Using Low-Cost Sensors for Fenceline Monitoring to Measure Emissions from Prescribed Fires

Area of Science:

• Environmental Science
• Atmospheric Chemistry
• Forestry Management

Background:

• Prescribed burning effectively reduces wildfire risk but releases harmful pollutants.
• Accurate quantification of prescribed fire emissions is crucial for atmospheric modeling and assessing community impacts.
• Traditional cumulative emission factors fail to capture real-time emission variability during combustion events.

Purpose of the Study:

• To evaluate and compare two methods for detecting prescribed fire plumes: an event detection algorithm and a random forest model.
• To determine which method provides more robust emission factor calculations.
• To assess the utility of fenceline monitoring for measuring prescribed fire emissions.

Main Methods:

• Implementation and evaluation of an event detection algorithm for plume identification.
• Development and application of a random forest model for prescribed fire plume detection.
• Utilizing fenceline monitoring equipment to collect emission data during prescribed fires.

Main Results:

• The random forest model achieved 61% detection accuracy with a 3% false positive rate.
• The event detection algorithm showed 51% accuracy and a 31% false positive rate.
• The random forest model demonstrated superior performance in plume detection compared to the event detection algorithm.

Conclusions:

• The random forest model offers a more robust framework for calculating emission factors from prescribed fires.
• Accurate plume detection is essential for understanding emission variability and impacts.
• This study highlights the value of fenceline monitoring in capturing real-time prescribed fire emissions.