Denoising Diffusion Probabilistic Models for Coastal Inundation Forecasting
Authors:
Kazi Ashik Islam,
Zakaria Mehrab,
Mahantesh Halappanavar,
Henning Mortveit,
Sridhar Katragadda,
Jon Derek Loftis,
Madhav Marathe
Abstract:
Coastal flooding poses significant risks to communities, necessitating fast and accurate forecasting methods to mitigate potential damage. To approach this problem, we present DIFF-FLOOD, a probabilistic spatiotemporal forecasting method designed based on denoising diffusion models. DIFF-FLOOD predicts inundation level at a location by taking both spatial and temporal context into account. It util…
▽ More
Coastal flooding poses significant risks to communities, necessitating fast and accurate forecasting methods to mitigate potential damage. To approach this problem, we present DIFF-FLOOD, a probabilistic spatiotemporal forecasting method designed based on denoising diffusion models. DIFF-FLOOD predicts inundation level at a location by taking both spatial and temporal context into account. It utilizes inundation levels at neighboring locations and digital elevation data as spatial context. Inundation history from a context time window, together with additional co-variates are used as temporal context. Convolutional neural networks and cross-attention mechanism are then employed to capture the spatiotemporal dynamics in the data. We trained and tested DIFF-FLOOD on coastal inundation data from the Eastern Shore of Virginia, a region highly impacted by coastal flooding. Our results show that, DIFF-FLOOD outperforms existing forecasting methods in terms of prediction performance (6% to 64% improvement in terms of two performance metrics) and scalability.
△ Less
Submitted 8 May, 2025;
originally announced May 2025.
Combination of Lidar Elevations, Bathymetric Data, and Urban Infrastructure in a Sub-Grid Model for Predicting Inundation in New York City during Hurricane Sandy
Authors:
Jon Derek Loftis,
Harry V. Wang,
Stuart E. Hamilton,
David R. Forrest
Abstract:
We present the geospatial methods in conjunction with results of a newly developed storm surge and sub-grid inundation model which was applied in New York City during Hurricane Sandy in 2012. Sub-grid modeling takes a novel approach for partial wetting and drying within grid cells, eschewing the conventional hydrodynamic modeling method by nesting a sub-grid containing high-resolution lidar topogr…
▽ More
We present the geospatial methods in conjunction with results of a newly developed storm surge and sub-grid inundation model which was applied in New York City during Hurricane Sandy in 2012. Sub-grid modeling takes a novel approach for partial wetting and drying within grid cells, eschewing the conventional hydrodynamic modeling method by nesting a sub-grid containing high-resolution lidar topography and fine scale bathymetry within each computational grid cell. In doing so, the sub-grid modeling method is heavily dependent on building and street configuration provided by the DEM. The results of spatial comparisons between the sub-grid model and FEMA's maximum inundation extents in New York City yielded an unparalleled absolute mean distance difference of 38m and an average of 75% areal spatial match. An in-depth error analysis reveals that the modeled extent contour is well correlated with the FEMA extent contour in most areas, except in several distinct areas where differences in special features cause significant de-correlations between the two contours. Examples of these errors were found to be primarily attributed to lack of building representation in the New Jersey region of the model grid, occluded highway underpasses artificially blocking fluid flow, and DEM source differences between the model and FEMA. Accurate representation of these urban infrastructural features is critical in terms of sub-grid modeling, because it uniquely affects the fluid flux through each grid cell side, which ultimately determines the water depth and extent of flooding via distribution of water volume within each grid cell. Incorporation of buildings and highway underpasses allow for the model to improve overall absolute mean distance error metrics from 38m to 32m and area comparisons from 75% spatial match to 80% with minimal additional effort.
△ Less
Submitted 2 December, 2014;
originally announced December 2014.
