Wildfire Air Quality Prediction: A Data-Driven Approach

Wildfire Air Quality Prediction: A Data-Driven Approach

Subhankar Dhar, Jerry Zeyu Gao
Copyright: © 2023 |Pages: 22
DOI: 10.4018/IJDREM.330148
Article PDF Download
Open access articles are freely available for download

Abstract

Wildfires are extremely harmful to the environment. While producing gaseous pollutants and particles that cause smoke, wildfires also release carbon dioxide (CO2), a greenhouse gas that will continue to warm the planet after the wildfire ends. This article delves into the impact of wildfires and air quality on human living conditions. The authors' machine learning models use wildfire data to forecast air quality with detailed indexes and geographic information during a wildfire. The work evaluates the performance of each machine learning model via statistical metrics like mean absolute error (MAE), R-squared (R2), and root mean squared error (RMSE). The experimental results used neural networks to predict a specific value for carbon monoxide (CO), ozone, and PM2.5. These are both promising and accurate, providing meaningful insight into air quality within a region. This work will be useful for cities, governments, citizens, and public safety.
Article Preview
Top

Preisler et al. (2015) highlighted the shortcomings of research when predicting wildfire impacts from PM2.5 concentration at ground-level monitors in California. While most researchers rely on satellite-based observational tools, this work combined models with an autoregressive statistical model, incorporating weather and seasonal factors to identify thresholds for predicting unusual events. The study focused on ground-based monitoring of PM2.5 levels, with data consisting of hourly values of PM2.5 and meteorological data. Data was gathered from the United States Department of Agriculture Forest Services. Unexpectedly, the study found that smoke plumes could identify seasonal wildfire influence with high accuracy.

Complete Article List

Search this Journal:
Reset
Volume 7: 1 Issue (2024): Forthcoming, Available for Pre-Order
Volume 6: 1 Issue (2023)
Volume 5: 2 Issues (2022): 1 Released, 1 Forthcoming
Volume 4: 2 Issues (2021)
Volume 3: 2 Issues (2020)
Volume 2: 2 Issues (2019)
Volume 1: 2 Issues (2018)
View Complete Journal Contents Listing