Inversion Of Evapotranspiration On Urban Land Surface Based On Remote Sensing Data

In the context of global climate change, because of drastic changes in urban land cover and enhancement of anthropogenic heat release, the urban heat environment deteriorated sharply. The urban heat island effect is becoming one of major urban environmental problems and has been widely concerned. Some subjects such as geography, ecology and environmental science, carried out extensive and in-depth research, the study includes the spatial characteristics and formation mechanism of the urban heat island. However, these studies rarely related to the evapotranspiration. Evapotranspiration is one of the important parameters which are used urban surface water cycle and energy cycle research, and is one of the important parameters in urban thermal environment study. It is of great scientific significance and practical value to study the evapotranspiration simulation and dynamic monitoring of the urban land surface and it contribute to a deeper understanding of the environmental problems of urban heat formation mechanism.In this paper, the core urban area of Changsha City was selected as study area. Firstly, a new algorithm called multi-source parallel model based on the urban surface energy balance characteristics has been proposed. Secondly, some parameterization methods for the model have been studied, and the inversion result of this algorithm is validated by comparison with the data of meteorological stations and estimation results of P-M model and SEBAL algorithm. Finally, the influence of land covers spatial structure and change on evapotranspiration of urban land surface, and the influence of evapotranspiration changes on the urban thermal environment have been analyzed.The results show that:(1) In support of multi-source remote sensing data, the inversion result of multi-source parallel model is satisfactory. Compared with the weather station measured data, the maximum bias is1.0mm, and the minimum bias is0.1mm. Compared with result of P-M model, different land cover types are present to a certain degree of underestimation, and the relative deviation does not exceed20%, compared with result of SEBAL algorithm, two results share good relations and the relation in vegetation area get maximum value.(2) Land cover changes resulted in a significant change in evapotranspiration. Conversions of forest land to building land from1993to2005resulted in significant reductions in evapotranspiration, and conversions of forest land to farmland in the same period resulted reductions in evapotranspiration, but the reduction degrade is mainly slight. Conversions of farmland to building land are widely distributed, and resulted in significant reductions in evapotranspiration. Conversions of farmland to forest land resulted in significant evapotranspiration increase.(3) Correlation between the intensity of urban heat island and surface evapotranspiration shows obvious negative correlation. From1993to2005, the area of urban region is significantly expanded, and the spatial distribution characteristics of evapotranspiration changed significantly, and resulted in the urban heat island effect exacerbated. The evapotranspiration in the region where urban heat island effect is not obvious is relatively high, and the evapotranspiration in the region where urban heat island effect is strong is relatively low, a significant negative correlation between the intensity of urban heat island and evapotranspiration can be found. From the classification map of heat island intensity in1993, it can be found that heat island effect mainly occurred in the central urban area, and the characteristic of spatial aggregation is obvious. The region of high heat island intensity expanded significantly from1993to2005, in the conditions of significant reduction of evapotranspiration, the heat island intensity obvious enhanced, but area slightly enhanced accounted for the main part.