RESOURCE-MANAGEMENT OF FORESTED WETLANDS - HURRICANE IMPACT AND RECOVERY MAPPED BY COMBINING LANDSAT TM AND NOAA AVHRR DATA

A temporal suite of NOAA Advanced Very High Resolution Radiometer (AVH RR) images, transformed into a vegetation biomass indicator, was combi ned with a single-date classification of Landsat Thematic Mapper (TM) to map the association between forest type and hurricane effects. Hurr icane effects to the forested wetland included an abrupt decrease and subsequent increase in biomass. The decrease was associated with hurri cane impact and the increase with an abnormal bloom in vegetation in t he impacted areas. Impact severity was estimated by differencing the b iomass maps before and immediately (3 days) after the hurricane. Recov ery magnitude was estimated by differencing the biomass maps from imme diately (3 days) after and shortly (1.5 months) after the hurricane. R egions of dominantly hardwoods suffering high to moderate impacts and of dominantly cypress-tupelos suffering low impacts identified in this study corroborated findings of earlier studies. Conversely, areas not reported in previous studies as affected were identified, and these a reas showed a reverse relationship, i.e., highly impacted cypress-tupe lo and low or moderately impacted hardwoods. Additionally, generated p roportions of hardwood, cypress-tupelo, and open (mixed) forests per e ach 1-km pixel (impact and recovery maps) suggest that regions contain ing higher percentages of cypress-tupelos were more likely to have sus tained higher impacts. Visual examination of the impact map revealed a spatial covariation between increased impact magnitudes and river cor ridors dominated by open forest. This spatial association was corrobor ated by examining changes in the percentage of open forest per I-km im pact pixel; the percentage of open forest peaked at moderate to high i mpacts. The distribution of recovery supported the impact spatial dist ribution; however, the magnitudes of the two indicators of hurricane e ffects were not always spatially dependent. Converse to univariate sta tistics describing all forested area within the basin, higher recoveri es tended to be related to higher percentages of hardwoods. Lower reco veries, on the other hand, tended to be related to forests with nearly equal percentages of hardwoods and cypress-tupelo.