-
- H24A-06: Global-to-Regional Climate Signal Extraction for Climate-Conditioned Projection of Hydrologic Extremes
-
NOLA CC
Author(s):Generic 'disconnected' Message
Julian Esler, Columbia University (First Author, Presenting Author)
Adam Nayak, Columbia University
Upmanu Lall, Columbia University
Understanding how flood risks may change in the future is crucial for protecting communities and planning for climate impacts. Traditional methods of flood risk analysis often assume that past flood patterns will continue in the future, failing to account for the influence of internal climate variability and changing climate conditions. Physics-based methods, such as using general circulation models (GCMs) are sensitive to climate conditions and variability, but have persistent errors and are inaccurate at fine spatial or temporal scales. Accordingly, decision-makers lack the tools required to plan future flood-protection measures. To address this need, we present a tool for projecting regionally-specific future flood risks. The tool has two primary components: (1) a machine learning model that extracts the modes of global climate variability that modulate regional precipitation and (2) a predictive model that projects this signal based on coarse-resolution information about future climate conditions. This projected signal is then integrated into a flood simulator to tailor simulations to the regional response to future climate conditions. This approach leverages the spatial specificity of traditional methods of flood risk analysis and the sensitivity of physics-based models to future climate conditions, to provide useful, regionally specific insights for managing future flood risk.
Scientific DisciplineNeighborhoodType
Enter Note
Go to previous page in this tab
Session
