MC1 DGVM Seasonal Fire Forecasting

James Lenihan, Raymond Drapek, Ronald Neilson, Christopher Daly, and Dominique Bachelet

USDA Forest Service Pacific Northwest Research Station (1-3,5), and Oregon State University (4)

For nearly a decade the Mapped Atmosphere-Plant-Soil System (MAPSS) team from the USDA Forest Service Pacific Northwest Research Station in Corvallis, Oregon has developed dynamic general vegetation models (DGVMs) for forecasting the potential impacts of climate change on terrestrial ecosystems. The MC1 DGVM includes a fire disturbance module that generates process-based simulations of fire occurrence, behavior, and effects. The fire module was developed to simulate fire effects over broad spatial scales and extended time frames to address critical issues in wildland fire and fuel management over the coming decades. These include the relative role of climate and human intervention (e.g., fire suppression) in the long-term dynamics of fuels and fire behavior, and tradeoffs between efforts to restore natural fire regimes and the sequestration of carbon to offset greenhouse gas emissions. More recently, the MAPSS team has also been using the state-of-the-art level of ecosystem and fire modeling in the MC1 DGVM to produce forecasts of fire risks and impacts over the conterminous United States for upcoming fire seasons.

The fire forecasting requires monthly updating of the observed climate across the United States, up to the most recent month available and updated every month thereafter. Observed data from a dense network of weather stations are first subject to rigorous quality control, and then interpolated to complex terrain by the PRISM model as a continuous climate surface on a grid of cells approximately 4-5 km on a side and then aggregated to a grid of about 50-kilometer per side. The future climate forecasts are available through cooperation with the International Research Institute for Climate Prediction (IRI) of Columbia University which provides monthly updates of 7-month future climate forecasts from five different general circulation models (GCMs) of the global atmosphere. These GCMs come from the University of Maryland (COLA), the University of Hamburg (ECHAM4.5), the National Weather Service’s Climate Prediction Center (NCEP), NASA’s Goddard Institute of Space Studies (NSIPP), and the Scripps Oceanographic Institute (ECPC).           

The fire forecasts begin by simulating ecosystem and fire dynamics over the United States from 1895 at a monthly timestep up through the most recent month and then shift into the future forecasts from each of the GCMs.  The MC1 model then simulates ecosystem dynamics and fire, including fire occurrence, fire area, and fire effects, through the 7-month forecast period.  The forecasts are combined into a consensus “fire risk” map showing the areas where all or only a few of the forecasts agree on where fires might occur and their general size. The fire risk maps and ancillary drought index maps are posted monthly on the MAPSS team’s web site.

After three seasons of fire forecasting, comparisons to observed fire data have shown MC1 fire forecasts to be reasonably accurate in terms of predicting both total area burned across the United States and the location of fire events on the 50-kilometer square resolution modeling grid.  However, discrepancies between observations and hindcasts have indicated the need for model improvements, especially in regard to the assumed availability of ignition sources and the prediction of fire occurrence in fine-fuel dominated vegetation. Analysis of the monthly weather forecasts and subsequently observed weather has also revealed contraints on the accuracy of the fire forecasts imposed by the skill of the climate models.

The MAPSS team’s fire risk forecasting activities are supported by funding from the National Fire Plan, and serve all four elements of the National Fire Plan goals. At the beginning of each fire season, the MC1 fire forecasts are presented to fire managers from all nine western Geographic Area Coordination Centers (GACCs) attending the Western National Seasonal Assessment Workshop (NSAW) sponsored by the Predictive Services Group of the National Interagency Coordination Center (NICC), and are routinely incorporated into NICC’s seasonal weather/climate/fuels outlooks for the western GACCs. And currently over 160 land managers from various resource agencies are alerted each month to new fire forecasts posted on the MAPSS web site via an ever-growing email list.

Fire Session - Thursday Afternoon

corresponding author:

James Lenihan
USDA Forest Service
Pacific Northwest Research Station
3200 SW Jefferson Way
Corvallis, OR 97331
541-750-7432
jlenihan@fs.fed.us