Introduction

In temperate forests, bark beetle attacks and fire are two dominant ecological processes which often interact, and these processes may both become more frequent with warming temperatures (Bentz et al. 2010). The spruce bark beetle (Dendroctonus rufipennis Kirby) outbreak on the Colorado Plateau in the 1990s (DeRose et al. 2013) was responsible for considerable mortality of Picea engelmannii. In stands dominated by Picea engelmannii, mortality was commonly 95% (and up to 99%) of stems ≥ 10-cm diameter at breast height (DBH; DeRose and Long 2007, 2012a, 2012b), but mortality in stands with high proportions of other species has been understudied. This extensive Picea engelmannii mortality is responsible for a pulse of deadwood (down woody debris) that is starting to accumulate on the ground. This large quantity of new deadwood will have far-reaching implications for carbon storage, habitat structure, regeneration, and potential fire behavior within these forests.

The impacts of wildfires in recent years have provoked a great deal of public concern, and mitigating fire severity has become a top priority for managers among many western forests (especially those affected by bark beetle outbreaks). The rapid rise in deadwood associated with snagfall may contribute considerably to future extreme fire behavior (e.g., longer flame length, greater crown fire potential) either overall or in discrete patches of high deadwood accumulation. Whether deadwood in forests with high levels of beetle-induced mortality should be reduced mechanically (i.e., by salvage logging) remains a pressing concern. Removal of snags may reduce fire risk but also reduce or eliminate ecosystem function (Janisch and Harmon 2002; Lutz and Halpern 2006; Thorn et al. 2020). Previous work in other forest types has shown that the results of salvage logging can sometimes increase fuel loads (i.e., Donato et al. 2006) and sometimes reduce subsequent fire risk (i.e., Povak et al. 2020).

Furthermore, our understanding of how spatial heterogeneity in trees, snags, and deadwood drives fire behavior is limited. Fire behavior is intricately linked to the total amount, size distribution, and the spatial arrangement of fuels (Miller and Urban 2000). Two-dimensional spatial patterning may be captured by a variety of field methods, but the three-dimensional spatial structure (i.e., stacking) of fuels is much more difficult to measure. The three-dimensional arrangement of fuels is particularly important to understanding potential fire behavior (Loudermilk et al. 2012; Hiers et al. 2009; Pimont et al. 2016; Jeronimo et al. 2020) as fuel mass and structure influence the physical process of combustion (combustion rate and efficiency) as well as crown fire potential (greater with higher flame lengths and a more continuous canopy). A variety of field methods have been used to quantify three-dimensional fuel structure and relate this to observed fire behavior (e.g., Hudak et al. 2020), but existing methods (LiDAR and three-dimensional bulk density sub-plots) are labor intensive, and it is difficult to obtain complete coverage of large study areas; in this study, we demonstrate an alternative method.

Deadwood contributes to ecological processes in addition to surface fuel dynamics (Harmon et al. 1986; Thorn et al. 2020). Between snagfall and deadwood’s eventual decomposition or combustion (Stenzel et al. 2019), it contributes to biodiversity by changing the microenvironment (Vrška et al. 2015), serving as substrate for seedlings and bryophytes (Taborska et al. 2015) and regulating carbon cycling dynamics (Harmon et al. 1986; Privetivy et al. 2018; Harmon et al. 2020). Because of the differential effects of large-diameter trees, snags, and deadwood on forest structure (Lutz et al. 2012; Lutz et al. 2013; Réjou-Méchain et al. 2014), it is important to specifically consider the effects and dynamics of larger pieces (sensu Lutz and Halpern 2006), rather than adopting an approach that ignores piece size.

As forest cover and fire behavior are patchy and decomposition is slow on the Colorado Plateau, deadwood can remain on the ground for centuries or millennia, causing beetle-mediated mortality events to be ecologically influential for long time periods. Although all pieces of deadwood contribute to fuel loading, the largest pieces contribute disproportionally. In a global study, Lutz et al. (2018) found that approximately half the aboveground live biomass was contained in the largest 1% of trees. One of our objectives was to extend this analysis to the deadwood to estimate what proportion of deadwood was from the largest trees. We also wanted to specifically examine the deadwood component contributed by the beetle-killed Picea. Our spatially explicit dataset allowed us to answer the following questions about large-diameter deadwood in Picea–Abies forests of Utah:

  1. 1)

    What is the distribution, spatial orientation, and spatial variation in biomass of standing snags and deadwood ≥ 10-cm diameter?

  2. 2)

    What proportion of the current snag and deadwood biomass is Picea engelmannii?

  3. 3)

    How do spatially explicit calculations of deadwood compare with traditional planar sampling methods?

Materials and methods

Description of study site

The Utah Forest Dynamics Plot (UFDP; Furniss et al. 2017, Bishop et al. 2019) is a 15.32-ha plot subdivided into a 20 m × 20 m grid located at a mean elevation of 3081 m in the mixed subalpine forests of Cedar Breaks National Monument, UT, USA (Cedar Breaks; Fig. 1). The UFDP is characterized by long, cold winters and short growing seasons. Situated above a 750-m deep natural amphitheater, the site exhibits strong local winds year round. Located in the Picea–Abies vegetation zone, the forest has a mixed species complement including widespread Populus tremuloides as well as the regionally rare, but locally abundant, Pinus longaeva (Table 1).

Fig. 1
figure 1

The Utah Forest Dynamics Plot (UFDP) is located in the Colorado Plateau ecoregion of the Intermountain West of North America (a). The high elevation (mean, 3081 m) results in cooler temperatures and higher precipitation than in the lower elevation portions of the Great Basin (b). The UFDP encompasses 15.32 ha (blue outline, of which, 13.64 ha analyzed here) and features a surveyed 20-m grid (c), with planar Brown’s transects measured along red lines (c). Panel a digital elevation model shows shading from the Shuttle Radar Topography Mission (2000); panel c topographic lines are a 5-m digital elevation model derived from LIDAR (2018). Shading in panel b is from a composite of Landsat 8 scenes (2014–2020). The orthophoto in panel c is from USDA National Agricultural Imagery Program (NAIP) data (2018)

Full size image
Table 1 Number of live stems and basal area of woody species ≥ 1 cm DBH in 13.64 ha of the Utah Forest Dynamics Plot in 2019
Full size table

The fire regime of forest types found in the UFDP is one of infrequent (> 100 years median fire return interval) mixed-severity fire (Romme et al. 2009; Heyerdahl et al. 2011). The presence of long-lived (> 1000 years old) Pinus longaeva indicates that there were considerable contiguous areas of low-severity or unburned fire refugia (sensu Kolden et al. 2012, Meddens et al. 2018a) as well as smaller refugia associated with rock outcrops (Kolden et al. 2015; Meddens et al. 2018b; Blomdahl et al. 2019). Dendrochronological analysis of establishment dates for shade-intolerant Pseudotsuga menziesii (Freund et al. 2014) and mass establishments of Abies bifolia, Picea engelmannii, and Populus tremuloides suggest that the last extensive fire in the UFDP was circa 1802. However, given the number of trees older than 200 years, the high-severity portions of the fire must have been patchy. Prior to that, establishment dates of the oldest living Abies bifolia, Picea pungens, and Pseudotsuga menziesii suggest a previous widespread fire in the late 1590s. Dendrochronological analysis of traumatic resin ducts in living Picea engelmannii and lower growth rates of Picea engelmannii versus other tree species indicate that the most extensive bark beetle attack on trees that remained alive in the UFDP was in 1999 (we did not analyze increment cores from trees that died prior to 2018). Other disturbances (aside from the Dendroctonus rufipennis outbreak of the 1990s) include periodic frost defoliation and other species of endemic bark beetles set amid the processes of post-fire succession (i.e., DeRose and Long 2007, 2012a; Birch et al. 2019a, 2019b; Furniss et al. 2020). Dendroclimatological analysis of Pinus longaeva from the UFDP confirms the region-wide extensive drought circa 1250–1345 (Cook et al. 2007). Plant nomenclature follows Flora of North America (1993+).

Data collection

Within the UFDP, all trees, shrubs, and snags ≥ 1 cm DBH were identified, measured, mapped, and tagged in 13.64 ha of the plot from 2014 to 2015 according to the protocols of the Smithsonian Forest Global Earth Observatory network (ForestGEO; Anderson-Teixeira et al. 2015, Lutz 2015, Lutz et al. 2018b). The UFDP was divided into 20 m × 20 m quadrats, with corners surveyed by total stations, and all trees, snags, and deadwood were referenced to the surveyed grid. Subsequent to plot establishment, each of the 23,268 original trees and 4387 snags were visited each year (2016–2019; demographic data in Table S1) to check for mortality and snagfall. New recruitment was also recorded. We measured the diameter and height of each new snag.

Within the UFDP, deadwood originating from the main stems of trees ≥ 10-cm diameter (both newly fallen in 2016–2019 and those pieces on the ground in 2015) was also mapped and assessed for decay class according to the Smithsonian ForestGEO deadwood protocol (Janík et al. 2018). This deadwood protocol, designed for angiosperm and gymnosperm species, facilitates comparison of deadwood among forests worldwide. Snags and deadwood are divided into five different decay classes that are congruent with existing studies of biomass and carbon content of the different decay classes (Harmon et al. 2008, 2013, 2020). See Janík et al. (2018) for protocol details and Lutz et al. (2020) for an example application in temperate forests.

Deadwood was tracked two ways, depending on tree diameter. For trees ≥ 10-cm diameter at the previous measurement, each piece of the main bole ≥ 1-m long was mapped, and pieces shorter than 1 m were ocularly combined with adjacent pieces. Length and diameter at both ends were measured (see Janík et al. 2018 for measurement details). Trees were considered to be fully “fallen” if the remaining stump height was < 1.37-m high. Otherwise, deadwood was mapped as the “broken top” of a standing snag. For trees and shrubs 1 cm ≤ DBH < 10 cm, the year of snagfall was recorded, but the smaller stems were not mapped and were assumed to have fallen as a single piece.

Distribution, volume, and biomass calculations

The biomass of live trees was calculated using equations selected from Chojnacky et al. (2014) and other sources (Brown 1976; Brown 1978; Halpern and Lutz 2013; Jenkins et al. 2003; Means et al. 1994; see Table S2 for listing of equations). We examined all available allometric equations that we could find for species within the UFDP and morphologically similar species, and we selected those that most closely represented growth form of trees at this high-elevation site. This was particularly important for pines (i.e., Brown 1978), as genus-level biomass equations were developed to perform well for large-statured Pinus species (such as Pinus ponderosa), and this would grossly overestimate biomass for short-statured Pinus in the UFDP (sensu Lutz et al. 2017a, 2017b). We verified the biomass values calculated by our selected equations with biomass estimates derived from volume and density calculations made from field measurements of intact deadwood (Fig. S1A). Many allometric equations for tree species have lower diameter limits (often ca. 3 cm DBH) that do not encompass our full data range, leading to very low (or even negative) biomass estimates for trees 1 cm ≤ DBH < 3 cm. We accordingly set minimum biomass values for small-diameter stems of the most common tree species as follows: Populus tremuloides and Sambucus racemosa—250 g; Abies, Juniperus, Pseudotsuga, Pinus edulis, and P. ponderosa—300 g; Picea engelmannii and Picea pungens—400 g; and Pinus longaeva and Pinus flexilis—500 g, and other shrub species—100 g (Table S2, Figure S1B). With the selected equations, we calculated biomass and determined the large-diameter threshold—the diameter cutoff that divides aboveground live biomass into two equal portions—a large number of small-diameter trees constituting half the biomass, and a much smaller number of large-diameter trees constituting the other half (sensu Lutz et al. 2018).

For snags that died relatively recently (decay class 1 or 2) and were still intact (first-order branches present, bark intact, and top diameter ≤ 10 cm), we estimated biomass with live tree equations and corrected with the species- and decay-class-specific relative density factors of Harmon et al. (2008). For snags that were in an advanced state of decay or were not intact, their biomass was estimated by modeling their volume as a conic frustum and using species and decay-class-specific density values from Harmon et al. (2008).

For deadwood, biomass was calculated for the entire originating tree and attributed to specific pieces that fell from that tree. This was done by weighting a piece according to its volume as a proportion of the total volume of the tree, including any remaining standing volume. Biomass was then corrected following the methods of Lutz et al. (2020). For pieces that could not be identified, we used an average density value calculated by proportionally weighting the identified pieces from the UFDP data (0.374 g cm−3 for decay class 1; 0.299 g cm−3 for decay class 2; 0.255 g cm−3 for decay class 3; 0.154 g cm−3 for decay class 4; and 0.139 g cm−3 for decay class 5).

We divided trees, snags, and deadwood into diameter classes (1 cm ≤ DBH < 10 cm, 10 cm ≤ DBH < 33.2 cm, ≥ 33.2 cm, and 20 m × 20 m quadrats) to examine the effects of the Dendroctonus rufipennis outbreak by considering the relative difference between biomass of susceptible (i.e., Picea engelmannii) and unsusceptible (i.e., all other) species. We also assessed the fall direction of the deadwood from the endpoint locations of the longest piece as this can reflect site-level factors driving the spatial arrangement of deadwood including wind direction and topography (Lutz et al. 2020).

Comparison with traditional deadwood sampling methods

In addition to the explicit mapping deadwood ≥ 10-cm diameter, we installed 1580 m of planar transects (Brown’s transects) consisting of 79, 20-m transects established between permanent markers (Fig. 1; Brown 1974). We measured surface fuels in four size classes (1 h, less than 0.25 in. (0.64 cm) in diameter; 10 h, between 0.25 and 1 in. (between 0.64 and 2.54 cm) in diameter; 100 h, 1 to 3 in. (2.54 to 7.62 cm) in diameter; and 1000 h, greater than 3 in. (7.62 cm) in diameter) as well as litter and duff at ten points on each transect (odd meters from 1 to 19 m). These transects, running between the permanent quadrat markers, were sampled in 2019 (Fig. 1). We calculated fuel loading by the methods of Brown (1974) as implemented by Cansler et al. (2019) with additional species data from Woodall and Monleon (2010).

We used a novel method to quantify a portion of the vertical structure of the fuel bed based on the deadwood map. We used the position of individual deadwood polygons (derived from field measurements of large-end and small-end diameters and length; Janík et al. 2018) to develop a “heat-map” depicting the number of intersecting logs which reflected the amount of stacked wood. This provides a map of the three-dimensional structure of surface fuels that may serve as a useful compromise between typical fuel sampling methods (Brown’s transects) and more accurate (but expensive and labor intensive) techniques such as ground-based LiDAR. We used the st_intersects() function from the sf package (Pebesma 2018) to count the number of pieces that intersected each piece of deadwood. For the background heatmap, we used the stat_density2d() function from ggplot2 (Wickhan 2016) to perform a kernel density estimation on a point layer generated by placing points at the intersection of two or more pieces of deadwood.

All calculations were performed in R version 3.6.2 (R Core Team 2020) using packages rgdal version 1.4-8 (Bivand et al. 2019), rgeos version 0.5-2 (Bivand and Rundel 2019), sp version 1.3-2 (Pebesma and Bivand 2005), and sf version 0.8-1 (Pebesma 2018).

Results

Distribution and spatial orientation of snags and deadwood

Live tree densities showed both considerable survival of large-diameter Picea engelmannii through the period of the Dendroctonus rufipennis outbreak as well as significant regeneration (Table 1). In 2019, there were 23,598 trees and 4712 snags present (Tables S1, S2). Considering only stems ≥ 10 cm DBH (11,452), 25% were snags (2505). Tabulation of biomass using site-specific equations showed that the large-diameter threshold was 33.2 cm, smaller than the 41.9 cm threshold derived using generic equations for North America (i.e., Chojnacky et al. 2014; Lutz et al. 2018). Total aboveground live biomass was 1624.6 Mg (119.1 Mg ha−1). There was a total of 4182 pieces of deadwood originating from 3776 different trees ≥ 10-cm diameter in 13.64 ha (Fig. 2, Table S3). Picea engelmannii represented 17.4% of deadwood biomass (Table S3), and only 5.7% of live basal area (Table 1).

Fig. 2
figure 2

Annual accumulation of deadwood ≥ 10-cm diameter in 13.64 ha of the Utah Forest Dynamics Plot, as well as deadwood present at the first census in 2015. For annual accumulations by species, see Table S2 and Figs. S2, S3, S4, S5, S6, S7

Full size image

Not considering 20 m × 20 m quadrats where deadwood was absent (generally on the rocky cliffs; Fig. 1), deadwood in 20 m × 20 m quadrats varied from 0.05 to 154.0 Mg ha−1 (mean of 41.8 Mg ha−1). Picea engelmannii deadwood biomass in each 20 m × 20 m quadrat varied from 0.003 to 105.5 Mg ha−1 (mean of 7.3 Mg ha−1; Fig. 3). Deadwood ≥ 10 cm did not exhibit strong directionality (Fig. 4a). However, larger diameter deadwood generally fell in northerly directions (Fig. 4b). This directionality most likely reflects the channeling of strong gusts by the local terrain (Fig. 1c). The mean amount of surface fuel measured by planar transects was 62.7 Mg ha−1, but the large wood component (1000 h fuels and larger; ≥ 7.62-cm diameter) was only 31.8 Mg ha−1, 75% of the deadwood mass estimated by explicit measurements.

Fig. 3
figure 3

Spatial variation of deadwood in the Utah Forest Dynamics Plot at the scale of 20 m × 20 m quadrats. Most trees fell as single pieces. Biomass of Picea engelmannii (PIEN) exceeded that of many other species, but the mean Picea engelmannii biomass was about 20% of that of other species combined

Full size image
Fig. 4
figure 4

Directional diagrams of deadwood ≥ 10-cm diameter (a) and large-diameter (≥33.2 cm DBH) deadwood (b) in the Utah Forest Dynamics Plot. Each concentric gridline of the rose diagram represents 2% of the total pieces

Full size image

Annual deposition of deadwood was low compared to the total deadwood pool (Fig. 5). Annual deposition was dominated by smaller diameter classes, with only a few large-diameter trees and snags transitioning to deadwood in each year (Tables S1 and S3, Fig. 5).

Fig. 5
figure 5

Diameter distribution of deadwood deposition in the Utah Forest Dynamics Plot in 2016–2019. Pre-establishment refers to deadwood from 2015 and before. Colors correspond to those in Fig. 2. Curves are based on the raw diameter distributions smoothed with a Gaussian density kernel. The square-root transformation of the y-axis shows the low frequency but cumulative effects of large-diameter inputs

Full size image

Contribution of Picea engelmannii

The distribution of Picea engelmannii aboveground biomass among live, snag, and deadwood pools differed from that of other species (Figs. 6, S8). For Picea engelmannii, live, snag, and deadwood biomass were approximately equal (7.9 Mg ha−1, 9.8 Mg ha−1, and 7.4 Mg ha−1, respectively; Fig. 6). For other species as a group, live biomass was about four times snag biomass (111.2 Mg ha−1 vs. 26.0 Mg ha−1), with deadwood biomass slightly exceeding snag biomass at 34.8 Mg ha−1. Compared to other species as a group, Picea engelmannii also showed a higher concentration of biomass in the large diameter class among trees, snags, and deadwood (Fig. 6).

Fig. 6
figure 6

Division of live tree, snag, and deadwood biomass pools in the Utah Forest Dynamics Plot by diameter class and by susceptibility to Dendroctonus rufipennis (i.e., Picea engelmannii vs. all other species). The diameter division at 33.2 cm DBH separates living biomass into two equal portions, with 1212 trees ≥ 33.2 cm DBH and 22,388 trees 1 cm ≥ DBH > 33.2 cm. Total live biomass was 119 Mg ha−1, total snag biomass was 36 Mg ha−1, and deadwood biomass was also 42 Mg ha−1. Wood identified as Picea but without designation of species was counted as Picea engelmannii

Full size image

Comparison with traditional planar fuel sampling

The traditional measurements of large-diameter fuels (Fig. 7d) were significantly less (31.8 Mg ha−1) than estimates from spatially explicit measurements (41.8 Mg ha−1). Analysis of stacked deadwood showed areas of high vertical fuel structure (Figs. 7 and 8). These areas were usually associated with Picea-Abies-Populus forest communities.

Fig. 7
figure 7

Fuel loading by fuel category as measured in 79, 20-m planar transects (Brown’s transects) in the Utah Forest Dynamics Plot in 2019. CWD–1000-h fuels or larger, defined as ≥ 3 in. (7.63 cm) in diameter at the point of line transect interception

Full size image
Fig. 8
figure 8

Vertical structure of surface fuels ≥ 10 cm diameter in the Utah Forest Dynamics Plot. Deadwood was locally variable in terms of simple biomass (Fig. 2) but also in terms of vertical structure (a). Colors indicate number of pieces of wood “stacked” on each other. The degree of vertical agglomeration was transformed into an intensity diagram (b), with white areas showing no overlap between pieces of deadwood and warmer colors indicating more stacking

Full size image

Discussion

Large-diameter (≥ 33.2-cm diameter) snags and deadwood were the principal components of snag and deadwood biomass despite representing only 11% and 16% of pieces, respectively (Fig. 6). Irrespective of species, the concentration of snags and deadwood in the large diameter classes reinforces the ecological significance of large-diameter trees even beyond their death (Erb et al. 2016; Körner 2017; Lutz et al. 2018, 2020). The low levels of biomass and productivity of the UFDP compared to temperate forests globally (i.e., Larson et al. 2008; Lutz et al. 2018) and the slower treefall dynamics of beetle mortality as opposed to fire mortality have resulted in a slow development of deadwood in this Picea–Abies forest (Fig. 5).

Within the UFDP, trees killed in the Dendroctonus rufipennis outbreak have not yet become a major contributor to deadwood due to the diverse species assemblage (and hence lower pre-outbreak Picea engelmannii relative abundance; Table 1; Fig. S8). Heterogeneous mixed species stands such as the UFDP appear to have higher resistance to bark beetle outbreaks (Churchill et al. 2013; DeRose and Long 2014). The continued transition of Picea engelmannii snags into deadwood will represent a growing, but still relatively minor component of the overall deadwood within the UFDP. But as drought can hasten the return of both epidemic bark beetle outbreaks and fire (Hart et al. 2014, Sherriff et al. 2001), future climates may see increased mortality (i.e., Lutz et al. 2009) and additional pulses of deadwood.

Fire and epidemic outbreaks of Dendroctonus rufipennis represent two of the principal ecological processes contributing to mortality and deadwood deposition in subalpine spruce-fir forests on the Colorado Plateau (Veblen et al. 1994). As most bark beetles in these forests are specialists on particular host tree species, the ecological effects of bark beetles on forest composition, snags, and deadwood will depend on the prevalence of the host species. In comparison to nearby pure stands of Picea engelmannii that experienced nearly complete overstory mortality (93%; DeRose and Long 2007), the UFDP exhibited considerable survivorship and regeneration (Table 1), similar to that seen in mixed-wood forests in other regions that have experienced epidemic Dendroctonus rufipennis outbreaks (Birch et al. 2019a, 2019b; DeRose and Long 2007). The persistence of living large-diameter Picea engelmannii contributes substantially to the overall structural heterogeneity of the forest (sensu Lutz et al. 2013). Standing dead and down Picea engelmannii will also contribute to structural heterogeneity and ecological function for years to come, as long fire return intervals and slow decomposition allow this dead biomass to persist for centuries.

In general, both bark beetle epidemics and fire occur on multi-centennial timescales with local epidemic outbreaks of Dendroctonus rufipennis reoccurring over intervals of approximately 116.5 years and high-severity fire returning every 149–212 years (USDA 2012; Veblen et al. 1994). The data from the UFDP is broadly in line with these estimates, although at the higher end of the range of disturbance return intervals. However, mean return intervals are perhaps less important than the particular sequence of disturbances, their intensities, and intervals between them (Veblen et al. 1994; Becker and Lutz 2016).

Value of explicitly mapping horizontal and vertical deadwood structure

The spatial heterogeneity of deadwood in the UFDP (Figs. 2, 5, S8) reinforces the importance of modernizing surface fuel field methods (Lutz et al. 2018b), especially when elevated log positions can increase deadwood persistence (Körner 2017). Large-diameter deadwood was a major constituent of fuel stacks, both because of longer lengths and also, presumably, because of slower decomposition rates for deadwood supported off the ground. Actual heterogeneity in surface fuels becomes unexplained variance when using planar sampling methods such as Brown’s transects, and this variance is ignored when multiple transects are averaged to estimate surface fuel loads among fuel size classes. While planar transects are a useful approximation of the average amount of surface fuels, two- and three-dimensional heterogeneity in surface fuels (Fig. 6b) are fundamental to inferring potential fire behavior (Pimont et al. 2016), especially in small areas where fire behavior may be extreme. This three-dimensional approach to surface fuels may also portend a more efficient way for managers to reduce surface fuels to mitigate fire risk. By understanding the heterogeneity in fuel structure that drives locally high potential flame lengths, fuel reduction, or prescribed fire treatments could focus more resources on less area, potentially maximizing the efficiency of limited time and resources that are available to mitigate fire risk. It also demonstrates that in heterogeneous forests such as this, more targeted salvage logging operations could extract a considerable proportion of available timber while only visiting a small fraction of a stand. This may provide a way to find compromise between the diverse, and often divergent, priorities of local stakeholders.

Conclusions

Tree mortality and subsequent snagfall have profound consequences for both snag and deadwood biomass, with large-diameter trees dominating biomass dynamics even beyond their death. Outbreaks of species-specific tree mortality agents (such as bark beetles) may have both immediate and delayed effect on deadwood, with the transition to deadwood being quite slow in subalpine environments, especially for large-diameter trees. Lengthy intervals between fires can allow deadwood to accumulate, evident in the high stand-level averages with a long right tail representing exceptionally high amounts of deadwood at local levels. These local maxima contribute diversity to the spatial heterogeneity in habitat and surface fuels, and the vertical stacking of pieces increases airflow which may potentially slow decomposition and lead to longer persistence times.