Predicting cyanobacteria bloom occurrence in lakes and reservoirs before blooms occur.

With increased global warming, cyanobacteria are blooming more frequently in lakes and reservoirs, severely damaging the health and stability of aquatic ecosystems and threatening drinking water safety and human health. There is an urgent demand for the effective prediction and prevention of cyanobacterial blooms. However, it is difficult to effectively reduce the risks and loss caused by cyanobacterial blooms because most methods are unable to successfully predict cyanobacteria blooms. Therefore, in this study, we proposed a new cyanobacterial bloom occurrence prediction method to analyze the probability and driving factors of the blooms for effective prevention and control. Dominant cyanobacterial species with bloom capabilities were initially determined using a dominant species identification model, and the principal driving factors of the dominant species were then analyzed using canonical correspondence analysis (CCA). Cyanobacterial bloom probability was calculated using a newly-developed model, after which, the probable mutation points were identified and thresholds for the principal driving factors of cyanobacterial blooms were predicted. A total of 141 phytoplankton data sets from 90 stations were collected from six large-scale hydrology, water-quality ecology, integrated field surveys in Jinan City, China in 2014-2015 and used for model application and verification. The results showed that there were six dominant cyanobacterial species in the study area, and that the principal driving factors were water temperature, pH, total phosphorus, ammonia nitrogen, chemical oxygen demand, and dissolved oxygen. The cyanobacterial blooms corresponded to a threshold water temperature range, pH, total phosphorus (TP), ammonium nitrogen level, chemical oxygen demand, and dissolved oxygen levels of 19.5-32.5 °C, 7.0-9.38, 0.13-0.22 mg L-1, 0.38-0.63 mg L-1, 10.5-17.5 mg L-1, and 4.97-8.28 mg L-1, respectively. Comparison with research results from other global regions further supported the use of these thresholds, indicating that this method could be used in habitats beyond China. We found that the probability of cyanobacterial bloom was 0.75, a critical point for prevention and control. When this critical point was exceeded, cyanobacteria could proliferate rapidly, increasing the risk of cyanobacterial blooms. Changes in driving factors need to be rapidly controlled, based on these thresholds, to prevent cyanobacterial blooms. Temporal and spatial scales were critical factors potentially affecting the selection of driving factors. This method is versatile and can help determine the risk of cyanobacterial blooms and the thresholds of the principal driving factors. It can effectively predict and help prevent cyanobacterial blooms to reduce the global probability of occurrence, protect the health and stability of water ecosystems, ensure drinking water safety, and protect human health.