The 21st Century Colorado River Hot Drought and Implications for the
Future
Bradley Udall1,3, Jonathan Overpeck2,3
1

Colorado Water Institute, Colorado State University, Fort Collins, CO 80521

2

Department of Geosciences, Department of Hydrology and Atmospheric Sciences and Institute
for the Environment, University of Arizona, Tucson, AZ 85721
3

Colorado River Research Group

Corresponding author: Bradley Udall (Bradley.udall@Colostate.edu)
Key Points:
•

Record Colorado River flow reductions averaged 19.3% per year during 2000-2014.
One-third or more of the decline was likely due to warming.

•

Unabated greenhouse gas emissions will lead to continued substantial warming,
translating to 21st century flow reductions of 35% or more.

•

More precipitation can reduce the flow loss, but lack of increase to date and large
megadrought threat, reinforce risk of large flow loss.

This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as an
‘Accepted Article’, doi: 10.1002/2016WR019638

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Abstract
Between 2000 and 2014, annual Colorado River flows averaged 19% below the 1906-1999
average, the worst 15-year drought on record. At least one-sixth to one-half (average at onethird) of this loss is due to unprecedented temperatures (0.9°C above the 1906-99 average),
confirming model-based analysis that continued warming will likely further reduce flows.
Whereas it is virtually certain that warming will continue with additional emissions of
greenhouse gases to the atmosphere, there has been no observed trend towards greater
precipitation in the Colorado Basin, nor are climate models in agreement that there should be a
trend. Moreover, there is a significant risk of decadal and multidecadal drought in the coming
century, indicating that any increase in mean precipitation will likely be offset during periods of
prolonged drought. Recently published estimates of Colorado River flow sensitivity to
temperature combined with a large number of recent climate model-based temperature
projections indicate that continued business-as-usual warming will drive temperature-induced
declines in river flow, conservatively -20% by mid-century and -35% by end–century, with
support for losses exceeding -30% at mid-century and -55% at end-century. Precipitation
increases may moderate these declines somewhat, but to date no such increases are evident and
there is no model agreement on future precipitation changes. These results, combined with the
increasing likelihood of prolonged drought in the river basin, suggest that future climate change
impacts on the Colorado River flows will be much more serious than currently assumed,
especially if substantial reductions in greenhouse gas emissions do not occur.
Index Terms: 1807 Climate Impacts; 1812 Drought; 1833 Hydroclimatology; 1880 Water
Management; 1884 Water Supply
Keywords : Colorado River Basin, climate change, Colorado River Compact,
megadrought,
1. Introduction
A large number of studies over the last twenty-five years have considered the future runoff of the
Colorado River (Figure 1) under climate change. Nearly all of these studies have cautioned that
future warming will deplete the flow of the river, but the results have varied from minor to major
[Nash and Gleick, 1991; Christensen et al., 2004; Milly et al., 2005; Brekke et al., 2007;
Christensen and Lettenmaier, 2007; National Research Council, 2007; Seager et al., 2007;
Barnett and Pierce, 2008; Ray et al., 2008; Barnett and Pierce, 2009; Rajagopalan et al., 2009;
Cayan et al., 2010; Reclamation, 2011; Harding et al., 2012; Seager et al., 2012; Vano et al.,
2012; Ficklin et al., 2013; Vano et al., 2014; Ayers et al., 2016; Milly and Dunne, 2016]. In
contrast, the latest U.S Government assessment implies little or no change is likely because
precipitation increases will be sufficient to maintain temperature-depleted flows [Reclamation,
2016]. Fifteen years into the 21st century, the emerging reality is that climate change is already
depleting Colorado River water supplies at the upper end of the range suggested by previously
published projections. Record setting temperatures are an important and underappreciated
component of the flow reductions now being observed.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Between the start of the drought in 2000 and the end of 2014, our analysis period, annual flow
reductions averaged 19.3% below the 1906 to 1999 normal period, and Lakes Mead and Powell,
the nation’s two largest reservoirs, ended the period at approximately 40% of maximum volume
despite starting the period nearly full [Wines, 2014; Colorado River Basin Stakeholders, 2015]
(Figure 2a). This drought has continued into 2015 and 2016 with higher, but still below normal,
flows estimated at 94% in 2015 and 94% in 2016 with unusual late season May and June
precipitation in both years that raised runoff by nearly 20% [Alcorn, 2015, 2016]. Despite these
smaller recent reductions, Lake Mead continues to decline and in May of 2016 it hit a level not
seen since its initial filling in the 1930s [James, 2016]. The overall Colorado River reservoir
system stores 4 times the annual flow of the river, one of the largest ratios in the world. This
storage provides a large drought buffer when full. However, when the reservoirs are low,
shortage risk can be high for years because high demands, now equal to 20th century average
flow, make it difficult to refill system storage [Reclamation, 2012]. While the multi-year
California drought has been garnering more national attention, the more slowly unfolding
Colorado River drought is every bit as serious and also has national and international
ramifications [Wines, 2014].
The Colorado River Basin encompasses seven states and northern Mexico and is home to 22
federally recognized tribes. The river provides municipal and industrial water for 40m people
distributed across every major Southwestern city both within and without the basin, including
Los Angeles, San Diego, Las Vegas, Phoenix, Tucson, Salt Lake City, Denver and the entire
Front Range of Colorado, Albuquerque and Santa Fe [Reclamation, 2012].
Continued low flows would result in additional declines at Lake Mead, eventually requiring
Lower Basin (Arizona, California, Nevada) water delivery shortages with mandatory cutbacks
imposed primarily on Arizona, but also Nevada and Mexico [Verburg, 2011]. At the same time,
Upper Basin (Colorado, New Mexico, Utah, Wyoming) water users would continue to endure
physical shortages from a lack of water. These initial Lower Basin Lake Mead delivery shortages
and Upper Basin physical shortages are manageable to a point; however, under current operating
rules with continued low flows during the next six to eight years Lake Mead would drop to
elevation 305 m (1000 feet) above sea level, resulting in a number of serious and unprecedented
problems [Collum and McCann, 2014].
In the Lower Basin, Arizona could theoretically lose its water allocation for the entire Central
Arizona Project canal, a critical $4.4B, 530 km cross-state 2 bcm/year water source for 4.7m
people, multiple sovereign Indian nations, and over 120,000 irrigated hectares [Glennon, 1995;
Colorado River Basin Stakeholders, 2015]. This canal currently relies on occasional but
uncertain ‘equalization’ releases from Lake Powell that only occur with irregular and rare large
Powell inflows. The extra water is delivered when Lake Powell reaches levels substantially
higher than Lake Mead, a use allowed under the 1922 Colorado River Compact section III (e)
and formalized most recently under rules established in a 2007 Record of Decision for
coordinated operations of Lakes Powell and Mead and for shortage sharing in the Lower Basin
[Department of Interior, 2007].

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Under normal operating rules, without these extra inflows, Lake Mead has excess outflows of
1.5 bcm per year, the so-called Lower Basin “structural deficit” [Collum and McCann, 2014].
The structural deficit was created in 1968 when Congress authorized the Central Arizona Project
(CAP). In order to obtain the support of the large California Congressional delegation, Arizona
agreed to rely on this unused, but in the long run unreliable water, because there was not enough
remaining unallocated Lower Basin water. The CAP had long been a desire of Arizona and the
state was willing to make this bargain despite its flaws [Johnson, 1977]. This same water is first
available for use by the Upper Basin under the Colorado River Compact, but heretofore has not
been developed for Upper Basin use. A plan to augment the Colorado River with flows from
outside the basin, discussed during the hearings on the legislation, but not included in the final
package due to opposition from potential source areas, was never revisited by Congress.
Reclamation in 2011 said that such augmentation was now unlikely.
The structural deficit only became a problem when the CAP was fully completed in the mid
1990s combined with the drought that began in 2000. Upper Basin demand growth has also
played a small role, although Upper Basin demands are still much less than forecast in 1968 for
the year 2000 [Tipton and Kalmbach, Inc, 1965; Johnson, 1977]. The recent Lake Mead declines
are strongly influenced by this imbalance, and solutions to this deficit have been a recent focus
of the Basin states and federal government [Central Arizona Project, 2016; Davis, 2016].
The Upper Basin also has serious issues, one of which ripples into the Lower Basin. When the
surface of Lake Mead declines to an elevation 305 m (1000 feet) above sea level, Lake Powell
will also be below its minimum power pool 75% of the time [Collum and McCann, 2014]. This
occurs in part because low Mead levels make ‘equalization’ releases from Powell more likely
thus driving Powell lower. Hydropower losses at Lake Powell could result in substantial rate
increases for irrigators who rely on the reservoirs for long term lower cost power contracts, and
would also dry up funding for basin-wide programs necessary for water delivery environmental
compliance [Adler, 2007; Collum and McCann, 2014]. Under such low reservoir conditions,
there is also a high likelihood that the Upper Basin states would have to curtail existing water
deliveries to cities such as Denver, Colorado Springs, Albuquerque and Salt Lake City in order
to make required deliveries to Lake Mead. Heretofore, largely because of the structure of the
Colorado River Compact, the Upper Basin and Lower Basin have been managed
separately. With permanent flow declines of approximately 20%, however, the required
deliveries to Lake Mead would become a hardship on the Upper Basin, as well as create Lower
Basin delivery shortages [Reclamation, 2007; Barnett and Pierce, 2009; Rajagopalan et al.,
2009]. The original compact, signed during one of the wettest periods in the last 450 years
[Woodhouse et al., 2006], did not envision how large scale flow declines would be managed
between the basins, and such declines could cause an allocation crisis between the Upper and
Lower Basins [Adler, 2008].

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Understanding the cause of, and reacting properly to, the ongoing drought is critical to the future
of the Southwest. Herein we investigate the role of precipitation versus temperatures as causes
of the current drought, provide temperature- and precipitation-based 21st century flow projections
and provide policy implications of these findings. Our approach separates the impacts of highconfidence temperature projections from those associated with the much lower-confidence
projections of future precipitation using a simple but powerful sensitivity technique. Moreover,
we make a novel – and important - case that there is a high likelihood that the impacts of
continued atmospheric warming will overwhelm any future increases in precipitation because
prolonged dry periods lasting multiple decades are likely to negate the beneficial impacts of
additional precipitation during other times.
Causes of the 2000-2014 Drought
The 2000-14 drought is defined by the lowest average annual flows for any 15-year period in the
historical record. To analyze this drought, gridded 4x4 km temperature and precipitation data
from 1896-2014 for the area above Lees Ferry were obtained from the Precipitation-Elevation
Regression on Independent Slopes (PRISM) model [Daly et al., 1994]. In addition, we obtained
reservoir contents and natural flows at Lees Ferry from the U.S. Bureau of Reclamation
(Reclamation) (Text S1). Lees Ferry is situated just below Lake Powell and is the Compact
dividing line between the Upper and Lower Basins. Approximately 85% of the flow originates
above Lees Ferry [Christensen and Lettenmaier, 2007].
Historically, Upper Colorado River Basin precipitation has been the main Colorado River runoff
driver such that high flow years (1920s, 1980s) were associated with high precipitation and low
flow years (1930s, 1950s) with low precipitation (Figure 2b-c). The current drought (our study
period is 2000-2014, but the drought is still on-going), with its modest -4.6% precipitation
decline and -19.3% flow decline, stands in stark contrast to the second-lowest 15-year flow
period (1953-1967), a precipitation-driven drought with averaged precipitation reductions of 6.1% per year and flow reductions of -18.1% per year (Figure 2b-c and Table 1). Compared to
the 1950s drought, the 2000s feature much more (near normal) winter precipitation (-8.6% 1950s
decline versus -2.7% 2000s) and significantly less summer precipitation (-3.6% 1950s decline
versus -6.4% 2000s). The 2000s precipitation decline is only 75% of the decline in the 1950s,
thus begging the question of why the recent drought was more serious. What has changed is that
temperatures in the runoff producing Upper Basin are now 0.9 °C above the 1896-1999 average
and are the highest in the gaged record; whereas temperatures during the 1953-1967 drought
were much cooler and only slightly above the 1896-1999 average (Figure 2d and Table 2). This
makes the current drought unprecedented in the gaged record.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

In contrast to the more precipitation-driven current California drought [Diffenbaugh et al., 2015;
Williams et al., 2015], lack of precipitation is only partially to blame for the Colorado River
runoff declines during the last 15 years. Instead, approximately a third, or more, of the recent
Colorado River flow reduction is most likely a result of record-setting warmth. Since 1988 an
increase in the frequency of warm years has been strongly associated with lower flows than
expected [Woodhouse et al., 2016], suggesting an important role for temperature in flow losses.
Such temperature-driven droughts have been termed ‘global-change type droughts’ and ‘hot
drought’, with higher temperatures turning what would have been modest droughts into severe
ones, and also increasing the odds of drought in any given year or period of years [Breshears et
al., 2005; Overpeck, 2013]. Higher temperatures increase atmospheric moisture demand,
evaporation from water bodies and soil, sublimation from snow, evapotranspiration (ET) from
plants, and also increase the length of the growing season during which ET occurs [Pitman,
2003; Weiss et al., 2009; Seneviratne et al., 2010; Seager et al., 2015a]. Warm season (April to
September) warming has been identified by models as especially important in reducing Colorado
River flows because of the increases in ET from longer growing seasons [Das et al., 2011].
Increases in measured vapor pressure deficits in the Southwest caused by warming and a
decrease in water vapor provide strong support for higher ET during the recent drought [Seager
et al., 2015b]. As increasing temperatures drive further drying, additional positive feedbacks are
possible in the form of lower humidity and less evaporative cooling, decreased cloudiness and
increased incident radiation, as well as decreased snow cover and more radiative heating [Betts et
al., 1996; Brubaker and Entekhabi, 1996; Pitman, 2003; Seneviratne et al., 2010]. In the 20th
century, droughts were associated almost exclusively with a lack of precipitation. In this
century, however, high temperatures alone can lead to anomalously dry conditions.
Estimates of 2000-2014 Temperature-induced Flow Loss
Over the last several years several studies specific to the Colorado River Basin have investigated
the specific relationships among temperatures, precipitation and flow in the basin using the
concepts of temperature sensitivity and precipitation elasticity [McCabe and Wolock, 2007;
Nowak et al., 2012; Vano et al., 2012, 2014; Vano and Lettenmaier, 2014]. Temperature
sensitivity is defined as the percent change in annual flow per degree rise in annual temperature.
Precipitation elasticity is defined as the fractional change in annual flow divided by the fractional
change in annual precipitation [Vano et al., 2012]. Note that elasticity has been studied for both
increases and decreases in precipitation, whereas sensitivity is typically investigated only for
temperature increases. These numbers can be determined empirically and through model studies.
Previous studies on temperature sensitivity and precipitation elasticity show that future impacts
to streamflow from increases in temperatures and changes in precipitation can be considered
separately using sensitivity and elasticity, and then added together to produce flow estimates
[Vano et al., 2014; Vano and Lettenmaier, 2014]. Considering these effects separately and
additively is a powerful conceptual tool for investigating climate change impacts because of the
ease in measuring the two variables for current impacts and the wide availability of temperature
and precipitation projections from global climate models for assessing future impacts. In
addition, the large differences in certainty associated with future changes in the two variables
(temperature will surely increase, whereas precipitation may increase or decrease – see below)
helps to set apart the risk of future changes in flow associated with each variable.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Vano et al. [2012, 2014], McCabe and Wolock [2007] and Nowak et al. [2012] provide multiple
estimates of the flow sensitivity of the Colorado River flow to temperature using three different
methods. Vano et al (2012, 2014) utilized 6 high-resolution, commonly used hydrology models
and two different temperature adjustment methods to obtain Lees Ferry temperature sensitivities.
They report an average sensitivity of -6.5% / °C warming with a one standard deviation range
from -3.0% to -10.0% / °C for the Upper Basin. Approximately 50% models show increasing
sensitivity and 50% decreasing sensitivity as temperatures warm so we elect to use a constant
sensitivity over all future temperatures. McCabe and Wolock (2007) constructed a simple water
balance model that infers an average temperature sensitivity of -8.9%/ °C and Nowak et al
(2012) found an empirical temperature sensitivity of -13.8%/ °C.
We use the complete one standard deviation range (-3% / °C to -10% / °C) of the Vano et al.
(2012, 2014) temperature sensitivity estimates as they were the most conservative and rigorous
of the three studies we investigated. Using this range, we found that recent warming of 0.9 °C
has likely already reduced river flows from -2.7% to -9% from the mean 1906-1999 flow. This
represents approximately one-sixth to one-half (average of one-third) of the total flow loss
during the 2000-2014 drought.
The higher temperature sensitivities of the two other studies suggest the actual Colorado River
temperature sensitivities are near the upper end and possibly exceed the Vano et al. (2012, 2014)
estimates. These higher sensitivities imply much greater temperature-induced losses during the
current drought (-7.9% to -12.3% vs -2.7% to -9%). Empirical results from the 2000-2014
drought also point to mid to high temperature sensitivities. Vano et al. (2012) report precipitation
elasticities ranging from 2 to 3 at Lees Ferry. Thus, using a mid-range precipitation elasticity of
2.5, the 2000-2014 annual -4.6% precipitation decline implies runoff reductions of -11.4%,
leaving the remaining -7.9% decline to be explained by other causes. If temperature were the
sole cause of this remaining decline, the inferred temperature sensitivity is -8.8% /°C. Using a
precipitation elasticity of 3.0 implies a temperature sensitivity of -6.2% /°C, very close to the
mid-range Vano et al., sensitivity. These temperature sensitivities imply large losses as
temperatures rise, the subject of the next section.
21st Century Flow Response to Changing Temperatures and Precipitation
For the analysis on how future temperatures and precipitation would affect runoff, and for
investigating how well current linked climate-hydrology models can reproduce the current
drought, we used Reclamation’s climate projection datasets [Brekke et al., 2013, 2014]. These
datasets use Coupled Model Intercomparison Project 3 and 5 (CMIP3, CMIP5 after the class of
climate models used) climate model projection data linked to the Variable Infiltration Capacity
hydrology model to produce flows from 1950 to 2099 (Text S2, Figures S2, S3)] [Liang et al.,
1996; Meehl et al., 2007; Taylor et al., 2012].

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

The same temperature sensitivity and precipitation elasticity numbers discussed above can be
used to estimate future flow reductions using climate model outputs under high (business-asusual, SRES A2 and RCP8.5) and moderate (somewhat reduced by mitigation, SRES A1B and
RCP4.5) greenhouse gas emissions to the atmosphere. By 2050, moderate and high emissions are
projected to yield Upper Basin mean warming of 2.6 to 2.8 °C (Figure 3), three times recent
warming, and by 2100, warming of 3.6 °C under moderate emissions and 5.4 °C under high
emissions. This warming implies total multi-model mean temperature-induced flow losses at
mid-range sensitivity of -6.5%/ °C of about -17% by mid-century and -25% to -35% at endcentury (Figures 4 and 5). The multi-model mean complete flow loss range over both periods
and both emissions is approximately -8% to -55% using the lower and upper temperature
sensitivities (Figures 4 and 5). As discussed above, there is little empirical evidence that the true
temperature sensitivity of flow to temperature increase is near the low sensitivity.
Temperature-induced losses may be somewhat buffered by projected additional precipitation that
can increase runoff by 2% to 3% for every 1% change in precipitation [Vano et al., 2012]. At
mid-century precipitation increases of +4 to +11% given a mid-range elasticity of 2.5 would
balance the range of temperature-induced flow losses at a mid-range – 6.5%/°C sensitivity
(Figure 5, right y-axis). At end-century, with the same sensitivity and elasticity, additional
precipitation increases of +4 to +20% would balance the range of possible temperature-driven
losses. At a higher -10%/°C sensitivity, the balancing precipitation would need to be as great as
+15% or more at mid-century and +25% or more at end-century. While these may seem like
relatively small increases in precipitation, and thus possible, they would represent a major and
unprecedented change in precipitation regime compared to the observed historical variation in
precipitation (Figure 2c). During the 20th century, for example, the wettest 10-year period (198397) had only a +8% precipitation increase. This unusual period was marked by major floods
downstream of Lakes Powell and Mead due to uncontrolled reservoir spilling and the near
catastrophic loss of the spillways at Glen Canyon Dam [Udall, 1983].
Vano and Lettenmaier [2014] argue that the sensitivity-based approach used in our projections
provides similar estimates of future streamflow to those generated with more computationallyintensive coupled-model methods, except for some (i.e., 10%) overstatement of flow reductions
at the highest levels of possible warming by 2100 (e.g., the business-as-usual SRES A2 scenario
used in the CMIP3 projections and the RCP8.5 in the CMIP5 projections). This would reduce the
end of century high emissions mean flow reductions shown in Figure 5 to a still very significant 45% by 2100.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Recent studies have suggested that CO2 fertilization may increase plant water efficiency thus
reducing future evapotranspiration which could serve to mitigate our projected losses [Milly and
Dunne, 2016; Swann et al., 2016]. Both studies call into question results that show large
portions of the globe drying in the 21st century [e.g., Dai, 2012; Cook et al., 2014]. However,
Milly and Dunne [2016] and Swann [2016] show that, despite this increase in plant water use
efficiency, the Southwestern US will still dry, a finding that is consistent with multiple global
assessments showing substantial drying risk to mid-latitude areas such as the Colorado River
Basin. Moreover, a recent Australian study found that higher evapotranspiration associated with
the increased plant growth stimulated by higher CO2 outweighed any CO2-related water-use
efficiency effect, and served to reduce streamflows in semi-arid regions [Ukkola et al., 2015], a
trend that must be exacerbated by the temperature-induced lengthening of the growing season.
These results suggest that plant physiological responses are likely consistent with our results, and
in any case, do not invalidate them.
Megadrought Risks to Flows
Megadroughts lasting decades in the Colorado River Basin have occurred in the past, with
resulting substantial flow reductions [Meko et al., 2007]. Multiple papers now suggest there is
high 21st century risk for megadrought in the American Southwest and that the risk will increase
as temperatures rise [Ault et al., 2014; Cook et al., 2015; Ault et al., 2016]. In addition, current
GCMs underrepresent the frequency of megadrought [Ault et al., 2012, 2013]. These findings
provide additional support for large flow reductions during at least multi-decadal drought periods
and suggest that current 21st century flow projections underrepresent this risk.
Significant Colorado River flow low losses occurred during previous multi-decadal
megadroughts. During the twelfth century, flow reductions of approximately -16% occurred
during one twenty-five year period [Meko et al., 2007]. Evidence indicates that hemispheric and
Southwest temperature anomalies were significantly smaller during past megadroughts than the
rapid on-going current warming that could easily exceed 4-5 °C by the end of century under
business-as-usual emissions [Salzer and Kipfmueller, 2005; Mann et al., 2009; Salzer et al.,
2014, Figure 5]. Using the additivity concepts discussed above, additional warming of 1, 2 or 3
°C beyond the historic twelfth century megadrought temperatures would have reduced the -16%
flow declines by an additional -6.5%, -13% or -19.5% at medium temperature sensitivity. These
additional reductions would have thus turned a -16% flow decline into declines of -21.5%, -28%
or -34.5%, losses near the middle of our projections.
There is recent strong evidence that continued warming over the next 80 years could increase the
risk of multidecadal drought [Ault et al., 2014, 2016; Cook et al., 2015]. Independent of the
added drought risk due to continued warming, the risk of a 35-year precipitation-deficit drought
later in this century exceeds 15% within a 50-year period (Ault et al., 2014). In contrast, with
continued anthropogenic warming, the risk of multidecadal megadrought in the Southwest
increases to over 90% over this century if there is no increase in mean precipitation; even if
modest precipitation increases do occur, the risk will still exceed 70% [Ault et al., 2014, 2016].
At medium warming (4 °C), 20% to 30% precipitation increases will be needed to reduce
megadrought risk below 50% and at high amounts of warming (>6 °C), it will take a ~40%
increase in precipitation to reduce megadrought risk below 50% [Ault et al., 2016]. These
changes in precipitation are huge and unlikely, and they would still only reduce megadrought
risk to below 50%.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Both the CMIP3 and CMIP5 Global Climate Models may not adequately reproduce the
frequency of occurrence of known past decadal and multi-decadal precipitation droughts [Ault et
al., 2012, 2013]. In the Colorado River Basin empirical evidence of this problem can be found in
the linked GCM-hydrology model results from Reclamation’s projections for the basin [Brekke
et al., 2014]. Approximately half of the CMIP5 models and one-quarter of the CMIP3 models
cannot simulate the 2000-2014 drought at any point in the 21st century (Text S3, Tables S1-S4).
This wet bias significantly affects the mean flows of drought-capable and non-drought capable
models. At the end of the 21st century the models unable to simulate the current drought are
much wetter (109% of 20th century average Lees Ferry runoff for CMIP3, 113% for CMIP5)
than the models that are able to simulate the current drought (85% of average runoff for CMIP3,
91% CMIP5) (Tables S1-S4). These flow differences are greater than 20%, and represent the
difference between serious management challenges and significant oversupply.
Risk-Based Framing of Future Runoff Projections
At present, some outputs from global climate models are ready to support reliable risk-based
policy while others are not as ready. A key novel aspect of our research is to provide more
insight into where confidence is warranted, and where it is not, with respect to projections of
future climate and flow change in the Colorado River Basin. In the case of the Basin, every
single moderate and high emissions model simulation agrees that temperatures will continue to
rise significantly with continued emissions of greenhouse gases to the atmosphere - this result is
robust, highly certain and well-suited for informing policy choices. The fact that observations
also show substantial warming only strengthens this assertion.
On the other hand, simulated future precipitation change in the Basin is clouded with much
greater uncertainty due to substantial disagreement among models and a highly uncertain ability
to simulate realistic change in key phenomena such as storm-track position or decadal and
longer-scale drought. Whereas climate models are in general agreement that cool season (warm
season much less certain) precipitation declines are likely in the Lower Colorado River Basin,
these same models disagree when it comes to the sign and amount of precipitation change that is
likely in the Upper Basin. This is because precipitation change in the Upper Basin will depend
heavily on the exact changes in the position of cool season jet stream and storm-tracks, two
aspects of climate change that are not simulated with confidence by global climate models
[Collins et al., 2013].
Moreover, there is strong evidence that the mean positions of both the jet stream and stormtracks are likely to push poleward, expanding the area of aridity in the Colorado River Basin, but
the amount of this expansion is poorly constrained [Collins et al., 2013]. Multiple studies,
including some focused on the American Southwest, suggest the proximate cause of this drying,
Hadley Cell expansion, is already well underway and will continue [Seager et al., 2007; Scheff
and Frierson, 2012; Feng and Fu, 2013; Norris et al., 2016; Prein et al., 2016].

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Our results regarding future changes in Colorado River flows agree with many previous studies
in suggesting climate change translates to flow reductions, although our work is generally not
directly comparable because we separate out high confidence temperature-related impacts from
the possible effects of much less certain and highly variable precipitation projections. However,
our work, as well as this larger body of literature, appears to be at odds with the recent
Reclamation projections for the Colorado River Basin, which are widely cited and used.
Reclamation’s projections use a global climate model output that is downscaled to drive a
hydrology model. It is worth understanding why our results emphasize substantially greater risks
along with apparently greater flow losses.
The 2011 CMIP3 climate change flow projections by Reclamation indicate a modest multimodel median flow decline of -9% by 2060 for the river, but with a wide range of outcomes from
flow increases to flow decreases [Reclamation, 2012] (Table S1). Reclamation’s most recent
CMIP5 projections show no change in mean and median basin-wide flow by 2070s
[Reclamation, 2016], but also embody a wide range of results. Compared to CMIP3, the CMIP5
results show increased precipitation, especially in the northern parts of the basin including
Northeast Utah, Northwest Colorado’s Yampa River and the Green River in Wyoming [Brekke et
al., 2014; Ayers et al., 2016] (Tables S1 and S3). The increased precipitation in the CMIP5
model runs compared to CMIP3 can be attributed to more southerly storm tracks in CMIP5 that
occur in late spring [Brekke et al., 2014].
Another issue arises in both the CMIP3 and CMIP5 datasets when GCM precipitation is adjusted
by the downscaling techniques necessary for off-line hydrology models. The first step in
Reclamation’s downscaling is a bias correction step. This step can add approximately 5% more
precipitation to the raw GCM precipitation, and this increase appears to not have a physical basis
[Reclamation, 2011; Brekke et al., 2013]. The final downscaling step, spatial downscaling, also
increases GCM precipitation, although there is at least a plausible physical explanation for some
of the increase: higher elevations in the Rockies receive large amounts of precipitation, but these
elevations are not properly modeled by the GCMs. In one study of the CMIP5 dataset after
downscaling, dry and average models show precipitation increases of approximately +~5%
from the raw GCM output, but the wettest models show +~10% increases, doubling future
precipitation increases from +10% to +20% [Lukas et al., 2014]. This extra precipitation is
manifested in a number of hydrology model runs that project huge and implausible flow
increases in some years that are 150% of the highest known flows in the 20th century (Text S4,
Figures S2, S3). The downscaling wetness problem has been identified, but has not been not
resolved [Lukas et al., 2014]. Reclamation acknowledges that the newer CMIP5 projections have
not been determined to be better or more reliable [Brekke et al., 2014]. It is noteworthy that
internally consistent GCM-only Southwest runoff projections almost uniformly produce
significant declines in both CMIP3 and CMIP5 runs [Milly et al., 2005; Seager et al., 2007,
2012; Koirala et al., 2014; Milly and Dunne, 2016].

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Our results are generally comparable to Reclamation’s most recent results when considering the
full range of our analysis when both precipitation and temperatures are included. However, our
focus and emphasis is on the large near-certain temperature-induced flow declines with a
separate analysis of precipitation. Reclamation, by contrast, has a focused on climate multimodel-ensemble median declines, including medians calculated across emission scenarios
[Reclamation, 2011, 2012]. Decision makers often treat these median outcomes as a proxy for
risk despite the fact that the median obscures the wide range of results and lumps wet and dry,
warm and hot, large and small emission increases and, most critically, near certain temperature
increases and very uncertain precipitation changes.
We assert that the large precipitation increases necessary to offset substantial temperatureinduced flow decreases appear unlikely to occur for a number of reasons. These reasons include
the potential for storm tracks to go north of the basin due to Hadley Cell expansion, the high
potential for megadrought to increase evaporation while reducing precipitation and runoff for
extended periods, the large size of the needed precipitation increases, especially when compared
to decadal historical increases, the consistent identification by global assessments of the
Southwest as an area likely to dry, and finally the lack of any trend over the last century or last
16 years (Figure 2c). Hence, we choose to focus on highly likely temperature-induced declines
with separate analysis of the precipitation needed to offset these declines.
5. Policy Implications and Solutions
The climate science take-home messages for Colorado River managers are thus: (1) there is little
doubt (i.e., high confidence) that temperatures will continue to increase as long as the emissions
of greenhouse gases to the atmosphere continue; (2) there is also high confidence that continued
temperature increases will cause river flows to decline, ranging from -11% to as much as -55%
by end of century under moderate to high emissions (Figures 4 and 5); (3) there is only low
confidence associated with the possibility of storms and precipitation in the Upper Basin
increasing enough to even partially offset the temperature-driven declines in river flows; (4) the
risk of multidecadal megadrought in the Basin is significant even in the absence of continued
anthropogenic climate change, and this risk rises substantially with continued global warming;
(5) the likelihood of drought and megadrought means that there will likely be decades-long
periods with anomalously low runoff even if there is an increase in precipitation relative to the
historical mean during some other periods due to anthropogenic climate change.
Temperature-driven threats to the flows of the Colorado are thus large and real. The only way to
curb substantial risk of long term mean declines in Colorado River flow is thus to work towards
aggressive reductions in the emissions of greenhouse gases into the atmosphere. Our work
shows that modest (e.g., RCP4.5) reductions in greenhouse gas emissions, while having better
outcomes than the business-as-usual future (e.g., RCP8.5), still imply large Colorado River flow
losses.
The record warm nature of the on-going Colorado River drought indicates that this drought is not
just a natural drought, and our work demonstrates that flows are unlikely to return to the 20th
century averages if we only wait. Unusually wet periods like the 1920’s and 1990’s will still
continue to occur, but they will co-occur with higher temperatures that will increase water
demand from plants, soil, snow and humans.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Climate models and theory suggest that flow reductions would be more severe in the Southern
portions of the Upper Colorado Basin affecting tributaries such as the San Juan, Dolores, and
Gunnison more severely, with smaller impacts to more northerly tributaries such as the Yampa
and Green [Ayers et al., 2016] . Such spatial distribution would provide additional water
management challenges in that the more southerly basins have in general more people,
infrastructure and uses. Such a distribution would create new localized water supply shortages in
addition to the overall basin-wide issues.
Other known threats to streamflows include the potential large scale loss of conifers [Breshears
et al., 2005; Adams et al., 2009; Allen et al., 2010, 2015] , and the impacts of dust on snow
[Painter et al., 2010; Deems et al., 2013]. These factors along with the observed and projected
temperature-induced Colorado River flow declines, the inability of many linked climatehydrology models to simulate persistent droughts, and the increasing likelihood of hot drought
and megadrought, all imply that future Colorado River water supply risk is high. It is imperative
that decision-makers begin to consider seriously the policy implications of potential large-scale
future flow declines. Stable 20th century Colorado River flow regimes may not reoccur for many
centuries – the time scale of climate system readjustment to the complete cessation of
greenhouse gas emissions [Solomon et al., 2009; Collins et al., 2013] .
The Colorado River declines do not stand alone as the only warming-related threat to
Southwestern water supplies. The Rio Grande also has a grim prognosis [Reclamation, 2013;
Elias et al., 2015]. The drought in California has garnered national attention, and multiple
studies have strongly implicated increasing temperatures as a contributor to these woes [Griffin
and Anchukaitis, 2014; Belmecheri et al., 2015; Diffenbaugh et al., 2015; Mann and Gleick,
2015; Seager et al., 2015a]. Southern California is particularly at risk, with a critical economy
and a very large population, all coupled with a large reliance on both climate-threatened in-state,
as well as Colorado River, water.
Adjusting to the new reality of rapid climate change will not be an easy or fast task; water
management and water policy change slowly. The Colorado River is managed by a complex set
of agreements, interstate compacts approved by Congress, international agreements, legislation
and court decrees set in place over the last 100 years [Verburg, 2011]. Most agreements were
derived from 20th century state-based negotiations with win/lose policy prescriptions that
minimized basin-wide considerations of economic prosperity and potential harm [Adler, 2008].
None expressly includes climate change risk management, nor the provision for flow reductions
that will be relentless on decadal timescales. New agreements often take years to put in place
[Department of Interior, 2007]. The recently proposed structural deficit solution [Central
Arizona Project, 2016], while important and laudable for the short term, will not solve the
problem of large scale flow losses. With reduced water supplies, much will have to change in
these agreements to address equity, economics and social concerns on regional, state, basin-wide
and even national levels. Climate change threats to western water supplies are very real, and
should prompt great concern and urgency among both water managers and the citizens of the
Southwest.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Acknowledgments and Data
We thank the U.S. Bureau of Reclamation for data assistance and comments, and M. Kelly, D.
Kenney, E. Kuhn, D. Lettenmaier, J. Robison, P. Troch, and C. Woodhouse for comments and
discussion. A portion of this research was supported by the Colorado Water Institute, National
Science Foundation, the NOAA Climate Assessment for the Southwest, and the U.S. Geological
Survey Southwest Climate Science Center. We acknowledge the Downscaled CMIP3 and
CMIP5 Climate and Hydrology Projections archive at: http://gdodcp.ucllnl.org/downscaled_cmip_projections as well as the World Climate Research
Programme’s Working Group on Coupled Modeling and the U.S. Department of Energy, Office
of Science. The views expressed in this article are those solely of the authors. All data used in the
analysis and conclusions in this paper are available from links in the Supporting Information.

References
Adams, H. D., M. Guardiola-Claramonte, G. A. Barron-Gafford, J. C. Villegas, D. D. Breshears, C. B. Zou, P. A.
Troch, and T. E. Huxman (2009), Temperature sensitivity of drought-induced tree mortality portends
increased regional die-off under global-change-type drought, Proc. Natl. Acad. Sci., 106(17), 7063–7066.
Adler, R. W. (2007), Restoring Colorado River Ecosystems: A Troubled Sense of Immensity, Island Press,
Washington.
Adler, R. W. (2008), Revisiting the Colorado River Compact: Time for a Change, J Land Resour. Envtl L, 28, 19.
Alcorn, B. (2015), Colorado Basin River Forecast Center Final Reservoir Forecasts September 1, 2015, NOAA,
Salt Lake City, UT.
Alcorn, B. (2016), Colorado Basin River Forecast Center Final Reservoir Forecasts June 16, 2016, NOAA, Salt
Lake City, UT.
Allen, C. D. et al. (2010), A global overview of drought and heat-induced tree mortality reveals emerging climate
change risks for forests, For. Ecol. Manag., 259(4), 660–684, doi:10.1016/j.foreco.2009.09.001.
Allen, C. D., D. D. Breshears, and N. G. McDowell (2015), On underestimation of global vulnerability to tree
mortality and forest die-off from hotter drought in the Anthropocene, Ecosphere, 6(8), art129.
Ault, T. R., J. E. Cole, and S. St. George (2012), The amplitude of decadal to multidecadal variability in
precipitation simulated by state-of-the-art climate models, Geophys. Res. Lett., 39(21), n/a-n/a,
doi:10.1029/2012GL053424.
Ault, T. R., J. E. Cole, J. T. Overpeck, G. T. Pederson, S. St. George, B. Otto-Bliesner, C. A. Woodhouse, and C.
Deser (2013), The Continuum of Hydroclimate Variability in Western North America during the Last
Millennium, J. Clim., 26(16), 5863–5878, doi:10.1175/JCLI-D-11-00732.1.
Ault, T. R., J. E. Cole, J. T. Overpeck, G. T. Pederson, and D. M. Meko (2014), Assessing the Risk of Persistent
Drought Using Climate Model Simulations and Paleoclimate Data, J. Clim., 27(20), 7529–7549,
doi:10.1175/JCLI-D-12-00282.1.
Ault, T. R., J. S. Mankin, B. I. Cook, and J. E. Smerdon (2016), Relative impacts of mitigation, temperature, and
precipitation on 21st-century megadrought risk in the American Southwest, Sci. Adv., 2(10), e1600873,
doi:10.1126/sciadv.1600873.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Ayers, J., D. L. Ficklin, I. T. Stewart, and M. Strunk (2016), Comparison of CMIP3 and CMIP5 projected
hydrologic conditions over the Upper Colorado River Basin, Int. J. Climatol., 36(11), 3807–3818,
doi:10.1002/joc.4594.
Barnett, T. P., and D. W. Pierce (2008), When will Lake Mead go dry?, Water Resour. Res., 44(3), W03201,
doi:10.1029/2007WR006704.
Barnett, T. P., and D. W. Pierce (2009), Sustainable water deliveries from the Colorado River in a changing climate,
Proc. Natl. Acad. Sci., 106(18), 7334–7338.
Belmecheri, S., F. Babst, E. R. Wahl, D. W. Stahle, and V. Trouet (2015), Multi-century evaluation of Sierra
Nevada snowpack, Nat. Clim. Change.
Betts, A. K., J. H. Ball, A. C. M. Beljaars, M. J. Miller, and P. A. Viterbo (1996), The land surface-atmosphere
interaction: A review based on observational and global modeling perspectives, J. Geophys. Res.
Atmospheres, 101(D3), 7209–7225, doi:10.1029/95JD02135.
Brekke, L., B. Harding, T. Piechota, B. Udall, C. Woodhouse, and D. Yates (2007), Review of Science and Methods
for Incorporating Climate Change Information into Reclamation’s Colorado River Basin Planning Studies,
Bureau of Reclamation.
Brekke, L., B. Thrasher, Maurer, Edwin P., and T. Pruitt (2013), Downscaled CMIP3 and CMIP5 Climate
Projections, Reclamation.
Brekke, L., A. Wood, and T. Pruitt (2014), Downscaled CMIP3 and CMIP5 Hydrology Projections, Reclamation.
Breshears, D. D. et al. (2005), Regional vegetation die-off in response to global-change-type drought, Proc. Natl.
Acad. Sci. U. S. A., 102(42), 15144–15148.
Brubaker, K. L., and D. Entekhabi (1996), Analysis of feedback mechanisms in land-atmosphere interaction, Water
Resour. Res., 32(5), 1343–1357.
Cayan, D. R., T. Das, D. W. Pierce, T. P. Barnett, M. Tyree, and A. Gershunov (2010), Future dryness in the
southwest US and the hydrology of the early 21st century drought, Proc. Natl. Acad. Sci., 107(50), 21271–
21276, doi:10.1073/pnas.0912391107.
Central Arizona Project (2016), Lower Basin Drought Contingency Plan, Central Arizona Project.
Christensen, N. S., and D. P. Lettenmaier (2007), A multimodel ensemble approach to assessment of climate change
impacts on the hydrology and water resources of the Colorado River Basin, Hydrol. Earth Syst. Sci.
Discuss., 11(4), 1417–1434.
Christensen, N. S., A. W. Wood, N. Voisin, D. P. Lettenmaier, and R. N. Palmer (2004), The effects of climate
change on the hydrology and water resources of the Colorado River basin, Clim. Change, 62(1–3), 337–
363.
Collins, M. et al. (2013), Long-term climate change: projections, commitments and irreversibility, in Climate
Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report
of the Intergovernmental Panel on Climate Change, edited by 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, Cambridge University
Press, Cambridge, United Kingdom and New York, NY, USA.
Collum, C., and T. McCann (2014), Central Arizona Project Board Report Agenda Number 9, Central Arizona
Project.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Colorado River Basin Stakeholders (2015), Colorado River Basin Stakeholders Moving Forward to Address
Challenges Identified in the Colorado River Basin Water Supply and Demand Study, Phase 1 Report.
Cook, B. I., J. E. Smerdon, R. Seager, and S. Coats (2014), Global warming and 21st century drying, Clim. Dyn.,
43(9–10), 2607–2627.
Cook, B. I., T. R. Ault, and J. E. Smerdon (2015), Unprecedented 21st century drought risk in the American
Southwest and Central Plains, Sci. Adv., 1(1), e1400082, doi:10.1126/sciadv.1400082.
Dai, A. (2012), Increasing drought under global warming in observations and models, Nat. Clim. Change, 3(1), 52–
58, doi:10.1038/nclimate1633.
Daly, C., R. P. Neilson, and D. L. Phillips (1994), A statistical-topographic model for mapping climatological
precipitation over mountainous terrain, J. Appl. Meteorol., 33(2), 140–158.
Das, T., D. W. Pierce, D. R. Cayan, J. A. Vano, and D. P. Lettenmaier (2011), The importance of warm season
warming to western U.S. streamflow changes, Geophys. Res. Lett., 38(23), L23403,
doi:10.1029/2011GL049660.
Davis, T. (2016), Big CAP cuts coming as 3-state water agreement nears, Ariz. Dly. Star. Available from:
http://tucson.com/news/local/big-cap-cuts-coming-as--state-water-agreement-nears/article_876e3aa6-6cf053ec-bd0c-95be8c6468ae.html (Accessed 6 June 2016)
Deems, J. S., T. H. Painter, J. J. Barsugli, J. Belnap, and B. Udall (2013), Combined impacts of current and future
dust deposition and regional warming on Colorado River Basin snow dynamics and hydrology, Hydrol
Earth Syst Sci, 17(11), 4401–4413, doi:10.5194/hess-17-4401-2013.
Department of Interior (2007), Record of Decision Colorado River Interim Guidelines for Lower Basin Shortages
and the Coordinated Operations for Lake Powell and Lake Mead, Department of Interior, Washington,
D.C.
Diffenbaugh, N. S., D. L. Swain, and D. Touma (2015), Anthropogenic warming has increased drought risk in
California, Proc. Natl. Acad. Sci., 112(13), 3931–3936, doi:10.1073/pnas.1422385112.
Elias, E. H., A. Rango, C. M. Steele, J. F. Mejia, and R. Smith (2015), Assessing climate change impacts on water
availability of snowmelt-dominated basins of the Upper Rio Grande basin, J. Hydrol. Reg. Stud., 3, 525–
546, doi:10.1016/j.ejrh.2015.04.004.
Feng, S., and Q. Fu (2013), Expansion of global drylands under a warming climate, Atmospheric Chem. Phys.,
13(19), 10081–10094, doi:10.5194/acp-13-10081-2013.
Ficklin, D. L., I. T. Stewart, and E. P. Maurer (2013), Climate Change Impacts on Streamflow and Subbasin-Scale
Hydrology in the Upper Colorado River Basin, edited by V. Shah, PLoS ONE, 8(8), e71297,
doi:10.1371/journal.pone.0071297.
Glennon, R. J. (1995), Coattails of the Past: Using and Financing the Central Arizona Project, Ariz St LJ, 27, 677–
756.
Griffin, D., and K. J. Anchukaitis (2014), How unusual is the 2012-2014 California drought?, Geophys. Res. Lett.,
41(24), 9017–9023, doi:10.1002/2014GL062433.
Guentchev, G., J. J. Barsugli, and J. Eischeid (2010), Homogeneity of Gridded Precipitation Datasets for the
Colorado River Basin, J. Appl. Meteorol. Climatol., 49(12), 2404–2415, doi:10.1175/2010JAMC2484.1.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Harding, B. L., A. W. Wood, and J. R. Prairie (2012), The implications of climate change scenario selection for
future streamflow projection in the Upper Colorado River Basin, Hydrol. Earth Syst. Sci., 16(11), 3989–
4007, doi:10.5194/hess-16-3989-2012.
James, I. (2016), Lake Mead declines to lowest level in history, Desert Sun. Available from:
http://www.desertsun.com/story/news/environment/2016/05/19/lake-mead-declines-new-recordlow/84597120/ (Accessed 6 June 2016)
Johnson, R. (1977), The central Arizona project, 1918-1968, University of Arizona Press, Tucson.
Koirala, S., Y. Hirabayashi, R. Mahendran, and S. Kanae (2014), Global assessment of agreement among
streamflow projections using CMIP5 model outputs, Environ. Res. Lett., 9(6), 64017, doi:10.1088/17489326/9/6/064017.
Liang, X., D. P. Lettenmaier, and E. F. Wood (1996), One-dimensional statistical dynamic representation of subgrid
spatial variability of precipitation in the two-layer variable infiltration capacity model, J. Geophys. Res.
Atmospheres, 101(D16), 21403–21422, doi:10.1029/96JD01448.
Lukas, J., J. Barsugli, N. Doesken, I. Rangwala, and K. Wolter (2014), Climate Change in Colorado A Synthesis to
Support Water Resources Management and Adaptation, Western Water Assessment.
Mann, M. E., and P. H. Gleick (2015), Climate change and California drought in the 21st century, Proc. Natl. Acad.
Sci., 112(13), 3858–3859, doi:10.1073/pnas.1503667112.
Mann, M. E., Z. Zhang, S. Rutherford, R. S. Bradley, M. K. Hughes, D. Shindell, C. Ammann, G. Faluvegi, and F.
Ni (2009), Global Signatures and Dynamical Origins of the Little Ice Age and Medieval Climate Anomaly,
Science, 326(5957), 1256–1260, doi:10.1126/science.1177303.
McCabe, G. J., and D. M. Wolock (2007), Warming may create substantial water supply shortages in the Colorado
River basin, Geophys. Res. Lett., 34(22), L22708, doi:10.1029/2007GL031764.
Meehl, G. A., C. Covey, K. E. Taylor, T. Delworth, R. J. Stouffer, M. Latif, B. McAvaney, and J. F. B. Mitchell
(2007), THE WCRP CMIP3 Multimodel Dataset: A New Era in Climate Change Research, Bull. Am.
Meteorol. Soc., 88(9), 1383–1394, doi:10.1175/BAMS-88-9-1383.
Meko, D. M., C. A. Woodhouse, C. A. Baisan, T. Knight, J. J. Lukas, M. K. Hughes, and M. W. Salzer (2007),
Medieval drought in the upper Colorado River Basin, Geophys. Res. Lett., 34(10).
Milly, P. C. D., and K. A. Dunne (2016), Potential evapotranspiration and continental drying, Nat. Clim. Change,
6(10), 946–949, doi:10.1038/nclimate3046.
Milly, P. C. D., K. A. Dunne, and A. V. Vecchia (2005), Global pattern of trends in streamflow and water
availability in a changing climate, Nature, 438(7066), 347–350, doi:10.1038/nature04312.
Moss, R. H. et al. (2010), The next generation of scenarios for climate change research and assessment, Nature,
463(7282), 747–756.
Nash, L. L., and P. H. Gleick (1991), Sensitivity of streamflow in the Colorado basin to climatic changes, J. Hydrol.,
125(3), 221–241.
National Research Council (2007), Colorado River Basin Water Management: Evaluating and Adjusting to
Hydroclimatic Variability, The National Academies Press, Washington, D.C.
Norris, J. R., R. J. Allen, A. T. Evan, M. D. Zelinka, C. W. O’Dell, and S. A. Klein (2016), Evidence for climate
change in the satellite cloud record, Nature, 536(7614), 72–75, doi:10.1038/nature18273.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Nowak, K., M. Hoerling, B. Rajagopalan, and E. Zagona (2012), Colorado River Basin Hydroclimatic Variability, J.
Clim., 25(12), 4389–4403, doi:10.1175/JCLI-D-11-00406.1.
Overpeck, J. T. (2013), Climate science: The challenge of hot drought, Nature, 503(7476), 350–351.
Oyler, J. W., S. Z. Dobrowski, A. P. Ballantyne, A. E. Klene, and S. W. Running (2015a), Artificial amplification of
warming trends across the mountains of the western United States, Geophys. Res. Lett., 42(1), 153–161.
Oyler, J. W., A. Ballantyne, K. Jencso, M. Sweet, and S. W. Running (2015b), Creating a topoclimatic daily air
temperature dataset for the conterminous United States using homogenized station data and remotely
sensed land skin temperature, Int. J. Climatol., 35(9), 2258–2279, doi:10.1002/joc.4127.
Painter, T. H., J. S. Deems, J. Belnap, A. F. Hamlet, C. C. Landry, and B. Udall (2010), Response of Colorado River
runoff to dust radiative forcing in snow, Proc. Natl. Acad. Sci., 107(40), 17125–17130.
Pitman, A. J. (2003), The evolution of, and revolution in, land surface schemes designed for climate models, Int. J.
Climatol., 23(5), 479–510, doi:10.1002/joc.893.
Prein, A. F., G. J. Holland, R. M. Rasmussen, M. P. Clark, and M. R. Tye (2016), Running dry: The U.S.
Southwest’s drift into a drier climate state, Geophys. Res. Lett., 2015GL066727,
doi:10.1002/2015GL066727.
R Development Core Team (2008), R: A Language and Environment for Statistical Computing, R Foundation for
Statistical Computing, Vienna, Austria.
Rajagopalan, B., K. Nowak, J. Prairie, M. Hoerling, B. Harding, J. Barsugli, A. Ray, and B. Udall (2009), Water
supply risk on the Colorado River: Can management mitigate?, Water Resour. Res., 45(8).
Rangwala, I., T. Bardsley, M. Pescinski, and J. Miller (2015), SNOTEL sensor upgrade has caused temperature
record inhomogeneities for the Intermountain West: Implications for climate change impact assessments,
Western Water Assessment, Boulder, CO.
Ray, A. J., J. J. Barsugli, K. B. Averyt, K. Wolter, M. Hoerling, N. Doesken, B. Udall, and R. S. Webb (2008),
Climate change in Colorado: a synthesis to support water resources management and adaptation, CU NOAA Western Water Assessment.
Reclamation (2007), Final Environmental Impact Statement Colorado River Interim Guidelines for Lower Basin
Shortages and Coordinated Operations for Lake Powell and Lake Mead,
Reclamation (2011), West-Wide Climate Risk Assessments: Bias-Corrected and Spatially Downscaled Surface
Water Projections, U.S. Department of the Interior, Bureau of Reclamation, Technical Services Center,
Denver, Colorado.
Reclamation (2012), Colorado River Basin Water Supply and Demand Study, U.S. Department of the Interior,
Bureau of Reclamation, Technical Services Center, Denver, Colorado.
Reclamation (2013), West-Wide Climate Risk Assessment: Upper Rio Grande Impact Assessment, U.S. Department
of the Interior, Bureau of Reclamation.
Reclamation (2016), SECURE Water Act Section 9503(c) – Reclamation Climate Change and Water., U.S.
Department of the Interior, Bureau of Reclamation.
Salzer, M. W., and K. F. Kipfmueller (2005), Reconstructed temperature and precipitation on a millennial timescale
from tree-rings in the southern Colorado Plateau, USA, Clim. Change, 70(3), 465–487.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Salzer, M. W., A. G. Bunn, N. E. Graham, and M. K. Hughes (2014), Five millennia of paleotemperature from treerings in the Great Basin, USA, Clim. Dyn., 42(5–6), 1517–1526.
Scheff, J., and D. M. Frierson (2012), Robust future precipitation declines in CMIP5 largely reflect the poleward
expansion of model subtropical dry zones, Geophys. Res. Lett., 39(18).
Seager, R. et al. (2007), Model Projections of an Imminent Transition to a More Arid Climate in Southwestern
North America, Science, 316(5828), 1181–1184, doi:10.1126/science.1139601.
Seager, R., M. Ting, C. Li, N. Naik, B. Cook, J. Nakamura, and H. Liu (2012), Projections of declining surfacewater availability for the southwestern United States, Nat. Clim. Change, 3(5), 482–486,
doi:10.1038/nclimate1787.
Seager, R., M. Hoerling, S. Schubert, H. Wang, B. Lyon, A. Kumar, J. Nakamura, and N. Henderson (2015a),
Causes of the 2011–14 California Drought*, J. Clim., 28(18), 6997–7024.
Seager, R., A. Hooks, A. P. Williams, B. Cook, J. Nakamura, and N. Henderson (2015b), Climatology, Variability,
and Trends in the US Vapor Pressure Deficit, an Important Fire-Related Meteorological Quantity*, J. Appl.
Meteorol. Climatol., 54(6), 1121–1141.
Seneviratne, S. I., T. Corti, E. L. Davin, M. Hirschi, E. B. Jaeger, I. Lehner, B. Orlowsky, and A. J. Teuling (2010),
Investigating soil moisture–climate interactions in a changing climate: A review, Earth-Sci. Rev., 99(3),
125–161.
Solomon, S., G.-K. Plattner, R. Knutti, and P. Friedlingstein (2009), Irreversible climate change due to carbon
dioxide emissions, Proc. Natl. Acad. Sci., 106(6), 1704–1709, doi:10.1073/pnas.0812721106.
Swann, A. L. S., F. M. Hoffman, C. D. Koven, and J. T. Randerson (2016), Plant responses to increasing CO 2
reduce estimates of climate impacts on drought severity, Proc. Natl. Acad. Sci., 201604581,
doi:10.1073/pnas.1604581113.
Taylor, K. E., R. J. Stouffer, and G. A. Meehl (2012), An Overview of Cmip5 and the Experiment Design, Bull. Am.
Meteorol. Soc., 93(4), 485–498.
Tipton and Kalmbach, Inc (1965), Water Supplies of the Colorado River, Upper Colorado River Commission,
Denver, CO.
Udall, J. (1983), Floods Reveal Water Policy Chaos, High Ctry. News, 15(18).
Ukkola, A. M., I. C. Prentice, T. F. Keenan, A. I. J. M. van Dijk, N. R. Viney, R. B. Myneni, and J. Bi (2015),
Reduced streamflow in water-stressed climates consistent with CO2 effects on vegetation, Nat. Clim.
Change, 6(1), 75–78, doi:10.1038/nclimate2831.
Vano, J. A., and D. P. Lettenmaier (2014), A sensitivity-based approach to evaluating future changes in Colorado
River discharge, Clim. Change, 122(4), 621–634, doi:10.1007/s10584-013-1023-x.
Vano, J. A., T. Das, and D. P. Lettenmaier (2012), Hydrologic Sensitivities of Colorado River Runoff to Changes in
Precipitation and Temperature*, J. Hydrometeorol., 13(3), 932–949, doi:10.1175/JHM-D-11-069.1.
Vano, J. A. et al. (2014), Understanding Uncertainties in Future Colorado River streamflow, Bull. Am. Meteorol.
Soc., 95(1), 59–78.
Verburg, K. O. (2011), Colorado River Documents 2008, Reclamation, Denver, Colo.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Weiss, J. L., C. L. Castro, and J. T. Overpeck (2009), Distinguishing Pronounced Droughts in the Southwestern
United States: Seasonality and Effects of Warmer Temperatures, J. Clim., 22(22), 5918–5932,
doi:10.1175/2009JCLI2905.1.
Williams, A. P., R. Seager, J. T. Abatzoglou, B. I. Cook, J. E. Smerdon, and E. R. Cook (2015), Contribution of
anthropogenic warming to California drought during 2012-2014, Geophys. Res. Lett., 42(16), 6819–6828,
doi:10.1002/2015GL064924.
Wines, M. (2014), Colorado River Drought Forces a Painful Reckoning for States, N. Y. Times, 5th January.
Woodhouse, C. A., S. T. Gray, and D. M. Meko (2006), Updated streamflow reconstructions for the Upper Colorado
River basin, Water Resour. Res., 42(5).
Woodhouse, C. A., G. T. Pederson, K. Morino, S. A. McAfee, and G. J. McCabe (2016), Increasing influence of air
temperature on upper Colorado River streamflow, Geophys. Res. Lett., 2015GL067613,
doi:10.1002/2015GL067613.

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Table 1: Winter/Summer/Annual Upper Basin Mean Water Year Precipitation
1953-1967

2000-2014

1896-2014

mm

mm

mm

total

Anomaly

Anomaly %
of mean

Winter (Oct-Mar)

176

-16

-8.6%

Summer (Apr-Sept)

184

-7

-3.6%

179

-12

-6.4%

191

100%

Total

359

-23

-6.1%

365

-17

-4.6%

383

100%

total
187

Anomaly

Anomaly % of
mean

Mm

% Avg

-5

-2.7%

192

100%

Table 2: Upper Basin Water Year Flows and Temperatures
Average Annual Flow

Average Annual Temperature

Period

bcm

% 1906-1999

Deg. C

°C Anomaly to
1896-1999

1953-1967

15.38

81.9%

7.0

0.2

2000-2014

15.15

80.7%

7.7

0.9

1906-1999

18.77

100.0%

6.8

0.0

1906-2014

18.27

97.3%

6.9

0.1

This article is protected by copyright. All rights reserved.

 Confidential manuscript submitted to Water Resources Research

Figure Captions
Figure 1. Map of the Colorado River Basin. Lower and Upper Basins, major U.S. cities receiving
Colorado River water, major tributaries, and Lakes Mead and Powell are shown. The Central
Arizona Project canal in red.
Figure 2. (a) Lakes Mead and Powell combined monthly contents. Upper Basin annual Colorado
River (b) runoff at Lees Ferry from 1906-2014, (c) precipitation and (d) temperatures from 18962014. Mead first filled in 1935, Powell in 1963 (Text S1). Two fifteen-year drought periods,
1953-1967 and 2000-2014, are highlighted and discussed in main text.
Figure 3. Probability Density Functions of Upper Colorado River Basin temperature projections
for mid-century and end-century under moderate (SRES A1B and RCP4.5) and high (SRES A2
and RCP8.5) emissions.
Figure 4. Probability Density Functions of Upper Colorado River Basin temperature-induced
flow reductions for mid-century and end-century with the three temperature sensitivities (-3%, 6.5%, -10%) and the two levels of emissions (Moderate: SRES A1B and RCP4.5 and High:
SRES A2 and RCP8.5).
Figure 5. Temperature-induced flow losses by model run (one per dot) with temperature
increases shown on horizontal axis. For each period (Mid-Century, End-Century) and emissions
type (Moderate, High), flow losses for each model run are shown with the 3 (Low = -3%/°C,
Medium = -6.5%/ °C, High = -10%/°C) temperature sensitivities. Black dots/circles are
averages/medians for each sensitivity. Precipitation increases needed to counteract flow losses at
right are based on 2.5 precipitation elasticity. Range for the temperature-induced losses during
2000-2014 drought are shown in shaded brown at the top (Text S5).

This article is protected by copyright. All rights reserved.

 

Accepted Articl

This article is protected by copyright. All rights reserved.

44
42

Cheyenne

●

40

Salt Lake City ●
Provo ●
●

Denver

●

38

●

Colorado Springs
Pueblo

Lake Powell

36

Las Vegas

●

Lake Mead
●

● Santa Fe
Albuquerque

34

Los Angeles
●

San Diego

●

Phoenix

32

●

Tucson

30

●

−125

−120

−115

−110

−105

This article is protected by copyright. All rights reserved.

−100

 

Accepted Articl

This article is protected by copyright. All rights reserved.

50
40
maf

2000s
Temperature
Dominated
Drought

1950s
Precipitation
Dominated
Drought

a − Mead + Powell Volume

30
20
10

25

Only Lake Mead

L. Powell Fills

Average

Linear Trend

Both Reservoirs

b − Upper Basin Natural Flows

maf

20

15

10

5
20

c − Upper Basin Precipitation

inches

18
16
14
12
10
48

d − Upper Basin Temperatures

°F

46

44

42
Annual Data

1900

1910

1920

This article is protected by copyright. All rights reserved.

1930

1940

1950

1960

1970

1980

Loess Smooth

1990

2000

2010

 

Accepted Articl

This article is protected by copyright. All rights reserved.

0.5
0.0

0.1

0.2

0.3

0.4

Mid−Century Moderate Emissions
Mid−Century High Emissions
End−Century Moderate Emissions
End−Century High Emissions

°C

2

4

6

Temperature Increase °C

This article is protected by copyright. All rights reserved.

8

10

 

Accepted Articl

This article is protected by copyright. All rights reserved.

Mid−Century

4e−07
0e+00

Density

8e−07

−3% /°C Moderate Emissions
−3% /°C High Emissions
−6.5% /°C Moderate Emissions
−6.5% /°C High Emissions
−10% /°C Moderate Emissions
−10% /°C High Emissions

0%

−10%

−20%

−30%

−40%

−50%

−60%

−70%

−80%

Flow Reductions

End−Century

4e−07
2e−07
0e+00

Density

6e−07

−3% /°C Moderate Emissions
−3% /°C High Emissions
−6.5% /°C Moderate Emissions
−6.5% /°C High Emissions
−10% /°C Moderate Emissions
−10% /°C High Emissions

0%

−10%

−20%

−30%

−40%

−50%

−60%

Flow Reductions

This article is protected by copyright. All rights reserved.

−70%

−80%

 

Accepted Articl

This article is protected by copyright. All rights reserved.

0%
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
● ●
●
●
● ●
● ●
●
●
●
●
● ●
● ●
●
●
●
● ●
●
●
●
● ●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●

●
●

−10%

−15%

2000−2014
Loss Range

●
●

−20%

−25%

−30%

●
●

●

●
●

●

●

●
●
●
●
●
●
●
●
●
●
●
●
●
●

●

●

●
●

●
●
●
●

●
●

●
●
●
●
●

●
●
●
●
●
●
●
●
●
●
●
●

●

−40%

●
●
●

●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●

●
●
●
●
●
●
●
●
●
●
●
●

●

●

●
●
●
●

●
●
●
●
●
●
●
●
●
●
●

●
●
●

●
●

+14%

●
●

●
●
●

+12%

●
●
●
●
●
●
●

●
●
●
●
●
●
●

●

+16%

●

●

●
●
●
●
●
●
●

●
●
●
●
●
●
●
●

+18%

●
●
●
●
●

●

●
●

−55%

+10%
●

●

●
●

●
●

−50%

+8%

●

●
●
●

●

+6%

●

●
●
●

●

−45%

●
●

●
●

●
●
●
●

●
●

+4%

●
●

●
●

●
●
●

●
●
●
●

●
●
●
●
●
●
●
●
●
●
●
●
●

●
●

●

●

+2%

●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●

●
●
● ●
● ●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●

●
●

−35%

●

●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
● ●

+20%

●

●

●
●
●

●

+22%

●
●
●
●
●
●
●

●

●

●
●

−65%

●
●

●

−60%
●

+24%

●
●
●

●

Low Sensitivity
Medium Sensitivity
High Sensitivity

●

●
●
●

+26%

●
●
●
●

●

−70%

+28%

●
●
●
●
●

−75%

−80%

+30%

Mid−Century
Moderate
Emissions

Mid−Century
High
Emissions

End−Century
Moderate
Emissions

End−Century
High
Emissions
●

+32%

●

2

4

6

8

10

2

4

6

8

10

2

4

6

8

10

Temperature Increase °C
This article is protected by copyright. All rights reserved.

2

4

6

8

10
●

Precipitation Required to Equal Flow Loss

−5%

0%