Monitoring Hurricane Rita Inland Storm Surge

By Benton D. McGee, Roland W. Tollett, and Burl B. Goree

Pressure transducers (sensors) are accurate, reliable, and cost-effective tools to measure and record the magnitude, extent, and timing of hurricane storm surge. Sensors record storm-surge peaks more accurately and reliably than do high-water marks. Data collected by sensors may be used in storm-surge models to estimate when, where, and to what degree storm- called high- surge flooding will occur during water marks, left behind future storm-surge events and to by flood waters. Identifying and calibrate and verify storm- qualifying high-water marks and surge models, resulting in a determining how well these marks better understanding of the represent the peak are often dynamics of storm surge. subjective. The quality of the high-water mark depends upon the type of mark, Introduction such as debris, seed, mud, or stain, and on Storm surge whether the mark was associated with created in a protected hurricanes occurs environment, such when winds push as the interior wall water up onto of a building, or the shoreline. in an unprotected Storm surge can environment, such be intensified by as an exposed bridge sustained winds, low piling or fence post. barometric pressure, High-water marks do excessive rainfall, and not record the date and high tides (http://geology. time of their creation nor com/articles/storm-surge. do they record the duration shtml, accessed September of the storm-surge event. 11, 2006). Historically, the magnitude of hurricane storm surge has been measured Methods by using water lines, In the days prior to the landfall of Hurricane Rita (fig. 1), an experimental water- level and barometric pressure sensor network (fig. 2) was deployed to record the magnitude, extent, and timing of inland storm surge and coastal flooding in the monitored area along the Louisiana and Texas Gulf Coast. East of Rita’s storm path, sensors recorded storm-surge water levels over 14 ft (4.3 258 Science and the Storms: the USGS Response to the Hurricanes of 2005

Arkansas Mississippi Alabama Louisiana

Sept. 24, 2005 Sept. 24, 2005 Monitored area 10 a.m. Florida

Texas Sept. 23, 2005 10 p.m.

Gulf of Mexico

Sept. 23, 2005 1 a.m. N Sept. 22, 2005 4 a.m. Figure 1. Map showing path of Hurricane Rita and study area (Hurricane Rita satellite imagery obtained from the National Aeronautics and Space Administration, 2006). Monitoring Hurricane Rita Inland Storm Surge 259

94º00’ 93º00’ 27 96 49 LOUISIANA LC2b 165 TEXAS (8.03) LA3 LC3 10 Beaumont LC4 LC2a (10.84) LA2 30º10’ Lake Charles 90 Lafayette Orange (8.93) (4.49) 10 LC6b LA8 B1 B19b 14 LA7 167 LC6a LC5 (3.86) (4.21) B20 Abbeville LF3 (6.93) 27 B10 B19a (5.04) Port Arthur 27 (9.52) LC8b LC7 Grand LF5 (11.15) B12 CalcasieuLake (7.38) Lake º LC12 LC9 Cameron (10.07) 29 50’ Sabine Lake B15b (7.52) LC11 (13.82) (13.34) (14.83) (14.68) LA9b LC13 82(14.90) LA9 Sabine Pass LC8a LA12 (10.73) (9.35) (10.62) LC10 LA11 82 White Lake LA10 (6.62) Intracoastal Gulf of Mexico City N (8.70) Vermilion Lake 0 10 20 Miles

0 10 20 Kilometers 29º30’

Land-surface elevation, in feet above the North Data-collection site and identifier, by sensor type America Vertical Datum of 1988 (NAVD 88) LC6b Barometric pressure

Greater than 30 LC5 Water level—Maximum surge elevation, in feet above NAVD 88, shown in parentheses

LC7 Barometric pressure and water level—Maximum surge elevation, in feet above NAVD 88, shown in parentheses

0 LC10 Sensor lost to hurricane

Figure 2. Map showing locations of storm-surge sensors in southwestern Louisiana and southeastern Texas. Maximum storm-surge elevation measurements are shown in yellow. 260 Science and the Storms: the USGS Response to the Hurricanes of 2005 m) above the North American Vertical Datum of 1988 (NAVD 88) along the Louisiana Gulf Coast and rates of water-level rise in excess of 5 ft (1.5 m) per hour. Comparisons between high-water marks and water-level data from nearby inundated sensors indicated that the sensors recorded storm-surge peaks more accurately and reliably than did high-water marks. Data collected by the storm-surge sensors may be used in computer models (1) to estimate when, where, and to what degree storm-surge flooding will occur in future events and (2) to calibrate and verify storm-surge models, resulting in a better understanding of the dynamics of storm surge. The study area was situated east of Rita’s storm path (fig. 1), in the right front quadrant of the storm, where the maximum storm surge was expected to occur. The monitored area covered approximately 4,000 mi2 (10,360 km2) and was generally bounded to the north by Interstate 10 (I-10), to the south by the Louisiana and Texas coastlines, to the east by the Vermilion River, and to the west by the cities of Beaumont, Orange, and Port Arthur, Tex. A total of 47 sensors (34 water-level sensors and 13 barometric pressure sensors) were deployed at 33 sites during September 22–23, 2005, prior to the landfall of Rita (fig. 2). Sensors were deployed along and inland of the Louisiana and Texas coasts, usually along waterways, from Sabine Pass, Tex., through Abbeville, La. The sensors were unvented pressure transducers capable of measuring and recording three parameters: absolute pressure, temperature, and internal battery voltage. Absolute Figure 3. Photograph showing sensor used to pressure is the force exerted by air (barometric pressure) record water level and barometric pressure. or water on the sensor. The force, or weight of water is converted to the height of water (water level) overlying the sensor. Inundated sensors were used to record storm- surge water levels, and noninundated sensors were used to record barometric pressure. Water-level data from inundated sensors were corrected for changes in barometric pressure by using data from a colocated barometric pressure sensor. If a barometric pressure sensor was not colocated with a storm- surge sensor, the data from the nearest barometric pressure sensor were used to correct the storm-surge data. The sensors recorded temperature and pressure every 30 seconds during the storm and for several days afterwards. The dimensions of the sensors were approximately 6 inches (15.2 cm) in length and 1 inch (2.5 cm) in diameter (fig. 3) and were encased in 1.5 inch (3.8 cm) by 18 inch (45.7 cm) metal pipes and strapped to permanent objects, such as piers and power poles (fig. 4). Storm-surge sensors were deployed on permanent structures at elevations that would likely be inundated by storm surge, and barometric pressure sensors Figure 4. Photograph showing metal pipe containing a water- were deployed on permanent structures at elevations that level sensor or barometric pressure sensor strapped to a power would not likely be inundated by storm surge. pole. Water-level data from inundated sensors also were corrected for salinity because salinity content increases the density, and therefore the weight, of the water. Corrections for salinity were based upon the location of the sensor in proximity to the coast. In general, sensors located in the southern part of the study area were categorized as measuring Monitoring Hurricane Rita Inland Storm Surge 261 salt water, sensors located in the northern part of the study area variability between individual sensors. Sensors were deployed were categorized as measuring freshwater, and sensors located at two existing USGS stream gages (one each near sites LA2 in the middle were categorized as measuring brackish water. and LA8) to assess the variability of water levels measured at Elevation surveys were conducted to relate all storm- sensors and USGS stream gages. Water-level data collected surge data, including water-level data from inundated sensors, by multiple sensors agreed closely with each other (fig. 5), high-water marks, and water-surface measurements, to sea and sensors agreed closely with water-level data from USGS level (sea level defined by the NAVD 88) (http://www.ngs. stream gages (fig. 6), as evidenced by the overlying plots noaa.gov/PUBS_LIB/NAVD88/navd88report.htm, accessed of the sensor and USGS stream gage data during periods of September 11, 2006). Elevation surveys were conducted by sensor inundation. using the Global Positioning System and differential levels. Water-surface levels, commonly referred to as “tape Quality assurance measures included (1) placing multiple downs,” were determined by measuring the distance from a sensors at a site, (2) placing a sensor near an existing U.S. reference mark, such as a bridge railing or pier surface, to the Geological Survey (USGS) stream gage, and (3) manually water surface at a specific date and time. Tape downs were measuring water-surface levels from reference marks. made at seven sites, namely LA2 (fig. 6), LA3, LA9, LC2a, Multiple storm-surge sensors were deployed at two sites (three LF3 (fig. 5), LC5, and LC8b, and all agreed closely with sensors at site LF3 and two sensors at site LA7) to assess the sensor data.

15.0 16 15.0 16 15 15 14.8 14.8 Barometric pressure 14 Barometric pressure 14 13 13 14.6 14.6 12 12 11 14.4 11 14.4 10 10 Tape-down elevation 14.2 9 14.2 9 8 8 Water-level elevation—3 water-level 14.0 14.0 sensors placed at different elevations 7 7 Water-level elevation—Water-level sensor track together during surge inundation 6 placed near USGS coastal streamgage 08012150 6 13.8 Tape-down elevation 13.8 Sensor 1 track together during surge inundation 5 5 13.6 4 13.6 4 Lowest recordable elevation 3 3 13.4 USGS coastal 13.4 2 2 streamgage Water-level elevation, in feet above NAVD 88 elevation, in feet above NAVD Water-level Barometric pressure, in pounds per square inch

Water-level elevation, in feet above NAVD 88 elevation, in feet above NAVD Water-level Sensor 3 Barometric pressure, in pounds per square inch 1 1 13.2 Sensor 13.2 Sensor 2 Lowest recordable elevation 0 0 -1 13.0 -1 13.0

Sept. 23, 2005 Sept. 24, 2005 Sept. 25, 2005 Sept. 26, 2005 Sept. 27, 2005 Sept. 28, 2005 Sept. 29, 2005 Sept. 23, 2005 Sept. 24, 2005 Sept. 25, 2005 Sept. 26, 2005 Sept. 27, 2005 Sept. 28, 2005 Sept. 29, 2005

Figure 5. Hurricane Rita storm-surge data at site LF3a. The Figure 6. Hurricane Rita storm-surge data at site LA2. The graph graph shows water-level elevations above North American shows water-level elevations above North American Vertical Vertical Datum of 1988 (NAVD 88) from multiple storm-surge Datum of 1988 (NAVD 88) from storm-surge sensor and U.S. sensors and tape-down measurement. Geological Survey stream gage and tape-down measurement. 262 Science and the Storms: the USGS Response to the Hurricanes of 2005

Results Conclusion

High-water marks were identified and surveyed at seven Sensors recorded storm-surge water levels over 14 ft sites, namely LA9b (fig. 7), LA12, LC2b, LC5, LC7, LC8a above NAVD 88 at Constance Beach (LC11), Creole (LA12), (fig. 8), and LF5, and were compared with nearby sensor data. and Grand Chenier (LA11), La., about 20 mi (32 km), 48 mi Comparisons between high-water marks and storm-surge (77 km), and 54 mi (87 km), respectively, east of Sabine Pass, peaks from inundated sensors varied. In general, high-water Tex., at approximately 2 a.m., September 24, 2005 (fig. 2). marks of high quality agreed closely with the storm-surge In general, storm-surge water levels increased eastward from peaks from the sensors, while high-water marks of lesser the Sabine River into southwest Louisiana. The magnitude quality were consistently lower than the storm-surge peaks of the storm surge was greatest near the coast and decreased from the sensors. For example, the high-water mark near site inland through the approximate latitude of I-10, about 35 mi LA9b (fig. 7) rated as “excellent” was approximately 0.2 ft (56 km) inland from the coast (fig. 2). Sensors reported rates (6.1 cm) lower than the storm-surge peak from the nearby of water-level rise during the storm-surge event in excess of storm-surge sensor, the high-water mark near site LC8a (fig. 5 ft (1.5 m) per hour at sites LC8b, LC9, and LC13. By using 8) rated as “good” was approximately 1 ft (30.5 cm) lower data from the sensors and digital land-surface elevation data, than the storm-surge peak from the nearby storm-surge sensor, a computer-generated map of storm-surge depth (fig. 9) in the and the high-water mark near site LF5 rated as “poor” was monitored area was created and indicated over 13 ft (3.9 m) of approximately 1.9 ft (57.9 cm) lower than the storm-surge water depth at 3 a.m. on September 24, 2005, along parts of peak from the nearby storm-surge sensor. the Gulf Coast of southwestern Louisiana (Dean Gesch, U.S. Geological Survey, written commun., 2006).

16 15.0 16 15.0 15 15 14.8 14.8 14 Barometric pressure 14 Barometric pressure 13 13 High-water mark 14.6 High-water-mark elevation 14.6 12 12

VD 88 High-water mark 11 High-water-mark elevation 14.4 11 14.4 VD 88 10 10 14.2 9 14.2 9 8 Water-level elevation 8 14.0 14.0 7 7 Water-level elevation 6 6 13.8 13.8 5 5 -level elevation, in feet above NA 13.6 4 Lowest recordable elevation 13.6 4 ater -level elevation, in feet above NA Lowest recordable elevation W 3 Barometric pressure, in pounds per square inch 3 ater 13.4

13.4 W 2 2 Barometric pressure, in pounds per square inch 1 1 13.2 13.2 0 0 13.0 -1 13.0 -1

Figure 7. Hurricane Rita storm-surge data at site LA9b. The Figure 8. Hurricane Rita storm-surge data at site LC8a. The graph shows water-level elevations above North American graph shows water-level elevations above North American Vertical Datum of 1988 (NAVD 88) from storm-surge sensor and Vertical Datum of 1988 (NAVD 88) from storm-surge sensor and high-water mark of excellent quality. high-water mark of good quality. Monitoring Hurricane Rita Inland Storm Surge 263

94º00’ 93º00’ 27 96 49 LOUISIANA LC2b 165 TEXAS LA3 LC3 10 Beaumont LC4 LC2a LA2 º Lake Charles 90 Lafayette 30 10’ Orange 10 LC6b LA8 B1 B19b 14 LA7 167 LC6a LC5 B20 Abbeville LF3 27 B10 B19a Port Arthur 27 LC8b LC7 Grand LF5 B12 CalcasieuLake Lake º LC12 LC9 Cameron 29 50’ Sabine Lake B15b LC11 LA9 LA9b Sabine Pass LC13 82 LC8a LA12 LA11 82 White Lake Orange LA10 Intracoastal N City Vermilion 0 10 20 Miles Lake Gulf of Mexico 0 10 20 Kilometers 29º30’ Storm-surge depth, in feet Land-surface elevation, in feet above the the North America Vertical Datum of 1988 (NAVD 88) Greater than 13 Greater than 30 12 11 10 9 8 0 7 6 Data-collection site and identifier, by sensor type 5 4 LC6b Barometric pressure 3 LC5 Water level 2 LC7 Barometric pressure and water level 1 0

Figure 9. Map showing locations of storm-surge sensors and computer-generated storm-surge depth, in feet, on September 24, 2005, at 3 a.m. in southwestern Louisiana and southeastern Texas.

Contact Information

Benton D. McGee, Supervisory Hydrologist ([email protected]); Roland W. Tollett, Hydrologist ([email protected]); and Burl B. Goree, Hydrologic Technician ([email protected]) U.S. Department of the Interior U.S. Geological Survey Louisiana Water Science Center 3535 S. Sherwood Forest Blvd., Suite 120 Baton Rouge, LA 70816