| Land surface temperature(LST)is an important indicator of climate change.Accurate estimation of the spatial and temporal distribution of LST is essential for enhancing the understanding of surface hydrological processes and predicting future drought conditions.The accuracy of the general land surface model simulation results is closely related to the quality of the atmospheric forcing data set,and the error of the forcing data set itself will be transferred to the simulation results.This study used three atmospheric forcing data sets,Climatic Research Unit-National Centers for Environmental Prediction(CRUNCEP),Global Soil Wetness Project(GSWP),and China Meteorological Forcing Dataset(CMFD),the National Meteorological Science Data Center's daily surface data set(V3.0),Moderate-resolution Imaging Spectroradiometer(MODIS)LST data sets,and the general land surface model Community Land Model 5.0(CLM5.0)as a research tool,the influence of the uncertainty of atmospheric forcing data sets on the simulated land surface temperature of the model was discussed.The characteristics of the spatial and temporal distribution of land surface temperature in the upper and middle reaches of the Yellow River Basin and the characteristics of land surface temperature simulation driven by different atmospheric forcing data sets were analyzed.Research shows:(1)From 2003 to 2010,there is a certain gap between the three sets of atmospheric forcing data.Regarding the temporal change of air temperature,the three sets of atmospheric forcing data have a more consistent trend,with smaller fluctuations and gaps,showing the highest GSWP data,followed by CRUNCEP,and the lowest CMFD.In terms of the temporal change of incident solar radiation,the three sets of atmospheric forcing data have a certain difference in the interannual change trend.The CRUNCEP data is also significantly different from the other two sets of data,which is the highest of the three sets.In autumn and winter,CMFD data is the lowest;in spring and summer,CRUNCEP data is the highest.Spatially,the three sets of data show the characteristics of higher temperatures in the southeast and lower temperatures in the northeast and west in all seasons.The spatial distribution of CRUNCEP and GSWP data is similar,the spatial distribution of air temperature of CMFD data is more detailed,and the difference between the two sets of data is larger.Incident solar radiation is characterized by lower southeast and higher Tibetan Plateau in the west.The CRUNCEP data is higher,the fluctuation is smaller,and the spatial distribution of CMFD data is larger.Uncertainty analysis found that the data in the eastern part of the study area was significantly better than the data in the west part of the study area.The uncertainty ofCMFD data is the smallest,GSWP data is the second,and CRUNCEP data is the largest.Therefore,the temperature and incident solar radiation data provided by CMFD data are the best in three data sets.(2)From 2003 to 2010,the changes trend of land surface temperature driven by three sets of atmospheric forcing data in the upper and middle reaches of the Yellow River Basin is more consistent,showing the highest CLM-GSWP data,followed by CLM-CRUNCEP,CLM-CMFD lowest situation.In terms of space,the three experimental schemes can better simulate the spatial distribution of LST that is closer to MODIS data.They all capture the spatial distribution characteristics of low LST in the west of the study area and high LST in the Guanzhong Plain.The characteristics of LST change with the change of latitude,season,topography,and land use type.In terms of uncertainty,the results of the three options show that the uncertainty in the east is significantly lower than that in the west.Except that the winter CLM-CMFD results are not the best,the simulation results of the CLM-CMFD schemes in the other three seasons are the best.After a comprehensive analysis,it is concluded that the CLM-CMFD scheme has the best LST simulation results,but the accuracy of its driving simulation of winter LST needs to be further improved.(3)The rank in time scale,spatial distribution characteristics,and the characteristics of the east and west uncertainties in three LST simulation results and the atmospheric forcing data sets are consistent.The overestimation of the simulation results driven by CMFD data in the Guanzhong Plain is mainly due to the uncertainty of air temperature in the atmospheric forcing data set.In the spring and autumn seasons,the LST simulation results of the three schemes all show a higher correlation with the two variables,while in the winter and summer,the correlation is lower;compared with the incident solar radiation,LST has a more significant correlation with the temperature,It shows that when CLM simulates LST,it is more affected by the air temperature in the atmospheric forcing data set.In general,the uncertainty of air temperature and incident solar radiation in the atmospheric forcing data sets will be transferred to the simulation results of the land surface temperature through the model.In the future,the accuracy of the atmospheric forcing data sets needs to be continuously improved to improve the accuracy of land surface temperature simulation in the upper middle reaches of the Yellow River Basin. |
PDF Full Text Request