Wildfire Suppression Funding and Costs

Initial response: The closest available state, local, or federal agency responds, regardless of jurisdiction

Statewide, initial fire response operates on the "closest forces" concept. JLARC staff analysis found this approach is common nationwide. This means that the closest available resources are dispatched to a fire regardless of which agency they belong to and regardless of which agency has jurisdiction over the land. The Department of Natural Resources (DNR) may respond to a fire on federal land, and federal agencies may respond in kind to fires on DNR land. Not all local fire agencies have such arrangements with DNR or federal agencies.

Extended response: Multiple agencies may be dispatched to large fires

If the initial response does not suppress or contain the fire, agencies typically will use an Incident Management Team (IMT). IMT members may come from different jurisdictions. They all are trained and certified to the same national standards for handling fires. The IMT has one incident commander and multiple sections with varying responsibilities. For example, the finance section is responsible for timekeeping and tracking costs, while the operations section directs the firefighters on the ground.

The IMT determines the resources needed to fight the fire. Dispatchers have protocols for the order in which they call resources: closer resources are called before those that are farther away. Large fires often cross jurisdictional boundaries and involve many agencies.

The State Fire Marshal may authorize state mobilization of resources to help a local fire agency with a large fire

If a local fire agency needs more resources to suppress a fire in its jurisdiction, it works with nearby agencies under Under mutual aid agreements, agencies help one another at no costmutual aid. If resources beyond mutual aid are needed, the local fire chief can ask the State Patrol Chief to authorize a mobilization, which will be coordinated by the State Fire Marshal (a part of State Patrol). More detail is in Appendix 2.

Under a mobilization, the Fire Marshal coordinates resources to help the district with the fire in its jurisdiction. The state is responsible for the costs associated with fire suppression, including personnel, equipment, and transportation. The Fire Marshal does not directly suppress fires, but does work with the IMT. A mobilization ends when the threat to the district is controlled, even if the fire continues outside the district (e.g., on federal land).

DNR takes a decentralized approach to fire suppression

DNR is responsible for direct fire suppression on state and privately owned forest lands subject to forest fire protection assessments. The agency delegates fire suppression and most supporting business operations to its six regions:

• Each of its six regions manage firefighting responsibilities, including dispatch and business operations, within the regional boundaries. The regions operate autonomously, but share firefighting staff and equipment during the fire season. A map of the regions is in Appendix 1.
• DNR also has a central wildfire division that provides support to the regions, negotiates agreements with federal agencies, and manages out of state dispatches.

View Executive Summary | Next Section

The Legislature asked JLARC staff to determine how the Department of Natural Resources (DNR) and the State Fire Marshal (Fire Marshal) share costs for fire suppression with federal and other agencies.

Initial response: Response provided without reimbursement

Under mutual aid agreements, agencies help one another at no cost during the initial response (about one work shift). Between state and federal agencies, this is known as reciprocal response, rather than mutual aid.

DNR has agreements for initial response and mutual aid with local fire agencies. Each DNR region manages the agreements with local fire agencies within its boundaries. The state's master agreement with federal agencies and Oregon includes reciprocal response.

Mutual aid and reciprocal response agreements are common nationwide. The underlying assumption for the agreements is that over time, all agencies mutually benefit, and that the approach helps to keep fires from growing. In fact, a common measure for wildfire agencies, including DNR, is to keep at least 95 percent of fires under 10 acres.

Extended response: Master agreement between federal agencies, Washington, and Oregon governs cost-sharing

Large fires often cross jurisdictional boundaries and involve resources from many agencies. Washington is party to a master agreement negotiated between the state, Oregon, and federal agencies that governs response and cost sharing. DNR is the state's signatory to the agreement. Both DNR and the Fire Marshal participate in cost sharing discussions with other agencies.

The agreement identifies how agencies can share costs and provides guidelines for use. The agencies mutually choose the cost-sharing method for each fire and document the decision in a fire-specific agreement. This approach is standard practice across the country. JLARC staff found that for a sample of fires between fiscal years 2010 and 2016, costs were shared according to guidance from the master agreement, as well as provisions in fire-specific agreements.

DNR, the Fire Marshal, and federal agencies have separate agreements with local fire agencies, other state agencies (Washington and other), and private vendors to provide supplies and services during a fire (see Section 3 for additional detail). Additional agreements address cost sharing and responses with other states (e.g., Idaho) and Canada.

Agencies report that their preferred cost-sharing method for large fires is based on the benefit received by each agency

In interviews, staff from DNR, the Fire Marshal, and the U.S. Forest Service (USFS) reported that they prefer to use a cost-sharing method called "percent of effort." It has the following elements:

• Pay for Benefit Received: Each agency pays a percent of the cost based on the resources expended on its lands (benefit received). Both federal and state staff stated that they consider this to be the fairest approach.
• Based on Estimated Total Costs: While each agency will pay staff and contractors it hired for actual costs, the cost-sharing approach is based on estimated total costs. DNR, the Fire Marshal, and the USFS report that they use estimates because it improves efficiency, and that waiting for actual expenditures could delay reimbursement by months or years. JLARC staff identified that some states also use estimates, while others use actual expenditures for cost sharing.
• Post-Incident Review: DNR regions and the U.S. Forest Service reported that the cost share calculation and supporting documentation are reviewed post-incident for accuracy, and that the process can take months for large fires. While neither DNR nor the Fire Marshal have agency-wide policies and procedures for conducting post-incident reviews, JLARC staff found records of those reviews when we analyzed fire incident records.

Since agencies pay a portion of the cost directly, this cost-sharing practice can reduce the amount of interagency reimbursements. For example, at the Mills Canyon fire shown below, the USFS received a total benefit of $5.6 million. But, it paid an estimated $4.5 million directly to staff and contractors, so its reimbursement to DNR was $1.1 million.

Exhibit 2.2: Each agency pays a percent of the cost based on the resources expended on its lands (benefit received). The 2014 Mills Canyon fire offers a real-life example.

Source: JLARC staff depiction of master agreement and Mills Canyon fire records.

Acres burned may not be a reliable method for understanding or sharing costs

The master agreement states that sharing costs based on acres burned is allowed but not preferred. Different types of land require different suppression efforts. The efforts vary in cost. For example, a fire line built with bulldozers has a different cost than one built by a hand crew. Both differ from dropping retardant from an aircraft. The effort related to these methods may have little to do with the size of the fire. As a result, acres burned may not be a good indicator of cost. Likewise, suppression cost per acre on its own is unlikely to be a useful measure for understanding cost.

Exhibit 2.3: Acres burned may not be a good indicator of suppression costs

Source: Northwest Interagency Coordination Center Annual Reports 2014 and 2015. Data reflects acreage and estimated costs for all agencies before cost share or reconciliation.

Previous Section | Next Section

The Legislature asked for information about fire suppression funding and costs for the Department of Natural Resources (DNR) and the State Fire Marshal (Fire Marshal). Below is a review of how each agency estimates costs of fires or mobilizations, the types of expenditures included, and whether expenditures are related to fires.

Agencies have different approaches for estimating costs of fires

DNR and the Fire Marshal routinely have expenditures that exceed their base appropriations for fire suppression or mobilization. The Legislature fully funds the work with supplemental and special appropriations. This approach is used by other western states and British Columbia.

Both agencies use a combination of actual and estimated expenditures in their supplemental requests:

• DNR manually compiles estimates from at least four different data systems, and makes adjustments based on the professional judgment of division and regional staff. This process is complicated by DNR's approach to documenting incident codesDNR may assign multiple codes to single fire and lacks a list of all codes (see Section 5). The estimates do not include any reimbursements expected from the cost share for large fires or fire complexes. A complex is multiple fires combined under a single incident command structure.
• The Fire Marshal creates estimates from spreadsheets that track the resources for each mobilization. These spreadsheets are based on time sheets, resource orders, agreements, and other source documentation.

Like all state agencies, DNR and the Fire Marshal submit their initial supplemental budget requests in September, before the end of the fire season. At this point, much of the information is incomplete. As the agencies update the information, the costs of some fires change, even if the overall fire season estimate stays relatively stable. Reasons for the changes include updating estimates with actual costs, shifting costs between fires within a fire A complex is multiple fires combined under a single incident commandcomplex, and changes to the state's portion of the cost share during the Cost share calculation and supporting documentation are reviewed for accuracy after the incident. See Section 2post-incident review.

Agencies have procedures to track actual fire costs separately from other expenditures

DNR and the Fire Marshal each use budget codes and management procedures to track fire-related expenditures and help ensure that costs can be associated with a fire, fire activity (e.g., Strategically positioning resources before a fire.prepositioning), or mobilization. For example:

• Each agency has a set of high-level codes that indicate expenditures are related to fire or mobilization.
• Each agency assigns a code to a fire, fire complex, mobilization, or other activity. Expenditures, including agency staff time, are charged to the code. Some DNR fires have multiple codes, as discussed in Section 5.
• For a large fire, each resource assigned to the fire is given a unique resource order number. For example, if a contractor provided two engines with crews, each would have its own number. The incident management team checks the number at the fire for accountability and tracking. Resources cannot be paid without the resource order number.
• Agency staff review invoices from contractors against documentation from the fire before payment. JLARC staff compared a sample of expenditures against source documentation and was able to verify the costs.

DNR and the Fire Marshal spent $434 million on fire-related activities between fiscal years 2010 and 2016

Fire-related activities include DNR's response to fires and false alarms, Strategically positioning resources before a fire.prepositioning resources based on weather and fire conditions, and the Fire Marshal's coordination of state mobilizations. Costs include firefighters and other personnel, aircraft, equipment, and camp supplies.

Exhibit 3.1: Annual spending by DNR and the Fire Marshal between fiscal years 2010 and 2016 (dollars in millions)

Agency	2010	2011	2012	2013	2014	2015	2016	Total	Notes
DNR	$26	$16	$13	$47	$31	$89	$146	$369	Major fund sources are State General Fund, the Disaster Response Account, and the Budget Stabilization Account. Contracted goods/services and salaries/benefits are the largest expenditure categories.
Fire Marshal	$2	$4	$4	$12	$3	$18	$22	$65
Total	$28	$21	$17	$59	$34	$107	$168	$434

Source: Agency Financial Reporting System (AFRS), dollars in millions. The totals may not match the sum of individual parts due to rounding. Fiscal years are July 1 through June 30, so fiscal year expenditures reflect fire activity in different calendar years. For example, expenditures for a fire in August 2015 are included in fiscal year 2016.

Majority of spending is on contracted goods and services, including contracts with local fire agencies

DNR and the Fire Marshal contract for services including suppression, fire camp operations, and equipment. Contracts — including those with local fire agencies — account for the largest portion of spending. Salaries and benefits are the second largest. For the Fire Marshal, most of the money spent on salaries and benefits is for volunteer firefighters who are temporarily paid as state employees during a mobilization,

Exhibit 3.2: Fire suppression and mobilization spending is largely contracts, salaries, and benefits

Source: Agency Financial Reporting System (AFRS) data, fiscal years 2010-2016.

DNR and the Fire Marshal establish agreements and rates with contractors before each fire season begins. In 2017, they began to use consistent wage and equipment rates. This avoids creating an incentive for contractors to work with only the higher-paying agency. DNR and the Fire Marshal also are beginning to use a federal system with lists of pre-approved contractors and negotiated rates.

Established procedures and controls are intended to prevent DNR and the Fire Marshal paying the same contractor for the same work

Contractors can work for different agencies at different times during a fire. There are procedures and controls in place at the fire intended to prevent a contractor from being paid by both DNR and the Fire Marshal for the same work. For example, the agencies use different resource order numbers when hiring contractors, and agency staff at the fire verify the hours worked and appropriate coding. Also, the resource order number changes at each fire.

Some expenditures are recouped from other agencies and responsible parties

The state has three ways to recoup suppression and mobilization expenditures: reimbursement from federal or other state governments, DNR's recoveries from private parties, and federal fire management assistance grants.

Exhibit 3.3: DNR and the Emergency Management Division (EMD) report that the state has recouped $72 million to date between fiscal years 2010 and 2016 for fire suppression

Source	Description	Credited to	Amount Received (FY10-16)
Reimbursement from federal and state partners (e.g., Oregon)	Reimbursement under cost sharing agreements Reimbursement for responses to out of state fires	DNR receives reimbursement and recovery if it is paid in the same fiscal period as the fire. Reimbursement received after the fiscal period is deposited into the State General Fund balance.	$34.6 million
Recoveries from private parties that negligently start fires	May be paid from homeowner's insurance policies May not recover full cost of suppression		$2.9 million
Fire Management Assistance Grants (FMAG) from the Federal Emergency Management Agency (FEMA)	Reimburse 75% of eligible expenditures Limited to fires that threaten to become a major disaster: 37 fires declared to be FMAG eligible between FY 2010-16	Deposited directly into the State Disaster Response Account.	$34.4 million

Sources: Reimbursement and recovery figures provided by DNR, FMAG data provided by the Emergency Management Division (EMD).

Costs may be recouped after the fiscal period in which the fire took place

There are multiple reasons that reimbursements may be submitted after the fiscal period in which the fire occurred. For example, fires can happen near the end of the fiscal year, agencies may need to review cost share documentation, or a cost recovery may be litigated. The time lag for reimbursements and recoveries makes it difficult to determine the actual cost of any fire season or fire.

The Emergency Management Division (EMD) of the Washington Military Department is responsible for submitting FMAG applications for the state. As of June 2017, the state had received partial FMAG reimbursement ($3.8 million) for fires that took place in the 2014 through 2016 fire seasons. EMD has not yet submitted some requests and others are awaiting FEMA review and approval. EMD reports that requests may be delayed due to the complexity of the application, number of fires, volume of records, delays in federal documentation, and competing operational responsibilities.

Local governments also may receive FMAG reimbursement. EMD has processed, and FEMA has paid, reimbursements for some local fire agencies and governments.

Previous Section | Next Section

DNR and the Fire Marshal prepare public reports and respond to legislative questions about costs of fires and mobilizations

Like other state fire agencies and the U.S. Forest Service (USFS), the Department of Natural Resources (DNR) and the State Fire Marshal (Fire Marshal) are frequently asked to provide information on fire season characteristics (e.g., location, cause, acres burned) and costs. While fire characteristic data is not needed for agency budget requests, it can inform policy discussions.

DNR staff reported that, over time, the agency has added fire characteristic fields to its data system in response to inquiries from individual legislators. However, this data has not been consistently maintained and is not used for management purposes.

Exhibit 4.1: DNR and the Fire Marshal present information about costs and characteristics to the Legislature

Source: DNR and Fire Marshal presentations to the Legislature, 2016.

The State Fire Marshal keeps records of its mobilizations

The Fire Marshal coordinates state mobilizations to help local fire agencies, and does not directly fight fires. There were 93 mobilizations from fiscal year 2010 through fiscal year 2016.

When a local fire agency requests a mobilization, the Fire Marshal receives initial fire characteristic data such as the location, complexity, size, and values (homes, infrastructure) at risk. The Fire Marshal staff reported that they file this information with other mobilization documents, but do not update it during or after the mobilization. The agency also has access to reports completed at each mobilization by the fire Incident Management Team. The Fire Marshal staff reported that the agency periodically uses characteristic data from these reports for internal and external reporting.

While the Fire Marshal maintains basic information about mobilizations (e.g., county, cost), it does not have a system to compile and report details about the mobilization (e.g., structures threatened or destroyed).

DNR's data on fire characteristics is often unreliable, incomplete, or unused

DNR is responsible for directly fighting fires. Due to the size and scale of its responsibilities, DNR is the state agency with the most information on wildfire characteristics.

Like agencies across the country, DNR reports characteristics of large fires to the National Wildfire Coordinating Group (NWCG). DNR does not always update its own data systems, so its information may either match or contradict NWCG data.

Concerns with DNR's data systems and data entry protocols

DNR collects fire characteristics for large and small fires in a database called Emergency Incident Report System (EIRS). It is a reporting system that is not used to manage fire suppression or track actual costs. The system has over a hundred fields, and staff consistently reported that most data is unused.

There are limited internal controls over data entry and updates. As a result:

• Basic characteristic data such as acres burned and incident type can be unreliable. EIRS entries may reflect the number of acres at the start of a fire, end of fire, or a point in between. DNR stated that the NWCG data for total acreage is more reliable than its own. DNR region staff also reported that they use different approaches for counting and classifying A report where a response was dispatched but no fire was found, or a controlled fire not requiring suppression.false alarms and fires that require limited response (e.g., extinguishing a campfire).
• Dozens of fields have incomplete data. Data on land cover, structures damaged or destroyed, and why a fire escaped initial attack is often not entered into EIRS. DNR staff stated that some fields, such as tree species involved in a fire, remain incomplete because neither the regions nor headquarters use the data.
• Some characteristics are unreliable or incomplete because staff lack guidance or appropriate choices. DNR staff stated that fields such as bushels of wheat or timber board feet burned lack data because staff do not know how to calculate the value. Other fields have limited options. For example, only one agency can be entered in the jurisdiction field even though fires often burn across jurisdictional boundaries.

Exhibit 4.2: DNR's data about acres burned is unreliable

Source: DNR presentation to GGIT, January 2016. DNR data from EIRS.

DNR lacks an efficient way to identify the costs of individual fires

DNR has a financial reporting system called DataMart that is separate from EIRS. It tracks fire expenditures by project codes. Each fire may have one or more project codes assigned to it. DNR does not maintain a centralized list of project codes.

• A fire may have more than one code for tracking costs and characteristics. For example, DNR may need to separately track costs for federal reimbursement or may have a code for costs shared across a fire A complex is multiple fires combined under a single incident commandcomplex.
• DNR lacks a single index of all project codes. While one code per fire is always entered in EIRS, other codes are recorded in multiple locations. Codes may be in the A physical box with the supporting documentation for the firefire box or in spreadsheets at headquarters or the region. The spreadsheets vary in format and content. Currently, the only way to match the fire name with all relevant project codes is to manually gather the information.
• The full cost of a fire (or group of fires) cannot be identified without project codes from all locations. DNR'S financial reporting system can be searched only by project code, not by fire name. Without a centralized list of project codes, there is no systematic way to identify costs.  For example, to identify the costs of all fires with the highest Fires are rated on a five point scale reflecting threat, fire behavior, and other factors. See Appendix 1.complexity in fiscal year 2016, DNR would need to manually compile all project codes assigned to 38 different incidents (see Appendix 1).

Legislative Auditor recommends improving fire data collection and reporting

The Legislative Auditor makes three recommendations:

1. DNR should refine its collection of key data elements and seek input from the appropriate Legislative committees.
2. DNR should improve the accuracy and reliability of the key data elements it collects.
3. DNR should develop a systematic and verifiable way to identify the costs of individual fires.

The Department of Natural Resources concurs with these recommendations. You can find additional details on the Recommendations tab.

Previous Section | Next Section

The Legislature requested information on how development, land ownership, and insurance affect fire suppression costs. To answer these questions, JLARC staff reviewed agency data and current research on suppression costs. The research can inform policy discussions about fires, but is not intended for budgeting or predicting fire costs.

Research indicates that development is one of several factors that may affect cost

Published research indicates that factors such as topography, weather conditions, fire size, and fuel buildup can influence the cost of a fire. Home construction in high fire-risk areas such as the Areas where homes are built in forests, grasslands, or shrub lands that are otherwise undeveloped.wildland-urban interface (WUI) also can affect the cost. In our literature search, JLARC staff noted that many articles and reports associate increasing fire costs with development in the WUI, but few estimate the magnitude of the relationship.

• Fire suppression costs can increase when homes are nearby. However, additional homes in a low-density area increase costs by a larger factor compared to areas that are already high-density.
• Additional effort spent protecting homes may reduce available resources for suppressing other fires. This may increase overall costs.

• Literature suggests (and Department of Natural Resources (DNR) staff reported) that suppression tactics on developed lands can be different than those employed on undeveloped lands. Tactics to prevent fire from reaching structures, such as digging a fire line in front of a home rather than taking advantage of a natural fire break like a road, may increase suppression costs.

Land ownership can affect suppression goals and land management decisions

Land ownership is not cited as a factor that directly affects fire suppression costs. But, ownership can affect decisions about how fires are fought. For example, DNR has suppression responsibility on state and private forest lands, while the U.S. Forest Service (USFS) is responsible for federally-owned land. The USFS has different fire management options, including discretion to allow remote fires to burn. Land ownership also influences land management decisions that can affect cost. That is, public and private landowners have discretion over preventive measures such as thinning trees, clearing debris, or creating A natural or manmade change in fuel (e.g., a gap in vegetation) that acts as a barrier so that fires be more readily controlled.fuel breaks.

Research reviewed by JLARC staff does not cite private insurance as a factor that influences fire suppression costs or activities

Existing research does not cite private insurance, such as homeowners insurance, as a factor that affects fire suppression costs. Information on fire insurance coverage is proprietary and not publicly available. It is impossible to determine from publicly-available information how many homeowners or landowners in Washington are covered for fire damage. JLARC staff found that fire insurance for timber is rare, and that only a small percentage is insured worldwide. In interviews, DNR staff told JLARC analysts that incident management teams do not inquire or know about insurance when deciding where and how to attack fires.

Still, homeowners insurance can play a role in other aspects of fire prevention and response:

• Some insurance companies may reduce fire risk through premiums and education. A 2016 study found that rate increases have not discouraged building in fire-prone areas, but that insurers are moving toward requiring safer landscaping standards or offering discounts for fire-safe practices.
• Some insurance companies in Washington contract with private fire-fighting companies to protect insured properties.
• If an insured homeowner negligently causes a fire, DNR may recoup fire suppression costs from the home insurance policy.

No research specifically focused on Washington's fires

The research cited above is national or focused on a few fires in specific states. To date, researchers have not focused on Washington, and the state has not contracted for such studies. Similar research likely could not be completed with the data Washington currently collects.

• Neither DNR nor the State Fire Marshal track home-building or other development data, so studies that rely on house locations cannot be replicated. Demonstrating a change in costs over time due to development would require historical data on homes in Washington's wildland-urban interface, which does not exist.
• DNR tracks whether the fire started on public or privately owned land, but this information is unreliable (see Section 4).

Previous Section | Next Section