Return to Table of Contents
Appendix J
Fire and Fuels

Table of Contents

Section I Description of the Alternatives

Section II FARSITE Report

Section III Fire Behavior Predictions and Modeling

Section 1

This section provides additional information on:
· the fuel management strategy of each alternative,
· how each alternative was developed for modeling purposes,
· how fuel treatments would be prioritized for implementation, and
· the design criteria used to analyze cumulative effects of the alternatives.

Alternative 1

Fuel Management Strategy. The three Forest Plans (USDA FS 1988; 1990; 1992a) were not designed to function as comprehensive fuels reduction or wildfire protection documents. Each Plan addresses and allows fuels management activities and defensible fuel profile zones (DFPZ) can be constructed under existing management direction contained in each of the Plans. Most fuels reduction work accomplished under the Plans is accomplished in association with logging operations and prescribed burning to reduce natural fuels accumulations.

Many of the silvicultural prescriptions that are implemented through timber harvest projects today (about 25,000 acres a year within the project area) are designed with the objectives of establishing fire-tolerant stands in terms of tree species, size and structure of the stand, and reducing the potential for crown fire initiation and spread. Flexibility in determining the location on which to conduct fire hazard reduction projects is generally limited to those areas selected through the normal Timber Sale Program of the three Forests, which can result in strategic placement if timber targets are also achievable.

The current level of the hazardous fuel reduction program treats approximately 16,000 acres a year. Most of the projects accomplished in this program are underburns aimed at either the maintenance of past fuel reduction areas or are associated with current harvest operations to reduce natural fuel accumulations, including those created through the harvest operations. While these projects are generally located in areas most at risk to large fires, coordination between the three National Forests to maximize the benefits of these activities may be difficult to attain under this alternative.

Map L displays areas that have been treated in recent years in the timber sale program, hazardous fuel reduction program, or the Forest Health Pilot Program, which is similar to what is being proposed under Alternatives 2, 3 and 4. Most areas shown on Map L are sufficient in size and location to be effective in reducing the rate of fire spread. Some of the smaller areas are shown because they are locations where future treatments could provide connectivity to establish control points. It is not known however, if this rate of accomplishment of strategically placed projects would continue in the future. Many of the projects displayed on Map L were accomplished through the "Salvage Bill" (Public Law 104-19), when emphasis was placed on treating areas of high fire hazard, and with Forest Health Pilot funds that may not be available in the future. Approximately 22,000 acres of DFPZs have been accomplished recently in the pilot project area.

Alternative 2

Fuel Management Strategy: This alternative reflects the strategy outlined by the Quincy Library Group, as summarized in the QLG Fuelbreak Strategy, which was adopted at the QLG meeting on January 30, 1997. This strategy places the highest priority on developing a network of defensible fuel profile zones (DFPZs) as the first stage of a comprehensive strategy to improve protection from large-scale, high-intensity wildfire (exemplified recently by the 46,800 acre Cottonwood Fire, Sierraville District/Tahoe National Forest and the 33,000 acre Clarks Fire, Plumas National Forest), by improving suppression efficiency.

Suppression efficiency would be improved under this strategy by creating an environment where wildfires would burn at lower intensities and where fire fighting production rates would be increased because less ground fuels and small diameter trees would need to be cleared for fireline construction or backfiring (removing the fuels under controlled burning conditions prior to the wildfire reaching the DFPZ). Aerial retardant application would also be more efficient under this strategy because the open canopy would allow the retardant to penetrate and be more effective at slowing fire spread in the light surface fuels.

Fourty to sixty thousand acres would be treated each year in strips of about one quarter mile wide, located where possible along existing roads. Each strip would provide areas of lower snag densities and down logs along the primary control point (generally a road) to provide an efficient base for suppression activities to occur more safely and efficiently.

After the initial network is constructed during the 5-year pilot project, the long term strategy may consist of additional strips that further sub-divide the network, or a shift to area fuel treatments. These area fuel treatments would also aimed at reducing fuel beds and fire ladders, but without the need to improve safety or efficiency of suppressing wildfires within these areas. Instead the objectives of the area treatments would be to reduce the intensity of a wildfire, thereby reducing the damage done to the stand by wildfire. Over the long term, maintenance of area fuel treatments and DFPZs is the goal.

Alternative 2 is consistent with the first goal of the fuel management strategy outlined in the Sierra Nevada Ecosystem Project (SNEP) Final Report to Congress (Weatherspoon and Skinner 1996). Specifically, the final report recommends that during short-term, planning and implementing, DFPZ networks should have a high priority for management of low- to middle- elevation Sierran forests (SNEP Volume I. p. 1486). In the long-term, this strategy could also provide the basis for achieving the second and third goals, which is to restore more of the ecosystem functions of low- to moderate - severity fire, thereby improving the health, integrity, and sustainability of the ecosystem.

The DFPZ network proposed in Alternative 2 is consistent with the goals and objectives of the Federal Wildland Fire Management Policy 1995. Specifically, criteria for establishment of DFPZs will be consistent with the principles, policies and recommendation of the report. When DFPZs are successful in allowing for the suppression of wildland fire, suppression costs will be less. The creation of DFPZs in the urban interface will assist local entities such as "fire safe councils" in achieving their goals. The use of prescribed fire to establishment and maintain DFPZs will be determined at the project level and take into consideration the natural role of fire. DFPZs will be linked together for continuity across district and forest boundaries and are proposed in areas adjacent to the Lassen Volcanic National Park and Caribou Wilderness to establish a protective buffer to enhance prescribed wildland fire use in those areas. Once Fire Management Plans are developed for the three Forests, the creation of DFPZs allows for the use of wildland fire in an appropriate mannner to protect, maintain, and enhance resources and allow it to be used in its natural ecological role. DFPZs can also be used to establish anchor points for future prescribed burning projects.

How The DFPZ Network was Developed . The strategic system of DFPZs (Figure 2.2) was developed to show general locations for the purposes of analyzing cumulative effects. The DFPZ network was originally developed by District employees at the seven Ranger Districts within the Pilot Project area. These District networks were then merged and reviewed together by fire management personnel from the seven Districts and the HFQLG Interdisciplinary Team, to insure that the network was as efficient as possible while taking into consideration places that needed to be avoided as required by the ACT. The most common factors considered when developing the DFPZ network was the location of existing roads and areas that would be effective at either controlling the forward or lateral spread of a wildfire. Spatial relationship to high value areas, such as Wildernesses, Off-Base and Deferred Areas, was considered as well as natural fire barriers. Another consideration was if most of a segment could be completed without having large breaks in the network from private land, riparian buffers and California spotted owl habitat areas (SOHA) and protected activity centers (PAC).

Project analysis would determine the exact locations of DFPZs. Adjustments may be made based on additional information at a finer scale that would improve the efficiency of DFPZs by placing them in areas that do not conflict with other resource needs (i.e. avoiding PACS, large cultural resource sites, etc.), and taking advantage of areas already provided with a wildfire control point.

How DFPZs Segments Would be Prioritized for Implementation. Because the network in this alternative crosses areas where the risk of fire occurring is variable, prioritizing the placement of DFPZs is critical. The criteria to prioritize which DFPZ segments would be constructed first would be consistent with some of the principles, policies and recommendations in the Final Report of the Federal Wildland Fire Management Policy and Program Review of 1995. Generally, DFPZs would be placed where they could improve protection of human life, and provide protection to property, natural, and cultural resources.

Specific elements used to rank the DFPZs are:
· Protection to populated or high use areas.
· Protection to private property, high value natural or cultural resources.
· Susceptibility of the area to stand replacement fires.
· Fire frequency of the area.
· Existing road systems that allows for efficient suppression response and product removal (sawlogs and biomass).
· Spatially contributes to the initial goal of establishing a network across the entire project area, before the subsequent goal of further subdividing the landscape.
· DFPZ could it be established with little modification of existing vegetation.
· Risk of cumulative watershed effects high or very high.
· Areas where the natural fire return interval is thought to be frequent.

Ranking DFPZ segments by these elements would result in groupings of segments that have higher priority. Beginning with the highest priority groups, an implementation schedule could then be developed using these criteria and other relevant factors to insure that the goal of establishing the network across the landscape is achieved.

What the DFPZ Network Looks Like Across the Landscape. Spatially, DFPZs are generally located on ridgetops, broad valley bottoms or flat terrain, and adjacent to urban interface or high use areas (i.e. campgrounds and highways). The DFPZs would not have the appearance of an abrupt edge, but rather blend into to the adjacent forest. The quarter mile width of the DFPZs may vary based on strategic importance, topography or other conditions. The completed DFPZ network would divide forested lands into many areas that would be less than 5000 acres, and a few that would be 20,000 to 30,000 acres, most areas would be between 5,000 to 20,000 acres. Approximately 12% of the public lands within the Pilot Project area would be comprised of DFPZs, if this project proceeds as modeled. The DFPZs would be established in approximately 13% of the forest types that are thought to have missed the most natural fire return intervals (ponderosa pine, eastside pine, mixed conifer).

DFPZ Description. Stands would be fairly open and dominated mostly by larger, fire tolerant trees. The openness of crown fuels creates a network of intermingled openings between the clumps of large trees, the absence of most small diameter trees and the low amount of surface fuel would produce a very low probability of sustained crown fire. DFPZs would be designed to blend into the adjacent forest, leaving lower canopy, down logs and snag levels adjacent to the primary control point (usually a road). While limited empirical information exists to evaluate what the threshold to limit crown fire spread might be, Weatherspoon and Skinner (1996) report canopy cover should not usually be more than 40%, although adjustments in stand density based on local conditions is appropriate. The density of tree crowns, measured by crown bulk density1 is one method that has been used to establish what the crown cover threshold might be, and this statistic could be calculated from typical stand exam data. Agee (1996) describes the relationship of crown bulk density and crown fire spread, along with limitations of these method of measurement.

Small group selections could be consistent with achieving the desired condition of the DFPZ, if they are placed on the fringe, away from the primary control point of the DFPZ and at a density that was consistent with achieving the criteria of maintaining 90% of the area in a condition that is not susceptible to torching. Weatherspoon (1996) reported that DFPZs will require periodic regeneration of portions of the zone and long-rotation, low-density versions of group selection might be the best silvicultural method for this purpose.

For this analysis, design criteria for DFPZ construction have been developed based on desired fire behavior characteristics of a completed DFPZ and other resource requirements (Table 1), in order to analyze the cumulative effects of the alternatives. Many site specific variables such as topographic position, exposure to predominate wind, stand structure and the amount and arrangement of surface fuels all influence fire behavior. Therefore, site specific prescriptions must be developed that will meet the fire behavior objectives of the DFPZ for weather conditions under which most large fires occur. This prescription may vary greatly depending on forest type and the other variables previously mentioned.



Footnote:
1 Crown bulk density is the mass to volume ratio of crown fuel biomass and crown volume.

Table 1. Alternative 2 DFPZ Design Criteria
Aerial Fuels
Surface Fuels
Eastside Pine Types 1. Overstory crowns are spaced at a distance that reduces potential for crown fire spread.2
2. Over 90% of the area, there is sufficient separation between the ground and crown fuels to prevent heat from surface fires from igniting the tree crowns.
3. Retain 3 of the largest snags/acre.
1. Up to 10 tons/ac of largest logs.
2. < 4' flame lengths or below the fire intensity threshold that would result in > 10% mortality in the stand.
3. Resistance to control reduced from current level.
4. Meet Regional Soil Quality Standards and Forest Plan requirements for small diameter fuel and ground cover.
Selected Timber Strata
(Preferentially selected for nesting by owls)
Strata types are; M4G, M4N, M5G, M5N, M6, P4G.
1. Overstory crowns are spaced at a distance that reduces potential for crown fire spread.
2. Retain 40% of the basal area, (largest of the healthy trees + culls).
3. If this does not yield canopy closure > or = 40%, make up the difference with trees 12-24 " dbh.
4. Over 90% of the area, there is sufficient separation between the ground and crown fuels to prevent heat from surface fires from igniting the tree crowns.
5. On an stand bases retain up to 8 of the largest snags/acre. (Grouping of snags on the fringe and lower densities along primary control feature (usually a road) to improve fire suppression safety and efficiency.)
Same as Eastside Pine types EXCEPT:
1. Up to 10 - 15 tons/ac of the largest logs.
Other Strata (Utilized, but not preferred by owls). P3G, P3N, P4P, M2G, M3P, M3N, M3G, M4P, R3P, R3G, R3N, R4G, R4N. Same as Selected EXCEPT:
1. Retain a minimum of 30% of the basal area.
2. At least 50 square feet of basal area per acre would be retained.
Same as Selected Timber Strata.
Forest Carnivore Habitat Network Eastside Zone:
1. Denning or resting habitat maintain 40% canopy closure.
2. Foraging or travel habitat maintain 25-40% canopy closure.
Westside Zone: 3
1. Denning or resting habitat maintain 60% canopy closure.
2. Foraging or travel habitat maintain 40% canopy closure.
Same as Select Timber Strata
SAT RHCAs Maintained at a level to enhance and protect riparian goals. Maintained at a level to enhance and protect riparian goals.
LSOG 4 and 5 Not entered. Not entered
ALSEs Maintained at a level to enhance and protect ALSE attributes. Maintained at a level to enhance and protect ALSE attributes.


Footnotes:
2 Agee (1996) provides information on how crown bulk density might be applied at the stand level to meet this objective.
3 Refer to Ecological Zone Map C in the Appendix. (Westside = Transition/Westside Central/Westsede Wet)


Alternative 3

Fuel Management Strategy. The strategy of this alternative relies less on DFPZs and more on mosaics of area fuel treatments to improve protection from large-scale, high-intensity wildfire. Approximately 30% of the treatment would be in the form of DFPZs and 70% in area fuel treatments. The goal of the area treatments would be to widen or replace many of the DFPZs in Alternative 2.

Fourty to sixty thousand acres of DFPZs and/or area fuel treatments would be constructed each year of the Pilot Project. The goal of this alternative is to reduce the number of acres burned and to reduce the damage done by wildfires on the acres that do burn. This would be accomplished by establishing DFPZs in areas that are most likely to be used as control points for fires that escape initial suppression action, and reducing the intensities at which fires burn by treating areas most at risk to large damaging fires with area fuel treatments. This strategy relies heavily on the assumptions that those forest types which have missed the most fire return intervals have an increased chance of stand-replacement fires (Brown 1985; van Wagtendonk 1985; Kilgore 1987; Arno in press) and that the forest types on southerly slopes have the highest fire frequency (Agee et al. 1990). There are several other assumptions and observations that guide this strategy as well. Within much of the project area, large fires often spread under the influence of the southwest winds. A small percent of the fires are likely to continue to escape control with initial suppression action. It is also assumed that the development of forest structures and landscape patterns comparable to those that developed under the more frequent fire regimes of the past will help to lessen the ecosystem disruption caused by severe fires that are beyond fire-suppression capabilities (Skinner and Chang 1996). The limited number of DFPZs constructed under this alternative are considered adequate to provide indirect control points. The assumption is also made that most fires that escape initial attack suppression action in terrain with steep slopes will continue to burn the most acres, with the highest intensity on the first day and controlled at the first major ridge. Therefore DFPZs in these locations will make for more efficient fire suppression action.

As with Alternative 2, this alternative is consistent with the goals of the fuel-management strategy outlined by Weatherspoon and Skinner (1996) in SNEP. In many cases it may be logical to implement an initial high-priority "low-density" DFPZ network along major ridges and main roads and in the vicinity of forest communities (SNEP; Volume II Chapter 56 p. 1482). The appropriate area would vary among landscapes depending on a number of factors including topography and aspect. They go on to say that subsequent efforts could be a combination of maintaining existing DFPZs, constructing new ones to break up the landscape into smaller blocks, and broadening existing DFPZs in conjunction with area fuel treatments as a longer term strategy.

Alternative 3 would also allow Fire Management Plans to be developed for the three National Forests that meets the goals of the Federal Wildland Fire Management Policy 1995. The main difference is the flexibility allowed in the placement of area fuel treatments may result in landscape patterns similar to what is thought to have occurred in low to moderate intensity fire regimes.

How the Location on DFPZs and Area Fuel Treatments were Developed. Like Alternative 2, the location of the DFPZs and area fuel treatments (Figure 2.3) for this alternative was developed for the purpose of analyzing cumulative effects. The DFPZ network mapped for Alternative 2 was used for the basis of the DFPZ segments that would remain in this alternative, based on a review by fire management personnel from each District and members of the HFQLG Interdisciplinary Team. The DFPZs that remain in this alternative are those believed to be the most beneficial in both the short and long-term, based on fire frequency and the potential for large fires. Some of the DFPZs in high risk and high hazard areas were removed from this alternative when it was thought that area treatment would better meet the goal of reducing area burned by high intensity fire. These evaluations were based on general knowledge of past fire occurrence and the existing hazard.

The locations of area fuel treatments (Figure 2.3) are based on a coarse scale analysis of past fire history in the Pilot Project area, in a search for areas most at risk to large damaging fires and within close proximity to existing road systems. They are generally located below 6500 feet elevation, on southeast to west facing slopes, in the eastside pine, ponderosa pine, mixed conifer pine and some mixed conifer-fir vegetation types.

Like Alternative 2, project level analysis would determine the exact locations of both DFPZs and area treatments. Project level considerations for DFPZ placement would be the same as the factors outlined in Alternative 2. The actual placement of area fuel treatments would be determined based on a finer scale analysis of areas with high fuel accumulations, dense small diameter trees and the spatial relationship of treatment areas to areas of low hazardous fuel conditions. The emphasis would be to locate area treatments in those areas mentioned above from the coarse scale analysis regardless of topographic location, although other forest types may also be treated.

What the Combination of Fuel Reduction Treatments Look Like Across the Landscape. Like Alternative 2, DFPZs and area treatments would be located near urban interface and other high use areas. The area treatments would vary in size from less than 50 acres to more than 1,000 acres, but would have a common design strategy. They would be arranged in a pattern that would reduce the intensity at which a fire is likely to spread. In broken terrain, they generally would be located on the upper portions of the southerly slopes, but may also extend to the base of a slope. DFPZs in this alternative are made up mostly of short segments and do not have the connectivity of DFPZs that would be developed under Alternative 2. The combination of DFPZs and area treatments in this alternative could result in approximately 12% of the public lands within the project area beginning to develop landscape patterns similar to what existed historically, and may treat 27% of the forest types that are thought to have missed natural fire return intervals.

How Area Fuel Treatments would be Prioritized for Implementation. In addition to ranking DFPZ segments as described in Alternative 2, watersheds would be ranked as to their susceptibility to large damaging fires and the values to be protected.

Specific elements used to rank watersheds for area treatments will be;
· Spatial relationship of watershed to urban interface, high use areas, or a major transportation corridor.
· Spatial relationship of watershed to private property, high value natural or cultural resources.
· Susceptiability of watershed to stand replacement fires.
· Fire frequency of the watershed.
· Risk of cumulative watershed effects high or very high.
· Areas where the natural fire return interval is thought to be frequent.

DFPZ and Area Fuel Treatment Description . In unsuitable owl habitat, DFPZs and area fuel treatments will be fairly open and dominated mostly by larger trees as they are in Alternative 2. Alternative 3 however does not allow as much open canopy in suitable owl nesting and foraging habitats. The goal in these habitat types is to retain a two layered canopy, remove small diameter trees and surface fuels that could initiate and permit crown fire spread. Like Alternative 2, project level evaluation will determine what the prescription will be in order to achieve that objective.Table 2. Alternative 3 and 4 DFPZ and Area Fuel Treatment Design Criteria

Table 2.  Alternative 3 and 4 DFPZ and Area Fuel Treatment Design Criteria
Alternative 3: Acres
DFPZs - 90,000-135,000
AFT -110,000 - 165,000
Alternative 4:
DFPZs - 64,5000
AFT - 60,500
Aerial Fuels Surface Fuels
Unsuitable Owl Habitat 1. Overstory crowns are spaced at a distance that reduces potential for crown fire spread.4
2. Over 90% of the area, there is sufficient separation between the ground and crown fuels to prevent the flames from surface fires from igniting tree crowns.
3. Retain 3 of the largest snags/acre.
1. Up to 10 tons/ac - largest logs.
2. DFPZs: < 4' flame lengths or below the fire intensity threshold that would result in >10% mortality in the residual stand.
3. Area Fuel Treatments : Flame lengths below the fire intensity threshold that would result in > 20% mortality in the residual stand.
4. Resistance to control reduced from current level.
5. Meet Regional Soil Quality Standards and Forest Land Management Plan requirements for small dia. fuel and ground cover.
Suitable Owl Foraging 5
Strata types are; P3G, P3N, M3N, M3G, R3G, R3N, R4G, R4N.
1. Retain 2 layered canopy.
2. Minimum 50% canopy cover.(As measured by densiometer) 
3. Minimum 12' from ground to base of live crown over at least 90% of area to prevent the flames from surface fires from igniting tree crowns.
4. On an stand basis retain up to 8 of the largest snags/acre. (Grouping of snags on the fringe and lower densities along primary control feature (usually a road) to improve fire suppression safety and efficiency.)
1. Up to 10-15 tons/ac -largest logs.
2. DFPZs: < 4' flame lengths or below the fire intensity threshold that would result in > 10% mortality in the residual stand.
3. Area Fuel Treatments : Flame lengths below the fire intensity threshold that would result in > 20% mortality in the residual stand.
4. Resistance to control reduced from current level.
5. Meet Regional Soil Quality Standards and Forest Land Management Plan requirements for small dia. fuel and ground cover.
Suitable Owl Nesting
Strata types are; M4G, M4N, M5G, M5N, M6, P4G.
Same as suitable owl foraging except maintain minimum 70% canopy cover.(As measured by densiometer)  Same as suitable owl foraging.
Forest Carnivore Habitat Network Eastside Zone:
1. Denning or resting habitat maintain 40% canopy closure.
2. Foraging or travel habitat maintain 25-40% canopy closure.
Westside Zone: 6
1. Denning or resting habitat maintain 60% canopy closure.
2. Foraging or travel habitat maintain 40% canopy closure.
Same as Select Timber Strata
SAT RHCAs Maintained at a level to enhance and protect riparian goals. Maintained at a level to enhance and protect riparian goals.
LSOG 4 and 5 Not entered. Not entered
ALSEs Not entered in Alternative 4.
Maintained at a level to enhance and protect ALSE attributes in Alt. 3.
Not entered in Alternative 4.
Maintained at a level to enhance and protect ALSE attributes in Alt. 3.


Footnotes:
4 Agee (1996) provides information on how crown bulk density might be applied at the stand level to meet this objective.
5 Complete description of design criteria for suitable foraging and nesting is located in Chapter 2
6 Refer to Ecological Zone Map C in the Appendix. (Westside = Transition/Westside Central/Westsede Wet)



Alternative 4

Fuel Management Strategy. The strategy of this alternative is similar to Alternative 3 although it accomplishes less fuel reduction treatments. Approximately 50% of the treatment would be in the form of DFPZs and 50% in area fuel treatments.

How the Location on DFPZs and Area Fuel Treatments were Developed. Like the other alternatives, DFPZs and area fuel treatments in this alternative (Figure 2.3) were developed to show possible general locations for the purpose of analyzing cumulative effects. The DFPZs and area fuel treatments mapped for Alternative 3 formed the base for this alternative. DFPZs and area fuel treatments located in ALSEs were removed, along with some of the area fuel treatments located in watersheds that were rated "low" on the Ignition Risk Map (Map H).

What the Combination of Fuel Reduction Treatments Look Like Across the Landscape. Like Alternative 2 and 3, DFPZs and area fuel treatment would still be located near the urban interface and high use areas. Area treatments would still form similar patterns as described in Alternative 3, however, they would only result in approximately 5% of the public lands within the project area beginning to develop landscape patterns similar to what existed historically, and may treat 11% of the forest types that are thought to have missed the most natural fire return intervals.

How DFPZs and Area Fuel Treatments would be Prioritized for Implementation. In addition to ranking DFPZ segments as described in Alternative 3, watersheds would be ranked according to their susceptibility to large, damaging fires and the values to be protected.

Specific elements used to rank watersheds for area treatments will be;
· Spatial relationship of watershed to urban interface, high use areas, or a major transportation corridor.
· Spatial relationship of watershed to private property, high value natural or cultural resources.
· Susceptiability of watershed to stand replacement fires.
· Fire frequency of the watershed.
· Risk of cumulative watershed effects high or very high.
· Areas where the natural fire return interval is thought to be frequent.

Area Fuel Treatment and DFPZ Description . This would be similar to Alternative 3, however approximately 97,000 fewer acres of area fuel treatments and approximately 6,000 fewer acres of DFPZ would be constructed and ALSEs would not be entered.

Alternative 5

Fuel Management Strategy. Similar to Alternatives 3 and 4, this strategy relies on mosaics of area fuel treatments to reduce the risk of moderate to high intensity wildfire and does not included DFPZs as a component of the strategy. Emphasis is placed on treating areas within and immediately adjacent to the urban interface and along major transportation routes.

Thirty to fourty thousand acres would be treated each year in the urban interface area, along major transportation routes and other areas where fire risk is high. This alternative proposes an expanded use of prescribed fire as a way of reintroducing fire as an important process in the ecosystem, as well as a way to reduce the excessive accumulation of fuels. The use of prescribed fire would be emphasized when it can be used to restore the ecological function of fire and achieve objectives or reducing potential for high intensity wildfire. When prescribed burning alone would not meet the objectives, thinning of the understory would be completed prior to prescribed burning.

The total acres that would be treated under this alternative is midway between Alternatives 3 and 4, approximately 200,000 acres during the pilot project. Also, a ecosystem analysis at the watershed scale would be required prior to recommending specific management practices needed to increase the fire resiliency of the forest and reduce the risk of large damaging fires.

Similar to the other action alternatives, Alternative 5 allows Fire Management Plans to be developed for the three National Forests to achieve the goals of the Federal Wildland Fire Management Policy 1995. The main difference is the increased emphasis on the use of prescribed fire.

This alternative is consistent with one of the goals of the fuel-management strategy outlined by Weatherspoon and Skinner (1996) in SNEP. Specifically this includes restoring the many functions of fire as an ecosystem process by using fire (SNEP; Vol. II Chapter 56 p. 1486). And, while alternative and supplemental methods must play a large part in needed restoration, they can not fully substitute and mimic the effects of fire.

How the Location of Area Fuel Treatments was Developed. Like the other Action Alternatives, the locations of area fuel treatments (Figure 2.5) was developed to show general locations for the purposes of analyzing cumulative effects. Like Alternatives 3 and 4, the locations of area treatments are based on the same coarse scale analysis for areas most at risk to large damaging fires as was used for Alternatives 3 and 4..

Like the other alternatives, site specific project analysis would determine the exact location of area treatments. Additionally this would be done after a watershed analysis was completed.

What the Area Fuel Treatments Would Look Like Across the Landscape. This would look very similar to Alternative 4. It is likely that the required watershed analysis would develop management practices similar to the area fuel treatments in Alternative 4 in the urban interface and areas of "matrix land" (lands without specific goals and objectives to maintain old forest values, aquatic goals or retain roadless characteristics). The rest of the forest (with the exception of roadless areas greater than 1000 acres) would have stand characteristics similar to those in Alternative 3 and 4, and higher canopy densities than Alternative 2. Overall area treatments in this alternative could result in approximately 7% of the public lands within the project area beginning to develop landscape patterns similar to what historically existed, and may treat 18% of the forest types that are thought to have missed the most natural fire return intervals.

How Area Fuel Treatments Would be Prioritized for Implementation. In addition to ranking DFPZ segments as described in Alternative 3, watersheds would be ranked according to their susceptibility to large, damaging fires and the values to be protected.

Specific elements used to rank watersheds for area treatments will be;
· Spatial relationship of watershed to urban interface, high use areas, or a major transportation corridor.
· Spatial relationship of watershed to private property, high value natural or cultural resources.
· Susceptiability of watershed to stand replacement fires.
· Fire frequency of the watershed.
· Risk of cumulative watershed effects high or very high.
· Areas where the natural fire return interval is thought to be frequent.

Area Fuel Treatment Description . In the urban interface and matrix lands area, treatments are expected to be fairly open and similar to unsuitable owl habitat described in Alternative 3. Other habitat areas or aquatic emphasis areas will generally have denser overstory canopy.Table 3. Alternative 5 Area Treatment Design Criteria

Table 3.  Alternative 5 Area Treatment Design Criteria
Acres of Area Fuel Treatments: 150,000 - 200,000 Aerial Fuels Surface Fuels
Matrix Lands.
Urban Interface.
1. Overstory crowns are spaced at a distance that reduces potential for crown fire spread.
2. Over 90% of the area, there is sufficient separation between the ground and crown fuels to prevent the flames from surface fires from igniting tree crowns.
4. On an stand basis retain up to 8 of the largest snags/acre. (Grouping of snags in areas not likely to be used as control feature (usually a road) to improve fire suppression safety and efficiency.)
1. Up to 10 tons/ac of the largest logs.
2. Urban Interface : < 4' flame lengths or below the fire intensity threshold that would result in > 10% mortality in the residual stand.
3. Matrix Lands : Fame lengths below the fire intensity threshold that would result in > 20% mortality in the residual stand.
4. Resistance to control reduced from current level.
5. Meet Regional Soil Quality Standards and Forest Land Management Plan requirements for small diameter fuel and ground cover.
Spotted Owl Home Range 1. Maintain at least 50% of each owl home range in habitat that is suitable for foraging and nesting. 
2. Over 90% of the area, there is sufficient separation between the ground and crown fuels to prevent the flames from surface fires from igniting tree crowns.
3. No reduction in overstory canopy closure in areas that do not meet suitable foraging or nesting.
4. No removal of dominant and co-dominate trees.
5. No reduction in the largest snags.
1. Up to 10 - 15 tons/ac of the largest logs.
2. Flame lengths below the fire intensity threshold that would result in > 20% mortality in the residual stand.
3. Resistance to control reduced from current level.
4. Meet Regional Soil Quality Standards and Forest Land Management Plan requirements for small diameter fuel and ground cover.
Forest Carnivore linkage between LSOG reserve network. 1. Canopy cover in the dominant and co-dominate trees shall not be reduced below 60%.
2. No reduction in canopy in areas with < 60% canopy closure.
Same as Spotted Owl home range.
Norther Goshawks, Great Gray Owl and Forest Carnivore PACs Hand removal of small diameter material unless a site-specific biological evaluation determines that other practices would better meet the immediate and long term needs. Maintained at a level to enhance and protect PAC attributes.
LSOG 1,2 & 3 Retain all dominant and co-dominate trees. Maintained at a level to enhance and protect LSOG attributes.
LSOG 4 and 5 Hand removal of small diameter material. Maintained at a level to enhance and protect LSOG attributes.
ALSEs Hand removal of small diameter material. Maintained at a level to enhance and protect ALSE attributes
SNEP Land Use Buffer Thin from below  Maintained at a level to enhance and protect aquatic goals.
SNEP community Energy Buffer. Timber removal only to protect human health and safety. Maintained at a level to enhance and protect aquatic goals. Use of prescribed fire minimized.
Roadless Areas Prescribed fire use only Prescribed fire use only

Section II

Some Effects of Fuel Treatment Patterns on Fire Growth:
A Simulation Analysis of Alternative Spatial Arrangements

Synopsis

Fire simulations completed through computer models were performed on artificial and actual landscapes to examine some effects of spatial fuel treatment patterns on fire growth and behavior. Artificial landscapes were used to isolate, as much as possible, relationships between spatial patterns and fire growth. Actual landscapes were then used to examine how fire might be affected by spatial fuel patterns when implemented under more realistic constraints and complex environmental conditions.

Artificial landscapes were used to contrast three patterns of fuel treatments: network, random, and overlap. The network pattern contained the largest contiguous areas of untreated fuels and the overlap pattern had the smallest. When fires of different sizes were simulated on each landscape, the network pattern had less affect on small fires than the other patterns. The overlap pattern provided the most involvement of treated areas with fires at all sizes or stages of growth. Spotting resulted in larger fires under all treatment patterns. When fires were allowed to become "large'' during longer simulations, the different treatment patterns produced only minor differences in fire sizes, but substantial differences in fire behavior within the burned area. The amount of area burned by the fastest rate of spread was reduced by the overlap pattern (with and without spotting) but not by the network or random patterns. This analysis supports the following observations.

· When fires are small, the network pattern is likely to have less of an effect on fire behavior or rate of fire growth and provide less benefit to suppression than the other patterns tested.

· When fires are large, the overlap pattern contributes to a reduction in the area burned by the fastest and most intense portion of the fire, thereby reducing fire severity and facilitating fire suppression efforts.

Two actual landscapes were selected to represent fuel and topographic conditions from the Plumas (near Quincy) and Lassen National Forests. Treatment patterns addressed three alternatives: no change from current practices (current conditions), linear treatments only, and area-based treatments. The second and third patterns are similar in concept to the network and overlap patterns on the artificial landscapes. Weather and fuel moisture inputs to the simulations were based on local weather stations for the 90th percentile of all fire weather recorded between July and September. The results of these simulations suggested:

· Treatments affected fire growth in a manner consistent with the trends indicated by the simulations on artificial landscapes. The computer simulations showed that the Plumas site (approximately 4000-6000 acres) demonstrated a more pronounced difference in the treatments compared to the Lassen site. This occurred because the Lassen site (approximately 5000-7000 acres) contained a large percentage of land that had been previously treated therefore the additional treatments had relatively minor impact on fire behavior.

· Little change in the distribution of fire behavior within the burned area was evident in these simulations. The likely reasons were 1) the fires were small compared to the treatments, meaning not enough treated area was burned to affect the fire behavior distributions, and 2) the environmental heterogeneity masked the effects of treatments on fire behavior. Fires were determined to be of reasonable sizes compared with historical fires, but still involved relatively small areas of treatment in the simulation.

Conclusions

This simulations suggested that the choice of spatial treatment pattern has potentially meaningful consequences to both fire size and behavior. The analysis also suggested that the greatest benefits (e.g. reduced fire size and severity) occurs when spacing or gap between treatments was small compared to the expected fire sizes. Small gaps between the treated areas tends to restrict fires at small stages of growth and requires frequent flanking of large fires to circumvent treated zones. Flanking fires spread more slowly and with lower intensity. Smaller untreated gaps between the treated areas can be largely independent of the amount of treatment when arranged on an area-based approach. A network approach, however, is constrained by the total treatment area because of the fixed width of the treatments. The simulations suggested that the network treatments provide benefits in the absence of suppression action. The implementation of these concepts will be subject to practical constraints and variability in environmental conditions and will likely produce varied results as indicated by fire simulations on actual landscapes.

Some Effects of Fuel Treatment Patterns on Fire Growth:
A Simulation Analysis of Alternative Spatial Arrangements

Introduction

Modifications to forest fuel complexes, as described by Van Wagtendonk (1996), Weatherspoon and Skinner (1996), and Agee et al. (in press), are designed to reduce the rate of spread and intensity of wildfires burning within the treated areas. They can also facilitate wildfire suppression by increasing the rate at which fireline can be constructed including removing the woody material and brush and reducing the amount of fireline needed. In other words the fire's rate of spread and intensity at which the fire burns are reduced and crews are building less fireline around a smaller fire (Finney et al. in press). The prescriptions for fuel treatment consist mainly of thinning the smaller trees from the stand to reduce the density of crown fuels and to sever the vertical continuity between the tree crowns and the ground surface. For reducing surface fire spread rate and intensity, prescribed fire is the most effective method of removing slash and pre-existing surface fuels (Van Wagtendonk 1996). It is generally accepted that fuel treatments that follow these guidelines would reduce local wildfire behavior and the severity of effects because only surface fires are likely to burn within the treated areas, leaving the crowns and the residual trees largely undamaged.

The influence that these treatments have in affecting the size of a wildfire and the consequences within the burned area are, however, not so easily described or understood. Fire growth and behavior across a landscape is not necessarily dependent on the conditions found within any given treatment unit. They are inherently spatial and dynamic, requiring the use of simulation techniques and observations as a means to determine how fires move in space and time relative to fuel patches.

Unfortunately little scientific knowledge exists, and few techniques are available, for studying the behavior of fires as they grow across heterogeneous landscapes. The simulation work to date has represented fire growth in the most general way possible, as a statistical or mathematical process, with little basis in fire behavior (Turner et al. 1989, Baker et al. 1991, Mladenoff et al. 1996, Gardner 1996, Bak et al. 1990, Malamud et al. 1998, McCarthy and Gill 1997). Without incorporating fire behavior as integral part of the modeling, these techniques could only address the most abstract properties of fires (i.e. sizes) and provide no information on behavior within the burned area or a physical explanation for their results. The studies essentially suggest only that fire sizes should be sensitive to spatial variation in fuels and confirm that small fires should be more frequent than large fires. Observations of burn patterns provide more concrete support for the idea that previously burned areas can limit fire sizes (Van Wagtendonk 1995, Parsons and Van Wagtendonk. 1996, Minnich and Chou 1997) and modify fire behavior (Van Wagtendonk 1996, Agee et al. in press).

For simulation experiments, if fire behavior is treated with as much detail as afforded by the current state of knowledge, the results can help provide a physical explanation for the following issues:

1. How and why fire behavior is affected by fuel changes within treatments,
2. How and why spatial treatment patterns affect fire growth and behavior,
3. How simulation results translate into practical guidelines and standards,
4. How fire suppression can interact with fire and fuel patterns in space and time.

This report describes the results of two simulation analyses of spatial treatment alternatives. The first analysis uses artificial landscapes. The artificial landscapes contained stylized treatment patterns that were generated by a computer program to control the variability in the fuel patterns. This helps to limit the variables that affect the response of fire growth and behavior. The spatial arrangements were selected to represent the basic characteristics of the range of treatment alternatives proposed or available for the study area (Plumas and Lassen National Forests). The second analysis was conducted on two actual landscapes, one from the Plumas, near Quincy, and the other from the Lassen. Actual landscapes were used to illustrate some likely consequences of the treatment alternatives with realistic treatment patterns and natural variability in fuels, weather, and topography.

Methods

The FARSITE simulation model (Finney 1998) was used to simulate fire growth and behavior for the artificial and actual landscapes. The simplifying assumptions used in this analysis were:

1. All analyses are conducted at a single point in time representing the landscape structure produced at the end of the treatment period (i.e. five years).
2. All treatment areas are represented as having homogeneous fuel characteristics,
3. No suppression activities are simulated or assumed in the modeling. Although technically feasible (demonstrated by Finney et al. in press), it was not practical for this analysis because of the many assumptions required concerning resource availability and tactics, and the emphasis of this study on the impact of treatment pattern on growth of larger fires. During the growth of large fires, suppression efforts have little affect on the forward rate of spread of fire. This forward spread is typically responsible for most of the severe affects to vegetation as well as being largely responsible for the amount of area burned by the fire.
4. Fires were simulated under fire weather conditions that historically contribute to the largest fires (approximately the 90th percentile). These fires have the most adverse effects and pose the greatest threats to safety and property. It was not the intent to simulate fire behavior under the worst fire weather conditions.
5. Fuel and vegetation maps provided for the actual landscapes at Plumas and Lassen sites were assumed to be correct and represent as accurately as possible the stated conditions.

Part I. Analysis of Fire Growth and Behavior on Artificial Landscapes

Three spatial arrangements were selected to represent the range of ideas for treating fuels across large landscapes. These scenarios were synthesized from the alternatives in order to illustrate their fundamental geometric implications for fire growth. Terrain was always assumed flat. Treatments were assumed to have reduced the spread rate to 1/10th that in the untreated fuels. All patterns were designed to treat approximately 15% of the available landscape area. This fraction was obtained from the cumulative proposed treatment area on the Plumas and Lassen National Forests of approximately 250,000 acres out of a total of around1,500,000 acres that are considered suitable for treatment. The three patterns selected for analysis were:

1. A regular network of linear treatments with a fixed width. The arrangement was represented by a hexagonal network with a 1/4 mile width (400m) (Figure 1). This produces an interior area of approximately 4000 acres, with sides 2700 meters (m) on center (a=2.59808s 2).

2. A random arrangement of linear treatments with random orientation, length and width. This represents a situation where treatments are located without consideration to their size, juxtaposition, or orientation.

3. A regular parallel pattern of linear treatments with calculated widths, gaps between treatment areas, and overlaps of treatement areas. This pattern was chosen to contrast the random arrangement that occurs in example 2. This pattern is organized to incorporate the influences of fire shape and spread rate on treatment size and separation from its neighbors. Each treatment modeled was ? mile long (400m) with the width determined by formulas relating the fire size to the relative spread rate in the treatments.

The analysis is believed to largely independent of the dimensions of the treatment patterns or the simulated fires, as long as the relative sizes of the fires and the treatments are similar. The analysis therefore, should reflect the general way that each pattern influences fire growth and behavior.

For each landscape, two scenarios were selected to address behavior of short-term/small fires and long term/large fires:

? The short-term fire scenario was used to examine the potential influence of the treatments on fires as they grow in size. As small fires increase in size, their area and forward extent is more likely to come in direct contact with the treated areas. This is relevant to interpretations of the direct influence of treatments on fire behavior and indirect influence on suppression efforts used to contain those fires.

? The long-term fire scenario was used to assess the characteristics of larger fires that escape initial attack. Fires tend to burn longer and become large during periods of extreme weather and when ignition "episodes'' deplete local resources for initial attack. This analysis was concerned with questions about how the fire interacts with the treated area in terms of its size and behavior after a fixed duration of nearly constant burning conditions (24 hours). It is relevant to understanding fire effects and to interpreting the potential use of the treated areas by suppression forces. The size of the fire after the fixed duration was dependent on the individual fuel patterns, and to some degree on the ignition locations.

Each artificial landscape was 10 kilometers (km) in length (6.25 miles) with a 10 m resolution for a total of 1000 x 1000 cells. This size was considered large enough to illustrate the representative influences of the different fuel treatment patterns at the specified level of coverage (approximately15%). Fire ignitions were considered as points on the landscape. For the short-term simulations, the fires were selected to have a length to breadth ratio of about 2:3 so that there would be directional differences in fire growth which differentiated the fire head, flank, and backing spread rates consistent with wind-driven fires. Fires were ``allowed'' or modelled to burn for 2 hours, although the exact shape and time-span would have little influence on the results of these simulations. Ignitions were located systematically on a square spacing every 1000 m within each landscape.

For the long-term simulations two complete sets of simulations were performed. One included the potential for spot burns and one similation did not include spotting. The fuel conditions used for each situation were simplified to eliminate ``noise'' or non-essential variables in the comparisons.
 

Condition
Untreated
Treated
Surface Fuels
10
(reduced to 1/10 spread rate
and intensity)
Canopy Cover
50%
30%
Crown Base Height
2m
4m
Stand Height
15m
15m
Crown Bulk Density
0.25 kg m-3
0.15 kg m-3

Weather conditions were assumed to be extreme. Winds were constant at 25 mph (at 20 ft) from a single direction (south) with fuel moistures at 4, 5, 8% for 1 hour, 10 hours, and 100 hours respectively.

For each set of simulations, two ignition points were used. These were subjectively located with relation to the treatment corridors on the network landscape. This was done to reflect the consequences of fire growth beginning on the leeward as well as the windward sides of the treatment area. These same two ignition points were used for modeling both the random and overlapping arrangement. For each simulation, a raster map of the fire spread rate (10 m resolution) was generated so that frequency distributions of area by spread rate could be calculated.

Results

Fire Progression and Influence of Treatments

Having the same treatment fraction, each landscape had the same probability of ignitions occurring within a treated area, either at random or in this case at a systematic spacing. For the small fire scenario, ignitions occurred with the same frequency in treated areas on each landscape (13/81 = 16%), and is essentially a sample estimate of the treatment fraction. The simulations showed, however, that wider spacing between treatments decreased the probability that a fire growing from a random ignition would be influenced directly by a treatment unit (Figure 1). The probability that the fire will spread into a treated area was inversely related to the size of the untreated area between treatments. The network approach was intersected by fewer fires when they were small (up to about 800 acres, or about 3000 meters long) compared to the other treatments (Figure 2 and Figure 3). As fire size increased the fire affected more of the network pattern. The overlap pattern had roughly the same affect on fires of all sizes because of the relatively small size of the gap between the treated areas. The random arrangement was generally somewhere between the network and the regular dispersed pattern in terms of fire contact.

The graphic depictions of the simulations demonstrated how the large fires were affected in different ways by the three treatment patterns (Figure 4 and Figure 5). Without spotting, the two fires simulated on the network arrangement were of very different sizes. The ignition on the ``lee'' side of the corridor was not strongly influenced by the network treatment and was the largest fire observed in any simulation. The ignition located immediately on the windward side was the smallest observed because most of the fire's spread was within the wide treatment zone. Even when spotting was considered the differences in the sizes of the fires was minimal. In this case fires were of roughly similar sizes because embers flew over the 400 m fuel treatment zones, starting new fires on the lee side again burning without being greatly affected by the treatments.

Without spotting, the fires in the random treatment essentially burned around the periphery of the isolated treatment units, locally restricting the forward spread of the fire throughout the unit. The large sizes of some of the random units seemed to be effective in trapping embers and reducing the forward advance of fires burning when spotting occurred.

The overlap pattern was effective in limiting the forward growth of the fire regardless of whether spotting occurred. With spotting, embers easily jumped the narrow treatment areas and started spot fires. However these fires were then limited in the forward direction by the numerous treatment units.

Fire Sizes

The average fire size of the two ignitions (with and without spotting) simulated for each landscape did not vary appreciably between treatment methods. The numerical differences are even less meaningful than are indicated by the numbers generated from the model because the areas were calculated only for the burned area, and did not include the islands remaining unburned within the outermost perimeter or the areas within concave edges. These areas are not typically deducted when the total area of a fire is calculated therefore the actual amount of land affected by a fire can is often less than the total acres reported. The addition of these unaffected area increase the overall fire size regardless of spotting. This would mainly affect fires that occur in the random and network patterns which have the most enclaves and concavities along their edges.

All treatments had the effect of reducing fire sizes with and without spotting enabled. The network and overlap patterns produced the smallest average fire sizes without spotting, but the random arrangement had the smallest combined size when spotting was enabled. The network pattern produced the greatest differences between the two simulated fires without spotting; there was little difference between the two fires when spotting was enabled.

Average Fire Area by Fire Type and Fuel Arrangement (acres: n=2)
Fire Type
Random
Network 
Overlap
Control
Spotting
1748
2499
2606
2936
No Spotting
1467
1054
1060
1774

Relative Fire Area of Fire Type and Fuel Arrangement
Fire Type
Random
Network 
Overlap
Control
Spotting
0.60
0.85
0.89
1.0
No Spotting
0.82
0.59
0.59
1.0

Fire spread rate distributions

The distributions of spread rate suggested greater differences due to treatment pattern than evidenced by the similar fire sizes (Figure 6 and Figure 7). The overlapped pattern produced a uniquely different distribution than fires burning on the random, network, or control landscapes. This unique distribution was characterized by a shift in burned area from the highest spread rates (in the heading direction) to the lowest spread rates (backing and flanking). This occurred with and without spotting and was evident in distributions of both absolute and relative frequencies. The network and random arrangements produced spread rate distributions similar to the control for simulations with and without spotting.

Without spotting the highest spread rates accounted for the most burned area because the heading spread was unconstrained by treatments and is responsible for a large amount of area burned by an elliptical fire (Catchpole et al. 1992). When spotting was enabled, all simulations suggested that the mode of the spread rate distribution (most common frequency) shifted back to a lower spread rate at about two-thirds of the range. This occurred because each spot fire ahead of the main front has a distribution of spread rates around it (including backing and flanking) that burn areas at a lower intensity than the faster spread of the heading portion of the main fire.

Part II. Analysis of Fire Growth and Behavior on Actual Landscapes

FARSITE simulations were performed on actual landscapes from the Plumas and Lassen National Forests. This afforded an opportunity to examine effects of treatments implemented under more complex environmental conditions and provide an analysis that is subject to such practical constraints as treatment location and unit size. The alternatives consisted of:

1. Alternative 1: Current Conditions. The fuel maps represented the current status of surface and crown fuels with no additional treatments.
2. Alternative 2: Linear Fuel Breaks. The fuel maps were changed to represent the conditions of actual proposed treatment areas in a linear fashion.
3. Alternative 3: Area-based Treatments along with linear fuel breaks. The fuel maps reflected the same area treated as in Alternative 2, with the length of linear fuel breaks reduced by about half. The remaining acreage was placed as area-based treatments that were arranged in overlapping patterns oriented perpendicular to the spread direction of most concern (SW-NE axis).

All treatments were located by Forest Service personnel in consideration of existing road access and topographic and statutory constraints on harvesting (i.e. CASPO). The treatments were modeled as:
 

Plumas
Lassen
FM Number
Custom #49
FB Standard #9
1hr (tons[t]/ac)
0.65
2.92
10hr (t/ac)
1.96
0.41
100hr (t/ac)
2.39
0.15
SAV 1 (1/ft)
1500
2000
Depth (ft)
0.47
0.2
XMx (%)
14
25
Crown Base Height (ft)
7
15

Fuel model (FM) 49 represents a more compact and lighter version of the standard slash FM 11 which describes a condition that could occur when prescribed burning has not been conducted after thinning or where fuels such as shrubs and small trees have increased over a period of 5-10 years after burning and thinning. The spread rate and intensity of FM 49 was about 1/3 that of FM 9 under the same environmental conditions. Fuel model 9 represents a condition that could occur when east-side ponderosa pine has been thinned and burned, temporarily lessening the influence of shrubs (with needle drape) on fire behavior. Fuel model 9 is about 3-4 times slower than a FM 2, which is typically used to represent the untreated areas on the east side that burn with much less intensity than normal. Crown base heights were raised moderately for each fuel treatment to reflect the silvicultural prescription of thinning from below.

The simulations reflected a fire which burned for approximately 3 days. This was done to capture the fire behavior during the peak burning period of each day. Weather conditions chosen for the simulations were based on the percentile summaries of records from local NFDRS stations and trends from a sample of the Pierce Creek RAWS station for the 1996 fire season. In general, the weather consisted of maximum temperatures of 85 degrees with a minimum humidity of 20%. Winds on the first day consisted of light winds from the southwest with down-canyon flow at night. Winds on the second day depicted a passing cold front; stronger southwest winds predominate for 1 burning period. Maximum 20-ft winds of 18 mph were used for the Plumas site and 25 mph for the Lassen. These were based initially on the 10-minute average winds for the 90thpercentile burning conditions on the Boulder Creek station (Quincy) and the Bogard station (Lassen). The windspeeds were increased to reflect the effect of gusts that commonly exceed the average by 2-5 times and have a large effect of fire behavior (see Figure 8). The initial fuel moisture conditions were also selected from the 90 th percentile fire weather from the respective NFDRS stations. The ignition points were located within the untreated fuels, such that fire would burn into treated areas. As with the artificial landscapes, the simulations were intended to suggest trends and permit a graphical interpretation of treatment effects on fire growth. The limited number simulations obviously do not cover the range of possibilities that could occur in wildland fire situations.

Results

Fire Progression and Influence of Treatments

The simulations for the Plumas and Lassen sites (Figure 9 and Figure 10) suggested how fire might be expected to respond to the treatments compared to the current conditions. Fuels at the Plumas site were dominated by patches of brush (FM 5) and timber (FM 9 & 19), with moderate overstory (10-50%) canopy cover. The Lassen site was composed of stands that had already treated (thinned and underburned, FM 9) and fuels yet to be treated (FM 2). Canopy cover was uniformly less than the 50% common for pine communities on flat lands.

At the Plumas site with current conditions (Alternative 1), the fire spread quickly up-slope under the influence of moderate to strong southwest winds (Figure 9a). Fire behavior would be characterized by torching trees that produced medium-range spotting (1/10-1/4 mile). The head of the fire entered a change in fuel type about 2/3 of the way up the ridge that caused the fire to slow considerably, resulting in a reduction of torching and spotting. Fingers of the fire originated from the flanks of the main fire because of local topography, fuels, and diurnal changes in winds. By comparison, fuel treatments implemented under Alternative 2 (FM 49) slowed the fire, increasing the time the fire took to move (with the wind) onto the steeper part of the slope (Figure 9b). After breaching the treatment zone, the fire resumed growth on the second day when the stronger southwest winds arrived and the fire then behaved similarly to the fire modelled in Alternative 1 (torching and spotting without the influence of the treatments).

Treatments implemented under Alternative 3 performed much like Alternative 2 but afforded repeated interruption of the fire as it traveled up slope because multiple treatment units were present to impede fire growth (Figure 9c). These treatments limited fire growth (forward spread with the wind and slope) and at the back and flanks where treatment units were also encountered fire. Fire behavior between treatments was similar to Alternative 1, characterized by torching and spotting.

The fires simulated on the Lassen site showed fewer differences resulting from treatments than at Plumas. Fire growth under current conditions suggested that the extensive treatments already in place were the primary cause of limited fire growth and behavior even under strong southwest winds (Figure 10a). Fire behavior was considerably milder in FM 9 than in the untreated areas with a FM 2. Little torching or spotting was calculated because of the increased distance from the ground to the crown base (lowest branches in the trees). Even with southwest winds, the fire progressed most rapidly to the south, having no treatment units to interrupt the growth of the fire perimeter. The combination of reduced spread rate in the treatments at the head of the fire, and faster rates of spread in untreated fuels at the flanks, caused the fire shape to be nearly round.

The fire simulated under Alternative 2 showed that a single fuel treatment zone delayed fire growth from spreading to the south, thereby considerably reducing the ultimate size of the fire in the southward direction (Figure 10b). Elsewhere, the addition of the linear fuel treatments had little effect, owing to the existence of an already vast area in a treated condition. Fuel treatments implemented under Alternative 3 (Figure 10c) had no effect on fire growth in this simulation, and resulted in fire growth nearly identical to that which occurred in Alternative 1. This occurred because all treatments were concentrated in the northeast portion of the landscape which were not reached by the fire. As with Alternative 2, the extensive treatment coverage existing under current conditions was more effective in limiting fire growth and behavior than the additional treatments applied under the alternatives.

Fire Sizes

The simulations for the Plumas site showed that, after 3 burning periods, fires were remained smaller under Alternative 3 than Alternative 2. The fire size in the Lassen simulations was smallest under Alternative 2 because the fire did not encounter the lands treated under Alternative 3.

Fire area (acres) by treatment alternative.
Plumas
Lassen
Alternative 1
6545
7639
Alternative 2
4901
5283
Alternative 3
4171
7699

Flame Length Distributions

Flame length was selected as the parameter for comparing fire behavior distributions for the simulations on actual landscapes because it integrates both the potential rate of spread and fuel characteristics (moisture and type). Despite the fact that the treatments did produce moderate fire behavior locally within the treatment, little change in the shape of the flame length distributions was found for the fires simulated on actual landscapes at either the Plumas or Lassen sites. This contrasts with the findings of the artificial landscapes. The most likely explanations for this difference are a combination of two factors:

? Only a small fraction of treated area (absolute acres and relative to fire size) was burned by the fires simulated for the Plumas and Lassen sites compared to those used on the artificial landscapes. Larger fires or the application of treatments in smaller units would probably have produced a greater distinction in fire behavior distributions in these simulations.

? The heterogeneous weather, fuel, and topographic conditions produced a greater magnitude of change to the distribution of fire behavior than did fuel treatment. This would happen when changes in weather conditions (reduced winds and higher nighttime humidity) and variation in topographic aspect effectively change fire behavior to the same or have an even greater affect than the changes in fuel type from the treatments. The spatial and temporal heterogeneity of the fire environment probably had a stronger effect on the fire behavior distributions than the fuels in these simulations.

A minor distinction between treatments at the Plumas sites showed that the upper limit of flame lengths were smallest under Alternative 3 (0-15 ft) when compared to the other alternatives (0-24 ft). Flame lengths greater than 15 feet, however, only accounted for about 1% of the burned area, probably in areas not burned by the fire under Alternative 3.

Discussion

The simulations on artificial landscapes identified some trends that have implications for the behavior, effects, and suppression of real fires. First, larger blocks of contiguous fuels between treatments allows ignitions to result in larger fires before the treatment units are encountered by the fire. This was most evident in the network pattern, but was also noticed in the random pattern. For suppression, this could mean that direct and parallel fireline construction tactics that are most often used for initial attack on small fires be of little benefit in areas treated with network or random patterns. The models further imply that fires that start in areas that have not been treated would be larger when contained and hence, a greater proportion of fires might escape initial attack compared to fires starting among the other treatment patterns.

Second, larger blocks of untreated fuel creates a greater potential for an uninterrupted forward fire spread, which produces the fastest spread rates, highest intensities, and least opportunity for suppression. The distributions of spread rate on the artificial landscape modeling suggested that forward spread is moderated most by an overlapping pattern and is not greatly affected by the network or the random patterns. This implies that the network pattern and random pattern would do little to change the effects that occur from large fires in terms of fire severity. If spotting is significant, fires can become large regardless of the treatment pattern used. However, the burned area inside the fire would still be modified by the overlapping pattern that interrupts the heading growth of the small spot fires.

The two "large'' fires simulated for each artificial landscape were used to suggest trends, not to produce statistically valid estimates of means or variances in fire sizes or fire behavior. The analysis is intended to highlight the relative differences and provide an explanation for the observed results. This was necessary due to the stylized landscapes used and the fact that they were not intended to be "real'' or to represent an exact landscape structure. Also, the weather and fuel moisture conditions were chosen to provide a reasonable standard to facilitate comparisons of the various simulation results and make them relevant. The specific results of the simulations in terms of fire size and spread distributions would change with different weather conditions, assumptions about fuel structure, and ignition locations. The simulations do however, establish some tendencies that can be used to evaluate the merits of the different treatment patterns.

One of the concerns raised over the linear network approach to fuel breaks (Van Wagtendonk 1996) was that spotting would breach the 0.25 mile break, rendering them ineffective in stopping extreme fires. This occurred in all treatments tested here, on both the artificial and actual landscapes. Van Wagtendonk (1996) however, did not find that spotting over 390m fuel breaks. This might be explained by the use of the constant 18 mph windspeed which does not allow for gusts that typically more than double the average wind speed and has a profound affect on fire behavior characteristics such as spotting (Figure 8).

The differences in results between the simulations on actual landscapes and those on the artificial landscapes illustrates how the effects of a complex real fire environment, with variable fuels, weather, and topography, can mask the results of fuel treatments implemented under the proposed alternatives. Fuel treatments will likely affect wildland fires depending on the weather and suppression actions. Bunnell (1998) reported that the spatial pattern of clearcuts amidst old-growth lodgepole pine forests were useful in suppressing the Little Wolf Fire in 1994 in western Montana. The clearcuts actually produced faster rates of spread than within the conifer forest, but the fire was of lower intensity (grass fire not a crown fire) and was easily suppressed. Spotting caused the fire to become large despite the suppression efforts and clearcuts, similar to the simulation results found in this analysis.

The locations of natural limitations on fire growth and behavior should be considered in the placement of treatment units. For example, treatments would be more effective if located to use natural barriers, such as rock outcroppings, to increase the size of barriers, augmenting the ability to control wildland fire on the landscape level. Similarly, treatment of fuels that already exhibit moderate fire behavior, like those that dominate at higher elevation, would be less likely to produce benefits to fire behavior on a landscape basis than to treatments of more hazardous types.Figure Captions

Figure 1. Artificial landscapes used to simulate growth of small and large fires. Ignitions at a regular spacing were allowed to burn for various time periods (2-hrs shown). Treated fuels are shown in blue and untreated in yellow. Fires contacted fewer treatments with the network pattern (B) or random pattern (C) than the overlap pattern (D).

Figure 2. The frequency of contact by fires with treatment units increased as fires got bigger. The network pattern produced the fewest contacts for a given size than the random or the overlap pattern. The overlap pattern produced a consistently high frequency of contact with fires of all sizes due to the smaller spacing between treatments.

Figure 3. The frequency of contact by fires with treatment units also increased as fires spread farther. The network pattern produced the fewest contacts for a given length than the random or the overlap pattern. The overlap pattern produced a consistently high frequency of contact with fires of all lengths due to the smaller spacing between treatments.

Figure 4. Simulations of fires for 24 hours under constant weather conditions on artificial landscapes showing patterns of spread rate within the burned area. Surface fire spread only without spotting produced different fire sizes depending on the ignition location relative to the treatment units.

Figure 5. Simulations of fires for 24 hours under constant weather conditions on artificial landscapes showing patterns of spread rate within the burned area. Fires were of similar size on all landscapes because of spotting.

Figure 6. Absolute frequency distributions of spread rate for the simulations on artificial landscapes. All treatments reduced fire sizes. Only the overlap pattern changed the shape of the distribution with and without spotting by shifting most of the burned area into slower spread rates.

Figure 7. Relative frequency distributions (normalized by area of each fire) show the same trends as the absolute frequency distributions. Only the overlap pattern changed the shape of the distribution with and without spotting by shifting most of the burned area into slower spread rates.

Figure 8. Relationship between maximum gust wind speed (mph) and 10-minute average for the Pierce Creek RAWS station for July-September 1996. Peak gusts exceeded the 10-minute average by roughly 2-5 times.

Figure 9. Fire progression maps for the fire simulations on the Plumas site for Alternative 1 (a), Alternative 2 (b) and Alternative 3 (c). Treatments are shown in yellow.

Figure 10. Fire progression maps for the fire simulations on the Lassen site for Alternative 1 (a), Alternative 2 (b) and Alternative 3 (c). Treatments are shown in yellow.

Figure 11. Flame length frequency distributions for the Plumas fire simulations reflect no changes in the shape of the distribution attributable to the treatments.

Figure 12. Flame length frequency distributions for the Lassen fire simulations reflect no changes in the shape of the distribution attributable to the treatments.

Section III

Fire Behavior Predictions and Modeling

Description of Components used to Analyze Wildfire Susceptibility

Wildfire susceptibility is the combination of the probability of a fire igniting (Risk) along with the intensity at which it will burn (Hazard).

Hazard -

For this analysis the FlamMap computer program was used to display relative surface fuel hazard within the planning area. FlamMap is a computer program that makes fire behavior calculations across a landscape using GIS data inputs for terrain and fuels. The purpose of FlamMap is to generate fire behavior data that are comparable across the landscape for a given set of weather and/or fuel moisture data inputs.

Fire behaviors modeled by FlamMap are surface fire, crown fire, and fuel moisture. The fire behavior models and their integration are identical to those included in FARSITE and described by Finney (1998). These models are used to make calculations for all cells of a raster landscape, independently of one another. That is, there is currently no process that accounts for fire movement across the landscape or among adjacent cells. FlamMap only calculates the instantaneous behavior of a fire occurring at each location given local weather inputs. Map K in the appendix displays surface fire by flame length categories. It does not display fire behavior potential of densely stocked stands.

Wildfire Susceptibility

The fire susceptibility analysis (Map I) was done for the Sierra Nevada Forest Plan Amendment Process to assist in evaluating potential susceptability of large, severe fires at the broad scale. It provides a relative rating of one area to another. Relative hazard ratings were developed from mean values of flame lengths by CALWATER planning watersheds based upon outputs from Flam Map with 90th percentile weather conditions (Map K). Mean flame lengths were normalized for each forest by developing separate criteria of high, medium, and low. Relative risk of wildfire was based upon two seperate layers; one was the maximum fire occurrence values by CALWATER planning watershed. Maximum fire occurrence values by watershed were rated the same for all forests as follows: Greater than 5 (High), 3-4 (Moderate) and less than 2 (Low). The second was based upon elevation zones corresponding to the high probability of high severity fire from the McKelvey (1996) analysis, and addtional analysis based on the proportion of acres burned in high intensity fires. The elevational zones approximate the boundary between lower elevation mixed-confier pine and pine forests, fir dominated mixed-conifer forest, and upper-montane red fir or Jeffery pine and subalpine areas.

Two forms of fire risk analysis were done for this analysis.

Fire occurrence areas (Map E) was developed for the purpose of displaying the relative probability of a fire starting, as compared to other areas. This map is based on past fire occurrence, each acre in a fire occurrence area has the same probability that a fire could start. This does not display what the probability of a fire continuing to burn is, only the probability of ignition.

The Ignition Risk Map (Map E) was developed by calculating the number of fire ignitions (regardless of fire size) per 1000 acres within watersheds. It was used to rank the risk of cumulative watershed effects (Map H).

0 - .0019 fires per 1000 acres = Low Risk
.002 - .0039 fires per 1000 acres = Moderate Risk
.004 plus fires per 1000 acres = High Risk

Limitations and Assumptions of BEHAVE and FARSITE Computer Programs

BEHAVE. BEHAVE is a computer generated fire behavior model that calculates various fire behavior characteristics such as rates of spread, flame lengths, and maximum spotting distance given input for fuel models, fuel moisture, windspeed, slope, and tree parameters. It describes fire behavior at the flaming front and assumes fuel, fuel moisture, wind and slope to be constant.

In this analysis, BEHAVE was used to evaluate expected fire behavior to estimate fire suppression effectiveness, and to estimate maximum spotting distance potential. Weather and fuel moisture inputs represent 90th percentile weather conditions within the planning area recorded between July and September.

FARSITE. Fire arrival and behavior at a given point on the landscape is dependent on the behavior and time of travel en route to that location (rate of spread). This means that fire growth projections should generally worsen with time and spread distance because errors will be compounded, regardless of the accuracy or resolution of temporal or spatial data. That is, unless errors to one extreme are compensated by equal errors to the opposite.

Logically, however, a fire growth simulation should be most accurate when using accurate data at high spatial and temporal resolution. An "optimum" resolution for each landscape parameter, fire behavior type, and simulation purpose, probably exists so that the pertinent variability is preserved without irrelevant detail. The sensitivity of spatial fire simulation to the resolution and qualities of different fire input parameters on the landscape, however, remains to be tested.

Winds, Weather, and Fuel Moistures

The present version of FARSITE is capable of using two types of weather and wind inputs, weather and streams and gridded weather and winds.

Weather and Wind Streams. As with previous versions, FARSITE can use the simplified weather and wind input streams. Here, the open-winds that are provided are assumed parallel to the terrain and spatially constant but can vary in speed and direction over time. Spatial variability in winds is accomplished only through use of multiple data streams. Wind speeds are adjusted for mid-flame height based on canopy characteristics and fuel model. The weather streams specify daily maximum and minimum temperatures and humidities and the elevation of the observations. The time of the maximum temperature is assumed to be coincident with the minimum humidity. This will probably not be accurate for the period where thunderstorms occur or during passage of weather fronts.

Temperature and humidity observations are interpolated with a sine-curve (sensu Rothermel et al. 1986) to acquire temperature and humidity throughout the day. A sine-exponential interpolation (Beck and Trevitt 1989) may be an improvement, but has yet to be tested. A lapse-rate (3.5F/1000ft) is used to adjust these observations to other elevations on the landscape. Daily precipitation amounts are included in the weather stream and assumed constant across the landscape. Solar radiation at the ground surface is computed using canopy coverage and terrain information; in this version of FARSITE, all other canopy characteristics (height, height to live crown base, crown bulk density, and foliar moisture content) are assumed spatially constant except where optional spatial data themes have been provided (height, height to live crown base, and crown bulk density).

Dead fuel moistures are calculated using the procedures implemented in BEHAVE (Rothermel et al. 1986, Hartford and Rothermel 1991). The calculations for daily fine fuel moistures (at 1400 hours) are different from the calculations for hourly fuel moistures at other times of the day. This results in an abrupt shift in the fuel moistures at 1400 hours. It is not known how critical this inconsistency is to the results of long term fire behavior spread simulations. Live fuel moistures are assumed to remain constant throughout the simulation unless manually changed. There are currently no general models for all species of live fuels that describe moisture variation either diurnally or seasonally.

The limitations to fire spread projections of the simplified weather data are not really known. Obviously, model results would be expected to suffer where strong interactions of wind and terrain are present. Furthermore, calculations that depend on fuel temperature and moisture may not be accurate where shadows are cast by topography, precipitation varies elevationally or spatially, or water availability is significantly is altered (e.g. higher fuel moistures near streams) .

Gridded Weather and Winds. With version 3.0 of FARSITE weather and/or winds are available in gridded formats. see Gridded Weather for an explanation of these data.

Spread Patterns

Fire spread patterns generated using Huygen's principle with an elliptical wave have been found to agree reasonably well with observed surface fire spread under relatively simple conditions (Anderson et al. 1982, French 1992). Wind changes produced fire spread shifts close to those observed for fires spreading in grass fuels with essentially no influence of topography. It is not yet confirmed how well Huygen's principle simulates fire growth on complex landscapes. Sanderlin and Sunderson (1975) were apparently the first to apply a perimeter expansion technique, now generally referred to as Huygen's principle, to simulating wildland fire spread. From a comparison of predictions with the perimeters observed from the Potrero wildfire (September 1973) in Southern California, they concluded that the technique was acceptable for fire growth modeling in complex situations. The only other indications from complex circumstances are from some preliminary validations of FARSITE in which observed fire spread patterns was compared against surface fire spread predicted by the model (Finney 1994, Finney and Ryan 1995, Finney and Andrews 1996). These early comparisons were promising but the many potential sources of error in the observed data (fuel maps, perimeter maps, weather data etc.) preclude definite conclusions. More validations are planned and will be necessary before the accuracy of the program is defined.

For practical purposes, the most important result of the FARSITE tests to date has been that spread rates for all fuel models tended to be overpredicted by the Rothermel spread equation (Rothermel 1972). Sanderlin and Sunderson (1975) made a similar observation and ascribed the cause to problems relating windspeed to elliptical dimensions. Some problems may be a result of inaccurate data on fuel moistures, fuel descriptions (e.g. models), and weather. Also, wind reduction factors for forested areas and lee-side topographic sheltering can undoubtedly cause errors for spread rate calculations on some parts of a landscape.

However, even assuming that all the input data is accurate, the problem with overprediction may persist. The scale of time and space-averaged winds (e.g. hourly) and spatially homogenized fuels within rasters may be too coarse to reflect fine-scale variability in fire environment (temporal or spatial) that keeps fire actually spreading at variable rates. This could force the average fire spread rate over large areas and long time spans to be overpredicted. The nonlinear relationship between windspeed, fire acceleration, and fire spread rate means that the average windspeed cannot be expected to predict the average spread rate (Richards 1993). Fluctuating wind directions also cause overprediction of spread in the heading direction because they reduce the eccentricity of the fire shape compared to the ellipse (see #6 below).

The simple approach to correcting the spread rates, perhaps too simplistic for complex landscapes, is to assign rate of spread adjustment factors to each fuel type (Rothermel and Rinehart 1983). These factors must be based on empirical observations of previous fires, or of phases of growth of the existing fire, in patches of homogeneous fuels. They should be based on the heading portion of the fire, given that spread in other directions is dependent on the elliptical dimensions. It would be possible however, to compute the spread rate for one fuel type in a mixture of fuels if the following were known: 1) the fractional distance occupied by fuel type, 2) the average spread rate for the fuel mixture, and 3) the individual spread rates of other fuel components of the mixture. Then the equation for the harmonic mean (Martin 1988, Fujioka 1985) can be solved for the unknown spread rate. The adjustment factors, however, may not be constant throughout the duration of a fire. Average spread rates may change if wind variations change frequency when compared to the conditions used to obtain the adjustment factor. For example, adjustment factors determined from fire spread before a cold front may be not be adequate during and after the front passes because there may be more variability in wind speed or direction than before. The FARSITE model provides a means to apply and change fuel-specific adjustment factors throughout the simulation.

A number of assumptions are critical to modeling fire growth using Huygen's principle. As discussed below, some of these assumptions are probably violated by current modeling methods. The degree to which a technical violation limits the practical application of a model, however, is not yet known. This is the critical question, because models will never be fully valid at all scales or for all purposes, but they can be useful if the limitations are clearly understood by the user. The following paragraphs present a discussion of some major assumptions of the modeling method used for FARSITE. A detailed treatment of the subject was also written by Andre and Viegas (1994).

1. Fire spread is elliptical. This is probably not strictly true. The shapes of fires are assumed to be elliptical under uniform conditions because this is mathematically convenient (Van Wagner 1969). Fire shapes under different conditions have variously been described as ovoid (Peet 1967), as a pair of ellipses (Albini 1976, Anderson 1983) or as fan-shaped (Byram 1959). Green et al. (1983) found the ellipse fit as well as more complex shapes, given its simplicity and the absence of more definitive data. An analysis of fire shapes by Richards (1993) suggested that neither ovoid, double ellipse, or fan-shaped fires can be explained simply from variations in windspeeds or directions acting on an otherwise elliptical spread pattern. Richards? methods however, made the assumption that fire spread was independent of the shape of the fire front which may not be supported (see #2 below). Even if the assumption of elliptical fire shapes in continuous fuels is true, however, fire shapes in fuels that are not continuous at the scale relevant to mechanisms of fire propagation will not be elliptical or intuitive (Green 1983). For example, a fire may spread only in the heading direction because of wide spacing between fuel patches and would have the shape of a rectangular strip. Fire shapes resulting from discontinuous fuels will not be adequately modeled by FARSITE.

2. Fire spread in any direction is independent of the shape of the fire front (i.e. points along a fire front can be considered independent sources of wavelets). Recent studies suggest that this is not correct (Weber 1989, Cheney et al 1993). Radiative heat transfer ahead of a spreading fire has long been known to depend on the shape and length of the fire front (Byram 1959). Radiation from a continuous line fire decreases as the distance from it, but as the square of the distance from a point source fire. The violation of the shape independence assumption limits the extent to which Huygens? principle can be simply applied to spreading fire and distinguishes fire spread from the travel of light; portions along light waves do not interact as can portions of a fire front. For fire growth modeling, this means that the existing shape and length of the fire front along any segment should affect the nature of heat transfer and spread rate along that segment. Therefore, a broad flank of a fire that becomes a heading fire as a result of a wind direction shift should assume a different shape than predicted by a Huygens? algorithm. This will not be reflected in the current FARSITE model. The practical effects of these problems on fire growth patterns produced in a simulation are however, not yet known.

3. Fire acceleration is fuel dependent but independent of fire behavior. Fire acceleration is defined as the rate of increase in spread rate from the current rate to an equilibrium spread rate under constant environmental conditions. In version 2.0 of FARSITE you can adjust the fire acceleration constants for each fuel type. The fire acceleration equations (Alexander et al. 1992) in FARSITE compute the average and ending rate of spread for a timestep. These are likely to be important where the simulation uses small time-steps (<10min), where fuels and topography are very heterogeneous (spatially), and winds are variable. The incorporation of acceleration means that fire spread rates will not immediately adjust to the equilibrium spread rates when conditions change. The rate of fire acceleration is dependent on a rate factor. The default rate for all fuel types in FARSITE is subjectively set at .115 (Alexander et al. 1992) to allow acceleration to 90% of equilibrium rates after 20 minutes from a point source fire. Line source fires are known to accelerate much faster (Johansen 1987). These factors can be adjusted in FARSITE, but there are no data to guide settings for these factors. Although the equilibrium spread rate is dependent on fuel conditions, the buildup or acceleration rate has been found to be fuel independent for a variety of fuel types (excelsior, pine needles, conifer understories). A single acceleration rate may not be accurate for all fuel types (McAlpine and Wakimoto 1991), especially between very different fuel types. Fire in grass fuels is expected to accelerate more rapidly than in slash fuels, but there are few data to guide these settings. Acceleration is presumed to be independent of the fire behavior or eventual spread rate. Thus, the same time is required in a given fuel type to achieve a steady-state spread rate regardless of the environmental conditions.

4. Fires will instantly achieve the expected elliptical shape when burning conditions change (e.g. wind speed or slope steepness). This assumption is probably acceptable for simulations with a time step longer than a few minutes. Laboratory experiments (McAlpine 1989) suggest that shape changes occur relatively rapidly compared to the time required for buildup in spread rate or intensity.

5. The elliptical shapes are fuel independent; shape (not size) is only determined by the resultant wind-slope vector. This assumption is probably acceptable because 1) empirical relationships between windspeed and elliptical dimensions suggest shapes are common to a variety of fuel types over a wide range of ambient windspeeds (Alexander 1985), and 2) the empirical coefficients for wind and slope effects on fire spread rates used in the Rothermel spread equation are dependent of fuel bed characteristics (Rothermel 1972). These coefficients are the unit vectors used to obtain the resultant wind-slope vector.

6. Variation in windspeed and direction at a higher frequency than the wind stream resolution, does not affect the elliptical fire shape. This is not technically correct, but the importance of its effect on fire growth patterns is not yet clear. Fluctuating wind directions decrease the length to breadth ratio of an otherwise elliptical fire (see Technical Documentation). This has the effect of overpredicting the heading spread of a fire at the expense of flanking spread. Some compensation for the overpredicted heading spread will be achieved through the rate of spread adjustment factors.

7. The origin of an elliptical fire is located at the rear focus of the ellipse. The focus is assumed as a starting point because it provides an implicit means to calculate backing spread rates (see Technical Documentation). Alexander (1985) reports that using the origin as the focus may underpredict the backing spread. At present, FARSITE allows the user to select a constant backing spread rate calculated from the spread rate under zero slope and wind for the given fuel type (Rothermel 1983).

8. The spread of a continuous fire front can be approximated using a finite number of points. The adequacy of this assumption is dependent on the spatial resolution required by the user and the resolution specified for the simulation (see Technical Documentation). It is assumed that a resolution can be specified that preserves the features of fire growth but ignores irrelevant spatial detail. This is dependent on the purpose and requirements for the simulation. The same concept is implicit in maps of fire growth made by direct observation; minor variations in fire position that result from rocks or small discontinuities in fuel are ignored. The relevant resolution probably decreases as the fire gets larger.

9. The FARSITE model is not designed to determine if a fire will spread. It is also not technically designed for modeling fire spread only by smoldering or by the rolling of burning debris, even though the resulting spread rate may be approximated by judicious use of the adjustment factors for a given fuel type (see fuel inputs). The FARSITE model cannot determine where or if a fire will cross a barrier (e.g. a creek or a vertical cliff) unless the resolution of the data are fine enough to reflect the areas where fire may cross.

Multiple Fires

Although FARSITE will handle up to 5,000 simultaneous fires, the fire spread patterns of neighboring fires will not necessarily be accurately represented because fire interaction with weather and fuels is not accounted for. For example, behavior resulting from ?back fires? set for suppression purposes, or a prescribed fire ignition pattern that is applied to ?draw? the fire together at different times, places, or stages of build-up will not be addressed by FARSITE. Extreme fire behavior, e.g. plume dominated fires, that are affected by feedback between the weather and fire behavior are not intended to be simulated by FARSITE. Users should not assume that ignition patterns used for prescribed burning will result in correct simulation of fire behavior!!

Crown Fire

The crown fire models of Van Wagner (1977, 1993, and Alexander 1988) have been implemented in FARSITE. This approach requires information on crown fuels and the forest canopy, including:

· Effective height to live crown base,
· Crown bulk density
· Tree height
· Foliar moisture content

Although FARSITE presently requires a canopy cover theme, the above crown-fuel characteristics must be constant for areas having canopy coverage, unless the optional crown fuel themes have been provided to the .LCP file.

Van Wagner (1993) notes that the height to live crown base is a difficult parameter to measure. The height to live crown base is really an "effective" number that incorporates ladder fuels (see Fahnestock 1970) and understory fuels such as small trees that assist the transition to crown fire. Thus, height to live crown base will not be a simple measurement in multi-storied stands.

Heat required to ignite the crown is based only on fuel sizes on the 100hr timelag and smaller (<3? diameter). This limitation may underestimate the potential for crowning or torching because larger woody fuels (1000 hr+) and their contributions to radiative and convective heating of overstorey fuels are ignored. Rothermel (1991, 1994) discusses the contribution of large woody fuels to the development of convection columns and consequent crown fire behavior.

The wind-slope vectoring for crown fire has not been tested and may not be realistic. FARSITE presently uses the wind-slope vector direction from the understory surface fire with midflame winds. The reason is that transition to crown fire in Van Wagner?s (1977, 1993) model is dependent first on the surface fire behavior that is determined by midflame winds. The problem with later combining an open-wind vector with the surface slope effect is that the range of data used to develop the slope coefficient in the Rothermel (1972) model may not be applicable to crown fire. The slope coefficient depends on fuel bed parameters not accounted for by the canopy fuels in which the fire is then burning. This situation needs further work and testing.

Spotting

The existing spotting models (Albini 1979, 1981, 1983a, 1983b, Morris 1987) were originally devised to predict the maximum distance burning embers would travel over flat and regularly undulating terrain. The maximum spotting distance is determined by the balance between particle size, burnout rate, and time or distance traveled. Smaller particles are lofted higher and transported further, but burnout sooner than larger particles. Thus, as published, Albini's equations for the maximum spotting distance cannot be implemented for complex topography because winds, terrain, and forest canopy can all vary.

At present only the model for spotting from torching trees (Albini 1979) is present in FARSITE. The purpose of the spotting capability of FARSITE is to compute the maximum distances that particles of different sizes would travel over complex landscapes. These indicate the potential distances ahead of the fire that spotting could be found, assuming winds vary only as a function of height above ground or as specified spatially by the weather/wind grid. Nevertheless, this greatly oversimplifies reality in mountainous terrain.

Depending on topography, Albini's equations may suggest the farthest spotting distances are produced by larger particles that aren't transported over deep ravines. The spotting model in FARSITE does not intend to predict the number of embers produced, or exact locations that embers will land, only the direction and distance embers might land.

Spotting is produced whenever some form of crown fire develops (torching and running crown fire). You must recognize, however, that the torching tree model of ember lofting was not intended for representing ember lofting from a running crown fire. It will likely underestimate both the ember sizes, lofting height, and ultimate spotting distances under conditions of running crown fire.
 
 

Return to Table of Contents