Methodology and Tools
Determination of Critical Stands and Stand Types
The critical stands are determined based on their most probable fire behavior and its comparison with a “natural” fire regime. Two factors are used to assess the criticality:
- The stands potential for sustaining and propagating a stand-replacing fire (i.e. a fire that kill essentially all the trees and leads to mineralization of a large part of the organic soil material. This factor is primarily a function of the available fuel load.
- The probability (or frequency) of a successful ignition of such a fire both from within the stand and sources outside of it (e.g. adjacent stands). This probability depends on the quality of the fuel load, but also on environmental conditions (precipitation, wind etc.).
It is necessary to consider both aspects, since neither the fuel load nor the ignition frequency alone is sufficient to determine critical stands with an unnatural fire regime. Subalpine or boreal forests can naturally have very high fuel loads because of rare fires and slow decomposition. But the low probability that successful ignition occurs over large areas reduces the fire risk (as defined as the product of extent and probability of a fire). In these forests, high fuel loads are not a deviation from a more or less natural fire regime and thus not considered critical. On the other end of the ecotone spectrum, in lower elevation with mediterranean climate such as the chaparral in southern California, the fire frequency and extent has not been changed significantly by fire suppression in the 20th century (Keeley et al. 1999), and it can be assumed that the fire regime is still close to its historical state (although the damage may have increased due to an increase in values at stake).
According to this reasoning, it is to be expected that the forests of intermediate elevation have experienced the most critical changes in their fire regime. They were historically subject to frequent fires with a return interval in the order of decades, and thus regular elimination of the ground and ladder fuels. Recent studies show that the return intervals have increased by as much as an order of magnitude (as a result of fire suppression), and the related fuel buildup results in fireline intensities that may well be comparable with those in boreal forests (Alexander 1982). Survival of trees under such conditions is highly unlikely, and total mortality over large areas is historically an rare fire pattern in the intermediate elevations. The Ponderosa pine and Mixed conifer forest types are thus determined to be the stands most affected by changes in the fire regime (PSW 1996) and therefore of particular interest for this study.
Defining and Assessing the Fire Resilience of a Stand
Obviously, not all fires can be prevented. The absolute fire-safety is not achievable, and the goal of fire risk management is a relative resilience, which is defined as the potential of a stand to resist fire impacts to an extent such that none of the long-term forest functions is lost.
The fire resilience is described exclusively in terms of the stand structure, although it is evidently also depending on factors with a temporal variability such as the season or the weather. It is assumed that over an extended period, they will be sufficiently frequent in a state where they do not contribute to the fire resilience (e.g. no precipitation), i.e. there will be times when the stand structure is the only effective factor to make a forest fire resilient.
Loss of long-term forest function results when long-term structural elements of the stand are lost. Within any given stand, the upper canopy represents the most important long-term structure, since it usually contains the oldest and largest trees. Other structures such as the ground vegetation or natural regeneration have also important functions for the stand as a whole, but they recover much faster if they get destroyed by fire (although repeated destruction can endanger the long-term stand functionality as well).
Therefore, for the purposes of the study, a stand is considered sufficiently fire resilient if the probability of a complete loss of the upper canopy is reasonably low. This is the case if the stand structure prevents a continuous propagation of a fire within the upper canopy and continuous lethal scorching of individual trees of the canopy as a result of ground fires crowning.
Whether a fire can spread from a point A to a point B in a spatial element depends on the heat flux along the vector AB. The heat flux is a function of three parameters: i) the heat of ignition of the fuel, ii) the rate of spread and iii) the mass of fuel available within a unit of space. Assuming that the first factor is constant, a minimum mass flow of fuel can be determined that is required to propagate the fire. Assuming further that the rate of spread is limited, it is possible to calculate a minimal fuel bulk density within the spatial unit under consideration. [For a more detailed discussion of this concept, see Agee 1996]. This concept can now be used to describe fire resilient stand structures, as is shown in Figure 1.
Fig. 1: Stand structures and fire spread. Only the structures depicted in Case V and Case VIII can reliably prevent continuous loss of upper canopy trees.
Figure 1 shows a very important fact that must be taken into account when fuel treatments are considered: to eliminate critical fire spread paths with regard to the upper canopy, it is not sufficient to reduce the density in the canopy layer or in the ladder fuel layer alone. Fuel densities must not exceed a critical level in the canopy AND the ladder layer. Any intervention must address not only the large trees that are to be protected, but also the smaller trees that function as fuel ladders. The ground layer is assumed to be lost in case of a fire and there is no effort to specifically modify this compartment for inherent fire-resilience.
Estimating or even measuring the fuel bulk density in the canopy is difficult, and the available data about this variable are very limited. For practical purposes, it is necessary to find a substitute density measure that can be determined more easily. Experience from real fires and use of simulation models (see for example Van Wagtendonk 1996) suggest that the canopy closure of a stand is suitable measure for assessing the probability of a fire spreading through the canopy; and based on the literature, it seems that below a canopy closure of 40%, a stand is likely not to propagate fires through the upper canopy even under adverse environmental conditions. Consequently, this value is used as a target for planning the intensity of the fuel treatments. However, with the methodology outlined below, it is possible to vary this target (e.g. increase it to 50% or more) if new findings should support such a change. More important than the raw number is the condition that it is not exceeded either in the upper canopy or in the ladder fuel layer.
Selection of a Silvicultural Management System
The need to include different vertical layers in the treatment requires that a silvicultural system is selected that extends to more than one tree size class. Pure age-class management systems usually focus on trees of ± equal height and are thus not suitable for this purpose. There are basically two concepts that can be used instead:
- Uneven-aged management based on diameter classes according to a DBQ model, where the number of trees in the stand decreases exponentially with increasing diameter. The assumption of this model is a constant survival probability for trees of all sizes and a positive correlation between diameter and height.
- Layer-oriented management based on height classes, where different layers are distinguished based on the stand height hdom.
The second concept does more directly address the problem of vertical fuel continuity (ladder fuels). However, it is more difficult to implement on an operational level, since it requires an artificial height stratification in addition to the more common strata species and diameter. For this reason, the DBQ model has been selected as a basis for the silvicultural management system.
Early simulations have shown that applying a pure DBQ model with a 40% canopy cover target and a diameter range of 40 inches and more (which is realistic in the Sierra Nevada) results in extremely flat diameter distributions, i.e. the survival probability must be assumed close to 100% to satisfy such a model. To avoid such an unrealistic assumption, the diameter range in the simulation was divided in two segments, one to which a pure DBQ model is applied (usually covering diameters between 2 and 30 inches) and one where a linear model is used (covering larger diameters). This “mixed” approach also helps in evaluating the long-term effects management policies such as the California Spotted Owl guidelines have on fire resilience.
Selection of Simulation Tools
The simulations are performed on a stand or inventory plot level, not on a landscape level. This and the goal of providing a generic framework that can be used by the interested public at large makes the US Forest Service Forest Vegetation Simulator (FVS) the most suitable modeling tool. FVS is available for a broad range of locations, and it is public domain. It is well known and accepted within the user community, and although there may be more accurate solutions for specific purposes, FVS offers an unparalleled flexibility for the purposes of the study.
However, FVS has a serious shortcoming with regard to fuel treatment simulations. Most local variants do not model natural regeneration automatically. Since these small trees represent an important source of ground and later also of ladder fuels, it is imperative that some kind of regeneration is simulated. The problem is that there is only very little information available about the regeneration patterns in most forests of the west. Fortunately, the use of a DBQ model alleviates some of this uncertainty. In the simulations, it is assumed that there is always enough regeneration to “keep the model going”, i.e. to provide enough trees to fill the smallest diameter class. Based on the low number of trees that is required to meet this condition (generally below 100 trees per acre every 20 years), this seems a reasonable assumption. If the real regeneration is more abundant, the surplus will be eliminated in the next intervention. If there is less regeneration in reality than necessary to ensure stand continuity, planting may be required.
The graphical user interface Suppose allows a pseudointeractive operation of FVS. This is convenient for individual simulation runs, but inefficient for a large number of runs. FVS has thus been used primarily in a batch mode, with all the treatment prescription supplied to it in a keyword file. The use of a mathematically extract function to determine the stand composition allows automation of the keyword building process, and a Tcl/Tk front-end user interface called “Keyword File Builder” has been developed to speed up the input process.
FVS output are text files, and Suppose allows to extract reports and graphs from them.
Besides, FVS also offers the possibility to write tree lists that can be translated for use with the Stand Visualization System SVS. The reports are further processed with a standard spreadsheet application.
References
- Agee, J. K. (1996): The influence of forest structure on fire behavior. In: Proceedings of the 17th annual Forest Vegetation Management Conference: 52 – 68.
- Alexander, M. E. (1982): Calculating and interpreting forest fire intensities. Can. J. Botany 60: 349 – 357.
- Keeley, J. E. et al. (1999): Reexamining Fire Suppression Impacts on Brushland Fire Regimes. Science 284: 1829-1831.
- PSW (1996): Appendices [to the] Draft Environmental Impact Statement.
- Van Wagtendonk, J. W. (1996): Use of a Deterministic Fire Growth Model to Test Fuel Treatments. Sierra Nevada Ecosystem Project, Final Report to Congress. Vol. II: 1471 – 1492. UC Davis. Davis, CA.