Spatiotemporal characterization and forecasting of coastal water quality in the semi-enclosed Tolo Harbour based on machine learning and EKC analysis

Characterizing and forecasting coastal water quality spatiotemporal evolution should be significant to ecosystem management. However, high-quality modeling their evolutions is rather challenging due complex dynamic mechanisms especially in a spatially temporally heterogenous semi-enclosed bay. To this end, study develops framework incorporating machine learning (ML) algorithms the Environmental Kuznets Curves (EKC) analysis model, analyze forecast variations of indicators for different subzones seasons Tolo Harbour Hong Kong. The application results indicate that developed ML-based with an accuracy range 0.672 ∼ 0.998 well-suited understanding harbour compared conventional approach. Furthermore, characteristics bay are analyzed discussed hydro-environmental Moreover, EKC also performed determining essential variables under 95% confidence interval Kong PCGDP projection then implemented model future prediction.