| At present,there are a variety of remote sensing monitoring soil moisture methods,mainly focused on optical remote sensing and microwave remote sensing.Considering all kinds of remote sensing monitoring soil moisture methods,this paper chooses TVDI(Temperature Vegetation Drought Index)to construct the inversion model of soil moisture.Based on the Landsat series satellite imagery data in 2007 and 2013,the vegetation coverage of the Yanhe River Basin was extracted according to the dimidiate pixel model and the surface temperature was extracted according to radiation conduction model.Using TVDI extracted soil moisture information of Yanhe River in summer and analyzed the spatial distribution of soil moisture.According to the analysis of spatial distribution of soil moisture in two phases,it is found that TVDI can well reflect the wet and dry condition of soil.Construction of the deep TVDI regression model based on the surface TVDI and return point in 2013,the study respectively extracted soil moisture content of 0-20 cm,20-40 cm,40-60 cm,60-80 cm and 80-100 cm according to soil water inversion model and obtain the spatial distribution of each soil layer.Using some measured soil moisture data as the verification point,the study verify the accuracy of the retrieval values of the soil water content.The main conclusions are as follows:(1)Compared with 2007,the percentage of low vegetation coverage(0-0.2),lower vegetation coverage(0.2-0.4)and high vegetation coverage(0.8-1)in Yanhe River Basin increased in 2013.The ratio of low vegetation coverage,lower vegetation coverage increased by 2% and 3.1%,but the ratio of high vegetation coverage increased by 18.35%.The ratio of moderate vegetation coverage(0.4-0.6)and slightly higher vegetation coverage(0.6-0.8)decreased by 5.63% and 13.17%.(2)From the results of two periods of temperature vegetation drought index inversion,the Yanhe River Basin showed a drought trend.Micro drought areas which mostly distributed in the northwest region widely distributed in the Yanhe River Basin and the numerical range of TVDI is mainly concentrated at 0.6 to 0.8.Showing edge distribution,Humid areas which mostly distributed in the southern region small distributed in the Yanhe River Basin and the numerical range of TVDI is mainly concentrated at 0 to 0.5.The study shows that soil moisture was a big difference under different ground cover types.The overall performance trend of TVDI which showed different distribution patterns under each ground cover types was: bare land>urban area>vegetation area.(3)Compared with 2007,the percentage of humid areas in Yanhe River in 2013 decreased by 1.096%,and the percentage of normal areas increased by 7.207%.The percentage of slightly drought areas in Yanhe River in 2013 increased by 2.421% and the percentage of drought and severe drought areasareas respectively decreased by 2.555% and 5.707%.In 2013,humid areas moved gradually to the north.The studys shows that after several years of vegetation restoration and the implementation of the project of returning farmland to forest and grass,the surface soil moisture condition has improved and the area of soil arid area has narrowed,and the drought in Yanhe River has been alleviated.But the humid areas of the southeastern region are gradually replaced by slightly drought area.The scope of the wet area was narrowed,but the scope of the drought area was expanded.(4)The inversion value of 0-20 cm soil water content is not high and the average inversion accuracy is 70.2%.Because of that the surface soil water content is susceptible to rainfall,solar radiation,wind speed,evaporation and other factors,the correlation between the inversion value of soil water content and the measured value is small.The inversion accuracy of 20-40 cm soil moisture content is highest and the average inversion accuracy is 80.6%.The soil environment is relatively stable at 20-40 cm,and the optical remote sensing has some penetrating power.The inversion of soil water content in 40-60 cm was highly correlated with the measured value,and the inversion precision is relatively high which the average inversion accuracy reached 72.6%.The accuracy of soil water content in 60-80 cm and 80-100 cm was low,and the average inversion accuracy respectively was 67.3% and 55.4%,which shows that the remote sensing inversion model of soil moisture can not accurately extract the soil moisture at the depth of the soil.The study shows that the method of optical remote sensing inversion can be used to extract the information of shallow soil moisture,but can not meet the requirements of the extraction of soil moisture in deeper soil... |
PDF Full Text Request