Playing with Fire
Striving for better fire predictions of fire weather and its effects on fire behavior
Spring in South Carolina is normally mild and moist, with weather that entices trumpet flowers to bloom- not burn. But in April last year, emergency managers had to evacuate thousands of people from their homes in the path of a Myrtle Beach wildfire. The fire, fueled in part by the weather, burned and jumped over highways and canals and caused $16 million in property damage.
ESRL's Sher Schranz (Global Systems Division, GSD) finds motivation in such unexpected fire behavior. As the Myrtle Beach fire demonstrated, protecting the public and firefighters requires a better understanding of how weather affects fire and fire affects weather.
"You need a model that couples weather physics and fire dynamics," Schranz said, and that's what her team is now working toward.
Such research is increasingly critical for public safety, according to a NOAA Science Advisory Board Report published last year, "Fire Weather Research: A Burning Agenda for NOAA." Fire seasons have recently become longer and the fires more severe, the authors noted, and they called for better fire data and fire weather information "to serve the American public."
Schranz has been helping NOAA and the National Weather Service (NWS) develop fire weather prediction tools for nine years. Her team developed the FX-Net workstation, which provides weather data to Incident Meteorologists (IMETs) during wildfires. IMETS use FX-net to brief fire behavior analysts and fire Incident Commanders onsite.
This summer for the first time, GSD will run an experimental, Western US weather model, the High Resolution Rapid Refresh (HRRR)/Chem/ Smoke at 3-km resolution during wildfire season, in support of smoke and fire management operations.
"The HRRR has plume rise and fire emissions dispersion analysis and forecasts that are a big step forward for firefighting operations," Schranz said. "The 3-km winds over complex terrain, which, combined with the smoke forecast, provides critical situational awareness for the movement of fire fighters and emergency evacuation planning and execution." The U.S. Forest Service and NWS FX-Net users across the continental United States will have access to the 3-km HRRR model data.
Schranz anticipates NOAA and its partners will continue to make significant improvements to fire weather prediction. She's building collaborations with researchers who model fire behavior at 1-meter resolution. Linking those fire models to weather models will necessitate extremely fine resolution weather models, down to 100 m, Schranz said, which will require resources for dedicated downscaling research and increased access to high performance computers.
She's also seeking to use new technologies such as unmanned aircraft systems (UAS) to collect over-the-fire data essential for model verification. In 2006, a NOAA pilot study showed that a UAS could collect data over a southern California wildfire, but the plane carried only chemical sensors. To understand how well various fire models are behaving, researchers need atmospheric instruments to capture heat flux, relative humidity wind and pressure fields. These fields will also be used by forecasters in real time. "We really need relative humidity observations at night," for example," Schranz said. "If relative humidity drops at night, fires can flare substantially or suddenly."