This study focuses on identifying GW-SW interaction locations in a tropical lake - Vellayani Lake (VL), Southwest India, utilizing stable water isotopes (s18O, sD) and chloride mass balance approach. The northern lake region was identified as a critical groundwater discharge “hotspot” with pronounced discharge (2.14×106 - 3.82×106 m³/yr), prompting targeted management interventions. This reaffirms the critical role of groundwater inflow in sustaining the lake’s water balance. Additionally, the application of machine learning (ML) techniques refined the classification of Lacustrine Groundwater Discharge (LGD) and non-LGD sites while predictive modeling utilizing Sensitivity Indices enhanced the understanding of prominent factors influencing lake volume. K-means clustering and Random Forest (RF) classification, achieved high accuracy (90%) and a kappa value of 0.8 in distinguishing groundwater discharge and non-discharge sites. Predictive modeling and sensitivity analysis revealed precipitation as the most influential factor, with a ±20% change causing a 16.69% variation in lake volume. Groundwater discharge exhibited a sensitivity index of 0.5320, further emphasizing its critical role in maintaining lake hydrological balance. This integrated approach provided valuable insights into the critical role of nearshore groundwater recharge in maintaining lake hydrological balance and facilitates the identification of suitable areas for groundwater recharge structures. For practitioners and policymakers, this integrated approach offers a robust framework for identifying critical GW-SW interaction zones, prioritizing groundwater recharge areas, and designing sustainable water management strategies, especially in data-scarce regions, paving the way for improved resource management in similar tropical lake environments.
Read full abstract