Using species distribution models to guide seagrass management

Seagrasses provide essential and global ecosystem services. However, due to natural and anthropogenic disturbance, seagrass meadows around the world have declined dramatically in recent decades. Researchers and managers have been calling for increased frequency and accuracy in the mapping of seagrass distributions to benefit seagrass conservation for over a decade, and we argue that a critical advancement in the management of seagrasses could come through increased and iterative model-based mapping of potential habitat as an accessory to seagrass monitoring. We further demonstrate how focus on errors of commission and omission during model interpretation could guide management activities, using the Texas Gulf Coast as a case study representing seagrass habitats worldwide. We used the species distribution modeling program Maxent to predict the location of suitable habitat for each seagrass species along the Texas Coast based on local current velocity, distance from boat launches (i.e., an index of human disturbance), nitrogen, light availability, salinity, and temperature. Models accurately predicted suitable habitat for all seagrass species, including Halodule wrightii (AUC = 0.830 ± 0.032), Thalassia testudinum (AUC = 0.901 ± 0.058), Syringodium filiforme (AUC = 0.911 ± 0.036), Halophila engelmannii (AUC = 0.865 ± 0.092), and Ruppia maritima (AUC = 0.868 ± 0.040). The relative importance of environmental factors differed between models. Distributions for H. wrightii and T. testudinum were most influenced by surface nitrate concentrations. S. filiforme, H. engelmannii, and R. maritima distributions were most influenced by benthic light availability. Human disturbances often lead to elevated nitrate concentrations and decreased benthic light availability, and our models generally predicted a lack of suitable habitat near sites characterized by abundant human development. We considered model errors of commission and omission for each species to identify candidate regions for seagrass transplantation and habitat restoration, respectively. Overall, we believe that the utility of the approach we have developed in the Texas Gulf Coast case study along extends beyond a single study site, and our methods will assist conservation of seagrass meadows worldwide.