Study Makes Storm Surge Predictions at Less Cost without Sacrificing Accuracy

60 Flares 60 Flares ×
Hurricane Ike storm surge 2008. (Photo credit:  NOAA)

(Click to enlarge) Hurricane Ike storm surge 2008. (Photo credit: NOAA)

Two scientists from the University of Texas provide an alternate modeling framework that incorporates selective detailed adjustments while calculations are in process to predict storm surge.

The addition of adaptive mesh refinement (AMR) algorithms to existing models addresses the need for quick and accurate information without being cost prohibitive. The researchers published their findings in the March 2014 edition of Ocean Modelling: Adaptive mesh refinement for storm surge.

Catastrophic weather incidents are increasing as is population growth along vulnerable coastlines, making improved forecasting a key initiative in the scientific community. Storm surge is a sudden and fast rise in sea level that accompanies tropical storms such as hurricanes and can cause devastating damage to both property and human life. Accurate predictions of the landward reach of storm surge is especially important for the Gulf of Mexico which has experienced extensive weather-related disasters. Since the Gulf is one of the major petroleum-producing areas of the United States, improved storm surge predictions can assist responders with hazards caused by broken pipelines and failures in storage structures along coastlines and the subsequent spreading of oil further inland.

Existing prediction models can accurately detail a coastal region but require very powerful and expensive computing resources. The most common surge forecasting systems in use today are the Sea, Lake and Overland Surges from Hurricanes (SLOSH) and Advanced Circulation (ADCIRC) models. SLOSH can quickly evaluate the storm surge potential from multiple storms, but has a limited domain size and resolution. ADCIRC uses an unstructured grid which allows for fine scale coastal features and incorporates the ocean basin in its predictions, but it is costly to run, requiring “a large amount of computing resources in order to compute ensemble forecasts without the degradation of their resolution benefits.”

The study’s team identified essential capabilities for storm surge forecasting as “ensemble based calculations and simulations containing resolution sufficient to capture multiple length scales.” The limitation of SLOSH and ADCRIC is that they “each have these capabilities independently but not simultaneously.” The new model presented in this study adapted GEOCLAW, which originally was developed as a tsunami modeling system at the University of Washington. The goals were to determine if using an AMR-based code, which is a core capability of GEOCLAW, could “satisfy both without overly sacrificing either need” and “bridge the gap between the numerical cost of the unstructured grid storm surge models and the efficient but unresolved models.”

The researchers’ new approach leverages AMR algorithms to “retain the resolution required to resolve coastal inundation but only when necessary so that ensemble calculations are still feasible.” Nested structured grids vary in time, space, and resolution quality to lower computational costs. The key benefit of AMR “is the ability to change resolution as the simulation progresses.” A number of refinements are written into the algorithm, including location and strength of the storm, variable friction, wind speeds, and other physics-based criteria that enable calculations to adapt while they are being processed.

In order to evaluate GEOCLAW as a potential forecasting tool, scientists used it to re-create a storm surge prediction for Hurricane Ike. The team compared their simulation results with ADCIRC predictions and gauge data taken during the actual storm. The GEOCLAW excelled in its ability to run numerous simulations at once. In a two hour period, ADCIRC ran four simulations while the GEOCLAW model ran over a thousand. While noting that neither model precisely predicted the actual amount of surge seen in the storm, simulations showed that AMR is “able to capture many of the fine-scale features near coastlines that ADCIRC is able to capture.” Additionally, AMR reduced the “amount of effort needed to create the detailed unstructured grids,” thus lowering “computational cost considerably while providing comparable accuracy.” The reduction in cost makes this framework especially appealing for those “regions at risk which do not have the type of resources required to create these detailed grids ahead of a storm, both in terms of personnel and software.”

The researchers concluded that the new model framework compared favorably, with the adaption being most accurate starting at 18 hours before the storm. Combined, these results suggest that AMR is a “compelling way to forecast storm surge with both ensemble calculation and high resolution capabilities.”

In their discussions, the researchers noted that “the use of AMR in storm surge forecasting will require additional work on the refinement criteria” to further balance resolution with cost. They also cautioned that GEOCLAW “not be taken as the optimal settings,” learning from this study that “these settings were changed and run multiple times leading to speedup of a factor of ten from the initial settings.” Acknowledging that more work needs to be done to determine the best settings, they suggest that future studies focus on identifying specific aspects of AMR that provide “the largest impact and adopt them into existing codes which are already in operational use.”

The study’s authors are Kyle T. Mandli and Clint N. Dawson (Ocean Modelling, 2014, 75, 36–50).


This research was made possible in part by a grant from BP/The Gulf of Mexico Research Initiative (GoMRI) to the Consortium for Advanced Research on Transport of Hydrocarbon in the Environment (CARTHE). Other funding sources included the Office of Naval Research Grant 676 N00014-09-1-0649, the ICES postdoctoral fellowship program, and the King Abdullah University of Science and Technology Academic Excellence Alliance.

The GoMRI is a 10-year independent research program established to study the effect, and the potential associated impact, of hydrocarbon releases on the environment and public health, as well as to develop improved spill mitigation, oil detection, characterization and remediation technologies.  An independent and academic 20-member Research Board makes the funding and research direction decisions to ensure the intellectual quality, effectiveness and academic independence of the GoMRI research.  All research data, findings and publications will be made publicly available.  The program was established through a $500 million financial commitment from BP.  For more information, visit

60 Flares Twitter 7 Facebook 45 Google+ 8 StumbleUpon 0 Pin It Share 0 Reddit 0 LinkedIn 0 Email -- 60 Flares ×