Marrack, L., 2015. Incorporating groundwater levels into sea-level detection models for Hawaiian anchialine pool ecosystems.
As sea levels rise, the distribution and community structure of coastal ecosystems are expected to change. In many coastal aquifers, fresher groundwater floats on top of denser saltwater and will rise with sea level. Under these conditions, ecosystems dependent on groundwater may shift inland as a result of inundation, changes in salinity, or both. Groundwater-fed anchialine pool habitats existing in porous coastal substrates around the world have not been assessed for sea-level rise impacts. As a first step toward examining ecosystem response to rising water levels, geospatial models were developed to detect anchialine pools on the island of Hawai‘i at current water levels and models were validated with known pool locations. Specifically, the objectives were to determine whether accounting for groundwater levels in the model improved pool detection, to identify the model that most accurately detected known pools, and to identify which pool features make some pools more likely to be detected than others. Six water level models were validated with the test data set of actual pool locations to determine how well they detected known anchialine pools. Water surface models that included groundwater levels were up to 37% better at detecting anchialine pools than corresponding models without groundwater levels. The model that included groundwater levels at mean higher high water was applied to 42 km of coastline where it correctly detected 62% of known pools. A generalized linear model showed that pools with surface areas greater than 5 m2 and pools without canopy were the most likely to be detected. Future predictive modeling of anchialine pool response to sea-level rise should include groundwater levels. Furthermore, geospatial models aimed at predicting ecosystem shifts due to sea level rise may be improved by including groundwater as a factor and should be validated using current ecosystem conditions.
