AVHRR imagery used to identify hurricane damage in a forested wetland of Louisiana

Abstract

Certain events provide a unique opportunity to test the monitoring capability of AVHBR imagery. On 26 August 1992, Hurricane Andrew passed through Louisiana, impacting a large area of forested wetlands. One response to the widespread defoliation resulting from the hurricane impact was an abnormal bloom of new leaves and new growth in the underlying vegetation between September and October. To capture this atypical phenology, a time sequence of AVHRR images was transformed into a normalized difference vegetation index, NDVI, as an indicator of vegetation changes in the forest impacted by the passage of a hurricane. Using geographic information system functions, three sites in the impacted forest were vectorized as polygons, and the inclusive pixels were extracted for subsequent graphical and univariate statistical analysis. Temporal curves of mean NDVIs for the three sites for before, during, and after the hurricane passage, and aggregate curves of the impacted forest to an undisturbed forest, were compared. These comparisons corroborated the atypical phenology of the impacted forested wetland and directly related the cause to the hurricane passage.