The relationship between hurricane wind fields and the associated disturbance in US forest ecosystems

Tropical cyclones are recurrent in the U. S. Gulf Coast and produce drastic disturbance to the forest ecosystem by altering forest structure, species composition, nutrient cycling, biomass accumulation, etc. The recurrent characteristics of these events demand a rapid yet reliable assessment of forest disturbance in order to provide better management decisions, as well as to evaluate damage to the landscape, biomass loss and the associated impacts to the regional carbon budget. In this study we present a methodology for rapid assessment of forest disturbance produced by tropical cyclones based on maximum sustained wind swaths (H*wind), MODIS (Moderate Resolution Imaging Spectroradiometer)-derived disturbances and field-measured tree mortality collected in Gulf Coast forests. MODIS images in May (the month of maximum greenness) before the disturbance and in May of the year following the disturbance were processed using spectral mixture analysis (SMA) using image derived end-members. The changes in non-photosynthetic vegetation (ΔNPV, related to wood, dead vegetation and surface litter) from one year to the next was used as the disturbance metric. A strong correlation was found between H*wind and MODIS ΔNPV for hurricanes Charley (2004), Katrina (2005), Rita (2005), and Gustav (2008). In turn, MODIS ΔNPV was shown to have a strong correlation with field-measured mortality. The forest disturbance estimates based on hurricane wind-field and MODIS ΔNPV agree with those published in our previous studies. This study establishes an important relationship that could be incorporated into earth system models to improve our understanding of the effect of tropical cyclones on terrestrial ecosystems and their associated feedbacks within the climate system.