A REEVALUATION OF THE GEORGIA AND NORTHEAST FLORIDA TROPICAL CYCLONE OF 2 OCTOBER 1898
Al Sandrik(1)and Brian Jarvinen(2)
On October 2, 1898, a tropical cyclone impacted the Georgia and Northeast Florida coastlines and inland regions. Current National Hurricane Center archives indicate the center of the tropical cyclone made landfall along the central Georgia coast near Sapelo Island with the intensity of a category two hurricane on the Saffir/Simpson hurricane intensity scale. A detailed search of historical records, including newspaper accounts and personal diaries and letters, as well as National Weather Service records, indicate that the storm made landfall both farther south and with greater intensity than the current records indicate. Because the height of the storm tide is directly correlated to the intensity of the hurricane, an effort was made to measure known high water marks in the region. Utilizing the high water mark information and the existing meteorological data, the National Weather Service's storm surge model, called SLOSH (Sea, Lake and Overland Surges from Hurricanes) and NOAA's astronomical tide model were run in a series of successive iterations until a "best fit" between the model output and the historical storm tide elevations and meteorological observations occurred. The "best fit" hurricane thus obtained would have central a pressure of 938 millibars and a radius of maximum winds of 21 statute miles. Pressure/wind relationships would yield a maximum 1-minute wind of approximately 135 miles per hour. The center of the "best fit" hurricane would make landfall on Cumberland Island, which would place the center approximately 30 statute miles south of the current archived track. The data suggests that this storm was a category four hurricane on the Saffir/Simpson hurricane scale and would be similar in intensity and size to hurricane Hugo which impacted the South Carolina coast in September of 1989. A correction to both the archived track and intensity is suggested.
In order to remain true to the historical reports and data, tide and surge heights are presented in feet above mean sea level (M.S.L.), pressure in millibars (mb), winds in miles per hour (mph) and distances in statue miles (mi).
The current archived track for the October 2nd, 1898 hurricane indicates the tropical cyclone made landfall along the central Georgia coast, near Sapelo Island, as a category two hurricane. The original storm track for this event was presented in the October 1898 edition of the Monthly Weather Review. This track was presented in several early 20th century studies and was only slightly modified from the original presentation, with the point of landfall left unchanged. This information was later incorporated into the archived data base, sometimes referred to as the "Best Track", maintained by the National Hurricane Center.
A study of historical tropical cyclone landfalls along the northeast Florida and Georgia coast was undertaken in 1997 to enhance the understanding of the landfalls in those areas. An exhaustive effort was undertaken to obtain period documents including unpublished records such as diaries and journals kept by the local residents and local newspaper accounts. The data are presented below working in a north to south direction.
At Savannah Georgia the wind continued from the northeast to east throughout the event with a maximum wind speed of 60 mph reported. The minimum pressure reported was 999 mb and the barograph trace did not show a well pronounced "barograph V" signature. Along the central Geogia coast the winds at Darien and St Simons Island are reported by observers to also have been from an easterly direction through out the event. More significantly the Caretaker at the Jekyll Island Club reported, "It commenced to blow about midnight Saturday from the north east, the wind gradually increased until it blew a perfect gale. At about noon the wind changed to the south east and blew the same velocity for the balance of the day." At no time does this observer report a wind shift to a westerly direction even though the archived track places landfall approximately 20 miles to the north of his position.
Along the south Georgia coast, in Camden county, we find the most significant reports of a wind shift with a brief calm occurring at approximately 11 AM. Mrs. Jennie Hopper Miller, of St. Marys, Georgia, recorded in her diary, "This morning the wind was from the northeast and it has come around until it is now from the southwest and is blowing as hard, if not worse than it was." Mr. John Hernandez, of the same town, related to a local historian, "... The wind was calm for a while, but then began to blow from the southwest (italics added by author)." Of note here are both the reports of a wind shift by both observers with a calm evident in one of the reports. Also in Camden County, Georgia the pilot boat, Maud Helen was grounded on the bluff at High Point, on Cumberland Island. This bluff has an average elevation of 20 feet above M.S.L. and a storm tide of that height is unlikely so far to the left of the archived storm track.
In north Florida, both Fernandina Beach and Jacksonville report a significant wind shift to the west. Fernandina Beach, which sits on the western portion of Amelia Island, is vulnerable to sound side flooding and indeed was inundated during the event. The extent of this flooding, as at High Point, is unlikely with a storm landfalling approximately 40 miles to the north of this position. The minimum barometric pressure at Jacksonville was 985 mb with a pronounced "barograph V" signature noted on the micro-barograph trace.
The center of the storm continued to move inland was identified at the town of Blackshear (see figure 1) between 2 and 3 PM and in the town of Douglas at approximately 7 PM. Based on the data presented it appears that the tropical cyclone made landfall on Cumberland Island, in Camden County and moved inland just to the north of St. Marys, Georgia and Fernandina Beach, Florida. The storm continued to move on a west-northwest heading with a speed of advance of approximately 15 mph.
The National Weather Services's, Sea, Lake and Overland Surges from Hurricanes, or SLOSH model, was run in successive iterations to obtain a "best fit" hurricane utilizing the track information as specified above.
In developing a basin for the SLOSH model, in this case the Brunswick, Georgia Basin, an intensive effort is undertaken to survey each basin for topography, bathymetry and historical storm tide information. The points derived from the historical data are then surveyed by the Army Corps of Engineers to ascertain the location's height relative to M.S.L. The points surveyed for the 1898 event and presented in this paper are Isle of Hope, Darien, and two locations in the city of Brunswick, all along the Georgia coast and north of the "corrected" track .
The storm tide, as presented in the hydrographs (figures 2a-d) is defined as the sum total of the astronomical tide and the storm surge generated by the hurricane. The astronomical tide for each location was derived by using the closest tidal prediction point in the NOAA Tide Prediction Program database. The surge heights were generated by the SLOSH model.
SLOSH model output is dependent upon several factors, the exact landfall point and angle of incidence to the coast, the speed of advance, the estimated pressure at the storm's outer radius, the central pressure, and the radius of maximum winds (rmw). Given the above information the track/angle of incidence, landfall point and speed of advance were known values and an estimated outer pressure of 1010 mb was obtained using the Savannah and Charleston barograph traces (not shown).
Since the maximum storm tide at each of the four locations was known the SLOSH model was run in successive iterations adjusting both the central pressure and the radius of maximum winds until the SLOSH and astronomical generated storm tide approximated the storm tide observed at the four locations. The "best fit" hurricane thus obtained would have a central pressure of 938 millibars and a radius of maximum winds of 21 statute miles. Pressure/wind relationships would yield a maximum 1-minute wind of approximately 135 miles per hour over portions of northern Camden County and southern Glynn County, Georgia.
If care is taken in extensively researching the historical record, it is possible to re-evaluate historical tropical cyclone landfall events in order to better understand and quantify the true hurricane history for a given location. In this particular case the data suggests a correction of the historical record to include a category four hurricane making landfall approximately 30 statute miles south of the current archived position and a correction to the archived data is recommended.
1. Corresponding author: Al Sandrik, National Weather Service Office, Jacksonville, Florida, 13701 Fang Dr., Jacksonville, Florida, 32218, E-mail: Al.Sandrik@noaa.gov
2. Tropical Prediction Center/National Hurricane Center Miami, Florida.