Algorithms for source location source location and chemical oxygen demand concentration over time along the river

The global use of freshwater has increased six-fold in the past 100 years and has been growing by about 1 % per year since the 1980s. Meanwhile, rapid urbanization, climate change, and inefficient water management have contributed to a 20 % decline in per capita freshwater availability in recent decades reducing the water resources even more. The scarcity of water in the coming decades is exacerbated by human actions, including thousands of accidental pollution spills and intentional illegal discharges into surface waters. To address this challenge, a Monitoring and Mitigation Framework (MMF) was developed to minimize damage from contaminant spills, identify contaminant sources in real time, and estimate Chemical Oxygen Demand (COD) flow concentration rates in real time. The MMF was specifically designed to detect environmental crimes, track companies responsible for contamination spills, and facilitate the scheduling of mitigation actions in coordination with law enforcement authorities for ongoing and future events. This real-time framework uses Inverse Estimation (IE) algorithms built with Artificial Neural Networks algorithms to detect contamination source location and estimate COD flow concentration at source location over time. This provides real-time evidence for identifying potential companies involved in environmental crimes according to local police of the Malaga City Council (Spain). Both algorithms were trained on a Learning Database build from over 13,000 simulations, and generated by an algorithm called Launching Multiple Scenarios algorithm also developed in this work. The MMF was developed and tested at Llobregat river (Barcelona, Spain), and it was considered 2621 test cases with different spill location and COD concentrations ranging from 0.1 to 80,000 mg/l. In those tests, IE algorithms were able to estimate COD concentration at the source with a mean absolute error (MAE) lower than 2.6 mg/l and a Mean Square Relative Error (MSRE) lower than 10 mg/l, correctly identifying the contamination source location in more than 85 % of cases. • Monitoring and Mitigation Framework (MMF) created for contamination spill response. • MMF module uses Artificial Neural Network algorithms. • ANN algorithms were trained on a Learning Database with 13,000 + simulations. • MMF tested with 2621 cases at Llobregat River, ranging from 0.1 to 80,000 mg/l COD. • IE algorithms achieved MAE under 2.6 mg/l, correctly identifying sources in 85 %+ cases.