Decision support for evaluation of fire danger
Paul F. Hessburg1, Keith M. Reynolds2, Robert E. Keane3, R. Brion Salter1, & Kevin M. James1
1
U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, Wenatchee,
WA 98801, USA. E-mail: phessburg@fs.fed.us.
2
U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, Corvallis, OR
97331, USA.
3
U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Missoula, MT
59808, USA.
_______________________________________________________________________
Introduction
Wildland fuels have been accumulating in western U.S. forests for at least the past 70 years
due mostly to 20th century settlement and management activities (Hessburg and Agee 2003).
As demonstrated by recent wildland fires, additional fuels are contributing to more intense
fire behavior and increasing fire resistance to control or containment. Consequently, property
and natural resources have been destroyed, costs of fire management have escalated, firedependent ecosystems have deteriorated, and risks to human life continue to escalate.
In this paper, we describe a decision-support application for interpreting data and synthesizing
information to evaluate existing forest fuel conditions. Danger from wildland fire is evaluated as
a function of fire vulnerability, potential wildfire severity, and risk of ignition. We demonstrate
use of the application with an example from the Rocky Mountain region in the State of Utah
(Map Zone 16), which represents a planning area of about 5 million ha and encompasses 604
watersheds. We also discuss considerations for extending the application to support strategic
planning at national and regional scales, and tactical planning at local scales.
Methods
We selected a single map zone as a proof of concept for our approach (Fig. 1). Map zone 16
falls almost entirely within the state of Utah. Within this study area, we evaluated fire danger
for all 575 complete subwatersheds in map zone 16. The average size of subwatersheds
was about 8,274 ha, ranging from a minimum of about 5,000 to a maximum of about 10,000
ha. For reference, a subwatershed represents the 6th level in the established US Geological
Survey watershed hierarchy.
Most spatial data layers for this study
were obtained from the LandFire
(www.Landfire.gov) project, and were
provided as 30-m-resolution raster
grids. The LandFire layers included
both base layers as well as derived data
layers produced by the FireHarm model
(ref).
The logic specification for evaluation of
fire danger was graphically designed
with the logic-modeling system,
NetWeaver Developer (Rules of Thumb,
Inc., North East, PA)1. Each topic in a
Figure 1. Map zones of the United States from the Earth
NetWeaver model represents an object
Surface Resources and Science (EROS) Data Center.
for which a proposition is to be
evaluated. For example, the fire danger topic, representing the top level in the model,
evaluates the proposition that danger of severe wildfire is low. The complete evaluation of
fire danger depends on the three topics, fire vulnerability, fire severity, and ignition risk, each
of which incrementally contribute to the evaluation of fire danger, as indicated by the union
operator (Fig. 2). Because the union operator specifies that premises incrementally
contribute to the proposition of their parent topic, low strength of evidence for one topic can
be compensated by strong evidence from others. Notice that, if the fire danger topic is
thought of as testing a conclusion, then the three topics on which it depends can be thought
of as its premises. Similarly, each of the three topics under fire danger has its own logic
specification that includes a set of subtopics or premises (Hessburg et al., in prep.).
Figure 2. The logic specification for fire danger.
Each topic evaluates the proposition of a low
condition (e.g., fire danger is low).
Evaluation of fire danger for all complete
subwatersheds in map zone 16 was
performed with the Ecosystem Management
Decision Support (EMDS) system (Reynolds
et al. 2003b). The NetWeaver logic engine
(Miller and Saunders 2002), a component of
the EMDS system, was used to perform
evaluations of fire danger and all dependent
topics. Each map product generated by the
analysis reports the strength of evidence for
the proposition associated with the map topic.
Results
There were pronounced differences in fire danger between subwatersheds in the northern
and southern portions of the study area (Fig. 3). Strength of evidence for low fire danger was
generally moderate in the north, but low in the south. Strength of evidence for low
vulnerability to wildland fire exhibited a very similar spatial pattern. The same general pattern
also is apparent with respect to strength of evidence for low ignition risk, but there also is a
higher proportion of subwatersheds in the north with strong evidence, and a higher
proportion of subwatersheds in the south with very low to no evidence.
Figure 3. Strength of evidence for low fire danger and its three premises in subwatersheds of map zone 16.
In general, high ignition risk was driven by the tendency for more severe summer drought in
the southern portion of map zone 16. Crown fire potential, fireline intensity, and spread rate
under a wildfire burn scenario were most influential under the fire severity topic, and fire
severity in general was most influential to the full evaluation of fire danger. Crown bulk
density was most influential to the evaluation of the fire vulnerability topic.
Discussion
The present study illustrates application of EMDS for evaluation of forest fuels at the spatial
extent of a USGS map zone. However, when data from the LandFire program eventually
provide national coverage for the U.S., there is no reason why an analysis of all
subwatersheds in the U.S. could not be done in the same fashion as we have illustrated
here. Moreover, it is a relatively simple matter, given such a base analysis, to summarize
such watershed-scale evaluations to various intermediate broader scales such as States or
geographic regions as a basic input for broad-scale planning.
At the other extreme, the present study also provides a starting point for finer-scale planning.
The present study has only examined the evidence for fire danger in subwatersheds of map
zone 16, but this information, by itself, is not necessarily sufficient for fuel-treatment planning.
Subwatersheds in the worst condition with respect to fuel conditions may not be the best
candidates for treatment. In particular, additional strategic or logistical factors such as proximity
to population centres, presence of endangered species, slope steepness, and road access all
might be taken into account in selection of specific watersheds within a management area for
fuel treatment. Such an approach was illustrated by Reynolds et al. (2004) using the EMDS
planning component, which uses a decision engine, for such purposes. In that study they
considered the compositional and structural integrity of forests along with contemporary fire
risks and the technical and economic feasibility of restoration. Carefully designed decision
models can not only assist with a more circumspect approach to selection of individual
treatment units, but can also help elucidate which of several treatment options may be most
suitable in a given unit, thus also providing support for the tactical level of planning.
Finally, given the widespread increase in danger of wildland fire throughout the western U.S.
over the past 70 years or more, the sustainability of these forest ecosystems is clearly at stake,
and we hope the role of decision-support systems such as EMDS in assisting with restoration to
improve or maintain sustainability is relatively obvious from this study. What is perhaps not quite
so obvious, however, is the interdependency between science, policy, and decision-support
systems such as EMDS. Although logic models, for example, are sometimes used for prediction,
they are fundamentally concerned with interpretation (Reynolds et al. 2003a). In other words,
what does the information mean? Meaning can be highly normative or highly subjective, and
usually falls somewhere in between the two extremes. As a result, virtually all interpretation
embeds some degree of subjectivity, so values and policy are inextricable aspects of a logic
model, at least to some degree. The practical implication is that successful application of most
decision-support systems to real-world situations ultimately depends on a close collaboration
between the scientific community that brings its facts to the table, and the policy makers that
need to reach decisions based on that information.
Reference
Hessburg, P. F. and Agee, J. K. 2003. An environmental narrative of Inland Northwest US
forests, 1800-2000. In: Fire and Aquatic Ecosystems. Special Feature: Forest
Ecology and Management 178:23-59.
Hessburg, P.F., Reynolds, K.M., Salter, R.B., James, K.M. and Keane, R.E. (in preparation)
Integrated Decision Support for Fire Danger Evaluation and Fuels Treatment
Planning. Forest Ecology and Management.
Hessburg, P.F., Salter, R.B., Richmond, M.B., and Smith B.G. (2000) Ecological subregions
of the Interior Columbia Basin, USA. Applied Vegetation Science 3: 163-180.
Miller, B.J., and Saunders, M.C. (2002) The NetWeaver reference manual. Pennsylvania
State University, College Park, PA.
Reynolds, K.M.; Hessburg, P.F. 2005. Decision support for integrated landscape evaluation
and restoration planning. Forest Ecology and Management 207: 263-278.
Reynolds, K.M., Johnson, K.N., and Gordon, S.N. (2003a) The science/policy interface in
logic-based evaluation of forest ecosystem sustainability. Forest Policy and
Economics 5: 433-446.
Reynolds, K.M., Rodriguez, S., and Bevans, K. (2003b). User guide for the Ecosystem
Management Decision Support System, version 3.0. Environmental Systems
Research Institute, Redlands, CA.