Simulating fuel treatment effects in dry forests of the western United States: testing the principles of a fire-safe forest - PDF

Please download to get full document.

View again

of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Engineering

Published:

Views: 7 | Pages: 12

Extension: PDF | Download: 0

Share
Related documents
Description
This file was created by scanning the printed publication. Text errors identified have been corrected; however some errors may remain Simulating fuel treatment effects in dry forests of the western
Transcript
This file was created by scanning the printed publication. Text errors identified have been corrected; however some errors may remain Simulating fuel treatment effects in dry forests of the western United States: testing the principles of a fire-safe forest Morris C. Johnson, Maureen C. Kennedy, and David L. Peterson Abstract: We used the Fire and Fuels Extension to the Forest Vegetation Simulator (FFE-FVS) to simulate fuel treatment effects on stands in low- to midelevation dry forests (e.g., ponderosa pine (Pinus ponderosa Dougl. ex. P. & C. Laws.) and Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) of the western United States. We evaluated treatment effects on predicted post-treatment fire behavior (fire type) and fire hazard (torching index). FFE-FVS predicts that thinning and surface fuel treatments reduced crown fire behavior relative to no treatment; a large proportion of stands were predicted to transition from active crown fire pre-treatment to surface fire post-treatment. Intense thinning treatments (125 and 250 residual trees ha -1 ) were predicted to be more effective than light thinning treatments (500 and 750 residual trees ha -1 ). Prescribed fire was predicted to be the most effective surface fuel treatment, whereas FFE-FVS predicted no difference between no surface fuel treatment and extraction of fuels. This inability to discriminate the effects of certain fuel treatments illuminates the consequence of a documented limitation in how FFE-FVS incorporates fuel models and we suggest improvements. The concurrence of results from modeling and empirical studies provides quantitative support for fire-safe principles of forest fuel reduction (sensu Agee and Skinner For. Ecol. Manag. 211: 83 96). Resume : Nous avons utilise le module complementaire sur le feu et les combustibles du simulateur de la vegetation forestiere (u Fire and Fuels Extension Forest Vegetation Simulator (FFE-FVS)) pour simuler les effets du traitement des combustibles sur peuplements clans des for-as seches situees a une altitude allant de faible a moyenne (p. ex. pin ponderosa (Pinus ponderosa Doug]. ex P. & C. Laws.), douglas vert (Pseudotsuga menziesii (Mirb.) Franco)) dans l'ouest des Etats-Unis. Nous avons evalue les effets des traitements sur le comportement d'un feu potentiel a la suite du traitement (type de feu) et sur le risque do feu (indice d'embrasement des cimes). La simulation predit que 1'eclaircie et le traitement des combustibles de surface reduiraient les feux de cime comparativement a 1'absence de traitements; une forte proportion de peuplements potentiellement sujets a un feu de cime avant d'avoir ete traites ne seraient plus sujets qu'a un feu de surface apres avoir ete traites. Des traitements d'eclaircie forte (125 et 250 arbres residuels ha -1 ) seraient plus efficaces que des traitements d'eclaircie faible (500 et 750 arbres residuels ha -1 ) selon les predictions. La simulation a pi-edit que le brulage dirige serait le traitement des combustibles de surface le plus efficace tandis qu'il n'y avait pas de difference entre 1'absence de traitement des combustibles de surface et la recuperation des combustibles. Cette incapacite a distinguer les effets de certains traitements des combustibles illustre la consequence dune limite documentee concernant la fawn dont la simulation incorpore Ies modeles de combustibles et nous suggerons des ameliorations. La convergence des resultats provenant de la modelisation et des etudes empiriques appuie de fagon quantitative 1es principes de securite incendie qui pronent la reduction des combustibles (sensu Agee et Skinner For. Ecol. Manag. 211: 83 96). Introduction Dry forest types prevalent in western North America historically exhibited high-frequency, low- to moderate-severity fire regimes (Agee 1993; Taylor and Skinner 1998). Past management practices such as livestock grazing, wildfire suppression, and timber harvest have modified the fuelbed characteristics (vegetation composition and structure) and fire behavior of these dry forest types (Weaver 1943; Bisweil 1959; Dodge 1972; Hessburg and Agee 2003; Hessburg et al. 2005). They now have large amounts of fuel loads and ladder fuels such as tall grasses, shrubs, tree branches, and understory trees (Parsons and DeBenedetti 1979; Bonnicksen and Stone 1982; Peterson et al. 2005). As a result, these forests are more susceptible to active crown fire and higher burn severity than they were historically (Laudenslayer et al. 1989; MacCleery 1995; Arno and Allison-Bunnell 2002). Fuel treatments are advocated to reduce fire hazard caused by increased stem densities in low- to moderate-severity fire regimes (Graham et al. 2004; Peterson et al. 2005). Rather Received 28 March Accepted 4 February Published at on 28 April M.C. Johnson and D.L. Peterson. USDA Forest Service, Pacific Northwest Research Station, 400 North 34th Street, Suite 201, Seattle, WA 98103, USA. M.C. Kennedy. College of the Environment, School of Forest Resources, University of Washington, Box , Seattle, WA , USA. Corresponding author: M.C. Johnson ( Can 1 For Res (2011) doi: /x11-032 Johnson et al than stop wildfires (Finney and Cohen 2003), treatments are meant to decrease fireline intensity (i.e., rate of heat energy released), reduce crown fire initiation, and support suppression operations (Agee 1996). Silvicultural thinning practices such as thinning from below and pruning and the subsequent removal of surface fuels are effective options to reduce stand density, remove ladder fuels, and increase stand heterogeneity (Graham et al. 2004). With millions of hectares of dry forests in the western United States requiring fuel treatment, forest and fire managers need recommendations and information to support science-based decision making for fuel management (Graham et al. 1999; Peterson et al. 2005; Schmidt et al. 2008). Agee and Skinner (2005) proposed four guidelines to assist managers in developing effective treatments to reduce crown fire hazard and to understand treatment consequences: reduce surface fuels, increase canopy base height, decrease canopy bulk density, and retain large fire-resistant trees. These principles of a fire-safe forest are based on our current knowledge of crown fire theory (e.g., Van Wagner 1977) and are intended to increase the resilience of stands to wildfire (Agee and Skinner 2005). Several constraints, including insufficient funding, logistics, and safety, prevent robust experimental testing of the effects of different fuel treatments (Finney and Cohen 2003). Consequently, most validation of fuel treatment effects on mitigating fire hazard is done post hoc without high-quality pre-wildfire data (Pollet and Omi 2002; Finney et al. 2006). As an alternative, simulation models provide quantitative prediction of the effects of modifying fuelbed characteristics on crown fire hazard and the probability of crown fire initiation (Graham et al. 2004). The Fire and Fuels Extension to the Forest Vegetation Simulator (FFE-FVS) (Reinhardt and Crookston 2003) simulates forest growth and potential fire hazard and fire type of different vegetation types in the United States. FFE-FVS is the standard simulation model used by most federal, state, and tribal government agencies (Dixon 2003). In this study, we evaluated how FFE-FVS predicts the effects of fuel treatments (i.e., combinations of thinning and surface fuel treatments) on simulated fire hazard and potential fire behavior type in a set of stands from dry forests of the western United States. This study extends conceptual and analytical work on fuel treatments and fire behavior from the fuel treatment guidebook (Johnson et al. 2007). Unlike other studies that have used FFE-FVS to examine treatment effects for only a few stands in specific geographical areas (Fiedler et al. 2001; Calkin et al. 2005; Skog et al. 2006), we used a large set of stand data from several geographic regions to perform a virtual test of the four principles of a fire-safe forest (Agee and Skinner 2005) across a broad range of dry forest conditions. Methods Model description: FFE-FVS We used FFE-FVS (version 6.21) (Reinhardt and Crookston 2003) to simulate the effects of thinning and surface fuel treatments on potential fire hazard and fire behavior at small spatial scales (tens of hectares). FFE-FVS can simulate a fire or estimate the potential effect of a fire under userspecified weather and fuel conditions. A simulated fire modifies stand and fuels conditions (e.g., kills trees, reduces fuel loading) and alters the trajectory of stand succession and fuel dynamics. FFE-FVS calculations of potential fire effects are conducted before the effects of a fire are simulated. FFE-FVS is a consolidation of two computer modules, FVS and FFE. FVS is an individual-tree, distance-independent, growth-and-yield model that simulates tree growth, mortality, and the effects of a variety of silvicultural treatments (Dixon 2003). Stands are the basic unit of management, and projections depend on interactions among trees within the stands. Twenty variants have been developed and calibrated to cover most forestlands in the United States (Fig. 1). For example, the Southern Oregon/Northeastern California (SO) variant was fit to data representing forest types (33 species or species groups) in southern Oregon and northeastern California (Keyser 2008). The variant is applicable to a variety of tree species, forest types, and stand structures in the Deschutes, Fremont, Winema, Klamath, Lassen, Modoc, Plumas, Shasta, and Trinity national forests and corresponding Bureau of Land Management and industry lands (Keyser 2008). Data used to develop these equations were derived from numerous forest inventories, silvicultural stand examinations, research plots, and tree plantation studies (Keyser 2008). Major species modeled in the SO variant include western white pine (Pinus monticola Dougl. ex D. Don), sugar pine (Pinus lambertiana Dougl.), Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco), white fir (Abies concolor (Cord. & Glend.) Lindl.), mountain hemlock (Tsuga mertensiana (Bong.) Carriere), incense cedar (Libocedrus decurrens Torr.), lodgepole pine (Pinus contorta Dougl. ex Loud.), Engelmann spruce (Picea engelmannii Parry ex Engelm.), noble fir (Abies magnifica Rehd.), and ponderosa pine (Pinus ponderosa Dougl. ex P. & C. Laws.). For this study, we used seven of the 20 FVS variants (Fig. 1) and present the results of two of those variants in detail (with the remaining five in the supplementary material 1 ). FFE simulates snag dynamics and woody fuel accumulation and decomposition through time using fuels information projected by FVS (Reinhardt and Crookston 2003). Fire behavior and crown fire hazard are computed using methods developed by Rothermel (1972), Albini (1976), and Scott and Reinhardt (2001). Using the stand characteristics, FFE-FVS calculates two indices of crown fire hazard: torching index and crowning index (Table 1). Torching index is the windspeed (kilometres per hour) required for crown fire initiation and crowning index is the windspeed (kilometres per hour) required to support an active crown fire. Lower values of each of these indices indicate increased fire hazard. FFE-FVS also uses the stand data to classify the stand as one of four types of potential fire behavior associated with increasing fire hazard: surface fire, conditional fire, passive crown fire, and active crown fire (Table 1). In a stand classified as potential surface fire, a fire is predicted to spread primarily within the surface fuels (dead branches, leaves, needles, low vegetation). With potential conditional fire, conditions for sustained active crown fire spread are met, but conditions for crown fire initiation are not. In such a stand, FFE-FVS predicts that if the 1 (Supplementary data are available with the article through the journal Web site (http://www.nrcresearchpress.comlcjfr). 102 0 Can. J. For. Res. Vol. 41, 2011 Fig. 1. There are 20 total FVS variants, each calibrated separately to a specific geographic area of the United States. We chose seven FFE- FVS variants to evaluate for the current study. The results for the East Cascades and Northern Idaho variants are presented in detail, with the results for the remaining variants presented in supplementary material. 1 mile = 1.61 km. 1 Table 1. The Fire and Fuels Extension to the Forest Vegetation Simulator (FFE-FVS) (Reinhardt and Crookston 2003) simulates forest growth and potential fire hazard (e.g, torching index) and fire type (e.g., active crown fire) of different vegetation types in the United States. Fire hazard Torching index Crowning index Potential fire type Surface fire Conditional surface fire Passive crown fire Active crown fire Description The 6.1 m wind speed (km h -1 ) at which a surface fire is expected to ignite the crown layer. This depends on surface fuels, surface fuel moisture, canopy base height, slope steepness, and wind reduction by the canopy The 6.1 m wind speed (km h -1 ) needed to support an active or running crown fire. This depends on canopy bulk density, slope steepness, and surface fuel moisture content Spreads primarily within the surface fuels Conditions for sustained active crown fire spread are met, but conditions for crown fire initiation are not. If the fire begins as a surface fire, then it is expected to remain so. If it begins as an active crown fire in an adjacent stand, then it may continue to spread as an active crown fire Individual or small groups of trees ignite, but solid flaming in the canopy cannot be maintained except for short periods The entire fuel complex becomes involved, but the crowning phase remains dependent on heat released from the surface fuels for continued spread fire begins as an active crown fire in an adjacent stand, it may continue to spread as an active crown fire (Scott and Reinhardt 2001). With potential passive crown fire (torching), a fire is predicted to occur when individual or small groups of trees ignite, but solid flaming in the canopy cannot be maintained except for short periods. In a stand classified as potential active crown fire, FFE-FVS predicts that the entire fuel complex becomes involved in a fire, but the crowning phase remains dependent on heat released from the surface fuels for continued spread (Van Wagner 1977). In this study, we evaluated simulated fuel treatment effects on the predicted torching index and the potential fire type classification. Johnson et al Fig. 2. Study design. Seven FFE-FVS variants were evaluated in total, with the results for East Cascades and Northern Idaho (shaded) presented here in detail. Results for the remaining variants are presented in the supplementary material. 1 For each variant, the pre-treatment fire type and torching index were identified by FFE-FVS for each stand. (a) The stands in each variant were partitioned by their pre-treatment fire types. (b) Twelve combinations of thinning and surface fuel treatments were simulated for each stand in each pre-treatment fire type for each variant (in the example above, there are 800 stands in the Northern Idaho variant classified pre-treatment as potential active fire type). Posttreatment simulated values were change in log torching index ( ti) (eq. 1) and post-treatment potential type (Post-trt fire type). Twenty-eight separate analyses were conducted for each of the response variables ( ti, post-treatment fire type), with the predictor variables of thinning treatment and surface fuel treatment. Table 2. Weather variables used to estimate fire behavior across all stands pre-treatment and post-treatment and weather parameters used in the simulation of prescribed fire as a surface fuel treatment. *Percentage of stand burned equals 75%. Stand examination data We downloaded data for stands from a relational database that contains measurements collected in the field from national forests (FSVeg). The database contains plot vegetation data from field surveys such as Forest Inventory Analysis data, stand exams, inventories, and regeneration surveys (USDA Forest Service 1992). Of these stands, we retained (41%) that met the following selection criteria: initial stand density 750 trees ha -1 (tph) (corresponding to the least intense thinning treatment in our study), slope 30% (standard maximum slope for fuel treatments), and species composition including at least 1% of dry tree species typically dominant in dry forests of the western United States (e.g., Pinus ponderosa, Pseudotsuga menziesii). We discarded several stands with low canopy fuels and dominated by hardwoods because FFE-FVS could not calculate fire hazard in these stands. We partitioned the stands into seven of the 20 FFE-FVS variants (Fig. 1): East Cascades, South Central Oregon and Northeastern California, Northern Idaho, Central Idaho, Central Rockies, Eastern Montana, and Blue Mountains. For each stand in each of the seven variants, we used FFE-FVS to predict the initial pre-treatment potential fire behavior type and torching index (Fig. 2a) with the same weather-related variables (e.g., fuel moisture and windspeed) for all variants (Table 2). Thinning and surface fuel treatments We developed a treatment matrix to emulate Agee and Skinner's (2005) four principles of a fire-safe forest (Fig. 2b). Four thinning treatments and three surface fuel treatments were programmed into FFE-FVS batch processing format for each variant. All four thinning prescriptions were to thin from below to a target residual density: 750, 500, 250, or 125 tph. Only trees up to 46 cm diameter breast height were removed in the simulation. For each thinning density, three surface fuel prescriptions were simulated: no action (all slash remained in the stand), extraction (all slash removed from the stand), and prescribed fire (trees 15 cm left in the stand, trees of cm had boles removed and branches left in the stand). Each treatment combination was simulated separately for each stand, and the torching index and potential fire behavior type were recorded for the first year following treatment. We generated an individual FFE-FVS projection for each stand and treatment combination. Simulation analysis The simulations performed by FFE-FVS are deterministic, and in our study, we have a nonrandom sample of stand examination data. Therefore, our analysis applies only to the set of stands evaluated and in the context of the simulation model structure. In effect, we have a census of the stands 102 2 Can. J. For. Res. Vol. 41, 2011 Fig. 3. Interaction plot showing the mean change in log torching index ( ti) for each combination of thinning (four levels) and surface fuel (three levels) treatments for the East Cascades variant for stands classified pre-treatment as (a) surface, (b) conditional, (c) passive, and (d) active fire types. Vertical lines represent ±2 SE for each treatment combination mean. Note that the x-axis is a categorical variable (thinning treatment) rather than a scalar. All plots are shown on the same y-axis to enable comparison of mean values among pre-treatment fire types. A positive value of mean ti indicates that torching index increases post-treatment, thereby decreasing fire hazard. The mean ti increases from the less intense thinning treatment (750 residual trees ha -1 ) up to the second most intense thinning treatment (250 residual trees. ha -1 ) and then levels off or decreases from 250 residual to 125 residual trees ha -1. Overall, the mean ti is higher for the prescribed fire surface fuel treatment than for the no action and extraction surface fuel treatments. East Cascades Pre treatment Fire Type available to us that meet our criteria. In our analysis of the simulated results, our conclusions rely on interpretation of how patterns change across the treatment combinations. We make suggestions for how the results may apply more generally to dry forest types of the western Unit
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks