Hurricane Storm Surge—Complex Dynamics, High Losses
Apr 23, 2015
Editor's Note: Storm surge—an abnormal and often substantial increase in sea level—frequently accompanies a low-pressure storm system, such as a hurricane, as the storm moves onshore. Potentially more devastating than wind, this high-loss sub-peril can impact multiple lines of business and account for a significant portion of total hurricane losses.
Over time, storm surge has proven to be a critical source of hurricane losses. AIR's enhanced storm surge module in the 2015 U.S. hurricane model, scheduled for release this summer, shows that when historical hurricanes are modeled at today's exposures and sea levels, storm surge loss to all properties—from an average annual loss perspective and before any insurance coverage is applied—equals approximately one-third of total loss, even for events at different loss levels.
The severity of storm surge depends on a combination of variables, including storm size, strength, speed, and path; tidal height and cycle; and coastal topography, bathymetry, and geology. The destructive impact of storm surge can extend from the immediate coast to many miles inland (in low-lying areas of Louisiana, for example), with steeper slopes and rougher terrain onshore causing more rapid surge attenuation than gradual slopes and smoother terrain.
In addition to the damage caused by inundation, storm surge also causes damage from the force of moving water, debris collision, sedimentation, and contamination. Storm surge can be deadly, too: According to the Federal Emergency Management Agency (FEMA), storm surge presents the greatest potential for loss of life from hurricane-related hazards.1
Dynamics of the Storm Surge Hazard
Within the chaotic hurricane environment, two basic, interrelated mechanisms cause storm surge: differences in atmospheric pressure and the winds that are generated by these differences.2 These fundamental processes interact with the ocean's surface in different ways. Wind generally results in a much greater increase in surge elevation than the pressure gradient in near-shore regions:
- Wind applies stress to the ocean's surface, generating a current and driving water approximately parallel to the direction of the wind, which results in a rise in the ocean's surface when the current interacts with the seafloor and the coastline.
- The atmospheric pressure gradient has a slightly more complex interaction with the ocean's surface. Because lower pressure is being applied to the ocean's surface at the center of the storm than at the periphery, the ocean's surface is higher at a storm's center. This impact on sea level, known as the "inverted barometer effect," is analogous to pressing down on a mattress with a large ring: Under the ring the mattress is lower because more pressure is being applied (as with a storm periphery), while in the center of the ring the mattress is higher (as in the center of a storm).3
Although atmospheric pressure and wind generate storm surge, a number of other variables modify the intensity and/or impact of storm surge. As a result, no highly accurate empirical relationship exists between wind intensity and water level because of the hugely important role coastal geography and other physical processes play in modifying the wind-driven surge. The shape of the coastline in the vicinity of landfall—as well as the shape and depth of the seafloor (bathymetry) and the shape and height of the coast (topography)—effectively determine the elevation of storm surge for a given event. In general, surge elevation will be higher in areas where the sea is shallower (e.g., continental shelves, large inland lakes, and bays) and in areas with concave coastlines (where surge can be contained by the coast) as opposed to straight coastlines. (See Figure 2.)
The physical characteristics of a given hurricane play a significant role in the intensity of storm surge as well. For example, an intense, fast-moving hurricane may have higher maximum winds than a slowly moving, more moderate hurricane, but the slower-moving hurricane might generate a higher surge due to the longer period of time winds interact with the ocean's surface. In addition, the angle at which a storm approaches the coastline can significantly alter the elevation of surge, particularly in areas of complex coastal geography. The overall size of a hurricane—with regard to wind field width and the distance from the center of the storm to the maximum winds—has an enormous impact on where the maximum surge will occur and what the extent of total surge inundation will be.
Even though the generation of storm surge is a purely meteorological process, storm surge can be greatly modified by the astronomical effects of the sun and moon via the tides. The roughly six-hour difference between peak storm surge striking at high tide versus low tide can be the difference between a zero-surge-loss event and a catastrophic-surge-loss event. The role tides play in modifying storm surge is dependent on geographic location, of course, as the difference between high and low tide (tidal ranges) varies greatly along the Gulf and East coasts, from under 2 feet in Texas to nearly 20 feet in Maine.
By their nature, all these surge modifiers interact with one another, resulting in an incredibly complex and dynamic system. To properly understand and predict surge, scientists and engineers use powerful numerical models to solve the equations that govern surge generation, modification, and dissipation. Which model and set of techniques are selected will depend on the ultimate use case. Some models are designed to understand small-scale surge processes that occur during the course of a single event; these require large, high-performance computing resources. Other, less computationally intensive models allow modelers to simulate thousands of events to better understand the larger-scale behavior of surge in a region. Regardless of the model, the modeler must use an accurate representation of a hurricane's meteorology and physical characteristics—as well as precise depictions of the coastline's geography, bathymetry, and topography—to effectively model storm surge.
Key Variables Affecting Loss Potential
Surge losses are a consequence of water being driven inland and inundating normally dry areas where exposure is found. The extent of inundation depends on the coastal topography, so flat coastal floodplains, such as those in eastern Texas and Louisiana, experience more inundation than steeper coastal areas, such as those in the Northeast. Additionally, areas of high coastal exposure concentration, such as large cities on the coast, are susceptible to large losses from moderate surge events due to the large accumulated value over a small area.
To demonstrate the influence of geographic location and exposure concentration on storm surge losses, we selected simulated hurricanes from the AIR model's catalog that have roughly comparable meteorological and physical parameters but that make landfall in different locations along the coast. Table 1 shows the narrow range of meteorological parameters used to select the example events.
|Parameters (Units)||Lower Limit||Upper Limit|
|Central pressure (mb)||965||970|
|Forward speed (mph)||14||16|
|Radius to maximum winds (miles)||20||25|
These parameters produced the events illustrated in 14 matching events from AIR's U.S. tropical cyclone catalog (see Figure 3). Although the search parameters are somewhat region-specific (Northeast hurricanes have fundamentally different characteristic than Gulf hurricanes), these events are realistic representations of landfalling tropical cyclones from Texas to North Carolina.5
Despite the similar event parameters and even similar landfall locations in some cases, losses for these events are very different, ranging from USD 21 million to nearly USD 5 billion, with a mean of about USD 1 billion.
Land inundation can be estimated for the 14 storms we selected by computing the area above sea level that is covered with water at maximum surge. Similar to loss figures, these numbers have a broad range: from 34 square miles (87 square kilometers), which is approximately half the size of Washington, D.C., to 5,896 square miles (15,271 square kilometers), slightly larger than the state of Connecticut.
Comparing losses to inundation area does not reveal a strong correlation, as shown by Figure 4.
When comparing the relationship between inundation area and losses in Figure 4 to landfall location in Figure 3, the roles of both coastal geography and exposure concentration become apparent. The four events with the largest amount of inundation—N, L, C, and I—all impact Louisiana and its extensive low-lying coastal wetlands. Because of their inaccessibility and natural vulnerability to storm surge, the wetlands of southern Louisiana contain little exposure and therefore are subject to relatively small losses, despite widespread impact from surge. Conversely, the top three loss events—K, M, and G—all have moderate inundation areas, but impact the highly developed Florida coastline. South Florida, while not as low-lying as southern Louisiana, has a low enough elevation in most parts to allow surge to penetrate inland, causing catastrophic losses due to the dense development and urbanization of many parts of the coastline.
The variability in both storm surge losses and storm surge inundation that is demonstrated from this relatively homogeneous set of hurricanes underscores the primary importance of the geographic context of the event over the existence of the event. A Category 3 hurricane that stays offshore has a very different impact than if the same storm were to make landfall in Miami. As evidenced here, a complex relationship between surge, geography, and losses exists. Events of similar intensity generate varying levels of surge inundation due to the local coastal geography, which in turn produces varying levels of losses due to the local concentration of exposure. The fundamental surge generation processes of winds and pressure interacting with the ocean's surface is the same in every hurricane, but surge interacting with the natural and built environment is what determines a surge's local impact.
Know the Risk to Own the Risk
Since 1900, many hurricanes have manifested with substantial, destructive, and deadly storm surges. For example, the Galveston Hurricane in 1900 generated a 15-foot storm surge, which destroyed thousands of homes and severed communication lines. More recently, the storm surge from Hurricane Katrina in 2005 resulted in more than 50 levee breaches, while the storm surge from Hurricane Sandy in 2012 caused unprecedented coastal flooding in New York and New Jersey (even though Sandy had been reclassified as an extratropical storm by landfall).
Understanding the potentially devastating impact of storm surge can help companies effectively plan for this serious risk. The new, hydrodynamic storm surge component of the updated AIR Hurricane Model for the United States provides a greatly refined and thoroughly validated view of storm surge loss potential, giving companies broad, accurate, and reliable storm surge information.
Although man-made defenses can help reduce the impact of storm surge, the risk of damage from this peril remains. In fact, in many locales the risk of insured loss is growing due to continued development along the coast and projected sea-level rise, a topic addressed in the second article in this storm surge issue, "The Growing Value of U.S. Coastal Property at Risk."
The final article in this special issue, "Insights into Storm Surge Vulnerability," discusses the importance of understanding the many aspects of vulnerability of U.S. building stock, as well as specialty lines, such as automobiles and pleasure boats, to storm surge to effectively manage U.S. hurricane risk.
2 Hurricane intensity relates to the difference between the lower pressure at the center of the storm and the higher pressure on the periphery: the greater the difference the stronger the winds.
3 The magnitude of the inverse barometer effect is a rise in ocean surface of approximately 2 inches for every 5 millibar drop in atmospheric pressure.
4 This range of central pressure represents approximately a Category 2 hurricane. See Holland, Greg, 2008. "A Revised Hurricane Pressure-Wind Model," Monthly Weather Review: http://dx.doi.org/10.1175/2008MWR2395.1.
5 Northeast cyclones have much larger radius to maximum winds than Gulf of Mexico cyclones, hence it's tough to choose a narrow set of parameters that will capture both Gulf and Northeast events.