20. Long-Term Ecological Research in the Nakdong River: Application of Ecological Informatics to Harmful Algal Blooms

In recent decades, the importance of long-term ecological research (LTER) has been highlighted because of the growing interest in global environmental changes. Specifically, LTER data allows one to track the history of target ecosystems (e.g., trends of particular ecological entities) and enables one to understand the causal relationships of ecosystem functioning. One ecological problem is harmful algal blooms (HABs) in freshwater environments. It is generally perceived that global warming and local eutrophication are responsible for serious and frequent HAB events, and various efforts have been made to explain and forecast HABs. LTER data for HABs typically consist of various forcing functions and variables; thus, the selection of appropriate data-analysis methods for a HAB database is necessary. This chapter presents a series of studies related to the prediction and elucidation of two HABs, such as summer cyanobacteria (e.g., Microcystis aeruginosa) and winter diatom (e.g., Stephanodiscus hantzschii) that occur in the regulated Nakdong River, South Korea. First, HABs, water quality, and zooplankton patterns were analyzed using self-organizing maps (SOMs). Those major factors that have a close relationship to HABs, i.e., water temperature, pH, and rainfall, were selected. We created a predictive model and control scenario for HABs using a variety of methods (evolutionary computation, artificial neural network) in the real world based on confirmed information. We also suggest potential further studies of the Nakdong River.

This chapter focuses on: (1) properties of the limnological dataset of the Nakdong River derived from Korean Long-Term Ecological Research (KLTER), (2) analysis and time-series modelling of KLTER dataset by means of machine learning techniques, and (3) benefits of applied ecological informatics for KLTER dataset.