Statistical Interpolation for Mapping Wastewater-Derived Pollutants in Environmental Systems: A GIS-Based Critical Review and Meta-Analysis

Effective management of wastewater discharges requires understanding the spatial distribution of pollutants both within engineered infrastructure and in receiving environments. However, spatial data sparsity constrains comprehensive assessment. This critical review examines the role of Geographic Information Systems (GIS) and statistical interpolation techniques in bridging these data gaps for wastewater-derived pollutants. Moving beyond a simple compilation of methods, this paper provides a synthesizing framework that categorizes and evaluates interpolation techniques-from deterministic and geostatistical approaches to emerging machine learning (ML) and hybrid models- based on their ability to address specific challenges in wastewater systems. A key contribution is a systematic review and meta-analysis following PRISMA guidelines, synthesizing evidence from 22 studies that directly compare interpolation methods for wastewater-relevant parameters (BOD5, COD, nutrients, heavy metals) in both engineered systems and impacted water bodies. Results indicate that machine learning methods significantly outperform traditional approaches, with a pooled 21% reduction in RMSE compared to Ordinary Kriging (95% CI: 15–27%). However, subgroup analyses reveal context dependency: ML advantages are most pronounced for organic pollutants (29% reduction) and data-rich environments (27% reduction with n > 100), while geostatistical methods remain competitive for physical parameters (8% reduction, non-significant) and data-sparse scenarios (12% reduction with n < 50). Co-Kriging achieves 15% RMSE reduction over Ordinary Kriging when auxiliary variables are available. The review explores applications in pollutant tracking, infrastructure planning, and environmental impact assessment, highlighting how integration of real-time sensor data (IoT) and remote sensing is transforming static maps into dynamic monitoring tools. Finally, a forward-looking research roadmap is presented, emphasizing hybrid modeling frameworks, digital twin integration, and improved uncertainty communication for decision support. By quantitatively synthesizing the current state-of-the-art and identifying critical knowledge gaps, this review aims to guide future research towards more intelligent, adaptive, and reliable spatial assessments of wastewater-derived pollutants.