Using generalized additive models to investigate factors influencing cyanobacterial abundance through phycocyanin fluorescence in East Lake, China.

Abstract

East Lake is a shallow lake (in Wuhan, China) where cyanobacteria blooms occurred frequently from 1970 to 1985. During the study period, all Carlson trophic state index values were > 50, indicating that East Lake is in a eutrophic state. In this study, phycocyanin concentrations were measured through phycocyanin fluorometry for rapid assessment of cyanobacterial abundance. The smoothing splines of the optimal generalized additive model (GAM) indicated that Secchi depth (SD), total phosphorus (TP) and dissolved oxygen (DO) concentrations, electrical conductivity (EC), chemical oxygen demand (COD), and ratios of total nitrogen (TN) to TP (TN:TP) were the main environmental factors in a moderate nonlinear relationship with cyanobacterial phycocyanin concentrations in East Lake. The shape of the GAM smoother can be used to quantify the relationship between a response variable and an explanatory variable in the scatterplot. Phycocyanin concentrations were sharply and negatively related to both SD and EC when the SD was 20-80cm and EC was > 270mg/L. Phycocyanin concentrations increased with concentrations of TP, DO, and COD. Phycocyanin concentrations increased sharply with TP concentrations when TP concentrations were > 0.10mg/L and approached to a constant when DO concentrations were > 8.20mg/L. Approximately, 85% of the phycocyanin concentrations were negatively correlated with TN:TP of < 26. In summary, organic compounds and TP were inferred to the key factors limiting the potential growth of cyanobacteria in East Lake. These change points/thresholds of smoothing splines of aforementioned variables may serve as a framework for managing the cyanobacterial growth.