基于多源数据的土壤水分反演及空间分异格局研究

Spatial Variation and Inversion of Soil Moisture Based on Multi-Source Data

温度植被干旱指数(TVDI)是一种通过反演土壤水分来反映土壤干旱状况的重要方法。在TVDI的基础上引入数字高程模型(DEM)数据对地表温度进行校正,能够减少地形起伏对能量二次分配的影响。用阈值将研究区分割成不同土地利用类型,结合野外同步实测数据,用分段反演模型反演渭-库绿洲土壤水分分布图,并对渭-库绿洲土壤水分的空间格局和分异规律进行分析。结果表明:(1)地形校正后的TVDI能够更好地反映土壤水分状况;(2)土壤水分总体上从西至东,由北向南降低,在绿洲内部较为稳定,在交错带变化较为剧烈,土壤水分垂直变异系数呈现幂函数递减的整体变化趋势;(3)针对研究区不同土地利用类型用分段反演模型进行反演,有效地实现优势互补,总体上提高了土壤水分的反演精度,在区域土壤水分研究中值得关注和应用。

[Objective] Soil moisture is a key variable in parameterization of a land surface. Soil moisture content varies spatially and temporally, and plays an important role in material and energy exchanges occurring in the earth-atmosphere interface. It is a basic indicator for prediction of crop growth and crop yield, and also an important parameter in the numerical weather prediction model that helps efficiently forecast improvement in the physics of land surface processes on regional or global scales. However, it is very difficult to obtain an accurate soil moisture parameter just through field measurement temporally and spatially, especially at large spatial scales. Remote sensing technology has widely been used in estimation of soil moisture contents for it features capability of observing a large area synchronously, timely and economically. In the present study, a total of 39 soil samples were collected from an oasis in the Weigan- Kuqa Delta, Xinjiang of China. [Method] Temperature Vegetation Dryness Index （TVDI） is an important tool that can be used to reflect soil moisture regime through inversion. In this study DEM data were introduced to correct the data of land surface temperature （ Ts ） on the basis of the researches on TVDI in an attempt to mitigate the impact of undulation in land relief on secondary distribution of energy. The study area was divided into different land use types by threshold. Based on the synchronously field measured data, a piecewise inversion model was used to inverse soil moisture distribution in the delta oasis, and then spatial distribution pattern of soil moisture in the studied area and rules of its variation were further analyzed. [Result] Results show as follows： （ 1 ） Based on TVDI, DEM was introduced for terrain correction. Comparing the data before and after the terrain correction demonstrates that Modified Soil-Adjusted Vegetation Index （ MSAVI ） and the determination coefficient of the field measured soil moisture content in the topsoil layer were higher after the correction than before the correction. After the correction, R2 rose from 0.44 to 0.53, and all were above the 0.01 significance level; （ 2 ） The soil moisture regime of the delta oasis reflected in the classification maps of TVD1 demonstrates an apparent rule of spatial variation of soil moisture that on the whole, soil moisture declines from west to east and from north to south. Based on the variation coefficient of the mean soil moisture contents in the six soil layers of the 0 ~ 100 cm soil profiles different in land use type in the study area, soil moisture variation coefficient of the oasis and ecotone was 28% and 36%, respectively. From the distribution of the soil moisture variation coefficients of the soil layers in the soil moisture vertical variation coefficient map, it is discerned that the closer to the surface, the higher the variation coefficient, and that the vertical variation coefficient of soil moisture varied with land use type even in the same soil layers. However, the overall trend was quite similar, that is, it declines with soil depth; （3） By comparing the piecewise inversion model with the original model in inversion, it is learnt that the determination coefficient between the measured soil moisture and that predicted using the piecewise model and original model was 0.72 and 0.64, with RMSE being 0.03 and 0.04, respectively. Obviously, the coefficient of determination increased by 0.08 and RMSE decreased by 0.01. [Conclusion] Based on all the above-described findings, it could be concluded as follows： （ 1 ） In.this study, DEM elevation data were used for terrain correction of TVDI, and the corrected TVDI better reflected the sol west to east and from north to south, and ecotone. Soil moisture vertical variation c function; （3） By comparing the piecewi that the piecewise inversion model should may effectively supplement each other, th 1 moisture regime; （ 2 ） On the whole, soil moisture declines from remains quite consistent in the oasis, but varies drastically in the oefficient varied as a whole, showing a declining trend of a power se inversion model and the original model in inversion, it is learnt be used to invert soil moisture in soils different in land use, which us improving accuracy of the inversion as a whole. It is, therefore, worthy of attention and application in researches on regional soil moisture regime