Adaptive agents for forecasting seasonal outbreaks of blue-green algal populations in lakes categorised by circulation type and trophic state

摘要

Adaptive agents for specific algal populations can be powerful tools for early warning and operational control of harmful algal blooms in lakes and drinking water reservoirs. One way of developing adaptive agents for algal populations is the extraction of generic rules from ecological time series data by means of evolutionary algorithms as suggested by RECKNAGEL ( 2003). This study aims to demonstrate that merged ecological data of categorised lakes allow evolutionary algorithms to induce rulebased agents that are generic for lake categories and facilitate short-term forecasting of blue-green algal blooms. Hybrid evolutionary algorithms (HEA) are applied in a rigorous k-fold cross-validation framework to develop agents for algal populations for categorised lakes. Three case studies are presented resulting in rule-based agents for forecasting 5- or 7- days-ahead abundances of Microcystis in the shallowpolymictic and hypertrophic lakes Kasumigaura and Suwa ( Japan) as well as in the warm-monomictic and hypertrophic lakes Hartbeespoort, Roodeplaat, and Rietvlei (South Africa), and abundances of Oscillatoria in the shallow-polymictic and hypertrophic lakes Veluwemeer and Wolderwijd (The Netherlands). Both explanatory and predictive validity of the resulting agents are tested by comparisons between predicted and measured data as well as sensitivity analyses.