{"title":"Australia-wide projections of extreme rainfall and flooding","authors":"C. Wasko, D. Guo, M. Ho, R. Nathan, E. Vogel","doi":"10.36334/modsim.2023.wasko","DOIUrl":null,"url":null,"abstract":": Engineering design, floodplain management, and water resources planning all require estimates of extreme rainfall and flooding. However, as we plan and design for the future, the historical records we have used in the past are no longer representative of the future due to climate change. Our climate system is experiencing many changes: rising temperatures are increasing the saturation vapor pressure increasing extreme rainfalls; changes in circulation patterns are shifting the frequency of rainfall events; and changes in the mean annual rainfall and time between rainfall events are impacting on the soil moisture conditions before a rainfall event. Hence, if we are to correctly specify the level of risk in future design and planning and decisions, all these changes need to be accounted for in our estimates of extreme rainfall and flooding. Here, we project extreme rainfall and flooding (in the form of frequency curves) across Australia’s diverse climate and, in doing so, develop a simple, robust methodology that can be readily used for flood projections. We first calibrate the rainfall-runoff model GR4J across 467 Hydrologic Reference Stations using observed rainfall, potential evapotranspiration (PET), and streamflow. The calibration uses a novel objective function which aims to match flood quantiles. The hydrological models across all catchments are then evaluated in terms of flood frequency, Nash-Sutcliffe Efficiency (NSE), and the trend in annual maxima, to ensure that the processes causing changes in flood frequency are captured. For use in future projections, rainfall and PET climate model data from four GCMs and four different bias-correction methods are obtained from the Australian Bureau of Meteorology (","PeriodicalId":390064,"journal":{"name":"MODSIM2023, 25th International Congress on Modelling and Simulation.","volume":"404 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"MODSIM2023, 25th International Congress on Modelling and Simulation.","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.36334/modsim.2023.wasko","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
: Engineering design, floodplain management, and water resources planning all require estimates of extreme rainfall and flooding. However, as we plan and design for the future, the historical records we have used in the past are no longer representative of the future due to climate change. Our climate system is experiencing many changes: rising temperatures are increasing the saturation vapor pressure increasing extreme rainfalls; changes in circulation patterns are shifting the frequency of rainfall events; and changes in the mean annual rainfall and time between rainfall events are impacting on the soil moisture conditions before a rainfall event. Hence, if we are to correctly specify the level of risk in future design and planning and decisions, all these changes need to be accounted for in our estimates of extreme rainfall and flooding. Here, we project extreme rainfall and flooding (in the form of frequency curves) across Australia’s diverse climate and, in doing so, develop a simple, robust methodology that can be readily used for flood projections. We first calibrate the rainfall-runoff model GR4J across 467 Hydrologic Reference Stations using observed rainfall, potential evapotranspiration (PET), and streamflow. The calibration uses a novel objective function which aims to match flood quantiles. The hydrological models across all catchments are then evaluated in terms of flood frequency, Nash-Sutcliffe Efficiency (NSE), and the trend in annual maxima, to ensure that the processes causing changes in flood frequency are captured. For use in future projections, rainfall and PET climate model data from four GCMs and four different bias-correction methods are obtained from the Australian Bureau of Meteorology (