Science Inventory

Sensor data cleaning and correction: Application on the AirNow Fire and Smoke Map

Citation:

Johnson Barkjohn, K., A. Holder, A. Clements, S. Frederick, AND R. Evans. Sensor data cleaning and correction: Application on the AirNow Fire and Smoke Map. To be Presented at American Association for Aerosol Research Conference, Albuquerque, NM, October 18 - 22, 2021.

Impact/Purpose:

Interest and concern about air quality has grown in recent years especially during wildfires. Simultaneously, growth in the popularity and use of air sensors across the US has also occurred. However, with so many sources of sensor and governmental data it can be challenging for users to interpret. In this presentation we will share the methods behind incorporating the PurpleAir sensor data on the AirNow Fire and Smoke map and how this map is used to convey important information to the public. This abstract is for a presentation that will be presented to researchers at the American Aerosols Association Conference Oct 18-22 who are interested in communicating aerosol information to the public in an effective way.

Description:

As smoke from wildfires becomes a larger public health concern, residents of smoke impacted communities frequently use lower-cost air sensors to provide more localized air quality data. However, it can be challenging for the public to understand data accuracy, the impact of averaging time, and ultimately which source to trust and how to interpret the data to protect health. In 2020, the United States Environmental Protection Agency (US EPA) and the US Forest Service began the sensor data pilot showing sensor data alongside permanent and temporary monitors on the AirNow Fire and Smoke map (fire.airnow.gov). Data cleaning methods and a correction equation were applied to PurpleAir sensor fine particulate matter (PM2.5) data to make it more comparable to AirNow monitors thus reducing data accuracy concerns and streamlining several discrete data sources to help users view and interpret data.  The original correction equation was evaluated under typical US ambient conditions and smoke impacted conditions with PM2.5 concentrations up to 250 µg m-3. However, smoke conditions were extreme during 2020 wildfires with some sites seeing PM2.5 concentrations over 1,000 µg m-3 with sensors underestimating concentrations above 250 µg m-3. A collocated dataset collected during the 2020 wildfire season was used to extend the correction equation into the higher concentration regime. Since US EPA primarily communicates air quality data using the Air Quality Index (AQI), a piecewise correction equation was selected to reduce the error at each AQI category break point. This correction enables PurpleAir PM2.5 data to be used with greater confidence and will provide the public with more spatially resolved data on the Fire and Smoke map. Although this abstract was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

Record Details:

Record Type: DOCUMENT ( PRESENTATION/ SLIDE)
Product Published Date: 10/22/2021
Record Last Revised: 02/02/2022
OMB Category: Other
Record ID: 353088