Estimated increase in inundation probability with confidence intervals for the Gulf of Mexico


Journal Title

Journal ISSN

Volume Title




A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Coastal Marine System Science.
The main objective of this research is to study the impact of sea level rise on the relative increase in frequency of inundation for the low-lying coastal zones of the Gulf of Mexico caused by storms of different sizes. The research is based on locations around the Gulf of Mexico that benefit from existing long term sea level records and are located near population centers: Galveston Pier 21, Galveston Pleasure Pier, Port Isabel, Rockport, Texas, Grand Isle, La, and Pensacola, Key West, and St Petersburg, Florida stations. The stations' long-term water level records are divided into a long term sea level trend, a tidal component and a stationary surge component. Several extreme value distributions, such as three and four parameters Burr, Dagum, log-logistic, and generalized extreme value distribution (GEV), are compared using multiple statistical measures for the modeling of maximum annual storm surges. While differences are small the GEV and log logistic distributions are selected for this work based on performance, sensitivity to the series outliers and ease of implementation. Increases in inundation frequencies are computed by combining the stations' respective annual maximum surge models with two possible sea level rise scenarios, a conservative linear continuation of the past century trend and a scenario based on the upper limit of the sea level range in the IPCC (Intergovernmental Panel on Climate Change) AR4 report (Assessment Report 4), i.e. the A1FI scenario. Differences in oceanographic setting are discussed and affect vulnerability to sea level rise. To compare vulnerability to sea level rise, the ratios of future and present exceedance probabilities are computed for a range of water levels. The locations' respective vulnerabilities to sea level rise are assessed by comparing the maximum ratios of future to present water level exceedance probabilities and the corresponding water levels. Water levels at maximum ratios have a strong correlation with most common moment- and quantile - based statistics of surges, except the maximum annual surges. This indicates that the results of this study are not overly sensitive to the most extreme values or largest surge on the record provided that the record includes at least one large surge. Statistical bootstrap methods are used to estimate 90% and 95% confidence intervals for increases in inundation probability. For most cases the confidence intervals show a substantial decrease in interval width for stations with lengths of datasets of 50 years or longer indicating a preferred data length provided that a large surge event is included. For all locations the lower bounds of the confidence intervals imply significant increase in exceedance probabilities for both sea level rise scenarios. While expected increases in inundation frequencies are substantial for all stations, the results show considerable variation depending on the sizes of the surges, the station locations and the sea level rise scenarios. Annual maximum water levels resulting from small storms/surges will have higher frequencies, typically by a factor of 3 or more, than the historical frequency of water levels resulting from large hurricanes. As a result more frequent, smaller storm surges may have a larger impact on coastal communities than the effects of the less frequent, larger storm surges. Ratios of the exceedance probabilities depend mostly on sea level trends and the shape of the curves of the exceedance probabilities. The relative importance of these parameters depends on the sea level rise scenario. For a continued linear sea level rise maximum ratios are strongly correlated to the sea level trends or vertical land motion. For the conservative sea level rise scenario the study's highest increase in water level exceedance probability of 17 times is computed for a water level of 1.23m above present mean sea level for Grande Isle, Louisiana. For higher rates of global sea level rise local subsidence becomes less important and the dominant factor becomes the range of the locations' surges. For the study's A1FI based sea level rise scenario, the highest increase in water level exceedance probability is over 100 times for a water level of 0.83m above present mean sea level for Key West, Florida. The results of this research provide coastal decision makers quantitative estimates of future inundation risks for two sea level rise scenarios and a calibrated method to compute such risks for more sea level rise scenarios. This research is relevant for engineers, planners, insurance executives, and others to take into account the increasing impacts of storm surges of various sizes as sea level rises. The results will help develop better insurance rates, plan structures, land-use zoning, and others as the century progresses. The models, methodology and estimates developed as part of this research may be used to estimate the time before specific locations may become economically uninhabitable due to surge inflicted damages as sea level rises. Particularly, it is expected that this work will allow better to quantify coastal vulnerability to sea level rise along the Gulf of Mexico.
Physical and Environmental Sciences
College of Science and Engineering