Photo by Adam Carter
Mapping Wildfire Risk to Structures and Other Human Developments
Section 7 of SB762 directs Oregon State University (OSU) to work with the Oregon Department of Forestry (ODF) to complete a statewide wildfire risk map for structures and other human developments (“statewide risk map”). The statewide risk map is publicly available on the Oregon Wildfire Risk Explorer, which serves as an interactive educational tool for Oregonians.
This webpage describes how OSU scientists assessed wildfire risk to create the map. The following sections explain the components of wildfire risk, describe how risk was calculated, and what it means for properties to be classified into risk categories.
Simplified flowchart of how the three data products developed by OSU will be used together to develop and enforce defensible space rules established under Section 8 of SB762. Here we emphasize how the risk to structures map can be used, in conjunction with the WUI map, to determine risk class.
The statewide risk map identifies where wildfires pose the most risk to structures and other human developments in order to inform the actions of state agencies in their efforts to reduce wildfire risk and mitigate future damage to Oregon communities. Scientists at OSU classified risk for all tax lots in Oregon into one of five classes: no, low, moderate, high, or extreme wildfire risk. Relevant state agencies will refer to the statewide risk map in conjunction with the statewide wildland-urban interface and social vulnerability maps to implement their responsibilities under SB762.
Components of Risk
Wildfire risk is a function of the three components of the wildfire risk triangle: burn probability, fire intensity, and the susceptibility of structures and other human developments.
Burn Probability - Modeled annual burn probability. Non-burnable areas include open water, barren ground, urban areas and some types of agricultural land.
Burn Probability
Burn probability is the average annual likelihood that a specific location will experience wildfire. Burn probabilities represent long-term averages and are not forecasts or predictions of where fire is going to occur in a specific year. Annual burn probabilities are primarily a reflection of regional climate patterns and vegetation types, but can be affected by land use, ignition patterns and other elements that are within human control.
For the statewide risk map, OSU scientists used FSim to estimate annual burn probabilities. The simulations used local weather records, up to date landscape conditions, and historical patterns of fire occurrence to simulate where fires ignite, how they grow, and how often the landscape burns. Average burn probabilities were generated from simulations of more than 10,000 annual scenarios to account for the wide variability in factors that influence fire occurrence. [4].
Mean Conditional Flame Length - Modeled fire intensity. Non-burnable areas include open water, barren ground, urban areas and some types of agricultural land.
Wildfire Intensity
Wildfire intensity is a measurement of the amount of energy produced by a fire, frequently reported as “flame length.” Fire intensity is driven by a number of factors including weather, topography, and fuel type. Fire intensity is an important component of risk because varying intensities can lead to different impacts to buildings. For instance, fires with flame lengths less than two feet are less likely to damage buildings because they can usually be controlled with hand tools and machinery, and are less likely to cast large ember showers. In contrast, fires with flame lengths greater than eight feet are much more likely to damage and destroy structures and other human developments because they can only be engaged with aerial resources when weather conditions allow, and are far more likely to cast far-reaching embers that spark new fires. As with burn probability, fire intensity was estimated using many simulation scenarios based on observed weather and our best understanding of up-to-date fuel conditions [7].
Susceptibility
The estimated damage to a structure is directly related to the expected intensity of a wildfire and the kind of vegetation in which it’s burning. We call the expected damage “susceptibility,” and it is measured using response functions. For instance, if a fire is burning in forested vegetation and the flame length at the location of a structure is five feet, the response function is 50, meaning that the building is expected to suffer a 50% loss in value.
How is Risk Quantified?
To quantify risk, we combine all three components - burn probability, fire intensity and susceptibility - together. To calculate wildfire risk, we multiplied burn probability times the appropriate response function based on modeled fire intensities. The result was a wildfire risk value for every tax lot in Oregon. Five risk categories were then defined based on the statewide range of eNVC values:
Why Do Risk Categories Matter?
The risk category associated with each tax lot will inform how multiple state agencies execute their responsibilities related to SB762. For instance, tax lots classified as ‘high’ or ‘extreme’ risk may be subject to defensible space regulations as defined by the State Fire Marshal. Property owners can view their risk category on the Oregon Wildfire Risk Explorer. If property owners believe that their risk has been mischaracterized, they can appeal the decision.
Property level risk information will also be used to inform updates to comprehensive and local zoning codes and building code standards.
The statewide structure risk map quantifies risk on all tax lots in Oregon, regardless of whether a structure or other human development is currently present.
References
- Day, M., 2020. Fireshed Registry.
- Department of Interior, Geological Survey, and U.S. Department of Agriculture., 2016. 40 Scott and Burgan Fire Behavior Fuel Models Layer https://landfire.gov/fbfm40.php
- Dunn, C.J., O’Connor, C.D., Abrams, J., Thompson, M.P., Calkin, D.E., Johnston, J.D., Stratton, R., Gilbertson-Day, J., 2020. Wildfire risk science facilitates adaptation of fire-prone social ecological systems to the new fire reality. Environ. Res. Lett. 15, 025001. https://doi.org/10.1088/1748-9326/ab6498
- Finney, M.A., McHugh, C.W., Grenfell, I.C., Riley, K.L., Short, K.C., 2011. A simulation of probabilistic wildfire risk components for the continental United States. Stoch. Environ. Res. Risk Assess. 25, 973–1000. https://doi.org/10.1007/s00477-011-0462-z
- Gilbertson-Day, J., Stratton, R.D., Scott, J.H., Vogler, K.C., Brough, A., 2018. Pacific Northwest Quantitative Wildfire Risk Assessment: Methods and Results.
- Oregon State University, 2019. Oregon Wildfire Risk Explorer. https://oregonexplorer.info/topics/wildfire-risk?ptopic=62.
- Scott, J.H., 2020. A deterministic method for generating flame-length probabilities., in: In: Hood, Sharon M.; Drury, Stacy; Steelman, Toddi; Steffens, Ron, [Eds.]. Proceedings of the Fire Continuum Preparing for the Future of Wildland Fire. May 21-24, 2018. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO, pp. 195–205.
- Scott, J.H., Thompson, M.P., Calkin, D.E., 2013. A wildfire risk assessment framework for land and resource management (No. RMRS-GTR-315). U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Ft. Collins, CO. https://doi.org/10.2737/RMRS-GTR-315
- Short, K.C., 2021. Spatial wildfire occurrence data for the United States, 1992-2018 [FPA_FOD_20210617] (5th Edition). https://doi.org/10.2737/RDS-2013-0009.5
- Thompson, M., Bowden, P., Brough, A., Scott, J., Gilbertson-Day, J., Taylor, A., Anderson, J., Haas, J., 2016. Application of Wildfire Risk Assessment Results to Wildfire Response Planning in the Southern Sierra Nevada, California, USA. Forests 7, 64. https://doi.org/10.3390/f7030064
- U.S. Department of Agriculture, Forest Service, 2022. Confronting the Wildfire Crisis: A Strategy for Protecting Communities and Improving Resilience in America’s Forests (No. FS-1187a).
- USDA Forest Service, 2021. Wildfire Risk to Communities. Wildfire Risk Communities. URL https://wildfirerisk.org/
Frequently Asked Questions
Why is OSU making the wildfire risk map?
Section 7 of Senate Bill 762 directed OSU to coordinate with the Oregon Department of Forestry to develop a statewide risk map for all structures and other human developments.
OSU is a trusted, non-biased source of wildfire risk information. Since 2006, the Oregon Wildfire Risk Explorer, a program of the Institute of Natural resources at OSU, has been a public source of geospatial data and wildfire risk information used in state, regional and local risk management applications. Scientists at OSU have demonstrated leadership in the development and application of wildfire risk science.
What data was used in the fire behavior models?
The fire behavior models operate on three general types of inputs.
Fuelscape - the modeling landscape relies on eight data layers from LANDFIRE that describe topography, canopy fuel characteristics, and surface fuel characteristics. LANDFIRE 2.0.0 served as the foundation, but was updated to reflect 2022 conditions based on wildfires and significant disturbances that have occurred since LANDFIRE 2.0.0 was released. In addition, the LANDFIRE data was modified with recommendations from local and regional fuels specialists during a 3-day fuels calibration workshop.
Historical fire occurrence - spatial wildfire ignition records from 1992 - 2020 were used to inform the timing and location of simulated ignitions. The Fire Occurrence Database (FOD; Short, 2021) includes all recorded ignitions from both natural and human causes from 1992 - 2018. The FOD was amended to include spatial fire data from 2019 and 2020 using state, local, and federal records.
Historical weather and fuel conditions – Burn probability was modeled using temperature, precipitation, and relative humidity and data collected from carefully selected Remote Automated Weather Stations (RAWS) Weather records for 2011 - 2021. Wind speed and direction data as collected from the same station but from a longer time period based on the individual stations’ records (e.g. 1985 - 2021). Fire intensity was modeled using the same weather variables from gridded weather data available at gridMET (https://www.climatologylab.org/gridmet.html).
Why are developed regions sometimes classified as unburnable?
OSU’s task was to evaluate wildfire risk and the potential impacts to structures. The distinction between a wildfire and a structure fire can be unclear, but, in general, when fire begins to be transmitted directly from structure to structure, as happens in more densely developed areas, it becomes a structure fire. Assessing the risk from structure fires requires a different process than the one established in SB762. However, the fire modeling conducted for SB762 did allow fires to grow into developed areas in order to refine estimates of where and how many structures and other human developments might be at risk.
What if another wildfire risk assessment identifies different levels of wildfire risk compared to this assessment?
This map of wildfire risk to structures and other human developments made by OSU scientists serves as the authoritative assessment required by SB762. There are many other wildfire risk assessments representing risk to part or all of Oregon, and they may present different results based on their specific objectives, the methods used to assess wildfire risk, the spatial scale and extent for which wildfire hazard was modeled, and the data that were input to the model. This risk assessment was specifically designed to meet the requirements and needs described in SB762. Oregon state agencies will refer to this risk assessment when executing their responsibilities in SB762.
How is the SB762 risk map similar or different from the wildfire risk maps that have previously been available from the Oregon Wildfire Risk Explorer?
The Oregon Wildfire Risk Explorer has been a source for a wide range of objective wildfire risk-related maps and data. Most recently, the Oregon Wildfire Risk Explorer has housed many outputs from the 2018 PNW Quantitative Risk Assessment (PNRA) in both the Advanced and Homeowners tools. Although both the 2018 PNRA and the SB762 risk map are quantitative risk assessments, there are some important differences. The three most important differences are:
- The 2018 PNRA was a comprehensive assessment that assessed risk to a wide range of resources and assets including wildlife habitat, infrastructure, drinking water, timber value and others. In contrast, SB762 directed researchers to only assess risk to structures and other human developments.
- The SB762 map quantifies risk at the tax lot level whereas the 2018 PNRA does not.
- The SB762 risk map is based on more contemporary landscape conditions and incorporates recent fires that may significantly alter local patterns of risk.
The public is encouraged to refer to the 2018 PNRA to learn more about how wildfire might affect additional resources and assets, but the SB762 risk map is the authoritative map related to risk to structures and other human developments. The 2018 PNRA map is in the process of being updated for both Oregon and Washington and will be available in 2023.
How will a recent wildfire impact wildfire risk values?
Recent wildfires will have variable impacts on property-level risk values according to the proximity of the fire to the property in question, how long ago it occurred, in what kind of vegetation the fire burned and where in the state it occurred. On one hand, a recent wildfire can reduce the amount of fuel available to subsequent fires in the same footprint, thereby reducing intensity. Reduced intensity can lead to a lower risk rating. Similarly, a fire footprint can affect fire spread and may result in fewer fires reaching the property in question, again resulting in a lower modeled risk. On the other hand, as a burned area re-vegetates and recovers it may actually become more flammable and lead to more fires, faster spread, and higher intensity, leading to increased risk.
Importantly, the modeled reduction in fire intensity and fire spread in recently burned areas does not last forever. The risk map will be updated at least every five years, and each time the landscape will be updated to account for changes in vegetation, including vegetation that may have regrown in historic fire footprints.
How is climate change addressed in the wildfire risk assessment process?
The wildfire risk assessment does not include future climate change projections. Rather, it recognizes that we are already living it and need dynamic assessments to continuously account for the changes and surprises we are seeing more routinely these days. The fire behavior simulations use observed, historical weather and climate data. However, compared to previous risk assessments, this map is based on more recent climate data and better accounts for emerging trends in temperature and precipitation. The risk assessment is designed to guide decision-making over the next 5 years; the next update will include the latest weather, climate and vegetation data in order to accurately reflect conditions contributing to wildfire risk.
What are the limitations of this risk assessment?
This wildfire risk map is focused on risk to structures and other human development and does not account for risk to additional resources and assets like infrastructure or drinking water. If there are questions concerning overall wildfire risk, please reference the 2018 PNW QWRA which does include these other resources and ecosystem services.
Risk assessments are not a forecast of what will happen in any given year. Even with the best science and analytics, it is impossible to predict when and where a wildfire will occur. Properties classified as moderate and low risk are still exposed to potential wildfire damage.
The landscape used to model wildfire behavior reflected by 2021 conditions, but significant changes in the landscape since then may have altered local risk values. As directed in SB762, risk to structures and other human development will be updated every five years.
Why might wildfire risk class differ between neighboring properties?
The wildfire risk to any one property reflects the surrounding landscape of that property which may or may not differ between neighbors. For instance, neighboring tax lots might be in different risk classes because one tax lot is situated closer to a large swath of flammable vegetation, elevating the burn probability and the fire intensity compared to the other tax lot that abuts largely paved, unburnable land. Additionally, there may be topographic features that could move fire towards one tax lot more often than another.
Did the fire models account for work already performed on properties to reduce wildfire risk?
It is possible that fire model outputs will capture the effect of wildfire risk reduction activities, but not necessarily.
Fuel reduction projects like thinning and prescribed burning might be accounted for depending on how well the activities were documented. The fuel data (LANDFIRE, 2016) used to model fire behavior was updated to account for fuel reduction treatments and disturbances that occurred through the end of 2021 where adequate spatial data was available. Generally, adequate spatial treatment data is only available from federal, state and, occasionally local agencies and organizations. In some cases, spatial treatment data was mapped and collected on private land, but generally that is not the case.
Defensible space activities adjacent to structures are unlikely to be accounted for in the fire models. Spatial descriptions of defensible space activities are generally not available, but even if they were, defensible space activities are usually conducted at too small a scale to be accounted for in the fire models. If a property-owner has already taken steps to reduce their wildfire risk, they may not see the risk level change in the maps. However, the property will be less susceptible from damage from wildfire and the owner is more likely to be in compliance with any of the regulatory elements described in SB762 and set by state agencies.