Use of GIS mapping and modeling approaches to examine the spatial distribution of seagrasses in Barnegat Bay, New Jersey

Due to the ecological importance of seagrasses and recent indications of disease and dieback, we have synthesized existing mapped survey information concerning the spatial and temporal distribution of seagrass beds (primarily eelgrass, Zostera marina) in Barnegat Bay, New Jersey. Mapped surveys from the 1960s, 1970s, 1980s, and 1990s were digitized and compiled in a geographic information system to facilitate analysis. Comparison of the earlier maps with the 1990s survey shows an overall decrease of approximately 2,000 to 3,000 ha in die area of seagrass beds. While there are indications of seagrass decline, due to the great difference in mapping methods used for each of the surveys, we are cautious in directly attributing the decrease in mapped eelgrass acreage to a large-scale dieback. We examined the extent to which light could be used to predict the distribution of seagrass in Barnegat Bay. Data on Secchi depth throughout the bay were combined with a modification of at) existing model (Duarte 1991) of the relationship between Z. marina compensation depths and light attenuation coefficients to predict the distribution of seagrasses in Barnegat Bay. When compared with mapped seagrass distribution in the bay, the model correctly predicts seagrass presence-absence over two-thirds of the time. The majority of the model error is due to errors of commission, i.e., the model predicts seagrass occurrence where it was not observed to occur. Most of this commission error is located in specific geographic areas (i.e., southern third of Little Egg Harbor and the western shoreline of the bay).