Friday, October 7, 2016

The "Protected Coast" is Now the Most Dangerous Place of All

Snapshot: Many people consider the stretch of coast from northeast Florida through Southern South Carolina, including the entire coast of Georgia, as a "protected coast," immune from coastal hazards. Hurricane Matthew now threatens to slam this coastline with a storm surge more severe than anywhere else the storm has impacted. The "Protected Coast" is now the most dangerous place of all, as Matthew turns back the time and delivers a knock-out punch similar to the vicious hurricanes that struck this region in the late 1800s.

The "Protected Coast"

Many people consider the stretch of coast from northeast Florida through Southern South Carolina, including the entire coast of Georgia, as a "protected coast," immune from coastal hazards.

And with good reason.

The coastal profile in this region bends quite smoothly in a concave fashion, creating a broad basin of water that extends west from the Gulf Stream Current. This concave coast does indeed appear protected compared to more vulnerable areas to the north and south.

Many people consider the concave-shaped coastline from northeast Florida to southwest South Carolina, including the entire coast of Georgia, a "protected coast," immune to coastal hazards.

To the north, the coastline of North Carolina juts out into the Atlantic, exposing it to more hurricane strikes. To the south, the Florida Peninsula reaches out towards Cuba, extending itself into a hurricane highway that wants to connect the Atlantic to the Gulf of Mexico. But in between these areas is a protected enclave, enabling folks in Savannah to sleep well at night, not too worried about the next hurricane strike.

Hurricane history from the past decades seems to support this notion. Hurricane Hugo slammed South Carolina north of Charleston in 1989, but the area south of Charleston survived with little damage. Hurricane Andrew struck South Florida in 1992, far away from this region, and even hurricanes Frances and Jeanne in 2004 inflicted little damage, as they passed to the south.

Longer-term scientific analysis of hurricane strikes also supports the idea that this region is protected. For example, when Dr. Barry Keim and others at Louisiana State University conducted a comprehensive analysis of hurricane strikes from Texas to Maine, they found that only two hurricanes since 1901 impacted St. Simons Island, Georgia. Neither of these storms were major hurricanes, defined as category-3 or higher storms.

In this same period, Cape Hatteras. North Carolina, observed 21 hurricane and three major hurricane strikes. Miami and Palm Beach, Florida, both observed 21 hurricane strikes and six major hurricane strikes in this same timeframe. 

Keim and others from Louisiana State University published a study that included return period analysis of hurricane strikes. This graphic shows the number of years that pass, on average, between major hurricane (red), hurricane (orange) and tropical storm (light green) strikes. In the 105-year period study, St. Simeon Island (city 26) observed two hurricanes, meaning the return frequency of a hurricane strike is approximately 52 years.

Likewise, from Cocoa Beach, Florida, to Tybee Island, Georgia, on the border with South Carolina, only one (Jacksonville Beach) of five cities had observed a major hurricane since 1901. So this study showed a strong geography to hurricane risk, with observed hurricane data providing a clear picture of drastically reduced risk along this concave-shaped coast.

As a graduate student in climatology at LSU, I cut my teeth on this and other data-driven studies conducted by Dr. Keim and other scientists. Such research fascinated me, I saw great value in it, and I endeavored to do similar analyses for storm surge. Indeed, this is a project I am still undertaking today.

Throughout the years of data-driven analysis, one of the lessons I have learned is the value of looking "beyond" the period of analysis to see what's on the other side. Every climate study needs to start somewhere, and this study by Keim and others starts in 1901. However, when we look just before that start date, into the late 1800s, we see an interesting pattern emerge along our Protected Coast.

(And this is of no fault to Keim and own storm surge research began with the year 1880, and I was sometimes asked if it bothered me that I just missed out on the 1875 Indianola Hurricane that struck Texas. But if I moved my start date up to 1870, then I would just miss another major's a never ending cycle.)

The Hyperactive Period of the Late 1800s

The late 1800s was a time of hyperactive hurricane activity along the Protected Coast. Catastrophic hurricanes frequently lashed this region with vicious winds and deadly storm surges.

Perhaps the statistic that best sums up the hyperactivity of this time is an obscure fact I found hiding in the back of a book in a Charleston library. Between 1888 and 1898, eight different hurricanes passed within 75 nautical miles of Hilton Head, South Carolina. That's eight hurricanes in 11 years.

The years 1893, 1894 and 1898 all observed two hurricanes in this area.

However, in the 117-year period extending from 1899 to 2015, only seven hurricanes passed within that area. This means the Hilton Head area, in southern South Carolina, near the Georgia border, observed more hurricane activity during an 11-year period in the late 1800s, than it did in the following 117 years.

The graphic below (a pic I snapped with my camera) shows this pattern. It reveals five hurricanes from 1899 to 1993, but hurricanes Irene (1999) and Charley (2004) must be added to the list, because they passed within 75 nautical miles of Hilton Head.

Graphic of tropical cyclones passing within 75 nautical miles of Hilton Head, South Carolina. From: Smith, David J., 1994: Hurricane Risk, Hilton Head, South Carolina, Climate Report G-18, 20 pp.

Not only did many hurricanes strike this region in the late 1800s, the most severe storms inflicted catastrophic storm surges. The three most severe examples are the back-to-back hurricanes in  1893 and the Great Hurricane of 1898.

Back-to-Back Hurricanes in 1893

On August 27, 1893, the Great Sea Island Storm generated a 12-foot storm tide (surge + waves) at Savannah Beach, Georgia, and water levels reaching 18.2 feet from combined storm tide and waves.

This storm devastated the region from Savannah to central South Carolina, inflicting many fatalities.

The Great Sea Island Storm of 1893 was the first of two hurricanes to devastate Charleston in that active tropical cyclone season. 

Here is a quote on the storm from NOAA Technical Memorandum NWS Hydro-19:

This major hurricane, which originated in the Cape Verde Islands, reached the Georgia coast on August 27. It was accompanied by a very high storm tide which submerged the islands along the Georgia and South Carolina coasts. Between 2,000 and 2,500 people lost their lives on the coastal islands and in the lowland between Tybee Island~ Ga. and Charleston, S.C. Property damage along the coast was estimated at $10 million. Damage in Savannah was placed at more than $1 million. Nearly every building on Tybee Island was damaged and the railroad to the island was wrecked. The highest tide known to have occurred in the county was estimated to have a range of 17.0 to 19.5 ft msl at Savannah Beach [18.2 ft. given in reference [7]].

Shockingly, another powerful hurricane struck the region on October 13, generating a storm tide of 14 feet at Pawley's Island and 12 feet at Charleston

The Great Sea Island Storm of 1893 generated storm tide and waves reaching 18.2 feet along the Georgia Coast and killed 1,000-2,000 people. Map Source: Unisys Corporation

A second major hurricane struck the region two months after the Great Sea Islands Storm, generating storm tides of 12-14 feet along the South Carolina Coast. Map Source: Unisys Corporation

The Great Hurricane of 1898

The Great Hurricane of 1898 was even more severe in terms of maximum storm surge levels along the Georgia Coast. This vicious storm generated a storm surge of 16 ft (4.88 m) at Brunswick, and a storm tide of 18 ft (5.49 m) at Sapels Lighthouse. The storm killed 179 people along the Georgia coast.

The Great Hurricane of 1898 generated a 16-ft (4.88 m) storm surge at Brunswick, Georgia, completely devastating the harbor.

This stretch of coastline is less vulnerable to hurricane strikes than areas to the north and south, however, these three powerful hurricanes in the late 1800s show us that when a hurricane does strike this area, the storm surge can be catastrophic.

Ironically, some properties of storm surge are counterintuitive. Places that are usually the safest, like harbors, bays and inlets, generate the highest storm surges. This is also true for concave-shaped coastlines, like the Protected Coast. The shape of this region and the relatively shallow bathymetry (offshore water depth), efficiently pile up storm surge, but first we need a powerful hurricane tracking in from the southeast, like Matthew. This region is particularly vulnerable to storms like Matthew, as the coastline is open to the east, the very direction from which Matthew is blowing powerful winds for several days.

The coastline of northeast Florida, Georgia and southern South Carolina efficiently builds up storm surge due to its concave shape, relatively shallow offshore water depth and openness to strong east winds, as a hurricane approaches from the south. Graphic: Hal Needham

The efficiency of the Protected Coast for enhancing surge levels was already seen yesterday evening, while Hurricane Matthew was centered off the coast of South Florida. Even though Matthew was located closer to Palm Beach than Jacksonville, storm surge levels in extreme northeast Florida (Fernandina Beach) and Georgia (Fort Pulaski) were already higher than any other NOAA Tides and Currents tide gauge.

This should cause concern, as Matthew would not make landfall in this region for more than 24 more hours after these data were collected (see map below).

Storm surge levels along the Southeast U.S. Coast at 6:00PM EDT on Thu Oct 6.
Water levels: NOAA Tides and Currents; Map: Hal Needham

The National Hurricane Center storm surge forecast has remained consistent since yesterday evening, with peak water levels of 7-11 feet above ground level predicted from northeast Florida to Southern South Carolina. See map below that I created yesterday evening (forecast is the same now).

Indeed, the Protected Coast is now the most dangerous place of all, and we can only hope that coastal populations in this region have heeded the mandatory evacuation and removed themselves from harm's way.

One of the take home lessons from Hurricane Matthew is that we should take the time to read deep into the history of coastal hazards. We should be concerned when we see hyperactive hurricane history in a location like Georgia and southern South Carolina, even if the past decades have been relatively quiet.

If you have questions or comments, feel free to email me at:

Thanks for taking time to read my posts!

-Hurricane Hal


  1. So interesting for a newcomer to a Charleston barrier island to read a long time perspective on this region. Thanks for posting this.

    1. Thank you for reading and for your feedback!

  2. Excellent and Informative post! Thanks for taking the time out to share with us.

  3. How did Unisys generate complete storm tracks for the 1893 storms, including the daily locations, strength, and tracks as they crossed the Atlantic? Without radar, satellite, hurricane hunter planes, etc., how did anyone in 1893 know where the actual storm center was at any given time, let alone the specific location, strength and movement of the center every single day over the open ocean?

    1. Obviously, the record back then isn't as precise as today's but here's a paper which answers your question if you want to learn about it.

    2. Great question, Fred! UNISYS is using data from HURDAT, a comprehensive archive of hurricane positions and intensities that goes back to 1851. Link:

      The HURDAT project is an effort by Chris Landsea and others at NOAA/ AOML to archive hurricane history. In addition to scientific observations, they use ship logs, journals, newspapers, and any other archives to construct hurricane history.

      Data improve through time, so we definitely have a more precise idea of Matthew's track than we do for a storm 100 years ago. Yet, HURDAT is the best dataset available and usually provides an accurate picture of what happened for historic storms.

      For "older" storms, data will be more accurate closer to the coast than out over the open Atlantic.

  4. The buoy offshore of Cape Canaveral reported 30 foot wave heights before it stopped transmitting. Those waves were moving with the leading edge of the hurricane towards South Carolina.

    Thanks for the history lesson which will help me understand the risks here in eastern North Carolina better.

    FYI I post at Dailykos as FishOutofWater. This is a very useful source to use to educate a broader audience. Thanks for writing this blog.

    1. Massive waves push towards shore and the second wave hits before the first wave recedes. This is sometimes called wave setup and leads to an increase in water along the coast.

      Also, storm surge data/ forecasts DO NOT include waves on top of the water. Storm surge refers to the rise in the height of the sea (mean sea level basically), but large waves will ride on top of the surge in some cases. By the time we get slightly inland the waves will be smaller.

  5. Very interesting! We just escaped from St. Simon's Island (on vacation - ha ha to us), and were told no one worried too much about strikes there, but everyone was boarding up, closing down, moving out. My opinion: hurricanes are just plain random and unpredictable. Period.

    1. Hurricanes definitely have a random element. This explains, to some degree, why two hurricanes may strike in two years after 30 years of tranquility.

      That said, certain meteorological factors create patterns that make hurricanes more likely at given times. This is why we often see back-to-back hurricanes in the record, or multiple hurricanes in just several years.

      Also, history has a way of repeating itself. I have been concerned for years about folks in Georgia and southern SC saying that are protected from hurricanes, when the deeper history (late 1800s) shows a hyperactive period in which the area was getting slammed.

  6. Great article. Studied some Marine science in college so very interesting article.Why does a shallow continental shelf cause more storm surge build up?

    1. Shallow bathymetry (water depth) leads to higher storm surge. This is because water is displaced into a "smaller" container and cannot redistribute itself through underwater currents.

      A cat-4 hurricane striking Hawaii will probably produce a storm surge of 6 feet or less, whereas the same storm (depending on size, etc.) could generate a surge greater than 20 feet along coastal Mississippi. However, the waves in Hawaii would be higher.

      You can read about the physical properties that generate storm surge in this book chapter I co-authored.

  7. Thank you very much for your scholarship, and the willingness to share it in a blog. While shopping for property in Savannah I came across the UGA site which shows the hyperactive hurricane period of the late 1800's which affected our coast. Thank you for highlighting the phenomenon. Given the long view the protected coast feels like wishful thinking. Seems like the surge might have not been as bad as predicted at Tybee and I would be interested to read your post storm analysis. After having lived 12' above sea level in Eastern North Carolina, where we had 7 named storms come ashore (eye wall) within 50 miles in 5 years, we decided to build in the Historic District.

    1. Thanks for your encouragement! I will analyze Matthew's water levels and place them in historic context within the coming days and weeks!

  8. Thank you for the great article. I am interested in why hurricane Mathew did not result in severe (or really any) storm surge south of cape canaveral, but devastated regions further north. Was it the storm track, the lower-tide during closest approach, the geography of the region (well defined barrier islands, few inlets, etc.), or something else. It would be great to understand if it was specifics of this particular storm, or if the region has some inherent advantages with respect to storm surge. I could speculate, but it would be great to hear from an expert!

    1. Yes, I am interested in this question, too, being from just south of the Cape myself. It looks like our surge was about 3-4' on the ocean side, and higher on the Banana River side. Much worse in Merritt Island, just 15 miles to our north. Great post, by the way, intriguing and disturbing!

    2. Great question! The lower storm surge heights in S Florida were due to a combination of storm track and coastal geography. 1) The eyewall, or area of most intense winds around the eye, stayed offshore south of the Space Coast, but was coming onshore to the north. This resulted in a lot more forcing of water where the eyewall was coming ashore. 2) Coastal geography has a big influence too, especially because of bathymetry, or offshore water depth, because shallow water increases surge heights. Consider this, the water 10 miles E of St. Augustine is around 63 ft deep, whereas the water 10 miles E of Fort Lauderdale is around 850 ft deep. The region from Palm Beach south to Miami has the deepest offshore water in the state; this suppresses storm surge height, but builds higher waves. Let me know if you have any other questions! Thanks! -HH

    3. Maria- storm surge is quite localized. Given the track of the storm and offshore water depth, it makes sense that there was a big storm surge gradient, or difference, between your area and regions 15+ miles to the north. Thanks for your encouraging comment! -HH