This project analyzed Klamath River water quality, focusing on how weather and river flow affect thermal stratification, crucial for aquatic health. Generalized Additive Models (GAM) captured seasonal, nonlinear interactions, while Random Forest models handled complex, multicollinear data, identifying key predictors. K-means clustering established temperature thresholds to detect stratification periods, providing insights for effective river health management.
Data Integration: Merged weather, flow, and water quality datasets, resolving missing values and aligning timestamps for cohesive analysis.
Exploratory Analysis: Uncovered seasonal and daily patterns affecting water temperature and stratification.
Nonlinear Modeling: Utilized GAM to examine complex interactions between weather and water conditions impacting stratification.
Predictive Modeling: Applied Random Forest to assess variable importance and predict stratification events.
Stratification Detection: Used k-means clustering to set temperature thresholds, identifying key periods of stratification and reverse stratification.