Predicting Coral Reef Carbonate Chemistry Through Statistical Modeling: Constraining Nearshore Residence Time Around Guam

Abstract

To accurately predict the impacts of ocean acidification on shallow-water ecosystems, we must account for the biogeochemical impact of local benthic communities, as well as the connectivity between offshore and onshore water masses. Estimation of residence time can help quantify this connectivity and determine the degree to which the benthos can influence the chemistry of the overlying water column. We present estimates of nearshore residence time for Guam and utilize these estimates to model the effects of benthic ecosystem metabolism on the coral reef carbonate system. Control volume and particle tracking approaches were used to estimate nearshore residence time. These estimates were paired with observed patterns in the reef carbonate system around Guam using water samples collected by NOAA’s National Coral Reef Monitoring Program. Model performance results suggest that when considering the effects of benthic metabolism on the carbonate system, it is paramount to represent the contact time of the water volume with the benthos. Even coarse estimates of residence time significantly increase model skill. We observed the highest predictive skill in models including control volume derived estimates of residence time, but only when those estimates were included as an interaction with benthic composition. This work shows that not only is residence time critically important to better predict biogeochemical variability in coral reef environments, but that even coarse hydrodynamic models can provide useful residence time estimates at management relevant, whole-ecosystem scales.