Spatiotemporal modelling of PM$_{2.5}$ concentrations in Lombardy (Italy) -- A comparative study

TL;DR

This study compares three different predictive models—geostatistical, additive mixed models, and random forest kriging—for estimating PM2.5 air pollution levels in Lombardy, Italy, demonstrating similar effectiveness despite methodological differences.

Contribution

It provides a comparative analysis of traditional and machine learning models for spatiotemporal air pollution prediction, highlighting their strengths and limitations.

Findings

All models effectively captured spatiotemporal patterns.

Models showed similar out-of-sample performance.

Variable local performance depending on conditions.

Abstract

This study presents a comparative analysis of three predictive models with an increasing degree of flexibility: hidden dynamic geostatistical models (HDGM), generalised additive mixed models (GAMM), and the random forest spatiotemporal kriging models (RFSTK). These models are evaluated for their effectiveness in predicting PM$_{2.5}$ concentrations in Lombardy (North Italy) from 2016 to 2020. Despite differing methodologies, all models demonstrate proficient capture of spatiotemporal patterns within air pollution data with similar out-of-sample performance. Furthermore, the study delves into station-specific analyses, revealing variable model performance contingent on localised conditions. Model interpretation, facilitated by parametric coefficient analysis and partial dependence plots, unveils consistent associations between predictor variables and PM$_{2.5}$ concentrations. Despite nuanced variations in modelling spatiotemporal correlations, all models effectively accounted for the underlying dependence. In summary, this study underscores the efficacy of conventional techniques in modelling correlated spatiotemporal data, concurrently highlighting the complementary potential of Machine Learning and classical statistical approaches.