions,deep lakes tend to have more stable ice cover duration than shallow lakes(2)To improve the simulation of autumn mixing processes,the SELF model was coupled with the FLake model,resulting in a revised FLake-SELF model that demonstrated improved accuracy in ice-on date simulations.Specifically,the root mean square error for ice-on date simulations was reduced by 13.8,7.1,and 5 days for Lake Qinghai,Lake Ngoring,and Lake Gyaring,respectively,compared to FLake alone.Using this revised model,historical ice-on date series for the three lakes from 1979 to 2021 were reconstructed.Trend analysis revealed that Lake Ngoring had the fastest delay rate in ice-on dates,at 4.5 days per decade,followed by Lake Gyaring at 3.8 days per decade,while Lake Qinghai had the slowest delay rate at 2.8 days per decade.(3)Single-factor numerical experiments on climate change revealed that air temperature increase is the primary cause of ice-on date delay,followed by water vapor pressure and wind speed.Moreover,air temperature is also a controlling factor for interannual variability of iceon dates.A comparative analysis between Lake Ngoring and Lake Qinghai showed that the difference in their ice-on dates is mainly attributed to Lake Qinghai’s higher temperature and slower warming rate than Lake Ngoring.Furthermore,numerical experiments conducted on lakes with different depths indicated that deep lakes have later ice-on dates,greater delay rates,and more significant interannual variability,making them more susceptible to climate change compared to shallow lakes.(4)The sensitivity analysis of meteorological factors revealed that air temperature,solar radiation,and lake depth are the most influential factors affecting the ice-on dates.Using different climate change scenarios in CMIP6,the future changes in ice-on dates for the three lakes from 2021 to 2100 were predicted.The results suggest that under the SSP1-2.6 and SSP2-4.5 scenarios,the rate of change of ice-on dates in the future is smaller than that from 1979 to2021.However,under the SSP3-7.0 scenario,the delay rate for Lake Ngoring and Lake Qinghai is greater than the historical value,and the frequency of Lake Qinghai not freezing increases after 2048,while the delay rate for Lake Gyaring remains similar to the historical value.Under the SSP5-5.8 scenario,the ice-on dates for Lake Ngoring and Lake Gyaring are delayed by 2months and 1.5 months,respectively,by 2100,and Lake Qinghai becomes a non-ice lake after2076.