An NCCOS-supported research team developed a new model to rapidly simulate flood patterns in Miami, Florida, and better manage storm hazards. The model introduces a novel approach to represent the infiltration of rainwater into the ground, improving flood predictions by tracking changes in the height of the groundwater table during storms.
Across Miami, the groundwater table fluctuates seasonally and with individual rainfall events, and the amount of flooding caused by rainfall is strongly dependent on groundwater height at the beginning of the storm. This model now makes it possible to accurately and rapidly predict spatial flood patterns on a storm-by-storm basis, which better assess risks across the region. This work also offers an opportunity for local managers to take a proactive approach, protecting lives and property from harm by anticipating areas most likely to flood, and planning more effective infrastructure investments and other risk reduction measures.
The research team demonstrated the model across a densely populated area roughly 1,000 square kilometers in size — an area larger than New York City — within Miami-Dade County, Florida, where flooding is managed by the South Florida Water Management District. To carry out this project, the team used a combination of data, including rainfall, water level gauge data, topography, land use, flood infrastructure, and groundwater levels. The researchers selected a four-day period during May of 2020, an episode described as the Memorial Day flood, as the basis of modeling and analysis.
The team showed that water levels in canal networks with gates can be predicted with an average accuracy of about 15 centimeters, and that the model speed is at least 30 times faster than real time. This level of accuracy is comparable to the best results previously reported by other researchers, while the model speed is many times faster. The model was configured to run on a graphical processing unit to achieve its fast speeds.
The team also presented additional model testing across the Houston, Texas, metropolitan region, where they demonstrated a range of model speeds greater than 100. This test showed that adding groundwater table height and infiltration information to the model reduced model speed by less than five percent.
Researchers increasingly recognize groundwater table height as a major consideration with respect to flood risk in coastal areas. This work underscored the results of previous studies documenting compound risks from rainfall events coinciding with high groundwater table heights. It also highlighted the limitations of infiltration as a flood reduction strategy, where today it is heavily relied upon in South Florida to quickly drain water that accumulates on the ground during rainfall events.
This work was led by a team of researchers from the University of California, Irvine and the University of Miami, and funded through the NCCOS Effects of Sea Level Rise (ESLR) Program. Learn more about their work here.
Citation: Sanders, B. F., Schubert, J. E., Martin, E. H., Wang, S., Sukop, M. C., & Mach, K. J. (2025). A fast flood inundation model with groundwater interactions and hydraulic structures. Advances in Water Resources, 204, 105057. https://doi.org/10.1016/j.advwatres.2025.105057
This work is authorized by the NOAA Authorization Act of 1992, Pub. L. 102-567 (Oct. 29, 1992); sec. 201(c), which authorizes appropriation for the Competitive Research Program to augment and integrate existing programs of NOAA and shall include efforts to improve predictions of coastal hazards to protect human life and personal property.