Wildfire Smoke Plume Injection Fractions and Pollution Impacts Across Western US
by Feng, Xu / UCLA Dataverse·Updated 27d ago
Available on 1 platform
Sign in to view source links and access this dataset
Description
Dataverse at UCLA hosts datasets and scripts for the study 'From Satellite Observations to Machine Learning: Predicting Plume Injection Fraction and Its Impacts on Smoke Pollution Across the Western United States'. The collection includes GEOS-Chem model simulations, GFED fire emissions, MISR plume height data, EPA and IMPROVE air quality observations, and random forest models for predicting daily plume injection fractions. The data was last updated on June 8, 2026.
Use Cases
Train machine learning models to predict plume injection fraction based on satellite-derived plume heights and fire emissions data.
Validate GEOS-Chem atmospheric chemistry model simulations against EPA and IMPROVE ground-based pollution observations.
Analyze the impact of wildfire smoke injection height on surface-level PM2.5 concentrations across the western United States.
Reproduce figures and analyses from the associated research manuscript using the provided scripts.
Strengths
Includes data from multiple authoritative sources: GEOS-Chem model outputs, GFED fire emissions, MISR satellite plume heights, and EPA/IMPROVE observations.
Covers a multi-year time period for plume injection fractions (2008-2020) and includes specific simulation experiments for 2020.
Provides the trained random forest models and input/output variables used for prediction, enabling reproducibility.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Data may reflect geographic bias inherent to its focus on the western United States and North America.
Provenance
Source
UCLA Dataverse, authored by Feng, Xu.
Collection Method
Combines atmospheric model simulations (GEOS-Chem), satellite observations (MISR), ground-based monitoring (EPA, IMPROVE), and machine learning model outputs.
Time Range
Plume injection fraction data spans 2008-2020; core model simulations are for 2020.
Freshness
Last updated 2026-06-08 06:23:42; freshness should be verified.
Geography
North America, with a focus on the western United States for pollution observations.
License restrictions are unknown and should be verified prior to use.