This study is focused on estimating belowground productivity of Spartina alterniflora, the dominant salt marsh macrophyte on the east coast of the United States. Belowground productivity of marsh plants contributes to coastal marsh resiliency both through sub-surface expansion, allowing marsh surface elevation to keep pace with sea level rise, and through soil stabilization, preventing lateral marsh erosion from wave action. Belowground productivity also contributes to soil organic matter, providing information on so called “Blue Carbon” dynamics. Site-wide estimates of belowground biomass are difficult to obtain because root:shoot ratios vary in response to environmental gradients and other factors that drive biomass partitioning. We developed analytical methods to estimate S. alterniflora belowground biomass from aboveground proxies, including aboveground biomass, plant foliar N, and remote-sensing derived spectral reflectance. We then scaled these vegetation biophysical models spatially through the use of free Landsat 8 satellite observations. Field data to parameterize this effort came from 4 salt marshes along the Georgia coast, 3 within the Georgia Coastal Ecosystems Long-Term Ecological Research domain on Sapelo Island and 1 on Skidaway Island, south of Savannah.
Results/Conclusions
We used a marsh inundation index to categorize satellite observations as representing either dry or flooded conditions. This separation was 80% accurate and subsequent removal of flooded observations provided a high quality vegetation time-series for downstream analyses. Next, we used Partial Least Squares Regression to build biophysical models from satellite data, including aboveground biomass (Explained Variance (EV) = 60%), foliar N (EV = 43%) and belowground biomass (EV = 46-52%). We applied our best fit model to create maps of spatial and temporal variation in S. alterniflora characteristics. This mapping suggested that foliar N was highest adjacent to upland areas and peaked in early summer, whereas production of above and belowground biomass began in the marsh interior early in the growing season, followed later by creek-bank production. Thus, Landsat 8 was useful for site-wide assessments of above and belowground vegetation characteristics in salt marshes. These results mirrors those from a previous study in the Sacramento-San Joaquin Delta, CA, where we applied similar methods to estimate belowground productivity and root:shoot ratios of Scheonoplectus acutus and Typha spp., marsh plants common to some tidal freshwater areas. As these methods develop, our goal is to map belowground productivity in coastal marshes more broadly.