关键词: 遥感, 地表蒸散, 空间尺度转换, SEBS, MODIS, 伊洛河流域

Abstract:

As a result of the existence of spatial heterogeneity and mixed pixels, there are salient spatial scaling effects in the evapotranspiration (ET) estimated from low resolution remote sensing data. This paper proposes a spatial scaling transformation method based on GARWLM (Grid Area Ratio Weighting Lumped Method). The new method gives the corresponding weights according to the area ratio of various land cover types and makes the scaling model from the fine resolution (HJ-1B) to the coarse resolution (MODIS) in a grid (1km); spatial scaling transformation from ET estimated from HJ-1B data to ET estimated from MODIS images is performed by the above steps. Finally, the correction of scale effect to ET estimated from coarse resolution data or MODIS is accomplished by regression. The studies in Yiluo River basin show that the average relative error between the estimated ET and ground measured data decreases from 23.5% to 12%. The correlation of the fine resolution increases from 0.339 to 0.604. The accuracy of ET corrected for scale effect has been greatly improved. The method provides a good idea to solve the scale problem in quantitative analysis and application of multi-source remote sensing data with large scale, and can be extended to remote sensing estimation of other parameters.

Key words: remote sensing, evapotranspiration, spatial scaling transformation, SEBS, MODIS, Yiluo River Basin