Submerged aquatic vegetation mapping in coastal Louisiana through development of a spatial likelihood occurrence (SLOO) model

Abstract

Determining the spatial distribution of coastal foundation species is essential to accurately determine restoration goals, predict the ecological effects of climate change, and develop habitat management strategies. Mapping the distribution of submerged aquatic vegetation (SAV) species assemblages, which provide important habitat resource and ecological services in Louisiana, has been difficult due to the dynamic nature of SAV occurrence and the limited water clarity across much of the coast. Species distribution models (SDMs) link ecological conditions species occurrence across landscapes, and can predict the distribution of species across un-sampled or hard to sample areas and support the development of habitat maps. To predict SAV distribution in coastal Louisiana, a SDM was developed and projected across the landscape to create a spatial likelihood of occurrence (SLOO) model describing the probability of SAV presence in aquatic habitats. SAV presence and absence data were examined from over 500 field observations in relation to physical and hydrologic variables, including exposure, turbidity, water level, and salinity. A binary logistic regression model (p < 0.0001) identified three significant predictors of SAV presence: mean winter salinity, exposure, and turbidity. As each of these variables increased, the probability of SAV presence in the summer growing season decreased. The spatial application of this SDM helps to predict the likelihood of occurrence across the coastal landscape, creating a valuable tool to describe un-sampled SAV habitat and estimate future changes in habitat availability.