Implementation of artificial neural networks in patterning and prediction of exergy in response to temporal dynamics of benthic macroinvertebrate communities in streams

Exergy is an effective, single measurement to express the information level of communities, while the trends of community dynamics are difficult to represent since communities consist of different species varying in a complex manner. Using the data concerning benthic macroinvertebrate communities collected from streams, we implemented artificial neural networks in patterning and predicting exergy by utilizing the networksʹ feasibility of information extraction and self-organization. The time development of exergy measured at the sample sites was patterned through training by the Kohonen network. The trained mapping was able to characterize the development trend of exergy at different groups of sample sites in different time periods. The on-time and time-delayed trainings on community–exergy relations were also conducted by the backpropagation algorithm, and it was possible to predict exergy at contemporaneous and subsequent sampling times.