Rocky Mountain Research Station RMRS-4853
Topic: Use of the Stratified Cost Index in WFDSS
Background: In FY 2006, the FS adopted a performance
measure designated the “Stratified Cost Index” (SCI).
In FY 2007, the SCI was also incorporated into the new wildland
fire decision support system (WFDSS). The SCI is simply a set of
regression equations developed to estimate suppression expenditures
on individual large wildland fires (fires greater than 300 acres).
That is, characteristics of the fire, such as fire size, the fire
environment (slope, aspect, energy release component, fuel model,
etc), housing values within proximity to the fire, and geographic
area are used in a statistical model to come up with an estimate
of the cost of the fire. Currently, there are three regression equations
incorporated into WFDSS -- two Forest Service models (one for the
western U.S. and one for the eastern U.S.) and one DOI model. These
cost models are built using historic information on suppression
expenditures and fire characteristics, using data for fires occurring
in FYs 2004-2007 from all geographic areas. The independent variables
(the fire characteristics used to predict cost) vary somewhat by
model (FS west, FS east, DOI). The models also include a regional
variable that accounts for differing costs across geographic areas
(defined as either FS regions or Geographic Area Coordination Centers).
The dependent variable (what we are trying to predict) in each of
the models is suppression expenditures per acre, which includes
expenditures by all federal agencies (FS and DOI) on the fire in
question. It does not include state expenditures unless those expenditures
are already captured in the federal accounting system. The FS regression
models were developed using data on fires where the FS was the protection
agency of record. Similarly, the DOI model was built using data
on fires where one of the DOI agencies was the protection agency
Use of SCI in WFDSS:
The SCI for a particular fire is calculated in the new WFDSS simply
by entering the latitude and longitude of the fire start and estimated
acreage burned for the alternative fire scenarios from the WFSA.
From this, all of the independent variables (fire characteristics)
in the model are calculated and entered into one of the three regression
equations, depending upon which agency is the designated protection
agency. The WFDSS system then calculates the SCI and provides a
range of values with which to compare actual suppression expenditures
(see example below).
The goal of the fire manager should be to keep
costs at or below the 50 percent level in order to cut agency suppression
expenditures. In this case, fires with similar characteristics,
in terms of size, fuel models, housing values, etc., on average
cost about $1,591 per acre. Keeping the actual fire costs at or
below this level will contribute towards cost containment goals.
If actual costs reach the yellow and red levels, this does not imply
that the fire is being poorly managed. However, the probability
that the fire will be reviewed after the fire season increases.
In these instances, the use of FSPro and RAVAR may be helpful in
determining if the fire poses higher risk than fires typical to
these conditions, or if strategies should be revised to address
identified cost concerns.
Contact: Krista Gebert, RMRS
Suppression Expenditures for Individual
Large Wildland Fires (WJAF, 2007)
Fuel Treatment Costs on Forest
Service Lands in the Western United States (WJAF, 2006)
of SCI in WFDSS
SCI for Fire Complexes within the WFDSS System
of SCI in WFDSS for Multi-Agency Incidents