Soil Moisture Memory And Its Response To Precipitation Over The Tibetan Plateau

Soil moisture is a key land surface factor,which plays an important role in the momentum,energy and material exchange between land and atmosphere.Soil moisture anomalies can persist for a long time,and have great effects on subsequent weather and climate.However,due to the complexity of land surface,soil moisture has large spatial variability.Study on the characteristic of soil moisture persistence over Tibetan Plateau?TP?is beneficial for deep understanding the land-atmosphere interaction over alpine cold region.In this paper,the atmospheric forcing field produced by China's regional surface meteorological element dataset was used to drive the Community Land Model version 4.5?CLM4.5?to perform off-line simulations over TP,and the simulation results were compared with observations at three stations?D66,TTH and Maqu,respectively?and Global Land Data Assimilation System?GLDAS?dataset to evaluate the performance of CLM4.5 in simulating soil water-heat transport and analyze the biases of simulated soil temperature and moisture and its causes.Based on the CLM4.5 simulations,the temporal and spatial distribution characteristics of soil moisture memory over TP were analyzed by different methods.The model was also used to simulate the response of soil moisture to different periods and different characteristics of precipitation,and the effect of precipitation on soil under different dry and wet conditions.The following main conclusions have been found:?1?CLM4.5 simulated soil temperature?average RMSE?3°C?is closer to the observation than the GLDAS data?averaged RMSE>6°C?,especially during the freezing-thawing?FT?period;The seasonal variation of soil moisture during FT is well simulated by CLM4.5,but there is a certain deviation between its value and observation(averaged RMSE<0.1 mm³·mm^-3);while,GLDAS-CLM2 cannot reflect the variation characteristics of soil moisture.The simulation biases of CLM4.5 mainly come from the atmospheric forcing field,and the biases of GLDAS-CLM2 are not only due to the uncertainty of the atmospheric forcing field,but also from the imperfection of the FT parameterizations.The effects of air temperature and precipitation on variations of soil temperature and moisture are different between the FT period and non-FT period.During the non-FT period,soil temperature simulation was mainly affected by air temperature.The contribution rate of air temperature bias to simulated soil temperature bias was greater than 50%.The variation of soil moisture was mainly affected by precipitation,and soil moisture bias was related to the bias of precipitation with 20%⁴0%contribution rate of bias.During the FT period,the effects of air temperature and precipitation on soil temperature and moisture were weakened by the interaction between soil water flow and heat transport.The variation of soil moisture was significantly affected by soil temperature,and its contribution rate was 10%²0%.The imperfection of FT parameterizations in the land surface model is one of the major causes of soil temperature and moisture biases during FT period,and its effect on soil water and heat transfer might be greater than impacts of atmospheric forcing field.?2?Comparing two methods,the soil moisture memory?SSM?calculated by the lagged autocorrelation method reflects:in the spatial distribution,SSM in the northwestern TP is longer than the southeastern region by 100 days;in terms of temporal distribution,SSM in the northwestern TP increases from March to May,while SSM in the southeastern TP is the highest in March,which is approximately 90¹50 days,SSM is shortened in May?<30 days?.The improved Markov method is based on the surface water balance equation,which calculates the SSM by considering the hydrothermal process.The spatial distribution of SSM obtained by this method is generally consistent with the statistical method,and SSM in southeastern TP is well described,this method can reflect the feedback between soil moisture and precipitation.?3?Sensitive experiments about the effects of FT process on the SSM results showed that,the effects of FT process on SSM is prominent during the soil thawing period from March to April,soil FT process can make the SSM longer.When the TP rainy season begins?approximately in May?,the SSM is disturbed by precipitation.?4?The ideal experiments with different precipitation characteristics and different wet-dry soil condition show that,the effect of precipitation on soil moisture is affected by precipitation content and its intensity.The greater precipitation leads to the more positive contribution to the soil moisture,under the condition of same amount of precipitation.For soil moisture at different depths,the influence of precipitation on shallow soil moisture is more obvious than that of deep layer.When the soil moisture is dry,more soil water is transported downward into the soil,and the positive contribution of precipitation to the soil moisture is lower than the normal condition.When the soil moisture is wet,the more soil moisture is transported from the shallow layer to the deep layer,and effect of precipitation on soil moisture at shallow layer is larger than the normal condition;the response of soil moisture at deep layer to precipitation is lagged.?5?In different stages of a freeze-thaw cycle,precipitation has different effects on soil moisture.The results of numerical experiments on the effects of precipitation on soil moisture in different periods show that the soil moisture anomalies caused by precipitation anomaly before freezing period can continue through the freezing-thawing process until the late stage of soil melting,and further cause anomalies in the surface non-adiabatic heating;precipitation during the freeze-thaw period have little effect on the following precipitation.,but the abnormal soil moisture can still last until the end of the melting process;precipitation after the end of melting has the most obvious effect on the soil moisture in the spring,especially after the onset of the plateau rainy season.Changes in soil moisture is about 0.05⁰.1mm³/mm³ magnitude.