Urban stormwater: A key pathway for microplastic pollution
By Amir Reshadi
Urban areas are increasingly recognized as major contributors to microplastics (MP) pollution, with stormwater runoff serving as a primary transport pathway. While many studies have documented the presence of MPs in stormwater, understanding the underlying causes and accurately forecasting contamination levels requires robust modeling efforts.
A recent effort by members of the Microplastics Fingerprinting project addresses this gap by introducing the first global dataset of stormwater MP concentrations, compiled from 107 case studies across 11 countries. In a recently published paper, Amir Reshadi and his co-authors integrated the case study dataset with global hydrometeorological and socioeconomic information to develop the first machine learning (ML) model capable of predicting MP concentrations downstream of urban stormwater catchments and identifying key factors driving contamination.
A Data-Driven Approach to Understanding Microplastics
The model’s strength lies in its incorporation of diverse environmental and socioeconomic factors, allowing it to capture the complex dynamics of urban MP pollution. Through scenario testing and feature interaction analysis, the ML approach validates and refines assumptions about MP sources and transport pathways
Key findings emphasize the importance of standardized MP definitions and consistent catchment data in improving predictive accuracy. Beyond MP size, hydrological, economic, meteorological, and human activity factors emerge as strong predictors of stormwater MP concentrations. The research reveals that highly urbanized catchments with intense human activity and hydrometeorological conditions conducive to MP transport exhibit significantly higher MP levels in stormwater.
Implications for Urban Planning and Policy
The findings highlight the urgent need to retrofit stormwater systems in densely populated areas. More importantly, the model serves as a useful tool for policymakers and urban planners, enabling data-driven prioritization of resources—particularly in regions with limited capacity for extensive field sampling. By replacing inefficient trial-and-error approaches, this model facilitates more effective and targeted MP mitigation strategies.
Forecasts project a 1.5 to 2.6-fold increase in MP leakage by 2040, underscoring the need for proactive management. Refining models to incorporate MP characteristics—size, shape, texture, and composition—can enhance predictive accuracy and guide targeted interventions. As MP-related risks to ecosystems and human health become clearer, effective monitoring strategies will be essential.
Reshadi et al.’s ML-based approach provides a scalable solution for estimating MP concentrations in urban areas with limited empirical data, reducing reliance on costly field sampling. This data-driven framework enables researchers to make continuous model improvements, improving accuracy over time. For decision-makers, modeling stormwater catchments can help them identify MP hotspots, develop effective mitigation strategies, and support more efficient sampling efforts.
Photo: Stormwater management pond by the City of Brampton.