Global analysis of mineral aerosol and industrial effects on lake water quality using spectral indices and machine learning.

The complex physical and chemical interactions between desert mineral aerosol, industrial pollutants, and lake water during drought periods pose significant challenges for accurately assessing lake water quality. This study analyzes Spectral Water Quality Index (S_WQI), Turbidity, Spectral Total Suspended Solids (S_TSS), and Spectral Secchi Disc Depth (S_SDD) from 1990 to 2024 as novel indices for water quality assessment. These indices replace conventional metrics (WQI, TSS, SDD) for monitoring lake water quality. These spectral indices, derived from satellite reflectance data (MODIS, Landsat), offer a new approach for monitoring water quality degradation caused by desert dust and industrial pollution. A Lake Dust Vulnerability Index (LDVI) is developed to assess lake sensitivity to desert aerosols in arid regions. The Bidirectional Long Short-Term Memory (BiLSTM) model, with correlations of R = 0.93, indicates that aerosols (PM₂.₅) during drought periods reduce water quality by up to 35 % in arid lakes like Urmia and Aral. Transboundary aerosols play a significant role in S_TSS, which exacerbates eutrophication by increasing chlorophyll a. By modeling dust-drought feedback cycles and integrating ERA5 data, this study provides a global framework for monitoring the world's lakes. These spectral indices, derived from satellite reflectance data, provide a groundbreaking approach for measuring water quality degradation caused by desert dust and industrial pollution.