Machine learning-based emulation of water quality index and pollution source apportionment in the Haraz river: A probabilistic risk assessment approach.

Linear SVR-lin best emulates deterministic RMS-WQI for Haraz River water, with sulfate (SO4 2−) the most influential input and RMS-WQI mostly “fair” (94.44%).

• Why it matters: Pollution from a nearby landfill and mining threatens river water quality, but managers need probabilistic risk assessment and reliable emulation of water-quality indices.
• What they did: Water samples upstream and downstream of landfill and mining sites were analyzed, then RMS-WQI was emulated using linear support vector regression; sulfate’s relative importance was quantified, and APCS-MLR performed source apportionment.
• The result: Geogenic hydrochemical processes dominated sources (64.66%), anthropogenic nutrient pollution contributed 35.34%, and children had the highest vulnerability with HI95th = 1.81.

AI summary of the abstract · may contain errors · not medical advice · read the paper.