**3. Applications of burn probability modeling & exposure analysis**

The design and functionality of simulation-based approaches span a range of intended applications, from modeling a specific fire event given an ignition to projecting wildfire likelihood and intensity at landscape scales across multiple fire seasons. Advances in burn probability modeling have enabled increasing sophistication and analytical rigor across a variety of wildfire management applications. Researchers and practitioners are able to, for instance, project near-term fire behavior using real-time weather information (Andrews et al., 2007) or to project wildfire behavioral changes in response to fuel treatments (Kim et al., 2009). In this section we focus on application of burn probability modeling and exposure analysis to support management of wildfire incidents and to support proactive hazardous fuels reduction treatments.

### **3.1 Incident management**

104 Novel Approaches and Their Applications in Risk Assessment

This framework quantifies risk in terms of relative net value change (NVC), or the percentage change in initial value resulting from interaction with fire. That is, response functions address relative rather than absolute change in resource or asset value. Response functions translate fire effects into NVC to the described HVRA. In response functions illustrated in Figure 3 NVC is based on fire intensity, which is a robust fire characteristic that integrates fuel consumption and spread rate, and is often used to estimate fire effects (Thompson et al. 2011a; Ager et al. 2007). In Figure 3 the response function varies according to categorical fire intensity levels, although the framework is perfectly amenable to definition of multivariate response functions incorporating additional geospatial

Fig. 3. Conceptual model for calculating wildfire risk (Modified from Calkin et al., 2010).

Characterizing fire effects has presented a major challenge to risk assessment, due to limited understanding of the spatiotemporal dynamics of ecological changes wrought by wildfire. Many past analyses focused on wildfire risk to commercial values, such as commercial timber (Konoshima et al. 2008), with a much more limited set focusing on broader nonmarket resource values and public infrastructure (Venn & Calkin 2011). There exist a variety of models that can estimate first-order fire effects such as tree mortality, soil heating, fuel consumption, and smoke production, although managers are generally more concerned with second- and third-order effects such as air quality, water quality, and habitat degradation (Reinhardt & Dickinson 2010). The management context, availability of appropriate models, and quality of spatial data will inform selection of the appropriate fire effects modeling approach (Reinhardt et al. 2001). In the absence of fire effects models, a

information that influence response to fire (see Section 4).

Development of suppression strategies for escaped wildland fires is subject to considerable uncertainty and complexity. Factors to balance include likely weather and fire behavior, topography, firefighter safety, and the availability and productivity of firefighting resources (ground crews, fire engines, air tankers, etc.). Of particular importance is the ability to project where and under what conditions fire is likely to interact with HVRAs. This information can help fire managers decide where aggressive fire suppression may be effectives to protect HVRAs, and where fires may have a positive impact in fire-prone ecosystems.

In the United States, all wildfires occurring on federal lands are cataloged within the Wildland Fire Decision Support System (WFDSS). WFDSS provides decision documentation and analysis functionality to describe the fire incident, create objectives and requirements, develop a course of action, validate key dependencies and evaluate risks (Noonan-Wright et al., 2011). The system combines a suite of fire behavior predictions with identification and quantification of values at risk to inform incident management considering safety, complexity, economics, and risk (Calkin et al., 2011).

The two primary risk-based analytical components within WFDSS are the Fire Spread Probability model (FSPro) and the Rapid Assessment of Values at Risk (RAVAR). FSPro calculates the probability of fire spread from a current fire perimeter or ignition point, for a specified time period. Burn probability maps are derived from simulating fire growth for thousands of statistically generated weather scenarios (Finney et al., 2011b). As implemented in WFDSS burn probabilities are mapped as probability zones, or contours, of similar burn probability; exterior contours have lower probability of fire occurrence than interior contours.

Figure 4 displays an FSPro analysis for the SQF Canyon Fire, a human-caused fire that ignited on September 20, 2010 in California in the Sequoia National Forest. The figure provides a 7-day projection of fire growth as of September 14, 2010. The fire spread probability contours, moving outward from the red center, correspond to intervals of >80%,

The Science and Opportunity of Wildfire Risk Assessment 107

inform managers regarding the likelihood of fire impacting HVRAS and assist in developing target fire containment perimeters. WFDSS supports risk-informed decision making by analyzing HVRA exposure to fire, allowing local managers to evaluate the likely impacts

 **Acres Threatened Kern County** 

**Count by Zone** 

Table 1. Estimates of Structure Values at Risk, as output by WFDSS-RAVAR, using data

Fuel management seeks to alter the quantity, spatial arrangement, structure, and continuity of fuels so as to induce desirable changes in fire behavior. Broadly speaking, fuel management activities are designed to reduce the risk of catastrophic fire, protect human communities, reduce the extent and cost of wildfires, and restore fire-adapted ecosystems. For a fuel treatment to function effectively it must first spatially interact with an actual wildfire, and second mitigate fire behavior according to design objectives (Syphard et al.,

Recognized principles for fuels management planning (Agee & Skinner, 2005) largely relate to individual treatments and their effects on localized fire behavior. Less understood is how in aggregate fuel treatments can affect landscape-scale processes of fire spread (Hudak et al., 2011). Prospective evaluation the influence of fuel treatments requires the estimation of altered fire behavior both within and outside of treated areas (Finney et al., 2007). Spatial fire growth models and burn probability modeling have emerged as useful tools for analyzing the influence of fuel treatments on topological fire spread, and to enable risk-

A workflow for fuel treatment planning includes identifying the purpose and need for treatments, simulating wildfire behavior across the current, untreated landscape to

> 80 % 47,894 47,894 290 290 \$43,399,080 \$43,399,080 60 – 80 % 12,029 59,923 215 505 \$32,175,180 \$75,574,260 40 – 60 % 14,062 73,985 289 794 \$43,249,428 \$118,823,688 20 – 40 % 15,602 89,586 208 1,002 \$31,127,616 \$149,951,304 5 – 20 % 24,995 114,582 297 1,299 \$44,446,644 \$194,397,948 1 – 5 % 53,989 168,571 794 2,093 \$118,823,688 \$313,221,636

**Cumulative Zone** 

67,980 679 \$101,665,338

**Value by Zone** 

**Cumulative Value** 

**Cumulative Acres** 

and prioritize suppression efforts accordingly.

**Acres by Zone** 

**Fire Spread Probability Zone** 

Expected Value (without suppression)

2011).

from Kern County, California.

**3.2 Hazardous fuels management** 

based analysis of fuel treatment effectiveness.

60-80%, 40-60%, 20-40%, 5-20%, 1-5%, and <1% of likely fire spread given the current fire location and perimeter.

The RAVAR analytic model produces two distinct map products and associated reports, inventorying mapped Critical Infrastructure (CI) and Natural and Cultural Resources (NCR). HVRAs identified in CI reports include private structures, recreation facilities, water supply systems, major power lines, pipelines, communication towers, and hazardous waste sites. NCR products focus on regionally identified natural resources and wildland management priorities, such as sensitive wildlife habitat and restoration priority areas. Table 1 provides example tabular RAVAR output quantifying the number and value of structures at risk according to FSPro Fire Spread Zones, using county tax records.

Fig. 4. FSPro run for the Canyon Fire in the Sequoia National Forest, California.

Figure 5 displays a close-up view of an FSPro-RAVAR analysis for the SQF Canyon Fire, which overlays geospatial identification of Critical Infrastructure on top of probability contours. (RAVAR maps are approximately 4' x 3' and are intended for poster display, generally making detailed displays on computer screens difficult.) The current fire perimeter is outlined in red, overlaid on top of associated probability contours of likely spread (see Figure 4). Threatened resources include private structures (black triangles), federal structures (green triangle), power transmission lines (inverted "T", dashed connector), and mine sites (pick and shovel). The green line demarcates the National Forest boundary, and yellow/red dots identify "hot" points from satellite images.

Together FSPro and RAVAR provide state-of-the-art exposure analysis, linking near real time probabilistic fire spread predictions with values at risk These analytical products

60-80%, 40-60%, 20-40%, 5-20%, 1-5%, and <1% of likely fire spread given the current fire

The RAVAR analytic model produces two distinct map products and associated reports, inventorying mapped Critical Infrastructure (CI) and Natural and Cultural Resources (NCR). HVRAs identified in CI reports include private structures, recreation facilities, water supply systems, major power lines, pipelines, communication towers, and hazardous waste sites. NCR products focus on regionally identified natural resources and wildland management priorities, such as sensitive wildlife habitat and restoration priority areas. Table 1 provides example tabular RAVAR output quantifying the number and value of

structures at risk according to FSPro Fire Spread Zones, using county tax records.

Fig. 4. FSPro run for the Canyon Fire in the Sequoia National Forest, California.

boundary, and yellow/red dots identify "hot" points from satellite images.

Figure 5 displays a close-up view of an FSPro-RAVAR analysis for the SQF Canyon Fire, which overlays geospatial identification of Critical Infrastructure on top of probability contours. (RAVAR maps are approximately 4' x 3' and are intended for poster display, generally making detailed displays on computer screens difficult.) The current fire perimeter is outlined in red, overlaid on top of associated probability contours of likely spread (see Figure 4). Threatened resources include private structures (black triangles), federal structures (green triangle), power transmission lines (inverted "T", dashed connector), and mine sites (pick and shovel). The green line demarcates the National Forest

Together FSPro and RAVAR provide state-of-the-art exposure analysis, linking near real time probabilistic fire spread predictions with values at risk These analytical products

location and perimeter.

inform managers regarding the likelihood of fire impacting HVRAS and assist in developing target fire containment perimeters. WFDSS supports risk-informed decision making by analyzing HVRA exposure to fire, allowing local managers to evaluate the likely impacts and prioritize suppression efforts accordingly.


Table 1. Estimates of Structure Values at Risk, as output by WFDSS-RAVAR, using data from Kern County, California.

#### **3.2 Hazardous fuels management**

Fuel management seeks to alter the quantity, spatial arrangement, structure, and continuity of fuels so as to induce desirable changes in fire behavior. Broadly speaking, fuel management activities are designed to reduce the risk of catastrophic fire, protect human communities, reduce the extent and cost of wildfires, and restore fire-adapted ecosystems. For a fuel treatment to function effectively it must first spatially interact with an actual wildfire, and second mitigate fire behavior according to design objectives (Syphard et al., 2011).

Recognized principles for fuels management planning (Agee & Skinner, 2005) largely relate to individual treatments and their effects on localized fire behavior. Less understood is how in aggregate fuel treatments can affect landscape-scale processes of fire spread (Hudak et al., 2011). Prospective evaluation the influence of fuel treatments requires the estimation of altered fire behavior both within and outside of treated areas (Finney et al., 2007). Spatial fire growth models and burn probability modeling have emerged as useful tools for analyzing the influence of fuel treatments on topological fire spread, and to enable riskbased analysis of fuel treatment effectiveness.

A workflow for fuel treatment planning includes identifying the purpose and need for treatments, simulating wildfire behavior across the current, untreated landscape to

The Science and Opportunity of Wildfire Risk Assessment 109

implemented treatments, as in Figure 6, here the analysis simulates hypothetical fuel conditions had treatments not been implemented. The actual fire perimeter is outlined in red, and probability zones reflect contours of likely fire spread as output from wildfire simulations, had the treatments not been implemented. Areas of positive probability (yellow, orange, red) reflect that the treatments were effective in preventing spread. Exposure analysis intersects probability zones with mapped HVRAs including US Forest Service structures, improved structures (identified from county tax records), and bull trout (*Salvelinus confluentus*) critical habitat. Quantification of reduced exposure can inform

Fig. 6. Illustration of reductions in burn probability as a function of percent of the landscape treated (Ager et al., 2007). "TRT-X" refers to different modeled scenarios in which X percent

estimates of fuel treatment effectiveness.

of the landscape is treated.

characterize hazard and risk, developing treatment strategies, and iteratively simulating and evaluating changes to wildfire hazard and risk stemming from the treatment. Primary variables comprising a treatment strategy include the size of individual treatment units, the placement/pattern of the treatments, the proportion of the landscape treated, and treatment longevity (Collins et al., 2010).

Fig. 5. Detail of RAVAR analysis for the Canyon Fire in Sequoia National Forest, California.

Ager et al. (2011) reviewed the development and use of ArcFuels, an integrated system of tools to design and test fuel treatment programs within a risk assessment framework. A number of fuel treatment case studies have employed the same basic analytical approach of comparative burn probability and intensity modeling across untreated/treated landscapes (Ager et al., 2010; Parisien et al., 2007). Figure 6 illustrates such a case study that investigated the influence of different treatment strategies on burn probability. Four scenarios, representing treating 0%, 10%, 20%, and 50% of the landscape were fed into wildfire simulation models to estimate impacts to burn and intensity probabilities.

In addition to evaluating prospective fuel treatments and informing treatment design, burn probability modeling can also be used to evaluate the effectiveness of previously implemented treatments. Field-based evaluations of fuel treatments have relied on the relatively rare occurrence of wildfires interacting with treatments. Of these treatments that have engaged wildfire, few have been subject to rigorous review to characterize treatment effectiveness (Hudak et al., 2011). Only recently has it been possible to estimate the spatial probabilities of landscape burning as a function of extant fuels treatments for real wildland fire-affected landscapes (Cochrane et al., in press). Figure 7 displays an example of burn probability modeling to analyze the impact of implemented treatments and their engagement with the School Fire. Rather than simulating the impacts of hypothetically

characterize hazard and risk, developing treatment strategies, and iteratively simulating and evaluating changes to wildfire hazard and risk stemming from the treatment. Primary variables comprising a treatment strategy include the size of individual treatment units, the placement/pattern of the treatments, the proportion of the landscape treated, and treatment

Fig. 5. Detail of RAVAR analysis for the Canyon Fire in Sequoia National Forest, California.

Ager et al. (2011) reviewed the development and use of ArcFuels, an integrated system of tools to design and test fuel treatment programs within a risk assessment framework. A number of fuel treatment case studies have employed the same basic analytical approach of comparative burn probability and intensity modeling across untreated/treated landscapes (Ager et al., 2010; Parisien et al., 2007). Figure 6 illustrates such a case study that investigated the influence of different treatment strategies on burn probability. Four scenarios, representing treating 0%, 10%, 20%, and 50% of the landscape were fed into wildfire

In addition to evaluating prospective fuel treatments and informing treatment design, burn probability modeling can also be used to evaluate the effectiveness of previously implemented treatments. Field-based evaluations of fuel treatments have relied on the relatively rare occurrence of wildfires interacting with treatments. Of these treatments that have engaged wildfire, few have been subject to rigorous review to characterize treatment effectiveness (Hudak et al., 2011). Only recently has it been possible to estimate the spatial probabilities of landscape burning as a function of extant fuels treatments for real wildland fire-affected landscapes (Cochrane et al., in press). Figure 7 displays an example of burn probability modeling to analyze the impact of implemented treatments and their engagement with the School Fire. Rather than simulating the impacts of hypothetically

simulation models to estimate impacts to burn and intensity probabilities.

longevity (Collins et al., 2010).

implemented treatments, as in Figure 6, here the analysis simulates hypothetical fuel conditions had treatments not been implemented. The actual fire perimeter is outlined in red, and probability zones reflect contours of likely fire spread as output from wildfire simulations, had the treatments not been implemented. Areas of positive probability (yellow, orange, red) reflect that the treatments were effective in preventing spread. Exposure analysis intersects probability zones with mapped HVRAs including US Forest Service structures, improved structures (identified from county tax records), and bull trout (*Salvelinus confluentus*) critical habitat. Quantification of reduced exposure can inform estimates of fuel treatment effectiveness.

Fig. 6. Illustration of reductions in burn probability as a function of percent of the landscape treated (Ager et al., 2007). "TRT-X" refers to different modeled scenarios in which X percent of the landscape is treated.

The Science and Opportunity of Wildfire Risk Assessment 111

Structured Elicitation Process 1. Articulate research question

5. Design elicitation protocol 6. Prepare elicitation protocol 7. Elicit expert judgment

judgment. Modified from (Knol et al., 2010; Kuhnert et al., 2010)

4. Select experts

8. Feedback

expert consensus is reached.

2. Identify and characterize uncertainties 3. Resolve scope and format of elicitation

Fig. 8. Eight major steps in organizing and implementing a structured elicitation of expert

Figure 8 presents a structured process for eliciting expert judgment. In the first step, a clear articulation of the research question(s) will inform the design and implementation of the study as well as the larger structure of the modeling process. This is followed by identification and characterization of uncertainties, which will influence choices regarding the type of experts and elicitation format. Resolving the scope and format of the elicitation entails identifying the number of experts to engage and the nature of the engagement (interview, group workshop, survey distribution, etc.), while considering available resources and other constraints. Selection of experts includes choices between generalists, subject matter experts, and normative experts (those with experience to support elicitation itself). Design of the protocol considers the type of information to be elicited, the most appropriate metric(s), the most appropriate elicitation mechanism, and how to clearly communicate information needs to avoid issues of linguistic uncertainty (Regan et al., 2002). Preparation of the elicitation protocol includes providing selected experts with sufficient information on the nature of the research problem and associated uncertainties, the scope and purpose of the elicitation, and the nature of the elicitation procedure itself. Lastly, the elicitation procedure is implemented, with opportunities for post-elicitation feedback and revision.

In terms of the wildfire management context, the research question generally involves assessing wildfire risk to HVRAs, potentially in a comparative risk framework to evaluate the effectiveness of alternative management actions (Figure 1). In the second step, wildfire management is subject to myriad sources of uncertainty, not all of which are necessarily best handled with expert judgment (burn probability modeling to capture environmental stochasticity, e.g.). Thompson & Calkin (2011) present a typology of uncertainties faced in wildfire management, and identify that with regard to knowledge uncertainty surrounding fire effects, expert systems are perhaps the most appropriate approach. In our past experience we have adopted group workshops, and assembled resource scientists (hydrologists, soil scientists, wildfire biologists, fire ecologists, etc.) as appropriate given the HVRAs being assessed (Thompson et al., 2011b). The elicitation protocol identifies response functions that quantitatively characterize resource-specific response functions as a function of fire intensity, and response functions are iteratively explored, justified, and updated until

Figure 9 illustrates expert-based response functions for two HVRAs with contrasting response to fire, mapped across six fire intensity level (FIL) classes. These response functions

Fig. 7. Analysis for the School Fire demonstrating the impact of implemented fuel treatments. Modified from (Cochrane et al., in press).
