UC Merced
UC Merced Previously Published Works
Title
Response of Sierra Nevada forests to projected climate-wildfire interactions.
Permalink
https://escholarship.org/uc/item/78x341q4
Journal
Global change biology, 23(5)
ISSN
1354-1013
Authors
Liang, Shuang
Hurteau, Matthew D
Westerling, Anthony LeRoy
Publication Date
2017-05-01
DOI
10.1111/gcb.13544
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
Accepted Article
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/gcb.13544
This article is protected by copyright. All rights reserved.
Received Date : 24-Mar-2016
Revised Date : 21-Oct-2016
Accepted Date : 25-Oct-2016
Article type : Primary Research Articles
Title: Response of Sierra Nevada forests to projected climate-wildfire interactions
Running head: Forest response to projected climate-wildfire
Authors: Shuang Liang
1
, Matthew D. Hurteau
2
*, A. LeRoy Westerling
3
Affiliation:
1
Intercollege Graduate Degree Program in Ecology and Department of Ecosystem Science
and Management, The Pennsylvania State University, 228 Forest Resources Building,
University Park, PA 16802 (sul248@psu.edu)
2
Department of Biology, University of New Mexico, MSC03 2020, Albuquerque, NM 87131
3
Sierra Nevada Research Institute, University of California, Merced, 5200 N. Lake Road,
Merced, CA 95343 (awesterling@ucmerced.edu)
*
Corresponding Author:
Matthew D. Hurteau
Tel: 505-277-0863
Fax: 505-277-0304
E-mail: mhurteau@unm.edu
Keywords: climate change, wildfire, carbon, forest community change, LANDIS-II, Sierra
Nevada
Type of Paper: Primary Research Article
Abstract
Climate influences forests directly and indirectly through disturbance. The interaction of
climate change and increasing area burned has the potential to alter forest composition and
community assembly. However, the overall forest response is likely to be influenced by
species-specific responses to environmental change and the scale of change in overstory
species cover. In this study, we sought to quantify how projected changes in climate and large
wildfire size would alter forest communities and carbon (C) dynamics, irrespective of
competition from non-tree species and potential changes in other fire regimes, across the
Sierra Nevada, USA. We used a species-specific, spatially explicit forest landscape model
(LANDIS-II) to evaluate forest response to climate-wildfire interactions under historical
(baseline) climate and climate projections from three climate models (GFDL, CCSM3 and
CNRM) forced by a medium-high emission scenario (A2) in combination with corresponding
climate-specific large wildfire projections. By late-century, we found modest changes in the
spatial distribution of dominant species by biomass relative to baseline, but extensive changes
in recruitment distribution. Although forest recruitment declined across much of the Sierra,
we found that projected climate and wildfire favored the recruitment of more drought-tolerant
Accepted Article
This article is protected by copyright. All rights reserved.
species over less drought-tolerant species relative to baseline, and this change was greatest at
mid-elevations. We also found that projected climate and wildfire decreased tree species
richness across a large proportion of the study area and transitioned more area to a C source,
which reduced landscape-level C sequestration potential. Our study, although a conservative
estimate, suggests that by late-century forest community distributions may not change as
intact units as predicted by biome-based modeling, but are likely to trend toward simplified
community composition as communities gradually disaggregate and the least tolerant species
are no longer able to establish. The potential exists for substantial community composition
change and forest simplification beyond this century.
Introduction
Climate influences forests directly through its differential effects on tree species and
indirectly through disturbance. Climate-induced drought stress (Williams et al., 2012) and
climate-enhanced large wildfire activity (Westerling et al., 2006; Westerling, 2016) are
anticipated to cause changes in forest community composition and productivity, which are
likely to affect carbon (C) dynamics (Williams et al., 2007; Lenihan et al., 2008; Loudermilk
et al., 2013; Gonzalez et al., 2015). However, there is often a disparity between changes in
forest community composition at the landscape scale and the magnitude of environmental
change (Jones et al., 2009; Bertrand et al., 2011; Zhu et al., 2012; Svenning & Sandel, 2015).
While environmental change can shift the fundamental niche space for successful
reproduction of some species, change of overstory species composition is much slower
because mature trees are typically more tolerant of a broader range of abiotic conditions
(Dolanc et al., 2013; Svenning & Sandel, 2015). Understanding how the interaction of
changing climate and climate-driven changes in disturbance regime will influence tree
species distributions is central to understanding how these factors will alter forest
communities across large landscapes.
Climate has long been identified as a primary control on species occurrence, with species-
specific environmental tolerance largely determining where species occur along a climate
gradient (Woodward et al., 2004; McKenney et al., 2007). Temperature and precipitation are
the climate variables that most directly affect vegetation biogeography. Changes in these
variables are anticipated to influence species distributions and thus community assemblage
(Woodward et al., 2004; Williams et al., 2007; Lutz et al., 2010). With the exception of
extreme events, such as ‘hotter drought’ (Williams et al., 2012; Allen et al., 2015), which can
cause sudden forest dieback, climatically driven changes in community composition are often
gradual and show disequilibrium with climate (Eriksson et al., 1996; Bertrand et al., 2011;
Svenning & Sandel, 2015). Change in forest species cover may be delayed relative to the rate
of climate change because long-lived tree species can persist on site even if conditions have
become unfavorable for recruitment (Svenning & Sandel, 2015). Moreover, in more
topographically diverse environments, montane forest species can move relatively short-
distances and remain in the same climate space (Loarie et al., 2009; Scherrer & Körner,
2011). In contrast to the time lag between changing environmental conditions and change of
the overstory species assemblage, regeneration is more responsive to climate change with
recruitment success affecting species distribution and forest community assemblage in the
long run (Grubb, 1972; Mok et al., 2012; Zhu et al., 2012).
Climate can also indirectly modify forested landscapes through its effects on wildfire regimes
(Westerling et al., 2011a; Littell et al., 2009). Interannual and decadal climatic variability has
been found to cause regional synchrony in large fire years and area burned in the western US
(Heyerdahl et al., 2008a; Westerling, 2016). The relationship between climate variation and
Accepted Article
This article is protected by copyright. All rights reserved.
wildfire activity varies regionally and includes factors such as timing of snowmelt (fire
season length), vegetation growth (fuel availability), and biomass moisture (fuel
flammability) (Heyerdahl et al., 2008b; Gill & Taylor, 2009; Trouet et al., 2010; Taylor &
Scholl, 2012). Increasing temperatures and earlier spring snowmelt have been linked to an
increase in the frequency of large wildfires in the western US and more frequent large
wildfires are projected with continued warming (Westerling et al., 2006, 2011a,b). As area
burned increases, the area burned by severe fire is likely to increase (Dillon et al., 2011;
Miller & Safford, 2012; Harris & Taylor, 2015), especially when fires burn in dry conditions
under extreme fire weather in forests that have homogenized structure resulting from past
management actions (McKelvey et al., 1996; Miller et al., 2009; Van Mantgem et al., 2013;
Collins, 2014).
The interaction of warmer, drier climate and increasing area burned, coupled with increasing
fire severity (e.g., proportion of trees killed) resulting from fire-exclusion (Miller et al., 2009)
and past logging activity (McKelvey et al., 1996) has the potential to alter forest composition
and community assembly. Species with a higher tolerance to drought and fire may eventually
gain competitive advantage over less-tolerant species within the community (Stevens et al.,
2015). However, the effects of projected climate-wildfire interactions on forest cover and
species distributions may vary as a function of scale. The substantial change of vegetation
types projected by biome-based simulation approaches (Bachelet et al., 2001; Hayhoe et al.,
2004; Lenihan et al., 2008; Gonzalez et al., 2010) may overestimate the potential for change.
The variability in species-specific tolerance to environmental change may drive community
composition change, without a concomitant shift in the vegetation type (Zhu et al., 2012;
Svenning & Sandel, 2013).
Given the climatic constraints on species recruitment and the dispersal limitations resulting
from the increasing extent of high-severity fires (Miller & Safford, 2012), delayed forest
recovery could impact forest C dynamics. Fire-induced tree mortality can transition a forest
from a C sink to a C source, lowering landscape-level C sequestration potential (Dore et al.,
2008). The time required for the burned forest to return to a C sink depends on post-fire
succession. If the successional pathway results in re-establishment of the pre-fire community,
forest growth will re-sequester the C lost from fire. However, if changes in climate and fire
regime slow or alter post-fire succession, the burned area may transition to a lower C state
community type and this transition from forest to shrubland or grassland can be reinforced by
subsequent burning (Hurteau & Brooks, 2011; Collins & Roller, 2013; Stephens et al., 2013;
Lauvaux et al., 2016).
The Sierra Nevada Mountains are occupied by a diversity of tree species and forest types,
with distributions being shaped by climate and wildfire patterns that tend to sort by elevation
(van Wagtendonk & Fites-Kaufman, 2006). Forest types range from low-elevation dry forests
and woodlands to mid-elevation mixed-conifer forests and high-elevation upper montane and
subalpine forests. Mid-elevation forest composition has been most impacted by fire-
exclusion, transitioning these forests from being dominated by drought-tolerant, fire-resistant
pines to drought-intolerant, fire-sensitive firs (McKelvey et al., 1996; Scholl and Taylor,
2010). Given the substantial latitudinal and elevational range that Sierra Nevada forests span
and the range of species-specific physiological tolerance to stressors, we asked the question:
how will forest composition and community assembly as well as associated C dynamics
change across the landscape under projected climate-wildfire interactions? We used a
species-specific, spatially explicit landscape modeling approach to evaluate the effects of
climate change and climate-driven changes in area burned on forest dynamics. We used
Accepted Article
This article is protected by copyright. All rights reserved.
climate projections from three climate models driven by a medium-high emission scenario in
combination with corresponding climate-specific large wildfire projections to evaluate the
effects of different climate-wildfire scenarios. We hypothesized that: 1) the change in
overstory species composition would be less extensive than the change in recruitment
because young individuals tolerate a more constrained range of abiotic conditions than
mature individuals, 2) projected climate and wildfire would favor the recruitment of drought-
tolerant species over drought-intolerant species and that this change would be greatest at mid-
elevations where drought-intolerant species comprise a majority of the forest community, 3)
projected climate and wildfire would result in communities with lower species richness, and
4) warmer, drier conditions and larger wildfires would transition more of the landscape to a C
source.
Materials and methods
Study area description
Our study area comprised approximately 3.4×10
6
ha of forested land over the Sierra Nevada
Mountains of California and Nevada, USA (Fig.1). Other vegetation types, such as shrubland
and grassland, are also distributed across the Sierra Nevada but these vegetation types were
not included in our simulations because of computational limitations. Our study area spans a
substantial elevation gradient (Fig. S1a). The gradual western slope constitutes the majority
of our study area, while the steep eastern slope occupies a narrow strip of the study area
(Fig.1). The climate is primarily Mediterranean with dry summers and wet winters (van
Wagtendonk & Fites-Kaufman, 2006). More than half of the total precipitation falls as snow,
with snowmelt from persistent snowpack providing a source of moisture into summer. Total
precipitation varies over the region, decreasing from north to south and from high to low
elevation. Precipitation is also higher on the western slope than on the eastern slope, due to
the rain shadow effect. Fire activity mainly occurs during the annual drought period when
there is little rain (Westerling et al., 2003; Syphard et al., 2011). Soils in the study area are
primarily classified as shallow, well-drained Entisols and Inceptisols, but some more
developed Alfisonls, Mollisols, and Andisols exist (NRCS, 2013).
Forest type varies by elevation (Fig. S2), with low-elevation forests being more moisture
limited and higher-elevation forests being more temperature limited. On the western slope of
the study area, the low-elevation forests and woodlands are primarily comprised of gray pine
(Pinus sabiniana), ponderosa pine (P. ponderosa) and oaks (Quercus spp.). The mid-
elevation forests are dominated by a mix of conifers including white fir (Abies concolor),
Douglas-fir (Pseudotsuga menziesii), ponderosa pine, Jeffrey pine (P. jeffreyi), sugar pine (P.
lambertiana) and incense-cedar (Calocedrus decurrens). The upper montane and subalpine
forests mainly consist of red fir (A. magnifica), western white pine (P. monticola), mountain
hemlock (Tsuga mertensiana), lodgepole pine (P. contorta) and whitebark pine
(P. albicaulis). On the eastern slope of Sierra Nevada, the forest communities are similar, but
typically occur at higher elevation because of lower precipitation. The primary vegetation
differences are a higher proportion of Jeffrey pine at mid-elevation in the eastside forests and
the lower elevation eastern woodlands are primarily comprised of singleleaf pinyon pine (P.
monophylla). Chaparral communities are persistent at some locations in the Sierra Nevada
(Keeley et al., 2005), however, the focus of our study was on tree species and we did not
include parameterization for shrub species.