For more than a century, the suppression of wildland fires occurring on federal lands has been the responsibility of several agencies, including the U.S. Forest Service, Bureau of Land Management and the National Parks Service. In an effort to effectively coordinate resources on a regional and national level, 11 Geographical Area Coordination Centers have been established to coordinate the dispatch of various federal, state and local resources (aircraft, engines, crews and the like) to fires and other emergency response incidents in the geographic areas they serve. But they also depend on good meteorological, fire potential and resource information to provide cost-effective, timely and proactive coordination of firefighting forces.

Historically, the National Weather Service has been the sole provider of fire weather information to fire suppression agencies. This information has been provided mainly in a narrative format, forecasting key meteorological factors such as temperature, humidity, wind and so forth, giving fire agencies a basic view of meteorological conditions that may affect or contribute to wildland fire activity. In the past, this narrative information has proved very helpful in making decisions to manage and allocate suppression resources.

But over the course of the very severe wildland fire season of 2000, nearly 8 million acres in the western United States burned, and federal suppression efforts cost over $1 billion. One outcome of this devastating fire season was a concentrated effort by government officials to establish a national program — Predictive Services — to better assess and forecast weather, fire potential and resource allocations. Predictive Services now is executed by meteorologists, fire analysts and intelligence officers within various U.S. government agencies, and since its inception in 2000, the program has directed development of a wide range of decision support tools to help resource managers.

One of the most recent innovations has been the Seven-Day Significant Fire Potential Product, otherwise known simply as “the Seven-Day Product,” which forecasts the daily probability of large fire occurrence for a given area for each of the next seven days. The Seven-Day Product was first developed at the Pacific Northwest GACC a few years ago and has since been adapted for use by several other GACCs in the western United States. By May this year the product will be in use at all 11 GACCs, providing nationwide coverage.

SIGNIFICANT FIRE POTENTIAL

In the past, forecasting wildland fire activity and severity has been fairly subjective, relying on the forecaster's past experience and gut feelings. One of the goals of Predictive Services with the Seven-Day Product was to have a more objective tool with which to forecast fire potential.

In the context of the Seven-Day Product, fire potential is assumed to be a fire scenario measured in terms of anticipated occurrence of large fires, complex fires or number of fires, coupled with management's capability to respond. Fire potential is influenced by factors including fuel conditions (such as fuel dryness), ignition triggers significant weather triggers and resource capability. “Significant” fire potential is defined as the likelihood a wildland fire scenario will require mobilization of additional resources from outside the area in which the fire scenario originates.

What constitutes a wildland fire scenario can vary greatly depending on how the term is defined, but the simplest way is in terms of fire size. This simple definition forms the basis for national implementation of the Seven-Day Product. Thus, the Seven-Day Product is a forecast of the likelihood that a fire will grow large enough to require elevated resource response from outside the fire event area. Because these types of events occur only about 5% of the time during the course of a normal fire season, the Seven-Day Product is considered to be forecasting a rare event.

Fire managers know that assessing large fire potential is a difficult task. There are many variables, including fuel dryness, fuel type, fuel continuity, weather triggers, terrain, resource capability and suppression strategy. Theoretically all of these variables and more should be expressed in a model that forecasts significant fire potential. However, the Seven-Day Product makes some general assumptions to avoid becoming unwieldy. First, fuel type and continuity is assumed to be constant within any particular assessment area. Second, the terrain is assumed to be constant. Finally, the suppression strategy is also assumed to be constant because on the majority of fires, with the exception of wildland use fires, a full suppression strategy is employed. Thus the conceptual model for significant fire potential becomes a function of fuel conditions (dryness only), ignition triggers, weather triggers, and resource capability. (See “Significant Fire Potential Conceptual Model,” below.)

FUEL DRYNESS

For the purpose of forecasting weather and fire potential, each GACC area of responsibility is broken down into smaller Predictive Service Areas. Some PSAs have been constructed around national forests and parks, others around various fire agency jurisdictions, and still some others according to different climate and fuel regimes. Remote Automated Weather Stations in the PSAs have been used for years, not only by the fire community in assessing real time weather information during wildfire incidents, but also by agencies such as the Forest Service and National Weather Service for forecasting, research purposes and climate analysis.

The National Fire Danger Rating System developed by the Forest Service in the late 1960s, with a first version release in 1972 and a later version in 1978, provided the fire community with a suite of indices that help predict fire activity and behavior at each weather station. Some of the many index outputs include the energy release component; ignition component; and one-, 10-, 100- and 1,000-hour dead fuel moistures (F1, F10, F100 and F1000 respectively).

In the past, fire agencies might have used thresholds such as the 90th or 97th percentile of the ERC as a simple way to determine potential fire activity. The problem with this method is that in many areas, large fires frequently occur below the 90th percentile, and therefore a significant number of large fire events would be missed. A quite practical way to determine fuel dryness is to relate fire danger indices with fire occurrence, using statistical methods as such studies have done in the past.

The fuel dryness concept employed in the Seven-Day Product, however, uses a three-tier threshold approach, incorporating indices such as ERC, F10, F100, or F1000 either by themselves or in combinations of two (usually ERC and one of the dead fuel moisture parameters). Once several weather stations have been selected to represent each PSA, a statistical analysis is performed to correlate historical NFDRS indices with historical fire occurrence data to determine fuel dryness thresholds for each PSA.

The three categories of fuel dryness are:

1. Moist

   This level of dryness equates to approximately to a less than 3% chance of having a large fire event, and is represented by the color green on a Seven-Day Product report.

2. Dry

   This category represents normal dryness of fuels, equates to a chance of having a large fire based solely on fuel dryness of between 3% and 8%, and is indicated in yellow on the report.

3. Very dry

   This level of dryness represents a greater than 8% chance of a large fire and is displayed as brown on the report.

In all categories, the chances of a large fire event are calculated based on the absence of significant weather or an ignition trigger. Essentially, the three levels represent below-normal, normal and above-normal chances for having a large fire in any given area over the coming seven days based on fuel dryness alone. It's easy to conclude that even with very dry fuels, the probability of a large fire is still quite low, but this simply reveals that fuel dryness by itself is inadequate in assessing large fire potential.

IGNITION TRIGGERS

Across the western United States, most fire starts during the summer months are due to lightning. However, in the more populated areas of the country where there is a large interface between urban areas and wildlands, fires attributed to human activity account for over half the total number of ignitions. In some portions of southern California, arson alone is responsible for 20% of large fires. Unfortunately, human activity in relation to large fire occurrence is rather unpredictable, so it must be left out of the simple model equation with a few exceptions.

In general, during the warmer months of the year, weekends and national holidays in the United States usually promote a higher level of human activity with regards to outdoor recreation. Most areas — especially in the western United States — experience some increase in fire activity during these periods, especially around Independence Day due to fireworks. Even though the human factor involved during these national holidays can be unpredictable, the increase in widespread outdoor recreation is quite predictable, and the Seven-Day Product factors in such predictable increases in widespread outdoor recreation.

Beyond the human factor, lightning is one of the primary parameters affecting the prediction of large fire potential. While not all lightning events cause large fires, the probability of a large fire occurring is a function of the spatial coverage of lighting strikes and whether significant precipitation has occurred in the vicinity of those strikes. (In addition, the type, concentration and receptivity of fuels determine how many strikes, on average, are required to cause a large fire in a particular area.)

Currently, the logistic regression equations developed at the Pacific Northwest GACC to determine probabilities of amounts of lightning per day, per area provide the most objective forecast of lightning downstrikes that may occur. Research, however, is ongoing, and will improve the Seven-Day Product as these types of events become more accurately predictable.

Meteorological factors, such as wind, have proved to be the catalyst necessary for the development of a large fire or for causing significant growth of existing large fires. But because these meteorological factors do not start fires themselves, correlating these variables with large fires is difficult. A statistical study can help determine what kind of relationship exists with large fire occurrence.

Wind is obviously a major contributor to large fire activity, but it's difficult to quantify due to the complex factors associated with velocity and direction. As with lightning, the fuel type and continuity become a function of how much wind is needed to significantly affect the growth of a fire.

RESOURCE CAPABILITY

To date very little research has been conducted in the area of resource capability, mainly because of the lack of reliable data. Practically speaking, it's obvious that as resources become less available, the probability of large fire activity increases, provided that new ignitions occur. The use of fixed-wing or rotary aircraft becomes the most effective means for having a successful initial attack because of an aircraft's ability to make large water or retardant drops and also because of its sphere of influence. Consequently, because aircraft are the primary apparatus used in initial-attack situations, the preliminary research possibly could be limited to discovering the relationship between large fires and the number of aircraft that are readily available per area, per day.

The Seven-Day Product model uses fuel dryness, weather and resource information to make the final forecast assessment regarding where and when large fires are likely to occur. The model assumes a layer approach to these variables, with fuel dryness serving as the foundation for the product. Once the fuel dryness has been accurately assessed, any significant meteorological parameters are factored in. Finally, resource capability becomes the last consideration in the overall product outcome.

The variables that are used to assess fuel dryness — the ERC, F10, F100 and so forth — are projected seven days into the future using regression equations that have been developed through the Desert Research Institute of Reno, Nev. The equations are based on various meteorological parameters that simulate the NFDRS indices generated by the Weather Information Management System.

To forecast these fire danger indices, the equations that DRI developed make use of gridded data from the Global Forecast System numerical atmospheric model. Model Output Statistics are generated by DRI and include fire danger indices and weather parameters, such as temperatures, humidity, winds and so forth. Model Output Statistics are produced daily for the 00Z and 12Z runs of the Global Forecast System model for various weather stations in each PSA at each GACC.

The Seven-Day Product report takes the form of a chart that contains fuel dryness information for each PSA, with any significant weather factors denoted by various symbols. (See page 31.) If significant weather is forecast to occur with either “dry” and or “very dry” fuels, red is used to highlight that particular area and the day is labeled as a high-risk day. A high-risk day is defined as a day where an ignition or significant weather trigger and an appropriate level of fuel dryness combine to create conditions that historically have resulted in a significant fire event for a particular PSA. In addition to the chart, the report includes a narrative section highlighting local weather and fuel conditions, as well as available resources for initial attack activity.

Currently, six of the 11 GACCs are making use of the Seven-Day Product either on a daily or every-other-day basis. It's expected that by May all 11 GACCs will be using the product to issue reports on a daily basis. Initial feedback from the field has been positive, and most users across California, Washington and Oregon fully support the product and use it daily. Other GACCs that have recently started using the product this year have had positive response as well.

Initial verification of fuel dryness as the primary predictor of the product is promising, but the number of high-risk days being forecast is still significantly more than would be reasonably expected. This is mainly due to the fact that the criterion that defines high-risk days has yet to be objectively determined and therefore is still primarily subjective.

Beginning next year, a full suite of verification parameters will be required by each GACC to determine the validity of the product. Parameters such as ERC, F100, F1000, fuel dryness, and high-risk days will be verified with hope that the product will continue to improve and become a more commonly used tool among the operational field.

With the availability of more and faster computing power and a variety of software applications that incorporate Geographical Information Systems mapping tools, several variants of the Seven-Day Product are likely to be developed. One idea that has been envisioned by Predictive Services is a regional or nationwide map showing dryness levels and any high-risk days for each of the seven days. This type of product would make the Seven-Day Product truly “national” not only in content but also in scope. Other products might include displays of the various output parameters from maximum temperatures, minimum humidity, ERC and so forth.

Tom Rolinski is a fire weather forecaster for the Bureau of Land Management. He is stationed at the Interagency Fire Weather Center, Riverside, Calif., which is operated by the California Wildfire Coordination Group.

SEVEN-DAY SIGNIFICANT FIRE POTENTIAL CONCEPTUAL MODEL

Weather Synopsis

A trough currently over California will give way to a strengthening area of high pressure over Arizona and New Mexico. This will cause temperatures to warm over much of the area by late in the weekend with a warming trend continuing into much of next week. Moisture from the southeast will increase over the Sierras by early next week, bringing a chance of thunderstorms to those areas by then.

Fire Potential

With the anticipation of warmer temperature later this weekend and next week, general fire activity will likely be increasing as fuel conditions worsen. Probabilities for large fire development will increase considerably over the higher terrain of the Sierra Range Tuesday and Wednesday as lightning activity increases.

Resource Capability

Currently, sufficient resources are available for light to moderate initial attack; however, should significant fire activity develop next week, additional resources will likely be requested.