Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2022 May 26.
Published in final edited form as: For Ecol Manage. 2018 Aug 29;430:460–471. doi: 10.1016/j.foreco.2018.08.034

Looking beyond the mean: Drivers of variability in postfire stand development of conifers in Greater Yellowstone

Kristin H Braziunas a,*, Winslow D Hansen a, Rupert Seidl b, Werner Rammer b, Monica G Turner a
PMCID: PMC7612775  EMSID: EMS145169  PMID: 35645456

Abstract

High-severity, infrequent fires in forests shape landscape mosaics of stand age and structure for decades to centuries, and forest structure can vary substantially even among same-aged stands. This variability among stand structures can affect landscape-scale carbon and nitrogen cycling, wildlife habitat availability, and vulnerability to subsequent disturbances. We used an individual-based forest process model (iLand) to ask: Over 300 years of postfire stand development, how does variation in early regeneration densities versus abiotic conditions influence among-stand structural variability for four conifer species widespread in western North America? We parameterized iLand for lodgepole pine (Pinus contorta var. latifolia), Douglas-fir (Pseudotsuga menziesii var. glauca), Engelmann spruce (Picea engelmannii), and subalpine fir (Abies lasiocarpa) in Greater Yellowstone (USA). Simulations were initialized with field data on regeneration following stand-replacing fires, and stand development was simulated under historical climatic conditions without further disturbance. Stand structure was characterized by stand density and basal area. Stands became more similar in structure as time since fire increased. Basal area converged more rapidly among stands than tree density for Douglas-fir and lodgepole pine, but not for subalpine fir and Engelmann spruce. For all species, regeneration-driven variation in stand density persisted for at least 99 years postfire, and for lodgepole pine, early regeneration densities dictated among-stand variation for 217 years. Over time, stands shifted from competition-driven convergence to environment-driven divergence, in which variability among stands was maintained or increased. The relative importance of drivers of stand structural variability differed between density and basal area and among species due to differential species traits, growth rates, and sensitivity to intraspecific competition versus abiotic conditions. Understanding dynamics of postfire stand development is increasingly important for anticipating future landscape patterns as fire activity increases.

Keywords: Stand structure, Variability, Forest development, Process-based modeling, Greater Yellowstone Ecosystem, Wildfire

1. Introduction

Large, high-severity, infrequent disturbances such as fires can shape landscape patterns of forest age, structure, and species composition for decades to centuries (Foster et al., 1998). Warming climate and concomitant increases in fire activity will likely reset forest succession across larger expanses of the western United States (Westerling et al., 2006; Abatzoglou and Williams, 2016; Westerling, 2016). Therefore, understanding how stands develop after fire is critical for anticipating future forest landscapes. This is particularly important in the Northern Rocky Mountains (USA), where decadal area burned increased 889% from the 1970s to the early 2000s (Westerling, 2016) and 34% of area burned across all forest types is stand-replacing fire (41% in subalpine and 25% in mid-montane forests; Harvey et al., 2016a). During large fire years, stand-replacing fire can exceed 50% of area burned (Turner et al., 1994). In the Greater Yellowstone Ecosystem (GYE) within the Northern Rocky Mountains, successional dynamics in subalpine forests have been influenced by infrequent (100−300 year fire return interval), high-severity (i.e., stand-replacing) fires throughout the Holocene (Romme and Despain, 1989; Millspaugh et al., 2004; Schoennagel et al., 2004; Whitlock et al., 2008; Higuera et al., 2011).

Among-stand variation in forest structure over stand development has received surprisingly little attention in studies of postfire stand trajectories (but see Kashian et al., 2005a, 2005b). Structure and function can vary considerably among stands of the same age (e.g., Turner, 2010), with substantial implications for carbon pools and fluxes (Litton et al., 2004; Turner et al., 2004; Bradford et al., 2008; Kashian et al., 2013), nitrogen pools and fluxes (Smithwick et al., 2009a, 2009b; Turner et al., 2009), wildlife habitat (Tews et al., 2004), and vulnerability to subsequent disturbances (Bebi et al., 2003; Seidl et al., 2016a). Due to high variation in stand structure following fire, simple descriptions of mean conditions within these forests might overlook important information about the ecological dynamics of a landscape (Fraterrigo and Rusak, 2008).

Initially distinct post-disturbance stands may converge over time due to competition and environmental constraints or follow distinct trajectories if the effects of initial post-disturbance regeneration and environmental heterogeneity persist over time (Glenn-Lewin and van der Maarel, 1992; Walker and del Moral, 2003; Tepley et al., 2013; Meigs et al., 2017). Postfire stand development pathways differ among species based on their fire adaptations, tolerances, and growth rates (Baker, 2009; Knight et al., 2014). For example, species that exhibit serotiny depend on a canopy seedbank that must be released by an environmental trigger such as fire (Crossley, 1956). Serotinous species [e.g., lodgepole pine (Pinus contorta var. latifolia), jack pine (Pinus banksiana)] can recruit in abundance following stand-replacing fire (Turner et al., 2004; Buma et al., 2013; Pinno et al., 2013; Edwards et al., 2015). In the Northern Rocky Mountains, postfire lodgepole pine densities vary widely as a result of broad-scale gradients in prefire serotiny (Tinker et al., 1994; Turner et al., 1997; Schoennagel et al., 2003). In contrast, other species must disperse into recently burned areas (Baker, 2009). Following severe stand-replacing fire, which kills all trees and consumes the shallow litter layer, tree seedling establishment occurs on mineral soil (Turner et al., 1997, 1999), and early seedling survival varies with climate (Harvey et al., 2016b; Stevens-Rumann et al., 2018).

Variation in early regeneration densities results in differing levels of competition for light and other resources in postfire stands, which in turn may differentially affect stand development pathways depending on species traits. For example, species that are tolerant of resource-limited conditions [e.g., shade-tolerant subalpine fir (Abies lasiocarpa) and Engelmann spruce (Picea engelmannii; Oosting and Reed, 1952; Alexander, 1987)] may continue to establish and survive in the understory for decades following disturbance (Veblen, 1986; Aplet et al., 1988), enabling convergence in stand density. Alternatively, species whose diameter growth rates are more sensitive to competition, such as lodgepole pine in comparison with other Rocky Mountain conifers (Buechling et al., 2017), may be likely to tend toward similar basal areas among stands of different densities. However, high-severity fire occurs in forests that span a broad range of climatic and topoedaphic conditions (Turner and Romme, 1994; Harvey et al., 2016a), and this abiotic heterogeneity may outweigh the effects of competition-driven convergence and instead maintain or increase variation among stands during postfire stand development.

Stand development unfolds over long periods of time and under changing climate, and models that can project variation in future stand structures are needed to explore these long-term trajectories and inform possible management practices. Models built on statistical relationships between environmental drivers and tree responses (i.e., empirical models) play an important role in forest management and form the basis for the development of more complex models based on mechanistic understanding of forest processes (Korzukhin et al., 1996). However, empirical models may fail to predict stand structures and forest landscapes under changing environmental drivers, whereas process-based models can improve projections of future forest conditions (Korzukhin et al., 1996; Cuddington et al., 2013; Gustafson, 2013; Reyer et al., 2015). Modeling ecological processes and variables at scales appropriate to phenomena, such as competition for resources at the individual-tree level, also allows broader-scale patterns to emerge from finer-scale interactions (Grimm et al., 2017; Scholes, 2017).

1.1. Objectives

We adapted and parameterized iLand, a process-based forest simulation model (Seidl et al., 2012) for four widespread conifer species in the Greater Yellowstone Ecosystem: Lodgepole pine, Douglas-fir (Pseudotsuga menziesii var. glauca), Engelmann spruce, and subalpine fir. We then conducted a simulation experiment to address the question: Over 300 years of postfire stand development, how does variation in early regeneration densities versus abiotic conditions influence among-stand structural variability for four conifer species widespread in western North America? We expected variation in early regeneration densities to drive structural variability among young stands and variation in abiotic drivers to become increasingly important as stands aged. We further expected early regeneration densities to influence stand structural variability of lodgepole pine for a longer period of time than other conifers, due to its wider range of initial stem densities resulting from variation in prefire serotiny.

2. Methods

2.1. Study area

The Greater Yellowstone Ecosystem in the Northern Rocky Mountains of the United States comprises 89,000 km2 (YNP, 2017) primarily in northwest Wyoming, and also in southeast Idaho and southwest Montana. The majority of the GYE is federally managed land, anchored by Yellowstone and Grand Teton National Parks and adjacent national forests, and natural processes such as disturbance and regeneration occur with minimal intervention across large areas within this relatively intact forested landscape. Forests in the GYE span a broad elevation gradient (1800 to 3050 m) and include mid-montane forests at lower elevations dominated by interior Douglas-fir, mid-elevation subalpine forests dominated by lodgepole pine, and higher-elevation subalpine forests dominated by Engelmann spruce and subalpine fir (Despain, 1990; Knight et al., 2014). Climate is warmer and drier at lower elevations, and cooler and wetter at higher elevations (mean annual temperatures range from −1.3 to 4.3 °C and mean annual precipitation from 444 to 1400 mm; Fig. 1). Soils in the central area of Greater Yellowstone are nutrient poor and largely derived from underlying rhyolitic or andesitic bedrock (Despain, 1990; Rodman et al., 1996). Soil inorganic nitrogen availability is very low, even following disturbance (Turner et al., 2007), and postfire tree nitrogen uptake is facilitated by associations with ectomycorrhizal communities (Douglas et al., 2005; Smithwick et al., 2009a). Rhyolite-derived soils, which cover most of Yellowstone's lodgepole pine-dominated central plateau, are coarser and less fertile than andesite-derived soils (Despain, 1990; Whitlock, 1993).

Fig. 1.

Fig. 1

Open in a new tab

Climate envelope for evaluation and simulation experiments, characterized by mean annual precipitation and mean annual temperature (derived from 1980 to 2015 daily climate data; Thornton et al., 2017) for each species. Each simulated stand is represented by one point within this climate space. Median climate conditions used in no among-stand variation scenarios (Regeneration varies and Neither vary) are shown in red. Subalpine fir and Engelmann spruce have the same median climate.

Forests in the GYE have been shaped by historical fire regimes (Romme, 1982). Fire return intervals are longer and percent serotiny of lodgepole pine is lower at higher elevations (Schoennagel et al., 2003). Of the four focal species, only mature Douglas-fir with its very thick bark is adapted to survive fire, and fire regimes in lower-elevation Douglas-fir forests are typically characterized as mixed severity (Baker, 2009). However, stand-replacing fires can occur in all forest types (Baker et al., 2007; Harvey et al., 2016a). Postfire regeneration densities following stand-replacing fire are extremely variable (Table 1) based on prefire stand conditions, local burn severity, and the spatial pattern of fires, which determine distances to seed sources (Turner et al., 1997, 1999). For example, regenerating lodgepole pine stands ranged from 0 to > 500,000 stems ha−1 11 years following stand-replacing fire (Turner et al., 2004). High-severity fire behavior in this region is primarily weather-driven (e.g., drought, wind; Schoennagel et al., 2004; Higuera et al., 2011).

Table 1.

Initial conditions and drivers for simulated stands. Median values indicate the no among-stand variation condition.

Douglas-fir
(n = 34)		 Lodgepole pine
(n = 70)		 Subalpine fir
(n = 38)		 Engelmann spruce
(n = 39)
Variable (units) Min-Max Mean (SE)
Median		 Min-Max Mean (SE)
Median		 Min-Max Mean (SE)
Median		 Min-Max Mean (SE)
Median
Time since fire (years)* – 24 (–) – 24 (–) 10–19 14 (0.4) 10–19 14 (0.5)
24 24 13 13
Postfire regeneration *
Density of trees, saplings, and seedlings (stems ha−1) 14–13,653 2224 (490) 33–344,075 21,446 (6546) 14–3154 268 (90) 14–11,997 610 (327)
1370 4050 83 83
Climate †
Mean annual temperature (°C) 1.1–4.3 2.9 (0.1) 0.0–2.6 1.2 (0.1) −1.2–3.1 1.0 (0.2) −1.3–3.1 0.8 (0.2)
3.0 1.1 0.7 0.6
Annual precipitation (mm) 444–787 637 (14) 629–1400 888 (22) 741–981 847 (11) 741–981 855 (11)
642 853 866 877
Daily global radiation (MJ m−2) 15.3–18.4 16.4 (0.1) 16.8–19.8 18.5 (0.1) 16.7–20.3 18.6 (0.2) 16.7–20.3 18.7 (0.2)
16.2 18.5 18.6 18.8
Daily vapor pressure deficit (kPa) 0.42–0.51 0.45 (0.00) 0.40–0.43 0.42 (0.00) 0.38–0.46 0.42 (0.00) 0.38–0.46 0.42 (0.00)
0.44 0.42 0.41 0.41
Soils
Effective depth (cm)‡ 83–152 102 (3) 83–137 114 (3) 83–152 111 (5) 83–152 106 (4)
106 106 86 86
Sand (%)‡ 30–71 50 (1) 52–56 54 (0) 30–56 48 (1) 30–56 48 (1)
48 54 53 53
Silt (%)‡ 21–49 33 (1) 30–33 32 (0) 30–49 35 (1) 30–49 35 (1)
33 33 32 32
Clay (%)‡ 8–22 17 (1) 13–15 14 (0) 13–21 17 (1) 13–21 17 (1)
19 14 15 15
Substrate§ – Andesite – Rhyolite – Andesite – Andesite
Open in a new tab
*

Postfire regeneration densities and time since fire from field data (Donato et al., 2016; Harvey et al., 2016b; Turner et al., 2016).

†

Climate data from Daymet Version 3 (Thornton et al., 2017), extracted using geographic coordinates of field data* or field-verified forest type (Simard et al., 2012).

‡

Soil depth and texture from CONUS-SOIL (Miller and White, 1998), extracted using geographic coordinates of field data* or field-verified forest type (Simard et al., 2012).

§

Soil substrate assigned based on parent material associated with each forest type (Despain, 1990; Knight et al., 2014). Substrate was used to derive relative fertility.

2.2. Simulation model

2.2.1. Model overview

We simulated stand (1 ha) development using the individual-based forest landscape and disturbance model iLand (Seidl et al., 2012), which integrates species-specific responses to environmental drivers such as light availability, temperature, precipitation, soil moisture, and nutrient levels. These environmental drivers limit seedling cohort establishment and modify sapling cohort growth and survival. Trees > 4 m in height are represented as individuals (2-m spatial resolution) in their responses to resource availability. Limitations to physiological processes (e.g., temperature and water limitation) are considered at a daily time step, and stand structure is updated annually. Nutrient levels (i.e., soil relative fertility) are expressed as plant available nitrogen and modify tree growth according to a species-specific nitrogen response class. Processes such as seed dispersal and competition for light are spatially explicit. Light availability for an individual tree is attenuated based on the heights and crown characteristics of neighboring trees. The growth allocation of individual trees to height and diameter responds dynamically to light competition within a species-specific range. In the absence of disturbance, variation in tree sizes and forest structure within stands (i.e., 100 × 100 m grid cells of homogeneous climate and soil conditions) emerges from these fine-scale tree-level dynamics, while among-stand variability is also driven by differences in environmental conditions between stands (Seidl et al., 2012). iLand has been demonstrated to work well in forested ecosystems in the Pacific Northwest (Seidl et al., 2012, 2014b) and in Europe (Seidl et al., 2014a; Silva Pedro et al., 2015; Thom et al., 2017), and Hansen et al. (2018) evaluated iLand's representation of early postfire regeneration dynamics in lodgepole pine and Douglas-fir forests in Yellowstone. Extensive model documentation is available at http://iland.boku.ac.at/ (Seidl and Rammer, 2018).

2.2.2. Model parameterization

We parameterized iLand for the four dominant conifer species in the Greater Yellowstone Ecosystem (Appendix A). Most species-specific traits and parameters were sourced from peer-reviewed literature and government reports, and a few parameters (e.g., height-to-diameter ratios, aging) were fit or iteratively derived by simulating stand development of initial conditions (see Seidl et al., 2012; Seidl and Rammer, 2018).

Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.foreco.2018.08.034.

2.2.3. Model evaluation

Because iLand had not previously been used for our focal species in the Greater Yellowstone Ecosystem, we undertook a three-stage evaluation of iLand to assess the model's ability to simulate stand structural development in the region, including single-species, mixed-species, and model comparison assessments (Appendix B). The model appropriately characterized monospecific stand structural trajectories and variability, encompassing a majority of independent field observations over 300 years of stand development (Figs. B.1 and B.2); reproduced expected successional trajectories from bare ground to Douglas-fir, lod-gepole pine, and spruce-fir forest types given appropriate seed source species composition and abiotic conditions (Fig. B.5); and responded consistently to variation in initial stem densities and environmental conditions with the forest growth model Forest Vegetation Simulator (FVS; Dixon, 2002; Crookston and Dixon, 2005) for at least 50 years of simulation (Figs. B.9−B.11). Overall, iLand performed well across multiple evaluations, supporting its use in studying the relative importance of factors influencing among-stand structural variability over hundreds of years of stand development in the Greater Yellowstone Ecosystem.

2.3. Initial conditions and drivers

We initialized model simulations using species-specific field data recorded in previously published studies of postfire regeneration following stand-replacing fires in the GYE (Table 1, Fig. 2; Donato et al., 2016; Harvey et al., 2016b; Turner et al., 2016). Individual trees (densities, diameters, and heights) and the number and height of sapling and seedling cohorts were initialized for each simulated 1-ha stand. Because our interest here was in long-term stand development once early postfire establishment had occurred, we began simulations with stands that were 10−25 years postfire (but see Hansen et al., 2018 for simulations of early postfire establishment).

Fig. 2.

Fig. 2

Open in a new tab

(Inset) Location of our study area within North America. (Main) Plot locations (Simard et al., 2012; Donato et al., 2016; Harvey et al., 2016b; Turner et al., 2016) within the Greater Yellowstone Ecosystem used for data on postfire regeneration, climate, and soils to simulate stand development in iLand. Fire years for postfire regeneration plots are indicated with differential shading, and additional climate and soil plots were identified to better encompass the expected range of variability in abiotic conditions for each species. Data sources: ESRI, Tele Atlas, National Atlas of the United States, YNP Spatial Analysis Center, and Monitoring Trends in Burn Severity (MTBS).

For each postfire stand from the same published studies, we extracted daily historical (1980−2015) climate drivers (temperature, vapor pressure deficit, precipitation, and radiation) from Daymet Version 3 (Thornton et al., 2017) and soil depth and texture from CONUS-SOIL (Miller and White, 1998) using geographic coordinates of plot centers (Fig. 2). Available field plots for lodgepole pine regeneration were widely distributed across Yellowstone National Park. However, some field plots for Douglas-fir, Engelmann spruce, and subalpine fir stands were in close proximity, resulting in replicated climate conditions, and higher-elevation Engelmann spruce-subalpine fir climates were underrepresented. To adequately simulate expected ranges of abiotic variability for each species, we extracted additional climate and soil conditions in the GYE based on field-verified forest type (see Figs. 1 and 2; Simard et al., 2012). Soil relative fertility was held constant by species to control for inconsistent field data on soil substrate, and simulated stands were assigned a fertility value based on associations between species and soil parent material in this region (i.e., rhyolite parent material for lodgepole pine and andesite parent material for Douglas-fir, Engelmann spruce, and subalpine fir; Despain, 1990; Knight et al., 2014).

Initial conditions and drivers represented a wide range of early regeneration densities and abiotic conditions for each species (Table 1, Fig. 1). Initial stem densities varied by at least two and up to four orders of magnitude. All species spanned a range of at least 2.5 °C mean annual temperature, 240 mm mean annual precipitation, and 50 cm effective soil depth. Mean annual temperature and precipitation ranges in our climate data were also representative of the longer-term (starting as early as 1881 through 2012) historical climate record from weather stations throughout the GYE (WRCC, 2018).

2.4. Simulation experiment

We conducted a 2-by-2 factorial simulation experiment to assess the influence of two distinct drivers of among-stand structural variability, early regeneration densities or abiotic (climate plus soil depth and texture) conditions. All stands were either simulated with observed variation in postfire regeneration and abiotic conditions or with no among-stand variation, resulting in four scenarios (Both vary, Abiotic varies, Regeneration varies, Neither varies). A representative abiotic and early regeneration stand was derived for each species based on the central tendency of observed drivers (i.e., median soil depth and texture, climate period with median vapor pressure deficit, stand with median postfire regeneration stem density; Table 1, Fig. 1). For scenarios with no among-stand variation in abiotic conditions or in early regeneration, the respective representative stand conditions were assigned to all simulated stands for a given species. Abiotic conditions were randomly assigned to early postfire regeneration densities by species in the Both vary scenario.

For each scenario, we simulated postfire development of mono-specific 1-ha stands for 300 years with no additional disturbance under historical climate conditions (n = 20 replicates per scenario). Climate year was randomly drawn with replacement from 1980 to 2015. Initial trees and saplings within a stand were the only seed source for subsequent tree recruitment.

2.5. Model outputs and analysis

Stand structure for each species was characterized using two metrics, stand density and stand basal area. Both metrics were calculated each year for trees > 4 m in height. The coefficient of variation (CV) was used to quantify variation in structure among stands of the same species and age (as in Kashian et al., 2005a, 2005b). As a relative estimate of variance, the CV enables comparisons among datasets with different means (Fraterrigo and Rusak, 2008). Because the coefficient of variation can be sensitive to low mean values, CVs were only calculated when mean stand density was ≥ 50 trees ha−1 and mean basal area was ≥ 2 m2 ha −1 (5% of approximate stand density and basal area of a mature stand).

For each species and scenario, mean stand density, mean stand basal area, and mean CV (among the n = 20 replicates) were calculated for each year. We first assessed differences in stand structure convergence among species in the Both vary scenarios, in which both abiotic conditions and early regeneration densities varied among stands, based on when CV declined below 50% and when mean stand density peaked. Past studies in this region have documented convergence at CVs slightly below 50% (Kashian et al., 2005b). We next assessed similarities and differences among the four scenarios for each species. For Abiotic varies and Regeneration varies scenarios, we characterized time periods at which abiotic drivers versus early regeneration densities were more important in influencing among-stand variation using the point of intersection of mean CVs and the overlap of CV ranges (minimum to maximum CV across the 20 replicates). Trends among species were compared based on time since fire in years. R statistical software (R Core Team, 2017) was used for all analyses of model outputs.

3. Results

3.1. Variation and convergence in among-stand structural trajectories (Both vary scenario)

When stands varied both in their early postfire densities and their abiotic conditions (Both vary scenario), among-stand CVs for stand density and basal area were highest at or near the beginning of the simulation. Stands eventually converged in density (Fig. 3) and in basal area (Fig. 4), but the timing of convergence differed among species. Stand density converged most rapidly for lodgepole pine (mean among-stand density CV declined below 50% by 83 years postfire), followed by subalpine fir (105 years), Engelmann spruce (106 years), and Douglas-fir (143 years; Fig. 3). Mean stand density (trees > 4 m height) peaked earlier in stand development for both Douglas-fir (468 trees ha −1 at 37 years postfire) and lodgepole pine (2960 trees ha −1 at 31 years postfire) compared to subalpine fir and Engelmann spruce (817 trees ha−1 and 1005 trees ha−1 at 149 and 143 years postfire, respectively). At peak stand density, among-stand variation was greater for Douglas-fir and lodgepole pine (mean CV = 108% for Douglas-fir and 112% for lodgepole pine) than for subalpine fir (mean CV = 32%) or Engelmann spruce (mean CV = 24%). By 300 years postfire, among-stand variability in density had declined to ≤ 45% mean CV for all species.

Fig. 3.

Fig. 3

Open in a new tab

(a-h) Mean stand density and mean among-stand density CV for the four simulation scenarios over 300 years of postfire stand development (n = 20 replicates of each). All metrics are for trees > 4 m in height. All vertical axes are on a log10 scale to facilitate comparison over time and among simulations. Convergence to CV = 50% is indicated by a dashed black line.

Fig. 4.

Fig. 4

Open in a new tab

(a−h) Mean stand basal area and mean among-stand basal area CV for the four simulation scenarios over 300 years of postfire stand development (n = 20 replicates of each). All metrics are for trees > 4m in height. The vertical axis for CVs (e−h) is on a log10 scale to facilitate comparison over time and among simulations. Convergence to CV = 50% is indicated by a dashed black line.

Basal area converged more rapidly than stand density for Douglas-fir and lodgepole pine, with mean among-stand basal area CV falling below 50% by 58 and 34 years postfire, respectively (Fig. 4). However, basal area and stand density converged at similar times for subalpine fir and Engelmann spruce (mean CV declined below 50% by 111 and 107 years postfire, respectively). By 300 years postfire, among-stand variability in basal area had declined to a mean CV of ≤ 23% for all species.

3.2. Influence of early regeneration densities versus abiotic conditions on stand structural variability (among-scenario comparisons)

Trajectories of mean tree density were similar over time across the four scenarios (Fig. 3a-d), but trajectories of among-stand variation in density differed among scenarios and species (Fig. 3e-h). Early in stand development, mean CVs for stand density were similar when early regeneration densities and abiotic conditions both varied (Both vary) and when only regeneration densities varied (Regeneration varies), but mean CVs were much lower when only abiotic conditions varied (Abiotic varies). When only early regeneration densities varied, mean CVs declined over time as stand density converged, eventually reaching a value lower than the Both vary scenario. In contrast, when only abiotic conditions varied, mean CVs were initially lower but declined less rapidly and in some cases increased over time. By 300 years postfire, Abiotic varies and Both vary scenarios had similar mean CVs, both of which were greater than mean CVs for Regeneration varies.

For all species, early regeneration densities were the most important driver of among-stand variation in density early in stand development (when Regeneration varies scenarios had the highest mean CV), and abiotic drivers were most important later in stand development (when Abiotic varies scenarios had the highest mean CV). However, species differed in both the postfire year at which the most important driver switched from early regeneration densities to abiotic conditions (point of intersection between Regeneration varies and Abiotic varies scenarios in Fig. 5i-j) and in the time period during which both drivers similarly influenced among-stand variation (overlap in CV ranges across all n = 20 replicates per scenario; Fig. 5).

Fig. 5.

Fig. 5

Open in a new tab

(a−h) Ranges of CVs (min to max) across n = 20 replicates each of Abiotic varies and Regeneration varies scenarios. (i−j) Timeline plot for all four species, showing time since fire years when Regeneration CVs > Abiotic CVs (red), years when Regeneration and Abiotic CV ranges overlapped (purple), and years when Abiotic CVs > Regeneration CVs (blue). The range of Regeneration varies and Abiotic varies overlap encompasses the first to last point of overlap, and in some cases includes non-overlapping years. Points indicate the point of intersection of the mean trajectories for each species. Psme = Douglas-fir, Pico = lodgepole pine, Abla = subalpine fir, and Pien = Engelmann spruce.

Variation in early postfire regeneration influenced among-stand variability in lodgepole pine density for a longer period of time than other species. Lodgepole pine Abiotic and Regeneration varies scenarios intersected at 217 years postfire (Fig. 5c). The point of intersection was earliest for subalpine fir stands (99 years postfire; Fig. 5e), followed by Engelmann spruce (149 years postfire; Fig. 5g) and Douglas-fir (174 years postfire; Fig. 5a).

Mean basal area also followed similar trajectories among all four scenarios (Fig. 4a-d), whereas among-stand variation in basal area differed among scenarios and species (Fig. 4e-h). In general, mean among-stand basal area variability behaved similarly to mean among-stand density variability over time for each of the four scenarios. Among-stand variation was initially high when only early regeneration densities varied (Regeneration varies) or when both regeneration and abiotic conditions varied (Both vary), but mean CV declined over time. When only abiotic conditions varied (Abiotic varies), among-stand variation was initially lower. However, the Abiotic varies mean CV equaled (point of intersection) and then surpassed the Regeneration varies mean CV over time.

The relative importance of abiotic conditions versus early regeneration densities as drivers of among-stand structural variability differed between basal area and density for a given species. For example, the point of intersection of mean CVs in Abiotic and Regeneration varies scenarios was earlier for basal area compared to stand density for all species except subalpine fir (Fig. 5i-j). Lodgepole pine had the earliest point of intersection (67 years postfire; Fig. 5d), and mean CVs for subalpine fir (Fig. 5f), Douglas-fir (Fig. 5b), and Engelmann spruce (Fig. 5h) intersected at similar times in stand development (125, 129, and 136 years postfire, respectively). For all species other than Douglas-fir, the overlap in CV ranges between Abiotic and Regeneration varies scenarios ended earlier compared to density (Fig. 5i-j). The ranges of Douglas-fir basal area CVs overlapped until the end of the simulation (300 years postfire).

4. Discussion

Here we show that variation in early postfire regeneration densities affects stand structural trajectories for decades to centuries for four widespread conifer species. Variation in early regeneration densities was particularly important in shaping long-term patterns of lodgepole pine stand densities. Among-stand structural variability was highest in young stands for all species, and stand structures converged with time since fire. Basal area converged more rapidly than stand density for Douglas-fir and lodgepole pine, but not for subalpine fir and Engelmann spruce. Differential responses among species correspond to variation in life history traits, growth rates, and sensitivity to intraspecific competition versus abiotic conditions. This study highlights the importance of understanding variability in early postfire regeneration and in young stand structures to anticipate future landscape patterns in ecosystems characterized by high-severity, infrequent disturbance regimes.

4.1. Variation and convergence in among-stand structural trajectories

Convergence of simulated stand density when both abiotic conditions and early regeneration varied (Both vary scenarios) is consistent with postfire chronosequence studies in this region (Kashian et al., 2005a, 2005b). Convergence occurs when initially dense stands undergo self-thinning as individual trees outcompete their neighbors (Peet, 1992), while establishment continues in sparser stands where light is available at the forest floor (Kashian et al., 2005b; Turner et al., 2016). Simultaneous self-thinning in some stands and infilling in others was evident early in simulations of both lodgepole pine and Douglas-fir. In field chronosequences, rapid declines in stand density (until 50 or 100 years postfire) and convergence have also been documented (Kashian et al., 2005b). In contrast, prolonged time to convergence and self-thinning in simulated subalpine fir and Engelmann spruce stands may be related to slower initial growth of these species during seedling and sapling stages, as well as alleviated light competition due to smaller initial stem numbers and narrower crowns (Oosting and Reed, 1952; LeBarron and Jemison, 1953; Alexander, 1987; Purves et al., 2007).

Past studies of lodgepole pine stand development in the GYE indicate that basal area increment converges more rapidly than stem density (Kashian et al., 2005b), but this was not observed for all species in our simulations. Lodgepole pine trees in sparser stands can exhibit a compensatory growth response to lower densities (e.g., Copenhaver and Tinker, 2014) and therefore tend toward similar basal area as denser stands. However, continued establishment and growth of the more shade-tolerant subalpine fir and Engelmann spruce under low light conditions may enable convergence in stand density and basal area simultaneously.

4.2. In fluence of early regeneration densities versus abiotic conditions on stand structural variability

Variation in early postfire regeneration densities was the most important driver of stand structural variability for at least 99 and up to 217 years, depending on the species. Over time, stands shifted from competition-driven convergence to environment-driven divergence, in which variability among stands was maintained or increased. This supports our expectation that variation in regeneration densities would drive variability among young stands, with variation in abiotic drivers becoming increasingly important as stands aged. As stands continue to age, it is also possible that processes such as mortality of trees from the first wave of postfire regeneration and continued infilling will contribute to homogenization of stand structures across different abiotic conditions later in stand development (as may be the case with subalpine fir and lodgepole pine in Fig. 3f,g). However, the long-lasting influence of early postfire regeneration is striking given the wide range of abiotic conditions. As fire frequency is likely to increase in the future (Westerling et al., 2011), the influence of variation in tree regeneration is also likely to increase relative to the effect of abiotic variation.

Our expectation that regeneration would influence stand structural variability of lodgepole pine for a longer period of time than other conifers was supported for density, highlighting the importance of species life history traits. Traits that favor abundant and rapid initial regeneration after fire, such as high prefire serotiny in lodgepole pine and rapid abscission schedule (e.g., jack pine; Greene et al., 2013), can have long-lasting effects on stand densities (Mason, 1915; Kashian et al., 2005b). Lodgepole pine also regenerates non-serotinously, but shading from an initial cohort may suppress subsequent regeneration and seedling growth (Lotan and Perry, 1983; Knight et al., 2014). Given smaller ranges of early postfire densities and prolonged periods of establishment and growth in the understory, species such as subalpine fir and Engelmann spruce may more rapidly overcome initial variation in regeneration. Of the three solely wind-dispersed species, subalpine fir is generally characterized as the most shade-tolerant and as a relatively prolific and regular seeder (Oosting and Reed, 1952; Alexander, 1987). Consistent with these traits, variation in early regeneration densities had the shortest influence on subalpine fir stand densities compared to Douglas-fir and Engelmann spruce.

In contrast, our expectation that regeneration would have a prolonged influence on lodgepole pine compared to other species was not supported for among-stand basal area. Differential growth rates and their sensitivity to environmental drivers and competition may explain differences among species. Lodgepole pine is relatively faster growing at young ages and radial growth increment is highly sensitive to crowding (Wykoff, 1990; Veblen et al., 1991; Buechling et al., 2017), resulting in accelerated growth of individual trees in sparser stands (Copenhaver and Tinker, 2014). Tree growth may also be enhanced given moderate annual precipitation (in the range of 600 to 900 mm yr−1; Buechling et al., 2017), and growing conditions are more limited at the upper (shorter growing season) and lower (drier) treeline (Knight et al., 2014). Among the four species, Douglas-fir radial growth increment is least sensitive to crowding (Buechling et al., 2017), potentially maintaining variability in stand basal area due to early regeneration densities for a longer period of time than in other species.

4.3. Implications and limitations of our model-based approach

Modeling studies can inform management decisions by serving as a benchmark of the historical range of ecosystem responses to drivers that are expected to change in the future (Seidl et al., 2016b). For example, the results of this study could be used to evaluate whether the relative importance of the abiotic template in influencing stand structural variability increases in scenarios of future interannual climate variability. This, in turn, may indicate whether forests are likely to be characterized by divergent stand structural trajectories under changing climate as well as changing fire regimes. Process-based models that integrate emergent, fine-scale responses to environmental drivers are particularly well suited to explore long-term landscape trajectories under no-analog conditions (Cuddington et al., 2013; Gustafson, 2013; Grimm et al., 2017). These models are powerful tools for characterizing ecosystem patterns and processes, generating new or improved understanding of underlying mechanisms, and facilitating a synthetic approach with field-based experiments to understanding complex systems (Jenerette and Shen, 2012; Bowman et al., 2015; Cottingham et al., 2017; Grimm et al., 2017; Rastetter, 2017; Seidl, 2017). Future studies could employ process-based models to disentangle the relative influence of individual abiotic drivers.

For simulating variability in ecosystems, it must be noted that models are still likely to underestimate real-world environmental variation. Stand structural variability is probably higher among real stands compared to our simulations, which did not include covariation between high regeneration densities and favorable abiotic conditions (Schoennagel et al., 2003; Turner et al., 2004; Donato et al., 2016; Harvey et al., 2016b; Stevens-Rumann et al., 2018), subsequent disturbances (e.g., insect outbreaks; Antos and Parish, 2002), interspecific interactions, and seed supply from neighboring stands. In addition, stands may reburn prior to the 300 years of stand development simulated in this study. Historically, extensive portions of reburned forest in Yellowstone were at least 300 years old (Romme and Despain, 1989), and higher-elevation (> 2400 m) forests were characterized by approximately 300-year fire return intervals (Schoennagel et al., 2003). However, as warming climate continues to drive increasing fire activity (Westerling et al., 2006, 2011; Abatzoglou and Williams, 2016; Westerling, 2016), it is increasingly likely that younger stands will reburn.

4.4. Variability is important for anticipating forest landscape structure and function

Variability characterizes many post-disturbance systems (e.g., Kashian et al., 2005b; Suzuki et al., 2009) and has substantial implications for future forests, but remains understudied (Fraterrigo and Rusak, 2008). Assumptions that all stands have the same value of aboveground carbon storage may be inappropriate with increasing fire activity and therefore increasing extent of young, structurally variable forest (Kashian et al., 2006, 2013). When variability is high, mean estimates of stand structure may be increasingly inaccurate predictors of ecosystem patterns and processes (Cottingham et al., 2000; Fraterrigo and Rusak, 2008), particularly if relationships between drivers and responses are nonlinear and interact across scales (Lovett et al., 2005; Peters et al., 2007). Increased forest landscape heterogeneity could also dampen the spread or severity of future disturbances (Bebi et al., 2003; Kulakowski and Veblen, 2007; Seidl et al., 2016a).

For forests adapted to high-severity, infrequent fire regimes, understanding causes and patterns of variation in early postfire regeneration appears critically important for anticipating long-term landscape variability. Early postfire regeneration (within the first two decades) affected structural trajectories of four widespread conifer species for decades to centuries in stands across a wide gradient of environmental conditions. Regeneration processes, such as seed supply, dispersal, establishment, and early seedling survival, are often highly sensitive to changes in disturbance regimes and environmental fluctuations (Kipfmueller and Kupfer, 2005; Larson and Kipfmueller, 2010; Harvey et al., 2016b; Kemp et al., 2016; Hansen et al., 2018; Stevens-Rumann et al., 2018). Early regeneration densities may forecast long-term trajectories in ecosystems that experience periodic high-severity disturbances (Turner et al., 1998), although it may be necessary to consider more than just the first few years of establishment (e.g., Peterson and Pickett, 1995; Gill et al., 2017). Ongoing research on the effects of changing climate, disturbance regimes, and other drivers of variation in early postfire regeneration is needed to anticipate future forest landscape patterns.

Supplementary Material

Appendix

Acknowledgements

KHB, RS, and MGT designed the study; KHB parameterized and evaluated the model in consultation with RS, WR, WDH, and MGT; RS and WR provided model code and ongoing development; KHB performed simulations and analyzed data; KHB and MGT wrote manuscript, and all authors contributed. We thank Anthony Ives, Chris Kucharik, Zak Ratajczak, and two anonymous reviewers for providing constructive feedback that greatly improved this paper. This research was funded by the Joint Fire Science Program (Project No. 16-3-01-4) and the University of Wisconsin-Madison Vilas Trust. KHB also acknowledges support from a University of Wisconsin-Madison Integrative Biology Department first-year fellowship. RS and WR acknowledge support from the Austrian Science Fund through START grant Y895-B25.

Data availability statement

The model output data that support the findings of this study and files to recreate model simulations are openly available in the Environmental Data Initiative (EDI) at DOI https://doi.org/10.6073/pasta/152ed98663904892d9d11903949cadb7.

References

  1. Abatzoglou JT, Williams AP. Impact of anthropogenic climate change on wildfire across western US forests. Proc Natl Acad Sci. 2016;113:11770–11775. doi: 10.1073/pnas.1607171113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alexander RR. Ecology, silviculture, and management of the Engelmann spruce - subalpine fir type in the central and southern Rocky Mountains. USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Agriculture Handbook No. 59, CO; Fort Collins: 1987. [Google Scholar]
  3. Antos JA, Parish R. Dynamics of an old-growth, fire-initiated, subalpine forest in southern interior British Columbia: Tree size, age, and spatial structure. Can J For Res. 2002;32:1935–1946. doi: 10.1139/x02-116. [DOI] [Google Scholar]
  4. Aplet GH, Laven RD, Smith FW. Patterns of community dynamics in Colorado Engelmann spruce-subalpine fir forests. Ecology. 1988;69:312–319. doi: 10.2307/1940429. [DOI] [Google Scholar]
  5. Baker WL. Fire ecology in Rocky Mountain landscapes. Island Press; Washington, DC: 2009. [Google Scholar]
  6. Baker WL, Veblen TT, Sherriff RL. Fire, fuels and restoration of ponderosa pine-Douglas fir forests in the Rocky Mountains USA. J Biogeogr. 2007;34:251–269. doi: 10.1111/j.1365-2699.2006.01592.x. [DOI] [Google Scholar]
  7. Bebi P, Kulakowski D, Veblen TT. Interactions between fire and spruce beetles in a subalpine rocky mountain forest landscape. Ecology. 2003;84:362–371. doi: 10.1890/0012-9658(2003)084[0362:IBFASB]2.0.CO;2. [DOI] [Google Scholar]
  8. Bowman DMJS, Perry GLW, Marston JB. Feedbacks and landscape-level vegetation dynamics. Trends Ecol Evol. 2015;30:255–260. doi: 10.1016/j.tree.2015.03.005. [DOI] [PubMed] [Google Scholar]
  9. Bradford JB, Birdsey RA, Joyce LA, Ryan MG. Tree age, disturbance history, and carbon stocks and fluxes in subalpine Rocky Mountain forests. Glob Chang Biol. 2008;14:2882–2897. doi: 10.1111/j.1365-2486.2008.01686.x. [DOI] [Google Scholar]
  10. Buechling A, Martin PH, Canham CD. Climate and competition effects on tree growth in Rocky Mountain forests. J Ecol. 2017;105:1636–1647. doi: 10.1111/1365-2745.12782. [DOI] [Google Scholar]
  11. Buma B, Brown CD, Donato DC, Fontaine JB, Johnstone JF. The impacts of changing disturbance regimes on serotinous plant populations and communities. Bioscience. 2013;63:866–876. doi: 10.1525/bio.2013.63.11.5. [DOI] [Google Scholar]
  12. Copenhaver PE, Tinker DB. Stand density and age affect tree-level structural and functional characteristics of young, postfire lodgepole pine in Yellowstone National Park. For Ecol Manage. 2014;320:138–148. doi: 10.1016/j.foreco.2014.03.024. [DOI] [Google Scholar]
  13. Cottingham KL, Rusak JA, Leavitt PR. Increased ecosystem variability and reduced predictability following fertilisation: Evidence from palaeolimnology. Ecol Lett. 2000;3:340–348. doi: 10.1046/j.1461-0248.2000.00158.x. [DOI] [Google Scholar]
  14. Cottingham KL, Fey SB, Fritschie KJ, Trout-Haney JV. Advancing ecosystem science by promoting greater use of theory and multiple research approaches in graduate education. Ecosystems. 2017;20:267–273. doi: 10.1007/s10021-016-0070-3. [DOI] [Google Scholar]
  15. Crookston NL, Dixon GE. The forest vegetation simulator: A review of its structure, content, and applications. Comput Electron Agric. 2005;49:60–80. doi: 10.1016/j.compag.2005.02.003. [DOI] [Google Scholar]
  16. Crossley DI. Fruiting habits of lodgepole pine. Canada Department of Northern Affairs and National Resources, Forestry Branch, Forest Research Division, Technical Note No. 35. Ottawa, Ont: 1956. [Google Scholar]
  17. Cuddington K, Fortin MJ, Gerber LR, Hastings A, Liebhold A, O'Connor M, Ray C. Process-based models are required to manage ecological systems in a changing world. Ecosphere. 2013;4:20. doi: 10.1890/ES12-00178.1. [DOI] [Google Scholar]
  18. Despain DG. Yellowstone vegetation: Consequences of environment and history in a natural setting. Roberts Rinehart, CO; Boulder: 1990. [Google Scholar]
  19. Dixon GE. Essential FVS: A user's guide to the Forest Vegetation Simulator. USDA Forest Service, Forest Management Service Center, CO; Fort Collins: 2002. [Google Scholar]
  20. Donato DC, Harvey BJ, Turner MG. Regeneration of montane forests a quarter-century after the 1988 Yellowstone Fires: A fire-catalyzed shift in lower treelines? Ecosphere. 2016;7:e01410. doi: 10.1002/ecs2.1410. [DOI] [Google Scholar]
  21. Douglas RB, Parker VT, Cullings KW. Belowground ectomycorrhizal community structure of mature lodgepole pine and mixed conifer stands in Yellowstone National Park. For Ecol Manage. 2005;208:303–317. doi: 10.1016/j.foreco.2004.12.011. [DOI] [Google Scholar]
  22. Edwards M, Krawchuk MA, Burton PJ. Short-interval disturbance in lodgepole pine forests, British Columbia, Canada: Understory and overstory response to mountain pine beetle and fire. For Ecol Manage. 2015;338:163–175. doi: 10.1016/j.foreco.2014.11.011. [DOI] [Google Scholar]
  23. Foster DR, Knight DH, Franklin JF. Landscape patterns and legacies resulting from large, infrequent forest disturbances. Ecosystems. 1998;1:497–510. doi: 10.1007/s100219900046. [DOI] [Google Scholar]
  24. Fraterrigo JM, Rusak JA. Disturbance-driven changes in the variability of ecological patterns and processes. Ecol Lett. 2008;11:756–770. doi: 10.1111/j.1461-0248.2008.01191.x. [DOI] [PubMed] [Google Scholar]
  25. Gill NS, Jarvis D, Veblen TT, Pickett STA, Kulakowski D. Is initial post-disturbance regeneration indicative of longer-term trajectories? Ecosphere. 2017;8:e01924. doi: 10.1002/ecs2.1924. [DOI] [Google Scholar]
  26. Glenn-Lewin DC, van der Maarel E. In: Plant Succession: Theory and Prediction. Glenn-Lewin DC, Peet RK, Veblen TT, editors. Chapman & Hall; London, UK: 1992. Patterns and processes of vegetation dynamics; pp. 11–59. [Google Scholar]
  27. Greene DF, Splawinski TB, Gauthier S, Bergeron Y. Seed abscission schedules and the timing of post-fire salvage of Picea mariana and Pinus banksiana. For Ecol Manage. 2013;303:20–24. doi: 10.1016/j.foreco.2013.03.049. [DOI] [Google Scholar]
  28. Grimm V, Ayllon D, Railsback SF. Next-generation individual-based models integrate biodiversity and ecosystems: Yes we can, and yes we must. Ecosystems. 2017;20:229–236. doi: 10.1007/s10021-016-0071-2. [DOI] [Google Scholar]
  29. Gustafson EJ. When relationships estimated in the past cannot be used to predict the future: Using mechanistic models to predict landscape ecological dynamics in a changing world. Landsc Ecol. 2013;28:1429–1437. doi: 10.1007/s10980-013-9927-4. [DOI] [Google Scholar]
  30. Hansen WD, Braziunas KH, Rammer W, Seidl R, Turner MG. It takes a few to tango: Changing climate and fire regimes can cause regeneration failure of two subalpine conifers. Ecology. 2018;99:966–977. doi: 10.1002/ecy.2181. [DOI] [PubMed] [Google Scholar]
  31. Harvey BJ, Donato DC, Turner MG. Drivers and trends in landscape patterns of stand-replacing fire in forests of the US Northern Rocky Mountains (1984-2010) Landsc Ecol. 2016a;31:2367–2383. doi: 10.1007/s10980-016-0408-4. [DOI] [Google Scholar]
  32. Harvey BJ, Donato DC, Turner MG. High and dry: Post-fire tree seedling establishment in subalpine forests decreases with post-fire drought and large stand-replacing burn patches. Glob Ecol Biogeogr. 2016b;25:655–669. doi: 10.1111/geb.12443. [DOI] [Google Scholar]
  33. Higuera PE, Whitlock C, Gage JA. Linking tree-ring and sediment-charcoal records to reconstruct fire occurrence and area burned in subalpine forests of Yellowstone National Park, USA. Holocene. 2011;21:327–341. doi: 10.1177/0959683610374882. [DOI] [Google Scholar]
  34. Jenerette GD, Shen W. Experimental landscape ecology. Landsc Ecol. 2012;27:1237–1248. doi: 10.1007/s10980-012-9797-1. [DOI] [Google Scholar]
  35. Kashian DM, Turner MG, Romme WH. Variability in leaf area and stemwood increment along a 300-year lodgepole pine chronosequence. Ecosystems. 2005a;8:48–61. doi: 10.1007/s10021-004-0067-1. [DOI] [Google Scholar]
  36. Kashian DM, Turner MG, Romme WH, Lorimer CG. Variability and convergence in stand structural development on a fire-dominated subalpine landscape. Ecology. 2005b;86:643–654. doi: 10.1890/03-0828. [DOI] [Google Scholar]
  37. Kashian DM, Romme WH, Tinker DB, Turner MG, Ryan MG. Carbon storage on landscapes with stand-replacing fires. Bioscience. 2006;56:598. doi: 10.1641/0006-3568(2006)56[598:CSOLWS]2.0.CO;2. [DOI] [Google Scholar]
  38. Kashian DM, Romme WH, Tinker DB, Turner MG. Postfire changes in forest carbon storage over a 300-year chronosequence of Pinus contorta-dominated forests. Ecol Monogr. 2013;83:49–66. doi: 10.1890/11-1454.1. [DOI] [Google Scholar]
  39. Kemp KB, Higuera PE, Morgan P. Fire legacies impact conifer regeneration across environmental gradients in the US Northern Rockies. Landsc Ecol. 2016;31:619–636. doi: 10.1007/s10980-015-0268-3. [DOI] [Google Scholar]
  40. Kipfmueller KF, Kupfer JA. Complexity of successional pathways in subalpine forests of the Selway-Bitterroot Wilderness Area. Ann Assoc Am Geogr. 2005;95:495–510. doi: 10.1111/j.1467-8306.2005.00471.x. [DOI] [Google Scholar]
  41. Knight DH, Jones GP, Reiners WA, Romme WH. Mountains and plains: The ecology of Wyoming landscapes. Second edition. Yale University Press; New Haven, CT: 2014. [Google Scholar]
  42. Korzukhin MD, Ter-Mikaelian MT, Wagner RG. Process versus empirical models: Which approach for forest ecosystem management? Can J For Res. 1996;26:879–887. [Google Scholar]
  43. Kulakowski D, Veblen TT. Effect of prior disturbances on the extent and severity of wildfire in Colorado subalpine forests. Ecology. 2007;88:759–769. doi: 10.1890/06-0124. [DOI] [PubMed] [Google Scholar]
  44. Larson ER, Kipfmueller KF. Patterns in whitebark pine regeneration and their relationships to biophysical site characteristics in southwest Montana, central Idaho, and Oregon USA. Can J For Res. 2010;40:476–487. doi: 10.1139/X10-005. [DOI] [Google Scholar]
  45. LeBarron RK, Jemison GM. Ecology and silviculture of the Engelmann spruce-alpine fir type. J For. 1953;51:349–352. [Google Scholar]
  46. Litton CM, Ryan MG, Knight DH. Effects of tree density and stand age on carbon allocation patters in postfire lodgepole pine. Ecol Appl. 2004;14:460–516. doi: 10.1890/02-5291. [DOI] [Google Scholar]
  47. Lotan J, Perry D. Ecology and regeneration of lodgepole pine. Vol. 606 USDA Forest Service Agriculture Handbook; Washington, DC: 1983. [Google Scholar]
  48. Lovett GM, Jones CG, Turner MG, Weathers KC. In: Ecosystem Function in Heterogeneous Landscapes. Lovett GM, Jones CG, Turner MG, Weathers KC, editors. Springer; New York, NY: 2005. Conceptual frameworks: Plan for a half-built house; pp. 463–470. [Google Scholar]
  49. Mason DT. The life history of lodgepole pine in the Rocky Mountains. Vol. 154 USDA Bulletin; Washington, DC: 1915. [Google Scholar]
  50. Meigs GW, Morrissey RC, Bace R, Chaskovskyy O, Cada V, Despres T, Donato DC, Janda P, Labusova J, Seedre M, Mikolas M, et al. More ways than one: Mixed-severity disturbance regimes foster structural complexity via multiple developmental pathways. For Ecol Manage. 2017;406:410–426. doi: 10.1016/j.foreco.2017.07.051. [DOI] [Google Scholar]
  51. Miller DA, White RA. A conterminous United States multilayer soil characteristics dataset for regional climate and hydrology modeling. Earth Interact. 1998;2:1–26. doi: 10.1175/1087-3562(1998)002<0001:ACUSMS>2.3.m;2. [DOI] [Google Scholar]
  52. Millspaugh SH, Whitlock C, Bartlein PJ, Wallace LL. In: After the Fires: The Ecology of Change in Yellowstone National Park. Wallace LL, editor. Yale University Press; New Haven, Connecticut, USA: 2004. Postglacial fire, vegetation, and climate history of the Yellowstone-Lamar and Central Plateau provinces, Yellowstone National Park; pp. 10–28. [Google Scholar]
  53. Oosting HJ, Reed JF. Virgin spruce-fir of the Medicine Bow Mountains Wyoming. Ecol Monogr. 1952;22:69–91. [Google Scholar]
  54. Peet RK. In: Plant Succession: Theory and Prediction. Glenn-Lewin DC, Peet RK, Veblen TT, editors. Chapman & Hall; London, UK: 1992. Community structure and ecosystem function; pp. 103–151. [Google Scholar]
  55. Peters DPC, Bestelmeyer BT, Turner MG. Cross-scale interactions and changing pattern-process relationships: Consequences for system dynamics. Ecosystems. 2007;10:790–796. doi: 10.1007/s10021-007-9055-6. [DOI] [Google Scholar]
  56. Peterson CJ, Pickett STA. Forest reorganization: A case study in an old-growth forest catastrophic blowdown. Ecology. 1995;76:763–774. doi: 10.2307/1939342. [DOI] [Google Scholar]
  57. Pinno BD, Errington RC, Thompson DK. Young jack pine and high severity fire combine to create potentially expansive areas of understocked forest. For Ecol Manage. 2013;310:517–522. doi: 10.1016/j.foreco.2013.08.055.. [DOI] [Google Scholar]
  58. Purves DW, Lichstein JW, Pacala SW. Crown plasticity and competition for canopy space: A new spatially implicit model parameterized for 250 North American tree species. PLoS One. 2007;2:e870. doi: 10.1371/journal.pone.0000870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Core Team, R. R: A language and environment for statistical computing. R Foundation for Statistical Computing; Vienna, Austria: 2017. [Google Scholar]
  60. Rastetter EB. Modeling for understanding v. modeling for numbers. Ecosystems. 2017;20:215–221. doi: 10.1007/s10021-016-0067-y. [DOI] [Google Scholar]
  61. Reyer CPO, Rammig A, Brouwers N, Langerwisch F. Forest resilience, tipping points and global change processes. J Ecol. 2015;103:1–4. doi: 10.1111/1365-2745.12342. [DOI] [Google Scholar]
  62. Rodman A, Shovic HF, Thoma DP. Soils of Yellowstone National Park. Yellowstone Center for Resources, YCR-NRSR-96-2; Yellowstone National Park, WY: 1996. [Google Scholar]
  63. Romme W. Fire and landscape diversity in subalpine forests of Yellowstone National Park. Ecol Monogr. 1982;52:199–221. doi: 10.2307/1942611. [DOI] [Google Scholar]
  64. Romme WH, Despain DG. Historical perspective on the yellowstone fires of 1988. Bioscience. 1989;39:695–699. doi: 10.2307/1311000. [DOI] [Google Scholar]
  65. Schoennagel T, Turner MG, Romme WH. The influence of fire interval and serotiny on postfire lodgepole pine density in Yellowstone National Park. Ecology. 2003;84:2967–2978. doi: 10.1890/02-0277. [DOI] [Google Scholar]
  66. Schoennagel T, Veblen TT, Romme WH. The interaction of fire, fuels, and climate across Rocky Mountain forests. Bioscience. 2004;54:661. doi: 10.1641/0006-3568(2004)054[0661:TIOFFA]2.0.CO;2. [DOI] [Google Scholar]
  67. Scholes RJ. Taking the mumbo out of the jumbo: Progress towards a robust basis for ecological scaling. Ecosystems. 2017;20:4–13. doi: 10.1007/s10021-016-0047-2. [DOI] [Google Scholar]
  68. Seidl R. To model or not to model, that is no longer the question for ecologists. Ecosystems. 2017;20:222–228. doi: 10.1007/s10021-016-0068-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Seidl R, Rammer W. iLand online model documentation. 2018. [accessed 12.29.17]. [WWW Document]. URL http://iland.boku.ac.at.
  70. Seidl R, Rammer W, Scheller RM, Spies TA. An individual-based process model to simulate landscape-scale forest ecosystem dynamics. Ecol Modell. 2012;231:87–100. doi: 10.1016/j.ecolmodel.2012.02.015. [DOI] [Google Scholar]
  71. Seidl R, Rammer W, Blennow K. Simulating wind disturbance impacts on forest landscapes: Tree-level heterogeneity matters. Environ Model Softw. 2014a;51:1–11. doi: 10.1016/j.envsoft.2013.09.018. [DOI] [Google Scholar]
  72. Seidl R, Rammer W, Spies TA. Disturbance legacies increase the resilience of forest ecosystem structure, composition, and functioning. Ecol Appl. 2014b;24:2063–2077. doi: 10.1890/14-0255.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Seidl R, Donato DC, Raffa KF, Turner MG. Spatial variability in tree regeneration after wildfire delays and dampens future bark beetle outbreaks. Proc Natl Acad Sci. 2016a;113:13075–13080. doi: 10.1073/pnas.1615263113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Seidl R, Spies TA, Peterson DL, Stephens SL, Hicke JA. Searching for resilience: Addressing the impacts of changing disturbance regimes on forest ecosystem services. J Appl Ecol. 2016b;53:120–129. doi: 10.1111/1365-2664.12511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Silva Pedro M, Rammer W, Seidl R. Tree species diversity mitigates disturbance impacts on the forest carbon cycle. Oecologia. 2015;177:619–630. doi: 10.1007/s00442-014-3150-0. [DOI] [PubMed] [Google Scholar]
  76. Simard M, Powell EN, Raffa KF, Turner MG. What explains landscape patterns of tree mortality caused by bark beetle outbreaks in Greater Yellowstone? Glob Ecol Biogeogr. 2012;21:556–567. doi: 10.1111/j.1466-8238.2011.00710.x. [DOI] [Google Scholar]
  77. Smithwick EAH, Kashian DM, Ryan MG, Turner MG. Long-term nitrogen storage and soil nitrogen availability in post-fire lodgepole pine ecosystems. Ecosystems. 2009a;12:792–806. doi: 10.1007/s10021-009-9257-1. [DOI] [Google Scholar]
  78. Smithwick EAH, Ryan MG, Kashian DM, Romme WH, Tinker DB, Turner MG. Modeling the effects of fire and climate change on carbon and nitrogen storage in lodgepole pine (Pinus contorta) stands. Glob Chang Biol. 2009b;15:535–548. doi: 10.1111/j.1365-2486.2008.01659.x. [DOI] [Google Scholar]
  79. Stevens-Rumann CS, Kemp KB, Higuera PE, Harvey BJ, Rother MT, Donato DC, Morgan P, Veblen TT. Evidence for declining forest resilience to wildfires under climate change. Ecol Lett. 2018;21:243–252. doi: 10.1111/ele.12889. [DOI] [PubMed] [Google Scholar]
  80. Suzuki SN, Kachi N, Suzuki JI. Changes in variance components of forest structure along a chronosequence in a wave-regenerated forest. Ecol Res. 2009;24:1371–1379. doi: 10.1007/s11284-009-0621-6. [DOI] [Google Scholar]
  81. Tepley AJ, Swanson FJ, Spies TA. Fire-mediated pathways of stand development in Douglas-fir/western hemlock forests of the Pacific Northwest, USA. Ecology. 2013;94:1729–1743. doi: 10.1890/12-1506.1. [DOI] [PubMed] [Google Scholar]
  82. Tews J, Brose U, Grimm V, Tielborger K, Wichmann MC, Schwager M, Jeltsch F. Animal species diversity driven by habitat heterogeneity/diversity: The importance of keystone structures. J Biogeogr. 2004;31:79–92. doi: 10.1046/j.0305-0270.2003.00994.x. [DOI] [Google Scholar]
  83. Thom D, Rammer W, Dirnbock T, Muller J, Kobler J, Katzensteiner K, Helm N, Seidl R. The impacts of climate change and disturbance on spatio-temporal trajectories of biodiversity in a temperate forest landscape. J Appl Ecol. 2017;54:28–38. doi: 10.1111/1365-2664.12644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Thornton PE, Thornton MM, Mayer BW, Wei Y, Devarakonda R, Vose RS, Cook RB. Daymet: Daily surface weather data on a 1-km grid for North America, Version 3. Oak Ridge National Laboratory Distributed Active Archive Center; Oak Ridge, Tennessee, USA: 2017. [DOI] [Google Scholar]
  85. Tinker DB, Romme WH, Hargrove WW, Gardner RH, Turner MG. Landscape-scale heterogeneity in lodgepole pine serotiny. Can J For Res. 1994;24:897–903. doi: 10.1139/x94-118. [DOI] [Google Scholar]
  86. Turner MG. Disturbance and landscape dynamics in a changing world. Ecology. 2010;91:2833–2849. doi: 10.1890/10-0097.1. [DOI] [PubMed] [Google Scholar]
  87. Turner MG, Romme WH. Landscape dynamics in crown fires ecosystems. Landsc Ecol. 1994;9:59–77. doi: 10.1007/BF00135079. [DOI] [Google Scholar]
  88. Turner MG, Hargrove WW, Gardner RH, Romme WH. Effects of fire on landscape heterogeneity in Yellowstone National Park. Wyoming. J Veg Sci. 1994;5:731–742. doi: 10.2307/3235886. [DOI] [Google Scholar]
  89. Turner MG, Romme WH, Gardner RH, Hargrove WW. Effects of fire size and pattern on early succession in Yellowstone National Park. Ecol Monogr. 1997;67:411–433. doi: 10.1890/0012-9615a997)067[0411:EOFSAP]2.0.CO;2. [DOI] [Google Scholar]
  90. Turner MG, Baker WL, Peterson CJ, Peet RK. Factors influencing succession: Lessons from large, infrequent natural disturbances. Ecosystems. 1998;1:511–523. doi: 10.1007/s100219900047. [DOI] [Google Scholar]
  91. Turner MG, Romme WH, Gardner RH. Prefire heterogeneity, fire severity, and early postfire plant reestablishment in subalpine forests of Yellowstone National Park Wyoming. Int J Wildl Fire. 1999;9:21–36. doi: 10.1071/WF99003. [DOI] [Google Scholar]
  92. Turner MG, Tinker DB, Romme WH, Kashian DM, Litton CM. Landscape patterns of sapling density, leaf area, and aboveground net primary production in postfire lodgepole pine forests, Yellowstone National Park (USA) Ecosystems. 2004;7:751–775. doi: 10.1007/s10021-004-0011-4. [DOI] [Google Scholar]
  93. Turner MG, Smithwick EA, Metzger KL, Tinker DB, Romme WH. Inorganic nitrogen availability after severe stand-replacing fire in the Greater Yellowstone ecosystem. Proc Natl Acad Sci USA. 2007;104:4782–4789. doi: 10.1073/pnas.0700180104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  94. Turner MG, Smithwick EAH, Tinker DB, Romme WH. Variation in foliar nitrogen and aboveground net primary production in young postfire lodgepole pine. Can J For Res. 2009;39:1024–1035. doi: 10.1139/X09-029. [DOI] [Google Scholar]
  95. Turner MG, Whitby TG, Tinker DB, Romme WH. Twenty-four years after the Yellowstone fires: Are postfire lodgepole pine stands converging in structure and function? Ecology. 2016;97:1260–1273. doi: 10.1890/15-1585.1. [DOI] [PubMed] [Google Scholar]
  96. Veblen TT. Age and size structure of subalpine forests in the Colorado Front Range. Bull Torrey Bot Club. 1986;113:225. doi: 10.2307/2996361. [DOI] [Google Scholar]
  97. Veblen TT, Hadley KS, Reid MS. Disturbance and stand development of a Colorado subalpine forest. J Biogeogr. 1991;18:707–716. doi: 10.2307/2845552. [DOI] [Google Scholar]
  98. Walker LR, del Moral R. Primary succession and ecosystem rehabilitation. Cambridge University Press; New York, NY: 2003. [Google Scholar]
  99. Westerling AL. Increasing western US forest wildfire activity: Sensitivity to changes in the timing of spring. Philos Trans R Soc B Biol Sci. 2016;371:20150178. doi: 10.1098/rstb.2015.0178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  100. Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW. Warming and earlier spring increase Western U.S. forest wildfire activity. Science. 2006;313:940–943. doi: 10.1126/science.1128834. [DOI] [PubMed] [Google Scholar]
  101. Westerling AL, Turner MG, Smithwick EAH, Romme WH, Ryan MG. Continued warming could transform Greater Yellowstone fire regimes by mid-21st century. Proc Natl Acad Sci. 2011;108:13165–13170. doi: 10.1073/pnas.1110199108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  102. Western Regional Climate Center (WRCC) Cooperative climatological data summaries: NOAA cooperative stations - temperature and precipitation. 2018. [accessed 8.14.18]. Stations 482375, 485115, 485345, 485355, 486440, 486845, 488315, 489025, and 489905. [WWW Document]. URL https://wrcc.dri.edu/Climate/west_coop_summaries.php.
  103. Whitlock C. Postglacial vegetation and climate of Grand Teton and southern Yellowstone National Parks. Ecol Monogr. 1993;63:173–198. doi: 10.2307/2937179. [DOI] [Google Scholar]
  104. Whitlock C, Marlon J, Briles C, Brunelle A, Long C, Bartlein P. Long-term relations among fire, fuel, and climate in the north-western US based on lake-sediment studies. Int J Wildl Fire. 2008;17:72–83. doi: 10.1071/WF07025. [DOI] [Google Scholar]
  105. Wykoff WR. A basal area increment model for individual conifers in the Northern Rocky Mountains. For Sci. 1990;36:1077–1104. [Google Scholar]
  106. Yellowstone National Park (YNP) Yellowstone resources and issues handbook: 2017. Yellowstone National Park, WY: 2017. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix
EMS145169-supplement-Appendix.pdf (2MB, pdf)

Data Availability Statement

The model output data that support the findings of this study and files to recreate model simulations are openly available in the Environmental Data Initiative (EDI) at DOI https://doi.org/10.6073/pasta/152ed98663904892d9d11903949cadb7.

RESOURCES