ChapterPDF Available

Drivers of Future Ecosystem Change in the US Pacific Northwest: the Role of Climate, Fire, and Nitrogen


Abstract and Figures

The US Pacific Northwest (PNW) is home to a wide array of ecosystems and climates. Because of its dry summers, the PNW was sensitive to large‐scale climate perturbations in the past and may be vulnerable to climate change during the coming century. In this chapter we explore the resilience and vulnerability of PNW ecosystems with the MC1 vegetation model. MC1 was run with historical and future climates using a range of projections and model configurations. Fire frequency increased ubiquitously, but the domain remained a carbon sink in most scenarios because of increased productivity in nonsummer months. Thresholds of summer drought were exceeded such that the maritime forests lost a large amount of carbon (24%) due to fire in the most extreme climate projection. Fire suppression changed the magnitude, but not the sign, of future changes, and may become less effective at limiting future burned area. The representation of nitrogen mainly affected tree‐grass competition in arid shrublands. Overall, simulations suggest the PNW has the potential to either sequester or emit approximately 1 pg C over the next century, or roughly 23 times the size of Oregon and Washington’s combined annual fossil fuel emissions.
Content may be subject to copyright.
Global Vegetation Dynamics: Concepts and Applications in the MC1 Model, Geophysical Monograph 213, First Edition.
Edited by Dominique Bachelet and David Turner.
© 2015 American Geophysical Union. Published 2015 by John Wiley & Sons, Inc.
In this study we used the MAPSS‐CENTURY 1 (MC1)
dynamic global vegetation model (DGVM) to assess the
impacts of climate change on ecosystem dynamics in the
US Pacific Northwest (PNW), defined here by the major-
ity of the states of Oregon and Washington (Figure7.1).
The seasonal migration of large‐scale pressure systems
over the northern Pacific Ocean results in generally cool
wet winters and warm dry summers across the region.
Despite this seasonal coherence, the PNW exhibits high
levels of climatic, topographical, and ecosystem diver-
sity. For example, some of the world’s highest biomass
forests are found in the western PNW, where high rain-
fall, mild winters, and infrequent fires allow for highly
productive forests and carbon accumulation [Smithwick
et al., 2002; Hudiburg et al., 2009]. However, the Cascade
mountain range induces a strong rain shadow to the east
that supports arid and semiarid flora. Upland sites east
of the Cascade Crest are dominated by lower‐density
mixed‐conifer forests with relatively frequent fires (return
intervals of 15−80 years) [Agee, 1996]. Lower elevations
are increasingly arid and dominated by mixtures of
woodlands, shrublands, and grasslands.
Drivers of Future Ecosystem Change in the US Pacific Northwest:
TheRole of Climate, Fire, and Nitrogen
Brendan M. Rogers1, Dominique Bachelet2, Raymond J. Drapek3, Beverly E. Law4,
Ronald P. Neilson3, and John R. Wells4
1 Woods Hole Research Center, Falmouth, Massachusetts,
2 Conservation Biology Institute, Corvallis, Oregon, USA
3 US Department of Agriculture (USDA) Forest Service, Pacific
Northwest (PNW) Research Station, Oregon, USA
4 Department of Forest Ecosystems and Society, Oregon
State University, Corvallis, Oregon, USA
The US Pacific Northwest (PNW) is home to a wide array of ecosystems and climates. Because of its dry
summers, the PNW was sensitive to large‐scale climate perturbations in the past and may be vulnerable to
climate change during the coming century. In this chapter we explore the resilience and vulnerability of PNW
ecosystems with the MC1 vegetation model. MC1 was run with historical and future climates using a range of
projections and model configurations. Fire frequency increased ubiquitously, but the domain remained a carbon
sink in most scenarios because of increased productivity in nonsummer months. Thresholds of summer drought
were exceeded such that the maritime forests lost a large amount of carbon (24%) due to fire in the most extreme
climate projection. Fire suppression changed the magnitude, but not the sign, of future changes, and may
become less effective at limiting future burned area. The representation of nitrogen mainly affected tree‐grass
competition in arid shrublands. Overall, simulations suggest the PNW has the potential to either sequester or
emit approximately 1 pg C over the next century, or roughly 23 times the size of Oregon and Washington’s
combined annual fossil fuel emissions.
0002548110.indd 91 8/6/2015 7:12:58 PM
Although parts of the PNW receive high rainfall, the
distinct seasonality and related summer droughts have
rendered the region sensitive to climate fluctuations in the
past. This is particularly true for regional fire regimes
[McKenzie et al., 2004; Gavin et al., 2007] and distribu-
tions of mesophytic and xerophytic taxa [Whitlock et al.,
2003]. For example, increased seasonality due to orbital
effects in insolation in the early Holocene (ca. 11,000
years BP) resulted in higher fire frequencies, higher alpine
treelines, lower lake and stream levels, and an expansion
of xerophtyic taxa [Whitlock et al., 2003]. Our primary
reference point for the PNW, the 20th century, has been
the wettest and least variable period during the past mil-
lennium [Millar et al., 2006]. Yet climate is projected to
change during the 21st century by magnitudes similar to
or greater than those of the past 20,000 years, and at
greater rates [Jackson and Overpeck, 2000; Mote and
Salathe, 2010]. Most climate projections include warmer,
drier summers, even if annual precipitation increases
[Climate Impacts Group, 2004; Littell et al., 2009; Mote
etal., 2008; Mote and Salathe, 2010]. We may therefore
expect some analogous and some wholly unexpected
alterations to PNW ecosystems compared to those in
the paleoecological record. Indeed, we may already be
witnessing climate‐induced alterations in disturbance and
mortality [Westerling et al., 2006; Van Mantgem etal.,
2009; Melilo et al., 2014]. Changes in state for the mesic
western forests, in particular, could signify a large regional
contribution to positive climate−carbon cycle feedbacks.
Information on potential ecosystem changes is valuable
for climate mitigation policies, biodiversity and afforesta-
tion efforts, parks management, fuels treatments, and
firefighting practices.
A number of studies have projected substantial changes
to vegetation [e.g., Thompson et al., 1998; McKenzie and
Halpern, 1999; Shafer et al., 2001; Hamann and Wang,
2006; Kerns et al., 2009], fire regimes [e.g., Flannigan et al.,
2000; McKenzie et al., 2004; Spracklen et al., 2009; Littell
et al., 2010], and carbon stocks [e.g., Raymond and
McKenzie, 2012] in the PNW using statistical approaches.
While highly informative, these approaches have several
limitations. For one, extrapolation issues are present in
almost all these statistical models, as they are applied to
novel climate assemblages not seen in the historical
records used to derive them. These models also do not
couple multiple related ecological proceesses, such as fire
frequency, fire severity, vegetation competition and suc-
cession, and carbon stocks, and therefore cannot project
g C m–2 y–1
> 150
Figure7.1 (a) Analysis regions within the Pacific Northwest study domain and (b) average county‐level timber
removals from 1965 to 2002. Regions in (a) were derived from level III ecoregions [Bailey, 1995]. The western
forests (WF) include the Coast Range, Klamath Mountains, Willamette Valley/Puget Trough, and North Cascades
ecoregions. The eastern forests (EF) include the East Cascades, Okanagan, and Blue Mountains ecoregions, and
the Columbia Plateau (CP) comprises its own analysis region. Harvest data in (b) were obtained from Washington
State DNR [2009] and Oregon Department of Forestry [2009]. For color details, please see color plate section.
0002548110.indd 92 8/6/2015 7:12:59 PM
changes to the totality of an ecosystem’s state. Although
current mechanistic models may misrepresent some
aspects of ecosystem dynamics due to choice of algo-
rithms, overparameterization, omission of processes, and
issues of spatial and temporal scaling, these models cou-
ple the majority of interrelated processes that often feed
back to each other in crucial and nonlinear ways. Thus,
they are arguably the best available tool for projecting
ecosystem form and function under novel future climates
and over large spatial domains [Diffenbaugh et al., 2003;
Gavin et al., 2007; Scheller and Mladenoff, 2007; Littell
etal., 2009].
Here we investigate the major drivers of change in the
PNW by running the MC1 DGVM in various configura-
tions under future climate scenarios, selected to bound
the uncertainties of future climate changes. We focus on
how the interactions between climate, fire, and nitrogen
drive change and resilience in the diverse PNW ecosys-
tems. Fire in the PNW’s western mesic forests is histori-
cally infrequent (typical fire return intervals of 150−500
years [Agee, 1996]) but intense and stand‐replacing.
Increases infire frequency, due to hotter and longer sum-
mers, have the potential to dramatically alter these forests.
East of the Cascade Range, in contrast, forests and shrub-
lands experience frequent fires. However, these fires have
been aggressively suppressed since the mid‐20th century,
resulting in forests with elevated fuel loads and higher fire
danger risk [Stephens et al., 2009]. We address the role of
fires and fire suppression by developing an intensity‐
based fire suppression rule in MC1, as described in Rogers
et al. [2011], and characterizing how the representation of
fire influences ecosystem projections.
All terrestrial systems are limited by nitrogen to some
degree [LeBauer and Treseder, 2008]. The demand for
nitrogen in photosynthetic enzymes has resulted in com-
plex interactions between plants and nitrogen‐fixing
bacteria that influence the competition between species
or plant functional types. We examine the controls that
nitrogen limitation places on the PNW’s future carbon
balance and vegetation compositions in MC1 by compar-
ing simulations with [Rogers, 2009] and without [Rogers
et al., 2011] nitrogen limitation. We also quantify the
effect of CO2 fertilization, interannual variability in pre-
cipitation, and choice of fossil fuel emissions scenario. We
end with a discussion on lessons learned for other regions,
advantages and deficiencies of MC1 compared to other
models, and suggestions for future model development.
7.2.1.Study Area
Our study domain covers the latitudinal range and
roughly the western three‐quarters of Oregon and
Washington (Figure 7.1). The longitudinal expanse was
selected to represent the majority of climatic and ecosys-
tem diversity within the region without unnecessarily
increasing computing demands. We aggregated level III
ecoregions [Bailey, 1995] into analysis regions, based on
similar vegetation communities, historical climates, and
responses to future climate projections (Figure 7.1a).
The western forests are characterized by a maritime cli-
mate with high rainfall; between 1971 and 2000 the region
averaged 2003 mm of mean annual precipitation (MAP)
[Daly et al., 2008]. Except for human‐dominated and
altered landscapes in the Willamette Valley, this region is
composed of high‐biomass conifer forests with infrequent
fires (return intervals between 150 and >500 years [Agee,
1996]), although much of it was subject to logging during
the 20th century (Figure7.1b). The mixed‐conifer eastern
forests are considerably drier (717 MAP), and burn much
more frequently (roughly every 15−80 years [Agee, 1996]).
The Columbia Plateau is the driest of the regions (295
MAP) with variable fire return intervals (7–200 years
[LANDFIRE Rapid Assessment, 2007; Barrett et al.,
2010]), dependent on fuel moisture and connectivity.
Thisregion is dominated by woodlands, shrublands, and
7.2.2.Model and Input Data
We used the MC1 DGVM to understand and project
ecosystem states in the Pacific Northwest. Described thor-
oughly in Chapter1 of this book, MC1 is a mechanistic
vegetation model that runs on a monthly timestep and is
able to capture the interactions between climate and veg-
etation types, disturbance, and ecosystem carbon balance.
Although somewhat similar to other large‐scale DGVMs
[e.g., Moorcroft et al., 2001; Cox, 2001; Sitch et al., 2003],
biogeographical categories in MC1 are generally more
nuanced and relate well to vegetation communities in the
PNW (Table7.1).
MC1 requires inputs of static soil composition (mineral
depth and sand, clay, and rock fragments), monthly
climate (precipitation, vapor pressure, mean temperature,
and mean daily minimum and maximum temperature),
and annual atmospheric CO2 concentrations. Soils data
were modified from Kern [1995], and historical climate
data (for 1895−2006) were obtained from the PRISM
model [Daly et al., 2008] at 30 arc‐second resolution
(approximately 0.6 km2). Future climate projections were
obtained from three general circulation models (GCMs)
used in the Intergovernmental Panel on Climate Change
(IPCC) Fourth Assessment Report (AR4) [IPCC, 2007]
and made available through the Coupled Model
Intercomparison Project phase 3 (CMIP3) [Meehl et al.,
2007]: CSIRO‐Mk3.0 [Gordon, 2002], MIROC 3.2 medres
[Hasumi and Emori, 2004], and Hadley CM3 [Johns et al.,
2003] (hereafter CSIRO, MIROC, and Hadley). We
selected projections forced by the Special Report on
0002548110.indd 93 8/6/2015 7:12:59 PM
Emissions Scenarios (SRES) B1, A1B, and A2 emission
scenarios [Nakicenovic et al., 2000]. GCM climate fields
were downscaled to our relatively fine‐scale grid using the
delta, or perturbation, method [Fowler et al., 2007].
Further details are given in Rogers et al. [2011].
Mote and Salathé [2010] compared these and 18 other
GCMs from the IPCC AR4 to observed climate in the
PNW (note that these authors analyzed CSIRO‐Mk3.5
instead of CSIRO‐Mk3.0). Each of the three we selected
displayed their own strengths and weaknesses relative to
the larger group. For example, Hadley exhibited one of
the lowest precipitation biases, both annually and season-
ally, but was relatively poor in representing the spatial
distribution of several meteorological fields and failed to
simulate a 20th‐century temporal trend in air tempera-
ture. MIROC, on the other hand, displayed one of the
lowest temperature biases but one of the highest precipi-
tation biases. CSIRO ranked highest of all models for its
20th‐century temperature trend, yet exhibited the highest
positive temperature bias.
7.2.3.Model Calibration
A number of model inconsistencies were corrected
[Rogers, 2009] and parameters were augmented for cali-
bration of MC1 to the Pacific Northwest. These included
biogeography thresholds (such as temperature limits for
temperate vs. subtropical vegetation and woody carbon
limits for forests vs. woodlands vs. shrublands) and
parameters regulating woody and herbaceous mortality,
productivity, evaporation, and transpiration. Vegetation
types for the historical analysis period (1971−2000) run
with full fire were compared to a potential vegetation
map from Küchler [1975], aggregated to 35 classes as part
of the Vegetation/Ecosystem Modeling and Analysis
Project (VEMAP) [Kittel et al., 1995]. Regional examples
of simulated vegetation types are given in Table7.1.
Carbon fluxes and stocks were compared to four data-
sets. We first calibrated net primary production (NPP),
mortality, and live and dead aboveground and below-
ground carbon pools to an aggregated database of peri-
odic Forest Inventory Analysis (FIA) plots in Oregon
from Hudiburg et al. [2009]. MC1 was run with full fire,
fire suppression, and no fire to emulate sites with distinct
disturbance histories. We then validated ecoregion‐
specific biomass using a gridded map of aboveground
forest carbon from Blackard et al. [2008]. We also vali-
dated old‐growth forests in the western forests using
observations from Smithwick et al. [2002], periodic FIA
plots from Hudiburg et al. [2009], and Environmental
Protection Agency (EPA) plots from the Oregon and
California (ORCA) regional carbon study [Sun et al.,
2004; Hudiburg et al., 2009]. We used a lower bound of
180 years to define old‐growth forests, where trees were
measured according to the Spies and Franklin or Van
Tuyl method [Spies and Franklin, 1991; Van Tuyl et al.,
2005]. To simulate old‐growth forests, fire was turned off
in MC1 to ensure no disturbance. The majority of all
plot data were collected between 1991 and 2001. We
therefore used model averages from 1991 to 2001 for
comparisons. Finally, we compared the seasonality of
NPP in MC1 to the MODIS Aqua satellite product
[Running et al., 2004] and eddy covariance data from the
Metolius Intermediate Pine site [Law et al., 2003]. In all
site‐level comparisons, model values were extracted from
the grid cell containing the centroid of a given observa-
tion. The model was calibrated in the same manner with
and without nitrogen limitation.
Table7.1 Regional examples of vegetation types in MC1.
MC1 Vegetation Type Regional Examples
Ice Barren rock, permanent snowpack
Tundra Alpine meadows
Subalpine forest Subalpine fir,a lodgepole pine,b mountain hemlock,c whitebark pined forest
Maritime conifer forest Douglas fir,e western hemlock,f Sitka spruce,g Pacific silver firh forest
Temperate conifer forest Ponderosa pine,i Douglas fir, lodgepole pine, grand fir,j white fir,k western juniperl forest
Temperate cool mixed forest Douglas fir−Garry oakm forest
Temperate warm mixed forest Douglas fir−Garry oak−Pacific madronen forest
Temperate deciduous broadleaf forest Garry oak, bigleaf mapleo forests
Temperate conifer woodland Ponderosa pine−western juniper woodland
Temperate shrubland Big sagebrushp−Bluebunch wheatgrass,q big sagebrush−Idaho fescuer
Temperate grassland Idaho fescue, Bluebunch wheatgrass
Subtropical mixed forest Douglas fir−Pacific madrone−tanoak forests
aAbies lasiocarpa; bPinus contorta; cTsuga mertensiana; dPinus albicaulis; ePseudotsuga menziesii; fTsuga heterophylla;
gPicea sitchensis; hAbies amabilis; iPinus ponderosa; jAbies grandis; kAbies concolor; lJuniperus occidentalis; mQuercus
garryana; nArbutus menziesii; oAcer macrophyllum; pArtemisia tridentate; qPseudoroegneria spicata; rFestuca idahoensis;
sLithocarpus densiflorus.
0002548110.indd 94 8/6/2015 7:12:59 PM
We developed a fire suppression rule based on simu-
lated fireline intensity, rate of spread, and energy release
component. If the energy release and either the fireline
intensity or rate of spread were above a given threshold,
fires were allowed to burn naturally in MC1. Otherwise
the fire was suppressed and the burned area was set at
0.06% of a grid cell. We calibrated these thresholds to two
sets of observations: (1) a domain‐wide burned area data-
set at 1° from 1980 to 2004 [Westerling et al., 2003], and
(2) the observation that roughly 95% of fires in the west-
ern US, have been suppressed since midcentury, and that
a small fraction of the escaped fires (2−5%) contribute
95% of the burned area [Graham et al., 1999]. The result
was model thresholds that represent levels at which fire-
fighting practices begin to lose effectiveness. For example,
fireline intensities over approximately 2−3 MW m1 gen-
erally are seen in passive and active crown fires, which
greatly hinder control efforts [Agee, 1996]. Fires spread-
ing faster than approximately 0.5 m s‐1 (1.1 mph) become
increasingly difficult to control, partly because of the dif-
ficulty in maintaining this speed through vegetated and
often steep terrain. We chose to compare MC1 with the
Westerling et al. [2003] dataset as opposed to a compara-
ble polygon‐based fire dataset for the US, the Monitoring
Trends in Burn Severity (MTBS) database [Eidenshink
etal., 2007], because MC1 is not expected to simulate fire
on exactly the same part of the landscape as in reality, but
rather capture patterns in large landscape blocks (e.g., 1°
grid cells) that contain comparable vegetation and are
influenced by similar climate patterns. The Westerling
dataset also extends further back in time than MTBS.
Finally, we compared simulated combustion factors
(fraction of preburn mass lost to combustion in a fire)
from 2002 in the Klamath Mountains ecoregion with
data from Campbell et al. [2007] on the 2002 Biscuit fire
in southwest Oregon. Combustion factors from individ-
ual carbon pools were combined using mass weighting
factors from both Campbell et al. [2007] and MC1.
7.2.4.Simulation Protocol
MC1 was spun up using a mean 1895−2006 climate for
up to 3000 years to establish carbon pools, and then with
a detrended and looped 1895−2006 climate time series for
an additional 3000 years to establish a dynamic equilib-
rium between fire and ecosystem properties. Historical
simulations began in 1895 and future scenarios, in 2007.
Fire suppression was initiated in 1940, which corresponds
to the decade when widespread fire suppression was initi-
ated [Pyne, 1982].
To assess the major drivers of change, simulations
were conducted with full fire (i.e., no suppression), fire
suppression, and no fire, as well as with and without
nitrogen limitation. All nine future projections (three
GCMs × three emission scenarios) were used for nitrogen
limitation runs, whereas only three projections with the
A2 emission scenario were used for simulations without
nitrogen limitation.
We also characterized the influence of atmospheric CO2
concentrations on ecosystem projections. MC1 simulates
CO2 fertilization through a logarithmic β‐factor on pro-
ductivity, transpiration, and C:N ratios [Bachelet et al.,
2001]. We tested the influence of CO2 by running the
model with a β‐factor of 0 (no enrichment) and 0.6
(greater enrichment) instead of the default value of 0.25.
Norby et al. [2005] report a β‐factor of 0.6 for experimen-
tal free‐air CO2 enrichment (FACE) sites, but following
results show NPP enhancement may decrease by over 60%
due to water and nitrogen restrictions [Norby et al., 2010].
Interannual variability in precipitation is positively
correlated with fire activity in many semiarid ecosystems
due to the interplay between fuel production in wet years
and severe fire weather in dry years [Veblen et al., 2000;
Westerling et al., 2003; Hessl et al., 2004; Schoennagel
etal., 2005]. To assess the impact of rainfall variability on
future burning and biomass, we placed a 3‐ and a 5‐year
filter on precipitation values from the CSIRO A2 pro-
jection by month. This preserved seasonality patterns
and long‐term means but significantly dampened the
interannual variability. Both of the above analyses were
conducted with nitrogen limitation and full fire in MC1.
Unless otherwise stated, our analyses are based on the
simulations with fire suppression and without nitrogen
limitation presented in Rogers et al. [2011]. Comparisons
between sets of simulations that differ in climate projec-
tion, treatment of fire, representation of nitrogen limita-
tion, and effects of atmospheric CO2 are used to elucidate
the major drivers of ecosystem change in the different
PNW vegetation zones. The historical analysis period
refers to 1971−2000 and the future period to 2070−2099.
7.3.1.Calibration and Historical Simulations
The PNW averaged 8.4°C and 1182 mm mean annual
precipitation (MAP) during the historical period. Although
the domain included distinctly different vegetation and
climate regimes (analyzed here by regions shown in
Figure 7.1a), there were consistent seasonal patterns of
cool wet winters and warm hot, dry summers (Figure7.2).
Simulated net primary productivity peaked in late spring/
early summer and again in early fall when available soil
water was relatively high and temperatures were warm
(Figure7.2). Low temperatures limited productivity in win-
ter, and water stress limited productivity in midsummer.
After calibration, MC1 compared favorably to histori-
cal conditions of vegetation types, carbon stocks, carbon
0002548110.indd 95 8/6/2015 7:12:59 PM
fluxes, and burned area. The spatial distribution of simu-
lated modal (most frequent) vegetation types run with full
fire were in good agreement with the potential vegetation
map from Küchler [1975] (Figure 7.3). Cohen’s kappa
statistic between modeled and observed vegetation was
0.53 when run without, and 0.56 when run with, nitrogen
limitation. Notable disagreements occurred in the central
Willamette Valley and grasslands of the Columbia
Plateau. However, Native Americans and European set-
tlers greatly modified fire regimes in the Willamette Valley
[Whitlock and Knox, 2002], which may have allowed for
the establishment and maintenance of mixed oak forests
and woodlands. Additionally, the Columbia Plateau is
particularly prone to both grass [Keane et al., 2008] and
woody [Belsky, 1996] encroachment. These attributes
may complicate the assignment of potential vegetation
communities within the two regions.
Calibrated carbon pools and fluxes compared well with
the FIA‐derived database from Hudiburg et al. [2009]
(Figure 7.4a), although MC1 simulated consistently
higher carbon pools and fluxes than the FIA plots
(overall bias in carbon pools of +25% and NPP of
+13%). Biases in carbon pools were highest in the
Willamette Valley (+90%), the East Cascades (+41%),
the Coast Range (+37%), and the West Cascades (+25%).
However, the FIA data include sites influenced by both
natural and anthropogenic disturbances [Van Tuyl et al.,
2005], whereas MC1 accounts only for natural fire
disturbance. Many areas within the western forests,
including the Coast Range, Willamette Valley, and to
some extent the West Cascades ecoregions, were subject
to heavy human influence during the 20th century,
including land conversion, logging (Figure 7.1), and
urbanization. These all act to decrease total ecosystem
carbon [Smithwick et al., 2002]. MC1 overestimated
mean total live carbon in the western forests (Figure7.4b)
in large part because it does not account for these human
influences. In the drier eastern systems, 20th‐century fire
suppression has increased woody and total carbon [Agee,
1996]. Consistent with this, live aboveground forest
Water stress (unitless)
mm month–1 g C month–1
Water stress NPP
Figure7.2 Mean monthly historical (1971−2000) and future climate projections (2070−2099) of surface air
temperature, precipitation, and simulated water stress and net primary productivity (NPP). Water stress is defined
as 100 × (1 – ws/PET), where ws = mean monthly available soil water and PET = potential evapotranspiration.
Figure adapted from Rogers et al. [2011].
0002548110.indd 96 8/6/2015 7:12:59 PM
Observations No N limitation N limitation
Subalpine forest
Maritime conifer forest
Temperate conifer forest
Temperate cool mixed forest
Temperate warm mixed forest
Temperate deciduous forest
Temperate conifer woodland
Temperate shrubland
Temperate grassland
Subtropical mixed forest
Figure7.3 Observed and simulated modal vegetation types calibrated with and without nitrogen limitation for
the historical (1971−2000) time period. In each case MC1 was run with full fire. Observations are derived from
a potential vegetation map from Küchler [1975]. Analysis regions are delineated by black outlines. The first two
panels are adapted from Rogers et al. [2011]. For color details, please see color plate section.
Total below live tree
Coarse root
Large wood
Fine root
Total above dead tree
Total above live tree
Live branch
kg C m–2
g C m–2y–1
kg C m–2
MC1 full fire
MC1 fire suppression
(a) (b)
kg C m–2
MC1 at FIA
MC1 at EPA
MC1 at smithwick
Observational live carbon
Observational dead carbon
MC1 live carbon
MC1 dead carbon
Figure7.4 Comparisons of carbon pools and fluxes between MC1 and observations: (a) FIA plots in Oregon from
Hudiburg et al. [2009] (5093 plots, 90.4% on public lands, with stand ages of 212 ± 134 years), (b) aboveground
live forest carbon from Blackard et al. [2008], and (c) old‐growth plots and MC1 run without fire (Smithwick data
[Smithwick et al., 2002] contain 37 plots on public lands with stand ages of 429 ± 257 years; EPA data [Hudiburg
et al., 2009] contain eight plots on public lands with stand ages of 417 ± 215 years; FIA data [Hudiburg et al.,
2009] contain 1607 plots, 98.2% on public lands with stand ages of 332 ± 123 years). Error bars represent
standard errors. Figure adapted from Rogers et al. [2011]. For color details, please see color plate section.
0002548110.indd 97 8/6/2015 7:12:59 PM
carbon in MC1 simulations with full and suppressed fire
bracketed the observations from Blackard et al. [2008] in
the eastern forests and Columbia Plateau (Figure7.4b).
Comparisons of carbon stocks in old‐growth forests
yielded mixed results depending on the data source
(Figure 7.4c), which may be in part due to differing
methods of plot selection [Van Tuyl et al., 2005].
Calibration of MC1 with and without nitrogen limita-
tion yielded similar accuracies in all fields except for sea-
sonal productivity. Whereas simulations without nitrogen
limitation were able to capture domain‐wide monthly NPP
patterns, those with nitrogen limitation exhibited signifi-
cant high biases in winter and low biases in summer
(Figure 7.5a,b). At the Metolius FLUXNET site [Law
etal., 2003] (Figure7.5c,d), simulated NPP with nitrogen
limitation dropped to zero in midsummer to late summer,
a time when observed NPP was declining but distinctly
positive. In MC1, nitrogen is made available through
atmospheric input (a function of precipitation), decay of
soil organic matter, and abiotic fixation (a function of
actual evapotranspiration). Monthly production is limited
to that which can be produced by a particular C:N ratio,
which is allowed to vary between fixed ranges. Conditions
of low rainfall and evapotranspiration in the PNW in mid-
summer rendered little nitrogen available for plant uptake,
and caused large reductions in modeled productivity.
With full fire, MC1 simulated approximately 8 times
more burned area during 1980−2004 compared to
Westerling et al. [2003] (Figure 7.6). This is consistent
with previous work with MC1 over the continental US,
which showed a model overestimation of burned area by
a factor of ~8 since 1960 [Neilson, 2004]. Before 1960,
MC1 accurately simulated annual continental burned
area. With our new fire suppression rule, MC1 closely
matched the long‐term observational mean, overestimat-
ing burned area from Westerling et al. [2003] by 1.4%
(Figure7.6). This included an underestimation of 13% in
both the western forests and Columbia Plateau, and an
overestimation of 28% in the eastern forests. We also note
that individual fire years were not always captured by the
model (Figure7.6). Although climate exerts a dominant
control on PNW fire behavior [Flannigan et al., 2000;
2003 2004 2005 2006 2007 2003 2004 2005 2006 2007
g C m–2 month–1g C m–2 month–1
g C m–2 month–1
2002 2003 2004 2005 2006
g C m–2 month–1
2002 2003 2004 2005 2006
No N limitation N limitation
Figure7.5 Comparisons of simulated and observed net primary production (a,b) across the domain from MODIS
[Running etal., 2004] and (c,d) at the Metolius Intermediate Pine eddy covariance site [Law et al., 2003] (44.4523°N,
121.5574°W). Simulations without (a,c) and with (b,d) nitrogen limitation are shown. Panels (a) and (c) are adapted
from Rogers et al. [2011].
0002548110.indd 98 8/6/2015 7:12:59 PM
Hessl et al., 2004; Bachelet et al., 2011], our model imple-
mentation was not able to capture the fine‐scale patterns
of ignition, weather, and suppression that were influenced
by county‐level demographics. Simula ted large (unsup-
pressed) fires accounted for 95.2% of the burned area,
closely approximating the observed distribution in western
U.S. forests [Graham etal., 1999]. MC1 tended to overes-
timate combustion factors compared to Campbell et al.
[2007]. The mean simulated mass‐weighted fraction of
biomass combusted in the Klamath Mountains in 2002
was 0.26, whereas Campbell et al. [2007] report high‐,
medium‐, and low‐severity fire combustion factors of
0.23, 0.17, and 0.14, respectively, for the 2002 Biscuit fire.
During the historical period the domain averaged 29.4
kg C m2 in total ecosystem carbon, 0.33% annual burned
area, 2.75 kg C m2 in combustion (biomass consumed
per unit burn area), and 8.95 g C m2 yr1 in biomass con-
sumed by fire (region‐level annual biomass losses to fire)
(Table7.2). The mesic western forests exhibited the high-
est ecosystem carbon (44.4 kg C m2) and mean combus-
tion (4.02 kg C m2), and the lowest annual burned area
(0.14%). The drier eastern forests displayed the highest
annual burned area (0.64%) and biomass consumed by
fire (21.9 g C m2 yr1), with moderate ecosystem carbon
(22.5 kg C m2). The lowest ecosystem carbon (12.2 kg C
m2), combustion (1.05 kg C m2), and biomass con-
sumed by fire (3.79 g C m2 yr1) was found in the arid
Columbia Plateau. We note that the inclusion of fire
suppression here caused large‐scale conversion from
shrublands and grasslands (shown in Figure7.3 with full
fire) to woodlands and forests.
7.3.2.Future Climate Projections
Downscaled future climate projections displayed both
similarities and differences in the seasonality and magni-
tudes of changes (as calculated from future 2070−2099
monthly means relative to 1971–2000 historical means)
(Figure7.2). Temperatures rose ubiquitously with larger
increases in summer than winter. Consistent with Mote
and Salathé [2010], increased precipitation generally
occurred in winter and decreased in summer months.
Comparatively speaking, the CSIRO climate projection
was warm and wet (+2.6°C and +176 mm MAP), MIROC
was hot and wet (+4.2°C and +82 mm MAP), and Hadley
was hot and dry (+4.2°C and 78 mm MAP). The latter
was especially influential in summer, when temperature
increases approached 8°C. This almost doubled mean
vapor pressure deficit across the domain in June, July,
and August from 0.89 kPa during the historical period to
1.64 kPa. These three GCM projections covered a wide
range in climate change space and are likely indicative of
the spectrum of possible future climates for the region.
The MIROC and Hadley climates lengthened the grow-
ing season and amplified the already strong seasonal
climatic cycles, thereby increasing simulated NPP during
the rainy season and decreasing NPP in summer because of
exacerbated drought stress (Figure 7.2). Under CSIRO’s
milder conditions, NPP increased and water stress decreased
year‐round. As a general trend, the seasonal amplitudes of
simulated plant functions and stressors were amplified by
future climates, thereby increasing summer drought and
susceptibility to fires, but also increasing productivity dur-
ing the rest of the year. These ecosystem responses have
been observed in the tree ring record [Villalba etal., 1994]
and predicted under projected future climates for lodgepole
pines in other parts of the American West [Smithwick et al.,
2009]. How these drivers influenced ecosystem dynamics in
MC1, however, differed greatly by region.
7.3.3.Future Ecosystem Changes by Region
The mesic high‐biomass western forests, typically con-
sidered stable with long fire return intervals, proved to be
the most vulnerable of the three PNW regions in MC1.
This was due entirely to fires. The maritime forests were
largely unable to benefit from increased winter precipita-
tion because, as has been observed [Harr, 1977], soils are
already saturated with winter. Instead, the western forests
suffered from more intense summer droughts that caused
significant increases in fire activity (Table7.2). Large fires
were simulated in years with summer droughts substan-
tially worse than those during the historical period. These
droughts occurred more often under CSIRO and MIROC
after 2070 (Figure7.7), and were mainly limited to the
southwest part of the region. Yet they occurred much
more frequently under Hadley throughout the 21st cen-
tury and caused fires across all of the western forests.
With fire turned off, the western forests gained carbon
under all three climate projections (+1.2 kg C m2 in
CSIRO, +0.5 kg C m2 in MIROC, and +1.3 kg C m2 in
1980 1985 1990 1995 2000 2005
% area burned
Observations MC1 full fire MC1 fire suppression
Figure 7.6 Domain‐wide annual burned area from observa-
tions [Westerling etal., 2003] and MC1 simulated with full fire
and fire suppression. Figure adapted from Rogers et al. [2011].
0002548110.indd 99 8/6/2015 7:13:00 PM
Hadley), due to warmer temperatures and increased pro-
ductivity. Although this increased production compen-
sated for greater fires in CSIRO and MIROC, and the
region remained a carbon sink, burned area and biomass
consumed by fire increased by an order of magnitude
under Hadley and the western forests lost nearly a quar-
ter of their large ecosystem carbon stocks (Table 7.2,
Figure7.7). The dependence on fire for an example grid
cell in the West Cascades can be seen in Figure7.8, where
one large fire disrupts the continued carbon sink in
MIROC, and a new regime characterized by frequent
fires resulted in a complete change of state under Hadley.
Hadley projections also displayed large‐scale forest
conversions (Figure7.9), with maritime coniferous forests
being replaced by temperate conifers. This implies that
fir‐hemlock‐spruce forests west of the Cascade Range
could be vulnerable to invasion by pines because of
reduced rainfall, increases in the seasonal amplitude of
temperature, and more fires. Taken together, simulations
for western forests under the Hadley climate constitute a
monumental change in state and somewhat resemble the
region’s climate, vegetation, and fire regimes of the late
Holocene thermal maximum [Whitlock et al., 2003]. This
includes drier‐adapted vegetation with more frequent
fires and a warmer and more variable climate, although
winter insolation and temperatures during the Holocene
thermal maximum were lower than today.
Other notable vegetation changes in the western forests
included the conversion of maritime conifer to warm
temperate and subtropical mixed forests in the central/
southern reaches due to elevated winter temperatures
(especially prominent in MIROC), and the widespread
loss of subalpine forests (Figure7.9). Regarding the lat-
ter, the ecotone between temperate and subalpine forests
increased by 253–786 m and was highly correlated with
mean domain‐wide increases in temperature (r2 = 0.92).
In the model this was due to biogeographic thresholds of
growing degree‐days.
Table7.2 Mean historical (1971−2000) and future (2070−2099) burned area, fire‐related carbon fluxes, and total ecosystem
carbon by region. Values are adapted from Rogers et al. [2011].
Domain/Variable Historical
All domains
Burned area (% per year) 0.33 0.57 (+76) 0.63 (+95) 1.34 (+310)
Combustionb (kg C m−2) 2.75 3.55 (+29) 3.74 (+36) 3.87 (+41)
Biomass consumedc (g C m−2 yr−1) 8.9 20.4 (+128) 23.7 (+165) 51.7 (+478)
Ecosystem carbond (kg C m−2) 29.4 33.0 (+12) 32.3 (+10) 25.5 (13)
Western forests
Burned area 0.14 0.37 (+161) 0.37 (+160) 1.83 (+1177)
Combustion 4.02 3.89 (3.1) 4.37 (+8.7) 4.44 (+11)
Biomass consumed 5.7 14.6 (+153) 16.2 (+182) 81.2 (+1313)
Ecosystem carbon 44.4 45.5 (+2.5) 45.2 (+1.7) 33.8 (24)
Eastern forests
Burned area 0.64 1.34 (+111) 1.54 (+141) 1.49 (+133)
Combustion 3.42 3.52 (+2.7) 3.74 (+9.1) 3.18 (7.2)
Biomass consumed 21.9 47.3 (+116) 57.6 (+163) 47.4 (+117)
Ecosystem carbon 22.5 28.1 (+25) 27.5 (+22) 22.1 (1.7)
Columbia Plateau
Burned area 0.36 0.26 (29) 0.32 (11) 0.46 (+28)
Combustion 1.05 2.96 (+182) 2.67 (+155) 2.21 (+111)
Biomass consumed 3.8 7.6 (+102) 8.6 (+126) 10.3 (+171)
Ecosystem carbon 12.2 18.1 (+47) 16.8 (+37) 15.4 (+26)
All domains (full fire)
Burned area (% per year) 1.94 1.92 (1.1) 2.01 (+3.9) 2.75 (+42)
Combustionb (kg C m−2) 1.25 1.64 (+31) 1.67 (+34) 2.24 (+80)
Biomass consumedc (g C m−2 yr−1) 24.2 31.5 (+30) 33.7 (+39) 61.5 (+154)
Ecosystem carbond (kg C m−2) 28.5 30.7 (+7.6) 30.1 (+5.5) 23.5 (18)
aPercent changes from historical are given in parentheses.
bCombustion is defined by the amount of biomass consumed and released to the atmosphere per unit burn area.
cBiomass consumed is defined by the regional loss of biomass to the atmosphere due to fires. Biomass consumed =
combustion/(fraction of area burned per year).
dEcosystem carbon includes all live and dead, above‐ and belowground carbon pools.
0002548110.indd 100 8/6/2015 7:13:00 PM
2000 2020 2040 2060 2080 2100
2000 2020 2040 2060 2080 2100
2000 2020 2040 2060 20802100
2000 2020 2040 2060 20802100
2000 2020 2040 2060 20802100
2000 2020 2040 2060 20802100
Ecosystem carbon
Biomass consumed by fire
kg C m–2
kg C m–2
Western forestsEastern forests
g C m–2 y–1
g C m–2 y–1
kg C m–2
kg C m–2
g C m–2 y–1
g C m–2 y–1
kg C m–2 kg C m–2
g C m–2 y–1
g C m–2 y–1
Figure7.7 Simulated time series of 21st‐century total ecosystem carbon, net ecosystem production (NEP), and biomass consumed by fire for (a,d)
CSIRO, (b,e) MIROC, and (c,f) Hadley projections in the western (a−c) and eastern (d−f) forests. Ecosystem carbon and biomass consumed by fire are
plotted as differences from historical means. Eleven‐year filters were applied to biomass consumed by fire and NEP, which are plotted on the righthand
axes. Note the change of scale for the Hadley projection of western forests.
0002548110.indd 101 8/6/2015 7:13:00 PM
In contrast to the mesic western forests, the xeric eastern
forests were the most resilient region in the PNW. Although
burned area and biomass consumed by fire more than
doubled in every scenario, this did not adversely impact
carbon stocks and the region continued to be either a car-
bon sink or essentially carbon‐neutral (Table 7.2,
Figure 7.7). Unlike the scenarios in western forests, the
scenarios that produced the greatest carbon losses from
fire did not necessarily experience the lowest overall car-
bon gains (Table 7.2). NPP more than compensated for
elevated fire activity because of increased atmospheric
CO2 and higher winter and fall temperatures: CSIRO aver-
aged 107%, MIROC 93%, and Hadley 107% increases in
NPP in the fall and winter months. As in the western for-
ests, fire regimes under Hadley quickly transitioned to a
new state, but this did not translate into biomass losses.
Vegetation composition was also relatively resilient;
although high‐elevation subalpine forests and some shrub-
lands in the south were lost, the region’s temperate conifer
forests (comprising 65% historically) remained unchanged
The arid Columbia Plateau was sensitive mostly to pre-
cipitation and displayed varying levels of carbon gains
that corresponded to woody biomass and precipitation:
increases in woody carbon were 144% in CSIRO, 111% in
MIROC, and 79% in Hadley, opposed to decreases in
grass carbon of 94% in CSIRO, 86% in MIROC, and 80%
in Hadley. Productivity was historically very low in sum-
mer. Trees and shrubs therefore benefited from elevated
temperatures in the nonsummer months and, in the case
of CSIRO and MIROC, precipitation. This resulted in
large‐scale conversions of shrublands to woodlands and
forest (Figure7.9). Combustion and region‐level biomass
losses to fire increased because of woody cover, despite
relatively stable fire frequencies (Table7.2).
The three major ecosystems in the PNW responded in
fundamentally different ways to climate forcing during
the 21st century. This was particularly evident in compar-
ing forests east and west of the Cascade Range. Eastern
forests are accustomed to frequent surface fires that do
not result in high levels of tree mortality [Agee, 1996,
1998]. MC1 was able to capture this, and despite
Figure7.8 Simulated time series of historical and 21st‐century total ecosystem carbon, net ecosystem production
(NEP), and biomass consumed by fire for an example grid cell in the West Cascades ecoregion (46.675°N,
122.341°W). NEP and biomass consumed are plotted on the right‐hand axes. An 11‐year filter was applied to
NEP values, and biomass consumed by fire was divided by 10 to display on the same scale.
1900 1920 1940 1960 1980 2000
2000 2020 2040 2060 20802100
2000 2020 2040 2060 2080 2100
2000 2020 2040 2060 20802100
Ecosystem carbon
(Biomass consumed by fire)/10
kg C m–2
kg C m–2
kg C m–2
kg C m–2
g C m–2 y–1
g C m–2 y–1
g C m–2 y–1 g C m–2 y–1
0002548110.indd 102 8/6/2015 7:13:00 PM
intensified fire weather and activity, no threshold was
crossed such that crown fires killed the majority of tree
biomass in the future. This was in stark contrast to west-
ern forests, whose wet forests with high fuel loads were
dried out sufficiently from the Hadley climate to allow for
large‐scale biomass losses from crown fires. In the
Columbia Plateau, trees and shrubs were able to outper-
form grasses in more productive future climates.
Ubiquitous responses to future climates included
increased productivity, mostly in the nonsummer
months, increased drought stress in the summer, elevated
fire activity, and conversion to warmer‐adapted vegeta-
tion. Historical legacies of fire and stand structure, and
severity of future conditions, ultimately determined
whether fires or productivity dominated the carbon
budget. On the whole, the domain gained 1.12 pg C
under CSIRO and 0.91 pg C under MIROC, but lost
1.23 pg C under Hadley by 2070−2099 when compared
to 1971−2000. To put this in context, 1 pg C is approxi-
mately one‐tenth of our current global annual fossil fuel
emissions [IPCC, 2013] and 23 times the size of Oregon
and Washington’s current combined annual emissions
[Waterman‐Hoey and Nothstein, 2007; Oregon
Department of Energy, 2010].
7.3.4.The Role of Fire and Fire Suppression
Fire proved to be one of the dominant drivers of eco-
system dynamics in future climates. With fire turned off,
total carbon increased throughout the 21st century in all
regions and all future climate projections, ordered by pre-
cipitation increases: the domain gained 6.5% (0.72 pg C)
under CSIRO, 4.2% (0.47 pg C) under MIROC, and 3.0%
(0.33 pg C) under Hadley (Figure 7.10). With fire acti-
vated in the model, both with and without human sup-
pression, thresholds were crossed in the Hadley scenario
such that the western forests burned extensively and the
domain emitted over 1 pg C to the atmosphere. Somewhat
counter‐intuitively, however, almost twice as much carbon
was gained in CSIRO and MIROC with fire suppression
than with fire turned off. This was due to the legacy of
carbon sequestration when suppression efforts began in
1940 (Figure7.10).
Fire activity intensified in all regions and with all sce-
narios (Table7.2, Figure7.10) because of more frequent
and intense summer droughts and, in (semi)arid regions,
elevated fuel loads. Domain‐wide mean combustion
increased (Figure7.10), but this was due mainly to more
wood‐fueled fires in the Columbia Plateau (Table 7.2).
Subalpine forest
Maritime conifer forest
Temperate conifer forest Temperate conifer woodland Temperate grassland
Subtropical mixed forest
Temperate shrublandTemperate cool mixed forest
Temperate warm mixed forest
Figure7.9 Simulated modal vegetation types with and without nitrogen for the historical (1971–2000) and future
(2070–2099) time periods from three climate projections. In each case MC1 was run with full fire. Maps without
nitrogen limitation are adapted from Rogers et al. [2011]. For color details, please see color plate section.
0002548110.indd 103 8/6/2015 7:13:00 PM
Because of the higher frequency of fires, the interannual
variability in combustion tended to decrease.
One major concern with fire suppression is that the
buildup of ladder fuels leaves forests more vulnerable to
high‐intensity and high‐mortality crown fires [Stephens
et al., 2009]. Despite the increasing carbon sink with
suppression, MC1 simulated increased future fire risk.
Burned area was, of course, higher in simulations with full
fire (6.0 times more in historical, 3.3 times in CSIRO, 3.2
times in MIROC, and 2.1 times in Hadley). However, rela-
tive increases in burned area and biomass consumed were
significantly higher with suppression (Table7.2). Because
Figure7.10 Historical and future burned area, fire severity (combustion), and ecosystem carbon simulated with
fire suppression, full fire, and no fire. Burned area and fire severity are shown only for simulations with fire sup-
pression, which was initiated in 1940. Shading indicates the historical period. Figure adapted from Rogers et al.
Ecosystem carbon (fire suppression)
Burned area Fire severity
Ecosystem carbon (full fire)
Ecosystem carbon (no fire)
Burned area (% area yr–1)
fire severity (kg C m–2 burned)
Ecosystem carbon (kg C m–2)
1900 1950 2000 2050 2100
0002548110.indd 104 8/6/2015 7:13:01 PM
fires were more intense in future simulations, fewer were
suppressed. For example, the ratio of domain‐wide bio-
mass consumption with fire suppression to that with full
fire increased from 0.37 for the historical period to 0.65,
0.70, and 0.84 for the CSIRO, MIROC, and Hadley cli-
mates, respectively (Table7.2). This decrease in the differ-
ence between the fire suppression and no‐suppression
cases suggests that fires in the future will be increasingly
difficult to fight with current techniques and resources.
7.3.5.The Influence of Nitrogen
Simulations with nitrogen limitation yielded some
overall patterns similar to those without, although a few
diversions between the two sets are of particular note.
The first involves vegetation changes in the Columbia
Plateau. Without nitrogen limitation, trees outperformed
grasses in more productive future climates so that forests
and woodlands replaced large expanses of shrublands
and grasslands. In contrast to this, when nitrogen limita-
tion was introduced, much of the Columbia Plateau tran-
sitioned to grasslands (Figure7.9) and only a minority
became woodlands. In MC1, grasses are better competi-
tors in a nitrogen‐limited environment compared to a
nitrogen‐unlimited one, and are better able to utilize new
nitrogen additions in the upper soil layers.
These latter trajectories, however, were also highly
dependent on fire. An example of this is shown in
Figure7.11. Here, historical conditions for two adjacent
shrubland grid cells were nearly identical. In one
(Figure 7.11a), two large fire events between 2017 and
2023 killed much of the woody biomass. Thereafter
grasses fueled frequent fires and kept woody growth to a
minimum, resulting in a stable grassland. In contrast,
these fires did not occur in the second grid cell
(Figure7.11b), where sustained woody growth resulted in
a woodland by the end of the century. This example high-
lights how slight variations in fire occurrence at particular
times of ecosystem development, a function of subtle
differences in climate, can determine the future state of
the system. It also demonstrates the concept of fire‐
mediated stable equilibria in savanna systems [Hirota
etal., 2011; Staver et al., 2011; Joubert et al., 2012].
Competition between trees and grasses is determined
by a number of factors, including how recruitment, pro-
ductivity, and mortality are influenced by climate, fire,
shading, and nutrient availability; the latter are mediated
by symbiotic relationships with nitrogen‐fixing bacteria.
While a number of studies have shown that the negative
impact of grasses on tree seedling survival is exacerbated
when nitrogen is added to the system [Davis et al., 1999;
Kraaij and Ward, 2006], others have found the opposite
[Siemann and Rogers, 2003]. The relationships may be
highly dependent on the specific system, species of inter-
est, and baseline nutrient loads. While the parameteriza-
tion of these processes in broad plant functional types is
still highly uncertain, these simulations expose how influ-
ential nutrient dynamics can be for future ecosystem
change and development.
The inclusion of nitrogen dynamics also impacted the
net carbon balance. In general, productivity increases
were dampened with nitrogen limitation, and the PNW
gained less carbon (increases of 0.28 pg C with CSIRO,
0.38 pg C with MIROC, and a decrease of 1.10 pg C with
Hadley) compared to simulations without nitrogen limita-
tion. This relationship has been found in other large‐scale
modeling studies [Thornton et al., 2007; Jain et al., 2009;
Gerber et al., 2013] as nitrogen supply is unable to keep
pace with increasing productivity demands. In some cases,
however, mineralization is increased by warmer soils and
carbon sinks are improved when nitrogen dynamics are
added to a model [Sokolov et al., 2008; Smith et al., 2014].
1900 1950 2000 2050 2100 1900 1950 2000 2050 2100
Biomass (g C m
Biomass consumed by
re (g C m
Biomass (g C m
Biomass consumed by
re (g C m
(a) (b)
Live tree carbon
Live grass carbon
Biomass consumed by re
Figure7.11 Time series of live tree carbon, live grass carbon, and biomass consumed by fire for two adjacent
grid cells simulated with the CSIRO A2 future climate with nitrogen limitation: (a) 43.340°N, 120.375°W;
(b)43.425°N, –120.383°W. Vertical lines indicate the start of the future time period.
0002548110.indd 105 8/6/2015 7:13:01 PM
7.3.6.Additional Drivers: Atmospheric CO2,
Precipitation Variability, and Emission Scenario
MC1 exhibited a minimal but noticeable sensitivity to
changes in the CO2 enrichment β‐factor. Compared to
control runs (with nitrogen limitation and full fire), simu-
lations with increased β resulted in higher future ecosys-
tem carbon stocks (average of 26.4 kg C m2 vs. 26.2 kg
Cm‐2 in control, averaged across all scenarios) and annual
burned areas (average of 3.9% vs. 3.6%), and slightly less
biomass combusted (average of 38.9 g C m2 vs. 39.8 g
Cm‐2). Decreased β runs displayed the opposite trends.
With an increasing β‐factor, the western and eastern forest
regions burned less whereas the Columbia Plateau burned
more: between a β‐factor of 0.0 and 0.6, increases in bio-
mass combusted between historical and future scenarios
decreased from 644% to 618% in the western Forests and
from 35% to 30% in the eastern Forests, and increased
from 13% to 20% in the Columbia Plateau, averaged
across all future scenarios. These regional differences
reflect the varying limitations on fire. With an increasing
β‐factor, water use efficiency and NPP increased. This
decreased fires in the forested regions because it allevi-
ated summer drought and decreased herbaceous coverage.
Yet a higher β‐factor increased fires in the Columbia
Plateau because it increased winter and spring fuel
production, particularly from grasses.
Precipitation variability is known to affect fire occur-
rence because of the interaction between fuel production
and drought. When calculated as the interannual varia-
bility relative to a scenario’s linear trend, the CSIRO
future projection displayed more (167 mm yr1), MIROC
displayed similar (149 mm yr1), and Hadley displayed
less (124 mm yr1) variability than the historical time
series (149 mm yr1). This did not appear to be a domi-
nant factor for levels of increased burning, however
(Table7.2). Nonetheless, when a 3‐year running average
filter was applied to CSIRO precipitation by month,
burned area decreased by 88% compared to historical.
Precipitation variability in future scenarios is likely a fac-
tor, but was outweighed by others such as overall drying
and productivity trends.
Although we chose to use projections driven by the
SRES A2 scenario, we also conducted simulations using
the B1 and A1B projections for all three GCMs with
nitrogen limitation and full fire. Comparatively speaking,
CO2 emissions are tracking the higher scenarios [Peters
etal., 2012; IPCC, 2013], and unfortunately A2 (or RCP
8.5) may be considered the most realistic. Except for
CSIRO, future ecosystem changes generally became more
severe with increasing emissions. For example, increases
in biomass consumed by fire in MIROC projections were
–0.91, 3.23, and 8.66 g C m2 yr1 with the B1, A1B, and A2
forcings, respectively, and in Hadley were 37.5, 44.82, and
45.88 g C m2 yr1. However, most of the domain’s mean
and spatial variation is captured within particular GCMs,
not emission scenarios. For example, in a two‐way analysis
of variance (ANOVA) for changes in mean temperature
and precipitation, within‐GCM variation yielded p values
of 0.015 and 0.0015, respectively, whereas within‐emission
scenario variation yielded p values of 0.63 and 0.014.
Spatially, the mean within‐GCM correlation coefficients
for changes in temperature, precipitation, and ecosystem
carbon were 0.98, 0.86, and 0.88, respectively. This is much
higher than the mean within‐emission scenario correlation
coefficients, which were 0.71, 0.15, and 0.47. These statistics
support the argument that, for this region, even when select-
ing only one simulation from an ensemble, signatures of
individual downscaled GCMs are stronger than particular
emission scenarios.
Using multiple scenarios and model configurations, this
study identified the major potential drivers of ecosystem
change in the US Pacific Northwest. Warming climates
tended to increase the land carbon sink, but fires that were
more difficult to suppress counteracted this to varying and
sometimes large degrees. Nitrogen limitation also tended
to suppress carbon sinks, and was a crucial factor for
grass‐tree competition in arid regions. The simulation of
nitrogen limitation is particularly uncertain. Future studies
focused on how nitrogen demand and limitation influences
tree‐grass competition, along with improved understand-
ing of nitrogen assimilation, turnover, and leaching, should
improve our ability to understand and model nutrient
influences. It should be noted that our simulations excluded
human land use, which has and will continue to be one of
the biggest drivers in populated areas and heavily managed
The sensitivity of maritime forests west of the Cascade
mountain range to fire and large‐scale forest type conver-
sion is perhaps the most surprising result. Although this
occurred in only one of three GCMs, it represented a
complete change of state toward drier‐adapted forests.
This was seen in the region at the height of the Holocene
thermal maximum, and is currently observed east of the
mountain range. In more recent work by Coops and
Waring [2011] and Hudiburg et al. [2013], these forests did
not display the same sensitivity to future drying as seen in
MC1 with the hot and dry Hadley climate. However, the
former did not explicitly include disturbance, which is the
sole driver of large carbon losses in the western forests in
our simulations. The latter study employed the Community
Land Model 4.0 (CLM) [Lawrence et al., 2011] and found
increases in both burned area and carbon stocks similar
to our simulations with CSIRO and MIROC. It is difficult
0002548110.indd 106 8/6/2015 7:13:01 PM
to know whether the lack of ecosystem resilience seen in
our simulations with Hadley in fact captures thresholds
that may be crossed in the future, as these conditions
could not be validated during the historical record.
The climate change effects shown here are likely applica-
ble for analogous systems worldwide. Warming‐induced
elevation increases in mountain vegetation communities
have been observed in many areas [Kelly and Goulden,
2008; Walther, 2004; Harsch et al., 2009], although this
may be limited by geomorphic/topographic characteris-
tics [Macias‐Fauria and Johnson, 2013]. The poleward
migration of vegetation seen in our study is being
observed [Parmesan and Yohe, 2003; Harsch et al., 2009]
and predicted [Malcolm et al., 2002; Gonzalez et al., 2010;
Koven, 2013] for a multitude of ecosystems worldwide.
The interacting effects of changing climate seasonality,
precipitation, fire, and human fire suppression seen in the
forested regions east of the Cascade Range likely apply
to other semiarid forests of the US mountain west and
pine forests in central and eastern Eurasia. Unless fire
weather severity increases to the point that crown fires
become substantially more common, these forests may
prove relatively resilient to climate change and continue
as carbon sinks.
This study also highlights the importance of drought
thresholds in historically mesic forests, which may have
the potential to cause a dramatic change in fire regime
and ecosystem dynamics. Boreal forests in North America
and parts of Eurasia (i.e., “dark taiga”) display similar
infrequent and stand‐replacing fires [Zackrisson, 1977;
Viereck, 1973; Gromtsev, 1996; Wirth, 2005], and may be
sensitive to drought thresholds in the coming century
[Soja et al., 2007; Balshi et al., 2009; Flannigan et al.,
2009]. Such thresholds could signify a change in state
toward more deciduous broadleaf boreal stands, which
generally store less carbon [Kasischke et al., 2000; Rogers
et al., 2014] but induce regional cooling through elevated
albedo [Randerson et al., 2006; Rogers et al., 2013].
MC1 has a number of advantages for this type of
analysis. Because it is largely mechanistic, the model is
capable of capturing the interacting and time‐varying
effects of temperature, precipitation, atmospheric CO2,
and evaporative demands on fire, soil hydrology, produc-
tivity, decomposition, lifeform competition, and vegeta-
tion communities. Statistical models provide an
alternative approach for projecting species distributions.
Yet these models are often extrapolated beyond the scope
of their historical climates used for fitting and are unable
to capture changing relationships in underlying mecha-
nisms. MC1 is also suitable for the regional and spatial
scale used in this study. Vegetation types simulated by the
model correspond to numerous ecoregion‐level commu-
nities that are easily separated by their functional com-
position, vegetation density, and bioclimatic habitats.
Simulating individual species introduces other large
potential errors, and MC1 was not designed to do this.
On the other hand, broad plant functional types used in
other land surface models [Woodward et al., 1998; Cox,
2001; Moorcroft et al., 2001; Sitch et al., 2003; Oleson
etal., 2013] are not aggregated into vegetation communi-
ties, and may not be as useful for predicting vegetation
distributions at this scale.
This research also highlighted limitations of the MC1
model, which can help inform future development. As in
other land surface and dynamic vegetation models, grid
cells in MC1 do not communicate with each other.
Although this formulation is computationally efficient
and arguably of little consequence at large spatial scales,
it becomes limiting at finer resolutions, particularly for
the simulation of hydrology and fire. The monthly
timestep of MC1 aids in its efficiency, but results in the
parameterization of processes that operate on a funda-
mentally shorter timescale. The most prominent of these
include photosynthesis, fire behavior, and precipitation
intensity and runoff.
In MC1, woody and herbaceous vegetation explicitly
compete for water, but delineation between forest types,
and between trees and shrubs, is layered on top of exist-
ing tree carbon pools using bioclimatic thresholds
[Bachelet et al., 2001]. Because shrubs and tree types do
not directly compete, carbon pools of an existing forest
type are automatically transferred to another if time‐
lagged climate thresholds are crossed. Some vegetation
changes may therefore be overly rapid, such as the large‐
scale conversions of maritime and subalpine forests in
our simulations. Future development would benefit from
shrub and forest type‐specific carbon pools and com-
petition. This would inevitably slow rates of vegetation
change, unless disturbances provided the catalysis
for invasion [Neilson, 1993]. Coupling with state‐and‐
transition models that provide detailed information on
landscape legacies, succession rules, and species‐specific
sensitivities to disturbance can increase realism in partic-
ular regions [e.g., Halofsky et al., 2013].
The representation of nitrogen in MC1 is arguable overly
simplistic for long‐term predictions. Because fixation is a
function of monthly evapotranspiration, nitrogen becomes
severely limiting in summer for (semi)arid systems in the
PNW, which factors into the seasonal representation of
production and future projections. In reality, however,
nitrogen is maintained within plant tissues throughout the
growing season, and nitrogen losses are reduced in more
nutrient‐limited environments [Chapin III et al., 2002].
Additionally, the storage and remobilization of nitrogen by
trees and perennial grasses allows them to decouple sea-
sonal growth from root uptake [Dickson, 1989; Millard and
Grelet, 2010]. There is also evidence that some conifers,
including pines, are able to fix nitrogen nonsymbiotically
0002548110.indd 107 8/6/2015 7:13:02 PM
[e.g., Bormann et al., 1993]. Nitrogen limitation, therefore,
should not be as responsive to monthly climate fluctua-
tions as it is in MC1. The model would benefit from more
explicit representation of nitrogen fixation, mineralization,
internal plant cycling, and atmospheric deposition.
The MC1 fire module is able to capture the effects of
monthly climate on various live and dead fuel classes and
the effects of both surface and crown fires. Fires impact live
and dead carbon pools and are found to be a dominant
driver of tree‐grass competition and carbon fluxes. There
are, however, a number of areas for improvement.
Incorporating submonthly climate, wind speed, and human
and natural ignitions could significantly improve the
temporal and spatial dynamics of fire occurrence and
behavior. Although more difficult to implement, cell‐to‐
cell communication is vital for the simulation of fire spread.
Algorithms from existing cellular automata models
[Hargrove et al., 2000; Yang et al., 2004; Yassemi et al.,
2008] could prove useful. Demographic information
(e.g., population density, land ownership) is important for
ignitions and suppression, and could improve the spatial
patterns of burning. We also found combustion factors to
be higher than observed, indicating that MC1 may be los-
ing too much biomass to fire for a given burned area.
Burned area in MC1 is heavily controlled by parame-
terized minimum and maximum fire return intervals
(FRIs), specific for individual vegetation types, that regu-
late burn fraction in relation to previous fires and the
assumed frequency. This prescription can strongly impact
ecosystem sensitivity to future climates. For example,
many grid cells in the western forests region did not expe-
rience fire in the historical period and were therefore sub-
ject to stand‐replacing fires in 100% of the grid cell once
a fire occurred in future projections, especially Hadley
(Figures7.7, 7.8, and 7.10). This likely caused some of
the large‐scale increases in burned area in the western
forests in all projections. In contrast, burn fractions of
future fires in arid systems east of the Cascade Range
were limited by historically frequent fire occurrences. In
a more realistic representation, FRIs would emerge from
mechanistically based fire ignition and spread functions.
In sum, MC1 offers an advanced modeling framework
suitable for investigating the response of ecosystems to
climate change, especially those that may be dependent
on fire, but contains many areas for future improvements.
Our work suggests that the PNW may continue to be a
carbon sink unless large conflagrations dramatically alter
the nature of existent forests. Although adaptation
options are still relatively limited, some general principles
are emerging that address multiple threats. For example,
higher levels of species diversity may increase vegetation
resilience to adverse changes in community type, carbon
stocks, and ecosystem services [Galik and Jackson, 2009].
Maintaining and establishing migration corridors may
sustain biodiversity and aid in inevitable species migra-
tions [von Hagen, 2009; Jantz et al., 2014]. Modeling exer-
cises such as this can potentially inform policies
surrounding carbon accounting and offsets, in which the
natural trajectory of ecosystem carbon under climate
changes needs to be accounted for.
We acknowledge the modeling groups, the Program for
Climate Model Diagnosis and Intercomparison (PCMDI)
and the WCRP’s Working Group on Coupled Modeling
(WGCM), for their roles in making available the WCRP
CMIP3 multimodel dataset. Support of this dataset is
provided by the Office of Science, US Department of
Agee, J. K. (1996), Fire Ecology of Pacific Northwest Forests,
Island Press, Washington, DC.
Agee, J. K. (1998), The landscape ecology of western forest fire
regimes, Northwest Sci., 72(17), 24–34.
Bachelet, D., B. R. Johnson, S. D. Bridgham, P. V. Dunn,
H. E. Anderson, and B. M. Rogers (2011), Climate change
impacts on western Pacific Northwest prairies and savannas,
Northwest Sci., 85(2), 411–429.
Bachelet, D., J. M Lenihan, C. Daly, R. P. Neilson, D. S. Ojima,
and W. J. Parton (2001), MC1: A Dynamic Vegetation Model
for Estimating the Distribution of Vegetation and Associated
Carbon, Nutrients, and Water Technical Documentation.
Version 1.0., General Technical Report, US Department of
Agriculture (USDA), Forest Service, Pacific Northwest
Research Station, Portland, OR.
Bailey, R. G. (1995), Description of the Ecoregions of the United
States, 2nd ed., USDA Forest Service, Washington, DC.
Balshi, M. S., A. D. McGuire, P. Duffy, M. Flannigan, J. Walsh,
and J. Melillo (2009), Assessing the response of area burned
to changing climate in western boreal North America using a
Multivariate Adaptive Regression Splines (MARS) approach,
Global Change Biol., 15(3), 578–600.
Barrett, S., D. Havlina, J. Jones, W. Hann, C. Frame, D.
Hamilton, K. Schon, T. Demeo, L. Hutter, and Menakis
(2010), Interagency Fire Regime Condition Class Guidebook,
Version 3.0, in Interagency Fire Regime Condition Class
(FRCC), USDA Forest Service; US Department of the
Interior; The Nature Conservancy. [online]; available from
Belsky, A. J. (1996), Viewpoint: Western juniper expansion: Is it a
threat to arid northwestern ecosystems? J. Range Management,
49(1), 53–59.
Blackard, J. A., M. V. Finco, E. H. Helmer, G. R. Holden,
M. L. Hoppus, D. M. Jacobs, A. J. Lister, G. G. Moisen,
M. D. Nelson, R. Riemann, B. Ruefenacht, D. Salajanu,
D. L. Weyermann, K. C. Winterberger, T. J. Brandeis,
R L. Czaplewski, R. E. McRoberts, P. L. Patterson, and
R. P. Tymcio (2008), Mapping US forest biomass using
0002548110.indd 108 8/6/2015 7:13:02 PM
nationwide forest inventory data and moderate resolution
information, Remote Sens. Environ., 112(4), 1658–1677.
Bormann, B., F. Bormann, W. Bowden, R. Pierce, S. Hamburg,
D. Wang, M. Snyder, C. Li, and R. Ingersoll (1993), Rapid
N2 fixation in pines, alder, and locust–evidence from the
Sandbox Ecosystem Study, Ecology, 74(2), 583–598.
Campbell, J., D. C. Donato, D. Azuma, and B. Law (2007),
Pyrogenic carbon emission from a large wildfire in Oregon,
United States, J. Geophys. Res Biogeosci., 112(G4), G04014,
Chapin III, F. S., M. C. Chapin, and P. A. Matson (2002), Principles
of Terrestrial Ecosystem Ecology, Springer‐Verlag, New York.
Climate Impacts Group (2004), Overview of Climate Change
Impacts in the U.S. Pacific Northwest, University of Washington,
Seattle, WA.
Coops, N. C., and R. H. Waring (2011), Estimating the vulnerabil-
ity of fifteen tree species under changing climate in Northwest
North America, Ecol. Model., 222(13), 2119–2129.
Cox, P. M. (2001), Description of the TRIFFID dynamic global
vegetation model, Technical Note 24, Hadley Centre,
Meteorological Office, Bracknell, UK.
Daly, C., M. Halbleib, J. I. Smith, W. P. Gibson, M. K. Doggett,
G. H. Taylor, J. Curtis, and P. P. Pasteris (2008),
Physiographically sensitive mapping of climatological tem-
perature and precipitation across the conterminous United
States, Int. J. Climatol., 28(15), 2031–2064.
Davis, M. A., K. J. Wrage, P. B. Reich, M. G. Tjoelker,
T. Schaeffer, and C. Muermann (1999), Survival, growth, and
photosynthesis of tree seedlings competing with herbaceous
vegetation along a water‐light‐nitrogen gradient, Plant Ecol.,
145(2), 341–350.
Dickson, R. (1989), Carbon and nitrogen allocation in trees,
Ann. Sci. Forestry, 46, S631–S647.
Diffenbaugh, N. S., L. C. Sloan, M. A. Snyder, J. L. Bell,
J. Kaplan, S. L. Shafer, and P. J. Bartlein (2003), Vegetation
sensitivity to global anthropogenic carbon dioxide emissions
in a topographically complex region, Global Biogeochem.
Cycles, 17(2), 1067.
Eidenshink, J., B. Schwind, K. Brewer, Z.‐L. Zhu, B. Quayle,
and S. Howard (2007), A project for monitoring trends in
burn severity, Fire Ecol., 3(1), 3–20.
Flannigan, M. D., B. J. Stocks, and B. M. Wotton (2000),
Climate change and forest fires, Sci. Total Environ., 262(3),
Flannigan, M., B. Stocks, M. Turetsky, and M. Wotton (2009),
Impacts of climate change on fire activity and fire manage-
ment in the circumboreal forest, Globa. Change Biol., 15(3),
Fowler, H. J., S. Blenkinsop, and C. Tebaldi (2007), Linking cli-
mate change modelling to impacts studies: Recent advances
in downscaling techniques for hydrological modelling, Int. J.
Climatol., 27(12), 1547–1578.
Galik, C. S., and R. B. Jackson (2009), Risks to forest carbon
offset projects in a changing climate, Forest. Ecol.
Management, 257(11), 2209–2216. doi:10.1016/j.
Gavin, D. G., D. J. Hallett, F. S. Hu, K. P. Lertzman,
S. J. Prichard, K. J. Brown, J. A. Lynch, P. Bartlein, and
D. L. Peterson (2007), Forest fire and climate change in
western North America: Insights from sediment charcoal
records, Frontiers Ecol. Environ., 5(9), 499–506.
Gerber, S., L. O. Hedin, S. G. Keel, S. W. Pacala, and
E. Shevliakova (2013), Land use change and nitrogen feed-
backs constrain the trajectory of the land carbon sink,
Geophys. Res. Lett., 40(19), 5218–5222.
Gonzalez, P., R. P. Neilson, J. M. Lenihan, and R. J. Drapek
(2010), Global patterns in the vulnerability of ecosystems to
vegetation shifts due to climate change, Global Ecol. Biogeogr.,
19(6), 755–768.
Gordon, H. B. (2002), The CSIRO Mk3 Climate System Model,
CSIRO Atmospheric Research Technical Paper 60, CSIRO
Atmospheric Research.
Graham, R. T., A. E. Harvey, T. B. Jain, and J. R. Tonn (1999),
The Effects of Thinning and Similar Stand Treatments on Fire
Behavior in Western Forests, General Technical Report PNW‐
GTR‐463, USDA Forest Service, Pacific Northwest Research
Gromtsev, A. N. (1996), Retrospective analysis of natural fire
regimes in landscapes of eastern Fennoscandia and problems
in their anthropogenic transformation, Forest Sci., 48,
Halofsky, J. E., M. A. Hemstrom, D. R. Conklin, J. S. Halofsky,
B. K. Kerns, and D. Bachelet (2013), Assessing potential cli-
mate change effects on vegetation using a linked model
approach, Ecol. Model., 266, 131–143 (available online at
S0304380013003281 (Accessed 9 October 2014).
Hamann, A., and T. Wang (2006), Potential effects of climate
change on ecosystem and tree species distribution in British
Columbia, Ecology, 87(11), 2773–2786.
Hargrove, W. W., R. H. Gardner, M. G. Turner, W. H. Romme,
and D. G. Despain (2000), Simulating fire patterns in
heterogeneous landscapes, Ecol. Model., 135(2–3),
Harr, R. (1977), Water flux in soil and subsoil on a steep for-
ested slope, J. Hydrol., 33(1−2), 37–58.
Harsch, M. A., P. E. Hulme, M. S. McGlone, and R. P. Duncan
(2009), Are treelines advancing? A global meta‐analysis of
treeline response to climate warming, Ecol. Lett., 12(10),
Hasumi, H., and S. Emori (eds.) (2004), K‐1 Coupled GCM
(MIROC) Description, K‐1 Model Developers Technical
Report 1.
Hessl, A. E., D. McKenzie, and R. Schellhaas (2004), Drought
and Pacific Decadal Oscillation linked to fire occurrence in
the inland Pacific Northwest, Ecol. Applications, 14(2),
Hirota, M., M. Holmgren, E. H. Van Nes, and M. Scheffer
(2011), Global resilience of tropical forest and savanna to
critical transitions, Science, 334(6053), 232–235.
Hudiburg, T., B. Law, D. P. Turner, J. Campbell, D. C. Donato,
and M. Duane (2009), Carbon dynamics of Oregon and
Northern California forests and potential land‐based carbon
storage, Ecol. Appl., 19(1), 163–180.
Hudiburg, T. W., S. Luyssaert, P. E. Thornton, and B. E. Law
(2013), Interactive effects of environmental change and man-
agement strategies on regional forest carbon emissions,
Environ. Sci. Technol., 47(22), 13132–13140.
0002548110.indd 109 8/6/2015 7:13:02 PM
IPCC (2007), Summary for policymakers, in S. D. Solomon,
D. Qin, M. Manning, Z. Chen, K. B. Averyt, M. Marquis,
M. Tignor, and H. L. Miller (eds.), Climate Change 2007: The
Physical Science Basis. Contribution of Working Group I to
the Fourth Assessment Report of the Intergovernmental Panel
on Climate Change, Cambridge University Press, Cambridge,
UK and New York.
IPCC (2013), Summary for policymakers, in T. F. Stocker,
D. Qin, G.‐K. Plattner, M. Tignor, S. K. Allen, J. Boschung,
A. Nauels, Y. Xia, V. Bex, and P. M. Midgley (eds.), Climate
Change 2013: The Physical Science Basis. Contribution of
Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change, Cambridge
University Press, New York.
Jackson, S. T., and J. T. Overpeck (2000), Responses of plant
populations and communities to environmental changes of
the late Quaternary, Paleobiology, 26(sp4), 194–220.
Jain, A., X. Yang, H. Kheshgi, A. D. McGuire, W. Post, and
D. Kicklighter (2009), Nitrogen attenuation of terrestrial car-
bon cycle response to global environmental factors, Glob.
Biogeochem. Cycles, 23, GB4028, doi:10.1029/2009GB003519.
Jantz, P., S. Goetz, and N. Laporte (2014), Carbon stock corri-
dors to mitigate climate change and promote biodiversity in
the tropics, Nature Climate Change, 4(2), 138–142.
Johns, T. C., J. M. Gregory, W. J. Ingram, C. E. Johnson,
A. Jones, J. A. Lowe, J. F. B. Mitchell, D. L. Roberts, D. M. H.
Sexton, D. S. Stevenson, S. F. B. Tett, and M. J. Woodage
(2003), Anthropogenic climate change for 1860 to 2100 simu-
lated with the HadCM3 model under updated emissions sce-
narios, Climate Dynamics, 20(6), 583–612.
Joubert, D. F., G. N. Smit, and M. T. Hoffman (2012), The role of
fire in preventing transitions from a grass dominated state to a
bush thickened state in arid savannas, J. Arid Environ., 87, 1–7.
Kasischke, E. S., K. P. O’Neill, N. H. F. French, and L. L.
Bourgeau‐Chavez (2000), Controls on patterns of biomass
burning in Alaskan boreal forests, in E. S. Kasischke and
B. J. Stocks (eds.), Fire, Climate Change, and Carbon Cycling
in the Boreal Forest, Springer, New York, pp. 173–196,
Keane, R. E., J. K. Agee, P. Fule, J. E. Keeley, C. Key,
S. G. Kitchen, R. Miller, and L. A. Schulte (2008), Ecological
effects of large fires on US landscapes: Benefit or catastro-
phe? Int. J. Wildland Fire, 17(6), 696–712.
Kelly, A. E., and M. L. Goulden (2008), Rapid shifts in plant
distribution with recent climate change, Proc. Natl. Acad.
Sci., 105(33), 11823–11826.
Kern, J. (1995), Geographic patterns of soil water‐holding
capacity in the contiguous United States, Soil Sci. Soc. Am.J.,
59(4), 1126–1133.
Kerns, B. K., B. J. Naylor, M. Buonopane, C. G. Parks, and
B.Rogers (2009), Modeling tamarisk (tamarix spp.) habitat
and climate change effects in the northwestern United States,
Invasive Plant Sci. Management, 2(3), 200–215.
Kittel, T. G. F., N. A. Rosenbloom, T. H. Painter, D. S. Schimel,
J. M. Melillo, Y. D. Pan, D. W. Kicklighter, A. D. McGuire,
R. P. Neilson, J. Chaney, D. S. Ojima, R. McKeown,
W. J. Parton, W. M. Pulliam, I. C. Prentice, A. Haxeltine,
S. W. Running, L. L. Pierce, R. R. Nemani, E. R. Hunt,
T. M. Smith, B. Rizzo, and F. I. Woodward (1995), The
VEMAP integrated database for modelling United States
ecosystem/vegetation sensitivity to climate change, J.
Biogeogr., 22(4–5), 857–862.
Koven, C. D. (2013), Boreal carbon loss due to poleward shift in
low‐carbon ecosystems, Nature Geosci., 6(6), 452–456.
Kraaij, T., and D. Ward (2006), Effects of rain, nitrogen, fire
and grazing on tree recruitment and early survival in bush‐
encroached savanna, South Africa, Plant Ecol., 186(2),
Kuchler, A. (1975), Potential Natural Vegetation of the United
States, 2nd ed., Map 1:3,168,000. American Geographical
Society, New York.
LANDFIRE Rapid Assessment (2007), Rapid assessment ref-
erence condition models, in LANDFIRE, USDA Forest
Service, Rocky Mountain Research Station, Fire Sciences
Lab; US Geological Survey; The Nature Conservancy.
Law, B. E., O. J. Sun, J. Campbell, S. Van Tuyl, and P. E.
Thornton (2003), Changes in carbon storage and fluxes in a
chronosequence of ponderosa pine, Global Change Biol.,
9(4), 510–524.
Lawrence, D. M., K. W. Oleson, M. G. Flanner, P. E. Thornton,
S. C. Swenson, P. J. Lawrence, X. Zeng, Z.‐L. Yang, S. Levis,
K. Sakaguchi, G. B. Bonan, and A. G. Slater (2011),
Parameterization improvements and functional and struc-
tural advances inVersion 4 of the Community Land Model,
J. Advances Model. Earth Systems, 3, M03001,
LeBauer, D. S., and K. K. Treseder (2008), Nitrogen limitation
of net primary productivity in terrestrial ecosystems is glob-
ally distributed, Ecology, 89(2), 371–379.
Littell, J. S., E. E. Oneil, D. Mckenzie, J. A. Hicke, J. A. Lutz,
R. A. Norheim, and M. M. Elsner (2009), Forest ecosys-
tems, disturbance, and climatic change in Washington
State, USA, in Washington Climate Change Impacts
Assessment: Evaluating Washington’s Future in a Changing
Climate, The Climate Impacts Group, University of
Washington, Seattle, WA.
Littell, J. S., E. E. Oneil, D. McKenzie, J. A. Hicke, J. A. Lutz,
R. A. Norheim, and M. M. Elsner (2010), Forest ecosystems,
disturbance, and climatic change in Washington State, USA,
Clim. Change, 102(1–2), 129–158.
Macias‐Fauria, M., and E. A. Johnson (2013), Warming‐
induced upslope advance of subalpine forest is severely lim-
ited by geomorphic processes, Proc. Natl. Acad. Sci., 110(20),
Malcolm, J. R., A. Markham, R. P. Neilson, and M. Garaci
(2002), Estimated migration rates under scenarios of global
climate change, J. Biogeogr., 29(7), 835–849.
McKenzie, D., Z. Gedalof, D. L. Peterson, and P. Mote (2004),
Climatic change, wildfire, and conservation, Conservation
Biol., 18(4), 890–902.
McKenzie, D., and C. B. Halpern (1999), Modeling the distribu-
tions of shrub species in Pacific northwest forests, Forest
Ecol. Management, 114(2–3), 293–307 (available online at
S0378112798003600; accessed 10/9/14).
Meehl, G. A., C. Covey, T. Delworth, M. Latif, B. McAvaney,
J. F. B. Mitchell, R. J. Stouffer, and K. E. Taylor (2007), The
WCRP CMIP3 multimodel dataset–a new era in climate
change research, Bull. Am. Meteorol. Soc., 88(9), 1383.
0002548110.indd 110 8/6/2015 7:13:02 PM
Melilo, J. M., T. C. Richmond, and G. W. Yohe (2014),
Climate Change Impacts in the United States: The Third
National Climate Assessment, US Global Change Research
Millar, C. I., R. Neilson, D. Batchelet, R. Drapek, and J. Lenihan
(2006), Climate change at multiple scales, in Forests, carbon,
and climate change: a synthesis of science findings, pp. 31–62,
Oregon Forest Resources Institute, Portland, OR, USA.
Millard, P., and G.‐A. Grelet (2010), Nitrogen storage and
remobilization by trees: ecophysiological relevance in a
changing world, Tree Physiol., 30(9), 1083–1095.
Moorcroft, P. R., G. C. Hurtt, and S. W. Pacala (2001), A
method for scaling vegetation dynamics: The ecosystem
demography model (ED), Ecol. Monographs, 71(4),
Mote, P., E. Salathe, V. Duliere, and E. Jump (2008), Scenarios
of Future Climate for the Pacific Northwest, Climate Impacts
Group, University of Washington, Seattle.
Mote, P. W., and E. P. Salathe (2010), Future climate in the
Pacific Northwest, Climat. Change, 102(1–2), 29.
Nakicenovic, N., J. Alcamo, G. Davis, B. de Vries, J. Fenhann,
S. Gaffin, K. Gregory, A. AGrubler, T. Y. Jung, and T. Kram
(2000), in R. Swart (ed.), Special Report on Emissions Scenarios:
Specia Report of Working Group III of the Intergovernmental
Panel on Climate Change, Cambridge University Press,
New York and Cambridge, UK.
Neilson, R. (1993), Transient ecotone response to climatic‐
change–some conceptual and modeling approaches, Ecol.
Applications, 3(3), 385–395.
Neilson, R. P. (2004), Impacts of climate change on Pacific
Northwest terrestrial ecosystems [online], available at http://
Norby, R. J., E. H. DeLucia, B. Gielen, C. Calfapietra, C. P.
Giardina, J. S. King, J. Ledford, H. R. McCarthy, D. J. P. Moore,
R. Ceulemans, P. De Angelis, A. C. Finzi, D. F. Karnosky,
M. E. Kubiske, M. Lukac, K. S. Pregitzer, G. E. Scarascia‐
Mugnozza, W. H. Schlesinger, and R. Oren (2005), Forest
response to elevated CO2 is conserved across a broad range of
productivity, Proc. Natl. Acad. Sci., 102(50), 18052–18056.
Norby, R. J., J. M. Warren, C. M. Iversen, B. E. Medlyn, and
R. E. McMurtrie (2010), CO2 enhancement of forest produc-
tivity constrained by limited nitrogen availability, Proc. Natl.
Acad. Sci., 107(45), 19368–19373.
Oleson, K. W., D. M. Lawrence, G. B. Bonan, B. Drewniak,
M. Huang, C. D. Koven, S. Levis, F. Li, W. J. Riley,
Z. M. Subin, S. C. Swenson, and P. E. Thornton (2013),
Technical Description of Version 4.5 of the Community Land
Model (CLM), NCAR Earth System Laboratory, Climate
and Global Dynamics Division, Boulder, CO.
Oregon Department of Energy (2010), Oregon Greenhouse Gas
Inventory – 1990−2007 [online], available at http://www.
Inventory_1990‐2007.htm (accessed 12/28/10).
Oregon Department of Forestry (2009), Hope
Page [online], available at
STATE_FORESTS/FRP/ (accessed 3/19/09).
Parmesan, C., and G. Yohe (2003), A globally coherent finger-
print of climate change impacts across natural systems,
Nature, 421(6918), 37–42.
Peters, G. P., G. Marland, C. Le Quere, T. Boden, J. G. Canadell,
and M. R. Raupach (2012), CORRESPONDENCE: Rapid
growth in CO2 emissions after the 2008‐2009 global financial
crisis, Nature Climate Change, 2(1), 2–4.
Pyne, S. J. (1982), Fire in America: A Cultural History of
Wildland and Rural Fire, Princeton University Press,
Princeton, NJ.
Randerson, J. T., H. Liu, M. G. Flanner, S. D. Chambers, Y. Jin,
P. G. Hess, G. Pfister, M. C. Mack, K. K. Treseder, L. R.
Welp, F. S. Chapin, J. W. Harden, M. L. Goulden, E. Lyons,
J. C. Neff, E. a. G. Schuur, and C. S. Zender (2006), The
impact of boreal forest fire on climate warming, Science,
314(5802), 1130–1132.
Raymond, C. L., and D. McKenzie (2012), Carbon dynamics of
forests in Washington, USA: 21st century projections based
on climate‐driven changes in fire regimes, Ecol. Applications,
22(5), 1589–1611.
Rogers, B. M. (2009), Potential Impacts of Climate Change on
Vegetation Distributions, Carbon Stocks, and Fire Regimes in the
US Pacific Northwest, master’s thesis, Oregon State University,
Corvallis, OR [online], available at http://scholarsarchive.
Rogers, B. M., R. P. Neilson, R. Drapek, J. M. Lenihan,
J.R. Wells, D. Bachelet, and B. E. Law (2011), Impacts of
climate change on fire regimes and carbon stocks of the U.S.
Pacific Northwest, J. Geophys. Research, 116, G03037,
Rogers, B. M., J. T. Randerson, and G. B. Bonan (2013), High‐
latitude cooling associated with landscape changes from
North American boreal forest fires, Biogeosciences, 10(2),
Rogers, B. M., S. Veraverbeke, G. Azzari, C. I. Czimczik,
S. R. Holden, G. O. Mouteva, F. Sedano, K. K. Treseder,
and J. T. Randerson (2014), Quantifying fire‐wide carbon
emissions in interior Alaska using field measurements and
Landsat imagery, J. Geophys. Res. Biogeosci., 119,
Running, S. W., R. R. Nemani, F. A. Heinsch, M. S. Zhao,
M. Reeves, and H. Hashimoto (2004), A continuous satellite‐
derived measure of global terrestrial primary production,
Bioscience, 54(6), 547–560.
Scheller, R. M., and D. J. Mladenoff (2007), An ecological clas-
sification of forest landscape simulation models:Tools and
strategies for understanding broad‐scale forested ecosystems,
Landscape Ecol., 22(4), 491–505.
Schoennagel, T., T. T. Veblen, W. H. Romme, J. S. Sibold, and
E. R. Cook (2005), Enso and pdo variability affect drought‐
induced fire occurrence in Rocky Mountain subalpine forests,
Ecol. Applicaions, 15(6), 2000–2014.
Shafer, S. L., P. J. Bartlein, and R. S. Thompson (2001), Potential
changes in the distributions of western North America tree
and shrub taxa under future climate scenarios, Ecosystems,
4(3), 200–215.
Siemann, E., and W. E. Rogers (2003), Changes in light and
nitrogen availability under pioneer trees may indirectly facili-
tate tree invasions of grasslands, J. Ecol., 91(6), 923–931.
Sitch, S., B. Smith, I. C. Prentice, A. Arneth, A. Bondeau,
W. Cramer, J. O. Kaplan, S. Levis, W. Lucht, M. T. Sykes,
K. Thonicke, and S. Venevsky (2003), Evaluation of ecosystem
0002548110.indd 111 8/6/2015 7:13:02 PM
dynamics, plant geography and terrestrial carbon cycling in
the LPJ dynamic global vegetation model, Global Change
Biol., 9(2), 161–185.
Smith, B., D. Wårlind, A. Arneth, T. Hickler, P. Leadley,
J. Siltberg, and S. Zaehle (2014), Implications of incorporat-
ing N cycling and N limitations on primary production in an
individual‐based dynamic vegetation model, Biogeosciences,
11(7), 2027–2054.
Smithwick, E. A. H., M. E. Harmon, S. M. Remillard, S. A. Acker,
and J. F. Franklin (2002), Potential upper bounds of carbon
stores in forests of the Pacific Northwest, Ecol. Applications,
12(5), 1303–1317.
Smithwick, E. A. H., M. G. Ryan, D. M. Kashian, W. H. Romme,
D. B. Tinker, and M. G. Turner (2009), Modeling the effects
of fire and climate change on carbon and nitrogen storage in
lodgepole pine (Pinus contorta) stands, Global Change Biol.,
15(3), 535–548.
Soja, A. J., N. M. Tchebakova, N. H. F. French, M. D.
Flannigan, H. H. Shugart, B. J. Stocks, A. I. Sukhinin,
E. I. Parfenova, F. S. Chapin, and P. W. Stackhouse
(2007), Climate‐induced boreal forest change: Predictions
versus current observations, Global Planet. Change, 56(3–4),
Sokolov, A. P., D. W. Kicklighter, J. M. Melillo, B. S. Felzer,
C. A. Schlosser, and T. W. Cronin (2008), Consequences of
considering carbon‐nitrogen interactions on the feedbacks
between climate and the terrestrial carbon cycle, J. Climate,
21(15), 3776–3796.
Spies, T. A., and J. F. Franklin (1991), The structure of natural
young, mature, and old‐growth Douglas‐fir forests in Oregon
and Washington, in L. F. Ruggiero, K. B. Aubry, A. B. Carey,
and M. H. Huff (eds.), Wildlife and Vegetation of Unmanaged
Douglas‐Fir Forests, USDA Forest Service Pacific Northwest
Research Station, Portland, OR, pp. 91–111.
Spracklen, D. V., L. J. Mickley, J. A. Logan, R. C. Hudman,
R. Yevich, M. D. Flannigan, and A. L. Westerling (2009),
Impacts of climate change from 2000 to 2050 on wildfire
activity and carbonaceous aerosol concentrations in the west-
ern United States, J Geophys Research, 114, 1998–2002.
Staver, A. C., S. Archibald, and S. A. Levin (2011), The global
extent and determinants of savanna and forest as alternative
biome states, Science, 334(6053), 230–232.
Stephens, S. L., J. J. Moghaddas, C. Edminster, C. E. Fiedler,
S. Haase, M. Harrington, J. E. Keeley, E. E. Knapp,
J.D. McIver, K. Metlen, C. N. Skinner, and A. Youngblood
(2009), Fire treatment effects on vegetation structure, fuels,
and potential fire severity in western US forests, Ecol.
Applications, 19(2), 305–320.
Sun, O. J., J. Campbell, B. E. Law, and V. Wolf (2004), Dynamics
of carbon stocks in soils and detritus across chronosequences
of different forest types in the Pacific Northwest, USA,
Global Change Biol., 10(9), 1470–1481.
Thompson, R. S., S. W. Hostetler, P. J. Bartlein, and K. H.
Anderson (1998), A strategy for assessing potential future
changes in climate, hydrology, and vegetation in the western
United States, US Geol. Survey Circular, (1153), 1–20.
Thornton, P. E., J.‐F. Lamarque, N. A. Rosenbloom, and N. M.
Mahowald (2007), Influence of carbon‐nitrogen cycle cou-
pling on land model response to CO2 fertilization and climate
variability, Global Biogeochem. Cycles, 21(4), GB4018,
Van Mantgem, P. J., N. L. Stephenson, J. C. Byrne,
L. D. Daniels, J. F. Franklin, P. Z. Fulé, M. E. Harmon,
A. J. Larson, J. M. Smith, A. H. Taylor, and T. T. Veblen
(2009), Widespread increase of tree mortality rates in the
western United States, Science, 323(5913), 521–524.
Van Tuyl, S., B. E. Law, D. P. Turner, and A. I. Gitelman (2005),
Variability in net primary production and carbon storage in
biomass across Oregon forests – an assessment integrating
data from forest inventories, intensive sites, and remote sens-
ing, Forest Ecol. Management, 209(3), 273–291.
Veblen, T. T., T. Kitzberger, and J. Donnegan (2000), Climatic
and human influences on fire regimes in ponderosa pine for-
ests in the Colorado Front Range, Ecol. Applications, 10(4),
Viereck, L. A. (1973), Wildfire in the taiga of Alaska, Quaternary
Research, 3(3), 465–495.
Villalba, R., T. Veblen, and J. Ogden (1994), Climatic influences
on the growth of sub‐alpine trees in the Colorado Front
Range, Ecology, 75(5), 1450–1462.
Von Hagen, B. (2009), Unexplored potentials of Northwest for-
ests, in T. A. Spies and S. L. Duncan (eds.), Old Growth in a
New World, Island Press, Washington, DC, pp. 286–299
Walther, G. R. (2004), Plants in a warmer world, Perspectives
Plant Ecol. Evoution. Systematics, 6(3), 169–185.
Washington State Department of Natural Resources (DNR)
(2009), Washington State Timber Harvest [online], availa-
ble at
Pages/washington_state_timber_harvest.aspx (accessed
Waterman‐Hoey, S., and G. Nothstein (2007), Washington’s
Greenhouse Gas Emissions: Sources and Trends, Washington
State Department of Community, Trade & Economic
Development, Energy Policy Devision [online], available at
Westerling, A. L., A. Gershunov, T. J. Brown, D. R. Cayan, and
M. D. Dettinger (2003), Climate and wildfire in the western
United States, Bulletin Am. Meteorol. Soc., 84(5), 595.
Westerling, A. L., H. G. Hidalgo, D. R. Cayan, and T. W.
Swetnam (2006), Warming and earlier spring increase
western US forest wildfire activity, Science, 313(5789),
Whitlock, C., and M. A. Knox (2002), Prehistoric burning in
the Pacific Northwest: human versus climatic influences, in
Fire, Native Peoples, and the Natural Landscape , Island Press,
Washington, DC. pp. 195–231.
Whitlock, C., S. L. Shafer, and J. Marlon (2003), The role of
climate and vegetation change in shaping past and future fire
regimes in the northwestern US and the implications for eco-
system management, Forest Ecol. Management, 178(1–2),
Wirth, C. (2005), Fire regime and tree diversity in boreal for-
ests: Implications for the carbon cycle, in D. M. Scherer‐
Lorenzen, P. D. C. Körner, and P. D. E.‐D. Schulze (eds.),
Forest Diversity and Function, Springer, Berlin and
Heidelberg. pp. 309–344,
0002548110.indd 112 8/6/2015 7:13:02 PM
Woodward, F. I., M. R. Lomas, and R. A. Betts (1998), Vegetation‐
climate feedbacks in a greenhouse world, Philos. Tranacts.
Royal Soc. London Series B‐‐Biol. Sci., 353(1365), 29–38.
Yang, J., H. S. He, and E. J. Gustafson (2004), A hierarchical
fire frequency model to simulate temporal patterns of fire
regimes in LANDIS, Ecol. Model., 180(1), 119–133.
Yassemi, S., S. Dragicevic, and M. Schmidt (2008), Design and
implementation of an integrated GIS‐based cellular autom-
ata model to characterize forest fire behaviour, Ecol. Model.,
210(1‐2), 71–84.
Zackrisson, O. (1977), Influence of forest fires on North
Swedish boreal forest, Oikos, 29(1), 22–32.
0002548110.indd 113 8/6/2015 7:13:02 PM
0002548110.indd 114 8/6/2015 7:13:02 PM
... 7. air quality impacts tied to wildfires, transport of pollutants from Asia, wood smoke, diesel, and agricultural dust (Dalton et al. 2013;Rogers et al. 2015). ...
... Forests already limited by water availability (mostly east of the Cascades) are expected to experience longer, more severe waterlimitation under reduced summer and early fall precipitation, resulting in decreased tree growth. Wildfire activity in the Columbia River basin is projected to increase in response to drier summer conditions that reduce the moisture of soil and fuels (Dalton et al. 2013;Rogers et al. 2015). It is estimated that the regional area burned per year will increase by roughly 900 square miles from 1970-2000 averages by the 2040s (Dalton et al. 2013). ...
... At the Columbia Basin scale in the 2030s water supply is projected to be sufficient to meet demands. However, at smaller watershed scales during some times of the year, for example in summer in in the Yakima basin, water supply is projected to be insufficient to meet demands ( Complex interactions between natural resource management policies and practices, regional development and climate change require ongoing scientific investigation (Dalton et al. 2013;Rogers et al. 2015). A wide range of Northwest US management decisions related to water rights, water storage infrastructure, nutrient management, cropping systems and tillage, rangeland management, timber harvesting, and wildfire management have potential to be informed by scientific understanding of the intersecting effects of regional decision-making and future climate change impacts (Rasmussen et al. 1998;Dalton et al. 2013;Rogers et al. 2015). ...
Full-text available
Scientific research is key to understanding complex environmental systems and informing decisions about natural resource management in the context of climate change. Environmental science research is, however, often conducted without active stakeholder engagement, and the result is typically development of new knowledge that does not directly serve the needs of individuals, industries and organizations that make decisions about environmental policy and resource management. Recent decades have seen rapidly expanding efforts to conduct environmental science research that directly informs government policies and private decision-makers’ management plans, yet significant barriers remain in the pursuit of usable climate science. Strategies for effective collaboration among researchers and stakeholders, who have diverse needs and expertise, are not well developed. Metrics are needed for evaluating approaches to usable climate science production. This research advances understanding of how to foster effective stakeholder engagement for usable climate science outputs, focusing on regional environmental modeling efforts based at universities. By tracking researchers’ perceptions about stakeholder engagement over the course of a 5-year project, assessing stakeholders’ information needs and perceptions of research and identifying characteristics of effective boundary-spanning organizations, this work suggests strategies for evaluating the effectiveness of efforts to produce usable climate science and identifies strategies for academic scientists to develop their capacity to bridge boundaries between research and decision-making.
Full-text available
Fires in the boreal forests of North America are generally stand-replacing, killing the majority of trees and initiating succession that may last over a century. Functional variation during succession can affect local surface energy budgets and, potentially, regional climate. Burn area across Alaska and Canada has increased in the last few decades and is projected to be substantially higher by the end of the 21st century because of a warmer climate with longer growing seasons. Here we simulated the changes in forest composition due to altered burn area using a stochastic model of fire occurrence, historical fire data from national inventories, and succession trajectories derived from remote sensing. When coupled to an Earth system model, younger vegetation from increased burning cooled the high-latitude atmosphere, primarily in the winter and spring, with noticeable feedbacks from the ocean and sea ice. Results from multiple scenarios suggest that a doubling of burn area would result in surface cooling of 0.23 ± 0.09 °C and 0.43 ± 0.12 °C for winter–spring and February–April time periods, respectively. This could provide a negative feedback to high-latitude terrestrial warming during winter on the order of 4–6% for a doubling, and 14–23% for a quadrupling, of burn area. Further work is needed to integrate all the climate drivers from boreal forest fires, including aerosols and greenhouse gasses.
Technical Report
Full-text available
In the West, thinning and partial cuttings are being considered for treating millions of forested acres that are overstocked and prone to wildfire. The objectives of these treatments include tree growth redistribution, tree species regulation, timber harvest, wildlife habitat improvement, and wildfire-hazard reduction. Depending on the forest type and its structure, thinning has both positive and negative impacts on crown fire potential. Crown bulk density, surface fuel, and crown base height are primary stand characteristics that determine crown fire potential. Thinning from below, free thinning, and reserve tree shelterwoods have the greatest opportunity for reducing the risk of crown fire behavior. Selection thinning and crown thinning that maintain multiple crown layers, along with individual tree selection systems, will not reduce the risk of crown fires except in the driest ponderosa pine (Pinus ponderosa Dougl. ex Laws.) forests. Moreover, unless the surface fuels created by using these treatments are themselves treated, intense surface wildfire may result, likely negating positive effects of reducing crown fire potential. No single thinning approach can be applied to reduce the risk of wildfires in the multiple forest types of the West. The best general approach for managing wildfire damage seems to be managing tree density and species composition with well-designed silvicultural systems at a landscape scale that includes a mix of thinning, surface fuel treatments, and prescribed fire with proactive treatment in areas with high risk to wildfire.
Full-text available
We investigated variation in carbon stock in soils and detritus (forest floor and woody debris) in chronosequences that represent the range of forest types in the US Pacific Northwest. Stands range in age from <13 to >600 years. Soil carbon, to a depth of 100 cm, was highest in coastal Sitka spruce/western hemlock forests (36±10 kg C m−2) and lowest in semiarid ponderosa pine forests (7±10 kg C m−2). Forests distributed across the Cascade Mountains had intermediate values between 10 and 25 kg C m−2. Soil carbon stocks were best described as a linear function of net primary productivity (r2=0.52), annual precipitation (r2=0.51), and a power function of forest floor mean residence time (r2=0.67). The highest rates of soil and detritus carbon turnover were recorded on mesic sites of Douglas-fir/western hemlock forests in the Cascade Mountains with lower rates in wetter and drier habitats, similar to the pattern of site productivity. The relative contribution of soil and detritus carbon to total ecosystem carbon decreased as a negative exponential function of stand age to a value of ∼35% between 150 and 200 years across the forest types. These age-dependent trends in the portioning of carbon between biomass and necromass were not different among forest types. Model estimates of soil carbon storage based on decomposition of legacy carbon and carbon accumulation following stand-replacing disturbance showed that soil carbon storage reached an asymptote between 150 and 200 years, which has significant implications to modeling carbon dynamics of the temperate coniferous forests following a stand-replacing disturbance.
Historical and geological data indicate that significant changes can occur in the Earth's climate on time scales ranging from years to millennia. In addition to natural climatic change, climatic changes may occur in the near future due to increased concentrations of carbon dioxide and other trace gases in the atmosphere that are the result of human activities. International research efforts using atmospheric general circulation models (AGCM's) to assess potential climatic conditions under atmospheric carbon dioxide concentrations of twice the pre-industrial level (a "2xCO 2" atmosphere) conclude that climate would warm on a global basis. However, it is difficult to assess how the projected warmer climatic conditions would be distributed on a regional scale and what the effects of such warming would be on the landscape, especially for temperate mountainous regions such as the Western United States. In this report, we present a strategy to assess the regional sensitivity to global climatic change. The strategy makes use of a hierarchy of models ranging from an AGCM, to a regional climate model, to landscape-scale process models of hydrology and vegetation. A 2xCO 2 global climate simulation conducted with the National Center for Atmospheric Research (NCAR) GENESIS AGCM on a grid of approximately 4.5° of latitude by 7.5° of longitude was used to drive the NCAR regional climate model (RegCM) over the Western United States on a grid of 60 km by 60 km. The output from the RegCM is used directly (for hydrologic models) or interpolated onto a 15-km grid (for vegetation models) to quantify possible future environmental conditions on a spatial scale relevant to policy makers and land managers.
Placing an upper bound to carbon (C) storage in forest ecosystems helps to constrain predictions on the amount of C that forest management strategies could sequester and the degree to which natural and anthropogenic disturbances change C storage. The potential, upper bound to C storage is difficult to approximate in the field because it requires studying old-growth forests, of which few remain. In this paper, we put an upper bound (or limit) on C storage in the Pacific Northwest (PNW) of the United States using field data from old-growth forests, which are near steady-state conditions. Specifically, the goals of this study were: (1) to approximate the upper bounds of C storage in the PNW by estimating total ecosystem carbon (TEC) stores of 43 old-growth forest stands in five distinct biogeoclimatic provinces and (2) to compare these TEC storage estimates with those from other biomes, globally. Finally, we suggest that the upper bounds of C storage in forests of the PNW are higher than current estimates of C stores, presumably due to a combination of natural and anthropogenic disturbances, which indicates a potentially substantial and economically significant role of C sequestration in the region. Results showed that coastal Oregon stands stored, on average, 1127 Mg C/ha, which was the highest for the study area, while stands in eastern Oregon stored the least, 195 Mg C/ha. In general, coastal Oregon stands stored 307 Mg C/ha more than coastal Washington stands. Similarly, the Oregon Cascades stands stored 75 Mg C/ha more, on average, than the Washington Cascades stands. A simple, area-weighted average TEC storage to I m soil depth (TEC,,,) for the PNW was 671 Mg C/ha. When soil was included only to 50 cm (TEC(50)), the area-weighted average was 640 Mg C/ha. Subtracting estimates of current forest C storage from the potential, upper bound of C storage in this study, a maximum of 338 Mg C/ha (TEC(100)) could be stored in PNW forests in addition to current stores.
The problem of forest fires has long been extensively dealt with in the scientific literature (Melekhov 1948; Sukachev 1975; Vakurov 1975; Spurr and Barnes 1984; and many others). However, it remains a pressing issue, especially in connection with the hot debate which arose between ecologists and economists, and which has even been discussed in the foreign news media in recent years (Abel 1985; Anonymus 1988). Nowadays, most forest ecologists consider fire to be one of the main ecological factors responsible for the structure and dynamics of the biotic components of taiga landscape, primarily natural forests. An analysis of the literature shows, however, that systematized data on fires were obtained either by dating fire scars on trees or by studying the archives. In the former case, a retrospective analysis of fire regime is usually restricted to 250–300 years (the age of the older tree generation); in the latter, a general analysis is made when only big fires are considered, without specifying areas and habitats. However, to assess the effect of fires on natural forests, detailed data covering a period of thousands of years should be collected with regard for area and the types of growth conditions under which forest-forming species develop. As well, patterns of the pyrogenic factor were discussed, with few exceptions (Furyaev and Kireyev 1979), regardless of forest landscape structure. It is impossible, however, to agree with the above authors, who argue that the problem of forest fires cannot be resolved for an individual biogeocenosis because a landscape basis is needed.