Prediction of Trophic State Index of Lentic Water Bodies Using Artificial Intelligence- A Case Study

Abstract

Eutrophication of lentic water bodies have significant impact on the natural environment and human life. The quantities of nitrogen, phosphorous and other biologically useful nutrients are the primary determinants of a water body’s Trophic State Index (TSI). The artificial intelligence (AI) methods have enhanced the estimation of Eutrophication status of lentic systems like fresh water lakes. This study uses three artificial intelligence methods such as M5 Model Tree, Support Vector Machine (SVM) for the prediction of Trophic states of Sasthamcotta fresh water Lake, Kerala. Four different cases were considered in the study via prediction of individual trpohic status (TSI-Ph, TSI-Ch1-a and TSI-Sechi Depth) along with the classical Carlson’s Trophic State Index (CTSI). Several statistical measures were used to quantify the performance of the three AI methods. Both the tree based algorithms were found to be successful in accurate prediction of TSI with Random Forest method as the best among them.