ABSTRACT

Beach ridge plains on four sides of the Gulf of Mexico produced a new sea level history for late Holocene time: St. Vincent Island, Florida panhandle; Sanibel Island, lower peninsula of Florida; Mesa del Gavilan, east of Brownsville, Texas; and Isla del Carmen, in the state of Campeche, Mexico.

On St. Vincent Island there are roughly 400 data points (elevation and relative age), each taken from a position on a beach ridge or in a swale. These ridges and swales can be grouped in sets; some sets stand high, and some are low. Some of the set boundaries mark rises or drops of sea level. Detailed topography, relative moment measures (sand grain size distribution), and linear-suite grain-size data all show the same changes: two drops and two rises since about 5,000 B.P. Vertical uncertainty is one meter or less; time uncertainty is less than about 50 years. High stands exceeded the present level by 1.0 m (3.3 ft) or more, and low stands were below it by 1.0 m (3.3 ft) or more. The latest change in this range was a rise, about 800 years ago.

The three other areas were compared with St. Vincent Island. The fit is very good, as far back as the record goes in each case. Areas of measurable (known) warping have been avoided in this work. Because the four study localities are spaced at distances up to more than 1,200 km (745 mi) apart, it does not seem possible that unknown warping could have affected the results. The large number of data points within each small area, the large distances between study sites, and direct use of the data (rather than statistical surrogates) provide very great assurance for the reliability of the result. The curve produced in this fashion has been designated the "Gulf of Mexico" late Holocene sea level curve.

Late Holocene sea level did NOT approach the present position asymptotically, but fluctuated both above and below it, with an apparent period of very roughly 2,000 years. The rate of short-term change was about 5 cm/yr (2 in./yr). Small changes (<50 cm; <20 in.) and very slow changes (e.g., < one cm/yr; < 0.4 in./yr) are not visible in the present data set.