Washington State 2012 Wildfire Smoke PM2.5 Exposure Estimates Using Blended Models

Washington State 2012 wildfire season data blending in-situ monitors, satellite observations, and chemical-transport model simulations to estimate population exposure to PM2.5 from wildfire smoke. The dataset, created by William Lassman, includes process data and code for ridge-regression and geographically weighted ridge-regression blending models. Performance of individual and blended exposure-estimation techniques was evaluated using leave-one-out cross-validation.

Use Cases

• Evaluate PM2.5 exposure estimation techniques based on the described blending of in-situ, satellite, and model data.
• Train or validate machine learning models for air quality prediction based on the described ridge-regression methods.
• Analyze spatial and temporal patterns of wildfire smoke exposure during a specific fire season based on the described geospatial blending.
• Assess the impact of monitor density on exposure model accuracy based on the described hypothetical scenario analysis.

Strengths

• Focuses on a specific geographic region and time period: Washington State during the 2012 wildfire season.
• Compares and blends three distinct exposure-estimation tools: in-situ measurements, satellite observations, and chemical-transport model simulations.
• Employs two specific modeling techniques for blending: ridge-regression and geographically weighted ridge-regression models.
• Includes process data and code, facilitating reproducibility of the analysis.

Limitations

• Row count, column definitions, and file formats are unknown, which limits suitability assessment.
• Last update date is unknown; freshness unverified.
• The description indicates blending provided only marginal improvements for this specific season due to high monitor density, potentially limiting its generalizability.

Provenance

Source

William Lassman

Collection Method

Blending of in-situ measurements, satellite-based observations, and chemical-transport model simulations using regression models.

Time Range

2012 wildfire season

Geography

Washington State, USA