Using generalized additive models for prediction of chlorophyll a in Lake Okeechobee, Florida

Abstract

AbstractLong‐term eutrophication data collected by the South Florida Water Management District at eight stations in the pelagic region of Lake Okeechobee since 1980 were combined with those from littoral region stations collected since 1986. These data were supplemented by those collected by the University of Florida in conjunction with the Lake Okeechobee Ecosystem Study. Generalized additive models (GAM) were fit to the chlorophyll a response for each of five previously documented‘ecological zones’, each using the same predictor variables for comparison between zones. Significant nonlinear increasing relationships between log10 TP and log10 chlorophyll a were found for littoral and near littoral regions of the lake, which change to decreasing at certain (high) levels of log10 TP. This change to a decreasing function suggests light limitation associated with high TP levels due to particle resuspension. The error distribution of the GAM compare favourably to those of linear models. The relationship of the chlorophyll a response to the predictors used in the GAM appear similar for littoral and near littoral regions, as they do for north pelagic and central pelagic ecological zones, with intermediate shapes of these functions found in the transition zone between these two groups.