Title : A novel approach to estimate water quality of inland aquaculture ponds using satellite data and WQI model
Abstract:
Assessment of water quality in aquaculture ponds on land is essential for maintaining the wellbeing and productivity of aquatic ecosystems. A popular tool for thoroughly assessing the quality of surface water is the Water Quality Index (WQI) model. In order to assess water quality parameters in inland aquaculture ponds, this work introduces a novel method that integrates Sentinel-2 Multispectral Instrument (MSI) satellite data and the WQI model, improving the accuracy and effectiveness of water quality monitoring. The proposed approach integrates satellite imagery and machine learning techniques to predict key water quality parameters. The WQI model is used concurrently to combine various parameter predictions into a single index, offering a comprehensive evaluation of water quality. This hybrid methodology provides a solid framework for overcoming classic WQI models' drawbacks, such as region-specific applicability. An extensive dataset of in-situ water quality measurements gathered from several aquaculture ponds is used to validate the methodology. Strong relationships between the prediction models for individual parameters and the field measurements are shown with r-squared (R²) 0.71, 0.9, 0.2 and Root Mean Square Error (RMSE) of 0.19, 1.5, 0.027 for pH, Salinity, Ammonia respectively. A machine learning model was explored to predict the water quality parameters and the outcomes are measured against ground truth values. This validation emphasises how accurate the suggested method is for calculating and assessing water quality conditions with R² and RMSE of 0.68 and 0.12 .This innovative approach allows resource management, environmental preservation, and sustainable aquaculture practices by overcoming the constraints of conventional methods with integration of satellite data and WQI model on global scale.
Keywords: Satellite data, WQI, Shrimp farming, Machine learning
