Detection of estuarine benthic macroinvertebrates communities using artificial neural network

Aquatic ecosystems depend heavily on benthic macroinvertebrates for their essential functions. They have become an indispensable part of pure water bio monitoring due to the widespread recognition of their capacity to recognize many kinds of anthropogenic stresses. A waterbody’s biological health can be determined by estimating the diversity and abundance of benthic macroinvertebrates. Macroinvertebrate biomonitoring depends on the manual identification process, which is time-consuming and expensive and requires highly qualified taxonomical experts. This research examines the effectiveness of the Artificial Neural Network (ANN) for the automatic detection of benthic macroinvertebrates. This study has investigated the prevalence and absence of fifteen species of benthic macroinvertebrates in a tropical estuary environment to enhance the extraction of information embedded in benthic macroinvertebrate data for assessing community dynamics. The station number, month, pore water salinity, and pore water p H are used to assess the detection of benthic macroinvertebrates employing Multilayer Perceptron Neural Network (MLPNN). It has been found that the accuracy rates of MLPNN in the prediction of fifteen species are very promising. The suggested model can be utilized to recognize macroinvertebrate species with high accuracy, and the resultant model can outperform manual taxonomist identification.