Research on algae blooms forecasting based on the multivariate data driven method: A case study of the Chaohu Lake

Abstract

The data driven is one of the main methods for forecasting algae bloom, which requires lots of continuous and accurate monitoring data. It is an effective way to increase sample data size by combining in-situ, and remote sensing data. The Chaohu Lake was taken as the case study. Based on water quality data (TLI), meteorological data (sunshine duration, temperature, wind speed, wind direction) and bloom grade data, provided respectively by remote sensing and in-situ monitoring, an artificial neural network was employed to build empirical data-driven models. The model accuracy was evaluated by algae bloom grade recognition rate and bloom trend recognition rate. The results showed that the bloom grade recognition rate of model driven by remote sensing data was better than others. Bloom trend recognition rate of model driven by in-situ data is higher than others. These results provide some insights for algae bloom forecasting.