基于蒸散发供需变化的干旱发生过程识别及其空间演变规律研究

Drought is one of the natural hazards that have the greatest impact on economic and social development, and its occurrence process is affected by precipitation and soil water content. Since the existing drought indices are unable to describe the difference of soil moisture caused by the variation of land cover, there are large uncertainties for the identified drought process. Meanwhile, the spatio-temporal dynamics of drought are often neglected in previous studies, and research on the spatial evolution of drought is lacked. Hence, it is urgently needed to further study the mechanism of drought process, improve the accuracy of drought identification, and explore the spatial evolution of drought. The project intends to estimate the actual and potential evapotranspiration based on satellite remote sensing data and meteorological data, and construct a new drought index based on the changes of actual and potential evapotranspiration since they have considered the impacts of various factors on drought such as the variation of land cover. Considering the spatio-temporal dynamics of drought, the project intends to quantify the thresholds for determining the correlation of drought events at adjacent times. With the central point of the most serious drought at the same time is defined as the drought-hit area, the probabilities for each direction of the diversion of drought-hit area are analyzed, and the spatial variations of the probability are also clarified. This will help to establish a probability prediction method for the spatial evolution trajectory of the drought-hit areas, and a new approach to early warning and prevention of drought using the difference occurrence time of drought in different areas is therefore proposed. This project will help identifying the drought process more accurately, enhancing drought early warning capability and reducing drought losses.

干旱是对社会经济影响最大的自然灾害之一，发生过程受降水及土壤含水量等因素影响，现有干旱指标因对土壤含水量受下垫面影响刻画不足，识别的干旱发生过程不确定性较大；同时对干旱时空关联性考虑不足，缺乏对干旱空间演变过程的认知，亟需进一步开展干旱发生机理研究，提高干旱识别精度，探究干旱空间演变规律。本项目拟基于卫星遥感及气象数据，反演实际及潜在蒸散发时空变化过程，利用蒸散发供需变化综合考虑下垫面差异等因素对干旱的影响，构建干旱指标，准确识别干旱发生过程。针对干旱时空关联性，界定相邻时刻干旱事件关联判断阈值。以干旱最严重区域的中心点作为干旱重灾区，分析干旱重灾区转移特征，探明其在各方向发生转移的概率及空间变化特征，建立干旱重灾区空间演变轨迹概率预报方法，利用不同地区干旱发生时间差进行干旱预警及预防。本项目研究将为更准确识别干旱发生过程，增强干旱预警能力，减少旱灾损失提供新的理论依据与技术支撑。

干旱是对社会经济影响最大的自然灾害之一，然而发生过程受土壤含水量、植被响应差异等多种因素影响，导致识别的干旱发生过程不确定性较大。一般来说，干旱的识别常借助于对降水、潜在及实际蒸散发变化情况的分析来进行判别，从而潜在及实际蒸散发计算结果的准确性将显著影响干旱识别的可靠性。本项目针对潜在及实际蒸散发计算方法进行了评价和改进，提出了一个不需要进行参数率定的蒸发互补理论算法，得到了一套全球实际蒸散发数据集；同时改进了蒸散发模拟中所必须的太阳辐射数据的计算方法，在中国区域的准确性比由联合国粮农组织推荐的常用公式提高了约33%。这些成果将显著提高干旱识别的准确性。另外针对干旱的时空关联性，基于Python语言开发了一套干旱时空特征识别算法，实现快速有效识别时空关联的干旱事件，界定干旱转移路径，从三维时空角度探究干旱变化特征，为准确量化干旱风险提供参考依据。本项目完成SCI学术论文5篇，参加国内外学术会议3人次。