| The accurate estimation of surface evapotranspiration is of great significance to the rational allocation of water resources,the construction of ecological environment and the development of agriculture and forestry in the study area.The eco-environment is fragile and water resources are very scarce in the windy sandy beach area of Maowusu desert.The study of evapotranspiration is of far-reaching significance to the improvement of the eco-environment in this area.Using Landsat 8 images of April,September,December 2016 and January,May and July 2017,based on SEBAL model,this paper explores the correlation between meteorological factors and surface evapotranspiration,studies the sensitivity of model parameters to surface evapotranspiration,introduces the characteristic space of surface temperature and vegetation index into the selection method of cold and hot pixels,and improves the parameters of SWCVR model to retrieve water.The optimum SEBAL model was constructed to estimate the regional evapotranspiration in the windy sandy beach area of Maowusu Desert.Finally,the land surface evapotranspiration calculated from the monitoring data of meteorological stations is used to validate the estimated results,so as to realize real-time and dynamic remote sensing monitoring of land surface evapotranspiration,and provide a basis for the research of agriculture and ecological environment in windy sand beach areas.The conclusions are as follows:(1)SPSS was used to analyze the correlation between meteorological factors and surface evapotranspiration.The results show that surface temperature and air temperature are significantly correlated with daily evapotranspiration,and the correlation coefficients R~2 are0.6344 and 0.6274,respectively.The correlation coefficient R~2 is 0.3307 and the correlation coefficient R~2 is 0.3374,while the correlation coefficient R~2 is 0.3374,which is negatively correlated with relative humidity.Through the magnitude of R~2,we can find that the influence of surface temperature>air temperature>relative humidity>wind speed on daily evapotranspiration.(2)Inversion of surface evapotranspiration based on SEBAL model.The sensitivity of water vapor content and hot spot surface temperature to daily evapotranspiration was analyzed by single variable analysis method.It was found that water vapor content was more sensitive to daily evapotranspiration.The accuracy requirement of 1 mm/day needs to guarantee the accuracy support of 0.5 g/cm~2 water vapor content,while the surface temperature of hot spots is affected by iteration calculation and is less sensitive to daily evapotranspiration.(3)Inversion of atmospheric water vapor content using Landsat8 thermal infrared channel data is studied.By modifying the parameters of existing inversion models,an inversion model of water vapor content in arid areas based on SWCVR algorithm is constructed.The root mean square error of the inversion results is 0.31 g/cm~2,which proves that the SWCVR algorithm has good inversion accuracy when applied to the retrieval of water vapor content in arid areas from medium-resolution thermal infrared remote sensing data.(4)Using Landsat 8 remote sensing image data,surface evapotranspiration was retrieved based on the optimized SEBAL model,and good retrieval accuracy was achieved.The average evapotranspiration of January 2,April 21,May 26,July 13,September 28 and December 17 were 0.56 mm/day,2.02 mm/day,3.52 mm/day,5.17 mm/day,3.31 mm/day and 0.72 mm/day,respectively,showing summer>autumn>spring>winter as a whole.In terms of spatial distribution,surface evapotranspiration is larger in the southeast and smaller in the northwest.The inversion results based on SEBAL model and optimized SEBAL model are validated and evaluated.The results show that the root mean square error of SEBAL model is 1.30 mm/day and the root mean square error of optimized SEBAL model is 0.90mm/day.It shows that the inversion results of optimized SEBAL model have better accuracy than SEBAL model. |
PDF Full Text Request