When people think about wildfire danger, they usually focus on hot weather, dry air, and strong winds.
But researchers believe another important warning sign may be hidden beneath our feet: the amount of moisture in the soil.
A team led by Salli Dymond, an associate professor in the School of Forestry at Northern Arizona University (NAU), is testing whether measuring soil moisture, along with the moisture inside trees and other forest plants, can improve wildfire forecasting.
If successful, the approach could help officials make faster and more accurate decisions about fire warnings, forest closures, and firefighting strategies.
This summer, Dymond is working with Salt River Project (SRP) and New Mexico startup Growvera to install moisture sensors throughout the southeastern Tonto National Forest in Arizona.
The sensors are attached to living plants, dead trees, fallen branches, and other forest materials that can fuel wildfires. Additional sensors are buried underground to measure soil moisture at different depths.
Over the next year, the research team will collect and analyze data from the sensors to see how well they reflect changing wildfire conditions.
Today, wildfire risk is mainly estimated using weather information such as temperature, humidity, and wind speed.
While these measurements are useful, Dymond says they do not always capture what is happening on the forest floor, especially in mountainous areas where local conditions can vary greatly.
Plants depend on water stored in the soil, and when the ground becomes dry, vegetation also dries out and becomes more likely to burn. Measuring both soil moisture and the moisture inside plants could provide a much clearer picture of how easily a wildfire might start or spread.
In the past, collecting this type of information was expensive because plant moisture sensors were costly.
However, advances in technology have made the equipment much more affordable. Growvera has developed a lower-cost sensor designed to measure moisture in plants, and the current project is testing whether it performs well enough for widespread use.
If the technology proves reliable, it could benefit both researchers and land managers. Forest agencies would gain access to more detailed information, while companies like Growvera could help make the technology widely available.
The researchers are also collecting data from a nearby weather station. By combining weather measurements with moisture readings from the soil and plants, they hope to better understand how forests respond to rainfall, changing temperatures, and humidity throughout the day.
The team expects to see that different plants react differently to changing conditions. Even forests located close together may experience very different moisture levels because rainfall can vary over surprisingly short distances.
One neighborhood may receive heavy rain during a summer storm, while another just a few kilometers away remains completely dry.
This local information could greatly improve wildfire management. Instead of issuing broad regional warnings, officials could identify neighborhoods or specific forest areas facing the highest fire danger. The data could also help firefighters decide whether to quickly extinguish a wildfire or allow a carefully managed fire to continue burning under safe conditions.
Graduate student Anastasiya Jonas from NAU is helping analyze the data, examining how soil moisture, plant moisture, dead vegetation, and weather conditions interact.
If the year-long trial is successful, Dymond hopes to expand the sensor network beyond the Tonto National Forest to fire-prone regions across the Southwest and eventually throughout other parts of the United States.
As wildfires become more frequent and destructive in many regions, researchers believe these small underground sensors could become valuable tools for providing earlier warnings, improving fire management, and helping protect forests, wildlife, communities, and firefighters from increasingly dangerous wildfire seasons.
