Content uploaded by Cameron W. Wobus
Author content
All content in this area was uploaded by Cameron W. Wobus on May 05, 2017
Content may be subject to copyright.
Contents lists available at ScienceDirect
Global Environmental Change
journal homepage: www.elsevier.com/locate/gloenvcha
Projected climate change impacts on skiing and snowmobiling: A case study
of the United States
Cameron Wobus
a
, Eric E. Small
b
, Heather Hosterman
a
, David Mills
a,⁎
, Justin Stein
a
,
Matthew Rissing
a
, Russell Jones
a
, Michael Duckworth
a
, Ronald Hall
a
, Michael Kolian
c
,
Jared Creason
c
, Jeremy Martinich
c
a
Abt Associates, 1881 Ninth Street, Suite 201, Boulder, CO, USA
b
University of Colorado Boulder, Geological Sciences, Boulder, CO, USA
c
U.S. Environmental Protection Agency, Climate Change Division, Washington, DC, USA
ARTICLE INFO
Keywords:
Climate change
Skiing
Snowmobiling
Snowmaking
Adaptation
ABSTRACT
We use a physically-based water and energy balance model to simulate natural snow accumulation at 247 winter
recreation locations across the continental United States. We combine this model with projections of
snowmaking conditions to determine downhill skiing, cross-country skiing, and snowmobiling season lengths
under baseline and future climates, using data from five climate models and two emissions scenarios. Projected
season lengths are combined with baseline estimates of winter recreation activity, entrance fee information, and
potential changes in population to monetize impacts to the selected winter recreation activity categories for the
years 2050 and 2090. Our results identify changes in winter recreation season lengths across the United States
that vary by location, recreational activity type, and climate scenario. However, virtually all locations are
projected to see reductions in winter recreation season lengths, exceeding 50% by 2050 and 80% in 2090 for
some downhill skiing locations. We estimate these season length changes could result in millions to tens of
millions of foregone recreational visits annually by 2050, with an annual monetized impact of hundreds of
millions of dollars. Comparing results from the alternative emissions scenarios shows that limiting global
greenhouse gas emissions could both delay and substantially reduce adverse impacts to the winter recreation
industry.
1. Introduction
Projected climate change through the 21st century will generate
warmer temperatures and changes in precipitation that are expected to
decrease the duration and extent of natural snow cover across the
northern hemisphere (e.g., Dyer and Mote, 2006; Brown and Mote,
2009; Diffenbaugh et al., 2013). A number of studies have examined
how climate change could influence seasonal snowpack in the western
United States (e.g., Barnett et al., 2005; Mote et al., 2005; Pierce and
Cayan, 2013), with the aim of understanding impacts on water
resources. However, the geographic extent and economic impacts of a
changing snowpack are likely to extend well beyond the western United
States (e.g., Hayhoe et al., 2007; Campbell et al., 2010). In particular,
large components of the winter recreation industry will face challenges
without reliable access to snow. This could threaten tens of millions of
current annual recreational visits and have important repercussions in
areas where winter recreation is central to economic activity.
This study follows a series of reports that quantify and monetize the
potential for climate change impacts to a number of sectors in the
United States (e.g., Walsh et al., 2014; U.S. EPA, 2015a,b). In particular,
U.S. EPA (2015a) quantifies and monetizes impacts from climate
change under a range of scenarios in terms of anticipated impacts to
human health and labor, electricity, forestry and agriculture, water
resources, ecosystems, and the built infrastructure. Here we model
potential changes in snowpack at sites across the United States, and
calculate the effects of changing season lengths on the number of
recreational visits and direct revenue associated with entrance fees. Our
study expands on previous impacts work related to winter recreation by
combining the geographic breadth of previous studies (e.g., Burakowski
and Magnusson, 2012) with the detail that has historically been applied
only to site- or regionally-specific studies (e.g., Burakowski et al., 2008;
Lazar and Williams, 2008, 2010; Scott et al., 2008; Dawson and Scott,
2013). We consider a range of future climate scenarios by examining
outputs from five global climate models (GCMs), two representative
http://dx.doi.org/10.1016/j.gloenvcha.2017.04.006
Received 30 November 2016; Received in revised form 14 April 2017; Accepted 17 April 2017
⁎
Corresponding author.
E-mail address: Dave_Mills@abtassoc.com (D. Mills).
Global Environmental Change 45 (2017) 1–14
0959-3780/ © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
MARK
concentration pathways (RCPs), and both mid- and late 21 st century
time periods. As a result, this study helps expand the type and number
of sectors with national-scale estimates of quantified and monetized
impacts from climate change.
The foundation of our method is a water and energy balance model
that accounts for simplified, but site-specific climatic and topographic
characteristics to project natural snow accumulation at 247 locations in
the continental United States (CONUS). For downhill skiing and
snowboarding, we combine results from this natural snow model with
projections of how resorts’abilities to make artificial snow will change
in the future. For Nordic (i.e., cross-country) skiing and for snowmobil-
ing, we use the model outputs without consideration for snowmaking,
since these activities are typically more reliant on natural snow. For all
three winter recreation activities, we estimate season length during a
baseline period, and then project the impact of climate change on
season lengths through the 21st century. We summarize these impacts
as anticipated changes in future season lengths, levels of recreational
activity, and entry-based expenditures. Although our results are specific
to the United States, our methods could be easily adapted and extended
to address other international regions (e.g., the Alps, Scandinavia, New
Zealand) where snow-dependent winter recreation is an important
cultural or economic activity.
2. Methods
To produce nationally informative results on climate change’s
potential impacts on winter recreation, we considered available sources
of recreational and meteorological data to select representative sites
across the CONUS. We then used a well-vetted snow model to project
natural snow accumulation and melt at each site. We used modeled
future temperatures to project changing snowmaking conditions at each
downhill skiing location. Snow modeling and snowmaking projections
were repeated for baseline and future climate conditions. We then
quantified and monetized the resulting changes in snow conditions by
integrating available baseline recreational and population data with a
series of reasonable, but simplifying, assumptions about future recrea-
tional participation and expenditures.
2.1. Recreational site selection
The first step in our work was to identify reliable recreational
participation data for multiple locations, ideally with multiple observa-
tions over time, which could be paired with observed and projected
climate data. To link climate projections to specific winter recreation
locations, we downloaded publicly available ski area information (i.e.,
polylines) to create footprints for winter recreation sites in the CONUS
(OpenSnowMap, 2016). We then merged this information with ski area
site names and locations from the National Oceanic and Atmospheric
Administration geospatial data portal (NOAA, 2016). We compared ski
area polygons to aerial photography to verify ski area names, ensure a
match of the area footprints, and record the type of skiing for each area
(i.e., downhill, cross-country, or both) −making any edits as necessary
in the review process. Our final sample included 247 ski locations,
distributed across the 6 National Ski Areas Association (NSAA) regions
and across private and public lands, as shown in Fig. 1.
2.2. Natural snow accumulation and snowmaking model
2.2.1. Utah energy balance model
We chose the Utah Energy Balance (UEB) model (Tarboton and
Luce, 1996), a physically-based model that simulates the water and
energy balance of a seasonal snowpack. We used several criteria in the
model selection process: (1) high computational efficiency, as the study
design required over 300,000 years of model simulations; (2) minimal
parameters, given the broad range of site conditions that exist across
the CONUS; and (3) acceptable performance. UEB is a single-layer snow
model, and thus is more efficient and has fewer parameters than more
complex, multi-layer models (e.g., Flerchinger et al., 1996). Even
though the UEB model is relatively simple, its performance is on par
with more complex models, as determined through snow model
intercomparison efforts (e.g., Rutter et al., 2009; Förster et al., 2014).
As discussed below, the selection of a meteorological dataset is even
more important than the model, given that uncertainty from forcing
data often exceeds that from errors due to model physics or parameters
(Raleigh et al., 2015). We used the current version of the model
(UEBveg; Mahat and Tarboton, 2012, 2014) to simulate natural snow
accumulation and snowmelt at two elevations, representing the bottom
and top of a ski area, for our selected locations. The UEB model tracks
three state variables: snow-water-equivalent (SWE), internal energy of
the snowpack, and snow surface age, the latter which affects surface
albedo.
For implementation, we set the vegetation fractional cover input to
zero to simulate the open areas that predominate on wide ski area
slopes. The shortwave radiation input is calculated based on date/time,
latitude, and slope angle and azimuth. The diurnal cycle of the surface
energy balance, and thus melt, is represented because we used hourly
meteorological forcing. However, we exported only daily SWE from the
UEB model.
2.2.2. Meteorological forcing and topographic adjustments
We used hourly North American Land Data Assimilation System
(NLDAS-2) meteorological forcing data (Xia et al., 2012) to drive the
UEB snow model. NLDAS-2 data were selected because they provide
physically-consistent forcing fields for the entire United States. NLDAS-
2 forcing data are provided on a 1/8th degree (∼12 km) grid for the
interval from January 1, 1979 through the present. No other multi-
decadal, high-spatial resolution, continental-scale datasets exist, and
thus NLDAS-2 data have been used in hundreds of snow and hydrology
studies (e.g., Sultana et al., 2014; Fu et al., 2015; Raleigh et al., 2015).
We used data for the following NLDAS-2 variables: air temperature,
specific humidity, wind speed, downward shortwave radiation, and
downward longwave radiation. The four non-precipitation variables
were generated on a 32-km grid at 3 hourly intervals as a part of the
National Centers for Environmental Prediction’s North American Re-
gional Reanalysis, and then interpolated to the NLDAS-2 grid (Cosgrove
et al., 2003). The NLDAS-2 precipitation data are based on daily
weather gauge values gridded to 1/8 °, utilizing information from the
Parameter-Elevation Relationships on Independent Slopes Model
(PRISM; Daly et al., 1994) and disaggregated through time using radar
analyses when available. The UEB snow model has previously been
driven with North American Land Data Assimilation System (NLDAS)-2
meteorological forcing data (as in this study), yielding a reasonable
time series of SWE as observed at individual California Department of
Water Resources Snow Telemetry (SNOTEL) sites (Sultana et al., 2014).
The accuracy of the NLDAS-2 input data affects the SWE simulated
by the UEB model. It is challenging to model SWE in mountain ranges:
precipitation, temperature, and radiation vary dramatically on length
scales from meters to kilometers, resulting in extreme spatial variability
in SWE (e.g., Clark et al., 2011). Even though NLDAS-2 data are
relatively high resolution (1/8th degree), it is impossible for this
dataset to represent the extreme range of conditions that exist within
individual grid cells. In order to represent topographic effects at scales
finer than the NLDAS grid cells, we applied site-specific adjustments in
temperature and precipitation as a function of elevation within each ski
area boundary. To do this, we extracted monthly climate normals for
1981–2010 from PRISM, and regressed both temperature and precipita-
tion against elevation for each cardinal direction within each ski area
boundary. We used these regression results to calculate an average
temperature and precipitation lapse rate for each ski area. Using these
calculated lapse rates, we adjusted the baseline NLDAS forcing to
estimate precipitation and temperature at the bottom and top of each
ski area.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
2
The topographic adjustment accounts for differences in elevation
between the ski area and the NLDAS model grid cell in which the ski
area exists. However, this adjustment does not account for the low bias
in NLDAS precipitation that exists in mountainous regions (e.g., Pan
et al., 2003; Argus et al., 2014; Fu et al., 2015). In two similar studies
that used NLDAS data as inputs for snow models, a comparison to
SNOTEL data was used to adjust the input precipitation to account for
this bias (Pan et al., 2003; Sultana et al., 2014). We take the same
approach here. We compared both precipitation and temperature from
NLDAS-2 to that measured at individual SNOTEL sites. We identified
SNOTEL sites within NLDAS-2 grid cells containing one of our target
skiing locations and that were within 100 m of the specified NLDAS-2
elevation. Only 27 SNOTEL stations met this criterion. Other SNOTEL
sites were too different in elevation, so the comparison would have
been obfuscated by topographic gradients in precipitation and tem-
perature. NLDAS-2 precipitation is 10% lower and temperature is 0.5 °C
warmer than observed at SNOTEL stations, averaged across these 27
sites. We accounted for both this 10% under-prediction relative to
SNOTEL and the documented 20% undercatch of precipitation at
SNOTEL gauges (e.g., Serreze et al., 2001) by multiplying the NLDAS
precipitation by 1.3 prior to use in the UEB model. We also subtracted
0.5 °C from the NLDAS temperature to remove the bias relative to
SNOTEL. The precipitation adjustment factor is smaller than that used
by Pan et al. (2003) and similar to that from Sultana et al. (2014).
The 27 SNOTEL sites used for this meteorological forcing adjust-
ment are all in the western United States. Homogenous monitoring
networks comparable to SNOTEL do not exist in the Midwest or East,
where meteorological observations are not made in environments and
at elevations similar to ski areas. Although SNOTEL stations are in the
western United States, we show below that the UEB model driven by
this bias-corrected NLDAS-2 forcing provides a reasonable ski season
length throughout the United States (see Section 3.1).The slope and
aspect of ski slopes each play an important role in controlling seasonal
snow accumulation and melt, due to the change in net solar radiation
per square meter depending on the incident angle of sunlight relative to
the surface. To account for these topographic influences, we also
calculated the mean slope and modal aspect of each ski area using a
90 m digital elevation model (USGS, 2008). Our modeling includes
snow accumulation and melt at the average slope and aspect for the top
and bottom of each skiing location.
2.2.3. Modeling hours suitable for snowmaking
Snowmaking is already a critical operational feature for many
downhill skiing locations, and helps to ensure that an area is open for
the Christmas/New Year holidays (e.g., Dawson and Scott, 2013).
Typically, downhill ski areas require between 400 and 500 h of suitable
snowmaking conditions before they can open (Robin Smith, TechnoAl-
pin, personal communication, October 13, 2016), which requires a wet
bulb temperature of 28 °F or less (e.g., Scott et al., 2003; Robin Smith,
TechnoAlpin, personal communication, October 13, 2016). We calcu-
lated cumulative hours of wet bulb temperature below 28 °F beginning
on October 1 of each year as a proxy for snowmaking potential at each
location. We calculated wet bulb temperature from NLDAS-2 humidity
and air temperature (e.g., Stull, 2011) at the lowermost elevation for
each location, under the assumption that most resorts need to cover
their lowest slopes to open. In each simulation year, we recorded the
Fig. 1. Map of skiing locations, by NSAA reporting region, to project climate change impacts on natural snow accumulation and potential snowmaking conditions.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
3
date when each location reached 450 h of cumulative snowmaking
conditions. We calculated an average date to reach 450 snowmaking
hours for each location from each of the individual years in the 30-year
climate dataset.
2.3. Climate change scenarios
Computational and resource constraints required that we select a
subset of GCMs from the full suite of the fifth Coupled Model
Fig. 2. Comparisons of UEB model SWE estimates with independent measures of season length from SNODAS, averaged over water years 2004–2010. Small dots represent individual
sites; large dots are regional averages with 1σerror bars.
Fig. 3. Baseline season lengths for cross-country skiing and snowmobiling.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
4
Intercomparison Project (CMIP5; Taylor et al., 2012) models. We chose
five GCMs (CCSM4, GISS-E2-R, CanESM2, HadGEM2-ES, and MIROC5)
with the intent of ensuring that (1) the subset captured a large range of
variability in climate outcomes observed across the entire CMIP5
ensemble, and (2) the models were independent and broadly used by
the scientific community. For each GCM, we chose two RCPs that
captured a range of plausible emissions futures. The RCPs, originally
developed for the Intergovernmental Panel on Climate Change’s Fifth
Assessment Report, are identified by their approximate total radiative
forcing in the year 2100, relative to 1750: 8.5 W/m
2
(RCP8.5) and 4.5
W/m
2
(RCP4.5). RCP8.5 implies a future with continued high emissions
growth with limited efforts to reduce greenhouse gases (GHGs),
whereas RCP4.5 represents a global GHG mitigation scenario.
To provide localized climate projections and to bias correct the
projections to improve consistency with the historical period, we used
the LOCA dataset (Pierce et al., 2014, 2015; USBR et al., 2016). The
LOCA downscaled dataset provides daily minimum and maximum
temperatures (T
min
and T
max
), and daily precipitation values at 1/16°
resolution from 2006 to 2100. For each climate scenario, we calculated
an average daily change factor for temperature and precipitation at
each grid cell by comparing 20 years of LOCA projections centered on
2050 and 2090 to a historical 1/16° gridded dataset from the period
1986–2005 (Livneh et al., 2015). We calculated these daily change
factors as a spatial average of nine 1/16° LOCA grid cells (3 × 3
window) surrounding each location.
We calculated hourly temperature change factors based on model-
projected changes in T
min
and T
max
by assuming these temperatures
occur at midnight and noon, respectively, and interpolating between
T
min
and T
max
values over the course of each day. These hourly changes
were then added to the baseline NLDAS-2 temperature time series. For
precipitation, we used the GCM outputs to calculate a multiplier to
apply to the hourly NLDAS-2 precipitation time series. In some cases,
the LOCA-modeled precipitation led to unrealistically high change
factors. To eliminate these outliers, we first discarded values that
exceeded the 90th percentile of all change factors for each station. We
then calculated daily change factors as a 31-day moving average ratio
of this filtered time series, and applied them to the NLDAS baseline.
Additional details regarding our GCM selection process, an overview of
the selected models, and processes for producing the relevant tempera-
ture and precipitation measures are provided in Supplementary in-
formation file #1.
2.4. Winter recreation activity, season length, and monetization approach
We combined the physical modeling, described above, with avail-
able recreational visit and entrance fee data to advance the under-
standing of how climate change affects winter recreation. First, we
created baseline recreational activity levels for downhill skiing, cross-
country skiing, and snowmobiling using NSAA and National Visitor Use
Monitoring (NVUM) program data (NSAA and RRC, 2016; USFS, 2016).
NSAA provides a comprehensive, annual dataset that uses survey data
from approximately 195 ski resorts to report downhill ski visits by state
(consistent with the NSAA survey, we use the term “visit”throughout to
represent one person’s activity for a single day of a given type of
recreation). To produce the downhill skiing baseline, we constructed a
decadal average of skier visits by state using visit data from the
2006–2007 season through the 2015–2016 season (NSAA and RRC,
2016). The United States Forest Service (USFS) uses onsite surveys to
Fig. 4. Average percent change in annual cross-country skiing and snowmobiling season lengths across GCMs. A) Results for RCP4.5 in 2050. B) Results for RCP4.5 in 2090. C) Results for
RCP8.5 in 2050. D) Results for RCP8.5 in 2090.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
5
estimate the level of recreation use on national forests through the
NVUM program. NVUM has a time-series of recreation visits and
expenditure data for cross-country skiing and snowmobiling, as well
as downhill skiing and other recreational activities (USFS, 2016). To
produce the cross-country and snowmobiling baseline, we used the
average of visits from two rounds of survey data (Round 1 from 2005 to
2009 and Round 2 from 2010 to 2014). For national forests that span
multiple states, we allocated baseline visits according to the distribu-
tion of the forest area in the respective states. The NSAA data are
representative of visits in 2011 and the NVUM data are representative
of visits in 2010.
We applied changes in season length, as modeled at each of the 247
locations in the CONUS, to estimate changes in winter recreation visits.
For downhill skiing season lengths, we incorporated snowmaking,
which is consistent with Scott et al. (2003, 2008).Wedefined the start
of each season as the earlier of 10‐cm SWE or 450 h of snowmaking at
the base of each location, and the end of each season as the last date
with 10 cm SWE at the upper elevation of each location. For Nordic
skiing and snowmobiling, the use of snowmaking is relatively uncom-
mon (Reese Brown, Cross Country Ski Areas Association, personal
communication, July 7, 2016) and, therefore, we did not incorporate
snowmaking into our analysis. We used the direct outputs from the UEB
model to simulate snowpack, and determined season length as the
difference between the first and last dates with 10 cm SWE at the base
elevation of each location.
Winter recreation use is strongly correlated to season length; this
correlation is particularly strong at the regional level for downhill
skiing (correlation coefficients for season length versus total annual
visits range from 0.60 in the Rocky Mountains to 0.99 in the Southeast
NSAA regions). To project potential impacts of climate change on
winter recreation activities, we assumed recreational visits will change
in direct proportion to the length of the associated recreational season.
This assumption is consistent with NSAA data that show the percent
visitation by month is approximately even throughout the ski season,
particularly in the Rocky Mountain region (NSAA and RRC, 2016).
Finally, we monetized the impacts of climate change on winter
recreation using ticket prices for downhill skiing and entry fees at
national forests for cross-country skiing and snowmobiling to reflect the
price of access to each recreational opportunity. For downhill skiing, we
used the average of reported adult ticket prices, by region, for the
2014–2015 through 2015–2016 seasons (NSAA and RRC, 2016).
Regional ticket prices ranged from approximately $59 in the Midwest
region to $127 in the Rocky Mountain region (NSAA and RRC, 2016; all
prices in year 2015 dollars). For comparability, we used entry fees from
the NVUM data to monetize projected changes in cross-country skiing
and snowmobiling visits in national forests. Entry fees are one
component of USFS’s NVUM trip spending profiles and include site
admission, parking, and recreation use fees (Stynes and White, 2005;
Dan McCollum, U.S. Forest Service, personal communication, October
12, 2016). The NVUM data ask visitors surveyed about trip-related
spending. We converted trip spending to visitor spending by dividing
the trip entry fees by the average number of people per trip. We use
these entry fee measures to monetize recreational impact given the
scale and goals of the analysis, and to avoid a number of more complex
economic issues discussed in greater detail in the Conclusions section.
For all three winter recreational activities, we first evaluated
impacts holding population constant over time to isolate the impact
of climate change. A second set of impact estimates were then
calculated to account for projected population growth. We used the
Integrated Climate and Land Use Scenarios (ICLUS) v2.0 (U.S. EPA,
2016) county-level, all-age population projections, to project 2050 and
2090 populations by state. To estimate future visitation, we multiplied
Fig. 5. Baseline and future cross-country skiing and snowmobile season length for Bretton Woods resort in New Hampshire. Boxplots represent the distribution of 30 annual UEB model
simulations for baseline conditions and each of the future scenarios specified. A) Results for RCP4.5 in 2050. B) Results for RCP4.5 in 2090. C) Results for RCP8.5 in 2050. D) Results for
RCP8.5 in 2090.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
6
the ICLUS state population projections by the average annual number of
visits in each state per resident calculated for the baseline period, and
then scaled the resulting product by the estimated proportional change
in season length for the activity in that state. The change in season
length used in this calculation represents an average of change for all
sites in each state, by activity type.
3. Results
Below, we summarize results for each NSAA region by RCP and time
period and provide detailed baseline and future snow modeling results
for representative sites across the United States. Detailed annual season
length results for all locations for the 21 simulations (one baseline and
20 future projections) are included in Supplementary information file
#2. We also provide detailed baseline and future winter recreational
activity levels under climate change scenarios. Detailed state-level
changes in winter recreation visits and dollars are included in
Supplementary information file #3.
3.1. UEB model validation
We used season length derived from the Snow Data Assimilation
System (SNODAS; Barrett, 2003) to validate our baseline simulations,
by examining how UEB-simulated season length varies from region to
region across the United States. SNODAS provides daily SWE at ∼1km
resolution nationwide, based on a multi-layer snow model that is forced
to be consistent with remotely-sensed observations of snow extent.
SNODAS data are available beginning in 2003, so we compared season
length (duration of SWE > 10 cm) at each of the 247 sites averaged
over the 7 overlapping seasons between SNODAS and our baseline
simulations (water years 2004–2010). The correspondence between
UEB and SNODAS season length estimates at the ski area scale is
reasonable given the coarseness of the NLDAS-2 forcing: including all
outliers, the r
2
is approximately 0.6 and there is little bias (Fig. 2). The
only clear difference between UEB and SNODAS is in the Pacific
Southwest, where the UEB season length is longer than SNODAS by
∼30 days.
We also compared UEB SWE with SNOTEL data at the 27 sites that
are within 1 km of ski areas. Although differences at individual SNOTEL
sites are in some cases substantial, the average season length from UEB
at these 27 sites is nearly the same as from SNOTEL (UEB season length
for these 27 sites is 112 +/−30 days; SNOTEL is 125 +/−40 days).
3.2. Baseline and projected season lengths
3.2.1. Cross-country skiing and snowmobiling
For cross-country skiing and snowmobiling, season lengths vary
from less than 1 week for some sites in the Northeast and upper
Midwest to more than 24 weeks for many sites in the western United
States (Fig. 3). These season length projections assume no adaptation
from snowmaking.
Average annual changes in cross-country skiing and snowmobiling
season lengths across the GCMs range from small increases in some
locations, to declines of more than 80% under the RCP4.5 scenarios in
2050 (Fig. 4A). In general, the most significant reductions in season
length occur in the upper Midwest and the Northeast, and the smallest
Fig. 6. Average date to reach 450 cumulative hours with conditions suitable for snowmaking (Twb < 28 °F), based on 30 years of baseline NLDAS-2 forcing data.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
7
reductions occur at locations in the central Rockies and Sierras. The few
locations with increases in season length are generally in arid regions of
the Southwest and parts of the upper Midwest. These increases in
season length are driven by projected increases in precipitation, which
offset projected increases in temperature by mid-century.
The general regional pattern of changes in cross-country skiing and
snowmobiling season length persists across GCMs into the late century
under the higher emissions scenario (i.e., RCP8.5 in 2090; Fig. 4D).
However, under this scenario a much larger fraction of the modeled
locations are projected to see average annual reductions from their
baseline season length of > 80% compared to the RCP4.5 estimates in
2090 (Fig. 4B).
Beneath these regional trends, there is substantial variability across
the GCMs. Fig. 5 illustrates projected changes in cross-country skiing
and snowmobiling season length at the Bretton Woods resort in New
Hampshire for each climate model/RCP combination. For this resort,
the average projected decrease in season length ranges from approxi-
mately 65% by 2050 under RCP4.5, to more than 90% by 2090 under
RCP8.5. While inter-annual variability remains high in 2050 under
some of the models (e.g., CCSM4, GISS), this variability effectively
collapses in the relatively unconstrained RCP 8.5 emissions scenarios,
particularly late in the century.
3.2.2. Downhill skiing and snowboarding
The length of the winter season for downhill skiing reflects the
combined influence of early season temperatures, which modulate
resorts’ability to make snow; and natural precipitation and tempera-
ture throughout the ski season, which control the water and energy
balance that drive the natural snowpack. Under baseline conditions,
locations with the highest base elevations (e.g., those in the central
Rocky Mountains) typically reach the 450 cumulative hours of snow-
making threshold by late October, whereas this snowmaking threshold
is not reached until late January or later in some locations in the
Southeast (Fig. 6). Including snowmaking, baseline season lengths for
alpine skiing range from just 1–2 weeks in some resorts in the
Southeast, to more than 28 weeks in the highest elevations of the
Rocky Mountains and Sierras (Fig. 7; Supplementary information File
#2).
Under climate change scenarios, the average date to reach the
cumulative 450 h snowmaking threshold increases by approximately
10–20 days by mid-century under RCP4.5, and by 30–70 days by late
century under RCP8.5 (Fig. 8). Winter recreation impacts are regionally
variable, with the largest delays in reaching this snowmaking threshold
occurring in the Pacific Northwest and smaller delays in the Rocky
Mountain region.
For most downhill skiing locations, opening prior to the Christmas
and New Year’s holidays is critical to remaining profitable and staying
in business (e.g., Dawson and Scott, 2013; Robin Smith, TechnoAlpin,
personal communication, October 13, 2016). While approximately 70%
of modeled downhill skiing sites can reach 450 h of snowmaking by
December 15 under baseline climate conditions, this percentage
declines markedly under each of the future scenarios (Fig. 9). By
2050 this percentage is reduced by nearly half under both RCPs. In
2090 the contrast is sharper, as only 23% of locations would meet the
December 15 date under the RCP4.5 scenarios and only 11% of
locations under the RCP8.5 scenarios.
Nationally, changes in projected downhill skiing season lengths
range from slight increases at a few areas (10 areas and 6 areas,
respectively, for RCP4.5 and RCP8.5 in 2050; and 4 areas for RCP4.5 in
2090) to declines of more than 80% under RCP8.5 in 2050 for some
Fig. 7. Modeled baseline season lengths for downhill skiing, including snowmaking.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
8
locations (Fig. 10). As with the cross-country skiing and snowmobiling
season length results, the general spatial patterns of changes in season
length are largely preserved under RCP8.5 in 2090, but are amplified
relative to the 2050 results. Specifically, the projected changes in
season length are most dramatic in the Northeast and upper Midwest,
and are less dramatic in the higher elevations in the Rockies and
Sierras. Further, in 2090 under RCP8.5, no areas are projected to have
an increased season and the smallest projected reduction in season
length is 15%.
As with the cross-country skiing and snowmobiling season results,
there is also considerable inter-model variability in climate change
results for downhill skiing season lengths. At Aspen Mountain, for
example, average season lengths decrease by 10–20 days under RCP4.5
in 2050 and by 25–75 days under RCP8.5 in 2090 (Fig. 11).
While Figs. 8 and 9 highlighted potential delays in opening dates
relative to the critical Christmas and New Year’s holidays, climate
change generally has a larger impact on closing dates than opening
dates across the combinations of RCPs and future years (Fig. 12). In the
most extreme reductions (RCP8.5 projections for 2090), the median
closing date is more than a month earlier than the baseline, moving
from early April to the end of February. In contrast, the largest shift in
the start date from the baseline involves several weeks from the
beginning to the end of December. Although we did not incorporate
detailed data on user visits by month, we do know that revenue from
spring break is important for some resorts, particularly in the Rockies.
Thus, scaling user visits and entry fee revenue linearly with changes in
season length (see Section 3.3) is likely to be conservative.
3.3. Quantifying and monetizing potential changes in future winter
recreation
3.3.1. Baseline winter recreation activity levels
Nationally, we estimated a baseline winter recreational activity
level of approximately 56.0 million downhill skiing visits from the
available NSAA data, with an additional 3.6 million cross-country ski
visits and 2.8 million snowmobile visits from the available NVUM data
(Table 1). Using calculated regional average adult weekend ticket
Fig. 8. Lost season days due to additional time required to reach 450 h of potential snowmaking time, by region. A) Results for RCP4.5 in 2050. B) Results for RCP4.5 in 2090. C) Results
for RCP8.5 in 2050. D) Results for RCP8.5 in 2090. (MW = Midwest, NE = Northeast, PNW = Pacific Northwest, PSW = Pacific Southwest, RM = Rocky Mountain, SE = Southeast; see
Fig. 1 for regions).
Fig. 9. Percentage of modeled areas able to reach 450 h of snowmaking before December
15.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
9
prices from the NSAA data, we monetized baseline downhill skiing at
$5.4 billion; using the conceptually-equivalent entry fees results in
$32.4 million for cross-country skiing and $12.6 million for snowmo-
biling.
3.3.2. Monetized impact holding populations constant
Holding population constant at baseline levels, we project climate
change would reduce national downhill skiing visits to 35.4 million
visits under RCP4.5 and 19.8 million under RCP8.5 by 2090; this is a
decrease of 20.6 million and 36.3 million visits from the baseline,
respectively (Table 1). Holding population constant, national cross-
country skiing visits would be projected to decrease to 2.7 million
under RCP4.5 and 1.5 million under RCP8.5 by 2090, and national
snowmobiling visits would decrease from approximately 2.8 million in
2010 to 1.9 million under RCP4.5 and 1.0 million under RCP8.5 by
2090.
3.3.3. Monetized impact including population growth
In our approach, population growth increases projected winter
recreation visits. As a result, population growth dampens the projected
adverse impacts of climate change on the winter recreation industry.
Nationally, downhill skiing visits decrease slightly after adjusting for
changes in climate and population to 52.8 million under RCP4.5 and
30.6 million under RCP8.5 by 2090; a decrease in 3.2 million and 25.4
million visits from baseline, respectively (Table 1). Under RCP4.5,
cross-country skiing visits increase slightly in 2050 and 2090, and
snowmobiling visits increase slightly in 2090 (Table 1). However, for
RCP8.5, which reflects a higher emissions scenario, the shortened
seasons overwhelm the increase in visits driven by population growth,
resulting in an overall decrease in recreational visits in both 2050 and
2090 (Table 1).
To clearly demonstrate the offsetting impact of population growth
on these recreational visit results, we aggregated state-level results for
downhill skiing to the NSAA regions and compared projected regional
visits with and without population growth for 2050 and 2090 (Fig. 13).
The effect of population growth on winter recreation visits is most
clearly seen in the results for the Rocky Mountain and Pacific Southwest
regions. Under the RCP4.5 scenarios, downhill skiing visits in these
regions increase in 2050 and 2090 when we account for the combined
impacts of climate change and population growth. However, when we
hold population growth constant and account for only the impacts of
climate change, downhill skiing visits decline in both regions in the
RCP4.5 scenario. In the RCP8.5 scenario, this impact is still observable
in these regions. In all cases, projected visitation at each time period is
larger when population change is included.
As shown in Table 1, holding the population constant at baseline
values, the projected impacts of climate change alone could result in the
loss of tens of millions of winter recreation visits with an undiscounted
annual impact measured in the billions of dollars. Integrating the
impacts of projected climate change and population growth complicates
these results, as seen in Fig. 13. In general, our assumption that winter
recreation visits will increase with population, all else equal, mitigates
but does not fully offset the projected adverse impacts of climate
change at a national level. Regional trends in projected population
growth are also critical. Specifically, the combination of relatively large
population increases in the Rocky Mountain and Pacific Southwest
regions, which have the highest average ticket prices, mitigate pro-
jected national-scale losses in visits and ticket revenue, especially under
Fig. 10. Average percent change in downhill skiing season length based on a combination of UEB model and snowmaking results. A) Results for RCP4.5 in 2050. B) Results for RCP4.5 in
2090. C) Results for RCP8.5 in 2050. D) Results for RCP8.5 in 2090.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
10
Fig. 11. Example output for Aspen Mountain showing change in season length for downhill skiing across all GCMs under A) RCP4.5 in 2050, B) RCP4.5 in 2090, C) RCP8.5 in 2050, and
D) RCP8.5 in 2090.
Fig. 12. Average baseline and projected season start and end dates for the downhill ski season, across all modeled resorts. Median opening date is represented by the red line at the bottom
of the box and whiskers plot, and median closing date is represented by the red line at the top of the box. Boxes enclose the middle 50% of the season length distribution, and whiskers
extend to the 5th and 95th percentiles. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
11
RCP 4.5 (Table 1).
4. Conclusions
Physical models that account for changes in natural snow and ski
resorts’ability to make snow demonstrate that season lengths for winter
recreation activities will decline at nearly all sites in the CONUS under
the considered future climate scenarios. In each region of the United
States, these impacts increase in severity over time for a given emissions
scenario, and also increase in severity with GHG emissions for a given
time period: impacts are more severe under RCP8.5 vs RCP4.5 at any
point in time and more severe in 2090 compared to 2050 for a given
RCP.
Underlying these national results, we found considerable variability
at all levels of the analysis, particularly with respect to the spatial
distribution of impacts. In general, sites at higher elevations (such as
the Rocky Mountains and Sierras) tend to be more resilient to projected
changes in temperature and precipitation, whereas sites at lower
elevations (generally in the upper Midwest and New England) are more
sensitive to climate change. Based on our modeling, the difference
between RCP4.5 and RCP8.5 could represent the difference between
preserving skiing and snowmobiling in the eastern half of the country
and losing these activities almost completely by 2090 (Fig. 10). When
these physical modeling results are monetized using current prices for
recreational entry while accounting for population change, we find that
the changes in winter recreation season lengths under RCP8.5 could
result in a loss of more than $2 billion annually for downhill skiing, and
an additional $5 million and $10 million for snowmobiling and cross-
country skiing, respectively (Table 1).
These results include a number of important caveats. On the
physical modeling side, our snow model was simplified to simulate
average conditions at the top and bottom of 247 areas across the
CONUS, and was driven by a relatively coarse-scale representation of
climate. The UEB model framework is flexible enough that it could be
refined on a site by site basis to generate an improved calibration for
each individual site. However, this added level of specificity would
have made both the data requirements and the computational burden
too high for this national study.
Our monetization approach also required a number of simplifying
assumptions. For example, not all downhill ski areas participate in the
NSAA data collection or are located on national forest lands, so our
impacts on estimated visits may be understated. Similarly, considerable
cross-country skiing and snowmobiling activity occurs outside of
national forests, so those impacted visit estimates are likely also
understated. Our entry fee also does not measure the implicit value of
winter recreation or the full monetary impact of these activities in a
specific region or collection of regions. While alternative economic
approaches could be incorporated to try to fully monetize the projected
impacts of climate change on winter recreation, we did not attempt to
do that here.
We also have not attempted to account for the complete loss of
recreational activity with the closure of facilities, as this would require
consideration and development of business models or general operating
rules that are beyond the scope of this study. However, our modeling
does suggest increased pressure on downhill ski facility operators in
general as sequences of what would currently be considered marginal
snow seasons increase over time. This is particularly relevant when
recognizing that revenue and profit for downhill ski operators is often
concentrated into the start and end of the current winter season (e.g.,
Christmas/New Year’s holiday and spring break). Over time, pressure
on these important revenue periods could result in a facility’s closure.
Since we have not attempted to project potential closures, our projected
estimates of downhill skiing visits could be conservative if visits to a
closed area are not transferred to those that remain open.
Finally, we have not accounted for the different types of substitution
that could arise with climate change. The impacts of climate change on
Table 1
National projected impacts in terms of visits by recreational activity averaged across models for different time periods and RCPs.
Baseline Impacts in 2050 Impacts in 2090
Visits Dollars Change in visits
(RCP4.5)
Change in visits
(RCP8.5)
Equivalent monetized
impact (RCP4.5)
Equivalent monetized
impact (RCP8.5)
Change in visits
(RCP4.5)
Change in visits
(RCP8.5)
Equivalent monetized
impact (RCP4.5)
Equivalent monetized
impact (RCP8.5)
National impacts of climate change holding population constant
Downhill skiing 56,028,000 $5,400,134,000 (16,131,000) (19,772,000) ($1,367,232,000) ($1,716,806,000) (20,608,000) (36,259,000) ($1,783,996,000) ($3,255,810,000)
Cross-country
skiing
3,590,000 $32,368,000 (636,000) (873,000) ($5,949,000) ($8,037,000) (935,000) (2,055,000) ($8,623,000) ($18,711,000)
Snowmobiling 2,821,000 $12,641,000 (681,000) (844,000) ($3,072,000) ($3,798,000) (912,000) (1,802,000) ($4,106,000) ($8,102,000)
Total 62,439,000 $5,445,143,000 (17,448,000) (21,489,000) ($1,376,251,000) ($1,728,641,000) (22,455,000) (40,116,000) ($1,796,725,000) ($3,282,623,000)
National impacts of climate change with anticipated population growth
Downhill skiing 56,028,000 $5,400,134,000 (6,845,000) (11,270,000) ($345,580,000) ($778,142,000) (3,228,000) (25,448,000) $125,888,000 ($2,029,791,000)
Cross-country
skiing
3,590,000 $32,368,000 220,000 (87,000) $1,635,000 ($1,060,000) 628,000 (1,106,000) $5,226,000 ($10,299,000)
Snowmobiling 2,821,000 $12,641,000 (113,000) (318,000) ($525,000) ($1,442,000) 83,000 (1,212,000) $352,000 ($5,469,000)
Total 62,439,000 $5,445,143,000 (6,738,000) (11,675,000) ($344,470,000) ($780,644,000) (2,517,000) (27,766,000) $131,466,000 ($2,045,559,000)
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
12
future winter recreation season lengths and associated conditions raise
the potential for three general types of substitution, including (1)
temporal, where the timing of future recreation will change, generally
shifting to later in the season; (2) spatial, where recreators will change
travel patterns to access different areas; and (3) activity, where some
recreators may switch to different recreational activities altogether. By
adjusting future recreation for projected season length, we are imposing
a strong assumption that captures some, but not all of these substitution
elements.
All of these caveats represent simplifications that were required to
complete this national-scale analysis. Despite these simplifications,
however, our approach represents an important step forward in that
it combines detailed physical modeling with a nationally consistent
monetization approach to evaluate how climate change might affect
this important industry in the United States. The methodology we have
employed in this study is also easily transferable, and could be refined
and adapted for further insight within the United States or for
applications elsewhere. For example, we could gather more detailed
meteorological, topographic and spending data from a single resort to
develop a site-specific model to dive deeper into the potential impacts
for a specific location. Alternatively, we could synthesize national-scale
meteorological and topographic data from other parts of the world to
develop scoping analyses of climate change impacts on winter recrea-
tion for other countries or regions. In any case, it is clear from this study
that climate change will have profound impacts on users’ability to
enjoy skiing and snowmobiling over the 21st century in the United
States. These impacts will ripple through the economies of regions that
depend strongly on these activities, and indicate significant challenges
for snow-dependent communities under these, and similar, climate
change scenarios.
Funding sources
This work was funded by the U.S. Environmental Protection
Agency’sOffice of Atmospheric Programs through contract No. EP-
BPA-12-H-0024.
Acknowledgements
This work greatly benefited from information provided by Robin
Smith of TechnoAlpin and Harry Lynk of the Aspen Skiing Company
with respect to snowmaking and downhill ski area operations. The work
also benefited from discussions with U.S. Forest Service (USFS) staff
including Dan McCollum, Don English, David Chapman, and Eric White
over the use and interpretation of different USFS recreational survey
data and reports. The views expressed in this article are solely those of
the authors, and do not necessarily represent the views of their
employers.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.gloenvcha.2017.04.006.
Fig. 13. Comparison of projected impact of climate change on downhill skiing visits holding population constant and allowing for population changes over time.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
13
References
Argus, D.F., Fu, Y., Landerer, F.W., 2014. Seasonal variation in total water storage in
California inferred from GPS observations of vertical land motion. Geophys. Res. Lett.
41 (6), 1971–1980. http://dx.doi.org/10.1002/2014GL059570.
Barnett, T.P., Adam, J.C., Lettenmaier, D.P., 2005. Potential impacts of a warming climate
on water availability in snow-dominated regions. Nature 438 (7066), 303–309.
Barrett, A., 2003. National Operational Hydrologic Remote Sensing Center SNOw Data
Assimilation System (SNODAS) Products at NSIDC. NSIDC Special Report 11.
National Snow and Ice Data Center. Digital media, Boulder, CO USA.
Brown, R.D., Mote, P.W., 2009. The response of northern hemisphere snow cover to a
changing climate. J. Clim. 22 (8), 2124–2145.
Burakowski, E., Magnusson, M., 2012. Climate Impacts on the Winter Tourism Economy
in the United States. Natural Resources Defense Council, New York.
Burakowski, E.A., Wake, C.P., Braswell, B., Brown, D.P., 2008. Trends in wintertime
climate in the northeastern United States: 1965–2005. J. Geophys. Res. 113, D20114.
http://dx.doi.org/10.1029/2008JD009870.
Campbell, J.L., Ollinger, S.V., Flerchinger, G.N., Wicklein, H., Hayhoe, K., Bailey, A.S.,
2010. Past and projected future changes in snowpack and soil frost at the Hubbard
Brook Experimental Forest, New Hampshire, USA. Hydrol. Processes 24 (17),
2465–2480.
Clark, M.P., Hendrikx, J., Slater, A.G., Kavetski, D., Anderson, B., Cullen, N.J., Kerr, T.,
Orn Hreinsson, E., Woods, R.A., 2011. Representing spatial variability of snow water
equivalent in hydrologic and land-surface models: a review. Water Resour. Res. 47.
http://dx.doi.org/10.1029/2011wr010745W07539.
Cosgrove, B.A., Lohmann, D., Mitchell, K., Houser, P.R., Wood, E.F., Schaake, J.C.,
Robock, A., Marshall, C., Sheffield, J., Duan, Q., Luo, L., Higgins, R.W., Pinker, R.T.,
Tarpley, J.D., Meng, J., 2003. Real-time and retrospective forcing in the north
american land data assimilation system (NLDAS) project. J. Geophys. Res. 108 (D22).
http://dx.doi.org/10.1029/2002jd0031188842.
Daly, C., Neilson, R.P., Phillips, D.L., 1994. A statistical-topographic model for mapping
climatological precipitation over mountainous terrain. J. Appl. Meteorol. 33,
140–158.
Dawson, J., Scott, D., 2013. Managing for climate change in the alpine ski sector. Tour.
Manag. 35, 244–254. http://dx.doi.org/10.1016/j.tourman.2012.07.009. Available:.
http://dx.org/. Accessed June 21, 2016 http://dx.org/.
Diffenbaugh, N.S., Scherer, M., Ashfaq, M., 2013. Response of snow-dependent
hydrologic extremes to continued global warming. Nat. Clim. Change 3 (4), 379–384.
Dyer, J.L., Mote, T.L., 2006. Spatial variability and trends in observed snow depth over
North America. Geophys. Res. Lett. 33 (16), L16503. http://dx.doi.org/10.1029/
2006GL027258.
Förster, K., Meon, G., Marke, T., Strasser, U., 2014. Effect of meteorological forcing and
snow model complexity on hydrological simulations in the Sieber catchment (Harz
Mountains Germany). Hydrol. Earth Syst. Sci. 18 (11), 4703–4720. http://dx.doi.org/
10.5194/hess-18-4703-2014.
Flerchinger, G.N., Baker, J.M., Spaans, E.J.A., 1996. A test of the radiative energy balance
of the shaw model for snowcover. Hydrol. Processes 10 (10), 1359–1367. http://dx.
doi.org/10.1002/(SICI)1099-1085(199610)10:10<1359:AID-HYP466>3.0.CO;2-N.
Fu, Y., Argus, D.F., Landerer, F.W., 2015. GPS as an independent measurement to
estimate terrestrial water storage variations in Washington and Oregon. J. Geophys.
Res. Solid Earth 120, 552–566. http://dx.doi.org/10.1002/2014JB011415.
Hayhoe, K., Wake, C.P., Huntington, T.G., Luo, L., Schwartz, M.D., Sheffield, J., Wood, E.,
Anderson, B., Bradbury, J., DeGaetano, A., Troy, T.J., 2007. Past and future changes
in climate and hydrological indicators in the US Northeast. Clim. Dyn. 28 (4),
381–407.
Lazar, B., Williams, M., 2008. Climate change in western ski areas: potential changes in
the timing of wet avalanches and snow quality for the Aspen ski area in the years
2030 and 2100. Cold Reg. Sci. Technol. 51 (2), 219–228.
Lazar, B., Williams, M., 2010. Potential impacts of climate change for US wasatch range
ski areas: projections for park city mountain resort in 2030, 2050, and 2075.
Proceedings of the International Snow Science Workshop 436–443.
Livneh, B., Bohn, T., Pierce, D., Munoz-Arriola, F., Nijssen, B., Vose, R., Cayan, D., Brekke,
L., 2015. A spatially comprehensive, hydrometeorological data set for Mexico, the
U.S., and Southern Canada 1950–2013. Sci. Data 2, 150042. http://dx.doi.org/10.
1038/sdata.2015.42.
Mahat, V., Tarboton, D.G., 2012. Canopy radiation transmission for an energy balance
snowmelt model. Water Resour. Res. 48 (1), W01534. http://dx.doi.org/10.1029/
2011WR010438.
Mahat, V., Tarboton, D.G., 2014. Representation of canopy snow interception, unloading
and melt in a parsimonious snowmelt model. Hydrol. Processes 28 (26), 6320–6336.
http://dx.doi.org/10.1002/hyp.10116.
Mote, P.W., Hamlet, A.F., Clark, M.P., Lettenmaier, D.P., 2005. Declining mountain
snowpack in western North America. Bull. Am. Meteorol. Soc. 86 (1), 39.
NOAA, 2016. GIS Data Sets. National Weather Service National Operating Hydrologic
Remote Sensing Center: Skiing Locations. National Oceanic and Atmospheric
Administration (Accessed June 21, 2016.) Available:. http://www.nohrsc.noaa.gov/
gisdatasets/.
NSAA, RRC, 2016. Kottke National End of Season Survey 2015/16: Final Report. National
Ski Area Association and RRC Associates.
OpenSnowMap, 2016. Ski Extracts from the OpenSnowMap Database. (Accessed June 21
2016) Available:. http://www.opensnowmap.org/iframes/data.html#osm.
Pan, M., Sheffield, J., Wood, E.F., Mitchell, K.E., Houser, P.R., Schaake, J.C., Robock, A.,
Lohmann, D., Cosgrove, B., Duan, Q., Luo, L., Higgins, R.W., Pinker, R.T., Tarpley,
J.D., 2003. Snow process modeling in the north american land data assimilation
system (NLDAS): 2. Evaluation of model simulated snow water equivalent. J.
Geophys. Res. 108 (D22), 8850. http://dx.doi.org/10.1029/2003JD003994.
Pierce, D.W., Cayan, D.R., 2013. The uneven response of different snow measures to
human-induced climate warming. J. Clim. 26 (12), 4148–4167.
Pierce, D.W., Cayan, D.R., Thrasher, B.L., 2014. Statistical downscaling using localized
constructed analogs (LOCA). J. Hydrometeorol. 15 (6), 2558–2585.
Pierce, D.W., Cayan, D.R., Maurer, E.P., Abatzoglou, J.T., Hegewisch, K.C., 2015.
Improved bias correction techniques for hydrological simulations of climate change.
J. Hydrometeorol. 16, 2421–2442. (Accessed March 15, 2017) Available:. http://dx.
doi.org/10.1175/JHM-D-14-0236.1.
Raleigh, M.S., Lundquist, J.D., Clark, M.P., 2015. Exploring the impact of forcing error
characteristics on physically based snow simulations within a global sensitivity
analysis framework. Hydrol. Earth Syst. Sci. 19 (7), 3153–3179. http://dx.doi.org/
10.5194/hess-19-3153-2015.
Rutter, N., Essery, R., Pomeroy, J.W., Altimir, N., Andreadis, K., Baker, I., Barr, A.,
Bartlett, P., Boone, A., Deng, H., Douville, H., Dutra, E., Elder, K., Ellis, C., Feng, X.,
Gelfan, A.N., Goodbody, A., Gusev, Y., Gustafsson, D., Hellström, R., Hirabayashi, Y.,
Hirota, T., Jonas, T., Koren, V., Kuragina, A., Lettenmaier, D.P., Li, W.P., Luce, C.,
Martin, E., Nasonova, O., Pumpanen, J., Pyles, R.D., Samuelsson, P., Sandells, M.,
Schädler, G., Shmakin, A.B., Smirnova, T.G., Stähli, M., Stöckli, R., Strasser, U., Su,
H., Suzuki, K., Takata, K., Tanaka, K., Thompson, E., Vesala, T., Viterbo, P., Wiltshire,
A., Xia, K., Xue, Y., Yamazaki, T., 2009. Evaluation of forest snow processes models
(SnowMIP2). J. Geophys. Res. 114, D06111. http://dx.doi.org/10.1029/
2008JD011063.
Scott, D., McBoyle, G., Mills, B., 2003. Climate change and the skiing industry in southern
Ontario (Canada): Exploring the importance of snowmaking as a technical
adaptation. Clim. Res. 23, 171–181.
Scott, D., Dawson, J., Jones, B., 2008. Climate change vulnerability of the US Northeast
winter recreation −tourism sector. Mitig. Adapt. Strat. Glob. Change 13, 577–596.
http://dx.doi.org/10.1007/s11027-007-9136-z.
Serreze, M.C., Clark, M.P., Frei, A., 2001. Characteristics of large snowfall events in the
montane western United States as examined using snowpack telemetry (SNOTEL)
data. Water Resour. Res. 37 (3), 675–688.
Stull, R., 2011. Wet-bulb temperature from relative humidity and air temperature. J.
Appl. Meteorol. Climatol. 50 (11), 2267–2269. http://dx.doi.org/10.1175/JAMC-D-
11-0143.1.
Stynes, D.J., White, E.M., 2005. Spending Profiles of National Forest Visitors NVUM Four
Year Report. USDA Forest Service Inventory and Monitoring Institute and Michigan
State University.
Sultana, R., Hsu, K.-L., Li, J., Sorooshian, S., 2014. Evaluating the Utah Energy Balance
(UEB) snow model in the Noah land-surface model. Hydrol. Earth Syst. Sci. 18,
3553–3570. http://dx.doi.org/10.5194/hess-18-3553-2014.
Tarboton, D.G., Luce, C.H., 1996. Utah Energy Balance Snow Accumulation and Melt
Model (UEB). Computer Model Technical Description and Users Guide. Utah Water
Research Laboratory and USDA Forest Service Intermountain Research Station,
Logan, UT (December) (Accessed November 7, 2016.) Available:. http://citeseerx.ist.
psu.edu/viewdoc/download?doi=10.1.1.364.2121&rep=rep1&type=pdf.
Taylor, K., Stouffer, R.J., Meehl, G.A., 2012. An overview of CMIP5 and the experiment
design. Bull. Am. Meteorol. Soc. 93, 485–498. http://dx.doi.org/10.1175/BAMS-D-
11-00094.1.
U.S. EPA, 2015a. Climate Change in the United States: Benefits of Global Action. EPA 420-
R-15-001. U.S. Environmental Protection Agency Office of Atmospheric Programs,
Washington, DC.
U.S. EPA, 2015. Climate Change Indicators in the United States, Fourth edition. EPA 430-
R-16-004. U.S. Environmental Protection Agency Office of Atmospheric Programs,
Washington, DC.
U.S. EPA, 2016. Updates to the Demographic and Spatial Allocation Models to Produce
Integrated Climate and Land Use Scenarios (ICLUS) (Version 2) (External Review
Draft). EPA/600/R-14/324. U.S. Environmental Protection Agency, Washington, DC.
USBR, NCAR, USGS, LLNL, SCU, Climate Analytics Group, CIRES, Climate Central,
USACE, and Scripps. 2016. Downscaled CMIP3 and CMIP5 Climate and Hydrology
Projections −Addendum. Release of Downscaled CMIP5 Climate Projections (LOCA)
and Comparison with Preceding Information. U.S. Bureau of Reclamation, National
Center for Atmospheric Research, Lawrence Livermore National Library, Santa Clara
University, Climate Analytics Group, Cooperative Institute for Research in
Environmental Sciences, Climate Central, U.S. Army Corps of Engineers, and Scripps
Institution of Oceanography. September. Data available: http://gdo-dcp.ucllnl.org/
downscaled_cmip_projections, (Accessed November 7, 2016).
USFS, 2016. National Visitor Use Monitoring Survey Results: U.S. Forest Service National
Summary Report −Draft −Data Collected FY 2011 Through FY 2015. United States
Forest Service (Accessed November 7, 2016.) Available:. http://www.fs.fed.us/
recreation/programs/nvum/pdf/508pdf2015_National_Summary_Report.pdf.
USGS, 2008. U.S. Geological Survey EROS Data Center. SRTM Global Digital Elevation
Model. U.S. Geological Survey.
Walsh, J., Wuebbles, D., Hayhoe, K., Kossin, J., Kunkel, K., Stephens, G., Thorne, P., Vose,
R., Wehner, M., Willis, J., Anderson, D., Doney, S., Feely, R., Hennon, P., Kharin, V.,
Knutson, T., Landerer, F., Lenton, T., Kennedy, J., Somerville, R., 2014. Our changing
climate. Chapter 2 in Climate Change Impacts in the United States: The Third
National Climate Assessment, Melillo, J.M., Richmond, Terese (T.C.), Yohe, G.W.,
(Eds.). U.S. Global Change Research Program. pp. 19–67. 10.7930/J0KW5CXT.
Xia, Y., Mitchell, K., Ek, M., Sheffield, J., Cosgrove, B., Wood, E., Luo, L., Alonge, C., Wei,
H., Meng, J., Livneh, B., Lettenmaier, D., Koren, V., Duan, Q., Mo, K., Fan, Y., Mocko,
D., 2012. Continental-scale water and energy flux analysis and validation for the
North American Land Data Assimilation System project phase 2 (NLDAS-2): 1.
Intercomparison and application of model products. J. Geophys. Res. 117. http://dx.
doi.org/10.1029/2011jd016048D03109.
C. Wobus et al. Global Environmental Change 45 (2017) 1–14
14
