EnviroHarvest-OptiNet: A Hybrid Optimization and Predictive Modeling Framework to Assess the Environmental Impact of Rainwater Harvesting

Sustainable water management through rainwater harvesting proves essential for Chennai's urban areas because they continuously face water scarcity together with flooding issues. The research presents EnviroHarvest-OptiNet as a combined optimization and predictive modeling tool that evaluates rainwater harvesting's environmental consequences. The system combines Geospatial Information System (GIS) mapping with hybrid optimization through Genetic Algorithm + Particle Swarm Optimization methods to select sites and XGBoost and LSTM machine learning algorithms for automated groundwater recharge prediction. The research analyzes ten rainwater harvesting sites that achieved an average optimization success rate of 0.97. The groundwater recharge prediction proficiently works through an ensemble model and XGBoost $(\mathrm{R}^{2}=0.9652)$ along with LSTM $(\mathrm{R}^{2}=0.9784)$ to achieve $\mathrm{R}^{2}=$ 0.9911. Research into centralized versus decentralized management systems demonstrates that central approaches maintain sustainability at 0.99 yet decentralized methods have a sustainability level of 0.95. A thorough environmental impact assessment outputs a sustainability score of 0.9814 which supports the effectiveness of rainwater harvesting systems for water preservation and flood management and natural ecosystem revival. Through the proposed framework policymakers obtain data-based decisions which advance their capabilities to strengthen water security and urban resilience.