PUBLICATIONS
Geophysical Research Letters
RESEARCH LETTER
10.1002/2014GL062039
Key Points:
• Temperature increases during drought
but the interpretation is disputed
• Temperature increase during drought
is partly due to surface feedbacks
• Radiative signature of drought differs
from a direct global warming signature
Supporting Information:
• Readme
• Figure S1
• Figure S2
• Figure S3
• Figure S4
• Figure S5
• Table S1
• Table S2
Correspondence to:
M. L. Roderick,
[email protected]
Citation:
Yin, D., M. L. Roderick, G. Leech, F. Sun,
and Y. Huang (2014), The contribution
of reduction in evaporative cooling to
higher surface air temperatures during
drought, Geophys. Res. Lett., 41, 7891–7897,
doi:10.1002/2014GL062039.
Received 30 SEP 2014
Accepted 23 OCT 2014
Accepted article online 28 OCT 2014
Published online 21 NOV 2014
The copyright line for this article was
changed on 16 January 2015 after original online publication.
This is an open access article under the
terms of the Creative Commons
Attribution-NonCommercial-NoDerivs
License, which permits use and distribution in any medium, provided the
original work is properly cited, the use is
non-commercial and no modifications
or adaptations are made.
YIN ET AL.
The contribution of reduction in evaporative cooling
to higher surface air temperatures during drought
Dongqin Yin1,2, Michael L. Roderick2,3,4, Guy Leech3, Fubao Sun2,4, and Yuefei Huang1
1
State Key Laboratory of Hydro-Science and Engineering, Tsinghua University, Beijing, China, 2Research School of Biology,
Australian National University, Canberra, Australia, 3Research School of Earth Sciences, Australian National University,
Canberra, Australia, 4Australian Research Council Centre of Excellence for Climate System Science, Canberra, Australia
Abstract
Higher temperatures are usually reported during meteorological drought and there are two
prevailing interpretations for this observation. The first is that the increase in temperature (T) causes an
increase in evaporation (E) that dries the environment. The second states that the decline in precipitation (P)
during drought reduces the available water thereby decreasing E, and in turn the consequent reduction in
evaporative cooling causes higher T. To test which of these interpretations is correct, we use climatic data
(T, P) and a recently released database (CERES) that includes incoming and outgoing shortwave and
longwave surface radiative fluxes to study meteorological drought at four sites (parts of Australia, US, and
Brazil), using the Budyko approximation to calculate E. The results support the second interpretation at arid
sites. The analysis also showed that increases in T due to drought have a different radiative signature from
increases in T due to elevated CO2.
1. Introduction
Although by no means universal, the near-surface air temperature (T) is commonly reported to be higher in
years with low precipitation (P) [Madden and Williams, 1978; Namias, 1960], especially in arid/semi-arid
regions. The consequent negative T-P relationship during a meteorological drought has long been
recognized [Koster et al., 2009; Trenberth and Shea, 2005], but there have been two very different
interpretations of the underlying cause/s. The first interpretative framework is that high T is the primary
forcing that causes the subsequent increase in evaporation (E) thereby drying the soil [Cai and Cowan, 2008;
Nicholls, 2004]. This interpretation holds that high T is the cause and increased E is the response. It is
consistent with the observed negative T-P relationship and is also widely held amongst the general public.
The second interpretative framework is more nuanced but common in the agricultural and hydrologic
sciences. It begins by distinguishing environments where evaporation is limited by available energy from
those where evaporation is limited by available water [Budyko, 1974; Roderick et al., 2009]. In a water-limited
environment there is minimal runoff, and any reduction in P during a meteorological drought will generally
result in a reduction in E and/or soil moisture [Jung et al., 2010; Lockart et al., 2009; Roderick and Farquhar,
2004]. A further consequence of the decline in E (and hence the latent heat flux, LE, with L the latent heat of
vaporisation) is less evaporative cooling. This implies that for the same net radiative flux, there will be a
concurrent increase in the sensible heat flux (H) thereby warming the air. This interpretation is supported by a
wide variety of flux-tower observations from water-limited environments showing that E does decrease and
H does increase during years with low P [Fischer et al., 2012; Jung et al., 2010; Meyers, 2001; Sun et al., 2008]. In
short, the agricultural-hydrologic interpretation is that a decrease in P in a water-limited environment can
be thought of as the forcing, and the reduction in E (and the associated reduction in evaporative cooling)
causes some of the increase in T.
However, in an energy-limited environment, the agricultural-hydrologic interpretation is quite different. In
those environments, E is more sensitive to changes in energy supply than water supply [Roderick and
Farquhar, 2011], and a larger fraction of the variations in P become variations in runoff. It is clear from the
distinction between water-limited and energy-limited environments that an understanding of the underlying
biophysics of drought requires that one go well beyond examining changes in T to also consider changes
in surface conditions/properties (e.g., changes in latent and sensible heat fluxes, changes in surface albedo)
as well as concurrent changes in atmospheric properties (e.g., less cloud during drought leading to more
solar radiation). Despite this, many climatological analyses of drought still focus on T and P because those
©2014. The Authors.
7891
Geophysical Research Letters
10.1002/2014GL062039
Figure 1. Location of the four study sites in relation to the aridity measure, ∂E/∂P (equation (4)). Inset shows mean annual climatic data (2001–2012) for the four sites.
data are routinely available from existing meteorological networks at the relevant spatial (100’s–10,000’s km2)
and temporal (seasonal-decadal) scales [Sheffield et al., 2009] for drought assessment. By comparison, data on
the surface radiative fluxes are not yet part of the standard climatological data collection efforts, and this has
hindered the interpretation of the drought-T correlation.
Recently, comprehensive radiative data derived from a common satellite source known as the Clouds and
Earth’s Radiant Energy Systems (CERES) program have been made available to the global geoscience
community by NASA [Loeb et al., 2012]. This 1° global database contains monthly observation-based
estimates of the four (incoming and outgoing shortwave and longwave) surface radiative fluxes beginning in
March 2000. The new CERES database presents a unique opportunity to test the agricultural-hydrologic
interpretation by examining individual surface radiative flux anomalies over regions known to have
experienced drought over the last decade.
In this paper we first revisit the relationship between annual T and P anomalies and then investigate how the
four surface radiative fluxes (CERES) vary with P over the same period. We use the Budyko approximation to
calculate E (from P and the radiative data). We examine those data at two sites that have experienced recent
prominent meteorological droughts (Texas, Murray Darling Basin) and compare with results from two
additional sites. If the above-noted agricultural-hydrologic interpretation holds over large spatial scales we
expect to find a different response pattern in components of the surface energy balance to a given variation
in P in a water-limited environment when compared to an energy-limited environment.
2. Materials
Of the four study sites (Figure 1), two were defined by the occurrence of prominent droughts over the last
decade in water-limited environments and include (i) parts of Texas in the southern United States [Seager
et al., 2014], and (ii) the Murray-Darling Basin (MDB) in southeast Australia [Verdon-Kidd and Kiem, 2009]. The
MDB is defined by the basin boundary, while in Texas we selected a region of near-homogenous aridity. We
selected a further site, of greater background aridity, located in southwest Australia (denoted SW Aus in
Figures and Tables), to evaluate the robustness of the relationships derived from Texas and the MDB. To
provide a contrast we also examine a (near-homogenous) humid site in Brazil which lies at similar latitudes
(~30°) to the other three arid sites. At these latitudes, all of the sites are more or less free of seasonal
YIN ET AL.
©2014. The Authors.
7892
Geophysical Research Letters
10.1002/2014GL062039
Figure 2. Relationship between Tm and (a) P, and (b) calculated E anomalies, at the four study sites (2001–2012). The driest and
wettest years are denoted for both the Murray-Darling Basin (MDB) (Dry: 2002, 2006, Wet: 2010) and for Texas (Dry: 2011, Wet:
2004). Regressions that are not significant (p > 0.05, see Table S1 for full statistical details) are shown as a dashed line.
complications arising from snow/ice cover, and all analyses are conducted using annual data. At each site we
obtained the annual means for the daily air temperature (Tm), daily maximum air temperature (Tx), and
annual P from the CRU TS3.21 database [Harris et al., 2014] for 2001–2012 (inclusive).
The analysis begins with the usual definition of the surface energy balance,
RS;i RS;o þ RL;i RL;o ¼ Rn ¼ LE þ H þ G
(1)
with the sum of the incoming and outgoing shortwave (RS,i, RS,o) and longwave (RL,i, RL,o) surface radiative
fluxes being equal to the net radiative flux (Rn) that is balanced by the latent (LE), sensible (H), and ground (G)
heat fluxes. Estimates of the four surface radiative fluxes were obtained from the CERES-EBAF Ed2.7 database
[Loeb et al., 2012] which provides a global monthly 1° record beginning in March 2000 and current to March
2013. We calculated the annual mean for each flux for the same period (2001–2012 inclusive) at the four
study sites. As we are using annual data we assume that G equals zero. To calculate H (=Rn LE) we first infer E
using the Mezenstev-Choudhury-Yang formulation [Choudhury, 1999; Mezentsev, 1955; Yang et al., 2008] of
the Budyko curve,
E¼
PE o
n
P þ E no
n1
(2)
with Eo defined as the water-equivalent of Rn (Eo = Rn/L). In that calculation we set n = 1.9 because that
parameter value reproduces the original Budyko curve [Donohue et al., 2011]. The annual E anomalies (ΔE)
were calculated as a function of the annual P and Eo anomalies (ΔP, ΔEo) using,
ΔE ¼ ð∂E=∂PÞΔP þ ð∂E=∂E o ÞΔE o
with the partial differentials calculated from equation (2) [Roderick and Farquhar, 2011],
∂E E
E no
¼
∂P P Pn þ E no
and,
∂E
E
Pn
:
¼
∂E o E o Pn þ E no
(3)
(4a)
(4b)
The numerical values of the partial differentials and other key climatic data at the four study sites are shown
in Figure 1.
Note that the partial derivatives have a straight-forward physical interpretation [Roderick and Farquhar, 2011;
Roderick et al., 2014]. In a water-limited (arid) environment, E mostly depends on P, and we expect ∂E/∂P → 1 and
∂E/∂Eo → 0, while in an energy-limited (humid) environment we expect the opposite. (See Figures S1–S3 for
global maps.) On that basis, three of the study sites (SW Australia, Texas, MDB) are considered arid/semi-arid,
YIN ET AL.
©2014. The Authors.
7893
Geophysical Research Letters
10.1002/2014GL062039
and for those sites we expect minimal
runoff with any variations in (annual)
P translating mostly into (annual)
variations in E. In contrast, in the humid
environment (Brazil), it is expected that
variations in P will mostly translate into
variations in runoff, and we expect E to
be more sensitive to variations in Eo than
P. To give a numerical example using
equation (3) and the data from Figure 1,
for the MDB we have ΔE = 0.75 ΔP
+ 0.07 ΔEo, while at the humid Brazil site
we have ΔE = 0.25 ΔP + 0.45 ΔEo.
3. Results
3.1. Relating Temperature to
Precipitation and Evaporation
Figure 3. Relationship between anomalies of P and of the surface radiative
and convective fluxes at the four study sites (2001–2012). (a) ΔRS,i, (b) ΔRL,i,
(c) αΔRS,i (change in RS,o caused by a change in RS,i), (d) ΔRL,o, (e) RS,i Δα
(change in RS,o caused by a change in α), (f) ΔRn, (g) calculated LΔE, and
(h) calculated ΔH. Regressions that are not significant (p > 0.05, see
Table S2 for full statistical details) are shown as a dashed line.
The relationship between annual Tm
and P anomalies (ΔTm, ΔP) at the four
study sites is shown in Figure 2a. The
driest and wettest years are denoted
for both the MDB (2002/2006 and
2010) and Texas (2011 and 2004).
During the prominent Texas drought
year (2011) Tm was around 1.2° C above
the decadal average. The absolute
magnitudes of the dry/wet anomalies
(~ ±200–300 mm yr1) were roughly
similar at all sites with the exception of
the very dry SW Australia site where the
anomalies were smaller. There is a strong
(and statistically significant) negative
ΔTm-ΔP relationship at all three arid sites
(SW Australia, Texas, MDB). (The same
results are evident for the ΔTx-ΔP
relationships, see Figure S5.) At all three
arid sites, a decrease in P resulted in a
more or less equal reduction in (the
calculated) E, and the resulting ΔTm-ΔE
relationships were qualitatively similar
to the earlier ΔTm-ΔP relationships
(Figure 2). In contrast, at the humid site
(Brazil) we find the opposite with
positive ΔTm-ΔP (Figure 2a) and ΔTm-ΔE
(Figure 2b) relationships although both
are weak and not statistically significant.
(The same relations were found using
other databases for E, see Figure S4.)
3.2. Relating Precipitation to the Surface Energy Balance
Figure 3 documents the relationship between annual P and surface energy balance anomalies at the
four sites. Note that in examining RS,o (=α RS,i, with α the albedo) we separated the anomaly into two terms
YIN ET AL.
©2014. The Authors.
7894
Geophysical Research Letters
10.1002/2014GL062039
(ΔRS,o = α ΔRS,i + RS,i Δα) to separately account for changes in incoming shortwave (α ΔRS,i, Figure 3c) and in
albedo (RS,i Δα, Figure 3e). Changes in the surface radiative fluxes with ΔP were primarily associated with
changes in RS,i and RL,o (Figures 3a and 3d). The key result is that drought was accompanied by increasing RS,i
(Figure 3a) at all sites irrespective of the background aridity. RL,o also increased during drought but only at the
three arid sites (Figure 3d). At the humid Brazil site there is no (statistically significant) relationship between
ΔRL,o and ΔP (Figure 3d) which was anticipated based on the earlier ΔTm-ΔP results at that site (Figure 2a).
Note that at the arid sites the increase in RS,i during drought is roughly equal in magnitude to the increase in
RL,o. The physical interpretation is that the increase in RS,i during drought is an important source of heat for
increasing the surface T (and thereby increasing RL,o) at the three arid sites. However, at the humid (Brazil) site
we observed a (qualitatively) similar increase in RS,i as P declined but that did not translate into an increase in
RL,o (i.e., surface T) (Figures 2a, 3a, and 3d).
Changes in the other radiative terms were not as prominent. In particular, there were consistent changes in
RS,o due to changes in RS,i at all sites but the overall magnitude is small (Figure 3c), while the changes in RS,o
due to changes in albedo showed large between-year and between-site scatter with little overall consistency
(Figure 3e). The relationship between ΔRL,i and ΔP also showed much scatter. While all sites showed a
decrease in RL,i during drought, that decrease is only statistically significant at the driest site (SW Australia)
(Figure 3b). Taken as a whole, there is minimal impact of drought on Rn at the Brazil and MDB sites (Figure 3f)
which emphasizes the point that a near-constant Rn during dry/wet years does not mean that each of the
individual surface radiative fluxes is also constant. At the remaining two arid sites (SW Australia, Texas), Rn
showed a (statistically significant) decrease during meteorological drought.
The Budyko-based estimates of changes in the two convective fluxes were as anticipated. At the three arid
sites LE showed a strong positive relationship with P, while the response was still positive but muted at the
humid Brazil site (Figure 3g). The magnitude of changes in LE is larger than those in Rn (compare Figures 3f
and 3g) which resulted in a strong (statistically significant) negative ΔH-ΔP relationship (Figure 3h) at all sites
irrespective of the background aridity. This implies a higher H during drought at all four study sites.
4. Discussion
The increase in air temperature (Tm) as E declined (Figure 2b) found at the three arid sites examined here is
consistent with previous analysis of both observations and climate model output [Schär et al., 1999;
Seneviratne et al., 2010; Sheffield and Wood, 2008; Teuling et al., 2009]. At the arid sites (SW Australia, Texas,
MDB) the decline in E is caused by a decline in P that results in less evaporative cooling and a consequent
increase in the sensible heat flux. We calculated E using the Budyko approximation (E = f(P, Eo)) and then
estimated H (=Rn LE) by balance. Our calculations of E are based on a steady state formulation [Roderick and
Farquhar, 2011] and ignore changes in soil moisture storage. During drought the soil moisture anomaly is
expected to be negative [Leblanc et al., 2009; Long et al., 2013]. This implies that the steady state formulation
we used would have likely underestimated E during drought and would in turn lead to an overestimation of
H. Hence, the calculated estimates for ΔE, LΔE, and ΔH (Figures 2b, 3g, and 3h, respectively) should only be
considered a useful first approximation. Despite that, the storage change would not alter the qualitative
nature of the derived relation and an increase in H during drought in arid regions has been confirmed by
numerous eddy flux observations [e.g., Fischer et al., 2012; Jung et al., 2010; Meyers, 2001; Sun et al., 2008]. In an
even more general sense, laboratory experiments where the energy supply is fixed demonstrate that as wet
soil dries there is an abrupt increase in T when E first begins to decline [Aminzadeh and Or, 2013]. This is the
well-known transition from stage I to stage II evaporation. In summary, our findings are in line with the
standard agricultural-hydrologic interpretation, and we conclude that some of the increase in T during
meteorological drought in an arid environment is the result of a land surface feedback to the atmosphere.
At the one humid site (Brazil) examined, there was a weak (and not statistically significant) decrease in Tm as P
declined that was the opposite of the pattern found at the arid sites (Figure 2a). Despite the different result,
the response of the turbulent fluxes at the humid site was qualitatively similar to that at the arid sites with
(slightly) less E and (slightly) more H during meteorological drought. The above-noted difference between
arid and humid regions emphasizes that the changes under meteorological drought cannot be solely
attributed to changes in a land surface feedback due to changes in surface moisture conditions (i.e., less E,
YIN ET AL.
©2014. The Authors.
7895
Geophysical Research Letters
10.1002/2014GL062039
2
Table 1. Annual Anomalies in the Surface Energy Balance (W m ) at the
1a
Four Study Sites Assuming a Fixed Reduction in P of 200 mm yr
ΔRS,i
ΔRL,i
αΔRS,i
RS,iΔα
ΔRL,o
ΔRn
LΔE
ΔH
SW Aus
Texas
Murray-Darling Basin (MDB)
Brazil
+11.8
6.2
+2.2
+2.3
+8.6
7.6
14.7
+7.3
+9.4
1.6
+1.8
+1.5
+10.0
5.5
12.9
+7.4
+8.3
0.5
+1.4
+0.8
+6.3
0.7
11.7
+11.0
+4.2
1.7
+0.7
+0.1
+1.0
+ 0.6
3.6
+4.3
more H). Instead, there must also be
changes in the individual radiative
fluxes that differ between arid and
humid regions.
To summarize the different drought
behavior between arid and humid
regions we use the site-specific
regressions (Figure 3 and Table S2) to
calculate the changes in the surface
energy balance for a given drop in P of
a
The anomalies are calculated using the site-specific slopes reported
200 mm yr1 (Table 1). To set a
in Figure 3 (see Table S2). Significant changes ( p < 0.05, Figure 3)
benchmark, if E were to decrease by the
are italicized.
same amount, the resulting decrease in
the latent heat flux (LΔE) would be around 15.5 W m2 which is reasonably close to the estimated drop at the
three arid sites (SW Australia, Texas, MDB). This again emphasizes the tight control that moisture supply (P)
has on E under arid conditions. However, at the humid (Brazil) site, the calculated change in latent heat flux
(= 3.6 W m2) is only a small fraction of the possible (= 15.5 W m2) change. At that site, most of the change
in P would become a change in runoff, as has been recently observed in a humid catchment in the Swiss alps
[Seneviratne et al., 2012]. That is again in accord with the standard agricultural-hydrologic interpretation.
During meteorological drought, all sites, irrespective of background aridity, experienced an increase (although
of different magnitudes) in the incoming shortwave irradiance. The key to a more complete understanding of
drought is to understand the ultimate fate of that additional shortwave irradiance. At the three arid sites there
was an increase in outgoing longwave irradiance of roughly equal magnitude implying that the extra
shortwave contributed to an increase in the surface T thereby increasing the outgoing longwave irradiance.
However, at the humid site, there was little change in either air temperature (Figure 2a, Brazil) or outgoing
longwave irradiance (Figure 3d, Brazil). Instead, much of extra shortwave at the humid site was partitioned into
the turbulent (LE + H) fluxes. Future research could use the CERES data to examine the radiative and turbulent
fluxes at a finer temporal scale (e.g., monthly) to try and resolve more details on the partitioning.
Another important point highlighted by the data (Table 1) is that in terms of a local (and transient) radiative
forcing, meteorological drought is characterized by both an increase in the incoming shortwave and a
decrease in the incoming longwave irradiance. The increases in incoming shortwave are statistically
significant at all sites, and such a finding is well known and physically sensible (i.e., less cloud leading to more
incoming shortwave irradiance during meteorological drought). The decline in incoming longwave during
drought was also found at all sites, although there was considerable scatter, and it was only statistically
significant at the driest (SW Australia) site (Figure 3b and Table 1). The CERES estimate is based on a radiative
transfer model that uses observed atmospheric profiles of temperature and humidity along with observed
cloud cover. In contrast, the agricultural and hydrologic communities have long used a number of simpler
empirical and semi-empirical models to estimate the incoming longwave radiation at the surface. (See
comprehensive summaries of different models by Duarte et al. [2006] and Lhomme et al. [2007].) Although not
all of those empirical models are identical, they would generally predict a decrease in incoming longwave
during meteorological drought because of less cloud cover and/or water vapor in accord with the more
complex CERES observation-based estimates.
This raises a key point relating to the source of variation in the radiative balance, and the consequent relationship
to T and to the enhanced greenhouse effect. At a global scale, the current conception is for the enhanced
greenhouse effect to increase the T at the surface over the next 100 years by increasing the incoming longwave
with little (globally averaged) change in incoming shortwave [Roderick et al., 2014; Russell et al., 2013]. That is very
different to the (transient) changes in the (local) surface radiative fluxes during meteorological drought (increase
in incoming shortwave, decrease in incoming longwave). In essence, the increase in T observed during
meteorological drought in arid regions is accompanied by a very different radiative signature from the increase
in T expected due to an enhancement of the greenhouse effect. With that insight we suggest that it may be
possible to separate the change in T due to increasing greenhouse gases from that due to forced/unforced
changes in atmospheric circulation using observations of the four individual surface radiative fluxes.
YIN ET AL.
©2014. The Authors.
7896
Geophysical Research Letters
Acknowledgments
We thank Graham Farquhar for comments on an earlier version of the
manuscript. We also thank the editor
and two referees for insightful comments that improved the manuscript.
This research was supported by the
Australian Research Council
(CE11E0098), the National Natural
Science Foundation of China
(91125018), and the China Scholarship
Council (201306210089).
M. Bayani Cardenas thanks Adriaan
Teuling and Rene Orth for their assistance in evaluating this paper.
YIN ET AL.
10.1002/2014GL062039
References
Aminzadeh, M., and D. Or (2013), Temperature dynamics during nonisothermal evaporation from drying porous surfaces, Water Resour. Res.,
49, 7339–7349, doi:10.1002/2013WR014384.
Budyko, M. I. (1974), Climate and Life, 508 pp., Academic Press, New York.
Cai, W., and T. Cowan (2008), Evidence of impacts from rising temperature on inflows to the Murray-Darling Basin, Geophys. Res. Lett., 35,
L07701, doi:10.1029/2008GL033390.
Choudhury, B. (1999), Evaluation of an empirical equation for annual evaporation using field observations and results from a biophysical
model, J. Hydrol., 216(1-2), 99–110.
Donohue, R. J., M. L. Roderick, and T. R. McVicar (2011), Assessing the differences in sensitivities of runoff to changes in climatic conditions
across a large basin, J. Hydrol., 406(3–4), 234–244.
Duarte, H. F., N. L. Dias, and S. R. Maggiotto (2006), Assessing daytime downward longwave radiation estimates for clear and cloudy skies in
Southern Brazil, Agric. For. Meteorol., 139(3–4), 171–181.
Fischer, M. L., M. S. Torn, D. P. Billesbach, G. Doyle, B. Northup, and S. C. Biraud (2012), Carbon, water, and heat flux responses to experimental
burning and drought in a tallgrass prairie, Agric. For. Meteorol., 166-167, 169–174.
Harris, I., P. D. Jones, T. J. Osborn, and D. H. Lister (2014), Updated high-resolution grids of monthly climatic observations-the CRU TS3. 10
Dataset, Int. J. Climatol., 34, 623–642.
Jung, M., M. Reichstein, P. Ciais, S. I. Seneviratne, J. Sheffield, M. L. Goulden, G. Bonan, A. Cescatti, J. Chen, and R. De Jeu (2010), Recent decline
in the global land evapotranspiration trend due to limited moisture supply, Nature, 467(7318), 951–954.
Koster, R. D., S. D. Schubert, and M. J. Suarez (2009), Analyzing the concurrence of meteorological droughts and warm periods, with implications for the determination of evaporative regime, J. Clim., 22, 3331–3341.
Leblanc, M. J., P. Tregoning, G. Ramillien, S. O. Tweed, and A. Fakes (2009), Basin-scale, integrated observations of the early 21st century
multiyear drought in southeast Australia, Water Resour. Res., 45, W04408, doi:10.1029/2008WR007333.
Lhomme, J. P., J. J. Vacher, and A. Rocheteau (2007), Estimating downward long-wave radiation on the Andean Altiplano, Agric. For. Meteorol.,
145(3–4), 139–148.
Lockart, N., D. Kavetski, and S. W. Franks (2009), On the recent warming in the Murray-Darling Basin: Land surface interactions misunderstood, Geophys. Res. Lett., 36, L24405, doi:10.1029/2009GL040598.
Loeb, N. G., J. M. Lyman, G. C. Johnson, R. P. Allan, D. R. Doelling, T. Wong, B. J. Soden, and G. L. Stephens (2012), Observed changes in top-ofthe-atmosphere radiation and upper-ocean heating consistent within uncertainty, Nat. Geosci., 5(2), 110–113.
Long, D., B. R. Scanlon, L. Longuevergne, A. Y. Sun, D. N. Fernando, and H. Save (2013), GRACE satellite monitoring of large depletion in water
storage in response to the 2011 drought in Texas, Geophys. Res. Lett., 40, 3395–3401, doi:10.1002/grl.50655.
Madden, R. A., and J. Williams (1978), The correlation between temperature and precipitation in the United States and Europe, Mon. Weather
Rev., 106(1), 142–147.
Meyers, T. P. (2001), A comparison of summertime water and CO2 fluxes over rangeland for well watered and drought conditions, Agric. For.
Meteorol., 106(3), 205–214.
Mezentsev, V. (1955), More on the calculation of average total evaporation [in Russian], Meteorol. Gidrol., 5, 24–26.
Namias, J. (1960), Factors in the initiation, perpetuation and termination of drought, International Association of Hydrological Sciences
Commission of Surface Waters Publication 51, 81–94. [Available from IAHS Press, Institute of Hydrology, Wallingford, Oxfordshire,
OX10 8BB, U. K.]
Nicholls, N. (2004), The changing nature of Australian droughts, Clim. Change, 63(3), 323–336.
Roderick, M. L., and G. D. Farquhar (2004), Changes in Australian pan evaporation from 1970 to 2002, Int. J. Climatol., 24(9),
1077–1090.
Roderick, M. L., and G. D. Farquhar (2011), A simple framework for relating variations in runoff to variations in climatic conditions and
catchment properties, Water Resour. Res., 47, W00G07, doi:10.1029/2010WR009826.
Roderick, M. L., M. T. Hobbins, and G. D. Farquhar (2009), Pan evaporation trends and the terrestrial water balance II. Energy balance and
interpretation, Geography Compass, 3, 761–780.
Roderick, M. L., F. Sun, W. H. Lim, and G. D. Farquhar (2014), A general framework for understanding the response of the water cycle to global
warming over land and ocean, Hydrol. Earth Syst. Sci., 18(5), 1575–1589.
Russell, G. L., A. A. Lacis, D. H. Rind, C. Colose, and R. F. Opstbaum (2013), Fast atmosphere-ocean model runs with large changes in CO2,
Geophys. Res. Lett., 40, 5787–5792, doi:10.1002/2013GL056755.
Schär, C., D. Lüthi, U. Beyerle, and E. Heise (1999), The soil-precipitation feedback: A process study with a regional climate model, J. Clim.,
12(3), 722–741.
Seager, R., L. Goddard, J. Nakamura, N. Henderson, and D. E. Lee (2014), Dynamical causes of the 2010/11 Texas-Northern Mexico drought*,
J. Hydrometeorol., 15(1), 39–68.
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(3–4), 125–161.
Seneviratne, S. I., et al. (2012), Swiss prealpine Rietholzbach research catchment and lysimeter: 32 year time series and 2003 drought event,
Water Resour. Res., 48, W06526, doi:10.1029/2011WR011749.
Sheffield, J., and E. F. Wood (2008), Global trends and variability in soil moisture and drought characteristics, 1950-2000, from observationdriven simulations of the terrestrial hydrological cycle, J. Clim., 21, 432–458.
Sheffield, J., K. M. Andreadis, E. F. Wood, and D. P. Lettenmaier (2009), Global and continental drought in the second half of the twentieth
century: Severity-area-duration analysis and temporal variability of large-scale events, J. Clim., 22(8), 1962–1981.
Sun, G., A. Noormets, J. Chen, and S. G. McNulty (2008), Evapotranspiration estimates from eddy covariance towers and hydrologic modeling
in managed forests in Northern Wisconsin, USA, Agric. For. Meteorol., 148(2), 257–267.
Teuling, A. J., et al. (2009), A regional perspective on trends in continental evaporation, Geophys. Res. Lett., 36, L02404, doi:10.1029/2008GL036584.
Trenberth, K. E., and D. J. Shea (2005), Relationships between precipitation and surface temperature, Geophys. Res. Lett., 32, L14703,
doi:10.1029/2005GL022760.
Verdon-Kidd, D. C., and A. S. Kiem (2009), Nature and causes of protracted droughts in southeast Australia: Comparison between the
Federation, WWII, and Big Dry droughts, Geophys. Res. Lett., 36, L22707, doi:10.1029/2009GL041067.
Yang, H., D. Yang, Z. Lei, and F. Sun (2008), New analytical derivation of the mean annual water-energy balance equation, Water Resour. Res.,
44, W03410, doi:10.1029/2007WR006135.
©2014. The Authors.
7897