Jill C. Trepanier

Combining Surge and Wind Risk from Hurricanes Using a Copula Model: An Example from Galveston, Texas

Consideration of climate-related impacts on coasts is important to ensure readiness for disaster response. Local risk of storm surge and strong winds from hurricanes affecting Galveston, Texas, is quantified using a bivariate copula model fit to observed data. The model uses a two-dimensional Archimedean copula. Parametric uncertainty (5th and 95th percentiles) is quantified using a Monte Carlo procedure. The annual probability of a hurricane producing winds of at least 50 ms−1 and a surge of at least 4 m is 1.7 percent with a 95 percent confidence interval of (1.33, 1.78) percent. The methodology can be extended to include inland flooding and can be applied elsewhere with available information.

Spatiotemporal patterns of extreme hurricanes impacting US coastal cities

US coastal cities are regularly subjected to destruction by tropical cyclones. The risk of tropical cyclone winds varies along the length of the coastline. We analyze landfalling North Atlantic basin tropical cyclones whose intensities are considered extreme relative to their landfall location. To be considered extreme, a tropical cyclone’s wind speed must exceed the 50-year return level for a given city. Of interest is the spatial and temporal patterns of these extreme hurricane wind events for fifteen coastal cities, which are organized into four coastal regions: Northeast Atlantic, Southeast Atlantic, Florida, and Gulf. Findings suggest that extreme hurricanes along the Florida and Atlantic coasts cluster in time, specifically decades, while there is no temporal clustering detected along the Gulf. Atlantic coast hurricane clusters are in part due to the likelihood of one intense hurricane impacting multiple coastal cities, which is unlikely to happen along the Gulf due to the alignment of the coast. It is also unlikely for an intense hurricane to impact multiple Florida cities as an extreme hurricane, suggesting a physical mechanism enables the temporal clustering seen here. The results of this work advocate for annual and decadal hurricane risk to include: (1) the likelihood of temporal clusters of extreme hurricanes along the Atlantic and Florida coasts and (2) extreme hurricanes impacting multiple cities along the Atlantic coast.

Using Synthetic Tropical Cyclones to Characterize Extreme Hurricanes Affecting Charleston, South Carolina

AbstractThe characteristics and conditions favoring extreme hurricanes remain largely unknown because of their small number in the observational record. Synthetic tracks are capable of providing a large, representative sample of these events, which provides an opportunity to further understanding of extreme characteristics as compared with those of more common tropical cyclones. The authors compare 300 synthetic extreme (100-yr event, ≥48.9 m s−1) and 300 common (5-yr event, ≤33.6 m s−1) tropical cyclones for Charleston, South Carolina, for differences in spatial, temporal, and other characteristics. Results suggest that extreme hurricanes have a more-defined spatial and temporal behavior, generally forming off the coast of Africa and making a direct landfall at Charleston. Common tropical cyclones sometimes make prior landfalls, may approach from either the Gulf of Mexico or the Atlantic Ocean, and often decay well before reaching Charleston. They are likely to occur through much of the hurricane season,...

Hurricane risk variability along the Gulf of Mexico Coastline

Hurricane risk characteristics are examined across the U. S. Gulf of Mexico coastline using a hexagonal tessellation. Using an extreme value model, parameters are collected representing the rate or λ (frequency), the scale or σ (range), and the shape or ξ (intensity) of the extreme wind distribution. These latent parameters and the 30-year return level are visualized across the grid. The greatest 30-year return levels are located toward the center of the Gulf of Mexico, and for inland locations, along the borders of Louisiana, Mississippi, and Alabama. Using a geographically weighted regression model, the relationship of these parameters to sea surface temperature (SST) is found to assess sensitivity to change. It is shown that as SSTs increase near the coast, the frequency of hurricanes in these grids decrease significantly. This reinforces the importance of SST in areas of likely tropical cyclogenesis in determining the number of hurricanes near the coast, along with SSTs along the lifespan of the storm, rather than simply local SST. The range of hurricane wind speeds experienced near Florida is shown to increase with increasing SSTs (insignificant), suggesting that increased temperatures may allow hurricanes to maintain their strength as they pass over the Florida peninsula. The modifiable areal unit problem is assessed using multiple grid sizes. Moran’s I and the local statistic G are calculated to examine spatial autocorrelation in the parameters. This research opens up future questions regarding rapid intensification and decay close to the coast and the relationship to changing SSTs.

Hurricane winds over the North Atlantic: spatial analysis and sensitivity to ocean temperature

Hurricanes pose serious threats to people and infrastructure along the United States Gulf and Atlantic coasts. The risk of the strongest hurricane winds over the North Atlantic basin is analyzed using a statistical model from extreme value theory and a tessellation of the domain. The spatial variation in model parameters is shown, and an estimate of the limiting strength of hurricanes at locations across the basin is provided. Quantitative analysis of the variation is done using a geographically weighted regression with regional sea surface temperature as a covariate. It is found that as sea surface temperatures increase, the expected hurricane wind speed for a given return period also increases.

Recording Tropical Cyclone Activity from 1909 to 2014 along the Northern Gulf of Mexico using Maritime Slash Pine Trees (Pinus elliottii var. elliottii Engelm.)

ABSTRACT Tucker, C.S.; Trepanier, J.C.; Harley, G.L., and DeLong, K.L., 2018. Recording Tropical Cyclone Activity from 1909 to 2014 along the Northern Gulf of Mexico using Maritime Slash Pine Trees (Pinus elliottii var. elliottii Engelm.). The temporally incomplete tropical cyclone (TC) observational record limits the understanding of recurrence intervals and the impacts of repeated TCs on coastal ecosystems. Growth of maritime trees near the Gulf of Mexico may be affected by high winds, precipitation, and storm surges from TCs. Proxy records, such as tree growth recorded in annual ring widths, can be used to extend TC records temporally but must be verified with the observational record first. This study develops chronologies of total ring width (TRW), earlywood ring width (ERW), and latewood ring width (LRW) from slash pine (Pinus elliottii var. elliottii Engelm.) trees located 2 km from an open saltwater bay in the Grand Bay National Estuarine Research Reserve in coastal Mississippi, U.S.A. These chronologies are compared with TC storm surge and wind speed records using superposed epoch analysis. The year after a TC occurrence, tree growth is significantly (p < 0.05) suppressed (narrower than average annual growth rings) for TRW, ERW, and LRW with respect to storm surge and shows no statistical significance with high wind speeds. Individual trees display suppression in growth for 1–6 years after TC occurrences. This study demonstrates that slash pine trees in close proximity to salt water can be used to produce chronologies of TC occurrences for the northern Gulf of Mexico region and thus can eventually be used to extended and supplement temporally the existing TC observational record.

Understanding the Influence of Tropical Cyclone Landfall Central Pressure and Accumulated Rainfall on Storm Surge Near New Orleans, Louisiana

ABSTRACT Trepanier, J.C.; Needham, H.F., and Ellis, K.N., 2018. Understanding the influence of tropical cyclone landfall central pressure and accumulated rainfall on storm surge near New Orleans, Louisiana. Tropical cyclone track, landfall central pressure, accumulated cyclone rain, and storm surge are assessed for five stations surrounding Lake Pontchartrain in Louisiana: Shell Beach, the New Orleans Lakefront, Frenier, Mandeville, and Slidell. Floods from 1901–2012 are analyzed for cases in which the event tracks, the landfall central pressure, the amount of inland flooding from precipitation, and the storm-surge inundation all lead to compound effects, as in Hurricane Katrina (2005) and Hurricane Isaac (2012). Data are from U-Surge, the National Hurricane Center, and the National Weather Service Cooperative Observer Program. A storm track climatology is provided showing the average track for all events and for events that reached at least 43 m s−1. The intense storms have a more concentrated track and approach from the south. Correlation and multivariate linear regression were used to evaluate the relationship among landfall central pressure, accumulated rainfall, and storm surge. For every 1 mbar increase in pressure when rainfall was held constant, the storm surge decreased by 0.024 m (p value <0.000). For every 1 mm of increase in rainfall when pressure was held constant, the storm surge increased by 0.006 m (0.0001). These two variables explain 63% of the variability in storm surge at this station. A percentage of change analysis from quantile regression was used to understand the amount of change in the storm surge based on the amount of change in the rainfall or the landfall central pressure. A small change in rainfall amounts or landfall central pressure can lead to a large change in the storm surge.

Using Proxy Records to Document Gulf of Mexico Tropical Cyclones from 1820-1915

Observations of pre-1950 tropical cyclones are sparse due to observational limitations; therefore, the hurricane database HURDAT2 (1851–present) maintained by the National Oceanic and Atmospheric Administration may be incomplete. Here we provide additional documentation for HURDAT2 from historical United States Army fort records (1820–1915) and other archived documents for 28 landfalling tropical cyclones, 20 of which are included in HURDAT2, along the northern Gulf of Mexico coast. One event that occurred in May 1863 is not currently documented in the HURDAT2 database but has been noted in other studies. We identify seven tropical cyclones that occurred before 1851, three of which are potential tropical cyclones. We corroborate the pre-HURDAT2 storms with a tree-ring reconstruction of hurricane impacts from the Florida Keys (1707–2009). Using this information, we suggest landfall locations for the July 1822 hurricane just west of Mobile, Alabama and 1831 hurricane near Last Island, Louisiana on 18 August. Furthermore, we model the probable track of the August 1831 hurricane using the weighted average distance grid method that incorporates historical tropical cyclone tracks to supplement report locations.