A Futuristic Approach to Subsurface-Constructed Wetland Design for the South-East Asian Region Using Machine Learning

This study investigates the optimized design of horizontal flow constructed wetlands (HFCWs) to enhance pollutant removal efficiency while minimizing surface area requirements, particularly in the Southeast Asian region. By refining the first-order removal rate coefficient (k) for organics and nutrients, the research aims to meet specific performance benchmarks across three scenarios, ensuring compliance with discharge or reuse standards. Utilizing a data set comprising 1680 entries, five machine learning models-multiple linear regression (MLR), eXtreme Gradient Boosting (XGBoost), random forest (RF), artificial neural network (ANN), and support vector regression (SVR)-were employed to predict k values. Pearson's correlation, heat maps, and ANOVA analysis identified the most influential parameters affecting k-value predictions. The k values ranged from 0.01 to 0.52 per day using the P-k-C* method, essential for effective pollutant removal. The SVR model demonstrated the highest predictive accuracy, with R-2 values of 0.91 for k BOD, 0.90 for k TN, 0.82 for k TKN, and 0.76 for k (TP). This optimization reduced standard deviations significantly, from 136.90% to 2.28%. Consequently, the required wetland area was reduced by up to 68% for biochemical oxygen demand (BOD), 60% for TN (total nitrogen), and 67% for TP (total phosphorus) in larger systems, supporting the tailored design of HFCWs to meet targeted discharge standards.