The industrial augmentation and unguided anthropogenic activities contaminate water sources in most parts of the world especially in developing countries like Pakistan. High concentration of pollutants in groundwater affects human, soil, and crop health badly. The present study was conducted to investigate groundwater quality for drinking and irrigation purposes in an industrial zone of Pakistan. A GIS tool was used to investigate the spatial distribution of different physico-chemical parameters. In this study, the average results exceeding World Health Organization (WHO) and National Environmental Quality Standards (NEQS) were found for pH 7.84, total dissolved solids (TDS) 1492 mg/L, phosphate 0.51 mg/L, dissolved oxygen (DO) 9.92% saturation, F-coli 6.48 colonies/100 mL, Na⁺ 366 mg/L, HCO₃⁻ 771 mg/L, sulfate 251 mg/L, chlorides 427 mg/L, total hardness (as CaCO₃) 292 mg/L, electrical conductivity (EC) 2408 μS/cm, iron (Fe) 0.48 mg/L, chrome (Cr) 0.50 mg/L, arsenic (As) 0.04 mg/L, total phosphorus (TP) 0.17 mg/L, sodium adsorption ratio (SAR) 9.76 (in meq/L), residual sodium carbonate (RSC) 9.28 meq/L, % ion balance 14.4 (in meq/L), percentage sodium ion (% Na⁺) concentration 58.9 meq/L, and water quality index (WQI) 69.0. The trend of cations and anions were (in meq/L) Na > Mg > Ca > K and HCO₃ > Cl > CO₃ > SO₄ respectively. Although the results of the present study showed poor conditions of the groundwater for drinking as WQI but and irrigation purposes as SAR, it needs to improve some more conditions for the provision of safe drinking water and irrigation water quality.

Analyzing three-dimensional excitation-emission matrix (3D-EEM) spectra through machine learning models has drawn increasing attention, whereas the reliability of these machine learning models remains unclear due to their “black box” nature. In this study, the convolutional neural network (CNN) for classifying numbers of fluorescent components in 3D-EEM spectra was interpreted by gradient-weighted class activation mapping (Grad-CAM), guided Grad-CAM, and structured attention graphs (SAGs). Results showed that the original CNN classifier with high classification accuracy may make a classification based on misleading attention to the non-fluorescence area in 3D-EEM spectra. By removing Rayleigh scatterings in 3D-EEM spectra and integrating convolutional block attention module (CBAM) in CNN classifiers, the correct attention of the trained CNN classifier with CBAM greatly increased from 17.6% to 57.2%. This work formulated strategies for improving CNN classifiers associated with environmental fields and would provide great help for water determination in both natural and artificial environments.