Wildfire behavior is complex and challenging to accurately predict. Winds, air temperature, humidity, and precipitation, for example, influence the flammability of fuel and largely determine the risk of fire ignition. In addition, wind speed and direction determine the rate of fire spread. Heat from wildfires causes rising air currents that strongly modify local weather patterns and create rapidly changing winds that may fan the fire. Compounding the problem, climate change is increasing the frequency, extent, and severity of fires by creating warmer and drier conditions.
As wildfire conditions continue to evolve in the fire-prone western United States, efficient and accurate modeling becomes more important than ever. However, existing global Earth system models have lacked the necessary resolution to accurately simulate wildfires. They also struggle to incorporate high-resolution fire emissions and leave out critical processes, such as how particles from the burning biomass (aerosols) interact with other gases and particles in ways that can affect cloud formation, precipitation, weather, and even regional climate.
LLNL researchers have aimed to solve this problem by enhancing existing modeling capabilities and connecting them with novel observational data to achieve a new framework for wildfire simulation. The framework focuses on large wildfires, which have a more significant impact on weather and climate through cloud formation and negative radiative forcing (atmospheric cooling).
The researchers developed a California regionally refined model (CARRM) configuration for the Department of Energy’s Energy Exascale Earth System Model (E3SM) with 3 km finest resolution—a vast improvement over the 100 km global resolution for E3SM Version 2 and the 25 km resolution for the United States RRM. They then incorporated advanced gas and aerosol chemistry features into CARRM in order to model the evolution of smoke in the stratosphere, as well as features to support simulation of smoke plume rise and surface temperature increases due to wildfire. Finally, the team linked hourly High-Resolution Large-Fire Database satellite data to CARRM to generate more accurate fire simulations.
Using several historical wildfires as test cases, the team compared E3SM-CARRM results to observational data. Results were promising, achieving reasonable accuracy without sacrificing computational efficiency (1 month of simulated time can be completed per day of computing time). With this throughput rate, multi-year ensemble simulations are affordable and uncertainty analyses feasible.
Coming soon
The team has established external collaborations with several wildfire observational data groups. With access to more observational datasets, the team continues to test and enhance the framework. Eventually, they plan to use it to model other significant events involving large-scale aerosol emissions, such as volcanic eruptions, nuclear winter, or climate-related geoengineering. Furthermore, they are linking this framework to downstream models to tackle important science and socioeconomic questions relating to energy infrastructure, the clean-energy transition, and human health.
