Dataset Information

Advancing methodologies for applying machine learning and evaluating spatiotemporal models of fine particulate matter (PM_2.5) using satellite data over large regions.

ABSTRACT: Reconstructing the distribution of fine particulate matter (PM_2.5) in space and time, even far from ground monitoring sites, is an important exposure science contribution to epidemiologic analyses of PM_2.5 health impacts. Flexible statistical methods for prediction have demonstrated the integration of satellite observations with other predictors, yet these algorithms are susceptible to overfitting the spatiotemporal structure of the training datasets. We present a new approach for predicting PM_2.5 using machine-learning methods and evaluating prediction models for the goal of making predictions where they were not previously available. We apply extreme gradient boosting (XGBoost) modeling to predict daily PM_2.5 on a 1×1 km² resolution for a 13 state region in the Northeastern USA for the years 2000-2015 using satellite-derived aerosol optical depth and implement a recursive feature selection to develop a parsimonious model. We demonstrate excellent predictions of withheld observations but also contrast an RMSE of 3.11 μg/m³ in our spatial cross-validation withholding nearby sites versus an overfit RMSE of 2.10 μg/m³ using a more conventional random ten-fold splitting of the dataset. As the field of exposure science moves forward with the use of advanced machine-learning approaches for spatiotemporal modeling of air pollutants, our results show the importance of addressing data leakage in training, overfitting to spatiotemporal structure, and the impact of the predominance of ground monitoring sites in dense urban sub-networks on model evaluation. The strengths of our resultant modeling approach for exposure in epidemiologic studies of PM_2.5 include improved efficiency, parsimony, and interpretability with robust validation while still accommodating complex spatiotemporal relationships.

SUBMITTER: Just AC

PROVIDER: S-EPMC7591135 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Chiang Mai encounters severe pollution during the wildfire season. Wildland firefighters encounter various hazards while engaged in fire suppression operations, which encompass significant exposure to elevated concentrations of air pollutants resulting from combustion, especially particulate matter. The adverse effects of wildfire smoke on respiratory health are a significant concern. The objective of this study was to examine the potential adverse effects of PM2.5 exposure on the respiratory function and DNA damage of wildland firefighters. This prospective cohort study conducted in Chiang Mai from January to May 2022 planned to evaluate the health status of wildland firefighters during the pre-peak, peak, and post-peak ambient air pollution seasons. The measurement of PM2.5 was done at every forest fire station, as well as utilizing data from the Pollution Control Department. Participants received general health examinations, spirometry evaluations, and blood tests for DNA damage analysis. Pair t-tests and multiple regression models were used to examine the connection between pulmonary function parameters (FVC, FEV1) and PM2.5 concentration, with a significance level of P < 0.05. Thirty-three peak-season and twenty-one post-peak-season participants were enrolled. Four pre-peak-season wildland firefighters had FVC and FEV1 declines of > 15%. Multiple regression analysis showed a negative association between PM2.5 exposure and FVC% predicted (- 2.81%, 95% CI - 5.27 to - 0.34%, P = 0.027) and a marginally significant negative correlation with FVC (- 114.38 ml, 95% CI - 230.36 to 1.59 ml, P = 0.053). The remaining pulmonary measures showed a statistically insignificant decline. There were no significant changes in DNA damage detected. Wildland firefighters suffered a significant decline in pulmonary function associated with PM2.5 exposure. Spirometry is crucial for monitoring and promptly identifying respiratory issues that occur during wildfire seasons. Further research is recommended to explore DNA damage alterations and their potential association with PM2.5.

Dataset Information

Advancing methodologies for applying machine learning and evaluating spatiotemporal models of fine particulate matter (PM_2.5) using satellite data over large regions.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets

Dataset Information

Advancing methodologies for applying machine learning and evaluating spatiotemporal models of fine particulate matter (PM2.5) using satellite data over large regions.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets

Advancing methodologies for applying machine learning and evaluating spatiotemporal models of fine particulate matter (PM_2.5) using satellite data over large regions.