报告题目：基于卫星遥感和混合模型的高分辨率蒸散发研究

报告时间：10月24日10:00-12:00

报告地点：A901

主持人：张永强 研究员

报告人简介

Diego Miralles 教授，2017年以来担任比利时根特大学水文与气候学教授，曾任职于英国、荷兰和美国。他的研究方向包括生态水文、陆-气相互作用、全球水文学、水文气候极值和遥感水文。他目前领导水文与气候团队，致力于理解水圈、生物圈和气候之间的相互作用及其对当前和未来社会的影响。他是欧洲研究委员会（ERC）Starting and Consolidator基金获得者，Clarivate全球高被引科学家。

摘要：

Terrestrial evaporation (E) is a crucial climate variable linking the water,carbon,and energy cycles. It regulates precipitation and temperature,influences feedbacks involving water vapor and clouds,and drives extreme events such as droughts, floods, and heatwaves. For water management, E represents a net loss of water, while in agriculture, transpiration determines irrigation demands. Despite its importance, global E estimates remain uncertain due to the scarcity of field observations, the complex coupling between physiological and atmospheric processes, and the challenges of retrieving E from satellite observations. These limitations have stimulated progress in modeling approaches that combine satellite data, in situ measurements,and advanced algorithms.

This presentation will review the evolution of evaporation science,with particular focus on its satellite remote sensing. The main existing approaches—ranging from thermal remote sensing to process-based and machine learning models—will be contextualized. Moreover,ongoing developments and future plans for the Global Land Evaporation Amsterdam Model (GLEAM),one of the most widely used global E datasets over the past 15 years,will be discussed,with emphasis on its applications in climate research and high-resolution studies.

The fourth generation of GLEAM (GLEAM4) estimates E and its components globally through a hybrid framework. It includes improved representations of interception, atmospheric water demand,soil moisture dynamics,and plant access to groundwater. GLEAM4 integrates machine learning techniques to capture evaporative stress, leveraging eddy-covariance and sapflow observations,while preserving water balance and thermodynamic consistency. By combining the interpretability of physics-based models with the adaptability of machine learning,it provides a scalable and physically consistent estimate of E across ecosystems.

Building on GLEAM4,a new generation of high-resolution evaporation datasets is under development to meet the growing demand for actionable information in agriculture,water management,and climate adaptation. These datasets will enable improved characterization of droughts,heatwaves,and water resource variability, particularly in  climate-sensitive regions,offering a valuable tool for managing water resources,mitigating climate impacts,and supporting a wide range of applications in hydrology,ecology,and climate science.