Automating water quality monitoring in aquaculture ponds: drone-based evidence from Tamil Nadu, India

Abstract

This study aimed to determine the effectiveness of drone-based autonomous systems, in combination with autonomous surface vehicles, autonomous underwater vehicles, Internet of Things systems and artificial intelligence applications, in real-time water quality monitoring. The study was performed over six months using experimental and control aquaculture ponds at the Aquaculture Research Facility in Pollachi, Tamil Nadu, India. Water quality parameters (dissolved oxygen, pH, turbidity, temperature, ammonia and nitrates) estimated using autonomous systems were compared to those monitored via conventional laboratory-based techniques. Machine learning algorithms were designed to use the assimilated data to identify and predict future anomalies and intervene. The average dissolved oxygen and pH in the experimental ponds (8.5 ± 0.7 mg l^-1 and 7.5 ± 0.2, respectively) was much higher than that in the control ponds (6.9 ± 1.1 mg l^-1 and 7.1 ± 0.4), and the turbidity was lower in the experimental ponds (80 ± 10 NTU) than in the control ponds (95 ± 15 NTU). After intervention, the dissolved oxygen levels in the hypoxic areas increased from 5.5 to 7.8 mg l^-1 following aeration, indicating a 20–25% increase in 30 min, while the level of ammonia decreased from 1.2 to 0.9 mg l^-1 in 24 h, indicating that the bio-remediation agents were effective. These improvements translated into an increase in fish survivability (95%) and biomass gain (25%). These outcomes show that autonomous monitoring with the help of autonomous vehicles and artificial intelligence-based decision support can help to ensure timely intervention, enhance resource utilisation and increase sustainability in aquaculture.