ASSESSING SPATIAL EROSION PATTERNS RESULTING FROM HURRICANE MARIA ON DOMINICA, LESSER ANTILLES, CARIBBEAN

Quantifying the changes resulting from a single large storm event in the context of geomorphic processes is a challenging task due to their unpredictability. This study assesses the erosional effects of Hurricane Maria on the island of Dominica by geochemical analysis of detrital samples. We collected sediment samples (<63 and 250-850μ) at 29 sites in July 2017 and January 2018 to document the effects of Hurricane Maria (landfall made in September 2017). Gamma spectroscopy was used to detect Pb-210, Be-7, and Cs-137 in the raw samples and their acid extractable grain coatings. Fractions were also analyzed using XRF and XRD. Meteoric Be-10 was measured by accelerator mass spectrometry. Additional data include repeat field photography and satellite imagery to observe the erosional and environmental changes over a longer (1-5 year) timescale. These methodologies provide insight into the spatial and temporal variation occurring across the island’s 20 watersheds.

XRF findings for clays on the island report, on average, mass compositions of 56% SiO2, 19% Al2O3 and 13% Fe2O3. These results agree with XRD peaks which are nearly uniform for all the samples, being primarily kaolinite and quartz. These data demonstrate consistency in spatial weathering patterns across the island. Preliminary isotopic data show the majority of post-hurricane samples are enriched in meteoric Be-10 compared to pre-hurricane samples suggesting storm induced erosion of shallow soil. Repeat photography of the channels reveals that most sites experienced significant vegetation loss and increased presence of cobbles and boulders after Hurricane Maria. Additionally, many channels noticeably increased in depth and width. Analysis of satellite imagery suggests nearly 700 total landslide events were triggered by Hurricane Maria, with a westerly bias (77%) across Dominica. On average, slides were higher in elevation (300-400m) but smaller in relative magnitude than those observed before the hurricane.