A storm surge database for the US Gulf Coast

Authors

  • Hal F. Needham,

    Corresponding author
    1. Department of Geography and Anthropology, Louisiana State University
    • Department of Geography and Anthropology, Southern Climate Impacts Planning Program (SCIPP), Louisiana State University, Baton Rouge, LA 70803, USA.
    Search for more papers by this author
  • Barry D. Keim

    1. Louisiana State Climatologist, Department of Geography and Anthropology, Louisiana State University,
    Search for more papers by this author

Abstract

Tropical cyclone-generated storm surges are among the most deadly and costly natural disasters to impact the United States. Unfortunately, no comprehensive storm surge dataset provides reliable, historical storm surge information for the United States. This paper introduces SURGEDAT, the first comprehensive surge database for the US Gulf Coast. SURGEDAT identifies the location and height of peak storm surge for 195 surge events since 1880. A total of 62 sources were utilized to construct this dataset, including 28 Federal Government sources, numerous academic publications, and more than 3000 pages of newspaper from 16 daily periodicals. Surge heights in this database range from Hurricane Katrina's 8.47 m surge to 1.22 m surges generated by 20 separate events. Spatial analysis reveals enhanced surge magnitudes and frequencies along the Central and Western Gulf Coast, as well as the Florida Keys, while reduced levels of surge activity were observed along the eastern Florida Panhandle and the West Coast of Florida. The methods utilized to create this database are thoroughly documented, a table provides the comprehensive list of surges, and a map provides the spatial distribution of peak surge events. SURGEDAT will be valuable to coastal stakeholders, including planners and emergency managers, coastal scientists, and the hurricane and storm surge research communities. Copyright © 2011 Royal Meteorological Society

1. Introduction

Tropical cyclone-generated storm surges are among the most deadly and costly natural disasters to impact the United States. The most extreme of these events have occurred along the Gulf Coast. The 1900 Galveston Hurricane, for example, claimed more than 6000 lives (Rappaport and Fernandez-Partagas, 1995), while Hurricane Katrina (2005), caused over $ 80 billion (US) in damage (McTaggart-Cowan et al., 2008).

Storm surges also dramatically alter the coastal environment, moving barrier islands and cutting channels through them, destroying forests and crops, inundating the coastline with salt water, and damaging habitats essential to support wildlife. Hurricane Beulah's (1967) 5.49 m surge, for example, slashed 21 cuts through South Padre Island, Texas, dramatically altering the landscape (Sugg and Pelissier, 1968). Storm surge and waves produced by strong hurricanes along the Louisiana Coast typically move barrier islands near 100 m landward (Stone et al., 1997).

Unfortunately, a thorough literature review reveals no comprehensive storm surge dataset providing reliable, historical storm surge information for the US Gulf Coast. As a result, coastal stakeholders may make decisions unaware of actual storm surge risks or past surge history. For example, planners in coastal communities may not accurately realize the risk of buildings and infrastructure to storm surge inundation. Also, physical scientists may study geomorphological processes, such as dune migration, without key information about past surges (Martinho et al., 2010). Furthermore, a storm surge dataset would be useful to storm surge modelers to validate hindcasts and improve forecasts (i.e. Kurian et al., 2009).

In the case of storm surge modelling, the availability of storm surge observations has proven essential for validating model runs. Multiple surge observations, for example, helped verify surge model runs for Hurricane Charley (2004) in Southwest Florida, using the FVCOM finite volume surge model, which previously ran successful simulations of storm surge scenarios in the Tampa Bay area (Weisberg and Zheng, 2006a).

This paper discusses the creation of the first comprehensive storm surge database for the US Gulf Coast, called SURGEDAT. This database was created in the image of HURDAT (Jarvinen et al., 1984). While HURDAT is a database of hurricane tracks and intensities, SURGEDAT identifies the location and height of maximum storm surge levels associated with hurricanes and tropical storms. Although this information does not depict the regional extent of specific surges, or historic surge levels at a particular location, these data are useful for understanding historical storm surge activity and surge height potential for the entire basin and regions within the basin.

2. Defining thresholds for surge height, geographic and temporal ranges

SURGEDAT utilizes a minimum surge threshold of 1.22 m (4 ft), thereby including all surges that equal or exceed this truncation level. This height was chosen because the National Hurricane Center associated this surge level with the landfall of a category one hurricane (sustained winds greater than 33 m/s) on the Saffir-Simpson Scale at the time that this research commenced (National Hurricane Center, 2002). After Hurricanes Katrina (2005) and Ike (2008) generated surge heights well beyond generalisations suggested by the Saffir-Simpson Scale, the National Oceanic and Atmospheric Administration dissociated this wind scale from suggested surge heights (National Oceanic and Atmospheric Administration, 2010).

The temporal range for SURGEDAT is the 131-year period from 1880 to 2010. This time frame is chosen because Elsner and Jagger (2007) note that the most reliable historical hurricane data are available beginning approximately in 1880. Also, Landsea et al. (2004) estimate that historical tropical cyclone records cover most of the US Gulf Coast by 1880, with the exception of southwest Florida.

The United States Gulf of Mexico (GOM) coastline is chosen as the geographic area of this analysis for several reasons. Storm surge inundations from tropical cyclones embody the most severe disasters along this coastline, in terms of both human and financial losses, and will likely cause the most catastrophic losses associated with potential climate change in this region (U.S. Global Change Research Program, 2009). This research establishes a climatological method on a relatively small scale that can be applied in future research at larger scales, such as the entire United States East Coast, or the Atlantic Basin (Jarvinen et al., 1984). Also, extreme GOM storm surges are generally considered larger than those recorded along the East Coast of the United States. The boundaries for this region extend from South Padre Island, Texas, eastward along the entire Gulf Coast to the southern tip of the Florida Everglades. The Florida Keys are included as the extreme southeast border of the GOM, following the example set by Keim and Muller (2009).

3. Creating the SURGEDAT database structure

A two-step process was utilized to create SURGEDAT. The first step assembled a database structure that contained every tropical cyclone that potentially generated a 1.22 m storm surge. The second step examined historical records to find the maximum storm surge level and location for each event in the database.

Every tropical cyclone that likely produced maximum sustained winds of at least tropical storm force (sustained winds greater than 17 m/s) along the US Gulf Coast was included in the database. Although the National Hurricane Center associated category 1 hurricanes with surges at least as high as 1.22 m (National Hurricane Center, 2002), tropical storm-force winds were chosen as the wind threshold because large or slow-moving tropical storms sometimes produce surges greater than 1.22 m. For example, Tropical Storm Frances (1998) only generated 21 m/s winds when making landfall along the Texas Coast, but the system generated a peak surge of 2.44 m at Matagorda Locks, Texas (Lawrence 1998). There are many examples like this in our database.

A 300 km buffer from the US Gulf Coast was utilized to incorporate tropical cyclones that potentially generated tropical storm force winds at the coast, even if they did not make landfall. Every tropical cyclone that passed within this buffer was included in SURGEDAT. Keim et al. (2007) generalize that winds of at least tropical storm force extend outward 240 km on the strong side of averaged-sized Category 3–5 hurricanes, however, because larger-than-average hurricanes produce larger wind fields, a 300 km buffer was adopted. Tropical cyclones that do not pass within 300 km of the US Gulf Coast are unlikely to generate tropical storm force winds at the coast and were therefore excluded from the database.

Hurricanes Anita (1977) and Lili (1996) are examples of storms that generated surges along the US Gulf Coast although they were centered near the outer edge of the 300 km coastal buffer. Anita generated a 1.68 m surge at South Padre Island, Texas, even though the hurricane made landfall in Mexico, about 270 km south of Brownsville, Texas (Lawrence, 1978). Although Lili's closest approach to the US was approximately 250 km, the hurricane produced a 1.83 m surge in the Florida Keys (Williams and Duedall, 2002).

Tropical cyclone best-track maps provided by the Unisys Corporation (2011) were utilized to identify tropical cyclones that passed within the buffer. These maps provide tropical cyclone tracks from the North Atlantic Basin from 1851 to date, from data made available by The National Oceanic and Atmospheric Administration's (NOAA) Tropical Prediction Center. The list of storms extracted from Unisys data were then verified with a dataset contributed by Landsea (2005), made available through the Atlantic Oceanographic and Meteorological Laboratory (AOML), which provided a list of hurricanes that impacted the US between 1851 and 2008. These two sources provided a list of 425 tropical cyclones that potentially produced storm surges along the US Gulf Coast.

Although most tropical cyclones included in this dataset produced no more than one storm surge, some produced two distinct surges, separated by at least 12 h. The most common double-surge pattern was an initial surge in southwest Florida or the Florida Keys, followed by a second surge event elsewhere in the GOM. These double surges were also included into SURGEDAT, adding 42 additional surge events to the database, and producing a total of 467 potential surge events along the US Gulf Coast.

4. Storm surge data sources

After the list of 467 potential storm surge events was created, we attempted to find the maximum surge level for as many events as possible through archival research. This research utilized 62 sources from several types of literature, including 28 Federal Government sources, numerous academic publications, and 16 newspaper sources.

Government documents include all relevant information published by agencies in the US Federal Government. Such sources include the National Hurricane Center, U.S. Army Corps of Engineers, United States Geologic Survey (USGS), National Oceanic and Atmospheric Administration Tides and Currents Program, and the National Weather Service. These sources are listed in Table I.

Table I. Federal Government publications utilized for historical storm surge research
Source and locationData date(s)Publication date(s)
National Hurricane Center, Archive of Hurricane Seasons http://www.nhc.noaa.gov/pastall.shtml1958–2009Current
National Oceanic and Atmospheric Administration (NOAA) Tides and CurrentsDepends on tidal gaugeCurrent
http://tidesandcurrents.noaa.gov/  
U.S. Army Engineer District, Galveston, Texas, 1968: Hurricane Beulah, Sept. 8–21 1967, 26 pp.19671968
U.S. Army Engineer District, New Orleans, Louisiana, 1975: Hurricane Carmen, 7–8 September 1974, 78 pp.19741975
U.S. Army Engineer District, Galveston, Texas, 1981: Report on Hurricane Allen, 3–10 August, 1980.19801981
United States Geologic Survey, Inland Storm-Tide Documentation Program2005–PresentCurrent
http://water.usgs.gov/osw/programs/storm_surge.html  
US Department of Commerce, Environmental Science Services Administration, “Some Devastating North Atlantic Hurricanes of the 20th Century”1900–19691970
National Weather Service, 1973: Preliminary Reports on Hurricanes and Tropical Storms, Hurricane Agnes, June 14–23, 1972, 193 pp.19721973
HURDAT reanalysis (until 1910)1851–19102004
See Landsea et al. (2004)  
HURDAT reanalysis (until 1920)1851–19202008
See Landsea et al. (2008)  
HURDAT reanalysis (until 1925)1851–19252009
See Landsea (2009)  
Harris DL. Characteristics of the Hurricane Storm Surge1926–19611963
Technical Paper No. 48 (Weather Bureau)  
Hurricane Elena, Gulf Coast, August 29- September 2, 1985; Prepared by Peter Sparks et al. (1991)19851991
Turpin RJB. Analysis of the Tropical Cyclone Threat at Key West. Published in Hurricane Havens Handbook for the North Atlantic Ocean, through the Royal Navy ExchangeVariable1982
U.S. Army Corps of Engineers, History of Hurricane Occurrences along Coastal Louisiana1559–19711972
Connor, 1956. Preliminary Summary of Gulf of Mexico Hurricane Data. Report from the New Orleans Forecast OfficePre-19561956
Roth D, 2010: Texas Hurricane History. National Weather Service, Camp Springs, Maryland.1527–20102010
National Weather Service Southern Region HeadquartersVariableCurrent
National Weather Service, Brownsville, TX. http://www.srh.noaa.gov/bro/VariableCurrent
National Weather Service, Corpus Christi, TXVariableCurrent
http://www.srh.noaa.gov/crp/  
National Weather Service Houston/Galveston, TXVariableCurrent
http://www.srh.noaa.gov/hgx/  
National Weather Service Lake Charles, LAVariableCurrent
http://www.srh.noaa.gov/lch/  
National Weather Service New Orleans/Baton Rouge, LAVariableCurrent
http://www.srh.noaa.gov/lix/  
National Weather Service Mobile, AL/Pensacola, FLVariableCurrent
http://www.srh.noaa.gov/mob/  
National Weather Service Tallahassee, FLVariableCurrent
http://www.srh.noaa.gov/tlh/  
National Weather Service Tampa Bay, FLVariableCurrent
http://www.srh.noaa.gov/tbw/  
National Weather Service Miami-South Florida, FLVariableCurrent
http://www.srh.noaa.gov/mfl/  
National Weather Service Key West, FLVariableCurrent
http://www.srh.noaa.gov/key/  

Academic publications include books, academic journal articles and online resources that generally provided indepth history of a specific storm in one location, or an overview of many storms that impacted a region. Examples include Roberts' (1969) Extreme Hurricane Camille, August 14th through 22nd, 1969, and Barnes' (2007) Florida's Hurricane History. These sources are listed in Table II.

Table II. Academic Publications
SourcePublication date(s)
The Monthly Weather Review, Journal of the American Meteorological Society1880–2008
Baker et al., The Social Impact of Hurricane Eloise on Panama City, Florida1976
Barnes, Florida's Hurricane History, second edition2007
Cline, Tropical Cyclones1926
Clyburn et al., A Furious Form of Wind: Hurricanes and Their Impact on Two Texas Coastal Cities1989
Dunn and Miller, Atlantic Hurricanes1960
Ellis MJ, 1988: The Hurricane Almanac—1988 Texas Edition.1988
Elsner, Hurricanes of the North Atlantic: climate and society1999
Keim and Muller, Hurricanes of the Gulf of Mexico2009
Manderson, Florida's Historical Hurricanes, published online by the International Hurricane Research CenterUnknown
Murnane and Liu, Hurricanes and Typhoons, Past, Present, and Future2004
National Public Radio, Deadly Hurricanes No Strangers to Gulf Coast (Website)2005
Roberts NC, Extreme Hurricane Camille, August 14th through 22nd, 19691969
Tannehill, Hurricanes1944
USA TODAY, Texas Hurricane History (Website)Unknown
Williams and Duedall, Florida Hurricanes and Tropical Storms, Revised Edition1997
Williams and Duedall, Florida Hurricanes and Tropical Storms, 1871–2001, Expanded Edition2002
Zebrowski and Howard, Category 5: The story of Camille, lessons unlearned from America's most violent hurricane2005

Newspapers provided local accounts of tropical cyclone conditions, often describing the storm surge height and damage. This research utilized daily periodicals published in cities or towns located on the US Gulf Coast, as well as some online newspapers that published storm accounts from syndicated national stories. Dates of newspaper availability depended on the specific publication. Newspapers from communities impacted by tropical cyclones were sought for three consecutive dates, beginning with the date of landfall. Table III gives a complete list of newspapers.

Table III. Newspapers utilized for historical storm surge database. These periodicals were ordered from libraries along the Gulf coast through the Interlibrary Loan program at Louisiana State University
Title of periodicalLocation
The Brownsville HeraldBrownsville, TX
The Corpus Christi CallerCorpus Christi, TX
The Galveston Daily NewsGalveston, TX
The Beaumont EnterpriseBeaumont, TX
The Lake Charles Daily PressLake Charles, LA
The Times-PicayuneNew Orleans, LA
The Daily HeraldBiloxi, MS
The Mobile RegisterMobile, AL
The Pensacola JournalPensacola, FL
The Panama City News-HeraldPanama City, FL
The Saint Petersburg TimesSaint Petersburg, FL
The Tampa TribuneTampa, FL
The Fort Myers News-PressFort Myers, FL
(or Tropical News) 
The Naples Daily NewsNaples, FL
Collier County NewsNaples, FL
The Key West CitizenKey West, FL

Approximately 780 newspaper dates were obtained through The Louisiana State University Inter-Library Loan (ILL) system, and approximately 260 newspaper dates were obtained through online newspaper archives, totaling approximately 1040 newspaper dates. Given the assumption that an average of three newspaper pages were read for each newspaper date, this portion of the research analysed over 3000 pages of newspaper from 16 different sources.

Although anecdotal accounts, such as historic newspapers, were most useful in the earlier years of this study, surges in the database cannot be categorized as exclusively anecdotal or scientific. Research for the vast majority of surges utilized information from a wide variety of sources to determine the height and location of peak storm surge. For surge events that incorporated both scientific and anecdotal data, scientific sources were usually given more weight and anecdotal accounts were generally used for verification. SURGEDAT utilizes a confidence rating to determine the level of credibility associated with each surge event. The rating ranges from one to five, with higher values reflecting more confidence. Surge events that utilize one credible scientific source that is not contradicted by other sources are ranked ‘3’ and surge events that utilize two credible scientific sources that validate each other are ranked ‘4’. Uncertainty regarding the influence of tides or waves generally lowered confidence levels by at least one category. A summarized list of these categories is provided in Table IV.

Table IV. Confidence levels assigned to classified storm surge events. Higher confidence levels indicate increased accuracy of storm surge estimation. Uncertainty regarding the influence of tides or waves generally lowered confidence levels by at least one category
Confidence LevelCategoryDescription
1Very lowAt least one anecdotal or scientific source is provided, but not from the area of peak surge (i.e. tide gauge located 50 km from area of peak surge). Also, major contradictions of sources fall into this category.
2LowAt least one anecdotal source from area of peak surge, or at least one scientific source that contradicts credible anecdotal accounts or other scientific sources.
3ModerateOne credible scientific source that is not contradicted
4HighTwo credible scientific sources validate each other
5Very highAt least two credible scientific sources, plus network of tide gauges from the area of peak surge

Surge research conducted on an unnamed, category-1 hurricane that struck the Central Texas Coast in September, 1941, reveals how various types of surge data were utilized to determine the height and location of peak surge. Six sources of data were utilized, including one federal source (U.S. Army Corps of Engineers), two academic publications, and accounts from three separate newspapers. The federal source and academic publications were less descriptive in nature, as these sources only provided the maximum surge levels. The newspaper accounts were a bit more descriptive, providing damage accounts, and two of the three articles also provided quantitative levels of maximum surge height. These varied accounts cross-validated each other well and revealed that the hurricane produced peak surge levels of 3.35 m at Matagorda, 3.02 m at Sargent, 2.74 m at the Freeport-Velasco Bridge, and 2.13 m at Galveston. A peak surge level of 3.35 m at Matagorda was entered into the database, as this highest surge observation was provided by the U.S. Army Corps of Engineers, a credible, scientific source, and the other sources supported this general surge level. Table V provides a summary of storm surge information related to this surge event.

Table V. Sample storm surge information provided by six separate sources for a surge event along the Central Texas Coast in September 1941
SourceInformation/QuoteMaximum surge height
Sumner HC, Monthly Weather Review, October 1941, pg. 265Quote: Tides had been somewhat above normal at Galveston since the minor disturbance of September 11–15 and began to rise again on the 21st, and more rapidly to a crest of 6.7 feet at 8 p.m. and 10 p.m. CST. On the 22nd, then falling to 5.0 feet at 1 p.m. of the 23rd. Tides rose again thereafter to a crest of 7.0 feet at 9 and 10 p.m. C.S.T. on the 23rd, after which they subsided rapidly.2.13 m (7 ft) at Galveston, TX
The Galveston Daily News, September 24, 1941, pg. 1Quote #1: Matagorda's 1250 residents were believed sheltered in safety. High tides reportedly spread through the entire town.No quantitative height, however, surge definitely inundated Matagorda, TX
 Quote #2: A rising tide, which was up to 7 feet above mean low by 10 o'clock, rose in drains and kept the rain water from draining. [This account was from Galveston, TX]2.13 m (7 ft) at Galveston, TX
The Freeport Facts, September 25, 1941, pg. 2Quote: At the height of the storm the nine-foot tide was up to the bottom of the Freeport-Velasco bridge and the high waves were splashing over the crossing.2.74 m (9 ft) at Freeport-Velasco Bridge
The Corpus Christi Times, September 24, 1941, pg. 1Quote: Constable AP Moore of Bay City returned from an inspection of Matagorda, a small town near the center of the hurricane, to report only minor damage to buildings there from the blow. Shallow backwater covered most of the town, he said, and residents were returning to their homes.No quantitative height, however, surge definitely inundated most of Matagorda, TX
Dunn and Miller, 1960: Atlantic HurricanesTable XVI on page 219 gives a maximum storm surge of 9.9 feet above m.s.l. at Sargent, TX, on September 23, 1941.3.02 m (9.9 ft) at Sargent, TX
U.S. Army Corps of Engineers, 1972: History of Hurricane Occurrences along Coastal LouisianaQuote: Tides were considerably high along the coast, a maximum of 11 feet occurring at Matagorda.3.35 m (11 ft) at Matagorda, TX

5. Estimating surge levels, referencing datums, removing tides and waves

Precise surge levels were obtained for events in which tide gauges measured the highest surge level. In those cases, the normal astronomical tide level was subtracted from the maximum observed water level to obtain the height of maximum storm surge. However, when tide gauge data were not available, or gauges were not located in the area of peak storm surge, maximum storm surge levels were estimated from scientific and anecdotal documentation. Generally, these estimations provided water levels above normal, sometimes rounded off to the nearest foot. For example, SURGEDAT lists a peak surge level of 1.22 m (4 ft) at Cedar Key, Florida for Tropical Storm Irene (1959), based on the following quote:

The highest tide at Pensacola was 1.6 ft. above normal but Cedar Keys, Florida, about 120 miles east of the storm's path, reported tides that were 4 feet above normal. (U.S. Weather Bureau 1959, pg. 1).

Some sources listed surge data as a range of values; in those cases the maximum level of this range was generally used. For example, SURGEDAT lists a peak surge value of 2.44 m (8 ft) for Hurricane Danny (1985), based on the following account:

The primary coastal affect was tides ranging from 2 to 3 feet above normal along the Alabama and Mississippi coasts to 5 to 8 feet above normal estimated along the Louisiana coast near and to the right of where the center moved on shore.” ( National Hurricane Center 1985, pg. 2).

A limitation of SURGEDAT is the inability to distinguish between datums, or set references of mean sea level (MSL). The most common datums referenced in coastal research are the National Geodetic Vertical Datum (NGVD 29), which refers to MSL in 1929, and the North American Vertical Datum (NAVD 88), which refers to MSL in 1988. Most events in the SURGEDAT dataset refer to water height above normal astronomical tide levels for the year of observation, however, a few events, like Tropical Storm Mitch (1998), reference height above NGVD 29 (Guiney and Lawrence 1998), while other events, like Hurricane Rita (2005), incorporate data that are referenced to NAVD 88 (McGee et al., 2006). Most surge events in the database incorporated surge observations from multiple sources, most of which did not specifically reference NGVD 29 or NAVD 88. Also, some sources did not measure water height above MSL, but instead used another reference, such as Mean Low Water (MLW), referring to the lowest tide level. The 3.54 m surge that peaked in Mobile, Alabama in 1916 incorporated data referenced to MLW (U.S. Army Corps of Engineers, 1972). Although these differences are not resolved in the final dataset, the SURGEDAT metadata file, located online, documents datum references when available.

Hurricane Rita provides an example of how datum decisions were made while building the database. In this case, the USGS provided a peak surge observation of 4.54 m referenced to NAVD 88 (McGee et al., 2006). However, other sources, such as Knabb et al. (2006), Beven et al. (2008), and Keim and Muller (2009) provided a peak surge level of at least 4.57 m, but did not reference this datum. Although we took this observation (referenced to NAVD 88) into consideration, we classified this event as a 4.57 m surge, not referenced to NAVD88, based on the general consensus of these sources. Because SURGEDAT included multiple sources for most surge events, it was not possible to assign a specific datum to each event. Furthermore, given irregular rates of subsidence along the Gulf Coast (Morton et al., 2005), it was not possible to equate events over time to a common datum.

Variations of water levels due to tidal influences were removed from the data as much as possible. When tide gauge data were available, predicted tide levels were subtracted from observed water levels to obtain the maximum surge value. For cases in which tide gauge data were not available, scientific and anecdotal documentation often distinguished between the maximum observed water level and the water height above normal, essentially differentiating between storm tide and storm surge. For example, Lawrence and Blake (2002) differentiate between a storm tide observation of 1.89 m (6.2 ft) and a storm surge observation of 1.55 m (5.1 ft) in Charlotte County, FL during Hurricane Gabrielle in 2001. The value of 1.55 m is entered into SURGEDAT as the best storm surge estimation.

In some cases, however, storm tide and storm surge were not differentiated. This limitation in SURGEDAT is partly due to the fact that the term, storm surge, was not in use during the early part of this study. A search through more than 600 pages of storm surge metadata utilized to construct SURGEDAT finds the term, storm surge, first used in 1961 (Dunn, 1961). In the earlier portion of this study the term, storm tide, is often used in the way storm surge is used today, essentially defined as water height above normal tide levels.

Fortunately, for cases in which the term, storm tide, is not clearly defined, tidal ranges along the US Gulf Coast are low, minimising potential errors in the database related to tidal influence on maximum water levels. The tidal range in this microtidal, semi-enclosed water body averages only 0.43 m (Coleman, 1988). This tidal range is much lower than some other basins that observe tropical cyclone-generated storm surges, such as the Bay of Bengal (2 + m tidal range), East Asia (3 + m tidal range), and Australia (4 + m tidal range in Western Australia, 2 + m tidal range in Eastern Australia), (Chittleborough J, (unspecified year)).

The influence of waves on maximum sea levels was also removed from the dataset as much as possible, as tropical cyclones in the Gulf of Mexico often generate large, destructive waves that elevate high-water marks beyond the range of storm surge. Hurricane Ivan, for example, produced enormous waves that damaged offshore energy platforms to a height of 27 m (Stone et al., 2005), however, the peak storm surge level was only 4.57 m at landfall (Stewart, 2005). Tests conducted using the ADCIRC storm surge model concluded that wave forces can elevate sea surface levels 30–50% (Weaver, 2004). In the coastal zone, large waves have also increased maximum water levels beyond the level of peak storm surge. Rafted debris and damaged trimlines, left behind by the combined forces of storm surge and waves, reveal that Hurricane Katrina generated maximum water levels exceeding 10 m in Biloxi and Waveland, Mississippi (Fritz et al., 2008). However, the peak surge level, which removes the influence of waves, was estimated at 8.47 m (Knabb et al., 2006).

The manner in which wave heights were removed from scientific surge observations depends upon the data type. Waves sometimes appear as noise on tide gauge graphs that use short temporal resolutions; in these cases, the surge height was visually estimated at the lower end of observed data, where MSL rose to produce a new baseline level. In regards to post-storm surge field work, the credibility of particular surge observations depends upon the practices followed by particular agencies or organisations. Generally, the highest quality surge data comes from sources that took precautions to discount the effect of waves on maximum water levels, such as measuring maximum water levels from the interior of flooded buildings. This practice was incorporated in many field work investigations, such as a team of researchers who conducted storm surge surveys following Hurricane Opal in 1995, concluding: ‘Still water mark elevations inside of buildings or tide gage maximums, which damp out breaking wave effects and are indicative of the storm surge, ranged from 5 to 14 feet above mean sea level’ (Mayfield, 1995, pg. 1). This source also concluded that waves increased the maximum water level approximately 3 m above the surge level, based upon the difference in elevation of a 2.53 m surge at the Panama City Beach Pier tide gauge and a 5.49 m debris line at the end of the pier.

The effect of waves was also minimized in the anecdotal data. Written accounts and historical photographs were analysed to find levels of standing water in surge events, not accounts that included descriptions of wave action. Because anecdotal sources are often quite descriptive, separating high-water marks associated with storm surge and wave run-up was usually possible. The following quote from the Freeport Facts (25 September 1941, pg. 1) describes a nine-foot storm surge overtopped with high waves: ‘At the height of the storm the nine-foot tide was up to the bottom of the Freeport-Velasco Bridge and the high waves were splashing over the crossing.’

6. Database characteristics

After examining data for the 467 hurricanes and tropical storms that could have created a surge of 1.22 m, this study identified the maximum storm surge location and height for 195 surge events. The number was reduced because many tropical storms and even a few hurricanes produced surges below the 1.22 m threshold, while other storms may have produced a surge above the threshold, but we could not find credible information about the peak storm surge level. Surge levels ranged in height from Hurricane Katrina's 8.47 m surge along the Mississippi Coast, to 1.22 m surges generated by 20 separate events. Table VI provides the completed storm surge database, including maximum surge height, peak surge location, surge height rank, as well as the name and year of the associated tropical cyclone.

Table VI. Locations and heights of peak storm surges along the US Gulf of Mexico Coast, 1880–2010. The latitude and longitude values are best estimates of peak surge locations and do not represent the extent of surge flooding. Confidence levels are based on a scale of 1–5, with higher values denoting higher confidence in surge height and location. Hurricanes Anita (1977) and Gilbert (1988) produced peak surges in Mexico, however, the location and height of their peak US surges are included in this table
RankSurge (m)NameYearMax Surge LocationStateLatLonConf
18.47Katrina2005Pass ChristianMS30.3036− 89.28643
27.5Camille1969Pass ChristianMS30.3133− 89.28642
36.1Galveston1900GalvestonTX29.2889− 94.78943
36.1Labor Day1935Long KeyFL24.8122− 80.82512
55.64Carla1961Port LavacaTX28.6122− 96.62033
65.55Eloise1975Dune Allen BeachFL30.3494− 86.24313
75.49Beulah1967South Padre IslandTX26.1086− 97.16362
85.33Ike2008Chambers CountyTX29.5994− 94.67083
95.18New Orleans1915Southeastern LA, S of New OrleansLA29.5797− 89.90172
104.88Cheniere Caminada1893Cheniere Caminada (W of Grand Isle)LA29.2097− 90.05062
104.88Unnamed1919W shore Nueces Bay (near Corpus Christi)TX27.8389− 97.49083
124.72Galveston1915GLS seawall (Galveston?)TX29.2889− 94.78942
134.66Frederic1979Gulf State ParkAL30.2497− 87.66013
144.63Unnamed1947Bay Saint LouisMS30.3025− 89.33083
144.63Betsy1965MS River- West Pointe a la HacheLA29.5697− 89.79753
164.57Indianola1886IndianolaTX28.5181− 96.49533
164.57Grand Isle1909Sea Breeze (Mouth of Bayou Terrebonne/Terr Bay)LA/MS29.2542− 90.59143
164.57Unnamed1910Key WestFL24.5456− 81.80253
164.57Unnamed1945Port LavacaTX28.6122− 96.62033
164.57Ivan2004Destin, FL to Mobile, ALFL/AL30.3422− 87.31063
164.57Rita2005CameronLA29.7953− 93.32473
224.48Unnamed1942Matagorda and Port LavacaTX28.6419− 96.32253
234.27Unnamed1906Galt, Santa Rosa CountyFL30.3575− 86.99973
234.27Unnamed1919Cow KeyFL24.5597− 81.73972
234.27Great Miami1926BagdadFL30.5969− 87.02813
234.27Opal1995Florida PanhandleFL30.3875− 86.79942
274.24Audrey1957west of CameronLA29.7697− 93.45252
284.11Donna1960Upper Matecumbe KeyFL24.9133− 80.63533
293.96Unnamed1933South Padre IslandTX26.1119− 97.16443
293.96Flossy1956Ostrica LockLA29.3683− 89.52973
293.96Wilma2005SW FloridaFL25.7051− 81.28582
293.96Gustav2008Southeast LA, near Bay GardeneLA29.5594− 89.70473
333.85Alicia1983San Luis Pass (between Freeport and Galveston)TX29.0767− 95.12474
343.81Terrebonne Parish1926Terrebonne ParishLA29.2461− 90.66172
353.75Lili2002Crewboat Channel (Calumet)LA29.5742− 91.42063
363.66Unnamed1886Johnson's BayouLA29.7499− 93.65893
363.66Great Miami1926Fort Myers and Punta RosaFL26.6367− 81.88112
363.66Unnamed1944NaplesFL26.1401− 81.80753
363.66Allen1980Port MansfieldTX26.5558− 97.42424
403.63Georges1998Fort MorganAL30.2244− 88.02083
413.54Unnamed1916MobileAL30.6775− 88.03813
423.47Unnamed1949Houston Ship Channel (Harrisburg)TX29.6878− 94.99222
433.35Unnamed1941MatagordaTX28.6881− 95.96583
433.35Kate1985Cape San BlasFL29.6689− 85.32753
453.32Unnamed1888Southeastern LA, near Fourchon BeachLA29.1033− 90.18641
463.2Tampa Bay1921TampaFL27.9403− 82.46014
463.2Carmen1974CocodrieLA29.2461− 90.66173
483.05Unnamed1894W Big Bend region of FL PanhandleFL29.9117− 85.37611
483.05Unnamed1896Fort Walton BeachFL30.3967− 86.62032
483.05Unnamed1896Cedar KeyFL29.1397− 83.04193
483.05Unnamed1903ApalachicolaFL29.7208− 84.98113
483.05Velasco1909Galveston and VelascoTX29.1081− 95.08813
483.05Alma1966New Port Richey to Cedar KeyFL28.9078− 82.69113
483.05Elena1985ApalachicolaFL29.7208− 84.98114
552.96Edith1971Cypremort PointLA29.7161− 91.87723
562.8Unnamed1916Corpus ChristiTX27.7997− 97.39113
562.8Celia1970Port Aransas BeachTX27.8301− 97.05064
582.79Claudette2003Freeport USGSTX28.9342− 95.29893
592.74Unnamed1888Cedar KeyFL29.1397− 83.04193
592.74Unnamed1910S Padre IslandTX26.1119− 97.16441
592.74King1950Bahia Honda (bridge), E of Big Pine KeyFL24.6547− 81.28171
592.74Betsy1965North Key LargoFL25.3128− 80.27643
592.74Josephine1996Levy CountyFL29.1397− 83.04192
592.74Dennis2005Apalachee BayFL30.1036− 84.02313
652.68Unnamed1929Key LargoFL25.0833− 80.43692
662.53Isidore2002Rigolets, LA and Gulfport Harbor, MSLA/MS30.2622− 89.40033
672.44Unnamed1901MobileAL30.6775− 88.03812
672.44Unnamed1901Port EadsLA29.0169− 89.17082
672.44Unnamed1923BiloxiMS30.3922− 88.88362
672.44Unnamed1941St. MarksFL30.1519− 84.20893
672.44Easy1950Clearwater to SarasotaFL27.5094− 82.72112
672.44Danny1985South Central LA CoastLA29.5875− 92.14693
672.44Juan1985CocodrieLA29.2461− 90.66173
672.44Andrew1992CocodrieLA29.2461− 90.66173
672.44Allison1995Wakulla through Dixie CountiesFL29.9801− 83.80863
672.44Earl1998Northwest FL- Big Bend area E of ApalachFL29.8053− 84.73333
672.44Frances1998Matagorda LocksTX28.6803− 95.98013
672.44Bret1999Port Mansfield PassTX26.5628− 97.27642
792.41Debra1959Morgan PointTX29.6731− 94.98783
802.38Unnamed1917Fort BarrancasFL30.3451− 87.29752
802.38Hilda1964CocodrieLA29.2461− 90.66172
822.26Flossy1956Laguna BeachFL30.2386− 85.92443
832.13Unnamed1886Sabine PassTX29.7194− 93.86393
832.13Unnamed1909S Padre IslandTX26.1119− 97.16443
832.13Apalachicola1915CarrabelleFL29.8439− 84.66443
832.13Unnamed1921West Bay (N side of W Galveston Island)TX29.3028− 94.89892
832.13Unnamed1929SeadriftTX28.4086− 96.71812
832.13Unnamed1934Galveston IslandTX29.2889− 94.78942
832.13Unnamed1942High IslandTX29.5506− 94.38583
832.13Ethel1960Quarantine BayLA29.3825− 89.51894
832.13Agnes1972Cedar Key and Port St. JoeFL29.8111− 85.30673
832.13Delia1973Galveston BayTX29.6994− 94.95173
832.13Jerry1989BaytownTX29.7219− 95.00753
832.13Chantal1989High IslandTX29.5506− 94.38583
832.13Andrew1992GoodlandFL25.9225− 81.64564
832.13Erin1995just west of Navarre BeachFL30.3422− 87.08012
832.13Charley2004Sanibel and Estero IslandsFL26.4489− 82.01943
982.07Baker1950ApalachicolaFL29.7208− 84.98113
991.99Danny1997HWY 182 W, b/t Gulf Shores & Fort MorganAL30.2306− 87.80012
991.99Ida2009Bay GardeneLA29.5594− 89.70473
1011.98Unnamed1931FrenierLA30.1075− 90.42312
1011.98Unnamed1932MobileAL30.6775− 88.03812
1011.98Florence1953PanaceaFL30.0225− 84.38583
1011.98Gladys1968HomosassaFL28.7844− 82.62783
1051.95Unnamed1940FrenierLA30.1075− 90.42313
1061.92Bob1979Gulfport and Harrison County CDMS30.3614− 89.09283
1071.83Unnamed1886IndianolaTX28.5181− 96.49532
1071.83Unnamed1897Sabine PassTX/LA29.7194− 93.86393
1071.83Unnamed1909S Padre IslandTX26.1119− 97.16442
1071.83Unnamed1912Point IsabelTX26.0744− 97.19893
1071.83Unnamed1920Lake Borgne and Mississippi SoundMS/LA30.2625− 89.39393
1071.83Unnamed1929Panama City to ApalachicolaFL29.9408− 85.40922
1071.83Unnamed1936Fort Walton Beach, Panama City, ValparaisoFL30.3014− 86.08283
1071.83Unnamed1943HWY 11, near Chef MenteurLA30.1275− 89.86722
1071.83Unnamed1946Punta Gorda to BradentonFL27.1292− 82.47283
1071.83Unnamed1948MS CoastMS30.3931− 88.93033
1071.83Unnamed1948Key WestFL24.5472− 81.78532
1071.83How1951Fort Myers BeachFL26.4501− 81.94972
1071.83TS Brenda1955MS CoastMS30.3931− 88.93033
1071.83Cindy1963Mouth of Calcasieu RiverLA29.7622− 93.34422
1071.83Dora1964YankeetownFL29.0186− 82.72643
1071.83Debbie1965Industrial Canal in New OrleansLA29.9881− 90.02173
1071.83TS #11965Apalachicola (in vicinity)FL29.7208− 84.98112
1071.83Fern1971FreeportTX28.9439− 95.30943
1071.83Chris1982Cameron ParishLA29.7608− 93.56253
1071.83Florence1988Bayou BienvenueLA29.9994− 89.86033
1071.83Keith1988Bradenton and Fort MeyersFL27.0653− 82.44753
1071.83Gilbert1988S. Padre IslandTX26.0733− 97.15473
1071.83Lili1996Florida KeysFL24.5617− 81.68113
1071.83Georges1998Florida KeysFL24.5594− 81.68473
1071.83Frances2004Pinellas CountyFL27.8411− 82.84033
1071.83Cindy2005SE LA, MS, Lakes Borgne & PontMS/LA30.1751− 89.69923
1071.83Alberto2006HomosassaFL28.7844− 82.62782
1341.78Matthew2004FrenierLA30.1075− 90.42313
1351.74Debra1978Atchafalaya Bay to Vermillion BayLA29.7489− 91.66583
1361.71Abby1968Everglades CityFL25.8456− 81.38782
1371.69Bill2003Bayou BienvenueLA29.9994− 89.86033
1381.68Unnamed1936Everglades CityFL25.8456− 81.38783
1381.68Unnamed1947Everglades CityFL25.8456− 81.38782
1381.68Anita1977Port Isabel Coast Guard/S. Padre Is.TX26.0728− 97.16693
1381.68Beryl1988Bayou BienvenueLA29.9994− 89.86033
1421.58Bonnie1986Sabine PassTX29.7194− 93.86393
1431.55Gabrielle2001Charlotte CountyFL26.9539− 82.09673
1431.55Hanna2002Gulfport HarborMS30.3614− 89.09283
1451.52Unnamed1920TampaFL27.9403− 82.46012
1451.52Freeport1932GalvestonTX29.2889− 94.78942
1451.52Unnamed1933South Padre IslandTX26.1119− 97.16442
1451.52Unnamed1933Port IsabelTX26.0775− 97.20562
1451.52Labor Day1935Cedar KeyFL29.1397− 83.04192
1451.52Unnamed1945Tampa (Garcia Avenue Bridge- now Eugene Holtsinger)FL27.9597− 82.46813
1451.52Unnamed1945Key LargoFL25.0836− 80.43722
1451.52TS Hazel1953Everglades CityFL25.8456− 81.38782
1451.52Esther1957MS CoastMS30.3931− 88.93033
1451.52Isbell1964Key WestFL24.5453− 81.81082
1451.52Alma1966Fort Myers BeachFL26.4501− 81.94973
1451.52Inez1966Big Pine KeyFL24.6719− 81.33972
1451.52Babe1977Southeastern LALA29.2675− 90.93443
1451.52Claudette1979Sabine Coast Guard StationTX29.7286− 93.87083
1451.52Alberto1994Okaloosa Island to DestinFL30.3897− 86.54972
1451.52Gordon2000Tampa Bay to Cedar KeyFL28.5739− 82.65532
1451.52Arlene2005Walton CountyFL30.3486− 86.23893
1451.52Emily2005S Padre IslandTX26.1119− 97.16443
1451.52Rita2005Key WestFL24.5478− 81.78473
1451.52Fay2008Everglades CityFL25.8456− 81.38783
1651.51Alex2010Port LavacaTX28.6122− 96.62032
1661.46Unnamed1949MandevilleLA30.3528− 90.06893
1671.43TS No. 11956BiloxiMS30.3922− 88.88363
1671.43Bertha1957West end of Vermillion BayLA29.7451− 92.10753
1691.37Unnamed1901GalvestonTX29.2889− 94.78942
1691.37Unnamed1924Cedar KeyFL29.1397− 83.04192
1691.37Gladys1955Corpus Christi BayTX27.7997− 97.39112
1691.37Candy1968Palacios (Matagorda County)TX28.6983− 96.21513
1731.34Unnamed1912MobileAL30.6775− 88.03813
1741.28Humberto2007Texas Point TCOON Tide GageTX29.6781− 93.83692
1751.25Unnamed1941Everglades CityFL25.8456− 81.38783
1761.22Unnamed1899CarrabelleFL29.8439− 84.66442
1761.22Unnamed1902Central TX CoastTX28.0575− 96.85831
1761.22Unnamed1916MobileAL30.6775− 88.03813
1761.22Unnamed1923ApalachicolaFL29.7208− 84.98112
1761.22Unnamed1934Port O'Connor to FreeportTX28.7264− 95.69782
1761.22Unnamed1938Cameron and Vermillion ParishesLA29.5953− 92.63013
1761.22Unnamed1943GalvestonTX29.2889− 94.78943
1761.22Unnamed1947Galveston (entrance of Galveston Channel)TX29.3356− 94.77032
1761.22Unnamed1947Tampa; Anna Maria IslandFL27.5256− 82.73783
1761.22Debbie1957Apalachee BayFL30.1033− 84.04423
1761.22Ella1958Texas and Louisiana CoastsTX/LA29.6794− 93.83833
1761.22Irene1959Cedar KeyFL29.1397− 83.04193
1761.22Abby1964Matagorda to FreeportTX28.8044− 95.54832
1761.22Jeanne1980South Texas CoastTX26.4653− 97.24423
1761.22Floyd1987Lower and Middle KeysFL24.6481− 81.41642
1761.22Arlene1993South and Central Texas CoastTX27.2742− 97.34972
1761.22Gordon1994Upper FL KeysFL25.0108− 80.51392
1761.22Mitch1998Lower Florida KeysFL24.5572− 81.70923
1761.22Katrina2005Extreme SW FLFL25.1294− 81.03693
1761.22Dolly2008S Padre Is., Port Mansfield, BrownsvilleTX26.3533− 97.20943

7. Geographical surge patterns

After these storm surge events were identified, the latitude and longitude of the peak surge level was estimated for each event. These values were identified through several different methods. The values for tide gauges were taken directly from the agency or organisation that deployed the gauge and usually only required a quick conversion from degrees-minutes-seconds to decimal degrees. For cases in which the surge peaked at a specific location or community, such as the 1.78 m surge that peaked at Frenier, Louisiana, during Tropical Storm Matthew in 2004 (Avila, 2004), coordinates were taken from a location along the shoreline that appeared exposed to the prevailing flow of surge. If possible, coordinates were taken from the most inward portion of concave-shaped surfaces, such as bays or harbours, especially if the prevailing surge likely approached such features from a perpendicular direction. In some cases, the best available literature describes the location of peak surge as a portion of coastline, such as Hurricane Ivan's surge, which inundated the coastline from Mobile Bay, Alabama to Destin, Florida, with a surge as high as 4.57 m (Stewart, 2005). In these cases, coordinates were generally taken from a location near the middle of the zone of peak surge, particularly in an exposed bay or harbour, if possible.

These coordinates were then used to plot all 195 events in a Geographic Information System (GIS), depicted in Figure 1. The unique point plotted for each event represents the estimated location of peak surge. Cartographically, circles depict all surge events; larger, darker circles represent larger magnitude surges. Hurricanes Anita (1977) and Gilbert (1988) produced peak surges in Mexico, however, this map includes the peak US surges produced by these storms, both of which occurred at South Padre Island, Texas (The Brownsville Herald, Brownsville, Texas, Friday, 2 September 1977; National Hurricane Center, 1988).

Figure 1.

The location and height of the 195 peak storm surges along the US Gulf Coast identified in SURGEDAT

Although this map plots the estimated location of maximum storm surge, storm surges often inundate vast portions of coastline. For example, although Hurricane Ike produced a maximum storm surge of 5.33 m at Chambers County, Texas (Berg, 2009), a high-water profile constructed from selected storm surge observations (Berg, 2009; Needham and Keim, 2011, see Fig. 2) reveals that more than 2 m of storm surge inundated more than 200 km of coastline. Hurricane Ike's storm surge also penetrated inland considerable distances; a surge level of 0.65 m was reported at Lake Charles, Louisiana (Berg, 2009), approximately 50 km from the Gulf of Mexico. A potential avenue of future storm surge research could include constructing a storm surge database that documents the full extent of flooding along the coast.

Figure 2.

High water profile for Hurricane Ike in the Gulf of Mexico, created from selected storm surge observations provided by Berg (2009). Published in Needham and Keim (2011)

Although this map predominantly depicts natural variability in storm surge observations, artificial flood control devices, such as levees, likely enhanced observed surge levels in some areas. This is particularly likely in southern Louisiana, along the Mississippi River, where levees often impede surge moving from east to west, causing localized increases in surge levels on the eastern banks of the levees (Westerink et al., 2008).

In general, the greatest level of storm surge activity, in terms of both surge frequencies and magnitudes, is observed along the central and western Gulf Coast, from approximately Dune Allen Beach, Florida, westward along the Alabama, Mississippi, Louisiana and Texas Coasts (Figure 1). This portion of coastline observed 24 surges ≥ 4 m. The Florida Keys experienced 4 surges ≥ 4 m. Locations in the Florida Panhandle east of Dune Allen Beach, as well as the entire West Coast of Florida, depicted noticeably lower storm surge frequencies and magnitudes. No surge events ≥ 4 m were observed in this region. However, Weisberg and Zheng (2006b) concluded that Tampa Bay is still vulnerable to catastrophic surges under worse-case scenarios.

The largest magnitude events, those ≥ 5 m, typically occurred along the Texas, Louisiana and Mississippi Coasts, where 7 of 9 of these events occurred. Pass Christian, Mississippi, observed the two highest surges, Hurricane Katrina's (2005) 8.47 m surge, and Hurricane Camille's (1969) 7.5 m surge. These extraordinary events are not likely coincidental at this location. The shallow waters south of the Mississippi Coast, the concave shape of Bay St. Louis, Mississippi, and the proximal location of the deltaic lobes from the Mississippi River, likely contributed to enhanced surge levels in this area. Chen et al. (2008) hypothesize that Hurricane Katrina may have produced a maximum surge height 4 m lower had the storm instead tracked over the wide, sloping continental shelf located off the coast of Alabama.

A comparison of storm surge activity and hurricane strike frequency reveals that areas which receive few strikes from major hurricanes also observe few storm surges, although areas that observe frequent major hurricane strikes do not necessarily observe the highest magnitude storm surges. Keim et al. (2007) calculated the return period of major hurricane strikes from 1901 to 2005 for locations from Texas to Maine, including 19 locations along the Gulf Coast, determining that a portion of Central Florida, from Panama City to Cedar Key, including Apalachicola, is the least likely stretch of Gulf coastline to observe a major hurricane strike. As expected, this area also observed limited storm surge activity. However, locations most likely to observe strikes from major hurricanes, from Fort Myers to Key West, Florida, and secondarily, from Mobile, Alabama to Pensacola, Florida, did not observe the highest magnitude storm surges. Instead, the Texas, Louisiana and Mississippi Coasts, which observe fewer major hurricane strikes, generally observed greater storm surge magnitudes.

The lack of correlation between frequent major hurricane strikes and enhanced storm surge heights makes sense when one considers that many hurricanes impacting south Florida cross the state from east to west, pushing offshore into the Gulf of Mexico. The majority of hurricanes that impact Texas, Louisiana and Mississippi, approach from the south or southeast (Keim and Muller, 2009), providing ample time for these storms to generate high storm surges as they cross the Gulf of Mexico. Therefore, although south Florida observes more major hurricane strikes than locations in the western Gulf of Mexico, surge heights are generally lower. These observations reveal that the average amount of time hurricanes spend over open water before striking a location may relate closer to storm surge magnitudes than a mere count of hurricane strikes. This provides insight that storm surge climatology is not merely an extension of hurricane landfall climatology.

8. Conclusion

SURGEDAT, the first comprehensive storm surge database for the Gulf of Mexico, identifies the maximum storm surge height and location of 195 surges along the US Gulf Coast since 1880. This dataset incorporated 62 sources of information, including 28 Federal government sources, numerous academic publications, and more than 3000 pages of newspaper, from 16 periodical titles.

The central and western Gulf Coasts, including the Coasts of Texas, Louisiana, Mississippi, Alabama, and the western Florida Panhandle, as well as the Florida Keys, have observed the highest storm surge frequencies and magnitudes, while the eastern Florida Panhandle and the West Coast of Florida have historically observed reduced surge activity. Relationships between frequent major hurricane strikes and increased surge magnitudes do not correlate as well as one might expect. The Texas, Louisiana and Mississippi Coasts have observed 7 of 9 surges ≥ 5 m, but observe fewer major hurricane strikes than southwest Florida, Alabama and the western Florida Panhandle. This strongly suggests that storm surge climatology is not merely derived from the climatology of hurricane strikes.

SURGEDAT provides data that will likely reduce human and economic losses as this information will improve coastal zone planning, emergency management preparation, and public education. Coastal environmental projects in fields such as biology, hydrology, geomorphology, and geology, will also benefit from this research, as such projects often incorporate maximum observed storm surges, or maximum potential surge height in a given location. Future research on long-term climate change will likely benefit from this baseline of past surge activity, while these data will be useful for the storm surge modellng community to validate model runs.

SURGEDAT is available electronically at <http://surge.srcc.lsu.edu>, along with detailed metadata files that explain all sources that were used to determine storm surge heights and locations. Improvements will be made as more accurate data become available and more storms occur. It is recommended that users consult this website to ensure they are using the most recent version of this dataset.

Acknowledgements

This research was funded by NOAA Grants NA080AR4320886 and EA133E-07-CN-0084.

Ancillary