Although nutrient reduction has been used for lake eutrophication mitigation worldwide, the use of this practice alone has been shown to be less effective in combatting cyanobacterial blooms, primarily because of climate change. In addition, quantifying the climate change contribution to cyanobacterial blooms is difficult, further complicating efforts to set nutrient reduction goals for mitigating blooms in freshwater lakes. This study employed a continuous variable Bayesian modeling framework to develop a model to predict spring cyanobacterial bloom areas and frequencies (the responses) using nutrient levels and climatic factors as predictors. Our results suggested that both spring climatic factors (e.g., increasing temperature and decreasing wind speed) and nutrients (e.g., total phosphorus) played vital roles in spring blooms in Lake Taihu, with climatic factors being the primary drivers for both bloom areas and frequencies. Climate change in spring had a 90% probability of increasing the bloom area from 35 km2 to 180 km2 during our study period, while nutrient reduction limited the bloom area to 170 km2, which helped mitigate expansion of cyanobacterial blooms. For lake management, to ensure a 90% probability of the mean spring bloom areas remaining under 154 km2 (the 75th percentile of the bloom areas in spring), the total phosphorus should be maintained below 0.073 mg·L-1 under current climatic conditions, which is a 46.3% reduction from the current level. Our modeling approach is an effective method for deriving dynamic nutrient thresholds for lake management under different climatic scenarios and management goals.
Keywords: Cyanobacterial blooms; Data-intensive statistical models; Global warming; Internal loading; Shallow lakes.
Copyright © 2023 Elsevier Ltd. All rights reserved.