Southern Illinois University When Will Lake Mead Go Dry Read the Article by Barnet, T.P. and D.W. Price, 2008. When will Lake Mead go Dry? Journal of Water
Southern Illinois University When Will Lake Mead Go Dry Read the Article by Barnet, T.P. and D.W. Price, 2008. When will Lake Mead go Dry? Journal of Water Resources Research, vol. 44, W03201. Your critique should:
1) Reference the article at the top of the critique
2) Briefly summarize the contents of the article
3) State major assumptions and contribution
4) Does the article has any limitations?
5) Do you agree with the results?
6) Describe the relevance of this material to CE 301.
Please answer these questions with your own words otherwise I will be penalized if I just copy from the article. The article is short ( 10 Pages) and I will attach it here as PDF. Click
Here
WATER RESOURCES RESEARCH, VOL. 44, W03201, doi:10.1029/2007WR006704, 2008
for
Full
Article
When will Lake Mead go dry?
Tim P. Barnett1 and David W. Pierce1
Received 27 November 2007; revised 22 January 2008; accepted 5 February 2008; published 29 March 2008.
[1] A water budget analysis shows that under current conditions there is a 10% chance
that live storage in Lakes Mead and Powell will be gone by about 2013 and a 50% chance
that it will be gone by 2021 if no changes in water allocation from the Colorado River
system are made. This startling result is driven by climate change associated with
global warming, the effects of natural climate variability, and the current operating status
of the reservoir system. Minimum power pool levels in both Lake Mead and Lake Powell
will be reached under current conditions by 2017 with 50% probability. While these dates
are subject to some uncertainty, they all point to a major and immediate water supply
problem on the Colorado system. The solutions to this water shortage problem must be
time-dependent to match the time-varying, human-induced decreases in future river flow.
Citation: Barnett, T. P., and D. W. Pierce (2008), When will Lake Mead go dry?, Water Resour. Res., 44, W03201,
doi:10.1029/2007WR006704.
1. Introduction
[2] A number of studies over the last 20 years have
suggested that there will be a decrease in runoff over
the Southwestern United States because of global warming.
The decrease will be caused by increasing temperatures
and evapotranspiration and decreasing precipitation. The
statistical/empirical studies [Revelle and Waggoner, 1983;
Nash and Gleick, 1991, 1993; Hoerling and Eischeid, 2007],
as well as climate model studies of the last few years [e.g.,
Milly et al., 2005; Christensen et al., 2004, Christensen and
Lettenmaier, 2006; Seager et al., 2007] all show a decrease
in runoff to the Colorado River (see caveats on climate
models below). The estimates of runoff reduction from
these studies are remarkably similar, and range between
10% and 30% over the next 30 50 years. The IPCC
Working Group II concludes there will be a 10 30% run
off reduction over some dry regions at midlatitudes during
the next 50 years with very high confidence [Intergovernmental Panel on Climate Change, 2008]. Current naturalized flow in the Colorado River is on the order of 15 million
acre feet (MAF, 1.233 109 m3) per year measured at Lees
Ferry (Figure 1), so these decreases will ultimately result in
a runoff reduction of 1.5 4.5 MAF/a from current levels,
which we assume leads to similar reductions in Colorado
River flow.
[3] The Colorado River is quite literally the lifes blood
of todays modern southwest society and economy. Given
the agreement about both size and timing of runoff reduction, it is important to examine what it will mean to the
people of the southwest and, especially, when they might
expect water shortage problems to appear. In its recent
report on Colorado River Basin water management, the
National Academy of Sciences [Committee on the Scientific
1
Scripps Institution of Oceanography, University of California, San
Diego, La Jolla, California, USA.
Copyright 2008 by the American Geophysical Union.
0043-1397/08/2007WR006704$09.00
Bases of Colorado River Basin Water Management, 2007]
notes future potential problems with availability of water in
the region. It calls for a comprehensive analysis of water
needs and uses in the region, but provides no analysis of the
timing or magnitude of potential problems. Hoerling and
Eischeid [2007] suggest water availability could soon fall
below critical levels but offer no temporal details. McCabe
and Wolock [2007] estimate climate changes will increase
chances of failure to meet water allocation requirements of
the Colorado Covenant, but their methods preclude estimates of just when this might happen.
[4] Our intent is to make a first estimate of when and how
the human-induced reduced runoff will impact people. We
simplistically state the question as when will Lake Mead
go dry? assuming there are no changes in water management strategies and sector-specific consumptive use. By
going dry, we mean when the live storage (the reservoir
space from which water can be evacuated by gravity) in
Lakes Mead and Powell becomes exhausted (Figure 2
summarizes the various storage levels in the Lakes). As we
shall see below, the answer is both startling and alarming.
[5] It is obvious that once long-term outflow exceeds
inflow the system is doomed to run dry. One of our
purposes in this work is to point out that currently scheduled
depletions (loss of water from consumptive use), along with
water losses due to evaporation/infiltration and reduction in
runoff due to climate change, have pushed the system into a
negative net inflow regime that is not sustainable. Another
purpose is to demonstrate how natural variability, i.e., the
chance of getting strings of dry years consistent with the
historical record, makes the system likely to run dry even
with positive net inflow. When expected changes due to
global warming are included as well, currently scheduled
depletions are simply not sustainable.
2. Methods
2.1. Water Balance Model
[6] The method is a simple water balance approach that
keeps track of water going into and out of the major
W03201
1 of 10
W03201
BARNETT AND PIERCE: WHEN WILL LAKE MEAD GO DRY?
Figure 1. Overview of the region of interest (31.2°
43.7°N, 104.0° 120.3°W), which is historically separated
into the upper basin (dots) and lower basin (gray).
Colorado River flow from the upper to lower basins is
measured at Lees Ferry.
reservoirs in the Colorado River system. The initial condition
for our study (Figure 2) is the amount of water currently
in live storage in the Lake Mead/Lake Powell system
(25.7 MAF above the dead pool as of June 2007; U.S.
Bureau of Reclamation Web page). We consider the two
reservoirs as a single storage unit, consistent with the U.S.
Bureau of Reclamation (USBR) plan to manage them
jointly [U.S. Bureau of Reclamation, 2007]. We assume
perfect management so that the amount of storage in each
W03201
reservoir above dead pool is manipulated to keep the storage
levels approximately the same in both reservoirs (see
caveats). The naturalized flow of the Colorado River at Lees
Ferry is 15 MAF/a over the period 1906 2005 (USBR Web
page, http://www.usbr.gov/lc/region/g4000/NaturalFlow/
current.html, accessed 10 January 2008), so we use this as
a working number, although on the basis of tree ring
reconstructions it is probably too high [Committee on the
Scientific Bases of Colorado River Basin Water Management, 2007], and does not reflect the drought of the last
7 years (see caveats).
[7] Today the Colorado system is, for all intents and
purposes, fully subscribed (see below) so any additional
consumptive use in the upper basin as now contemplated
(Figure 3), or reduced runoff into the river due to climate
change, must be covered by existing storage. We consider
human-induced reductions in runoff of 10 to 30%, in
accordance with estimates from global climate models and
statistical analysis, and take these reductions to be linear in
time over the next 50 years (i.e., runoff slowly decreases
until it reaches a total reduction of, say, 10% below current
levels in 2057). We first do a simple deterministic analysis
that does not include the complicating factors of runoff
variability, evaporation, and infiltration, in order to more
clearly isolate the effect of human-induced climate change
on the reservoirs. We then do a probabilistic analysis of the
likelihood of the reservoir storage becoming exhausted,
using Monte Carlo simulations with a water budget model,
and allowing for evaporation and infiltration as well as the
stochastic nature of the river flow itself.
[8] We tested the water budget model by comparing it to
the results obtained by Harding et al. [1995], who modeled
a severe sustained drought episode on the Colorado River
using a sophisticated river network model based on an
enhanced version of USBRs Colorado River model, CRSS.
The results (Figure 4) show the simulated, combined storage
from Harding et al. [1995] versus that from the water
Figure 2. Total reservoir storage in Lakes Mead and Powell (million acre feet) as a function of lake
surface elevation above mean sea level (feet). (We retain the units commonly used in the operation of
these reservoirs; data are from Colorado River Open Source Simulator, release 1.0, 2007, http://
www.onthecolorado.org/cross.cfm). Arrows indicate the maximum storage possible in each lake, the
amount present on 13 June 2007, the minimum needed to enable hydroelectric power generation, and the
minimum below which no more water can be extracted from the reservoir by gravity (dead pool).
Live storage is all current storage above the dead pool elevation.
2 of 10
W03201
BARNETT AND PIERCE: WHEN WILL LAKE MEAD GO DRY?
W03201
Figure 3. Historical water use (solid line) and scheduled future depletions (dashed line, 2008 2060) of
the Colorado River system. Superposed lines for the upper and lower basins show the best fit least
squares linear trend over the period 1960 2004. Note the abrupt change in water availability for the
lower basin states.
budget analysis used here. The differences are due principally to our neglect of smaller storage units within the
Colorado system. At any rate, the agreement suggests the
method is adequate to address the large-scale water budget
issues considered here.
[9] We tried three different methods to generate synthetic
time series of Colorado River flow consistent with the
historical record (Appendix A), including a simple firstorder autoregressive (AR-1) approximation, fractional
Gaussian noise (fGn), and a new Fourier-based technique
described in Appendix A. Overall, our results are robust
with respect to the method used, as the water budget effects
are large compared to differences in detail of the synthetic
flows. The plots shown here are made using fGn, since the
more familiar index sequential method (ISM) does not
correctly sample variability consistent with the historical
record (see Appendix A). Synthetic time series generated
with fGn also exhibit long-term persistence, which has been
shown to be important for correctly simulating the statistics
of hydrological processes [e.g., Phatarfod, 1989; Pelletier
and Turcotte, 1997; Wang et al., 2007; Koutsoyiannis and
Montanari, 2007].
2.2. Future Depletions
[10] Future depletions are taken from published USBR
schedules (appendices C and D of U.S. Bureau of Reclamation [2007]) over the period 2008 2060. In Figure 3
these are compared to historical water use (obtained from
http://www.usbr.gov/lc/region/g4000/uses.html, accessed
14 November 2007). Total scheduled depletions rise from
13.5 MAF/a in 2008 to 14.1 MAF/a by 2030. We also include
in the Monte Carlo results water loss due to evaporation and
changes due to infiltration (the 1971 2004 average evaporation was 0.894 and 0.516 MAF/a for lakes Mead and
Powell, respectively, while infiltration was +0.005 and
0.312 MAF/a (N. Yoder, USBR, personal communication,
2007)). Although the amount of evaporation and infiltration
change with lake level, possibly providing a negative feedback as the lake area shrinks, evaporation is also likely to
increase in the future as temperatures warm, and infiltration is
a second-order quantity compared to the other mechanisms
included here. Accordingly, in this work we have simply kept
the value of evaporation/infiltration constant at 1.7 MAF/a.
As a sensitivity test, we tried scaling evaporation with Lake
surface area, and found it made little difference to our results;
human-induced reductions in runoff overwhelm the Lake
surface area-dependent changes in evaporation.
3. Results
[11] In section 3.1 we begin with deterministic estimates of
when the live storage will be depleted by global warming
driven runoff reductions alone, without the outside impacts
of evaporation and natural variability in the river flow. This
approach is simplistic but gives an immediate feel for the
scope of the climate change problem and how it relates to
reservoir storage. In section 3.2 we then extend the analysis
to more realistic, probabilistic estimates of the same quantities but allowing for the additional impacts of natural
climate variability on runoff, as well as the effects of
evaporation and infiltration. A summary of the factors
included in each calculation is shown in Table 1.
3.1. Deterministic Estimates
[12] The above noted climate models and statistical
studies projected decreases in runoff that can be used to
compute the future decline in river flow in MAF, year by
year. We start by assuming a current steady state where
Figure 4. Reconstruction of combined Lakes Powell and
Mead storage (MAF) during the sustained severe drought
episode of the late 1500s from Harding et al. [1995]
(crosses) and this study (circles).
3 of 10
BARNETT AND PIERCE: WHEN WILL LAKE MEAD GO DRY?
W03201
W03201
Table 1. Summary of Factors Included in the Various Calculationsa
Probabilistic
Estimates?
Evaporation
and
Infiltration
Included?
Given in
Terms of
Net Inflow?
Climate
Change
Included?
Management
Strategies
Considered?
Deplete to
Power Pool or
Dead Pool
Location of
Results
10% Chance
to Deplete
by Year
50% Chance
to Deplete
by Year
No
No
Yes
Yes
Yes
Yes
Yes
no
no
yes
yes
yes
yes
yes
no
no
no
no
yes
yes
no
yes
yes
yes
yes
no
yes
yes
no
no
no
no
no
no
yes
dead
power
dead
power
dead
dead
dead
section 3.1 (start)
section 3.1 (end)
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
NAb
NA
2014
2010
2014c
2013c
2025d
2036
2021
2028
2017
2028c
2021c
2048d
a
For simulations that include climate change, the quoted years are for a 20% reduction in runoff over the next 50 years.
NA means not applicable.
For a net inflow of 1.0 MAF/a.
d
For a cut in requested water deliveries by 25%.
b
c
inflow to the reservoirs is equal to their discharge. In reality
the Lake Mead is currently being overdrafted by about
1 MAF (T. Labonde and J. Shields, Update for Green River
Basin Advisory Group, 2004, available at http://waterplan.
state.wy.us/BAG/green/briefbook), so our assumption of
steady state is highly conservative. We simply integrate
the annual reductions in runoff in time, assuming the
changes are temporally linear and levels of consumption
are constant, to determine how many years until the existing
live storage is gone. We find live storage will be depleted
completely 23 40 years from now, or sometime in the
span 2030 to 2047, for runoff reductions of 30 10% over
50 years, respectively.
[13] For further discussion, we take the median runoff
reduction, from the above studies, as 0.06 MAF per year.
This corresponds to a 20% decrease in runoff (3.0 MAF)
50 years from now, and yields approximately 29 years left,
or calendar year 2036, before the combined Mead and
Powell system is at dead pool elevation. Sensitivity studies
showed the dates vary by roughly 10 years around 2036 by
assuming larger/smaller 50 year runoff reduction rates or
that the 20% runoff reduction will happen soon/later than
2050. The time to dead pool elevation is not very sensitive
to the details and assumptions of the runoff estimates. One
can also vary the date depending on when one assumes the
warming impacts to set in. Recent studies show the global
warming impacts have been operative in the Southwest for
some decades [Barnett et al. 2008], but we make the
conservative assumption they start in 2007. Perhaps most
important are the initial conditions at the reservoirs for start
of the calculations; we used the current state as of June
2007. At this time the system had about 50% of its total
possible storage.
[14] In addition to water, both reservoirs are important
sources of hydroelectric power. Together the two reservoirs
can produce about 10,000 gW h. What do the runoff
reductions mean to the availability of that latter resource?
As of June 2007 there was a total, between both reservoirs,
of approximately 15 MAF of water above the minimum
power pool level, which is the reservoir elevation below
which the power generation turbines cannot safely operate
(Figure 2). Carrying through the same type of analysis as
above showed that there is a 50% chance the minimum
power pool elevation would be reached in around 2021;
only 14 years into the future. At that point (or before), there
would be an abrupt drop in the abilities of the reservoirs to
generate hydroelectric power.
3.2. Probabilistic Estimates
[15] The previous results neglected the natural variability
in river flow associated with weather (wet/dry years) and
short-term climate variability (e.g., El Nin?o/La Nin?a). Using
ten thousand realizations of river flow (statistically consistent with historic variability from 1906 2005 and tree ring
flow estimates over approximately the last 1250 years),
coupled with the deterministic linear runoff trend described
above, allowed us to construct cumulative distribution
functions (CDFs) for the depletion of the current live
storage. Future depletions were taken from the USBR
schedules shown in Figure 3, while evaporation plus infiltration was taken fixed at 1.7 MAF/a, as noted previously.
[16] The results are given in Figure 5 (left). The solid
curve shows the likelihood of reservoir storage levels falling
to the dead pool elevation with no runoff reduction. In the
absence of curtailed water delivery, there is a 50% chance
the system will go dry by 2037. This is driven by the sum of
depletions (14 MAF/a by 2030) plus evaporation/infiltration (1.7 MAF/a) being larger than runoff into the system
(15.05 MAF/a, the average over the period 1906 2005).
[17] Included also in Figure 5 (left) are the cases where
climate change decreases runoff into the river by 10%
(crosses) and 20% (circles). The probability of depleting
both reservoirs live storage is 50% by 2028, if we account
for natural variability and a 20% decrease in runoff (which
would be fully realized in 2057). The results are rather
insensitive to changes in runoff reduction. The different
methods of modeling the natural variability all give essentially the same results (Figure 5, right).
[18] All of these numbers are somewhat more pessimistic
than the deterministic analysis because they include evaporation/infiltration as well as allowing for natural variability
in the river flow. The answers, being expressed in probabilistic format, allow the user to determine the risk levels in
any decision process they undertake.
[19] The probabilistic analysis for minimum power pool
levels is shown in Figure 6. There is a 50% chance the
minimum power pool levels will be realized by about 2017,
in the absence of management responses. This result is
rather insensitive to changes in runoff, at least in the near
term. At any rate, the associated drops in power production
would be precipitous in time as turbine intakes went dry. It
4 of 10
W03201
BARNETT AND PIERCE: WHEN WILL LAKE MEAD GO DRY?
W03201
Figure 5. Cumulative distribution function (CDF) showing the probability of Lakes Mead and Powell
reservoir levels falling to dead pool elevation by the indicated year. (left) Case where only natural
variability is affecting river flow (solid curve) and cases where climate change produces a decrease in
runoff of 10% (curve with crosses) and 20% (curve with circles). (right) CDFs obtained with four
different methods of simulating natural runoff variability for the case with a 20% reduction in runoff.
ISM, index sequential method; AR-1, first-order autoregressive process; fGn, fractional Gaussian noise;
FRRP, Fourier reconstruction and randomized phase. See Appendix A for details.
seems clear that the threat to power produc…
Purchase answer to see full
attachment
