Study On Antarctic Snowfall Reconstruction And Its Impact Using Multi-source Data

The mass balance of the Antarctic Ice Sheet(AIS)has an important influence on global sea-levels due to global warming.Snowfall is the most important process of mass gain,but changes of snowfall and its influence on the AIS mass balance and surface mass balance have not been studied in depth.Owing to the harsh environment,there is no snowfall observation network on the AIS,and the continuous snowfall data are almost blank especially on the Antarctic plateau.Most studies only use indirect snowfall data to carry out relevant research on the AIS.This not only affects the research on Antarctic snowfall and its changes,but also hinders the research on the relationship between snowfall and ice mass balance.Furthermore,it restricts the improvement of the simulation accuracy of climate model.In this study,a new method is present to construct the Antarctic snowfall data using multi-source data(such as Cloud Sat reflectivity data,ICESat data,MODIS cloud mask data and ERA5 data)in order to make up for the blank of continuous snowfall data on the AIS,improve the understanding of influence of snowfall on the ice mass balance and surface mass balance,and improve the accuracy of assessment result for snowfall data from reanalysis.The new method improves the accuracy of Antarctic snowfall data,and make up the blank of continuous snowfall data on most regions of the AIS.The main research contents and achievements of this paper include:(1)Multi-conditional constraint method has been used in this paper to retrieve Antarctic snowfall data based on the reflectivity data.By analyzing the problems in Cloud Sat product,the Antarctic snowfall data is re-retrieved using the multi-source data based on the multi-condition constraint.The multi-conditional constraint method includes: using ICESat data to construct Antarctic DEM data to evaluate the surface roughness to determine near surface layer position,using MODIS cloud mask data to help determine snowfall conditions,and using ERA5 temperature data to determine precipitation patterns.The results suggest that,the re-retrieved Antarctic snowfall data is closer to the real snowfall data,and the deviation is reduced by 63% compared with the original Cloud Sat snowfall product.The large deviation between the original data and the real snowfall data indicates that it is necessary to improve the accuracy of Antarctic snowfall data.(2)A new method for calculating continuous snowfall data on the AIS at a synoptic timescale is proposed base on the Cloud Sat snowfall data.The continuous snowfall data fills the gap of direct snowfall data over most areas of the AIS.On the basis of the retrieved snowfall data and the characteristics of Cloud Sat footprints,the feasibility of constructing synoptic snowfall data using Cloud Sat data has been evaluated.The temporal resolutions of synoptic snowfall data on the different regions of the AIS are also examined.The result suggests that synoptic snowfall data constructed by this method has good accuracy and greatly enrich the existing snow observation database.(3)The accuracy of ERA5 snowfall data has been evaluated at different timescales using snowfall data constructed in this study and indirect snowfall data.The results show that the reconstructed snowfall data are agreed with ERA5 snowfall data at different timescales and ERA5 has a high capture rate of snowfall events(?71.5%).However,the evaluation results based on indirect snow data showed that the indirect snowfall data only have significant correlations with reanalysis data on the monthly and annual scale,and the correlation is insignificant at a synoptic time scale.This indicates that the indirect snowfall data cannot represent the real synoptic snowfall data,and it is necessary to use real snow data to evaluate the accuracy of the model.Furthermore,even if no Cloud Sat data have been assimilated into ERA5,two datasets show good agreement that enhances the reliability of these two snowfall datasets.(4)The influence of snow on the change of ice mass balance and surface mass balance,and the influence of drifting snow on the relationship between snowfall and snow accumulation are analyzed.The results show that the average wind speed and maximum wind speed of extreme snowfall events are higher than usual,and some extreme snowfall events do not correspond to the snow accumulation.32% of the extreme accumulation events occurred without snowfall.At the long timescales,the ice mass is consistent with the snowfall in most regions of the AIS.(5)Mass loss of the AIS has a significant relationship with the several sea conditions in the Southern Ocean.Through the correlation analysis,the results show that in different ecological areas of the Southern Ocean,ice mass loss has significant fertilization effect on chlorophyll concentration and net primary productivity.Such fertilization mechanism will improve the carbon stock of the ocean suggesting a important negative feedback effect on global climate change.