Study On The Variation Of Surface Energy Flux Distribution In The Taklimakan Desert Using Remote Sensing Data

The unique interactions between land and atmosphere in the Taklimakan Desert contribute significantly to regional and global climates and atmospheric circulation.It plays an important role in the climate formation in the arid regions of northwest China and the development of the East Asian monsoon circulation,and also plays an important role in the occurrence and development of severe weather?such as sandstorms?in the northwest,so it is necessary to fully understand the changes in the surface energy flux in desert area.The research on the energy flux of land surface process has been carried out earlier in the world,but with the deepening of the research and the expansion of the research area,the traditional observation of each component of the surface energy flux based on the point scale is not applicable.The development and maturity of remote sensing technology provides a technical means for the study of regional scale,and it is of great significance to the large-scale observation and dynamic monitoring of surface energy flux.In this paper,remote sensing technology is used as the main means to input SEBS?Surface Energy Balance System?model by using various original data and product data of MODIS?Moderate-resolution Imaging Spectroradiometer?image,combined with the scheme of surface feature parameters and meteorological raster data in accordance with the actual situation of the underlying surface of the study area.The spatial and temporal distributions of surface radiant flux,soil heat flux,sensible heat flux and latent heat flux in the Taklimakan Desert and its surrounding areas in the 12months of 2017 were analyzed,and the components of the surface energy flux in the study area were analyzed and discussed.The main conclusions of the study are as follows:?1?From the perspective of time distribution,the net radiation,soil heat flux and sensible heat flux simulated by SEBS model show the highest seasonal variation in summer and the lowest in winter.The overall average of the four seasons of net radiant flux in desert areas are:348.2 W·m^-2,453.4W·m^-2,376.1W·m^-22 and238.2W·m^-2,and the simulation results of net radiant flux are compared.Well,the observation error with the Tazhong station is between-9.6 and 16.5W·m^-2,and the winter error is higher than that in summer.The overall mean value of soil heat flux is94.8W·m^-2 in summer,92.4W·m^-2 in spring,77.4W·m^-2 in autumn,and 55.9W·m^-2 in winter.The error between the upper simulated value and the actual observed value is between±10W·m^-2,and the simulation results are generally overestimated except winter.The variation of sensible heat flux is consistent with the net radiation.The average values of sensible heat fluxes in the desert,spring,summer,autumn and winter are:229.8W·m^-2,296.9W·m^-2,193.4W·m^-2,83.1W·m^-2,the simulation results of sensible heat flux are better and the error of observation results is smaller.The simulation results of the model for latent heat flux are ideal.The overall average value of latent heat flux is 26.2W·m^-2 in spring,40.9W·m^-2 in summer,and 30.6W·m^-2 in autumn 22.8W·m^-2,which is inconsistent with the observations at Tazhong Station,overestimating the actual latent heat flux in the Tazhong area.The annual error is basically around 20W·m^-2.?2?From the perspective of spatial distribution,the net radiation of water is the largest,the bare soil and desert area is the second,and the alpine ice sheet is the smallest,and the difference of net radiation flux in different types of underlying surfaces is closely related to the surface albedo.Due to different calculation methods,lakes are the areas with the highest soil heat flux,while areas with high vegetation coverage have large surface moisture,and the corresponding surface albedo has low soil heat flux,so the oasis and farmland areas on the edge of the desert are slightly higher than the soil heat flux values in the desert areas.For the sensible heat flux,the sensible heat flux values of the oasis and forest land distributed in the desert edge are slightly higher than those in the desert area,and the sufficient water vapor exchange conditions on the lake make the latent heat and heat sensitivity small.?3?Further analysis of the simulation results shows that sensible heat flux occupies the dominant position in the surface energy flux consumption in the Taklamakan desert area.In one year,the sensible heat flux accounted for 34.16%⁷2.35%of the net radiation,and the summer sensible heat flux accounted for more than 55%⁷0%.In the winter simulation results,the proportion of unclosed energy increases,and the soil heat flux accounts for about 20%of the net radiation throughout the year.?4?Quantitative analysis is carried out by comparing the results of the surface energy components simulated by the SEBS model with the actual observations of the Tazhong station through four evaluation indicators.The best fit is the net radiant flux and sensible heat flux,the coefficient R² of the model simulation is 0.99.The deviation of bias is the least,which is 1.25;the root mean square error of net radiation flux and sensible heat flux is slightly higher than that of soil heat flux;the average absolute error MAE results show that the results of other energy balance components are good except latent heat flux,and the root mean square error of net radiation flux and sensible heat flux is slightly higher than that of soil heat flux;NSE results show that the net radiation and sensible flux model simulation results are closer to the actual values.On the whole,using the SEBS model in the desert area,the net radiant flux and sensible heat flux result are the best,which is closer to the actual value;the simulation of the soil heat flux in the desert underlying surface is not much wrong,but the overall trend is poor;the simulation of latent heat flux is poor and needs to be further strengthened.?5?Using the local sensitivity analysis method to test the sensitivity of the seven input parameters of the SEBS model,the parameters that have a great influence on the net radiant flux results are:NDVI,surface albedo and surface temperature;there are four parameters with high contribution to the simulation results:air temperature,NDVI,surface albedo and surface temperature.Parameters affecting sensible heat flux simulation are air temperature,air pressure,wind speed,and surface temperature.