军事知识服务平台
气候变化影响下年最大洪峰变化的统计归因分析
Statistical Attribution Analysis of Annual Flood Magnitude Changes Under the Influence of Climate Change
摘要 年最大洪峰变化的归因分析,是提升水文极端事件应对韧性的关键;然而,目前鲜有研究建立气候变化与历史观测的年最大洪峰变化之间的因果关联.本研究基于流域属性水文气象数据集(CAM‑ELS)涵盖的671个流域观测及全球洪水预警系统(GloFAS)径流再分析数据,量化分析年最大洪峰与气象因子之间的响应关系.首先对比分析两种统计建模方法,其次通过面板回归模型分离出各气象因子的独立影响并解析气候因子的调控机制,最后量化历史气候演变趋势下气象因子对年最大洪峰变化的贡献度.结果表明:(1)平均降水量与极端降水量每提升1%,年最大洪峰对应的径流深分别增加1.97%和0.35%.(2)ENSO不同位相下洪峰驱动机制存在差异,拉尼娜年份年最大洪峰对应的径流深对极端降水和极端融雪量的回归响应系数分别降低约0.02和0.04,厄尔尼诺年份对极端融雪量的回归响应系数提高约0.03.(3)在历史洪峰变化的归因中,平均降水量、极端降水量和极端融雪量的贡献率分别为52.28%、37.70%和 28.49%.(4)对比分析表明GloFAS 再分析数据高估了平均降水量变化对洪峰的影响,同时低估了极端降水量和融雪量变化的贡献.本研究揭示了年最大洪峰对气象因子的响应规律,对于洪水预报预警具有重要参考价值.
Abstract:
Attribution analysis of the changes in annual peak floods is crucial for enhancing resilience against hydrological ex‑treme events.However,few studies have established a clear causal link between climate change and observed variations in annual peak floods.In this study,observations from 671 catchments in the Catchment Attributes and Meteorology for Large-sample Stud‑ies(CAMELS)dataset and streamflow reanalysis data from the Global Flood Awareness System(GloFAS)were used to quantify the relationships between annual peak floods and meteorological drivers.First,two statistical modeling methods were employed for com‑parative analysis.Then,a panel regression model was used to isolate the independent effects of each meteorological factor and ex‑plore the regulatory mechanisms of climatic factors.Finally,the contribution of meteorological factors to the changes in annual peak floods under historical climate trends was quantified.The results show that a 1%increase in average precipitation and ex‑treme precipitation leads to increases of 1.97%and 0.35%in runoff depth for annual peak flood magnitude,respectively.Distinct flood-driving mechanisms were identified under different ENSO phases.Compared with all years,the sensitivity of runoff depth for annual peak discharge to extreme precipitation and extreme snowmelt decreased by approximately 0.02 and 0.04,respectively,dur‑ing La Niña years,whereas the sensitivity to extreme snowmelt increased by about 0.03 during El Niño years.Attribution analysis indicates that the contributions of changes in mean precipitation,extreme precipitation,and extreme snowmelt to historical varia‑tions in annual peak floods are 52.28%,37.70%,and 28.49%,respectively.Compared with observations,the GloFAS reanalysis data overestimate the influence of changes in mean precipitation on annual peak floods,while underestimating the contributions of changes in extreme precipitation and snowmelt.This study reveals the response patterns of annual maximum flood peaks to meteo‑rological drivers,providing valuable insights for flood forecasting and early warning.
刊 名 水文 ISTICPKU
年,卷(期): 2026, 46(5)
分类号: P339 P343.1
机标分类号: P467 S162.5 P339
在线出版日期: 2026年11月9日
基金项目: 52379033:国家自然科学基金,