Estimating above-ground biomass of salt marsh vegetation using hyperspectral measurements of canopy level reflectance

Various biophysical parameters of vegetation can provide researchers with key information regarding the health of the environment at multiple scales. Items of particular interest, such as biomass, can be important indicators of plant health, and thus are useful for assessing the overall condition of an ecosystem. However, techniques for estimating biomass are often invasive and destructive to the very vegetation targeted by the study. Remote sensing has the advantage of being minimally invasive and has been shown many times to be a highly useful tool for estimation of biophysical parameters in well-regulated and controlled habitats. Areas such as corn and soy fields have produced very high correlation between biomass and spectral response. Unfortunately, limited research has been done to apply these techniques to other, more diverse areas. This research focuses on estimation of these parameters in estuarine marsh environments where species composition can be highly complex and diverse. The Analytical Spectral Device (ASD) field spectro-radiometer was used to collect spectral data of 100 random points distributed throughout the Grand Bay National Estuarine Research Reserve (GBNERR), Mississippi. The vegetation was destructively sampled and above-ground dry-weight biomass values were determined. Spectra were processed using a number of vegetation indices applied to other areas and separated by species. Then, species-specific signatures were compared to biomass values and showed promising correlation levels (r2 values were generally above 0.5). Though not as high as correlation levels in other environments, this is understandable given the complex nature of the ecosystem and may be a highly useful tool for research managers.