Forest Ecology and Management 434 (2019) 289–302

Contents lists available at ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier.com/locate/foreco

Spatial dynamics of tree group and gap structure in an old-growth ponderosa pine- black oak forest burned by repeated wildfires T ⁎ Natalie C. Pawlikowskia, Michelle Coppolettab, Eric Knappc, Alan H. Taylora,d, a Department of Geography, The Pennsylvania State University, University Park 16803, b United States Department of Agriculture, Forest Service, 159 Lawrence Street Quincy, CA 95971, United States c United States Department of Agriculture, Pacific Southwest Research Station, 3644 Avtech Drive, Redding, CA, United States d Department of Geography, Earth and Environmental Systems Institute, The Pennsylvania State University, University Park 16803, United States

ARTICLE INFO ABSTRACT

Keywords: Knowledge of how tree groups and gaps are formed and maintained in frequent-fire forests is key to managing Spatial pattern for heterogeneous and resilient forest conditions. This research quantifies changes in tree group and gap spatial Canopy gap structure and abundance of ponderosa pine (Pinus ponderosa) and California black oak (Quercus kelloggii) with Forest dynamics stand development after wildfires in 1990 and 1994 in an old-growth forest in the Wilderness, southern Resilience Cascades, California. Forest demography and tree group and gap structure were quantified by measuring, Pinus ponderosa mapping, and aging trees in six 1-ha permanent plots in 2000 and 2016. Tree recruitment, mortality, and growth Quercus kelloggii fi Wilderness were estimated using demographic models and spatial characteristics including gap structure were identi ed Restoration using an inter-tree distance algorithm and the empty space function. Potential fire behavior and effects in 2016 Recruitment were estimated to determine if the current forest would be resilient to a wildfire in the near future. Tree mortality Stand density and basal area in both 2000 and 2016 resembled reference conditions for pre-fire suppression frequent-fire forests in the western United States. Wildfires initially promoted California black oak regeneration via sprouting, but oak regeneration from seed declined relative to ponderosa pine over the post-fire period. In 2000, ∼15% of trees were classified as single tree groups and an average tree group had 6 trees (range 2–38) which increased to 9 trees (range 2–240) in 2016. Small groups (2–4 trees) had similar-aged trees while larger groups were multi-aged. By 2016, single tree groups decreased by ∼30%, and the size, density, and intensity of clustering within tree groups increased, with an average tree group size of 9 (range 2–240) in 2016. Rates of post-fire regeneration, particularly the high rate for ponderosa pine, drove spatial dynamics in tree group and gap structure. Although the size and frequency of canopy gaps were similar in 2000 and 2016, the density of seedlings and saplings in gaps was higher in 2016, and large gaps were being fragmented by gap filling. Potential fire behavior predicted surface fire and low overstory tree mortality, suggesting the current forest would be resilient to a wildfire. However, burning will be necessary in the future to reduce the demographic pressure of ponderosa pine, promote black oak, and to maintain and create future spatial heterogeneity. Cumulatively, this research demonstrates that wildland fires under certain conditions can maintain and restore fire resilience in ponderosa pine forests reducing the negative ecological consequences related to past fire exclusion.

1. Introduction resilience to chronic fire (Churchill et al., 2013). Under a regime of frequent fire, a stand, at any one point in time, consists of a mosaic of Ponderosa pine (Pinus ponderosa) and mixed pine-oak forests in the groups of even/uneven aged overstory trees, widely spaced individual western USA evolved with frequent fire and were historically resilient trees, dense patches of regeneration, and unoccupied canopy gaps to fire. In this context, resilience represents a capacity for forests to (Cooper, 1960; Larson and Churchill, 2012; White, 1985). This group reorganize and return to a prior condition after disturbance while ab- and gap structure promotes understory species and lifeform diversity sorbing environmental changes such as climate change (Folke et al., (Dodson et al., 2008; Laughlin et al., 2011; North et al., 2005) and 2010; Millar et al., 2007). A characteristic feature of ponderosa pine provides a range of habitats that supports a diverse assemblage of bird forests is structural heterogeneity at small-scales which imparts and small mammal species (Buchanan et al., 2003; Latif et al., 2015). A

⁎ Corresponding author at: Department of Geography, The Pennsylvania State University, University Park 16802, United States. E-mail addresses: [email protected] (M. Coppoletta), [email protected] (E. Knapp), [email protected] (A.H. Taylor). https://doi.org/10.1016/j.foreco.2018.12.016 Received 10 October 2018; Received in revised form 30 November 2018; Accepted 9 December 2018 Available online 20 December 2018 0378-1127/ © 2018 Elsevier B.V. All rights reserved. N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302 group and gap structure also creates spatially heterogeneous surface USDI, 1998), and there evidence that reintroduction of fire has suc- and canopy fuel conditions which burn with highly variable effects that cessfully restored presettlement conditions exists for select con- inhibit spread of intense surface or crown fire while providing protected temporary forests with restored fire regimes and limited anthropogenic sites for tree regeneration (Fry and Stephens, 2010; Taylor et al., 2014). disturbance (Larson et al., 2013; Parks et al., 2015; Harris and Taylor, The structure and resilience of frequent-fire ponderosa pine and 2015). Yet, there is also increasing concern and mounting evidence that pine-oak forests in the western USA was altered significantly by Euro- in areas where forest structure has been dramatically altered, re- American settlement in the mid to late 19th century (Covington and introducing wildfire could exacerbate the problem (Thompson and Moore, 1994; Taylor, 2004). Logging, grazing, and especially fire ex- Spies, 2010, Crotteau et al., 2013, Coppoletta et al., 2016, Harris and clusion led to an increase in tree density and forest canopy cover, re- Taylor, 2017). The initial effects of a re-entry burn (the first fire fol- duction in gap area by infilling of trees, and an increase in both surface lowing a period of fire exclusion) strongly influence fire effects in and canopy fuel (Knapp et al., 2013; Scholl and Taylor, 2010; Taylor subsequent reburns (Harris and Taylor, 2017; Parks et al., 2015; Walker et al., 2014). In logged areas, the largest and most fire-resistant trees et al., 2018). High-severity fire effects in recent re-entry burns have led were often removed (Laudenslayer et al., 1989; Naficy et al., 2010) to high-severity fire effects in reburns that can initiate long-term shifts resulting in dense stands of intermediate and smaller sized trees. Fire from forest to vegetation types such as grassland or shrubland exclusion also shifted species composition of pine-dominated and mixed (Coppoletta et al., 2016; Guiterman et al., 2018; Savage and Mast, pine-oak forests toward conifer dominance, and oaks have declined 2005). In more anthropogenically-disturbed, fire-suppressed areas, ac- where they have been overtopped by a dense conifer canopy (Cocking tive management – such as prescribed burning or mechanical thinning – et al., 2012, 2014; McDonald and Tappeiner, 2002). Fire-exclusion re- can help restore a negative feedback and may be needed before re-in- lated forest changes have increased susceptibility of ponderosa pine and troducing wildland fire back on the landscape (Ritchie et al., 2007; pine-oak forests to high-severity fire, insect attacks, drought-induced Harris and Taylor, 2017). tree mortality, and climate change (Guarín and Taylor, 2005; Hurteau Contemporary reference sites that can guide active management of et al., 2014; Preisler et al., 2017). Fire exclusion, logging and other western forests are rare (Stephens and Gill, 2005; Larson et al., 2013). forest management practices are still prevalent in western landscapes These sites are valuable, since they allow for quantitative measures of with millions of hectares in need of restoration (Churchill et al., 2013; forest structure (Barbour et al., 2009) and provide important insights Hessburg et al., 2015). Creation of forest structures that are both re- into healthy ecosystem processes (Binkley et al., 2007; Stephens et al., sistant and resilient to fire is central to restoration goals in frequent-fire 2008). The Beaver Creek Pinery, an old-growth forest in the Ishi forests throughout western North America (e.g., Gaines et al., 2010; Wilderness, may serve as an important contemporary reference site for Roccaforte et al., 2010). The basis for manipulating forest structure to ponderosa pine-black oak forests. Due to limited anthropogenic dis- restore resilience is a documented shift away from historic (pre-Euro- turbances in the 20th century and a history of frequent, low-moderate American settlement) forest structure and composition and the in- severity fire, this forest was resilient to two wildfires in the 1990s. creased susceptibility of forests to high-severity fire that can shift forest Further, these fires preserved a resilient forest structure by thinning to an alternative stable state of grassland or shrubland through repeated ingrowth resulting from fire exclusion, promoting tree regeneration, high-severity fire (Airey Lauvaux et al., 2016; Coppoletta et al., 2016; and generating spatially heterogenous tree patterns thought to be si- Reily et al., 2017). milar to forests before fire suppression (Taylor, 2010). Here, we char- Before fire suppression, interactions between frequent fire and stand acterize changes in this resilient tree group and gap structure over a 22- structure maintained a gap and tree group structure in mixed ponderosa year period. pine-California black oak (Quercus kelloggii, hereafter black oak) forests Specifically, our objectives were to: (1) quantify post-fire tree re- in lower montane forests in the , southern Cascade, and cruitment, mortality, and growth for ponderosa pine and black oak Klamath Mountains in California (Skinner et al., 2018a; Skinner et al., during stand development; (2) quantify the spatial and structural at- 2018b; van Wagtendonk et al., 2018). Ponderosa pine and black oak are tributes of tree groups and gaps and how they have changed since the both well adapted to wildfire. Ponderosa pine has thick insulating bark, wildfires; (3) determine if historic tree group development was pre- is fire-resistant, and regenerates from seed; black oak has a thinner dominantly even or uneven aged; and (4) evaluate whether the current bark, sprouts from the base and crown after being injured or top-killed forest structure is resilient to fire by estimating potential fire behavior by fire, and also regenerates from seed (Barton, 2002; Hammett et al., and effects. To fulfill our objectives, we used repeat measurements of 2017; Nemens et al., 2018). Mixtures of pine and black oak within a forest characteristic in six large (∼1 ha) permanent plots that were stand are thought to increase forest resilience to fire (Skinner, 1995). sampled in 2000 and 2016. Black oak litter is highly flammable and promotes rapidly spreading, low-intensity, surface fire (Engber and Varner, 2018), and black oak in 2. Method mixed stands breaks up crown fuel continuity thereby reducing crown fire spread (Skinner et al., 2018a). The tree group and gap structure is 2.1. Study site also thought to be important for regeneration and persistence of black oak, particularly large diameter trees, as establishment and growth of Our study was conducted in the Beaver Creek Pinery, an ∼100 ha this species benefit from high-light environments (Long et al., 2016). forest on a plateau in the Ishi Wilderness, in the southern Cascades, With fire exclusion, ponderosa pine has increased and black oak has California (Fig. 1). The elevation of the Beaver Creek Pinery is 850 m. declined, altering fire-vegetation interactions and the potential to re- Ponderosa pine and black oak are the canopy dominants. Common store a resilient group and gap structure in highly affected forests understory shrubs include Arctostaphylos manzanita subsp roofii, A. (Cocking et al., 2014, 2012; Nemens et al., 2018). viscida, Ceanothus integerrimus, C. prostratus, C. lemmonii, Frangula cali- Managed wildfire has been proposed as a key restoration tool to fornica, Mahonia aquifolium, and Toxicodendron diversilobum. The Beaver increase structural heterogeneity and fragment fuels in frequent-fire Creek Pinery is underlain by Cohasset gravelly loam soil (USDA Web forests altered by fire suppression (Larson et al., 2013; Naficy et al., Soil Survey) and average annual temperature and precipitation in Co- 2010; North et al., 2012). Use of managed wildfire is particularly ap- hasset (800 m, 20 km southwest) is 16 °C and 145 cm, respectively. propriate in large remote tracts of forest where use of mechanical fuel Most precipitation occurs between October-April. The Yana (Yahi) tribe treatment to alter fire behavior is constrained by cost, access, terrain, lived in the region for 3000 years prior to European settlement in the and legal or regulatory requirements (Larson et al., 2013; North et al., mid-19th century, and Ishi, the last tribal member living in what is now 2015). In designated wilderness, fire (prescribed, managed wildfire) is the Ishi wilderness, emerged from the area in 1911 (Kroeber, 1961). the primary tool available for vegetation and fuels management (USDA- The Yana used fire to promote plant growth for food, fiber, and game

290 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

mapped new tree recruitment, and recorded tree mortality. In both sample years, we counted saplings (< 5 cm dbh and > 1.4 m tall) and seedlings (0.2–1.4 m tall) in 10 m × 10 m grid cells in each plot, and recorded if oaks had established from seed or sprout. Oak regeneration was assumed to be from seed unless stems were highly aggregated or in close proximity to the remains of a top-killed oak. We also estimated canopy cover above each cell (open 0–33%, intermediate 33–66%, closed > 66%) and estimated ground cover of shrubs, forbs, and grasses (0%, < 1%, 1–5%, 5–25%, 25–50%, 50–75%, and 75–100%).

2.3. Stand characteristics and development

To characterize post-fire stand development and tree population dynamics, we estimated rates of recruitment, mortality, and loss/gain of basal area for ponderosa pine and black oak between 2000 and 2016. Recruitment and mortality were estimated following Condit et al., 1995, 1993: Ln ()N − Ln (N ) mortality rate (%year −1 ) = b s × 100 t

Ln ()N − Ln (N ) recruitment rate (%year −1 ) = e s × 100 t

where Nb,Ne, and Ns are the number of trees (≥5 cm DBH) alive in 2000, total number of trees living in 2016, and the number trees sur- viving between 2000 and 20016, respectively; t is the time interval (16 years). Basal area (BA) lost or gained by recruitment or mortality was identified using the following equations: (BA /BA ) BAloss rate (%year −1) = l b × 100 Fig. 1. Location of the Beaver Creek Pinery in in the Ishi t Wilderness, within (Top inset), the location of the six stem map plots (Top and Bottom) measured in 2000 and 2016 and severity of ((BAr + BAg)/BAb ) BAgain rate (%year −1 ) = × 100 the 1990 Campbell fire (Top) and 1994 Barkley fire (Bottom) as identified by t the Monitoring Trends in Burn Severity Program (http://www.mtbs.gov/). where BAl,BAb,BAr and BAg are the basal area values for trees that died during the census period, trees alive at the beginning of the study, trees (Anderson, 2005). Burning was used, in particular, to promote black recruited during the census period, and trees that survived during the oak growth and acorn production, and to improve rates of acorn col- census period, respectively. lection (Long et al., 2016). The Beaver Creek Pinery was never logged Changes in the populations for trees and BA were inferred from the and became a designated Wilderness Area within the Lassen National population growth rate (λ)(Condit et al., 1999) using data for both Forest in 1984. sampling dates: Fire was a frequent disturbance in the Beaver Creek Pinery before fi re suppression was imposed in Lassen National Forest in 1905. The −1 Ln (N2016) − Ln (N2000 ) λtrees (%year ) = × 100 mean and median point fire return interval in the Beaver Creek Pinery t were 16- and 12-years, respectively, and burns occurred mainly late in fi −1 (BA2016 − BA2000)/BA2000 the growing season (Taylor, 2010). Four res burned in the Beaver λBA (%year ) = × 100 Creek Pinery in the 20th century, specifically in 1901, 1924, 1990 and t 1994; the most recent burns in 1990 (Campbell) and 1994 (Barkley) where N and BA are the number of trees or the basal area of species at burned at predominantly low- to moderate-severity with some mod- each time point. erate- to high-severity patches in the 1990 Campbell Fire (Fig. 1). These Comparisons in density, basal area, and rates of recruitment and fi res killed mainly smaller diameter trees and created canopy gaps fa- mortality between species over the census period were made using vorable for ponderosa pine and black oak regeneration (Taylor, 2010). paired t-tests (density, basal area) or non-parametric Kruskal-Wallis H tests (recruitment, mortality). We used a relaxed probability of sig- 2.2. Field sampling nificance (p = 0.1) for comparisons because the number of plots was small (n = 6). Post-fire stand development was studied in six large (0.8–1.0 ha) The influence of density and distance from focal tree to competitors plots established in 2000, six years after the last fire. Plot locations were on tree mortality was identified by calculating the H index (Hegyi, selected based on the presence of large diameter (≥1.0 m diameter at 1974). This tree-to-tree index assumes that relative dbh and distance breast height [dbh]) trees, similar site conditions, and stand structure between the focal and competitor tree(s) describes their competitive that captured the range of variation in old-growth structure in the interaction and is commonly used to assess competition where spatial Beaver Creek Pinery. Plots were re-sampled in 2016, 22 years post-fire. data for each tree is recorded (Contreras et al., 2011; Vernon et al., In 2000, trees (≥5 cm dbh) within each plot were mapped and all 2018). Competition was assessed for three populations: trees alive in living trees were aged by coring them at 0.3 m above the ground, cross- 2000 that survived to 2016, trees alive in 2000 that did not survive to dating the tree-rings and using the date of the inner most tree ring as 2016, and trees alive in 2016. The H index was calculated using the the tree age (Taylor, 2010). In 2016, we re-measured live trees (dbh), following equation:

291 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Dj 2.4.2. Gap characteristics n D ∑ i Gap characteristics (frequency, area) in 2000 and 2016 were j=1 Dist ij quantified using the empty space function F(t) to calculate distances to nearest tree (Diggle, 2013) from each cell in a 0.5 m × 0.5 m grid in where D is the diameter of the competitor, D is the diameter of the j i each plot. Distances were pooled in 3 m classes to create a gap size and focal tree, and Dist is the distance between the focal and competitor ij total open canopy area distribution for 2000 and 2016. We tested for tree. Competitor trees were defined as any tree within 12 m of the focal differences in the shape of these distributions using a Kolmogorov- tree and we applied a 12 m buffer to avoid artificially low competition Smirnov two-sample test. Gaps were identified using a threshold dis- indices for trees near the plot edges. Competition effects (measured by tance of 9 m from a tree stem, as suggested by Churchill et al. (2017) the H index) on trees that survived the census period vs trees that died and Clyatt et al. (2016), since this distance is beyond the influence of during the census period were compared using a t-test. Competition was the canopy and tree root systems and should permit successful pon- also compared for trees that survived the census period. derosa pine regeneration and growth (Simard et al., 2012; York et al., 2004). We followed Clyatt et al. (2016) in quantifying gaps where gaps were ≥9 m from the nearest tree and then buffered by 5 m to account 2.4. Spatial patterns and dynamics for tree crown extension into openings. Areas 9 m or closer to plot edges were also removed to prevent overestimation of gap area. Open area in 2.4.1. Group characteristics each plot for each sample year was calculated by subtracting crown Tree group and gap structure and dynamics were determined by area from plot area. Differences in mean gap size and open area in 2000 quantifying the spatial structure of tree populations within plots in and 2016 were identified using separate paired t-tests and a relaxed 2000 and 2016. Tree group characteristics were identified using an probability of significance (p = 0.1) because the number of plots was inter-tree distance (d) algorithm over a range of spatial scales. Trees small (n = 6). were assigned to a group within a given radius ≤d, while trees at Gap dynamics were identified by estimating rates of crown exten- distances > d were considered single tree groups (Plotkin et al., 2002). sion by the overstory, gap infilling from newly established trees, and Since edge effects have been shown to be minimal in identifying gap creation from tree mortality with the following equations: clumping patterns (Churchill et al., 2013; Plotkin et al., 2002; Yamada and Rogerson, 2003), an edge correction was not used. Tree group sizes (Gap − Gap )/Gap infilling (%year −1) = 2000 both 2000 × 100 were identified for five functional categories for ponderosa pine forests regen t identified by Churchill et al. (2017): (1) individual (1 tree); (2) small – – – (Gap − Gap )/Gap (2 4 trees); (3) medium (5 9 trees); (4) large (10 15 trees); and (5) creation (%year −1 ) = 2016 both 2016 × 100 extra-large clump (≥16 trees) at inter-tree distances of 1–12 m. We t compared tree group frequency for both 2000 and 2016 using a 6-m (CD − CD )/CD inter-tree distance, a distance that corresponds with the average crown crown extension (%year −1) = 2016 2000 2000 × 100 t radius of trees in mature and old growth ponderosa pine forests (Abella and Denton, 2009; Churchill et al., 2013; Sánchez Meador et al., 2011). where Gap2000 is gap area in 2000, Gapboth is the gap area that re- Characterization at this distance also permitted direct comparison with mained open between 2000 and 2016, Gap2016 is gap area in 2016, group characteristics in ponderosa pine forests in the Pacific Northwest CD2000 is the portion crown diameters from trees in 2000 extending into (Churchill et al., 2013, 2016, 2017), the Southwest (Abella and Denton, Gapboth (i.e. unoccupied gap area), CD2016 is overlap between Gapboth 2009; Sánchez Meador et al., 2011, 2009; Schneider et al., 2016) and and 2016 crown diameters of surviving trees, and t is the time interval the Rocky Mountains (Brown et al., 2015; Clyatt et al., 2016). (16 years). Crown diameter for this analysis was estimated using spe- In addition to quantifying tree group spatial characteristics using cies-specific diameter-crown width equations (Keyser, 2008) calibrated local spatial metrics, they were also quantified using a global statistic, with field measurements. Ripley’s K. Ripley’s K is based on point pattern analysis, and we used the standardized L-function to identify the type, scale, and intensity of 2.4.3. Regeneration and gap characteristics ’ pattern for trees in 2000 and 2016. Ripley s L-function was calculated The relationship between canopy cover and abundance of tree re- for trees in each plot using the Spatstat package (Baddeley and Turner, generation was identified by comparing seedling and sapling density in 2005) in R version 3.3.2 (R Core Team, 2013) and spatial patterns were 10 m × 10 m cell canopy cover classes in 2000 and 2016. Comparisons assessed at 1 m intervals up to a maximum radius (r) of one quarter the were made using a non-parametric Kruskal-Wallis test with a Dunn’s length of the shortest plot dimension. An envelope of complete spatial post-hoc test. Infilling of regeneration into gaps between 2000 and randomness (CSR) was generated using 999 simulations and values of L 2016 was also graphically assessed by mapping regeneration density of (r) above the envelope denote clumping, values below signify a uniform ponderosa pine and black oak in each cell and calculating the area pattern, values within the envelope represent a random pattern. Ana- occupied by density class in each plot. Changes in seedling and sapling lyses were performed for all six plots for populations of at least 15 trees abundance were identified using paired t-tests and a relaxed probability in the following categories: (1) all trees alive in 2000 and 2016; (2) of significance (p = 0.1) because the number of plots was small (n = 6). ponderosa pine alive in 2000 and 2016; (3) black oak alive in 2000 and 2016. 2.5. Potential fire behavior To assess if historical tree group development followed an even- or uneven-age model of group development, we identified the age struc- Potential fire behavior following 22 years of post-fire succession ture of trees in the tree group sizes in 2000. A narrow range of tree ages after the 1994 wildfire was determined for each plot using Crown Mass within groups would support an even-aged model whereas a wide range in the Fuel Management Analyst Plus software (FMA, Carlton, 2004). in ages would indicate groups were comprised of multiple age classes. Crown Mass uses tree lists from each plot to estimate potential fire We used a 20 year age range as a cutoff for even-aged (≤20 years) and behavior (i.e. rate of spread, flame length, fire type) and fire effects (i.e. un-even aged (> 20 years) groups to account for variation in the probability of mortality, percent crown scorch) under specified burning number of years to reach coring height. The proportion of even and conditions. Potential fire behavior depends on several factors including uneven aged groups was calculated for each tree group size except for the type and quantity of fuel, fuel moisture content, and fire weather the individual trees. conditions (Rothermel, 1983; Ryan and Reinhardt, 1988). To account for potential variability in fire behavior estimates, we ran simulations

292 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302 for three fire weather conditions: the 50th, 80th, and 98th percentile area loss for black oak (Table 3). Competition influenced tree mortality values of the energy release component (ERC) during the fire season over the study period. Trees that died over the census period experi- (May 1–September 31) (Table A1). ERC is sensitive to variation in fuel enced greater competition as measured by the H index than those that moisture values and how they influence potential fire intensity. Tem- survived to 2016 (H = 3.06 vs 2.01, p = 0.04), while surviving trees perature, wind speed, and fuel moistures for ERC percentiles were ex- had similar levels of competition in 2000 and 2016 (H = 2.01 vs 1.97, tracted using Fire Family Plus (Main et al., 1999) (Table A1). Daily ERC p = 0.37). percentiles and weather variables from the nearby (20 km southwest, elevation 800 m) Cohasset remote automated weather station were used to derive potential fire behavior estimates. Surface fuel char- 3.1.2. Seedlings and saplings acteristics in each plot were estimated using photo-series (Blonski and From 2000 to 2016, the density of ponderosa pine seedlings and Schramel, 1981), and photo-series values were used by FMA to estimate saplings increased (p < 0.1) while for black oak, seedlings decreased a standard fuel model. Photo-series estimates were most similar to and sapling density did not change (p ≥ 0.1) (Table 1). In 2000, 87% of Timber Litter model 08 (TL08, Scott and Burgan, 2005), so this surface black oak regeneration was from seedling establishment rather than fuel model was used for estimating potential fire behavior in all plots sprouts and this percentage decreased to 35% by 2016 suggesting lower (Table A2). A standard fuel model was used for potential fire behavior survival of oak seedlings compared to oak sprouts. estimates rather than field measurements because standard fuel models have been calibrated with fire behavior under conditions similar to those used in this study. Canopy fuel variables were estimated by FMA 3.2. Spatial patterns from the diameters and crown ratios of trees in the tree list of each plot (van Wagner, 1993, 1977) (Table A2). 3.2.1. Tree group characteristics and dynamics The proportion of trees in group sizes at different inter-tree dis- tances varied between 2000 and 2016. On average, the proportion of 3. Results individual and small tree groups at inter-tree distances of 1–3 m de- clined, while those in medium tree groups increased, and the propor- 3.1. Stand characteristics and development tion of extra-large tree groups increased at all inter-tree distances (1–12 m) (Figs. 3 and 4). Average tree group size was 5.5 trees per 3.1.1. Overstory trees group in 2000 and 9.2 trees in 2016. Maximum group size increased On average, tree density and basal area for ponderosa pine and from 38 trees to 240 trees as post-fire regeneration grew into the tree black oak were similar (p > 0.1) in 2000 to 2016 (Table 1), but there size class. Although the proportion of group types changed over time at were important demographic differences between species (Table 2). the 6-m inter-tree distance, the frequency of group types was statisti- Ponderosa pine had higher rates of recruitment and lower rates of cally similar in both years (p = 0.77, Kolmorov-Smirnov two sample − mortality than black oak (p < 0.05) over the study period. On average, test) (Fig. 5). The average number of groups per plot was 18 ha 1 −1 −1 −1 the number of living stems (λstems) of ponderosa pine increased (range 11–27 ha ) and 22 ha (range 15–30 ha ) in 2000 and 2016, − − (3.7% yr 1) while those of black oak decreased slightly (−0.01% yr 1) respectively (Table A3). (Table 2). As expected, ingrowth of both ponderosa pine and black oak The age range of trees varied with group size (Fig. 6). Using the were concentrated in the smallest size-classes (5–15, 15–25 cm, sap- criteria of ≤20 years of age between trees for an even-aged group, lings) as saplings recruited into the tree size classes. Mortality was size- ∼70% of small tree groups were even-aged while fewer medium (30%), dependent and not evenly distributed among size-classes. Mortality and no large (0%) or extra-large tree groups (0%) were even-aged. rates for ponderosa pine were highest for small (5–15 cm) and large The global spatial statistic, Ripley’s L(r) revealed a significant pat- (100–120 cm) diameter trees. In contrast, mortality rates for black oak tern of clumping in all plots in 2000 and 2016; however, the scale and were highest for larger diameter trees (65–75 cm) (Fig. 2). intensity of tree clumping differed among plots, by species, and by year. Basal area remained relatively stable due to limited overstory For plots 1, 3, and 6 the clump intensity for all trees and ponderosa pine mortality and large increases in tree density (Table 3). On average, the trees increased from 2000 to 2016 (Fig. 7). For plots 2, 4, and 5 the − rate of basal area increase (0.29% yr 1) was higher for ponderosa pine intensity of clumping was similar but the scale of clumping was smaller − than black oak (−0.79% yr 1) and plots with the highest rates of basal in 2016. In plots with both species, black oak was more intensely area increase in ponderosa pine experienced the greatest rates of basal clumped than ponderosa pine, except in plot 4.

Table 1 Characteristics of live and standing dead trees (≥5 cm dbh), density of seedlings (0.2–1.4 m tall), and saplings (> 1.4 m tall, < 5 cm dbh), and average shrub and herbaceous cover in six ∼1 ha plots in the Beaver Creek Pine.

Plot

1 2 3 4 5 6 2000 2016 2000 2016 2000 2016 2000 2016 2000 2016 2000 2016 − Ponderosa pine Basal area (m2 ha 1) Live 13.57 12.64 28.20 27.47 27.80 29.45 26.91 30.76 28.56 33.89 16.04 16.14 Snags 1.43 2.93 1.88 5.06 0.08 1.40 1.95 0.91 2.69 1.45 5.05 6.22 − Tree Density (ha 1) Live 82 410 103 106 65 90 139 163 184 176 64 281 Snags 17 13 16 30 1 2 13 15 34 14 17 22 − Seedlings (ha 1) 675 3231 691 9151 41 804 121 1553 1760 3019 730 1872 − Saplings (ha 1) – 1556 – 1563 – 75 3 239 – 104 – 640

− California Black oak Basal area (m2 ha 1) Live 0.91 0.97 1.78 2.05 7.62 7.21 2.88 1.86 1.26 0.64 2.05 1.91 Snags – – 0.84 0.21 0.52 1.51 0.37 0.32 – 0.65 0.59 0.55 − Tree Density (ha 1) Live 1 1 26 44 63 59 19 14 3 2 7 10 Snags – – 173 8 11 4 2 – 1 3 2 − Seedlings (ha 1) 60 50 342 50 252 14 61 4 41 26 73 63 − Saplings (ha 1) 1 2 7046 27 23 4 – – 13 1 3 Average Shrub Cover (%) 5.1 16.4 21.2 34.9 19.0 37.1 4.8 15.1 22.3 28.0 6.1 29.2 Average Herb Cover (%) 43.3 22.9 24.4 11.2 27.6 30.4 20.9 24.3 32.5 14.4 48.0 23.1

293 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Table 2 Number of living trees (stems ≥ 5 cm) on each census date and the rate (%) of mortality and recruitment over the study period (2000 to 2016).

Plot

1 2 3 4 5 6

Ponderosa pine Living Stems2000 (Nb) 83 103 59 139 148 64

Living Stems2016 (Ne) 411 106 83 163 141 281

Mortality (Nb-Ns) 1212 2 8 137

Recruitment (Ne-Ns) 340 15 26 32 6 224 − Mortality rate (% year 1) 0.98 0.77 0.22 0.37 0.57 0.72 − Recruitment rate (% year 1) 10.9 0.95 2.35 1.37 0.27 9.97 −1 λ stems (% year ) 10.0 0.18 2.13 1.00 −0.30 9.24

California black oak Living Stems2000 (Nb) 1 26 57 19 3 8

Living Stems2016 (Ne) 1 44 54 14 2 10

Mortality (Nb-Ns) 0 3 7 6 1 3

Recruitment (Ne-Ns) 0 214 1 0 5 − Mortality rate (% year 1) 0 0.76 0.82 2.37 2.53 2.94 − Recruitment rate (% year 1) 0 4.05 0.48 0.46 0 4.33 −1 λ stems (% year ) 0 3.29 −0.34 −1.91 −2.53 1.39

3.2.2. Gap characteristics and dynamics and the highest flame lengths had the highest density of small diameter On average, the rate of gap infilling from tree regeneration trees (< 15 cm dbh). Average probability of mortality was highest − − (1.44% yr 1) and crown extension into gaps (1.05% yr 1) exceeded (100%) for saplings (< 5 cm dbh, > 1.4 m tall) and lowest (1.1%) for − rates of new gap creation (0.58% yr 1). large trees (> 30 cm dbh) (Table 4). The low probability of mortality Despite the higher rate of infilling, there was no difference in the and low crown scorch (11%) for large trees, even under extreme frequency distribution and sizes of gaps in 2000 and 2016 (p > 0.05) weather conditions, masks species differences – the probability of top- (Figs. 3 and 8); however, total gap area (> 9 m from tree was greater kill was higher for black oak (21.5%) than ponderosa pine (7.3%). between the two times (p = 0.07) than in 2000. 4. Discussion 3.2.3. Gap characteristics and regeneration dynamics Area occupied by regeneration increased between 2000 and 2016 4.1. Stand characteristics (Fig. 9) and this was mainly due to the increase of ponderosa pine. There was a decline in area occupied by black oak regeneration Repeat low- and moderate-severity wildfires in the Beaver Creek (Table 1). In 2016, both ponderosa pine and black oak seedlings were Pinery in the 20th century preserved and reinforced a tree group and more abundant beneath canopy gaps (< 33% cover) than in either in- gap structure. Fires in 1901 and 1924 were recorded by fire scars termediate or closed canopy conditions (p < 0.05, Kruskal Wallis H- throughout the Beaver Creek Pinery and the presence of 80–250 year test). The same pattern was observed for ponderosa pine, but not black old trees in each plot suggest these fires were low to moderate in se- oak saplings. verity (Taylor, 2010). After a 66 year period without fire, the Beaver Creek Pinery burned again in 1990 and 1994. Six-years after the last 3.3. Potential fire behavior fire, the forest was comprised of tree groups of different sizes consisting of mainly intermediate and large diameter trees, interspersed with gaps. − Fire behavior modeling predicted potential flame lengths < 1 m Live tree density and basal area after the fires averaged 126 trees ha 1 − under all weather conditions (Table 4). Plots with passive crown fire and 26.3 m2 ha 1. These values fell within the range of values for tree

Fig. 2. Diameter class distribution (bars) of all non-surviving trees in six 1 ha plots in 2016 and size-specific mortality rates (lines) for ponderosa pine (dark green) and California black oak (light green) trees (≥5 cm) that died during the study period (2000–2016). Mid-point values for each 10 cm class are shown on the x-axis.

294 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Table 3 − Basal area (m2ha 1) of living trees (stems ≥ 5 cm) on each census date and the rate of basal area loss (from mortality) and gain (from growth and recruitment) over the study period.

Plot

1 2 3 4 5 6

Pinus ponderosa Living BA2000 (BAb) 13.7 28.2 25.7 26.9 23.0 16.0

Living BA2016 (BAe) 12.7 27.5 26.8 30.8 27.1 16.1

Loss of BA (BAl) 4.98 2.09 1.41 0.66 0.53 3.26

BA recruitment (BAr) 1.26 0.05 0.09 0.08 0.05 1.01

BA growth (BAg) 2.72 1.34 2.42 4.48 4.58 2.35 − BA loss rate (% year 1) 2.27 0.46 0.34 0.15 0.14 1.27 − BA gain rate (% year 1) 1.82 0.31 0.61 1.06 1.26 1.31 −1 λ BA (% year ) −0.46 −0.16 0.27 0.91 1.11 0.04

Quercus kelloggii Living BA2000 (BAb) 0.91 1.78 7.01 2.88 1.00 2.05

Living BA2016 (BAe) 0.97 2.05 6.56 1.86 0.51 1.91

Loss of BA (BA1) 0 0.20 1.21 1.18 0.52 0.62

BA recruitment (BAr) 0 0.07 0.11 0.003 0 0.01

BA growth (BAg) 0.06 0.4 0.65 0.16 0.03 0.47 − BA loss rate (% year 1) 0 0.7 1.08 2.56 3.25 1.89 − BA gain rate (% year 1) 0.41 1.65 0.68 0.35 0.19 1.46 −1 λ BA (% year ) 0.41 0.95 −0.4 −2.21 −3.06 −0.43

− − density (48–315 trees ha 1) and basal area (16–53.9 m2 ha 1) recorded wildfires in the Beaver Creek Pinery maintained conditions similar to from historic and contemporary reference forests in California (Scholl forests before fire suppression (Oliver, 2001; Skinner et al., 2018b). In and Taylor, 2010; Stephens et al., 2015; Taylor et al., 2014), Oregon contrast, current density and basal area in the Beaver Creek Pinery are − (Hagmann et al., 2013) and pine-oak forests in the southwest at, or below, the average density (343–757 trees ha 1) and basal area − − − (52–166 trees ha 1, 8.0–17.6 m2 ha 1)(Fulé et al., 2002; Moore et al., (27.2–69.9 m2 ha 1) of living trees in other conifer and pine-oak forests 2004; Rodman et al., 2017). Twenty-two years after the last fire, that have not burned since the late 19th or early 20th century or his- − − average tree density (226 trees ha 1) and basal area (27.5 m2 ha 1) torical reference stands (Collins et al., 2011; Knapp et al., 2013; North were still within the range of reference forest values indicating that et al., 2007; Taylor et al., 2014). This reduced density and basal area is

Fig. 3. Maps of tree groups and canopy gaps in 2000 (top row) and 2016 (bottom row) in six ∼1 ha stem map plots in the Beaver Creek Pinery. Background shading indicates the distance to the nearest tree (empty space function) and regions > 9 m away from any tree are outlined. Darker green circles represent ponderosa pine and light green circles California black oak.

295 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Fig. 4. Mean proportion of trees (stems ≥ 5 cm dbh) in different tree group sizes at 1–12 m intertree distances for six ∼1 ha plots in the Beaver Creek Pinery. Data shown in this figure are reported in Table A3.

Nakashizuka et al., 2000; Runkle, 2000; Taylor et al., 2004; van Mantgem et al., 2009). Size-dependent mortality in these forests ap- pears to be related to competitive effects from neighboring trees. Tree mortality rates were higher for focal trees at closer distances to other trees and for the smallest and largest diameter stems. Other mortality agents such as insects and pathogens, and warming temperatures may have contributed to the higher rates in these size-classes though specific causes of tree death were not determined. The overall demographic response since the wildfires has been a strong increase in ponderosa pine and a small decline in black oak, leading to greater dominance by ponderosa pine. The post-fire demographic responses of ponderosa pine and black oak in the Beaver Creek Pinery are consistent with longer term com- positional shifts observed in mixed stands that have not burned for several decades or even a century (Cocking et al., 2014, 2012). Shade tolerant, fire-intolerant and fast growing conifers such as Douglas-fir Fig. 5. Pooled frequency of individual, small, medium, large, and extra-large (Pseudotsuga menziesii), white fir(Abies concolor), and incense-cedar tree groups in 2000 (left panel) and 2016 (right panel) at a 6 m intertree dis- tance. (Calocedrus decurrens) have established abundantly in black oak woodlands and mixed pine-oak forests during long fire-free periods and these species can overtop and replace black oak (Cocking et al., 2014, fi likely the direct result of multiple recent res. 2012; Hammett et al., 2017). Re-entry burns after long periods of fire Post-fire forest development reveals different demographic re- suppression can have highly variable effects on pine-oak regeneration sponses by ponderosa pine and black oak. Fires initially promoted black and dominance. Higher severity fires reduce conifer dominance and oak regeneration from seed and via sprouting, a pattern observed for stimulate black oak regeneration through resprouting and new estab- fi black oak after other res (Crotteau et al., 2013). By 2016, many post- lishment (Cocking et al., 2014, 2012), and there is high survival of fire sprouts had grown into small trees while seedlings that established black oak sprouts after top kill in high-severity reburns (Hammett et al., ∼ from seed decreased by 75%, and the number and area occupied by 2017; Nemens et al., 2018). Compositional shifts leading to increased black oak seedlings declined over the census period. Post-fire height dominance of oaks has been observed in burns and reburns of mod- growth rates of black oak sprouts are high, and faster than height erate- and high-severity in California, the southwest and northern growth rates of conifers, providing an initial advantage to oaks in early Mexico and is characteristic of sprouting species in mixed pine-oak stand development under open conditions (Crotteau et al., 2013; forests (Barton and Poulos, 2018; Cocking et al., 2014, 2012; Fulé et al., Hammett et al., 2017; Nemens et al., 2018). In contrast, there was a 2000). Shifts in relative dominance are more muted after low-severity large increase in ponderosa pine regeneration by 2016, with seedlings fire (Cocking et al., 2014) and response patterns in the Beaver Creek five-fold and saplings 200-fold more abundant than in 2000. Average Pinery were consistent with a low-severity fire pattern. Yet, the robust annual percentage population growth rates for ponderosa pine were post-fire demographic response of ponderosa pine suggests that more also three-fold higher than for black oak. Ponderosa pine basal area also frequent (< 22-year intervals) low-severity fires may be necessary to increased over the study period, while black oak basal area declined. retain black oak in these old-growth stands by reducing ponderosa pine Tree mortality rates for both species were size-dependent, a pattern establishment. identified for other conifer and deciduous forest species in temperate forests in North America and East Asia (Canham and Murphy, 2017;

296 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Fig. 6. Box plots of the range in ages of trees (≥5 cm dbh) for different tree group sizes in each stem map plots in the Beaver Creek Pinery in 2000. The dashed line at 20 years represents the threshold used to distinguish between even (< 20 years) and un-even (> 20 years) aged groups.

4.2. Spatial patterns and dynamics tree groups, but their density was likely still high in 2000. Tree numbers in the large tree group sizes were augmented further by growth of A spatial structure comprised of individual trees, tree groups, and seedlings and saplings into the tree size class between 2000 and 2016. − canopy gaps is a hallmark of historical reference conditions in frequent- Although the maximum number of trees group 1 in the Beaver Creek fire ponderosa pine and pine-oak forests. The two low-moderate se- Pinery is higher than in other pine-oak and pine dominated frequent- − verity wildfires in the Beaver Creek Pinery in the 1990s re-enforced a fire forests, the average number of tree groups (18 ha 1) is quite similar − tree group and gap structure generated by the 1924 and earlier fires to forests in the eastern Cascades (∼19 ha 1)(Churchill et al., 2017, − that has some characteristics similar to reference forests despite forest 2013), the southwest (∼15 ha 1)(Abella and Denton, 2009), and the − changes caused by decades of fire suppression. The average (global) Rocky Mountains (∼18 ha 1)(Brown et al., 2015; Clyatt et al., 2016). spatial pattern of trees in the Beaver Creek Pinery was clumping at Canopy cover increased after 22-years of post-fire forest develop- scales < 0.4 ha with local variability in clump size, a spatial structure ment, with gap creation rates lower than rates of infilling and canopy identified in other frequent-fire old-growth pine-oak and pine-domi- extension. Gap areas in the Beaver Creek Pinery decreased slightly from nated forests in the western United States (Fry et al., 2014; Larson and 2000 to 2016 and infilling of gaps by trees was evident, causing frag- Churchill, 2012; Taylor, 2004; White, 1985). mentation of large gaps (> 600 m2) into smaller gaps (< 600 m2) (e.g. Fig. 3, plot 1, 6; Fig. 8). Nevertheless, gaps areas in the Beaver Creek Pinery were larger compared to frequent-fire forests that have experi- 4.2.1. Group and gap characteristics enced a century or more of fire suppression, Forests that have not Average tree group size in the Beaver Creek Pinery was similar to burned for a century or more have mainly small gaps (< 250 m2) and historic and contemporary reference conditions, but maximum tree more continuous forest cover because gaps have been filled by trees – −1 group size was larger (Beaver Creek Pinery = 38 240 trees group ). (Fry et al., 2014; Lydersen et al., 2013; Skinner, 1995). −1 Maximum values for trees group were only 40 in the eastern Creating and maintaining a diverse pattern of canopy gaps and tree Cascades (Churchill et al., 2017, 2013), 25 in the southwest (Abella and groups in both relatively even-aged, single cohort stands and un-even Denton, 2009; Sánchez Meador et al., 2011), and 26 in the Rocky aged stands with few older trees remaining, can create natural fuel Mountains (Brown et al., 2015). Moreover, compared to the Beaver breaks that reduce fire risk, thin out undesirable overstory species, – Creek Pinery, proportions of trees in large (10 15) and extra-large promote understory diversity, and over time create a more hetero- ≥ ( 16 trees) groups at a 6-m inter-tree distance was also smaller (e.g. genous age-structure (Churchill et al., 2013). In many western forests, Churchill et al., 2017; Larson and Churchill, 2012; Sánchez Meador particularly those impacted by past management and fire-suppression, ff et al., 2011). Some of these di erences among studies could be related active forest management is needed to restore or promote structural ff to variation in the minimum tree size cuto . heterogeneity. In these cases, the values reported here and in the sup- fi Site productivity and re frequency may explain the high maximum plemental information for a contemporary ponderosa pine-black oak −1 trees group in the Beaver Creek Pinery (Churchill et al., 2017; Clyatt reference site can assist managers in identifying specific thresholds for fi et al., 2016; Lydersen et al., 2013). Frequent- re forests growing on the size and frequency of tree groups and canopy gaps. The results of highly productive sites tend to support higher tree density and basal our study also suggest that, once stand structure is restored, multiple, area, and larger tree groups than on less productive sites (Churchill successive low-moderate severity fires can be used to maintain het- et al., 2017). Soil type (Cohasset Loam), rapid height-gain, and basal erogeneity and resiliency. area gain rates in the Beaver Creek Pinery between 2000 and 2016 The age of trees in tree groups provides insights into pattern-process ff indicate the site is productive. Fire suppression e ects on stand struc- relationships and mechanisms of group development. Trees were more ture would also contribute to high tree density in the extra-large tree likely to be even-aged in small groups (2–4 trees) and uneven-aged in group. Fires burned in the Beaver Creek Pinery at a median interval of larger groups (≥5 trees). These results support findings by White fi fi 12 years before re suppression and there was a 66-year re-free period (1985), and spatial analyses of tree ages in other forests that show large – fi (1924 1990) before the 1990 and 1994 res (Taylor, 2010). Re- tree groups consisting of overlapping groups of even-aged trees (Beaty cruitment of seedlings and saplings into the canopy after 1905 was high and Taylor, 2007; Scholl and Taylor, 2010). In frequent-fire forests, fi due to lower re frequency and some trees grew large enough to survive even-age groups of seedlings often establish on bare mineral soil where the 1990 and 1994 fires (Taylor, 2010). These fires thinned the dense

297 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Fig. 7. Global spatial patterns identified by Ripley’s L(r) for: (1) all living trees (≥5 cm dbh) (left column), all living ponderosa pine trees (middle) and all living California black oak (right column) at inter-tree distances of r. The spatial pattern is clumped when values exceed the envelope of complete spatial randomness (CSR – shaded grey), random within the envelope, and uniformly spaced when values are below the envelope. overstory tree(s) have been killed by fire (Cooper, 1961, 1960). Partial extreme weather conditions. Surface and canopy fuel loads in 2016 seedling and sapling mortality in subsequent fires and additional re- were still moderate, and only saplings and small trees (< 30 cm dbh) cruitment would create large uneven-aged tree groups comprised of had high probabilities of mortality; large trees (> 30 cm dbh) were fire- small even-aged groups (Beaty and Taylor, 2007; Scholl and Taylor, resistant under modeled fire behavior. The low-moderate fire severity 2010). Interestingly, half of the extra-large (≥16 trees) tree groups in effects predicted for the Beaver Creek Pinery are consistent with po- the Beaver Creek Pinery had a majority of trees in a single 20-year age tential fire behavior predicted for pine dominated frequent-fire forests class that established after 1920. This is likely related to the relatively in the western US before fire exclusion (Hessburg et al., 2005; Taylor large numbers of trees that established during the 1924 to 1990 fire-free et al., 2014; Van de Water and North, 2011) and demonstrates a rela- period near pre-existing individual trees creating large uneven-aged tively long-lasting effect of wildfire on forest structures that influence groups with a high percentage of trees in a single age-class. potential fire behavior. This suggests that if managers initially invest time and resources towards restoring heterogenous structure, these 4.3. Potential fire behavior restored forests may retain resilience to fire for an extended period of time, thereby extending the maintenance interval of these sites beyond Twenty-two years after the wildfires, fuels and vegetation in the historical fire frequencies. If management generated a similar degree of plots were predicted to support low-severity fire under moderate and heterogeneity across a greater proportion of historically frequent-fire

298 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

North et al., 2012; Calkin et al., 2015).

4.4. Conclusions

At present, forest conditions within the Beaver Creek Pinery plots are within the historical range of variability for frequent-fire forests in the western U.S. These conditions have been created and reinforced by repeated wildfires that occurred in the recent past and earlier in the 20th century. The fire effects experienced in the Beaver Creek Pinery were, in part, be due to the particular weather and fuel moisture con- ditions, as well as the heterogeneous stand structure present at the time of these wildfires. Our results support a perspective that wildfire can maintain and restore resiliency by preserving latent structure (such as legacy trees and canopy openings) in forests with relatively limited anthropogenic disturbances (Larson et al., 2013). Old-growth forest structures were resilient to fire despite nearly 70 years of fire exclusion; once restored, a tree group and gap forest structure will likely persist for an extended period following the last wildfire. Modeling of potential fire behavior for a fire that may occur in the near future also suggests that large, old, overstory trees are currently resilient to fire but seed- lings, saplings and small diameter trees would be killed. Maintaining relatively frequent and variable fire intervals into the future with Fig. 8. Pooled frequency of gap sizes in 2000 (white) and 2016 (black) for six wildfire or prescribed fire, in this Wilderness and elsewhere, will be ∼1 ha stem map plots in the Beaver Creek Pinery. needed to reduce the demographic pressure of ponderosa pine, promote black oak, and to maintain and create future spatial heterogeneity in western forests wildfire, burning under moderate weather conditions vegetation and fuels (Ryan et al., 2013). Cumulatively, this research fi may become a more feasible management strategy for more than just demonstrates that wildland res, under certain conditions, can main- fi unharvested forests in remote wilderness areas (Larson et al., 2013, tain and restore re resilience in unmanaged ponderosa pine and pine- oak forests and reduce the negative ecological consequences associated

Fig. 9. Maps of regeneration density for both ponderosa pine and California black oak seedlings and saplings in 2000 (top row) and 2016 (bottom row) in six ∼1ha stem map plots in the Beaver Creek Pinery. Darker shading represents higher density of regeneration. Overstory tree groups are gray and partially transparent.

299 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Table 4 with past fire exclusion (Franklin and Johnson, 2012; Naficy et al., Potential fire behavior and fire effects for saplings (< 5 cm dbh, > 1.4 m tall), 2016; North et al., 2012; Taylor et al., 2016). small trees (5–30 cm dbh), and large trees (> 30 cm dbh) under different ∼ weather conditions in six 1 ha plots in the Beaver Creek Pinery. Surface fuels Acknowledgements in each plot were constant (TL08, Scott & Burgan 2005) in all simulations.

Weather Percentile 50th 80th 98th Support for this research was provided by the Lassen National Forest, Pacific Southwest Research Station, and the Joint Fire Science Fire Type 1 Passive Passive Passive Crown Program (15-1-07-19). We thank Bob Carlson Patrick Doyle, Bobette Crown Crown ff 2 Passive Passive Passive Crown Jones, Lucas Harris, Cerena Brewen, Shannon Drennan, Je Mcfarland, Crown Crown Chelsea Phillips, and Molly West for help with logistics and field work. 3 Surface Surface Surface 4 Surface Surface Surface Appendix A. Supplementary material 5 Surface Surface Surface 6 Passive Passive Passive Crown Crown Crown Supplementary data to this article can be found online at https:// doi.org/10.1016/j.foreco.2018.12.016. Surface Rate of Spread (km/ 1 5.5 6.1 6.1 hr) 2 5.8 6.3 6.3 3 4.5 5.1 5.5 References 4 5.1 5.5 5.5 5 5.1 5.5 5.5 Abella, S.R., Denton, C.W., 2009. Spatial variation in reference conditions: historical tree 6 5.1 5.5 5.6 density and pattern on a Pinus ponderosa landscape. Can. J. For. Res. 39, 2391–2403. fi Surface Flame Length (m) 1 0.8 0.8 0.9 Airey Lauvaux, C., Skinner, C.N., Taylor, A.H., 2016. High severity re and mixed conifer 2 0.8 0.9 0.9 forest- dynamics in the southern . USA. For. Ecol. Manage. – 3 0.8 0.8 0.9 363, 74 85. Anderson, M.K., 2005. Tending the Wild: Native American Knowledge and the 4 0.8 0.9 0.9 Management of California’s Natural Resources. University of California Press. 5 0.8 0.9 0.9 Baddeley, A., Turner, R., 2005. Spatstat: an R package for analyzing spatial point patterns. 6 0.8 0.9 0.9 J. Stat. Softw. 12, 1–42. Barbour, M., Kelley, E., Maloney, P., Rizzo, D., Royce, E., Fites-Kaufmann, J., 2009. Average Probability of Mortality (StDev) Present and past old-growth forests of the Basin, Sierra Nevada, US. J. Saplings (< 5cm, > 1.4 m) 1 100 (0) 100 (0) 100 (0) Veg. Sci. 13. 2 100 (0) 100 (0) 100 (0) Barton, A.M., 2002. Intense wildfire in southeastern Arizona: transformation of a 3 100 (0) 100 (0) 100 (0) Madrean oak–pine forest to oak woodland. For. Ecol. Manage. 165, 205–212. 4 100 (0) 100 (0) 100 (0) Barton, A.M., Poulos, H.M., 2018. Pine vs. oaks revisited: conversion of Madrean pine-oak 5 100 (0) 100 (0) 100 (0) forest to oak shrubland after high-severity wildfire in the Sky Islands of Arizona. For. 6 100 (0) 100 (0) 100 (0) Ecol. Manage. 414, 28–40. Beaty, R.M., Taylor, A.H., 2007. Fire disturbance and forest structure in old-growth mixed Small Trees (5–30 cm) 1 80.7 (12.4) 81.2 (12.4) 82.1 (12.4) conifer forests in the northern Sierra Nevada. California. J. Veg. Sci. 18, 879–890. 2 81.3 (14.0) 82.2 (14.2) 82.8 (14.6) Binkley, D., Sisk, T., Chambers, C., Springer, J., Block, W., 2007. The role of old-growth 3 84.9 (14.7) 85.4 (15.5) 86.4 (13.4) forests in frequent-fire landscapes. Ecol Soc. 12, 18. 4 73.8 (13.0) 74.6 (14.4) 75.2 (15.1) Blonski, K.S., Schramel, J.L., 1981. Photo series for quantifying natural forest residues: 5 48.2 (11.2) 48.5 (11.4) 49.4 (12.1) southern Cascades, northern Sierra Nevada. Gen. Tech. Rep. PSW-56. Berkeley, Calif.: 6 97.6 (30.4) 98.1 (32.5) 98.3 (33.3) U.S. Depart. Agri. For. Serv., Pacific Southwest For. Range Exp. Stn. 145, p. Brown, P.M., Battaglia, M.A., Fornwalt, P.J., Gannon, B., Huckaby, L.S., Julian, C., Cheng, Large Trees (> 30 cm) 1 0 (0) 0 (0) 0 (0) A.S., 2015. Historical (1860) forest structure in ponderosa pine forests of the northern 2 0 (0) 0 (0) 0 (0) Front Range. Colorado. Can. J. For. Res. 45, 1462–1473. 3 0.1 (1.0) 0.3 (1.7) 0.6 (2.0) Buchanan, J.B., Rogers, R.E., Pierce, D.J., Jacobson, J.E., 2003. Nest-site habitat use by 4 0.2 (2.2) 0.2 (2.5) 0.2 (2.7) white-headed woodpeckers in the Eastern Cascade Mountains. Washington. 5 0 (0) 0 (0) 0 (0.1) Northwest. Nat. 84, 119. 6 4.2 (9.9) 6.3 (11.6) 7.8 (12.9) Calkin, D.E., Thompson, M.P., Finney, M.A., 2015. Negative consequences of positive feedbacks in US wildfire management. For. Ecosyst. 2, 1. Crown Scorch (%) Canham, C.D., Murphy, L., 2017. The demography of tree species response to climate: sapling and canopy tree survival. Ecosphere 8, e01701. Saplings (< 5cm, > 1.4 m) 1 100 (0) 100 (0) 100 (0) Carlton, D. Fuels Management Analyst Plus Software, Version 3.02; 2004. 2 100 (0) 100 (0) 100 (0) Churchill, D.J., Larson, A.J., Dahlgreen, M.C., Franklin, J.F., Hessburg, P.F., Lutz, J.A., 3 100 (0) 100 (0) 100 (0) 2013. Restoring forest resilience: from reference spatial patterns to silvicultural 4 100 (0) 100 (0) 100 (0) prescriptions and monitoring. For. Ecol. Manage. 291, 442–457. 5 100 (0) 100 (0) 100 (0) Churchill, D.J., Jeronimo, S.M., Larson, A.J., Fischer, P., Dahlgreen, M.C., Franklin, J.F., 6 100 (0) 100 (0) 100 (0) 2016. The ICO approach to quantifying and restoring forest spatial pattern: im-

Small Trees (9–30 cm) 1 61.2 (0) 63.2 (1.4) 65.4 (1.6) plementation guide. ICO App. 3, 45p. Churchill, D.J., Carnwath, G.C., Larson, A.J., Jeronimo, S.A., 2017. Historical Forest 2 60.5 (0) 65.7 (2.6) 67.7 (4.1) Structure, Composition, and Spatial Pattern in Dry Conifer Forests of the Western 3 70.3 (6.2) 72.2 (8.8) 74.2 (9.4) Blue Mountains. Gen. Tech. Report, USDA For. Serv, Oregon. 4 55.7 (3.2) 58.4 (5.8) 60.6 (7.9) Clyatt, K.A., Crotteau, J.S., Schaedel, M.S., Wiggins, H.L., Kelley, H., Churchill, D.J., 5 15.8 (0) 16.8 (1.8) 18.7 (4.2) Larson, A.J., 2016. Historical spatial patterns and contemporary tree mortality in dry 6 95.0 (24.9) 96.4 (27.9) 96.9 (29.4) mixed-conifer forests. For. Ecol. Manage. 361, 23–37. ff fi Large Trees (> 30 cm) 1 10.9 (6.1) 10.9 (6.1) 10.9 (6.1) Cocking, M.I., Varner, J.M., Knapp, E.E., 2014. Long-term e ects of re severity on oak–conifer dynamics in the southern Cascades. Ecol. Appl. 24, 94–107. 2 9.1 (8.4) 9.1 (8.4) 9.1 (8.4) Cocking, M.I., Varner, J.M., Sherriff, R.L., 2012. California black oak responses to fire 3 13.7 (12.2) 13.7 (12.2) 13.7 (12.2) severity and native conifer encroachment in the Klamath Mountains. For. Ecol.

4 10.0 (8.2) 10.0 (8.2) 10.0 (8.2) Manage. 270, 25–34. 5 10.7 (7.1) 10.7 (7.1) 10.7 (7.1) Collins, B.M., Everett, R.G., Stephens, S.L., 2011. Impacts of fire exclusion and recent 6 11.1 (10.3) 11.6 (11.4) 12.1 (12.2) managed fire on forest structure in old growth Sierra Nevada mixed-conifer forests. Ecosphere 2, art51. Condit, R., Ashton, P.S., Manokaran, N., Lafrankie, J.V., Hubbell, S.P., Foster, R.B., 1999. Dynamics of the forest communities at Pasoh and Barro Colorado: comparing two 50- ha plots. Philos. Trans. R. Soc. B 1739–1748. Condit, R., Hubbell, S.P., Foster, R.B., 1995. Mortality rates of 205 Neotropical tree and shrub species and the impact of a severe drought. Ecol. Monogr. 65, 419–439. Contreras, M.A., Affleck, D., Chung, W., 2011. Evaluating tree competition indices as predictors of basal area increment in western Montana forests. For. Ecol. Manage.

300 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

262, 1939–1949. Lydersen, J.M., North, M.P., Knapp, E.E., Collins, B.M., 2013. Quantifying spatial patterns Cooper, C.F., 1960. Changes in vegetation, structure, and growth of southwestern pine of tree groups and gaps in mixed-conifer forests: reference conditions and long-term forests since white settlement. Ecol. Monogr. 30, 129–164. changes following fire suppression and logging. For. Ecol. Manage. 304, 370–382. Cooper, C.F., 1961. Pattern in Ponderosa Pine forests. Ecology 42, 493–499. Main, W., Pananen, D., Burgan, R. FireFamily 1988. USDA For. Serv. Gen. Tech. Rep. NC- Coppoletta, M., Merriam, K.E., Collins, B.M., 2016. Post-fire vegetation and fuel devel- 138; 1999. opment influences fire severity patterns in reburns. Ecol. Appl. 26, 686–699. McDonald, P.M., Tappeiner, J.C. California’s Hardwood Resource: Seeds, Seedlings, and Covington, W.W., Moore, M.M., 1994. Southwestern Ponderosa forest structure: changes Sprouts of Three Important Forest-Zone Species, Gen. Tech. Rep. PSW-GTR-185. since Euro-American settlement. J. For. 92, 39–47. Albany, CA: Pacific Southwest Research Station, Forest Service, U.S. Department of Crotteau, J.S., Morgan Varner, J., Ritchie, M.W., 2013. Post-fire regeneration across a fire Agriculture. 39 p; 2002. severity gradient in the southern Cascades. For. Ecol. Manage. 287, 103–112. Millar, C.I., Stephenson, N.I., Stephens, S.L., 2007. Climate change and forests of the Diggle, P.J., 2013. Statistical Analysis of Spatial and Spatio-Temporal Point Patterns, future: managing in the face of uncertainty. Ecol. App. 17, 2145–2151. Third – Peter J. Diggle - Google Books, 3rd ed. CRC Press, Boca Raton, Florida. Moore, M.M., Huffman, D.W., Fule, P.Z., Covington, W.W., Crouse, J.E., 2004. Dodson, E.K., Peterson, D.W., Harrod, R.J., 2008. Understory vegetation response to Comparison of historical and contemporary forest structure and Southwestern thinning and burning restoration treatments in dry conifer forests of the eastern Ponderosa Pine forests. For. Sci. 50. Cascades. USA. For. Ecol. Manage. 255, 3130–3140. Naficy, C., Sala, A., Keeling, E.G., Graham, J., DeLuca, T.H., 2010. Interactive effects of Engber, E.A., Varner, J.M., 2018. Predicting Douglas-fir sapling mortality following historical logging and fire exclusion on ponderosa pine forest structure in the prescribed fire in an encroached grassland. Restor. Ecol. 20, 665–668. northern rockies. Ecol. Appl. 20, 1851–1864. Folke, C., Carpenter, S.R., Walker, B., Scheffer, M., Chapin, T., Rockstrom, J., 2010. Ecol. Naficy, C.E., Keeling, E.G., Landres, P., Hessburg, P.F., Veblen, T.T., Sala, A., 2016. Soc. 15, 20. Wilderness in the 21st century: a framework for testing assumptions about ecological Franklin, J.F., Johnson, K.N., 2012. A restoration framework for federal forests in the intervention in wilderness using a case study of fire ecology in the rocky mountains. Pacific Northwest. J. For. 110, 429–439. J. For. 114, 384–395. Fry, D.L., Stephens, S.L., 2010. Stand-level spatial dependence in an old-growth Jeffrey Nakashizuka, T., Iida, S., Tanaka, H., Shibata, M., Abe, S., Masaki, T., Niiyama, K., 2000. pine – mixed conifer forest, Sierra San Pedro Mártir. Mexico. Can. J. For. Res. 40, Community dynamics of Ogawa forest reserve, a species rich deciduous forest, central 1803–1814. Japan. Vegetatio 103, 105–112. Fry, D.L., Stephens, S.L., Collins, B.M., North, M.P., Franco-Vizcaino, E., Gill, S.J., 2014. Nemens, D.G., Varner, J.M., Kidd, K.R., Wing, B., 2018. Do repeated wildfires promote Contrasting spatial patterns in active-fire and fire-suppressed Mediterranean climate restoration of oak woodlands in mixed-conifer landscapes? For. Ecol. Manage. 427, old-growth mixed conifer forests. PLos One 9, e88985. 143–151. Fulé, P.Z., Covington, W.W., Smith, H.B., Springer, J.D., Heinlein, T.A., Huisinga, K.D., North, M., Collins, B.M., Stephens, S., 2012. Using fire to increase the scale, benefits, and Moore, M.M., 2002. Comparing ecological restoration alternatives: Grand. For. Ecol. future maintenance of fuels treatments. J. For. 392–401. Manage, Canyon, Arizona. North, M., Innes, J., Zald, H., 2007. Comparison of thinning and prescribed fire re- Fulé, P.Z., García-Arévalo, A., Covington, W.W., 2000. Effects of an intense wildfire in a storation treatments to Sierran mixed-conifer historic conditions. Can. J. For. Res. 37, Mexican Oak-Pine forest. For. Sci. 46, 52–61. 331–342. Gaines, W.L., Harrod, R.J., Dickinson, J., Lyons, A.L., Halupka, K., 2010. Integration of North, M., Oakley, B., Fiegener, R., Gray, A., Barbour, M., 2005. Influence of light and soil Northern spotted owl habitat and fuels treatments in the eastern Cascades, moisture on Sierran mixed-conifer understory communities. Plant Ecol. 177, 13–24. Washington. USA. For. Ecol. Manage. 260, 2045–2052. North, M.P., Stephens, S.L., Collins, B.M., Agee, J.K., Aplet, G., Franklin, J.F., Fulé, P.Z., Guarín, A., Taylor, A.H., 2005. Drought triggered tree mortality in mixed conifer forests 2015. Reform forest fire management: agency incentives undermine policy effec- in , California. USA. For. Ecol. Manage. 218, 229–244. tiveness. Science 349, 1280–1281. Guiterman, C.H., Margolis, E.Q., Allen, C.D., Falk, D.A., Swetnam, T.W., 2018. Long-term Oliver, W.W., 2001. Can we create and sustain late successional attributes in interior persistence and fire resilience of oak shrubfields in dry conifer forests of Northern ponderosa pine stands? In: Ponderosa Pine Ecosystems Restoration and Conservation: New Mexico. Ecosystems 21, 943–959. Steps toward Stewardship. Ogden, UT, pp. 99–103. Hagmann, R.K., Franklin, J.F., Johnson, K.N., 2013. Historical structure and composition Parks, S.A., Holsinger, L.M., Miller, C., Nelson, C.R., 2015. Wildland fire as a self-reg- of ponderosa pine and mixed-conifer forests in south-central Oregon. For. Ecol. ulating mechanism: the role of previous burns and weather in limiting fire progres- Manage. 304, 492–504. sion. Ecol. Appl. 25, 1478–1492. Hammett, E.J., Ritchie, M.W., Berrill, J.P., 2017. Resilience of California black oak ex- Plotkin, J.B., Chave, J., Ashton, P.S., 2002. Cluster analysis of spatial patterns in periencing frequent fire: regeneration following two large wildfires 12 years apart. Malaysian tree species. Am. Nat. 160, 629–644. Fire Ecol. 13, 91–103. Preisler, H.K., Grulke, N.E., Heath, Z., Smith, S.L., 2017. Analysis and out-year forecast of Harris, L., Taylor, A.H., 2015. Topography, fuels, and fire exclusion drive fire severity of beetle, borer, and drought-induced tree mortality in California. For. Ecol. Manage. the Rim Fire in an old-growth mixed-conifer forest, Yosemite National Park, USA. 399, 166–178. Ecosystems 18, 1192–1208. R Core Team. A Language and Environment for Statistical Computing. In: R Foundation Harris, L., Taylor, A.H., 2017. Previous burns and topography limit and reinforce fire for Statistical Computing. Vienna Austria; 2013. severity in a large wildfire. Ecosphere 8, e02019. Reily, M.J., Dunn, C.J., Meigs, G.W., Spies, T.A., Kennedy, R.E., Bailey, J.D., Briggs, K., Hegyi, F., 1974. A simulation model for managing jack-pine stand. In: Growth Models for 2017. Ecosphere 8, 1–28. Tree and Stand Simulation. Royal College of Forestry, Stockholm, Sweden, pp. 74–90. Ritchie, M.W., Skinner, C.N., Hamilton, T.A., 2007. Probability of tree survival after Hessburg, P.F., Agee, J.K., Franklin, J.F., 2005. Dry forests and wildland fires of the in- wildfire in an interior pine forest of northern California: effects of thinning and land Northwest USA: contrasting the landscape ecology of the pre-settlement and prescribed fire. For. Ecol. Manage. 247, 200–208. modern eras. For. Ecol. Manage. 211, 117–139. Roccaforte, J.P., Fulé, P.Z., Covington, W.W., 2010. Monitoring landscape-scale pon- Hessburg, P.F., Churchill, D.J., Larson, A.J., Haugo, R.D., Miller, C., Spies, T.A., North, derosa pine restoration treatment implementation and effectiveness. Restor. Ecol. 18, M.P., Povak, N.A., Belote, R.T., Singleton, P.H., Gaines, W.L., Keane, R.E., Aplet, 820–833. G.H., Stephens, S.L., Morgan, P., Bisson, P.A., Rieman, B.E., Salter, R.B., Reeves, G.H., Rodman, K.C., Sánchez Meador, A.J., Moore, M.M., Huffman, D.W., 2017. Reference 2015. Restoring fire-prone Inland Pacific landscapes: seven core principles. Landsc. conditions are influenced by the physical template and vary by forest type: a Ecol. 30, 1805–1835. synthesis of Pinus ponderosa-dominated sites in the southwestern United States. For. Hurteau, M.D., Bradford, J.B., Fulé, P.Z., Taylor, A.H., Martin, K.L., 2014. Climate Ecol. Manage. 404, 316–329. change, fire management, and ecological services in the southwestern US. For. Ecol. Rothermel, R.C. How to predict the spread and intensity of forest and range fires. Gen. Manage. 327, 280–289. Tech. Rep. INT-143. Oggen, UT: US Department of Agriculture, Forest Service, Keyser, C.E., 2008. Inland California and southern Cascades (CA) Variant Overview – Intermountain Forest and range Experinement Station. 161p; 1983. Forest Vegetation Simulator. Fort Collins CO. Runkle, J.R., 2000. Canopy tree turnover in old-growth mesic forests of eastern North Knapp, E.E., Skinner, C.N., North, M.P., Estes, B.L., 2013. Long-term overstory and un- America. Ecology 81, 554–567. derstory change following logging and fire exclusion in a Sierra Nevada mixed-con- Ryan, K.C., Reinhardt, E.D., 1988. Predicting postfire mortality of seven western conifers. ifer forest. For. Ecol. Manag. 310, 903–914. Can. J. For. Res. 18, 1291–1297. Kroeber, T. Ishi in Two Worlds: A Biography of the Last Wild Indian in North America – Ryan, K.C., Knapp, E.E., Varner, J.M., 2013. Prescribed fire in North American forests and Theodora Kroeber – Google Books; 1961. woodlands: history, current practice, and challenges. Front. Ecol. Environ. 11, Larson, A.J., Belote, R.T., Cansler, C.A., Parks, S.A., Dietz, M.S., 2013. Latent resilience in e15–e24. ponderosa pine forest: effects of resumed frequent fire. Ecol. Appl. 23, 1243–1249. Sánchez Meador, A.J., Moore, M.M., Bakker, J.D., Parysow, P.F., 2009. 108 years of Larson, A.J., Churchill, D., 2012. Tree spatial patterns in fire-frequent forests of western change in spatial pattern following selective harvest of a Pinus ponderosa stand in North America, including mechanisms of pattern formation and implications for northern Arizona, USA. J. Veg. Sci. 20, 79–90. designing fuel reduction and restoration treatments. For. Ecol. Manage. 267, 74–92. Sánchez Meador, A.J., Parysow, P.F., Moore, M.M., 2011. A new method for delineating Latif, Q.S., Saab, V.A., Mellen-Mclean, K., Dudley, J.G., 2015. Evaluating habitat suit- tree patches and assessing spatial reference conditions of Ponderosa Pine forests in ability models for nesting white-headed woodpeckers in unburned forest. J. Wildl. Northern Arizona. Restor. Ecol. 19, 490–499. Manage. 79, 263–273. Savage, M., Mast, J.N., 2005. How resilient are southwestern ponderosa pine forests after Laudenslayer, W., Darr, H., Smith, S., 1989. Historical effects of forest management crown fires? Can. J. For. Res. 35, 967–977. practices on eastside pine communities in northeastern California. USDA For. Serv. Schneider, E.E., Sanchez Meador, A.J., Covington, W.W., 2016. Reference conditions and Gen. Tech. Rep. RM-267, 37–387. historical changes in an unharvested ponderosa pine stand on sedimentary soil. Laughlin, D.C., Fulé, P.Z., Huffman, D.W., Crouse, J., Laliberté, E., 2011. Climatic con- Restor. Ecol. 24, 212–221. straints on trait-based forest assembly. J. Ecol. 99, 1489–1499. Scholl, A.E., Taylor, A.H., 2010. Fire regimes, forest change, and self-organization in an Long, J.W., Goode, R.W., Gutteriez, R.J., Lackey, J., Anderson, M.K., 2016. Managing old-growth mixed-conifer forest, Yosemite national park, USA. Ecol. Appl. 20, Califronia black oak for tribal ecocultural restoration. J. For. 115, 426–434. 362–380.

301 N.C. Pawlikowski et al. Forest Ecology and Management 434 (2019) 289–302

Scott, J.H., Burgan, R.E., Standard Fire Behavior Fuel Models : A Comprehensive Set for USA, 1600–2015 CE. Proc. Natl. Acad. Sci. USA 113, 13684–13689. Use with Rothermel â€TM s Surface Fire Spread Model of Agriculture Standard Fire Taylor, A.H., Vandervlugt, A.M., Maxwell, R.S., Beaty, R.M., Airey, C., Skinner, C.N., Behavior Fuel Models : A Comprehensive Set for Use with Rothermel ’ s Surface Fire 2014. Changes in forest structure, fuels and potential fire behaviour since 1873 in the Spread Model. Gen. Tech. Report, USDA For. Serv; 2005. Lake Tahoe Basin. USA Appl. Veg. Sci. 17, 17–31. Simard, S.W., Beiler, K.J., Bingham, M.A., Deslippe, J.R., Philip, L.J., Teste, F.P., 2012. Thompson, J.R., Spies, T.A., 2010. Factors associated with crown damage following re- Mycorrhizal networks: mechanisms, ecology and modelling. Fungal Biol. Rev. 26, curring mixed-severity wildfires and post-fire management in southwestern Oregon. 39–60. Land. Ecol. 25, 775–789. Skinner, C.N., 1995. Change in spatial characteristics of forest openings in the Klamath U.S.D.A, U.S.D.I of, 1998. Final Report on Fire Management Policy. Washington DC. Mountains of northwestern California. USA. Landsc. Ecol. 10, 219–228. van Mantgem, P.J., Stephenson, N.L., Byrne, J.C., Daniels, L.D., Franklin, J.F., Fulé, P.Z., Skinner, C.N., Taylor, A.H. Southern Cascades Bioregion. In: van J.W. Wagtendonk, J.W., Harmon, M.E., Larson, A.J., Smith, J.M., Taylor, A.H., Veblen, T.T., 2009. NG Sugihara, Scott L. Stephens, AE Thode, K.E. Schafer, J. Fites-Kaufmann (eds.), Widespread increase of tree mortality rates in the western United States. Science 323, 2018. Fire in California's Ecosystems. Univ of California Press, Berkeley. pp 195-218; 521–524. 2018. van Wagner, C.E., 1977. Conditions for the start and spread of crown fire. Can. J. For. Res. Skinner, C.N., Taylor, A.H. and Agee, J.K., Briles, C.E., Whitlock, C.L. Klamath mountains 7, 23–34. bioregion. In: van J.W. Wagtendonk, J.W., NG Sugihara, Scott L. Stephens, AE Thode, van Wagner, C.E., 1993. Prediction of crown fire behavior in two stands of jack pine. Can. K.E. Schafer, J. Fites-Kaufmann (eds.), 2018. Fire in California's Ecosystems. Univ of J. For. Res. 23, 442–449. California Press, Berkeley. pp 171–194; 2018. Van de Water, K., North, M., 2011. Stand structure, fuel loads, and fire behavior in ri- Stephens, S.L., Fry, D.L., Franco-Vizcaino, E., 2008. Wildfire and spatial patterns in forests parian and upland forests, Sierra Nevada Mountains, USA; a comparison of current in northwestern Mexico: the United States wishes it had similar fire problems. Ecol. and reconstructed conditions. For. Ecol. Manage. 262, 215–228. Soc. 13, 12p. van Wagtendonk, J.W., Sugihara, N.G., Stephens, S.L., Thode, A.E., Shaffer, K.E., Fites- Stephens, S.L., Lydersen, J.M., Collins, B.M., Fry, D.L., Meyer, M.D., 2015. Historical and Kaufman, 2018. J. Fire California’s Ecosyst. current landscape-scale ponderosa pine and mixed conifer forest structure in the Vernon, M.J., Sherriff, R.L., van Mantgem, P., Kane, J.M., 2018. Thinning, tree-growth, Southern Sierra Nevada. Ecosphere 6, 1–63. and resistance to multi-year drought in a mixed-conifer forest of northern California. Stephens, S.L., Gill, S.J., 2005. Forest structure and mortality in an old-growth Jeffery For. Ecol. Manage. 422, 190–198. pine-mixed conifer forest in north-western Mexico. For. Ecol. Manage. 205, 15–28. Walker, R.B., Coop, J.D., Parks, S.A., Trader, L., 2018. Fire regimes approaching historic Taylor, A.H., 2004. Identifing forest reference conditions on early cut-over lands, Lake norms reduce wildfire-facilitated conversion from forest to non-forest. Ecosphere. Tahoe Basin. USA. Ecol. Appl. 14, 1903–1920. White, A.S., 1985. Presettlement regeneration patterns in a southwestern ponderosa pine Taylor, A.H., 2010. Fire disturbance and forest structure in an old-growth Pinus pon- stand. Ecology 66, 589–594. derosa forest, southern Cascades. USA. J. Veg. Sci. 21, 561–572. Yamada, I., Rogerson, P.A., 2003. An empirical comparison of edge effect correction Taylor, A.H., Jinyan, H., ShiQiang, Z., 2004. Canopy tree development and undergrowth methods applied to K-function analysis. Geogr. Anal. 35, 97–109. bamboo dynamics in old-growth Abies-Betula forests in southwestern China: a 12- York, R.A., Heald, R.C., Battles, J.J., York, J.D., 2004. Group selection management in year study. For. Ecol. Manage. 200, 347–360. conifer forests: relationships between opening size and tree growth. Can. J. For. Res. Taylor, A.H., Trouet, V., Skinner, C.N., Stephens, S., 2016. Socioecological transitions 34, 630–641. trigger fire regime shifts and modulate fire-climate interactions in the Sierra Nevada,

302