Doctor of Philosophy (PhD)


Oceanography and Coastal Sciences

Document Type



This dissertation evaluates the hypothesis that hurricane activity levels in the North Atlantic during the late Holocene have been driven by latitudinal movements of the North Atlantic circulation system. Multi-millennial sedimentary proxy records, based on the occurrence of overwash clastic layers, provide clear evidence of abruptly alternating periods of hurricane landfall frequency for Nicaragua and Belize. Three Belizean transects exhibit an Active period (hyperactivity) occurring from ~2000-6000 cal yr BP, although dating is inconsistent across the transects. An Active period covering the last 500 years is found at one location. The Nicaraguan record, derived from three transects covering >90 km of coastline, consistently displays an Active period covering the last 800 years, preceded by a Quiet period that lasts until at least ~2800 years BP, before which time environmental factors render the sites insensitive. For both coastlines the calculated strike frequency increased by a factor of 3-12 during Active periods. The Barbados depositional record is characterized by sudden shifts from organic to clay, attributed to increased aridity, with the arid periods being roughly contemporaneous with the Active periods occurring in Belize and Nicaragua, as well as periods of southern residency of the Intertropical Convergence Zone. Latitudinal movements of a unified North Atlantic circulation system were probably the driver of these changes, with southern migration increasing both landfall frequency and aridity regionally. When correlated with published records, the timing of activity regime changes identified from our sites indicates that periods of increased hurricane activity proceed across the North Atlantic in a time-transgressive manner, with the Caribbean hyperactive period preceding that of the Gulf of Mexico. The Active period for Nicaragua beginning ~850 years BP is anti-phase with a recently published model, predicated upon basin-wide synchroneity in activity patterns. This discrepancy possibly results from differences in spatial coverage, as correlations between hurricane landfall and track patterns indicate three distinct groupings resulting from atmospheric conditions. The basin-wide pattern is derived from locations contained within a single (Atlantic coast) track set, while our time-transgressive model is derived from sites within both the Caribbean and Gulf of Mexico track sets.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Liu, Kam-biu