考虑土壤水分和蒸发比时空异质性的蒸散发遥感估算方法改进

Evapotranspiration (ET), which governs the water cycle and energy transport among the biosphere, atmosphere and hydrosphere as a controlling factor, plays an important role in hydrology, meteorology, and agriculture, such as in prediction and estimation of regional-scale surface runoff and underground water, in simulation of large-scale atmospheric circulation and global climate change, in the scheduling of field-scale field irrigations and tillage. There are many problems in the existing remotely sensed estimation methods, because of the limitations of observation. These problems were mainly brought about by neglecting the temporal and spatially heterogeneity of soil moisture and evaporative faction (EF) in the remotely sensed estimation. This project will rely on the equipments which have been installed during WATER and will be installed during Hi-WATER. Many intensive observations of hydrological and meteorological elements will be obtained, such as latent heat flux, sensible heat flux, soil moisture and temperature, wind speed, air temperature, and so on. This study aims at improving the problems brought about by neglecting the temporal and spatially heterogeneity of soil moisture and EF and providing estimations with high precision. A parameterization scheme for soil evaporation will be proposed based on the spatial distribution of soil moisture collected by the wireless sensor network and the soil evaporation divided from ET with the help of stable isotope observation system. The spatial and temporal variation variability of EF will be analysised based on SiB2 and Bowen ratio energy balance observation system. A revised temporal scaling method will then be proposed refer to the heterogeneous EF. Finnally, a reliable ET estimate would be supplied to the researches on eco-hydrological process at watershed scale.

蒸散发（ET）是地球系统能量和物质循环的核心，清楚的认识蒸散发，对了解大范围内能量平衡和水分循环具有重要意义。由于观测手段有限，目前的遥感估算方法还存在着众多问题还没有被合理解决。本项目将依托2012年开展的HiWATER试验所架设的仪器，对实验场内水文气象要素进行加密观测，旨在改进目前蒸散发遥感估算中存在的忽略土壤水分和蒸发比（EF）的时空异质性所带来的局限，并对估算结果进行合理验证。根据无线电传感器网络获得的土壤水分时空分布信息,结合稳定同位素观测系统分离土壤蒸发和植被蒸腾量,建立土壤蒸发参数化方案。借助陆面过程模型SiB2的模拟和波文比能量平衡系统的观测，获得EF的时空变化特征，改进时间尺度扩展方法。最终获得的高精度的日蒸散发量，将为流域的生态水文过程的研究提供更为可靠的数据支持。

本项目研究区位于黑河盆地，中国第二大内陆河流域。研究所用的遥感数据和野外气象、水文以及水热通量观测数据均来自黑河流域生态水文过程综合遥感观测联合试验（HiWATER）项目。研究区内的加密观测矩阵架设了17台涡动相关观测系统(EC)和自动气象站(AMS)为本项目提供了充足的数据支持，保证了本项目的顺利完成。.本项目研究者改进了一个基于彭曼公式（Penman-Monteith, P-M equation）的双层蒸散发模型。设计了一个基于贝叶斯反转的分离式的参数估计策略，通过分别使用农田生长季与非生长季的观测数据，达到分离作物蒸发与土壤蒸腾，达到对植被蒸腾子模型中的参数和土壤蒸发子模型的参数分别估计。然后研究者采用研究区内17个涡动和自动气象站所在样地的土壤水分，冠层高度，叶面积指数3组数据，利用聚类分析，找到相似程度较高的地块，将通过贝叶斯反转获得的7个样地上的6组参数值插补到其他整个研究区上，从而估算整个研究区内潜热通量。用研究区内17个涡动相关获取的数据作验证，结果表明，遥感估算的均方根误差（RMSE）小于20 W m-2。为何获取区域尺度更高精度的日蒸散发产品，研究者提出了一个基于蒸发比（EF）检测算法的改进的时间尺度扩展方法。利用一个监测算法判定EF的稳定性，该稳定性检测算法的流程图见图7。EFEC不稳定的时段和保持稳定的时段被区分开来，并分别积分。与固定蒸发比法和变化蒸发比法相比，改进方法的误差更小，RMSE小于0.6 mmd-1，估计值与观测值的线性关系更好，相关系数（Corr.）大于0.8。该方法使用简单，可以在区域尺度上获得更高精度的蒸散发产品。