Skip to content U.S. Department of Transportation/Federal Highway AdministrationU.S. Department of Transportation/Federal Highway Administration

Office of Planning, Environment, & Realty (HEP)
PlanningEnvironmentReal Estate

HEP Events Guidance Publications Awards Contacts

Climate Variability and Change in Mobile, Alabama

Appendix D: Additional Information on the Sea Level Rise and Storm Surge Analyses

D.1. Factors Not Considered in Sea Level Rise Analysis

Several factors that can affect local sea level rise were not considered in this study because they were not considered to likely significantly impact the results, or due to resources constraints. These factors are described in this appendix and include:

Modeling coastal sedimentation and erosion is quite complex and outside the scope of this study, but might be considered in future studies. Sedimentation and erosion are not major factors for parts of the coastline that are naturally rocky or artificially hardened, such as the port and downtown areas of Mobile. However, sedimentation and erosion may be important for parts of the coastline that are composed of soft sediments, such as sandy beaches and marshes. These vulnerable conditions do exist in large parts of the study area. In some places, vertical accretion of marshes is able to keep pace with the rate of sea level rise. However, it is unclear whether marshes will be able to keep pace with the scenarios of 0.75 and 2.0 meters GSLR explored in this study. In addition, sedimentation and erosion, particularly following major storms, could lead to changes in the coastal landscape in some parts of the Mobile region. For example, barrier islands such as Dauphin Island are particularly vulnerable to storm damage and storm-induced erosional and depositional processes, which may dramatically change their topography.

Gravitational changes from ice sheet loss could affect sea level rise in Mobile, however, they were not considered in this study. A large reduction in the mass of the Greenland and Antarctic ice sheets would affect regional sea level around the world. Regions close to where the ice sheets have shrunk will experience a reduction in sea level due to a reduction in the gravitational attraction from the mass of the ice sheet. In contrast, the regions farthest away from the shrinking ice sheets will tend to experience an increase in regional sea level. Mitrovica et al. estimate that if the West Antarctic Ice Sheet were to collapse, local sea level rise along the U.S. coast would be roughly 20 to 25% greater than global sea level rise due to gravitational changes. However, gravitational changes from ice sheet loss were not considered in this study because they are considered a second-order and smaller effect compared to global sea level rise from ice sheet melting. Furthermore, the understanding of the interplay between gravitational effects from the Antarctic and Greenland ice sheets is highly uncertain.

Changes in circulation patterns and ocean density were also not considered. These factors could potentially affect sea level. A decrease in ocean density due to warming or an increase in salinity will tend to reduce regional sea level. Changes in wind-patterns can lead to local and regional differences in sea level. However, the long-term influence of ocean circulation and density changes is likely to be near zero in the central Gulf region. Therefore, these factors were not examined in this study.

D.2. Methodology to Estimate Subsidence and Uplift

This appendix details the methodology used to estimate subsidence (downward land surface motion) and uplift (upward land surface motion) as part of the projected sea level rise analysis. USGS estimated subsidence and uplift rates using Interferometric Synthetic Aperture Radar (InSAR) data together with a series of stable survey benchmarks and tide gages. In summary, the InSAR data provided vertical movement data for most of the study area, while the benchmark data helped to augment the InSAR data outside the spatial domain of the InSAR data.1 InSAR data were used where possible, because they are spatially continuous and possess relatively high accuracy. A spatially complete data set of vertical motion from these two datasets was arithmetically added to a high resolution Digital Elevation Model based on LIDAR data2 to estimate the vertical position of the land surface out to 2050 and 2100.

D.2.1. Interferometric Synthetic Aperture Radar Data

InSAR data provided vertical movement data for most of the study area. Land surface deformation maps, showing rates of uplift and subsidence, were developed for the Mobile region using analysis of satellite-derived multi-temporal Interferometric Synthetic Aperture Radar (InSAR) data. The Temporarily Coherent Point (TCP) InSAR (TCP-InSAR) methodology developed by Zhang and Lu (2010a,b) was utilized. In brief, high precision maps of the land elevation at two different points in time were compared to determine the rate of vertical motion. USGS used Spaceborne SAR data from the European ERS-1/2 satellites3 during 1992 and 1999 and the Japanese ALOS/PALSAR4 during 2007 and 2010 for this purpose. USGS concluded that the ERS-1/2 deformation results were more reliable than those from the ALOS/PALSAR images due to the persistent orbital errors in the ALOS InSAR images, even after taking into account a correction procedure to reduce the orbit errors on the ALOS/PALSAR deformation results.

D.2.2. Benchmark Data

Benchmark data helped to augment the InSAR data outside the spatial domain of the InSAR data. The Mississippi Department of Transportation (MDOT), National Geodetic Survey (NGS), and contracted surveyors have established a network of precise elevations of stable benchmarks from the Florida panhandle, through Alabama and Mississippi, to Louisiana. The partners include MDOT, Alabama Department of Transportation (ALDOT), Florida Department of Environmental Protection (FLDEP), and many others. The finalized elevations of this benchmark network are not yet available, but the 2009-10 field adjusted elevations were made available by NGS for use in this study. USGS compared vertical change at tide gages, benchmarks, and Continuously Operating Reference Stations (CORS) to vertical change contours by Holandahl and Morrison (1974) in Mobile and Baldwin Counties, Alabama. Vertical change rates reported by Shinkle and Dokka (2004) were considered. In all, vertical change rates for 75 benchmarks were provided by USGS based on surveys from 1969 compared to 2009 or 1984 compared to 2010, depending upon location. In addition to uncertainties associated with the benchmark survey measurements, the usefulness of both the benchmark and CORS data are somewhat limited due to their spatial incompleteness.

D.3. Vertical Land Surface Rates

This appendix provides the vertical land surface rates in millimeters per year for the benchmark surveys and corresponding InSAR Data in Mobile and Baldwin Counties.

Table 46: Vertical Land Surface Rates (mm/yr) for Benchmark (BM) Surveys and Corresponding InSAR Data in Mobile and Baldwin Counties

These are preliminary results from an analysis by K. Van Wilson, USGS, which were provided in draft reports to FHWA and the ICF project team.

ID Code Latitude5 Longitude BM Survey6 ERS 7 PALSAR8
BG0094 30.522861 -87.493417 -0.8 1.0 -2.2
BG0098 30.517778 -87.480583 -0.7 0.7 -1.5
BG0099 30.518111 -87.463750 -0.4 0.6 -2.8
BG0100 30.521917 -87.453389 -0.4 1.1 ND
BG2485 30.527056 -87.512583 -0.4 1.7 -2.5
BG2487 30.531389 -87.519111 -0.3 -0.1 -1.8
BG2495 30.542167 -87.568583 -0.4 0.1 -2.8
BG2500 30.559528 -87.593417 0.0 0.2 -0.2
BG2506 30.567694 -87.633889 -0.6 0.7 -0.5
BG2508 30.567389 -87.661778 -0.7 0.7 -0.8
BG2512 30.563639 -87.682917 -0.6 0.2 -0.7
BG2513 30.564000 -87.698750 -0.6 1.1 -1.9
BG2516 30.562611 -87.713667 -0.6 1.5 1.1
BG2517 30.566944 -87.716389 -0.6 1.1 -0.4
BG2521 30.603611 -87.738194 -0.7 0.4 -0.2
BG2528 30.623444 -87.750750 -0.6 0.4 0.2
BG2532 30.632776 -87.759193 -0.9 1.3 -1.1
BG2534 30.632889 -87.790750 -0.9 1.4 1.3
BG2536 30.646389 -87.814167 -0.8 1.2 1.4
BG2537 30.647944 -87.828583 -0.8 1.4 2.5
BG2538 30.654944 -87.844583 -1.0 1.6 1.7
BG2540 30.673222 -87.852333 -1.2 1.3 1.5
BG2542 30.676667 -87.866944 -1.1 1.3 1.8
BG2544 30.671722 -87.883306 -1.3 1.0 -0.6
BG2546 30.670833 -87.897778 -1.0 1.8 1.1
BG2556 30.676750 -87.977444 -4.1 0.8 -1.6
BH0144 30.682667 -88.007000 -0.9 1.1 -1.0
BH0145 30.680750 -88.001028 -1.5 1.4 -1.8
BH0116 30.523333 -88.206667 -0.7 0.0 -4.0
BH0117 30.524417 -88.202722 -0.8 -0.4 -2.2
BH0120 30.546472 -88.174028 4.4 -0.6 0.1
BH0239 30.474528 -88.351167 -0.7 ND ND
BH0240 30.476306 -88.342278 -0.8 ND ND
BH0243 30.486944 -88.310278 -0.9 ND ND
BH0246 30.493803 -88.283297 -1.1 -0.1 ND
BH0250 30.498889 -88.266667 -0.8 0.1 ND
BH1465 30.666028 -88.049861 -0.8 -0.3 -2.5
BH1466 30.669389 -88.045000 -0.5 -0.2 0.4
BH1468 30.682028 -88.040889 -0.5 0.1 -1.2
BH14699 30.690083 -88.041167 -0.8 0.2 0.2
BH159859 30.692944 -88.031028 -1.9 0.0 -1.6
BH159959 30.692833 -88.031917 -1.7 0.6 -1.5
BH1602 30.675750 -88.042222 -0.4 0.7 -1.2
BH1605 30.632056 -88.105194 -0.7 0.6 1.5
BH0117 30.524417 -88.202722 -0.8 -0.4 -2.2
BH0137 30.545000 -88.171111 -0.8 -0.2 -1.1
BH0138 30.531667 -88.170556 -0.8 -0.4 -2.4
BH0251 30.490556 -88.169722 -0.7 0.0 -2.3
BH0256 30.446389 -88.166389 0.5 0.2 -1.8
BH0259 30.403611 -88.148611 0.0 -0.6 -2.9
BH1722 30.546806 -88.173750 -0.8 -0.5 0.5
BH1723 30.501667 -88.169722 -0.6 0.1 -2.9
BH1724 30.485556 -88.169444 -0.6 0.0 -2.7
BH1725 30.485556 -88.169444 -0.7 0.0 -2.7
BH1726 30.462500 -88.168889 -0.6 -0.6 -4.5
BH1733 30.432500 -88.160833 -0.4 -0.2 -2.6
BH1734 30.377222 -88.159722 0.5 0.2 -0.3
BH1735 30.371111 -88.145833 0.0 0.3 -1.0
BH1736 30.350942 -88.121363 -0.4 0.1 -0.1
BH1737 30.338056 -88.128889 -0.3 -0.6 -0.1
BH1740 30.310278 -88.137778 -0.7 1.7 -1.7
BH1741 30.299167 -88.133889 -0.9 0.2 -2.5
BH1742 30.289135 -88.128657 -1.9 -0.8 -3.1
BH1743 30.290000 -88.128889 -1.9 -0.7 -3.0
BH1744 30.277778 -88.122500 -0.5 -1.7 -3.8
BH1745 30.265000 -88.115833 -0.9 ND -4.4
BH1748 30.254167 -88.111667 -0.9 ND -5.0
BH1749 30.253611 -88.112500 -0.8 ND -4.4
BH1750 30.251111 -88.095000 -0.8 ND -5.3
BH175110 30.249722 -88.076389 -0.8 ND -5.0
BH175260 30.249444 -88.076667 -1.0 ND -5.0
BH175560 30.249672 -88.076022 -1.0 ND -4.7
BH175660 30.249582 -88.075489 -0.8 ND -5.9
BH175760 30.249167 -88.075556 -0.8 ND -5.5
BH175860 30.248806 -88.074944 -0.8 ND -4.7
BH176060 30.247900 -88.075283 -1.1 ND -4.5
BH176160 30.248028 -88.076278 -0.8 ND -4.8

D.4. Supplementary Sea level Rise Exposure Statistics

This appendix provides supplementary exposure statistics for the scenario-based analysis of future sea level rise.

Table 47: Supplementary Sea Level Rise Exposure Statistics

Scenario
Transportation Mode Asset Criticality 2050 - 30 cm 2100 - 75 cm 2100 - 200 cm
Highways Roads (miles) Critical 9 of 209 (4%) 11 of 209 (5%) 26 of 209 (13%)
Not Critical 0 of 284 (0%) 0 of 284 (0%) 3 of 284 (1%)
Evacuation Routes (miles) Not Critical 5 of 367 (1%) 7 of 367 (2%) 21 of 367 (6%)
Rail Rail (miles) Critical 2 of 196 (1%) 2 of 196 (2%) 40 of 196 (20%)
Not Critical 0 of 118 (0%) 0 of 118 (0%) 24 of 118 (21%)
Rail Points (#) Critical 0 of 5 (0%) 0 of 5 (0%) 2 of 5 (40%)
Not Critical 4 of 12 (33%) 4 of 12 (33%) 6 of 12 (50%)
Pipelines Pipelines (miles) Critical 3 of 426 (1%) 7 of 426 (2%) 13 of 426 (3%)
Not Critical 2 of 226 (1%) 4 of 226 (2%) 13 of 226 (6%)
Ports Ports (#) Critical 0 of 26 (0%) 0 of 26 (0%) 4 of 26 (15%)
Not Critical 3 of 48 (6%) 6 of 48 (13%) 18 of 48 (38%)
Transit Facilities Critical 0 of 2 (0%) 0 of 2 (0%) 1 of 2 (50%)
SDE Facilities (#) Not Critical 0 of 193 (0%) 0 of 193 (0%) 10 of 193 (5%)
Bus Stops (#) Not Critical 0 of 907 (0%) 0 of 907 (0%) 15 of 907 (2%)
Bus Routes (miles) Not Critical 0 of 126 (0%) 0 of 126 (0%) 5 of 126 (4%)
MODA Stops (#) Not Critical 0 of 22 (0%) 0 of 22 (0%) 0 of 22 (0%)
Bike Routes (miles) Not Critical 1 of 132 (1%) 2 of 132 (2%) 13 of 132 (10%)
Airports Mobile Downtown Airport (mi2) Critical 0 of 3 (1%) 0 of 3 (2%) 0 of 3 (3%)
Mobile Regional Airport (mi2) Critical 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%)
Other Medical Facilities (#) Not Critical 2 of 45 (4%) 3 of 45 (7%) 5 of 45 (11%)

D.5. Caveats, Gaps, and Replicability of Sea Level Rise Analysis

This appendix discusses the assumptions and simplifications used in this study's analysis of future sea level rise. These assumptions and simplifications should be taken into account when considering the results of the analysis or applying the methodology elsewhere.

First, this analysis assumes that future subsidence and uplift rates will remain constant over the next century. Since there are currently no robust predictions of changes in the soft sediment dynamics and tectonics that define the spatial heterogeneity of vertical change in the region, this assumption of constant vertical change is valid for this first-order analysis of potential inundation.

Second, the analysis does not take into account vertical addition or subtraction of sediment through coastal engineering, nor does it account for changes in the vertical accretion rate of wetlands.

The approach to sea level rise mapping used here is appropriate for initial exposure assessment. To more rigorously assess the exposure of the region's transportation to changes in land forms due to soft sediment dynamics, a sediment transport and erosion model (e.g., XBeach11) should be deployed. To more rigorously assess the impact of changes in wetland elevation and distribution, a model such as SLAMM (Sea Level Affecting Marshes Model12) could be used.

Third, the potential inundation analyses presented here do not account for small-scale protective barriers (e.g., sea walls and pumping systems) that are intended to prevent long-term flooding. These systems are generally only of very localized importance in the Mobile area. In areas where they are more prevalent and critical at the regional level as a whole (e.g., New Orleans), particularly where the land surface is at or below sea level, the SLAMM model can be used to account for them, provided that survey data or as-built information about these engineering structures is available.

Fourth, as noted previously, the LSLR scenarios used in this study do not account for changes in oceanic or atmospheric circulation or ocean density since these factors are likely to be of secondary importance in this region. In locations such as southern California, where changes in ocean circulation have had a strong influence on LSLR,13 it would be more important to take these factors into account. Information about the regional net importance of these factors can be obtained from the periodic Intergovernmental Panel on Climate Change assessments14 as well as the National Climate Assessment.15

Finally, due to the large uncertainties associated with future GSLR, it is not essential to use high resolution uplift and subsidence data (e.g., from InSAR) for initial exposure screening, as was done in this study. Coarse resolution analyses of local contributions to SLR can be obtained where tide gauge data are available for a few decades by subtracting out the estimated GSLR. This would be suitable for an initial assessment evaluating which assets are exposed to sea level rise. However, it would be useful for InSAR data to be available for the entire U.S. coastline to contribute to detailed vulnerability assessment and adaptation planning beyond the Mobile area.

D.6. Detailed Case Studies

This appendix provides more detailed summaries of the five case studies described in Section 2.8.1. These storms represent a sampling of the different types of storms that Mobile experiences, including a thunderstorm and tornado event, a hailstorm, a heavy rain event, a moderate hurricane (Georges), and an intense hurricane (Katrina).

The extra-tropical case studies include:

The hurricane case studies include:

D.6.1. Study 1: Severe Thunderstorms and Tornado Outbreak, November 15, 2006

Storm Development

Severe thunderstorms strong enough to produce six tornadoes struck the Mobile region on November 15, 2006.16 These severe thunderstorms developed due to a strong southerly jet stream aloft that steered a low pressure system at the surface into Alabama. Figure 97, on the left, is a surface map of the southeastern United States for November 15, 2006 at 6:00 am (CST) showing the key features for the storm's development.17

As illustrated by the blue markers in Figure 97, the wind ahead of the cold front was from the south bringing warm moist air from the Gulf waters into Alabama. Warm moist surface air feeds thunderstorm development by encouraging convective activity.18 Also noticeable in Figure 97 is the north-to-south gradient in the upper-level jet stream above Alabama, which is ideal for intensifying the surface low pressure system.

Figure 97: Surface Maps of the November 15, 2006 Storm at 6:00 am and 6:00 pm

This figure shows surface maps of the Mobile region showing atmospheric pressure surrounding the November 15, 2006 severe thunderstorms. At 6:00 am, before the storms, there is a low pressure system visible to the west of Mobile and moving toward the city. A low pressure system is stationed above the region. At 6:00 pm, a strong jet stream is visible.This figure shows surface maps of the Mobile region showing atmospheric pressure surrounding the November 15, 2006 severe thunderstorms. At 6:00 am, before the storms, there is a low pressure system visible to the west of Mobile and moving toward the city. A low pressure system is stationed above the region. At 6:00 pm, a strong jet stream is visible.

The key meteorological conditions for this storm development were: (1) a strong jet stream aloft, (2) a surface cold front associated with a low pressure system, and (3) warm moist surface air. As detailed earlier, this is a typical example of a severe storm event in Mobile, Alabama.

The surface low pressure system is marked with an "L" and the associated cold front is illustrated by the thick blue line with blue triangles. The black contour lines indicate the surface pressure in millibars (mb), and the blue lines attached to the blue circles illustrate wind direction and strength. The right figure is a 500 mb pressure map (500 millibars is about halfway up the lower atmosphere) also for November 15, 2006 at 6 pm (CST) illustrating the strong jet stream aloft, shown by the blue lines with triangles, the tightly spaced black atmospheric pressure lines, and the 'dip' in the black atmospheric pressure lines below Louisiana and Arkansas

Storm Event Data

Precipitation: 5.7 inches (14.3 centimeters) of rain fell at Mobile Regional Airport on November 15, 2006 occurring between 5:00 am and 3:00 pm (CDT).19 Rainfall rates peaked at over 1.5 inches per hour (3.8 centimeters per hour) at 8:00 am and 11:00 am. The total rainfall from this storm event is equivalent to the total average November rainfall in Mobile, Alabama.

Radar-estimated precipitation amounts for the region are shown in Figure 98. This figure illustrates how precipitation varied across the region and shows the hotspots of heavy rainfall for southern Alabama and Mississippi. The precipitation estimates appear to be lower than the measured precipitation at Mobile Regional Airport, but radar-derived precipitation estimates may not be as accurate as the observation data.

Figure 98 : Radar-Derived Precipitation Estimates for the Gulf Region, November 15, 2006, 12:00 pm

(Source: NWS, 2009a)

This image shows radar-derived precipitation estimates from the November 15, 2006 storm. A band of particularly heavy rainfall is visible stretching from Biloxi, MS, over Mobile, and into central Alabama.

Streamflow: Three USGS streamflow gages in Mobile County showed a marked increase in stream discharge on November 15, 2006. The hourly measured discharge increased between 2- and 40-fold compared to the long-term average for that day.20

Temperature: The temperature on November 15, 2006 was considerably warmer than the days surrounding it, with a high of 73°F (23°C) and a low of 52°F (11°C) (compared to 66°F (19°C) and 40°F (4°C), respectively).21 This is likely due to the warm moist Gulf air traveling into Alabama.

Wind: The highest observed wind gust recorded in Mobile during the storm was 52 miles per hour (83 kilometers per hour), recorded at 7:42 am on November 15, 2006. The fastest 2-minute sustained wind speed observed was 41 miles per hour (66 kilometers per hour). The average wind speed for the day was 13.6 miles per hour (21.8 kilometers per hour), compared to an average value for the surrounding week of 6.1 miles per hour (9.8 kilometers per hour). If the day of the storm is not included in this average then the average wind speed for the week was 5.2 miles per hour (8.3 kilometers per hour).22

Table 48: Summary of Peak Storm Event Data for November 15, 2006 for the Mobile, Alabama Region (*11/16)

Variable Value at Peak Storm Intensity Value averaged across surrounding days
Maximum Surface Temperature 73°F 66°F
Minimum Surface Temperature 52°F 44°F
Precipitation Total 5.7 in (> 2 year event) -
Peak Precipitation Rate 1.87 in/hr -
Hourly Streamflow Discharge
Crooked Creek 922 ft3/s( > 2 year event) 10 ft3/s
Chickasaw Creek* 5,660 ft3/s( > 2 year event) 120 ft3/s
Fowl River* 85 ft3/s(< 2 year event) 25 ft3/s
Wind Gust 52 mph -
Sustained Wind 41 mph -
Surface Pressure 992 mb 1008 mb
Tidal Datum
Dauphin Island (MLLW) 2.07 ft (< 10 year event) -
Dauphin Island (MHHW) 0.70 ft (< 10 year event) -
Pensacola (MLLW) 1.99 ft (< 10 year event) -
Pensacola (MHHW) 0.50 ft (< 10 year event) -

Surface Pressure: The average surface pressure in Mobile on November 15 was observed at 992 millibars, representing a drop of about 16 millibars compared to the surrounding days.23

Storm Highlights

  • 6 thunderstorms produced F-0 to F-2 tornadoes
  • Severe wind gusts at or above 57 mph (91 kph)
  • The NWS Office in Mobile Alabama issued five flash warnings prompted by the extremely heavy rainfall of up to 4 to 8 inches (10 to 20 centimeters) across the Mobile region

Water Level: Dauphin Island tidal station recorded peak water levels at 12:45 pm on November 15 of 2.07 feet (63.09 centimeters) above the mean lower-low level,24 1.5 feet (45.7 centimeters) higher than the expected water level of 0.57 feet (17.37 centimeters). At 3 pm on November 15, the Dauphin Island tidal station measured peak water levels of 0.70 feet (21.34 centimeters) above the mean higher-high level. 25

The tidal station in Pensacola FL recorded similar water level disturbances, with 2 feet (61 centimeters) of water above the mean lower-low level at 3:30 pm, compared to the expected water height of 0.5 feet (15 centimeters).26 The records also indicate 0.74 feet (22.56 centimeters) of water above the mean higher-high level at 3:30 pm, compared to the expected water height of -0.76 feet (-23.16 centimeters).

Storm Damage

Strong winds and tornadoes caused the majority of storm damage. This strong storm spawned six tornadoes near Mobile (see Figure 15).27 In addition, the storm caused several extreme straight-line wind gusts. During the morning of November 15, four extreme wind gusts were reported in Mobile of 52, 43, 38, and 35 miles per hour (83, 69, 61, and 56 kilometers per hour).28 The storm caused blocked roadways from debris and fallen trees and power lines. The NWS estimates that the storm's six tornadoes caused between $0.5 million and $1 million of damage.29

In addition to wind damage, transportation infrastructure was impacted by flooding caused by very heavy rain. Two heavy hour-long episodes accounted for over sixty percent of the rain. These heavy morning downpours caused flooding in many roads, streets, creeks, and streams in Mobile.

Figure 99: November 15, 2006 Storm Reports

Source: NWS, 2009a

This figure shows a map of the United States, with dots for where tornadoes and strong winds were reported on November 15, 2006. There are several dots indicating wind reports around the Mobile area and southeastern Alabama, along with a handful of tornado reports.

Figure 100: November 15, 2006 Tornado Tracks and Intensities

Source, NWS, 2009a. See Glossary for definition of tornado intensity scale (Fujuita Scale, F1, F0)

This figure shows where tornadoes occurred in Alabama on November 5, 2006. There were no tornadoes in Mobile County, but two in northern Baldwin County and one directly north of Mobile County.

D.6.2. Study 2: Severe Hailstorm, March 5, 1998

Storm Development

Thirteen severe thunderstorms developed in the Mobile region on March 5, 1998. The storms brought hail to the Mississippi/Alabama region, with hailstones ranging between the size of a dime to the size of a baseball.30 A number of key meteorological conditions led to the development of this storm, including: a strong west-to-east jet stream aloft, the existence of cold, dry air in the middle layer of the atmosphere, vertical wind shear, and strong potential for convective thunderstorms.31 In addition, a high pressure system over Florida brought warm moist air into Alabama, an important ingredient for severe thunderstorms in Alabama.

Figure 101: Surface Weather Map for March 5, 1998, 6:00 am

Source: NOAA, 1998

This figure shows a surface weather map of the United States for March 5, 1998, when there was a severe hailstorm in the Mobile area. A high pressure system over Florida is visible, which pushed air into Alabama

Figure 102: Zonal (West to East) Flow Diagram in Atmosphere

Source: NWS, 2009b

This figure shows a map of the southeastern United States for March 5, 1998, when there was a severe hailstorm in the Mobile area. In the figure, there is visible moisture converging above Alabama.

Storm Event Data

Precipitation: In total, 0.64 inches (1.63 centimeters) of rain were observed at Mobile Regional Airport on March 5, 1998.32 The majority of rain fell between 12:00 pm and 8:00 pm, peaking at 12:00 pm at a rate of 0.29 inches per hour (0.74 centimeters per hour).33 Though the precipitation amounts were minimal, hail stones ranging from 0.5 to 2 inches (1.3 to 5 centimeters) in diameter were reported in the area, with hail approaching 3 inches (8 centimeters) in diameter reported northwest of Leakesville, Mississippi.34 Hail accumulation in Leakesville, where the most severe hail damage occurred, was about 6 to 12 inches (15 to 30 centimeters).35

Streamflow: Hourly discharge measurements at the stream gages in Mobile County were not abnormally high, which is expected given the low precipitation amounts in the storms, hourly discharge measurements at the stream gages in Mobile County were likewise not significant.36 However, it is possible the precipitation from this event may have helped to saturate the soil and contribute to a much larger discharge event that was recorded in association with a precipitation event three days later.

Temperature: The temperature in Mobile on March 5, 1998 ranged from a low of 55°F (13°C) to a high of 75°F (24°C). These temperatures were warmer than previous days.37 It is possible the warm moist Gulf air advecting into the region contributed to this increase in temperature.

Winds: The highest observed wind gust in Mobile during the storm was 18 miles per hour (29 kilometers per hour) at the Mobile Regional Airport. The Mobile National Weather Service (NWS) Cooperative Observer Program (COOP) station measured a high 2-minute sustained wind of 24 miles per hour (38 kilometers per hour).38 The average wind speeds for the day were 6.4 and 7.5 miles per hour (10.2 and 12.0 kilometers per hour) at the two stations, respectively, compared to an average value for the surrounding week of 8.5 and 12.4 miles per hour (13.6 and 19.8 kilometers per hour).39

Table 49: Summary of Peak Storm Event Data for March 5, 1998 for the Mobile, Alabama region (including nearby Leakesville, Mississippi)

Variable Value at Peak Storm Intensity Value averaged across surrounding days
Surface Temperature - Max. 75°F 52°F
Surface Temperature - Min. 55°F 37°F
Precipitation Total 0.64 in ( < 2 year storm) -
Precipitation Rate 0.29 in/hr -
Hail Size (Diameter) 2 in -
Hourly Streamflow Discharge
Crooked Creek 51 ft3/s (< 2 year event) 15 ft3/s
Chickasaw Creek 308 ft3/s (< 2 year event) 300 ft3/s
Fowl River 40 ft3/s ( < 2 year event) 20 ft3/s
Wind Gust (@ Airport) 18 mph -
Sustained Wind
Mobile 16 mph -
Mobile Reg. Airport 24 mph -
Surface Pressure 1007 mb 1007 mb
Tidal Datum
Pensacola (MLLW) 1.62 ft ( < 10 year event) -

Surface Pressure: Surface pressure at both the Mobile Regional Airport and the Mobile weather stations averaged 1007 millibars on March 5, 1998, which was a typical value within that week.40

Water Level: The mean lower-low water level at the Pensacola tide station (the only tidal station with available data for March 1998) was unremarkable on March 5, 1998, peaking at 1.62 feet (49.38 centimeters) at high tide at 5:18 pm, close to the expected level of 1.08 feet (32.92 centimeters).41 Similarly, the mean high-high water level did not demonstrate any significant difference from expected levels.42

Storm Damage

This storm caused approximately $60,000 of damage in the Leakesville area.43 The severe hail chipped paint, dented house siding, stripped trees, and destroyed satellite dishes.44 In addition, nearly every vehicle that encountered the hail experienced damage.45

Figure 103: Image from the March 5, 1998 Hailstorm

Source: NWS, 2009b

This picture was taken in the Mobile region after the hailstorm. A layer of white hailstones one half to one inch think covers the roads and is hitting cars and infrastructure.

D.6.3. Study 3: Heavy Rain Event, April 4-5, 2008

Storm Development

On April 4, a line of intense storms moved east across central Alabama producing significant rainfall for the Mobile region. The storms developed in response to strong upper level north-to-south winds slowly steering a surface-level cold front into Mobile.46 Additional contributors to the severity of the storm included the warm moist air from the Gulf that was pulled into Mobile ahead of the cold front and the presence of vertical wind shear.47 Figure 104, on the left, shows a surface map of the United States for April 5, 1998 at 6 am (CST) demonstrating the cold front passing across Mobile, illustrated by the thick blue line with blue triangles.48 The green shading indicates precipitation. The right image in Figure 104 is a 500 millibar pressure map for April 5, 1998 at 7 am (CST) illustrating the strong jet stream aloft, shown by the blue lines with triangles moving in a northeast direction. A dip in the black atmospheric pressure lines below Louisiana and Arkansas is also evident.49

Figure 104 : Surface Maps of the United States for April 5, 1998 at 6:00 am and 7:00 am (CST)

This figure shows surface maps of the United States for April 5, 2008 at 6:00 and 7:00 am, right before a heavy rain event. The maps show a cold front passing across Mobile, bringing rain behind it. A strong jet stream is also visible, as indicated by a dip in the atmospheric pressure lines below Louisiana and Arkansas.This figure shows surface maps of the United States for April 5, 2008 at 6:00 and 7:00 am, right before a heavy rain event. The maps show a cold front passing across Mobile, bringing rain behind it. A strong jet stream is also visible, as indicated by a dip in the atmospheric pressure lines below Louisiana and Arkansas.

Storm Event Data

Precipitation: 8.32 inches (21.13 centimeters) of rain were observed at Mobile Regional Airport over a 15 hour period from 7:00 pm on April 4, 2008 to 10:00 am on April 5, 2008.50 At its peak, the rate of rainfall was 1.71 inches per hour (4.34 centimeters per hour) (at around 10 am).51

Streamflow: Hourly stream discharge spiked in the afternoon and evening of April 5 at three USGS streamflow gages in Mobile County.52 In that spike, discharge increased between 10- and 25-fold compared to the long-term average for that day.

Temperature: The high temperature on April 4-5, 2008 was 82°F (28°C), with a low of 60°F (16°C).53 Overall Mobile temperatures cooled off after the storm, as would be expected with a frontal event.

Wind: The highest observed wind gust in Mobile during the storm was 31 miles per hour (50 kilometers per hour) at Mobile Regional Airport.54 The fastest 2-minute sustained wind observed was 23 miles per hour (37 kilometers per hour).55 The average wind speeds for April 4th and 5th were 13.2 miles per hour (21.1 kilometers per hour) and 7.8 miles per hour (12.5 kilometers per hour), respectively, compared to an average value for the surrounding week of 7.5 miles per hour (12.0 kilometers per hour).56 If the day of the storm is not included, average wind speeds for the week were 6.4 miles per hour (10.2 kilometers per hour).

Table 50: Summary of Peak Storm Event Data for April 5, 2008 for the Mobile, Alabama Region

Variable Value at Peak Storm Intensity Value averaged across surrounding days
Surface Temperature - Max. 82°F 77°F
Surface Temperature - Min. 60°F 61°F
Precipitation Totals 8.32 in over 15 hr (~5 to 15 year storm) -
Precipitation Rate 1.71 in/hr -
Hourly Streamflow Discharge
Crooked Creek 610 ft3/s ( < 2 year event) 12 ft3/s
Chickasaw Creek 4,330 ft3/s ( < 2 year event) 275 ft3/s
Fowl River 1,420 ft3/s ( < 2 year event) 40 ft3/s
Wind Gusts 31 mph -
Sustained Winds 23 mph -
Surface Pressure 1003-1004 mb 1009 mb
Tidal Datum
Dauphin Island 1.58 ft ( < 2 year event) 0.71 ft
Mobile Docks 2.04 ft (< 2 year event) 0.97 ft
Pensacola 1.58 ft ( < 2 year event) 0.95 ft

Surface Pressure: The average surface pressure in Mobile on April 4 and 5 was approximately 1004 millibars, representing a very minimal drop from average non-storm surface pressures of 1009 millibars.57

Water Level: At 7:05 am on April 5, the Dauphin Island tide station observed peak water levels of 1.58 feet (48.16 centimeters) above the mean lower-low level, more than twice the expected water level of 0.71 feet (21.64 centimeters).58 At 8 am, Dauphin Island tide station recorded peak water levels of 0.38 feet (11.58 centimeters) above mean higher-high level.59

Similar patterns were observed at the other Mobile tidal stations. At 7:48 am on April 5, the Mobile Docks tide station observed peak water levels of 2.04 feet (62.18 centimeters), compared to the expected water level of 0.97 feet (29.57 centimeters).60 On April 5 at 1:24 am, Pensacola, Florida recorded 1.28 feet (39.01 centimeters) above the mean lower-low level compared to the expected water level of 0.70 feet (21.34 centimeters). Soon after noon time, Pensacola observed peak water level of 0.32 feet (9.75 centimeters) above mean higher-high levels, about 0.63 feet (19.20 centimeters) above expected levels.

Storm Damage

The April 4-5, 2008 storm moved slowly, inundating the region with varying amounts of rain. The majority of rain gauges in the area recorded around 8 inches (20 centimeters) of rain, but some reported close to 12 inches (30 centimeters) during the storm.61 The rain caused flooding in the streets of downtown Mobile, submerging multiple vehicles.62 The heavy rains also overwhelmed two wastewater pumping stations and caused over 13 million gallons (49 million liters) of sewage to spill into Mobile Bay.63 The sewage spill tainted water in the area for several days. The storm, identified as a 25-year storm by the National Weather Service, also downed trees and power lines, causing 7,600 homes to lose power.64 Throughout Alabama, the severe storm spawned tornadoes, damaging trees and buildings.

D.6.4. Study 4: Hurricane Georges, September 28, 1998

Storm Track and Intensification

Hurricane Georges began as a tropical depression on September 15, 1998, four hundred miles south-southwest of Cape Verde.65 As the storm traveled westward, it steadily intensified, developing into a tropical storm on September 16 and reaching hurricane strength by September 17. On September 19, when it was just a few hundred miles east of the Caribbean, Hurricane Georges' strength peaked as a Category 4 storm with winds of 150 miles per hour (240 kilometers per hour).66 Hurricane Georges caused damage in Puerto Rico, Dominican Republic, Haiti, and Cuba as it traveled towards the Gulf of Mexico, weakened at one point by the mountainous terrain of the Dominican Republic and Haiti.

On the left, this figure shows the storm track of Hurricane Georges, which struck Mobile in 1998. The storm developed in the tropics and moved over Hispaniola and Cuba into the Gulf of Mexico. It peaked as a Category 4 over the ocean, but hit Mobile as a Category3 storm.

Hurricane Georges entered the Gulf of Mexico on September 25, traveling north-northwest at an average speed of 11 miles per hour (18 kilometers per hour).67 The storm began to strengthen as it moved into the warm waters of the Florida Straits moving in a west-northwest track. Sea surface temperatures in the Gulf near the track of Hurricane Georges were estimated to be 81.7°F (27.6°C).68 Sea surface temperatures typically must be at least 82°F (28°C) for a storm to develop and maintain its strength.69 Climatological monthly mean sea surface temperatures in the Gulf for the month of July range from about 77°F (25°C) to 86°F (30°C) with cooler temperatures towards the coastlines.

Figure 105 : Storm Track of Hurricane Georges (left) and Infrared Image of Georges (right)

On the right, the figure shows an infrared image of Georges. The image shows that Georges was a large hurricane and hit Mobile nearly directly. The eye of the storm is centered just west of Mobile over Biloxi, Mississippi.

Georges made U.S. landfall near Biloxi, Mississippi around 6:30 am on September 28, 1998 as a Category 2 storm. The storm moved slowly over land and reached Mobile in the early morning of September 29.70 Because the storm moved so slowly, Alabama experienced significant torrential rains and coastal storm inundation.71 Figure 105 shows Georges' storm track approaching the Gulf Coast, where the color denotes the storm's Saffir-Simpson intensity rating.72 The image at the right is an enhanced infrared image of Georges that shows the shape and activity of the storm soon after hitting land.73

Figure 106: Total Rainfall from Hurricane Georges (Sept. 26-30, 1998)

This figure shows a map of Alabama and total rainfall levels associated with Hurricane Georges. The area directly around Mobile received about 15 inches of rain, with a small area upstream of Mobile that received 25 inches of rain from the storm. Rainfall levels of 10, 5, and 2 inches fell in the rest of the state, though no rainfall occurred in the northern third of Alabama.

Source: NOAA, 2011j

Storm Event Data

Precipitation: Total precipitation at Keesler AFB in Biloxi, where Georges made landfall, was recorded at 9.18 inches (23.32 centimeters).74 Total precipitation measured across much of southern Alabama ranged between 20 and 30 inches (51 and 76 centimeters) over the duration of the storm.75 Mobile Regional Airport recorded 15.02 inches (38.15 centimeters) of total precipitation while downtown Mobile recorded 13.13 inches (33.35 centimeters).76

Streamflow: Hourly discharge at the three Mobile stream gages peaked on September 28 and 29 far above their typical levels. The stream gage at Chickasaw Creek experienced the most dramatic discharge increase to over 19,000 cubic feet per second (570 cubic meters per second), up to 180 times its typical levels.77

Wind: At landfall, maximum sustained surface winds were 105 miles per hour (168 kilometers per hour), which is a Category 2 storm on the Saffir-Simpson Scale.78 Brief wind gusts up to 125 miles per hour (200 kilometers per hour) were observed at Keesler AFB in Biloxi.79 Winds in Mobile the next day were sustained at 57.5 miles per hour (92.0 kilometers per hour), with gusts observed up to 80 miles per hour (128 kilometers per hour) at Dauphin Island.80 Sustained winds at Mobile Regional Airport were 51 miles per hour (82 kilometers per hour), with peak gusts of 83 miles per hour (133 kilometers per hour), while measured winds at Mobile's Brookley Field were sustained at 54 miles per hour (86 kilometers per hour) with peak gusts of 62 miles per hour (99 kilometers per hour).81

Surface Pressure: Central storm pressure was 964 millibars at landfall, and increased slowly as it approached Mobile to 986 millibars.82

Storm Surge and Water Level: At landfall in Biloxi, storm surge was as high as 8.8 feet (268.2 centimeters).83 Figure 107 displays the storm surge measured at locations around the Mobile region. Storm surge ranged between 5 to 10 feet (152 to 305 centimeters) in Mobile County, with measured storm surge of 8.5 feet (259.1 centimeters) in Downtown Mobile.84 The highest storm surge in Alabama was recorded in west Mobile Bay at 9.3 feet (283.5 centimeters).85 Dauphin Island experienced storm surge of 5 feet (152 centimeters) on the bay side and 6.6 feet (201.2 centimeters) on the Gulf side, and the Mobile Bay Causeway experienced 8 to 9 feet (244 to 274 centimeters) of storm surge.86 High water marks near landfall were between 7 and 10 feet (213 and 244 centimeters).87 In Mobile, high water marks ranged between 7 and 8 feet (213 to 244 centimeters), while sites in Mobile County experienced high water marks up to 10.3 feet (313.9 centimeters).88 Every coastal river and stream from Mississippi to the Florida panhandle experienced serious, life-threatening flooding.89

Table 51: Summary of Storm Event Data for Hurricane Georges

Variable Value at Peak Storm Intensity
At Landfall (Biloxi, MS) In Mobile
Precipitation Totals 9.18 in 10-20 in (for a given day, approximately a 10 year storm)
Hourly Streamflow Discharge
Crooked Creek - 1,070 ft3/s (> 2 year event)
Chickasaw Creek - 19,900 ft3/s ( ~ 25 year event)
Fowl River - 3,340 ft3/s (< 5 year event)
Wind Gusts 125.4 mph 83 mph
Sustained Winds 105 mph 57.5 mph
Central Storm Pressure 964 mb 986 mb
Highest observed water level 12.60 ft 10.33 ft
Storm Surge 8.8 ft 8.5 ft (> 10 year event)

Figure 107 : Hurricane Georges representative storm surge above NGVD for Mobile, Baldwin, Jackson, and Escambia Counties

Adapted from U.S. Army Corps of Engineers, 1999

This figure shows the storm surge from Hurricane Georges that affected Mobile, Baldwin, and surrounding counties. The figure indicates that Mobile was hit with 8-10 feet of storm surge, particularly at the mouth of the Mobile Bay, in downtown Mobile, and on the southern coat. Dauphin Island recorded storm surges of 5 to 6 and a half feet.

Storm Damage

Georges caused severe flooding along the Gulf Coast from Mississippi to Florida, including the Mobile region. Downtown Mobile was heavily flooded as a result of heavy precipitation and high storm surge (see Textbox below). This resulted in inundated and blocked roadways. The Mobile Bay Causeway was fully inundated, disabling transportation across the bay between Mobile and Baldwin Counties.

Picture of a home on the dog river. The lawn of the house is littered with debris, including 3 large beached sailboats and one smaller boat.Picture showing an SUV in a flooded parking garage, with water more than covering the car's tires and side.

D.6.5. Study 5: Hurricane Katrina, August 29, 2005

Storm Track and Intensification

Hurricane Katrina was one of the most destructive hurricanes to hit the United States.90,91 The storm formed from the combination of a tropical wave, an upper-level trough, and the mid-level remnants of Tropical Depression Ten.92 Hurricane Katrina began its early development on August 23 as a tropical depression about 175 miles (280 kilometers) southeast of Nassau, Bahamas.93 On August 24, the tropical depression became a tropical storm as it headed towards the Bahamas. In the early evening of August 25, the storm strengthened to a Category 1 hurricane with sustained winds of 80 miles per hour (128 kilometers per hour) just before making landfall in Florida between Hallandale Beach and North Miami Beach. Hurricane Katrina crossed the southern tip of Florida through the night and then began to re-intensify once over the warm waters of the Gulf (sea surface temperatures were 2°F to 4°F (1°C to 2°C) above normal).94

From August 25 to August 31, Hurricane Katrina slowly turned to the northwest and north as a mid-level ridge that had been situated over Texas weakened. As Hurricane Katrina moved towards landfall, the upper atmosphere conditions and the above-normal sea surface temperatures aided in Katrina's continued intensification into a major hurricane. In addition, the vertical wind shear (i.e., caused by changes in the atmospheric wind direction and/or strength with altitude) was less than normal, which is conducive to hurricane intensification. On August 28, Hurricane Katrina became a Category 5 hurricane with peak winds speeds near 175 miles per hour (280 kilometers per hour) and a central pressure of 902 millibars. Katrina was a large storm extending out about 105 miles (168 kilometers) from its center with tropical storm force winds extending out another 100 miles (160 kilometers).

Figure 108: Storm Track and Infrared Image of Hurricane Katrina

On the left, this figure shows the storm track of Hurricane Katrina, which impacted the Gulf Coast in 2005. The storm developed in near the Bahamas and moved across the southern tip of Florida and into the Gulf of Mexico. It peaked as a Category 5 storm over the Gulf and made landfall in New Orleans as a strong Category 3 storm.On the right, the figure shows an infrared image of Katrina. The image shows that Katrina was a very large hurricane, which hit Mobile although the eye was not centered on it.

On the morning of August 29, Hurricane Katrina made landfall in Plaquemines Parish, Louisiana as a strong Category 3 hurricane with wind speeds of about 127 miles per hour (203 kilometers per hour) and a central pressure of 920 millibars. Hurricane Katrina made its final landfall near the Louisiana-Mississippi border at 9:45 am local time with winds reported at near 121 miles per hour (194 kilometers per hour).

Storm Event Data

Observations from Hurricane Katrina are provided below for both the site of landfall (near Slidell, LA and Gulfport, MS) and Mobile, AL.

Precipitation: Slidell, LA and Gulfport, MS, the two cities closest to Katrina's point of landfall, measured up to 11 to 12 inches (28 to 30 centimeters) of total precipitation on August 29.95 Mobile reported a total of 3.8 inches (9.7 centimeters) of rainfall.96

Streamflow: At the three Mobile County stream gage sites, hourly discharge measurements demonstrate a drop a sharp increase on August 29-30.97 These spikes were an increase of 4-, 16-, and 20-fold over typical discharge values at Crooked Creek, Chickasaw Creek, and Fowl River, respectively.

Wind: As Katrina made landfall at the Louisiana-Mississippi border, the hurricane had maximum sustained winds of 120.8 miles per hour (193.3 kilometers per hour), with gusts up to 123 miles per hour (197 kilometers per hour), which is within Category 3 wind speeds.98 Mobile experienced maximum sustained winds of 66.7 miles per hour (106.7 kilometers per hour) and peak gusts of 84 miles per hour (134 kilometers per hour).99

Surface Pressure: Katrina's central pressure at landfall near Buras was 923 millibars.100 The surface pressure in Mobile was measured at 984 millibars at 11:45 am, two hours after Katrina made landfall.101

Water Level: Hurricane Katrina caused a storm surge of 24 to 28 feet (7 to 9 meters) at the Slidell, LA / Gulfport, MS coastline.102 Mobile County observed large storm surges of 10 to 15 feet (3 to 4.6 meters), and the storm caused flooding several miles inland from the Gulf coast along Mobile Bay, where there was a storm surge of 8 to 12 feet (2. to 3.7 meters. Dauphin Island recorded storm surge of 6.63 feet (2.02 meters) and Mobile State Docks recorded storm surge of 11.45 feet (3.49 meters).103 High water marks surveyed in Mobile reached 11 to 12.5 feet (3.6 to 3.8 meters).104 On the northeastern portion of Mobile Bay, flooding elevations reached over 13 feet (4 meters), north of Fairhope.105

Table 52: Summary of Storm Event Data for Hurricane Katrina

Variable Value at Peak Storm Intensity
At Landfall (Slidell, LA / Gulfport, MS) Mobile, Alabama
Precipitation Totals 11-12 in. 4 in (~ 2 year storm)
Hourly Streamflow Discharge
Crooked Creek 330 ft3/s (8/29) ( < 2 year event)
Chickasaw Creek 3,140 ft3/s (8/30) ( < 2 year event)
Fowl River 677 ft3/s (8/30) ( < 2 year event)
Wind Gusts 123 mph 84 mph
Sustained Winds 121 mph 67 mph
Central Storm Pressure 928 mb 984 mb
Highest observed water level 27.9 15 ft
Storm Surge 24-28 ft 10-15 ft (~ 25 year event)

Storm Damage

Mobile County experienced significant damage from Hurricane Katrina, primarily in the form of coastal flooding and storm surge. Storm surge on Dauphin Island destroyed or damaged dozens of homes. On Dauphin Island, flooding elevations were between 8.5 and 11.5 feet (2.6 and 3.5 meters).106 In the city of Mobile, typical flood depths were on the order of 11 to 12.5 feet (33 to 3.8 meters), causing severe inundation and shut down of most major roadways.107 These are close to the highest levels ever recorded in Mobile.108 Downtown Mobile was entirely inundated, causing authorities to issue a dusk-to-dawn curfew for the area. The Mobile Bay Causeway was fully inundated, as it was during Georges, disabling transport across the bay.109

The Mobile area also experienced debris damage from oil rigs during Hurricane Katrina. Dauphin Island experienced damage from an offshore oil rig that washed up on the shore. In addition, an oil rig under construction along the Mobile River in Alabama was dislodged and carried 1.5 miles (2.4 kilometers) north where it struck the Cochrane Bridge just north of downtown Mobile.110

Damage Images from Hurricane Katrina in Mobile, Alabama

Severe flooding occurred in downtown Mobile.

Photos courtesy of wunderground.com; Used with permission.

Picture showing downtown Mobile flooded. Only the tops of trees and some lamp posts are visible above the flowing water.Picture showing downtown Mobile flooded. Only the tops of trees and some lamp posts are visible above the water.Picture showing downtown Mobile flooded. Only the tops of trees and some lamp posts are visible above the choppy water.

D.7. Projected changes in U.S. and Global Storm Events

The United States experiences a wealth of storm activity including severe thunderstorms producing tornadoes, nor'easters producing intense winds and precipitation, and tropical storms and activities. The duration, frequency, location, and intensity of storm events in the United States will likely evolve in response to changes in climate. Studies suggest that storm events will increase in severity across the United States. This appendix discusses projected changes in storm events across the United States and world in the coming century, including extreme precipitation events, tropical storms and hurricanes.

D.7.1. Extreme Precipitation Events

By the end of the century, extreme precipitation events in the United States are projected to increase in both frequency and intensity. By the end of the century, heavy precipitation events that currently have a 5% chance of occurring in a given year are projected to have a 7 to 25% chance of occurring.111 Extreme heavy downpours are also projected to produce more precipitation.112 Meanwhile, light precipitation events are projected to become less frequent. Taken together, these projections suggest that total annual precipitation may not change significantly, but the way that precipitation is delivered might be different.113

Scientists do not entirely understand the physical mechanisms that contribute to storm intensification, particularly when characterizing the influence of natural variability such as El Niño events and other large-scale circulation patterns. However, some mechanisms in the development and/or steering of U.S. storms are well-understood, such as the location and strength of the jet stream.114

D.7.2. Tropical Storms and Hurricanes

The recent scientific consensus on tropical cyclonic activity suggests tropical cyclones may globally decrease in frequency but increase in intensity. This consensus suggests that the globally averaged intensity of tropical cyclones will increase by 2 to 11% by the end of the century but the globally averaged frequency will decrease by 6 to 34%.115 This consensus is further supported by a recent study which estimates an increase in the number of more severe (Category 4 and 5) storms, but a decrease in the total number of tropical storms and hurricanes in the tropical Atlantic.116

There is observational evidence since about 1970 that the intensity of tropical cyclone activity has been increasing, correlated with increases in tropical sea-surface temperatures in the North Atlantic (a tropical cyclones refers to specific stages of hurricane development from tropical depression to tropical storm whereupon the storm is named to hurricane).117 There is also evidence of an increase in extreme wave height over the past two decades, associated with more frequent and intense hurricanes.118 However, the World Meteorological Organization (WMO) Sixth International Workshop on Tropical Cyclones in 2006 agreed that "no firm conclusion can be made" about anthropogenic influence on tropical cyclone activity because "there is evidence both for and against the existence of a detectable anthropogenic signal in the tropical cyclone climate record."119 Recently, there is growing confidence in the model projections that climate change may increase hurricane strength, but it is still unclear how the overall frequency of occurrence might change.120

D.8. Detailed Methodology for Scenario-based Storm Surge Analysis

This appendix describes in detail the methodology used to conduct a scenario- and model-based analysis of storm surge in the Mobile region.

The projected increase in hurricane intensity has the potential to increase flooding from coastal storms striking the Mobile area. As tropical storms and hurricanes approach Mobile, generally from a southerly direction, multiple physical properties promote flooding in the region:

The scenario- and model-based analysis included the following steps, which correspond to the sections of this appendix:

D.8.1 Selection of Storm Surge Scenarios

Although there is a strong theoretical basis underpinning the scientific assertion that the intensity of hurricanes will increase in the future, it is difficult to probabilistically estimate the number and intensity of hurricanes that will strike the Mobile region over the 21st Century. Therefore, as with the sea-level rise analysis, scenarios were selected to analyze the implications of a wide range of storms that could plausibly strike Mobile. For this analysis, records from historic storms were selected to use as the basis in developing these storm scenarios.

There were two main questions that the scenario-based analysis attempted to address:

In selecting the storms, historical storms were chosen that met the following criteria:

After reviewing records of all landfalling hurricanes in the Mobile area over the past few decades, the 1998 Hurricane Georges was selected to address Question #1, and the 2005 Hurricane Katrina was selected to address Question #2.

Using Hurricanes Georges and Katrina as base storms, 11 storm scenarios (see Table 53) were developed by adjusting certain characteristics of the storm parameters to simulate what could happen under alternate conditions. For the Georges simulations, all four sea level rise scenarios (0 meters, 0.3 meters, 0.74 meters, and 2.0 meters (0, 1.0, 2.5, and 6.6 feet)) were examined. Results for ADCIRC are reported relative to Mean Sea Level.123 For the Katrina simulations, the modeling considered different adjustments, including shifting the path of Katrina so that it hit Mobile directly, intensifying the storm, and adding in 0.75 meters (2.5 feet) of sea level rise. Two of the 11 scenarios were hindcasts of Georges and Katrina. They were used to validate the model and to serve as a basis from which to build the other 9 scenarios.

Anatomy of the "perfect storm"

A number of highly damaging storms have struck the Mobile area. However, none of them have had all of the characteristics that would maximize the impact. Below are features that would contribute to a "perfect storm."

  • High winds. Although flooding is generally the greatest source of damage in a hurricane, winds are responsible for most of the storm surge that occurs. Damage to buildings increases exponentially in proportion to the wind speed. Hurricane Katrina was a Category 3 (winds of 111-30 mph) storm at landfall, whereas the strongest hurricanes reach Category 5 (winds greater than 156 mph).
  • Slow movement. The slower the eye of the hurricane moves, the more time it has to pile up water against the coastline, thereby enhancing the storm surge. Although Hurricane Frederic was a Category 3 storm at landfall, its surge was similar to that of the weaker, Category 2 Hurricane Georges, due in part to Frederic's relatively rapid forward speed.
  • Large size and long travel distance (i.e., fetch). The larger the wind field, the more water will be pushed against the coastline, thereby exacerbating the surge. The same is true for the fetch, in which the farther the storm travels across open water prior to landfall the greater its surge will tend to be.
  • High precipitation. The precipitation from the storm can exacerbate the storm surge by creating inland flooding that meets the marine surge in the estuary or further up the rivers that feed into the estuary.

Table 53 : Storm Scenarios

Name Sea level rise Track Shift Amplification Question Addressed124
Georges-Natural None No None Baseline
Katrina-Natural None No None Baseline
Georges-Natural-30cm 0.3 m No None (1)
Georges-Natural-75cm 0.75 m No None (1)
Georges-Natural-200cm 2.0 m No None (1)
Katrina-Natural-75cm 0.75 m No None (1), (2)
Katrina-Shift125 None Yes None (2)
Katrina-Shift-75cm 0.75 m Yes None (2)
Katrina-Shift-ReducedPress-75cm 0.75 m Yes Central pressure reduced according to Knutson and Tuleya (2004) (2)
Katrina-Shift-MaxWind None Yes Max. wind speed sustained through landfall (2)
Katrina-Shift-MaxWind-75cm 0.75 m Yes Max. wind speed sustained through landfall (2)

Explanation of scenario names:

Figure 109 : Original Track of Hurricane Katrina

The image shows the observed track of Hurricane Katrina. Each dot represents the approximate location of the NOAA National Hurricane Center six-hour advisory bulletin used in the model simulations. kph = knots per hour. The times are UTC.

This figure shows the original track of Hurricane Katrina, which came up through the Gulf Coast and made landfall near New Orleans, Louisiana.

Figure 110 : Shifted Track of Hurricane Katrina

This figure shows the shifted track of hurricane Katrina used in the Advanced Circulation (ADCIRC) model for this study. The new track is shifted eastward to directly strike Mobile.

This image shows the shifted track of Hurricane Katrina that corresponds to five of the scenarios that were explored in this study.

D.8.2. Advanced Circulation Modeling

Simulations of storm-induced water levels (i.e. storm surge) were performed using the ADvanced CIRCulation model, ADCIRC.129 This finite-element hydrodynamic code is robust, well-developed, extensively-tested, and highly adaptable to a number of coastal-ocean processes. The storm simulations presented in this study were performed using the two-dimensional, depth integrated (2DDI) form of ADCIRC assuming barotropic forcing only (i.e. no density-driven flows). While the ADCIRC model is capable of applying a variety of internal and external forcings, including tidal forces and harmonics, inflow boundary conditions, density stratification, and wave radiation stresses, only the meteorological forcing input was used to drive the storm-induced flows and water levels.

The generalized wave continuity and momentum equations were discretized using finite-element techniques on an unstructured network of nodes and connective elements. These techniques allow for varying degrees of resolution throughout the computational domain, and provide great flexibility when modeling flows within complex or irregular boundaries. The organization of nodes and elements, and the information that the nodes contain (i.e., horizontal coordinates, elevation, roughness, etc.), is collectively referred to as the ADCIRC mesh. A diagram of the mesh used in this study is provided in Figure 111. A portion of this mesh (shown in red), extending from the Northwest Florida panhandle to the Mississippi-Alabama border, extends above mean sea level to accommodate storm surge inundation and flood modeling. Resolution in much of Mobile's metropolitan area ranges from 165 to 495 feet (50 to 150 meters). It is much coarser offshore in order to reduce the computational demand and computer run-time. Meshes such as these, which are quite time intensive to produce, are available for most of the U.S. coastline.130

Figure 111 : Diagram of the ADCIRC Mesh (SSv31L) Used in this Model Study of Hurricanes Georges and Katrina

This diagram shows the mesh used in this study for ADCIRD modeling. A portion of the mesh, highlighted in red, extends from the Northwest Florida panhandle to the Mississippi-Alabama border and extends above mean sea level to accommodate storm surge inundation and flood modeling. Data points are concentrated near the coasts, with coarser modeling offshore.

The ADCIRC mesh used in this study is composed of 446,459 discrete nodes with connective nodestrings creating 866,496 triangular mesh elements. Each node in the ADCIRC mesh contains at least two layers of information: the node's horizontal coordinate, which is given by latitude and longitude when performing simulations in spherical coordinates; and the node's elevation relative to mean sea level (MSL).

The model takes into account the slope of the land in its estimation of surge. However, since the surge model does not take into account wave breaking and wave run-up, it does not take into account the effect of the slope of the land on waves. The effect of the slope of the land is accounted for in the wave model (STWAVE, described below), but this does not feed back to the ADCIRC model since the two models are run asynchronously.

The ADCIRC storm simulations considered in this validation study were driven by meteorological forcing data extracted from six-hour advisory forecast and observation reports issued by the NOAA National Hurricane Center (NHC). The meteorological data required by the ADCIRC model includes a time code, the latitude and longitude of the eye, the maximum observed wind speed in knots, the minimum sea level pressure in millibars, and the radius, in nautical miles, from the center of the storm to a specified wind intensity (e.g., 34, 50, 64, or 100 knots) in each of the storm's four quadrants. These data must be assembled in a modified Automated Tropical Cyclone Forecast (ATCF) best track format as described in Luettich and Westerink (2004). An asymmetric hurricane vortex formulation131 based on a Holland-type gradient wind model132 was used to estimate the wind and pressure field of the storm. The Garratt (1977) formula was used to convert wind speed to an applied wind stress. These data are spatially interpolated onto the ADCIRC mesh, and a linear interpolation was used to map six-hour advisory data to each intermediate time that the model performs its calculations133 falling between advisory information. A general schematic of this process is provided in Figure 111.

Figure 112: A Representative Model Schematic for Meteorological Coupling in ADCIRC Storm Simulations 134

This figure illustrates how meteorological coupling occurs in ADCIRC storm simulators. First, ADCIRC obtains six-hour advisory storm data from the National Hurricane Center. Second, it estimates wind and pressure fields using a Holland-type model. Third, it interpolates storm winds to mesh. Finally, the ADCIRC simulation is performed.

D.8.3. Advanced Circulation Model Testing

Hindcast simulations of storm-induced water levels using the ADCIRC hydrodynamic model were completed for Hurricanes Georges and Katrina. As noted above, these simulations were driven by historical parameters and tracks. They were used to evaluate the model's ability to accurately reproduce the spatial distribution and peak storm-induced water levels of historical events and to assign a quantitative measure of accuracy to model predictions. This evaluation is indicative of the veracity of simulations of the surge response to various future storm scenarios for which no equivalent comparators exist for model-data verification.

Hurricane Katrina

The hindcast ADCIRC simulation of Hurricane Katrina was initiated using data from August 27, 2005 at 1800 hrs UTC. By this time, Katrina had already made an initial landfall on the southeast coast of Florida, had weakened in intensity, and was beginning its first of two rapid intensification periods.135 Hurricane Katrina made its final landfall at 1110 hrs UTC on August 29, 2005 near Grand Isle, Louisiana. At landfall, Katrina had maximum sustained winds ranging from 117 to 126 knots, as well as stronger gusts, and minimum central pressures ranging from 918 to 923 millibars.136 An overview of Katrina's historical track is provided in Figure 109. In this figure, each dot represents the location of the eye of the storm when an NHC advisory bulletin was issued providing basic storm parameters such as location, maximum winds, minimum sea level pressure, and radius to specific wind speed intensities in each of the storm's four quadrants. The first dot denotes the beginning of the model simulation.

The total duration of the Katrina hindcast simulation is 2.75 days, including an initial 0.5-day ramp period to avoid instabilities associated with model spin-up. The initial starting time for the simulation was chosen to capture all relevant storm effects during its residence time in the Gulf of Mexico prior to landfall, including any setup along the Alabama shelf that may elevate pre-storm water levels above the astronomical tide. For evaluation purposes, a 3.75-day simulation was performed and no measurable differences were obtained over much of the simulated storm surge hydrograph.

Please see Appendix D.6.5 for more information on this storm.

Hurricane Georges

Hurricane Georges started as a tropical wave off the west coast of Africa. After a week of development, the storm reached its peak intensity of 135-knot winds at 0600 hrs UTC on September 20, 1998 in the Lesser Antilles. Shortly thereafter, the storm diminished in strength prior to making its first of several landfalls in the Lesser Antilles. A succession of weakening and re-intensification ensued while the storm passed through the Lesser Antilles and over Puerto Rico, until it was significantly weakened as it passed over the mountainous terrain of the island of Hispaniola and later Cuba. Upon entering the Gulf of Mexico, a modest re-intensification from 65 to 90 knot one-minute sustained wind speeds was accompanied by a gradual reduction in forward speed. Georges made landfall near Biloxi, Mississippi early on the morning of September 28, 1998. At landfall, Georges had maximum sustained winds of 90 knots and a minimum central pressure of 964 millibars. A more thorough report of Georges' history and characteristics can be found in Guiney (1999). The hindcast simulation of Georges began on September 26, 1998 at 0300 hrs UTC. An overview of Georges' historical track is provided in the Figure 113 below, where each dot represents the location of the eye of the storm when an NHC advisory bulletin was issued. The first dot denotes the beginning of the model simulation.

Figure 113 : A General Overview of Hurricane Georges' Historical Track through the Caribbean Basin and Gulf of Mexico

The ADCIRC model simulation begins on 9/26/98 at 0300 UTC. Each dot represents the approximate location of the NHC six-hour advisory bulletin used in the baseline model simulation. The dates correspond to eye position at 0300 hrs UTC.

This figure shows how Hurricane Georges moved through the Caribbean and the Gulf of Mexico. Georges is represented by a dot on the map, showing the date at which the National Hurricane Center issued an advisory. At the first dot, on September 26, Georges had just passed over Cuba and was west of the Florida Keys. On September 27, Georges was in the middle of the Gulf of Mexico. On September 28 and 29, Georges made landfall on the Gulf Coast. The ADCIRC modeling simulation began with the first dot.

The total duration of the Georges hindcast simulation is 3.5 days, including an initial 0.5-day ramp period to avoid instabilities associated with model spin-up. The initial starting time for the simulation was chosen to capture all relevant storm effects during its residence time in the Gulf of Mexico prior to landfall, including any setup along the Alabama shelf that may elevate pre-storm water levels above the astronomical tide. For evaluation purposes, a 4-day simulation was performed and no measurable differences were obtained over much of the simulated storm surge hydrograph.

Please see Appendix D.6.4 for more information on this storm.

Evaluation Metrics

Two metrics were used to evaluate the ability of ADCIRC to accurately reproduce storm-induced water levels: time series of measured and simulated water levels at discrete spatial locations (i.e. storm surge hydrographs); and spatial distributions of measured high water marks (HWMs137) on land. Measured water levels were provided, where available, by a number of gages maintained by the NOAA's Center for Operational Oceanographic Products and Services (CO-OPS).138 FEMA (2006a) provided measured HWMs for Katrina and USACE (1998) provided measured HWMs for Georges. The ADCIRC-simulated maximum water levels at each node in the mesh were interpolated to the geographic coordinates of the measured HWMs for comparison.

There are discrepancies associated with the comparison of measured and simulated HWMs, as HWMs are measured relative to a survey datum, while model output is provided relative to the MSL tidal datum. There are only two locations in Mobile County where the relationship between these two datum is known: Dauphin Island and Mobile State Docks. However, only the value at Dauphin Island is published by NOAA CO-OPS, because the values at Mobile State Docks, do not meet NOAA/NGS vertical accuracy requirements. The approximate error associated with this datum discrepancy may be +/- 0.59 feet (0.18 meters) for Katrina, where HWMs are provided in NAVD88. The approximate error may be closer to +/- 0.17 feet (0.05 meters) for Georges, where HWMs are provided in NGVD29139. Time-series water level comparisons were performed relative to a consistent survey datum (NAVD88). All tidal datum values correspond to the National Tidal Datum Epoch of 1983 - 2001. The comparison of high water marks for Georges and Katrina are presented here, but discussed in more detail in the following sections. The discussion focuses on Dauphin Island and Pensacola since those are the two nearby gages that had the most continuous records during the two storms.

Figure 114: Water Levels at Dauphin Island, Alabama during Hurricane Katrina

The predicted tide (red dash) is the level that would have been expected in the absence of the storm. The measured level (green circles) is what was actually observed. The simulated level (blue line) is the water elevation predicted by the model, driven by observed winds and atmospheric pressure, but without accounting for tides. The predicted and measured levels in feet relative to Mean Higher High Water were obtained from the NOAA Tides & Currents data repository using CO-OPS station 8735180. Note that the simulated water levels shown do not include values obtained during the initial 0.5-day model ramping period.

This figure plots the water levels observed at Dauphin Island during Hurricane Katrina along with the water levels simulated by the ADCIRC model. The model reproduces the shape, duration, and elevation of the peak storm surge.

Figure 115: Water Levels at Pensacola, Florida during Hurricane Katrina

The predicted tide (red dash) is the level that would have been expected in the absence of the storm. The measured level (green circles) is what was actually observed. The simulated level (blue line) is the water elevation predicted by the model, driven by observed winds and atmospheric pressure, but without accounting for tides. The predicted and measured levels in feet relative to Mean Higher High Water were obtained from the NOAA Tides & Currents data repository using CO-OPS station 8729840. Note that the simulated water levels shown do not include values obtained during the initial 0.5-day model ramping period.

This figure plots the water levels observed at Pensacola during Hurricane Katrina along with the water levels simulated by the ADCIRC model. The model reproduces the general shape and duration of the peak in the observed surge, but under-predicts the observed peak water level by about 2 feet.

Figure 116: Water Levels at Dauphin Island, Alabama during Hurricane Georges

The predicted tide (red dash) is the level that would have been expected in the absence of the storm. The measured level (green circles) is what was actually observed. The simulated level (blue line) is the water elevation predicted by the model, driven by observed winds and atmospheric pressure, but without accounting for tides. The predicted and measured levels in feet relative to Mean Higher High Water were obtained from the NOAA Tides & Currents data repository using CO-OPS station 8735180. Note that the simulated water levels shown do not include values obtained during the initial 0.5-day model ramping period.

This figure plots the water levels observed at Dauphin Island during Hurricane Georges along with the water levels simulated by the ADCIRC model. The model reproduces both the shape and elevation of the peak storm surge; however, the rising and receding limbs of the simulated hydrograph do not correspond precisely to the observations. This is due to the absence of astronomical tidal oscillations in the model.

Figure 117: Water Levels at Pensacola, Florida during Hurricane Georges

The predicted tide (red dash) is the level that would have been expected in the absence of the storm. The measured level (green circles) is what was actually observed. The simulated level (blue line) is the water elevation predicted by the model, driven by observed winds and atmospheric pressure, but without accounting for tides. The predicted and measured levels in feet relative to Mean Higher High Water were obtained from the NOAA Tides & Currents data repository using CO-OPS station 8729840. Note that the simulated water levels shown do not include values obtained during the initial 0.5-day model ramping period.

This figure plots the water levels observed at Pensacola during Hurricane Georges along with the water levels simulated by the ADCIRC model. The model reproduces the general shape and duration of the peak in the observed surge, but under-predicts the observed peak water level by about 1 foot.

For Hurricane Katrina at Pensacola, the ADCIRC model reproduces the general shape and duration of the peak in the observed surge hydrograph, but underpredicts the observed peak water level by about 2 feet (0.6 meters). Note that this location is the most distant from the storm in the hindcast analysis. At Dauphin Island, the model appears to be in good agreement with measurements: the model reproduces the shape, duration, and elevation of the peak in the storm surge hydrograph.

For Georges, the performance of the model at Pensacola and Dauphin Island appears to be generally consistent with the hindcast simulation of Katrina. The model captures the general shape and duration of the peak in the storm surge hydrograph at Pensacola, but underpredicts the peak water level by about 1 foot (0.3 meters. At Dauphin Island, the model reproduces both the shape and elevation of the peak storm surge; however, the rising and receding limbs of the simulated hydrograph do not correspond precisely to the observations due to the absence of astronomical tidal oscillations in the model. The tidally-related error of the model simulations presented here is roughly ± 0.7 feet (0.2 meters), which is half the tidal amplitude in Mobile.

A quantitative error analysis covering the duration of both storm simulations was conducted. For Katrina, the root mean square (RMS) errors at Pensacola and Dauphin Island are 1.2 feet (0.4 meters) and 1.0 feet (0.3 meters), respectively. The Percent of Peak (POP), which is the RMS error divided by the maximum observed (measured) water level, was also computed. It is a more useful measure of model accuracy. The POP RMS errors for Katrina at Pensacola and Dauphin Island are roughly 20% and 16%, respectively. For Georges, the RMS errors between simulated and measured water levels at Pensacola and Dauphin Island are 1.38 feet (0.42 meters) and 0.88 feet (0.27 meters), respectively. The corresponding POP RMS errors for Georges are 29.3% and 17.6%. Note that both the RMS and POP RMS errors have similar magnitudes for both hindcast simulations. Given the number of sources of uncertainty discussed below (e.g., lack of tides, wave breaking, and river runoff in the simulations), these uncertainties are quite low for this type of hindcast.

The spatial range of differences between simulated and observed high water marks for both storms is shown in Figure 118 and Figure 119 below.

Figure 118 : A Comparison of Measured and Simulated High Water Mark Errors (in feet) in Mobile and Baldwin Counties during Hurricane Katrina

The comparison error is computed as (simulated) - (measured) such that positive and negative values suggest an over prediction and under prediction of water levels, respectively. A datum discrepancy between measurements (NAVD88) and simulated high water marks (MSL) may introduce differences on the order of +/- 0.7 feet.

This figure shows a map of Mobile bay with color-coded dots indicating the difference between observed high water and those simulated by the ADCIRC model for Hurricane Katrina. The ADCIRC model tended to under-predict high water marks on Dauphin Island, the southern coast of Mobile, and the Mobile State Docks. The model slightly over-predicted a high water mark near downtown Mobile and north of the Mobile Bay.

Figure 119 : A Comparison of Measured and Simulated High Water Mark Errors (in feet) in Mobile and Baldwin Counties during Hurricane Georges

The comparison error is computed as (simulated) - (measured) such that positive and negative values suggest an over prediction and under prediction of water levels, respectively. A datum discrepancy between measurements (NAVD88) and simulated high water marks (MSL) may introduce differences on the order of +/- 0.7 feet.

This figure shows a map of Mobile bay with color-coded dots indicating the difference between observed high water and those simulated by the ADCIRC model for Hurricane Georges. The ADCIRC model tended to slightly over-predict high water marks on Dauphin Island and the southern coast of Mobile.

As noted in the discussion above, there are measureable differences in both the comparison of storm surge hydrographs and HWMs for Hurricanes Georges and Katrina. These differences may be attributed to a number of simplifications, or assumptions, applied to the model scenarios or to deficiencies in the hydrodynamic model itself. These possible causes are described below:

Note that increases in global sea level will not necessarily cause a corresponding one-to-one increase in peak storm surge elevations at all locations due to such factors as: non-linear variations in the forces increasing storm surge (such as wind setup) and forces resisting storm surge (such as bottom friction).

D.8.4. Wave Modeling

The wave characteristics accompanying each of the storm surge scenarios were simulated using a state-of-the-art model, STeady State spectral WAVE (STWAVE). It is a flexible, robust model for nearshore wind-wave growth and propagation. STWAVE is a steady-state, finite difference, spectral model based on the wave action balance equation. STWAVE simulates depth-induced wave refraction144 and shoaling,145 current-induced refraction and shoaling, depth- and steepness-induced wave breaking, diffraction,146 wave growth based on wind input, and wave-wave interaction and white capping that redistribute and dissipate energy in a growing wave field. Recent upgrades to the model include wave-current interaction and steepness-induced wave breaking. STWAVE was written by the U.S. Army Corps of Engineers Waterways Experiment Station (USACE-WES). It is one of the most widely used models to compute waves in coastal environments, based on wind and bottom topography. Model details can be found in Smith et al. (2001).

For each scenario, the STWAVE model was run following the ADCIRC model. The coupling between the models was asynchronous. The wind fields used to drive STWAVE were derived from the Holland-type model that was used to drive the ADCIRC model. Waves were simulated over both open water and the land simulated by ADCIRC to be inundated.

Note that Dauphin Island currently helps to protect the mainland by attenuating waves generated out in the open Gulf. Some of that attenuation may be diminished if the topography of the island is reduced through erosion from prior storm wave action. Changes in morphology of the island were not taken into account in the simulations performed in this study.

Ideally, validation simulations would have been performed, as was done for the ADCIRC model. Unfortunately, wave data sufficient to compare to the STWAVE model are not available for hurricanes in the Mobile region. However, the STWAVE model has been extensively tested against observations in other contexts. It has proven to give relatively good approximations of wave characteristics under a variety of conditions. Documentation on tests of STWAVE run under a range of conditions can be found at the USACE Coastal and Hydraulics Laboratory site.147

D.8.5. Exposure Mapping

Finally, a Geographic Information System was used to overlay inundation under each of the storm surge scenarios on top of the locations of the critical assets defined in Task 1 of the Gulf Coast Study. The analysis took into account the specific elevations of land on which each asset sits, although it did not consider the height of each asset. Thus, an asset is considered "inundated" if its location is inundated, but the asset itself is not necessarily overtopped.

D.9. Supplementary Storm Surge Exposure Statistics

This appendix provides supplementary exposure statistics for the scenario-based analysis of future hurricane storm surge.

Table 54 : Supplementary Storm Surge Exposure Statistics for Georges and Katrina Natural Simulations, and Georges Sea-Level Rise Simulations

Scenario
Mode Asset Criticality Georges-Natural Katrina-Natural Georges-Natural-30cm Georges-Natural-75cm Georges-Natural-200cm
Highways Roads (mi) Critical 55 of 209 (27%) 58 of 209 (28%) 58 of 209 (28%) 63 of 209 (30%) 83 of 203 (40%)
Not Critical 7 of 284 (2%) 8 of 284 (3%) 8 of 284 (3%) 11 of 284 (4%) 20 of 284 (7%)
Evacuation Routes (mi) Not Critical 35 of 367 (10%) 38 of 367 (10%) 38 of 367 (10%) 46 of 367 (12%) 71 of 367 (19%)
Rail Rail (mi) Critical 111 of 196 (57%) 116 of 196 (60%) 114 of 196 (59%) 119 of 196 (61%) 132 of 196 (68%)
Not Critical 31 of 118 (26%) 31 of 118 (26%) 31 of 118 (26%) 32 of 118 (27%) 41 of 118 (35%)
Rail Points (#) Critical 4 of 5 (80%) 4 of 5 (80%) 4 of 5 (80%) 4 of 5 (80%) 5 of 5 (100%)
Not Critical 10 of 12 (83%) 10 of 12 (83%) 10 of 12 (83%) 10 of 12 (83%) 10 of 12 (83%)
Pipelines Pipelines (mi) Critical 14 of 426 (3%) 15 of 426 (3%) 15 of 426 (3%) 24 of 426 (6%) 50 of 426 (12%)
Not Critical 10 of 207 (5%) 10 of 207 (5%) 10 of 207 (5%) 13 of 207 (6%) 33 of 207 (16%)
Ports Ports (#) Critical 24 of 26 (92%) 24 of 26 (92%) 24 of 26 (92%) 24 of 26 (92%) 24 of 26 (92%)
Not Critical 43 of 48 (90%) 44 of 48 (92%) 44 of 48 (92%) 44 of 48 (92%) 44 of 48 (92%)
Transit Facilities Critical 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%)
SDE Facilities (#) Not Critical 29 of 193 (15%) 32 of 193 (17%) 32 of 193 (17%) 56 of 193 (29%) 73 of 193 (38%)
Bus Stops (#) Not Critical 64 of 907 (7%) 81 of 907 (9%) 69 of 907 (8%) 147 of 907 (16%) 209 of 907 (23%)
Bus Routes (mi) Not Critical 10 of 126 (8%) 11 of 126 (8%) 10 of 126 (8%) 15 of 126 (12%) 23 of 126 (18%)
MODA Stops (#) Not Critical 2 of 22 (9%) 6 of 22 (27%) 4 of 22 (18%) 20 of 22 (91%) 22 of 22 (100%)
Bike Routes (mi) Not Critical 15 of 132 (11%) 16 of 132 (12%) 16 of 132 (12%) 20 of 132 (15%) 31 of 132 (24%)
Airports Mobile Downtown Airport (mi2) Critical 0 of 3 (4%) 0 of 3 (5%) 0 of 3 (5%) 0 of 3 (7%) 0 of 3 (15%)
Mobile Regional Airport (mi2) Critical 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%)
Other Medical Facilities (#) Not Critical 0 of 45 (0%) 0 of 45 (0%) 0 of 45 (0%) 0 of 45 (0%) 4 of 45 (9%)

*Exposure statistics reflect the percent of the assets in the exposure zone.  These statistics do not necessarily represent the assets that are actually overtopped by storm surge.

Table 55 : Supplementary Storm Surge Exposure Statistics for Katrina Sea-Level Rise and Shifted Simulations

Scenario
Mode Asset Criticality Katrina-Natural-75cm Katrina-Shift Katrina-Shift-75cm Katrina-Shift-MaxWind Katrina-Shift-MaxWind-75cm Katrina-Shift-ReducedPress-75cm
Highways Roads (mi) Critical 69 of 209 (33%) 95 of 209 (46%) 114 of 209 (55%) 140 of 209 (67%) 149 of 209 (75%) 124 of 209 (60%)
Not Critical 15 of 284 (5%) 23 of 284 (8%) 29 of 284 (10%) 39 of 284 (14%) 41 of 284 (15%) 33 of 284 (12%)
Evacuation Routes (mi) Not Critical 52 of 367 (14%) 87 of 367 (24%) 107 of 367 (29%) 144 of 367 (39%) 157 of 367 (43%) 123 of 367 (33%)
Rail Rail (mi) Critical 127 of 196 (65%) 140 of 196 (71%) 144 of 196 (73%) 150 of 196 (77%) 154 of 196 (79%) 146 of 196 (74%)
Not Critical 33 of 118 (28%) 44 of 118 (37%) 56 of 118 (47%) 78 of 118 (66%) 83 of 118 (70%) 65 of 118 (55%)
Rail Points (#) Critical 5 of 5 (100%) 5 of 5 (100%) 5 of 5 (100%) 5 of 5 (100%) 5 of 5 (100%) 5 of 5 (100%)
Not Critical 10 of 12 (83%) 11 of 12 (92%) 11 of 12 (92%) 12 of 12 (100%) 12 of 12 (100%) 11 of 12 (92%)
Pipelines Pipelines (mi) Critical 44 of 426 (10%) 51 of 426 (12%) 54 of 426 (13%) 62 of 426 (15%) 67 of 426 (16%) 56 of 426 (13%)
Not Critical 25 of 207 (12%) 36 of 207 (17%) 38 of 207 (18%) 47 of 207 (23%) 49 of 207 (24%) 39 of 207 (19%)
Ports Ports (#) High 24 of 26 (92%) 24 of 26 (92%) 25 of 26 (96%) 26of 26 (100%) 26 of 26 (100%) 25 of 26 (96%)
Not Critical 44 of 48 (92%) 48 of 48 (100%) 48 of 48 (100%) 48 of 48 (100%) 48 of 48 (100%) 48 of 48 (100%)
Transit Facilities Critical 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%) 1 of 2 (50%)
SDE Facilities (#) Not Critical 66 of 193 (34%) 86 of 193 (45%) 105 of 193 (54%) 131 of 193 (68%) 140 of 193 (73%) 115 of 193 (60%)
Bus Stops (#) Not Critical 171 of 907 (19%) 286 of 907 (32%) 374 of 907 (41%) 622 of 907 (69%) 653 of 907 (72%) 483 of 907 (53%)
Bus Routes (mi) Not Critical 18 of 126 (14%) 32 of 126 (26%) 46 of 126 (36%) 78 of 126 (62%) 83 of 126 (65%) 60 of 126 (47%)
MODA Stops (#) Not Critical 20 of 22 (100%) 22 of 22 (100%) 22 of 22 (100%) 22 of 22 (100%) 22 of 22 (100%) 22 of 22 (100%)
Bike Routes (mi) Not Critical 23 of 132 (18%) 40 of 132 (30%) 53 of 132 (40%) 67 of 132 (51%) 69 of 132 (53%) 59 of 132 (45%)
Airports Mobile Downtown Airport (mi2) Critical 0 of 3 (9%) 2 of 3 (65%) 2 of 3 (90%) 3 of 3 (100%) 3 of 3 (100%) 3 of 3 (98%)
Mobile Regional Airport (mi2) Critical 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%) 0 of 4 (0%)
Other Medical Facilities (#) Not Critical 2 of 45 (4%) 7 of 45 (16%) 14 of 45 (31%) 17 of 45 (38%) 18 of 45 (40%) 14 of 45 (31%)

*Exposure statistics reflect the percent of the assets in the exposure zone.  These statistics do not necessarily represent the assets that are actually overtopped by storm surge.

D.10. Caveats, Gaps, and Replicability of Storm Surge Analysis

This appendix discusses the assumptions and simplifications used in this study's analysis of future storm surge, as well as lessons learned that could be useful in extending the results to other locations.

D.10.1 Assumptions and Simplifications

Not all factors affecting storm surge were taken into account in this study. For example, the study did not account for river flooding that often accompanies strong storms and tends to contribute to storm surge. Nor did it account for changes in beach profiles.

River Flooding

In an estuary such as Mobile Bay, river flooding that often accompanies strong storms will tend to contribute to storm surge. Riverine inputs to the northern head of the Bay can increase significantly during heavy precipitation events. To help minimize computational complexity, river flooding was not directly accounted for in this study's numerical simulations. As a result, the results presented in this study will tend to be lower bounds for surge levels near rivers.

Changes in Beach Profiles

If any of the surges associated with the shifted Katrina scenarios were to strike the Mobile area, they would lead to significant changes in beach profiles. For example, the larger surges would batter Dauphin Island, perhaps opening new cross-island channels, or even removing significant parts of the island. The simulations presented here are most robust for the present state of the shoreline. In a subsequent study, it would be useful to simulate the erosional and depositional effects of major storms and long-term sea level rise on the morphology of the barrier island using a model such as XBeach. The surge and wave modeling used in this study could then be repeated to assess the impacts of the changes in coastal morphology on storm surge and waves on both the island and the mainland.

D.10.2 Lessons Learned

In the process of conducting this series of storm surge and wave modeling simulations, several lessons were learned that may be useful in extending the results from this study to other locations while minimizing resource requirements.


1 In the western-most areas of Mobile County and the western end of Dauphin Island, actual data values were extended outward into data voids to build the interpolation surface.

2 LIDAR data provided by the City of Mobile, 2010

3 ESA, 2012

4 ERSDAC, 2006

5 The horizontal datum is the WGS84.

6 Interpolated from vertical change rate surface developed from vertical change rates of 1969-2009 and 1984-2010 benchmark surveys.

7 Interpolated from vertical change rate surface developed from ERS vertical change rates from 25 satellite images between July 1992 and December 1999.

8 Interpolated from vertical change rate surface developed from PALSAR vertical change rates from 13 satellite images between June 2007 and August 2010.

9 BM about 1.2 miles (1.9 kilometers) south of Alabama State Docks tide gage at Mobile, which has an estimated vertical rate of 0.02 inches/year (0.5 millimeters/year).

10 BM near Dauphin Island tide gage, which has an estimated vertical change rate of -0.5 inches/year (-1.2 millimeters/year).

11 http://oss.deltares.nl/web/xbeach/

12 http://warrenpinnacle.com/prof/SLAMM/index.html

13 Bromirski et al., 2011

14 See, for example, figure 5.15a and 5.16b from the IPCC Fourth Assessment Working Group 1 report, which show the regional distribution of decadal and longer-term sea level rise trends: http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-chapter5.pdf

15 Please see the U.S. Global Change Research Program's web page on the National Climate Assessment: http://www.globalchange.gov/what-we-do/assessment/nca-overview

16NWS, 2009a. NWS Forecast office of Mobile/Pensacola analysis of this storm event.

17 NWS, 2009a

18 In these situations, convective activity may intensify if there is a layer of stable cold air above. However, the November 15 soundings of Birmingham, Alabama did not demonstrate the existence of a layer of stable cold air above (NOAA, 2011i). Soundings which demonstrate how temperature and humidity change from the surface to aloft are generated by NOAA website of archived meteorological data.

19 NCDC, 2011a. Meteorological storm data was collected from the National Climatic Data Center which archives the National Weather Service Cooperative Observer Program observational meteorological data and includes hourly precipitation.

20 USGS, 2011a,b,c. United States Geological Survey (USGS) provides hourly discharge information by stream site within the United States.

21 NCDC, 2011b. Meteorological storm data was collected from the National Climatic Data Center which archives the National Weather Service Cooperative Observer Program observational meteorological data and includes surface temperature, wind, and surface pressure.

22 NCDC, 2011b

23 Ibid.

24 NOAA, 2011a

25 Ibid.

26 NOAA, 2011c

27 NWS, 2009a

28 Ibid.

29 Ibid.

30NWS, 2009b. Forecast office of Mobile/Pensacola analysis of this storm event.

31NWS, 2009b. Also evident was a 'dip' in the zonal air flow over Arkansas and Louisiana at 700 mb (air situated at 700 mb is between the surface and 500 mb).

32 NCDC 2011a

33Ibid.

34 NWS 2009b. Leakesville, Mississippi is about 60 miles northwest of Mobile, Alabama.

35 Ibid.

36 USGS, 2011a, b, c

37 NCDC, 2011b

38 Ibid.

39 Ibid.

40 NCDC, 2011b

41 NOAA, 2011c

42 Mean high-high water level is the average of the higher high water height of each tidal day observed over the National Tidal Datum Epoch (tidesandcurrents.noaa.gov/datum-options.html)

43 NWS, 2011b

44 Ibid.

45 Ibid.

46 NWS, 2011c

47 Ibid.

48 NOAA, 2011e

49 NOAA, 2001f

50 Ibid.

51 Ibid.

52 USGS, 2011a

53 NCDC, 2011a

54 Ibid.

55 Ibid.

56 NCDC, 2011a

57 Ibid.

58 NOAA, 2011a

59 Ibid.

60 NOAA, 2011b

61 NWS, 2011c

62 CNN, 2008

63 Smith, 2008

64 Gordon, 2008

65 U.S. Army Corps of Engineers, 1999

66 Ibid.

67 United States Department of the Interior, 2000

68 Ibid. The sea surface temperatures were averaged from Sea-Viewing Wide field-of Sensor (seaWiFS) satellite data.

69 NASA, 2003, Recipe for a Hurricane. http://www.nasa.gov/vision/earth/environment/HURRICANE_RECIPE.html

70 Though Biloxi is just 60 miles from Mobile, they have different shoreline characteristics. Biloxi sits directly on the Gulf of Mexico, while Mobile is inset on Mobile Bay, with some barrier islands between the Gulf and the inlet. The differences may affect storm surge and so the locations are considered separately in this analysis.

71 U.S. Army Corps of Engineers, 1999

72 NOAA, 2011g

73 NOAA, 2011h

74 U.S. Army Corps of Engineers, 1999

75 Ibid.

76 Ibid.

77 USGS, 2011a, b, c

78 U.S. Army Corps of Engineers, 1999

79 Ibid.

80 Ibid.

81 U.S. Army Corps of Engineers, 1999

82 Guiney, 1999.

83 U.S. Army Corps of Engineers, 1999. All storm surge data presented here is referenced to the National Geodetic Vertical Datum (NGVD) of 1929. FEMA requested COE use this datum which is the same used in the construction of the topographic charts published by USGS

84 Ibid.

85 Ibid.

86 Ibid.

87 Ibid.

88 U.S. Army Corps of Engineers, 1999

89 Ibid.

90 NOAA, 2005a

91 NOAA, 2005b

92 Ibid.

93 NOAA, 2005b. The remainder (?) of the "Storm track and Intensification" draws from this source.

94 NOAA, 2005a

95 Knabb et al., 2006

96 Ibid.

97 USGS, 2011a

98 Knabb et al., 2006

99 Ibid, Winds become dangerous to road maintenance, truck operations, and other road users at around 39 mph and are very dangerous at 74 mph (OFCM, 2002).

100 Knabb et al., 2006

101 Ibid.

102 FEMA, 2006b. All observations discussed here are referenced to the North American Vertical Datum of 1988

103 Ibid.

104 Ibid.

105 Ibid.

106 FEMA, 2006a

107 Ibid.

108 Kieper, 2011. The highest value ever recorded was 11.60 feet on 5 July 1916.

109 Ibid.

110 Knabb et al., 2006

111 USGCRP, 2009

112 USGCRP, 2009

113 USGCRP, 2009

114 USCCSP, 2008b

115 Knutson et al., 2010

116 Bender et al., 2010

117 USCCSP, 2008b

118 USCCSP, 2008b

119 WMO, 2006

120 NRC, 2010c

121 CCSP, 2008

122 Karl et al., 2008

123 Note that the mapping of potential inundation due to long-term sea level rise is conducted relative to Mean Higher High Water (MHHW) whereas the ADCIRC results are shown relative to Mean Sea Level (MSL). We show long-term inundation relative to MHHW so that the impacts above the tidal cycle are evident. We show the short-term flooding results relative to MSL since we do not know what the tidal stage of any particular future storm will be. MHHW is 0.77 feet above MSL at Mobile State Docks.

124 The two questions being address are: (1) What are the implications of a moderate hurricane striking the region with a higher sea level? (2) What are the implications of a strike by a larger hurricane than the region has experienced in recent history?

125 The term "shift" indicates an eastward shift of the storm track. This is used to explore the potential for a direct hit of a major hurricane on the Mobile area.

126 The ReducedPress, MaxWind, and shift scenarios were applied only to Katrina.

127 The intensity of a hurricane is defined in part by its central pressure. The lower the central pressure, the more intense it is.

128 A 1%/year increase in atmospheric CO2 concentration corresponds roughly to the IPCC B1 scenario through approximately 2060 (Goose et al. 2010).

129 Luettich et al., 1992; Luettich and Westerink, 2004; Westerink et al., 1994

130 ADCIRC, 2011

131 Mattocks and Forbes, 2008; Mattocks et al., 2006

132 Holland, 1980

133 The model computes all parameters.

134 After Blain et al., 2007

135 Knabb et al., 2006

136 USDOC, 2005

137 The HWM for a particular storm is distinct from other regularly occurring tidal indices. For example, the Mean Higher High Water mark corresponds to the average of the higher high water height of each tidal day observed over the National Tidal Datum Epoch.

138 NOAA, 2011

139 The basis for this error analysis is provided in a report produced by South Coast Engineers for ICF titled: "Hydrodynamic Model Testing and Validation for Two Historical Storms: Hindcast Simulations of Hurricanes Katrina and Georges." It is available upon request from ICF.

140 When a wave breaks against the shore it runs a distance horizontally up the beach slope.

141 Dean and Dalrymple, 1991

142 FEMA, 2000

143 Holland, 1980; Mattocks and Forbes, 2008; Mattocks et al., 2006. An asymmetric two-dimensional wind field formulation such as this one is more realistic than a circular one. The wind field does not account for the range of factors that can contribute to spatial inhomogeneities (e.g., squall lines; temporary lulls) that arise in typical storms. The aggregate effect of these factors is, however, relatively well represented.

144 Refraction refers to the bending of a wave due to a change in speed along the length of the wave.

145 Shoaling refers to the interaction between the seafloor and the wave.

146 Diffraction refers to the bending of waves around obstacles in their path.

147 USACE, 2011. Results from Ponce de Leon Inlet, Florida, Willapa Bay, Washington, and Grays Harbor, Washington are available at: http://chl.erdc.usace.army.mil/chl.aspx?p=s&a=SOFTWARE;9

Updated: 10/31/2014
HEP Home Planning Environment Real Estate
Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000