Abstract
As a dynamic system, salt marshes are highly productive habitats. Exceedingly dependent on specific tolerances to tidal regimes, temperature, elevation, and salinity, the plant composition and structure of a salt marsh are highly variable. Due in part to their visible spatial patterns and the impracticality of large scale field observations, aerial and satellite remote sensing has provided an effective alternative for studying these plant communities. This research focused on the use of hyperspectral remote sensing to advance an overall understanding of salt marsh ecosystem productivity, diversity, and abundance within specific National Estuarine Research Reserves (NERRs) using AISA (Airborne Imaging Spectrometer for Different Applications) imagery. |The overall goal of my thesis was to create a synoptic overview and site comparisons of salt marsh vegetation for seven NERRs, within their respective growing seasons, using three remote sensing vegetation indices. Additionally, within and between site geospatial patterns were examined using two landscape metrics. The seven sites were located in a range of climates, tidal, and river discharge regimes between Texas and Delaware. |Across all sites, the NDVI and MSAVI indices were well suited for distinguishing structural differences and illustrating variations in relative biomass and plant vigor at each site. They both showed latitudinal and longitudinal shifts in histogram distributions, relatable to nutrient, tidal, and climate regimes. These indices were also useful for discriminating features unique to individual sites, such as exposed, intertidal mud flats (Georgia site) and robust Phramites australis stands (Delaware site). |The VARIgreen index, more closely captured the photosynthetic capacity and "greenness" of salt marsh vegetation, and was more sensitive to phenological changes and foliage pigment densities. This index was more responsive to spring greening and late summer senescence. As a result, VARIgreen's distributional patterns appeared more sensitive to growing season patterns.|Patches per unit area (PPU) was an effective landscape metric for measuring patchiness of vegetation index classes for the entire landscape. The Texas, Georgia, and South Carolina sites displayed the highest PPU values, while sites in Mississippi and West Florida had the lowest values. Although PPU illustrated the abundance of different patches in a single area, it was biased towards dominant classes. In comparison, the Shannon Diversity Index (SDI) captured the degree of equitability of all classes of patches. The largest class diversity and SDI values were found in the Delaware and West Florida sites.|Ultimately, the vegetation indices and landscape metrics helped to show spatial variations at each site. As a result, site characterizations, unique to each site helped to explain the structure and productivity of the eight datasets.