Journal of Hydrology: Regional Studies 5 (2016) 6679

Contents lists available at ScienceDirect

Journal of Hydrology: Regional

Studies
journal homepage: www.elsevier.com/locate/ejrh

Variation of reference evapotranspiration in the central

region of Argentina between 1941 and 2010
A.C. de la Casa , G.G. Ovando
Facultad de Ciencias Agropecuarias, Universidad Nacional de Crdoba, Argentina, Ing Agr. Flix Aldo Marrone 746 Ciudad
Universitaria, Crdoba, Argentina

a r t i c l e

i n f o

Article history:
Received 14 May 2015
Received in revised form 5 October 2015
Accepted 13 November 2015
Keywords:
Reference evapotranspiration
Climate change
Climate variables
Precipitation

a b s t r a c t
Study region: Changes in reference evapotranspiration (ETo ) may have important consequences for agricultural suitability in the central region of Argentina. Annual ETo variation
was assessed, in terms of both territory and time, for the 7 decades between 1941 and 2010,
analyzing the behavior of the 4 atmospheric variables which determine it: temperature,
vapor pressure, wind speed and cloud cover.
Study focus: The inuence of each variable on ETo was evaluated from a multiple regression
model and a simple correlation analysis, using climate data from the observation network,
and repeating this analysis using interpolated variables. In this grid scheme, linear relationships were determined between ETo and the different key atmospheric variables, plus
precipitation (PP), and the t test was applied to establish the statistically signicant sectors
(P < 0.1). Then, those areas with a signicant trend change (P < 0.1) were determined by the
MannKendall test. Finally, the interception of the grids was performed to establish their
joint occurrence.
New hydrological insights for the region: Most of the region analyzed (>91%) presents a nonsignicant variation of ETo over time, with a mostly non-signicant change of each driving
variable, regarding both its relationship with ETo and its own trend of change. The benecial
change in agricultural suitability reported for this water-limited region was found to be
produced almost exclusively by increasing PP.
2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
The worldwide rise in air and ocean temperatures, widespread melting of snow and ice and increased sea levels are
evidence of climate system warming in recent decades (Bates et al., 2008). Despite the close link between the energy
balance and the water budget, however, while the rate of atmospheric evaporative demand has increased in some regions,
matching the warming (Tabari et al., 2011), Hobbins et al. (2004) paradoxically reported decreasing rates in other territories
(Bandyopadhyay et al., 2009; Chattopadhyay and Hulme, 1997; Xu et al., 2006). Changes in evapotranspiration (ET) are of
great importance because this is not only one of the components of the hydrological cycle in a region (Wang and Dickinson,
2012), but also has a direct impact on agricultural suitability (de la Casa and Ovando, 2014). Knowledge of ET variations in
time and space is crucial for understanding the interactions between different vegetation covers and the atmosphere (Keane

Corresponding author. Fax: +54 351 4334114.

E-mail address: delacasa@agro.unc.edu.ar (A.C. de la Casa).
http://dx.doi.org/10.1016/j.ejrh.2015.11.009
2214-5818/ 2015 The Authors. Published by
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Elsevier

B.V.

This

is

an

open

access

article

under

the

CC

BY-NC-ND

license

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

67

et al., 2002), better managing a territorys water resources (Raupach et al., 2009), or assessing the phenomenon of drought
and its associated environmental impacts (McVicar and Jupp, 1998).
The evaporative demand of the atmosphere, represented either by the potential ET (PET) (Penman, 1948; Thornthwaite,
1948) or the reference ET (ETo ) (Allen et al., 1998, 2006), is a complex biophysical process inuenced by four major meteorological variables: net radiation, vapor pressure, wind speed and air temperature (Donohue et al., 2010). A conceptual entity
more than an operative one, the PET or ETo (hereinafter ETo only) is difcult to measure directly, and so different calculation
methods have been developed, for models with one or more of these variables. Comparing the performance of different
predictive models that estimate ETo , Penmans combined method with the use of this complete set of variables was the one
that produced the most convincing estimates of the dynamics of potential evaporation (Donohue et al., 2010).
In a comprehensive review of trends in global rates of evaporation and evapotranspiration, McVicar et al. (2012) argue that
these generally tend to decrease, mainly due to the aerodynamic component (resulting from the combined effects of wind
speed (u2 ) and vapor pressure (ea )) and, secondly, as a result of changes in the radiative component. Through a worldwide
meta-analysis, they consider it very likely that decreased u2 is widespread, although, in order to assess its inuence on ETo
reduction, they argue the need to comprehensively analyze the variation of the four major meteorological elements involved
in atmospheric water demand.
There have been numerous changes in weather conditions worldwide during the twentieth century, shown by long-term
trends in global average air temperature (Trenberth et al., 2007), vapor pressure (Durre et al., 2009), precipitation (New
et al., 2001), net radiation (Wild, 2009) and wind speed (McVicar et al., 2010, 2012). In Australia, both temperature and
precipitation have increased on average in the last 3 or 4 decades (Bureau of Meteorology, 2007), as well as vapor pressure,
while wind and net radiation have decreased (Roderick et al., 2007; McVicar et al., 2008). All these atmospheric changes
may explain changes in evaporative demand (Donohue et al., 2010).
The relative importance of the four primary meteorological variables that control the trend of evaporative demand may
change both absolutely and relatively, and very probably vary both spatially and temporally. Tang et al. (2011) determined
that the decline in ETo in the Haihe River Basin in China can be attributed to variations of net radiation, ea , u2 and temperature
(T). While the increase in temperature determines a signicant increase in ETo over time, the negative rates of the other
variables that regulate the process explain the general negative trend of ETo . Thus, Zheng et al. (2009) interpret the paradox
of pan evaporation in the Haihe River Basin, noting that, while the increase in temperature may have increased evaporation,
this was offset by the decrease in wind speed and solar radiation, combined with the increase of water vapor pressure in
the region.
One of the main advantages of Penmans method is its ductility to represent atmospheric water demand under different
climatic conditions and to translate the complex combination of the constituent elements. Among the disadvantages, the
model presents operational difculties in terms of the availability of information needed for its implementation. This is
exacerbated when seeking to simultaneously evaluate the temporal and spatial variation of ETo , since meteorological networks change over time with the appearance and disappearance of some stations, or because the availability of instruments
is not uniform, etc. Recognizing this difculty, Irmak et al. (2012) developed a method for estimating the full set of variables required to calculate Penman-ETo , using only daily values of maximum temperature (Tmax ), minimum temperature
(Tmin ) and precipitation (PP). The ETo obtained in this way for Platte River Basin (Nebraska, USA) presented a linear trend
of 0.36 mm yr1 between 1893 and 2008, a decline that was attributed to a signicant increase in PP (100 mm during the
same period), reducing solar and net radiation.
In a study of agricultural potential variation in the central region of Argentina, de la Casa and Ovando (2014) consider
that the greatest potential at present can be attributed to increased precipitation since the 1940s. In contrast, the change of
ETo was not signicant in more than 90% of this territory and, where it was signicant, the variation was ambiguous because
in some sectors the trend is positive and in others negative.
To further analyze ETo behavior in the central region of Argentina, a territory that is limited in its evaporative capacity
primarily by water availability (Donohue et al., 2007), this study evaluates ETo variation between 1941 and 2010. The
temporal and spatial changes in atmospheric demand are assessed through the variation of the four meteorological variables
responsible (T, ea , u2 and cloud cover (N)) and also of PP. In addition to assessing the inuence of using cloud cover instead
of sunshine duration to determine radiation balance, the analysis proceeds in two spatial scales: local, which strictly uses
meteorological data recorded by the observation network, and the other regional, with information obtained from a process
of geographic interpolation.
2. Materials and methods
2.1. Data
The climate data used in this study were provided by the National Weather Service of Argentina (Smn, 1958, 1975,
1981, 1986, 1992, 2007, 2012 the last 2 series published digitally). These statistics present monthly mean values of weather
variables for a ten-year period, with a complete series from 1941. Thus, time series of climatic values were available for 7
consecutive periods (i.e., 19411950, 19511960, 19611970, 19711980, 19811990, 19912000 and 20012010). Fig. 1
shows the study area in the central region of Argentina, with the geographic location of the weather stations used in the
period 20012010, and Table 1 shows the list of weather stations from which data were used in each decade.

68

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

Fig. 1. Geographical location of weather stations in the central region of Argentina used in the period 20012010.

Changes within the observation network during the long period analyzed may have a negative inuence on the results. To
prevent this, the number of weather stations used in each 10-year period was adjusted to be similar and, in order to achieve
a more homogeneous and reliable data base, the geographic information analysis was optimized by including only those
weather stations which were present in 4 or more decades during the analysis period, as Table 1 shows. Also, the maximum
number of stations in the central sector of Argentina was included, to cover the entire territory of Crdoba province.

2.2. Reference evapotranspiration (ETo ) calculation

The reference evapotranspiration (ETo ) was estimated by the PenmanMonteith method, following the guidelines established by the FAO (Allen et al., 1998, 2006), with the following expression:
ETo =

900
u2 (es ea )
0, 408 (Rn G) +  T +237

 +  (1 + 0, 34u2 )

(1)

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

69

Table 1
Weather stations, grouped by province, used in each 10-year period of the study with their altitude.
Province

Weather station

Altitude (masl) Decade

19411950 19511960 19611970 19711980 19811990 19912000 20012010

BUENOS AIRES PERGAMINO

JUNIN AERO
CRDOBA OBS.
CRDOBA
LABOULAYE AERO
MARCOS JUAREZ AERO
PILAR OBSERVATORIO
RIO CUARTO AERO
VILLA DOLORES AERO
VILLA MARIA RIO SECO
GENERAL PICO AERO
LA
SANTA ROSA AERO
PAMPA
VICTORICA
CHEPES
LA
LA RIOJA AERO
RIOJA
VILLA MERCEDES
SAN
SAN LUIS AERO
LUIS
CERES AERO
SANTA
ESPERANZA
FE
RAFAELA
ROSARIO AERO
SAUCE VIEJO AERO
SANTIAGO DEL ESTERO
S.
DEL
ESTERO

65
81
425
137
114
338
421
569
341
145
191
312
658
429
486
713
88
38
100
25
18
199

x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x

x
x
x
x
x
x
x
x
x
x
x
x
x
x

x
21

x
20

x
x
x
x
x

x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
22

x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
22

x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
21

x
x
x
x
x
x
x
x
x
x

x
x
x
x
x
x
x

x
x
x
x

x
x
x
x

x
x
x
17

x
x
x
16

x
x

where ETo is the daily reference crop evapotranspiration rate (mm d1 ), Rn the net radiation (MJ m2 d1 ), G the soil heat
ux (MJ m2 d1 ),  the slope of the saturation vapor pressuretemperature curve (kPa K1 ),  the psychometric constant
(kPa K1 ), T the average air temperature (K), u2 the wind speed at 2 m height (m s1 ), es the saturation vapor pressure (kPa),
and ea the actual air vapor pressure (kPa), and 900 is a constant for a hypothetical reference crop with an assumed crop
height of 0.12 m, a xed surface resistance of 70 s m1 and an albedo of 0.23.
While mean monthly air temperature, vapor pressure, and wind speed were obtained from climatic statistics (SMN), solar
radiation is not measured systematically in the network at all. Instead, some stations record sunshine duration (h), which
is used to calculate the relative sunshine duration (h/H), making use of the astronomical duration of the suns brightness
(H), but this is not general, and these data are more available only in recent years. Cloudiness (N), however, is a widespread
variable observed in the entire region and in all decades, so this was used to represent the exchange of radiative transference.
The N (in Okta) was used to estimate relative sunshine duration (h/H) with the following expression (de la Casa and Ovando,
2014):
h
= 5.85 N + 88.62
H

(2)

A special analysis (see Fig. 2) was made in order to assess the inuence of using N instead of h/H in the ETo calculation.
2.3. Analytical procedure
The whole analysis was performed in a step sequence. First, the particular inuence of each variable on ETo was evaluated
from a multiple regression model (MR) (Chattopadhyay and Hulme, 1997; Cong et al., 2009) and with a simple correlation
analysis, using exclusively climate data from the observation network. This procedure was termed the local mode analysis.
To extend the evaluation to the entire region, a geographic information system was developed (Petersen et al., 1995). The
temporal and spatial variability of climatic factors that control ETo were analyzed using cartographic products developed
with the Surfer program (Golden Software, 1997), in the same way as de la Casa and Ovando (2014). Digital maps were
developed for each variable and for ETo , using geostatistical spatial interpolation techniques (Burrough, 1986). For spatial
representation of the results, the linear kriging method was used for all variables (Tabios and Salas, 1985). The interpolated
values were extracted with a grid of 78 longitudinal columns and 100 latitudinal rows (9.8 9.8 km), corresponding to a
section of the central region of Argentina (28 37 latitude S and 60 67 longitude W), which fully includes the territory of
the province of Crdoba and also peripheral or edge information.The relief of Central Argentina is mostly at, so the altitude
differences between weather stations were not considered in the data interpolation. To the west of Crdoba Province there
is a mountainous area (Sierras de Crdoba), with a maximum height of over 3000 masl. To avoid the inuence of elevation
on the ETo regime (Mc Vicar et al., 2010), the sector over 1000 m in altitude was excluded from the analysis using a mask.
Then, in the regional mode, the correlation and regression procedures were repeated but using the interpolated variables
to analyze the inuence of the regionalization. Next, in the 78 100 grid scheme, linear relationships were determined

70

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679
1700

1700

1600

1600
1500
y = 0.903x + 141.3
R = 0.897

1400

ETo RM (mm yr -1)

ETo RM (mm yr -1)

1500

1300

1300

1200

1200

1100

1100

1000

y = 0.976x + 48.27
R = 0.946

1400

1000

1000

1100

1200

1300

1400

1500

1600

1700

1000

1100

1200

ETo PM (mm yr-1)

1700

1600

1600

ETo RM (mm yr -1)

ETo RM (mm yr -1)

1500

1600

1700

1500

y = 0.901x + 135.0
R = 0.953

1400

1400

ETo PM (mm yr-1)

1700

1500

1300

1300

1300

1200

1200

1100

1100

1000

y = 0.942x + 76.44
R = 0.983

1400

1000

1000

1100

1200

1300

1400

ETo PM (mm yr-1)

1500

1600

1700

1000

1100

1200

1300

1400

1500

1600

1700

ETo PM (mm yr -1)

Fig. 2. Annual reference evapotranspiration estimated by the PenmanMonteith method (ETo PM) with the four basic variables (T, ea , u2 and h/H) on
the abscissa and the values calculated having replaced h/H for its estimate with N on the ordinate, for climate data of the decades 19912000 (left) and
20012010 (right). The lower two graphs, the same as the above but without data of General Pico weather station.

between ETo and the different key atmospheric variables, including precipitation (PP), and the t value was used to establish
those sectors where angular coefcient is statistically signicant (P < 0.1).
Subsequently, for each time series of the atmospheric variables, a nonparametric MannKendall analysis was carried out
at each grid point to determine those sectors with a signicant change over time at a 10% probability level (Gilbert, 1987).
The MannKendall (S) test was calculated using the following formula:
S=

n
n1 


sgn xj xk

(3)

k=1 j=k+1

where xj and xk are the annual values for the decades j and k, respectively, where j > k, and
1
sgn (xj xk ) = 0
1

if xj xk > 0
if xj xk = 0
if xj xk < 0

Finally, interception of the grids with statistical signicance was performed to establish their joint occurrence.
3. Results and discussion
In the absence of a direct measurement of solar radiation, its value is often estimated using the relative sunshine duration
(h/H) and the extraterrestrial radiation. As the cloudiness condition (N) was used here to estimate relative sunshine hours
(h/H), this substitution in the estimation of ETo was evaluated rst. For this purpose, we used climate information of the
last two decades (19912000 and 20012010) because, unlike the previous periods, N and h/H records are both available at
most meteorological stations in the region. Fig. 2 depicts the ratio between annual ETo estimated by the PenmanMonteith
method using the four basic variables on the abscissa, and the ETo values obtained having replaced h/H with its estimate
using N on the ordinate, for the central region of Argentina.
The relationship between the ETo estimated with each method tends to identity line for both decades, although with less
dispersion in 20012010, as the regression and determination coefcients in Fig. 2 indicate. Systematically, the estimation
error is reduced when the information of General Pico meteorological station is removed from the analysis, which indicates
a local effect for both periods.
In the next step, a multiple regression analysis (MR) similar to Chattopadhyay and Hulme (1997); Cong et al. (2009) was
performed to assess the four variables inuence on ETo estimation, using ETo as the dependent variable and air temperature

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

71

ETo RM (mm yr-1)

1800

1600
y = 0.952x + 58.8
R = 0.952
1400

1200

1000

800
800

1000

1200

1400

1600

1800

ETo PM (mm yr-1)

1941-1950

1951-1960

1961-1970

1971-1980

1981-1990

1991-2000

2001-2110

1:1

Fig. 3. Relationship between annual reference evapotranspiration (ETo PM) and the values estimated with a regression model (ETo RM) using climatic
values of temperature (T), actual vapor pressure (ea ), wind speed (u2 ), and cloudiness (N) as independent variables for 7 successive decades from 1941 to
1950 to 20012010.

(T), actual vapor pressure (ea ), wind speed (u2 ), and cloudiness (N) as the independent variables. This procedure was repeated
independently for each decade with observed climate information only and is therefore described as a local mode analysis.
Fig. 3 shows the ETo calculated by the PenmanMonteith model (PM) and those estimated with the MR model for each of
the seven ten-year periods, and the tted regression curve for the complete set of data analyzed.
This graphical analysis, which is complemented with the statistical information in Table 2, shows that the MR model
based on the four climate elements captured most of the observed variability of ETo , despite the strong non-linear behavior
of the evapotranspiration phenomenon (Wang et al., 2014). The similarity between R2 and R2 adjusted by the degrees of
freedom reveals that the model behaves well using four predictors, with R2 values between 0.93 and 0.98 and estimation
errors between 19 and 33 mm yr1 . While T and u2 show a positive contribution to ETo , ea shows an inverse relationship,
as can be seen in Table 3. The behavior of the regression coefcient for N is not so constant and in some decades does not
reach statistical signicance. As this model represents more strictly the variation of ETo throughout the region, the lower
(higher) cloud cover and air humidity to the west (east) of this region would explain the higher (lower) ETo rate, typical of

28

28

Latitude S ()

30

1600

30

1500

32

32
1400

1300

34

34
1200

1100

36

36
1000

66

64

62

60

66

64

62

60

Longitude W ()

Fig. 4. Variation of mean annual reference evapotranspiration (ETo , mm yr1 ) for 19411950 (left) and 20012010 (right) in the central region of Argentina.
The mountain region with an altitude higher than 1000 m asl in grey.

72

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

Table 2
Multiple regression model results to estimate reference evapotranspiration (ETo ) using temperature (T), actual vapor pressure (ea ), wind speed (u2 ), and
cloudiness (N) as independent variables (climatological observed data only).
Decade

Terms

Std. Err. of B

R2

R2 adj

Std. Err. of estimate

19411950

Intercept
u2
N
ea
T
Intercept
u2
N
ea
T
Intercept
u2
N
ea
T
Intercept
u2
N
ea
T
Intercept
u2
N
ea
T
Intercept
u2
N
ea
T
Intercept
u2
N
ea
T
Intercept
u2
N
ea
T

256.1
24.7
30.7
56.9
91.9
191.2
22.0
34.4
60.3
100.3
191.2
22.0
34.4
60.3
100.3
27.4
21.4
12.9
67.1
108.8
132.7
20.1
2.9
66.0
102.0
33.8
18.3
54.0
60.5
98.0
9.3
25.4
21.5
60.3
99.7
162.0
22.3
13.7
62.9
99.9

138.0
2.3
13.3
4.8
6.7
133.8
2.2
20.7
4.5
5.5
133.8
2.2
20.7
4.5
5.5
219.5
2.5
31.4
6.1
9.5
169.8
2.2
26.3
5.3
7.5
142.4
2.0
25.3
3.7
4.6
143.4
2.2
33.9
4.3
4.7
55.4
0.8
8.3
1.8
2.4

1.855
10.517
2.299
11.831
13.694
1.428
9.882
1.661
13.319
18.118
1.428
9.882
1.661
13.319
18.118
0.125
8.719
0.410
10.966
11.420
0.782
9.237
0.109
-12.438
13.683
0.238
9.154
2.135
16.395
21.121
0.065
11.350
0.633
14.058
21.087
2.924
27.322
1.655
35.094
41.908

0.0821
0.0000
0.0353
0.0000
0.0000
0.1713
0.0000
0.1151
0.0000
0.0000
0.1713
0.0000
0.1151
0.0000
0.0000
0.9020
0.0000
0.6871
0.0000
0.0000
0.4457
0.0000
0.9149
0.0000
0.0000
0.8162
0.0000
0.0540
0.0000
0.0000
0.9492
0.0000
0.5394
0.0000
0.0000
0.00405
0.00000
0.10029
0.00000
0.00000

0.953

0.941

24.6

21

0.980

0.975

24.7

20

0.969

0.961

31.2

22

0.934

0.918

33.3

22

0.941

0.926

28.8

21

0.976

0.968

19.2

17

0.984

0.978

19.8

16

0.952

0.951

28.3

139

19511960

19611970

19711980

19811990

19912000

20012010

All

Table 3
Regression coefcients for the local and interpolated Linear Regression models to ETo estimation and the correlation between each variable and ETo
discriminated by decade.

Decade

Regression coefcients
u2
N

Local mode
24.7
30.7
19411950
19511960
22.0
34.4
22.0
34.4
19611970
21.4
12.9
19711980
20.1
2.9
19811990
18.3
54.0
19912000
25.4
21.5
20012010
Regional mode (altitude <1000 masl)
25.9
25.3
19411950
27.9
13.9
19511960
20.2
49.7
19611970
19.0
32.1
19711980
20.0
6.0
19811990
19.6
62.2
19912000
27.6
23.5
20012010

Correlation coefcients
u2
N

ea

56.9
60.3
60.3
67.1
66.0
60.5
60.3

91.9
100.3
100.3
108.8
102.0
98.0
99.7

0.283
0.572
0.333
0.282
0.182
0.112
0.431

57.4
57.7
57.7
60.1
66.2
61.1
61.5

94.3
96.0
100.1
100.2
103.8
101.6
102.5

0.049
0.481
0.138
0.088
0.169
0.327
0.066

ea

0.499
0.534
0.554
0.587
0.440
0.123
0.204

0.402
0.293
0.236
0.355
0.213
0.211
0.222

0.256
0.485
0.579
0.291
0.409
0.543
0.515

0.560
0.734
0.554
0.730
0.487
0.081
0.378

0.363
0.277
0.229
0.235
0.001
0.128
0.147

0.333
0.506
0.528
0.369
0.521
0.549
0.563

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

73

Table 4
Fraction of central region of Argentina with positive (+)/negative () trend and signicant (s)/non signicant (ns) change of reference evapotranspiration
(ETo ) and for the related variables (cloud cover, N; wind speed at 2 m, u2 ; vapor pressure,ea ; air temperature, T; and precipitation, PP); between 19411950
and 20012010. Fraction of territory with z > 1000 m asl is 0.022.
Variable

Territory z < 1000 masl

Statistical signication of S
ns
s (P < 0.10)

ETo
u2
N
ea
T
PP

0.395
0.755
0.336
0.877
0.591
0.978

0.914
0.912
0.749
0.797
0.850
0.434

0.002
0.005
0.193
0.000
0.033
0.000

0.062
0.061
0.036
0.181
0.095
0.544

0.583
0.223
0.642
0.101
0.387
0.000

0.064
0.066
0.229
0.181
0.128
0.544

the semiarid (humid) climates there. The spatial variation of ETo can be observed in Fig. 4 for the decades 19411950 and
20012010.
The introduction of N in the local model, however, appears signicant only for the 19411950 decade, marginally for
19511960, 19611970 and when the entire data set is analyzed, revealing a minor predictive ability compared to the other
three driving variables. The narrower range of N probably limits its ability to differentiate ETo behavior between regions,
and also because its observation is less systematic.
Having interpolated each of the four predictive variables into a 78 100 grid in the central region of Argentina, the same
analytical procedure was followed to obtain an interpolated model to estimate ETo in a cell-by-cell scheme. The regression
coefcients obtained for each decade and also the correlation coefcients between each of these variables individually and
ETo are shown in Table 3, to compare the local with the regional mode.
Despite some minor differences, the coefcients in Table 3 are very similar to each other when the local and interpolated
modes are compared, not only in their absolute values but also in the sign of the coefcients, so that the interpolated mode
can be considered as a reliable extension tool to express the territorial behavior of ETo . Also, assessing the variation over
time of dependent and independent variables is easier this way, because the data matrix remains the same throughout the
period. The air temperature shows a positive, generalized and principal inuence on ETo over time, as the regression and
correlation coefcients indicate, in agreement with Tang et al. (2011). As the sign indicates, ea shows a negative inuence
on ETo , and thus the more humid the atmosphere is, the lower is the ET rate, but the correlation is very weak in this case.
Zheng et al. (2009) showed that an increase in temperature resulted in an increase of pan evaporation, but that this effect
was partially compensated by the increase in vapor pressure.
In some decades, cloudiness also presents a signicant inverse relationship, indicating that the cloudier the climate, the
less solar energy is available, intrinsically reducing the evaporative capacity of the atmosphere (Irmak et al., 2012). Certainly,
as humid climates necessarily present more cloud cover, the evaporative potential is more reduced, as occurs in the east
central region of Argentina. In Fig. 4, annual ETo values are depicted for the decades 19411950 and 20012010, showing
the characteristic increase of evapotranspiration from east to west, and also the increasing aridity in the region.
Next, the particular inuence of each variable on the variation of ETo over time was established to analyze their spatial
extent. The linear relationships between ETo and the related climate variables were established to analyze the inuence one
by one of the main parameters that might cause variation in ETo . As rain increased signicantly between 1940 and 2010 in
the central region of Argentina (de la Casa and Ovando, 2014), this was also examined. Fig. 5 depicts the slope of the linear
regression line and those areas with a signicant relationship according to the t value (t must be 2.015 or 2.015 for
n = 7 and P < 0.1). The region where the altitude is more than 1000 masl is also depicted in these maps, indicating those areas
excluded from the analysis because of the particular inuence of the wind regime there (McVicar et al., 2010).
While the relationships of ETo with N, ea and PP are predominantly negative in almost all the region, the inuence of
u2 and T over ETo , on the other hand, is positive. This is similar to the results shown in Table 3, but suggests that each
variable exercises its regional inuence in a particular way. In detailed analyzes of the relationships between ETo and all the
meteorological variables, looking for the sensitivity of ETo to the variables on an annual basis, Irmak et al. (2012), over a 116yr period in the Platte River Basin (central Nebraska, USA), determined that ETo was signicantly and inversely correlated
to precipitation and relative humidity, and signicantly and positively correlated to maximum temperature (Tmax ), mean
temperature (T), vapor pressure decit (DPV), solar radiation (Rs ), and net radiation (Rn ). These results closely match those
of this study. In much of the region, it is likely that both positive and negative inuences may compensate each other and
so the net contribution to ETo would not be apparent.
Fig. 6 depicts those areas in the central region of Argentina that have a signicant increase or decrease of N, u2 , ea , T and
PP during the analysis period, and the fractions of territory for each of these categories are shown in Table 4. The information
about ETo change is also shown in Fig. 6 and Table 5, to analyze the regional correspondence between them. While these
maps show that there is a general absence of a signicant change over time in the territory, with the exception of rainfall,
each variable has its own particular behavior in each region, without order or continuity.
Despite the sharp increase of rain in the region between 1941 and 2010 (Viglizzo et al., 1995; de la Casa and Nasello,
2010; de la Casa and Ovando, 2014), this is not matched by cloudiness, since cloud cover decreases toward the east. The trend
related to rain behavior is positive only for ea in almost all the territory and the signicant variation is exclusively positive.

74

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679
28

28

(a)

(b)
30

Latitude S ()

Latitude S ()

30

32

32

34

34

36

36

66

64

62

60

66

Longitude W ()
28

62

60

28

(c)

(d)
30

Latitude S ()

30

Latitude S ()

64

Longitude W ()

32

32

34

34

36

36

66

64

62

60

66

Longitude W ()

64

62

60

Longitude W ()

28

(e)

Latitude S ()

30

32

34

36

66

64

62

60

Longitude W ()

Fig. 5. Slope for the relationship between reference evapotranspiration (ETo ) and cloud cover, N (a); wind speed at 2 m, u2 (b); vapor pressure, ea (c); air
temperature, T (d); and precipitation, PP (e) on an annual time step for 7 decadal periods from 1941 to 2010 in the central region of Argentina ( (green)
and (red) indicate those areas where the slope of the regression line is positive or negative signicant at the 90% condence level, respectively). Also
shown in grey the mountain region with an altitude higher than 1000 masl.

Since these small signicant changes are spatially neither uniform nor systematic, and the relationship of some variables
with ETo is direct while of others it is inverse, their inuence over evapotranspiration is unlikely to become apparent, as
Fig. 6(e) shows.
Finally, to explain why there is no signicant change in ETo over the period analyzed in most of the central territory of
Argentina, we decided to make a joint spatial assessment for each variable to establish those areas that show signicant
inuence on ETo and also exhibit a signicant temporal trend. Under the grid structure, it is feasible to implement a carto-

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

75

28

28

(b)

(a)
30

Latitude S ()

Latitude S ()

30

32

32

34

34

36

36

66

64

62

60

66

Longitude W ()
28

60

(d)

30

Latitude S ()

Latitude S ()

62

28

(c)

30

32

32

34

34

36

36

66

64

62

60

66

64

62

60

Longitude W ()

Longitude W ()
28

28

(e)

30

(f)

30

Latitude S ()

Latitude S ()

64

Longitude W ()

32

32

34

34

36

36

66

64

62

Longitude W ()

60

66

64

62

60

Longitude W ()

Fig. 6. Central region of Argentina with signicant increase (in green) or decrease (in red) of cloud cover, N (a); wind speed at 2 m, u2 (b); vapor
pressure,ea (c); air temperature, T (d); precipitation, PP (e); and reference evapotranspiration, ETo (f); for annual values between 1941 and 1950 and
20012010 (S must be 13 or 13 for n = 7 and P < 0.10). In blue, those areas where the relationship with reference evapotranspiration (ETo ) is also
signicant. Also shown in grey, the mountain region with an altitude higher than 1000 masl.

graphic analysis, combining in each cell the information of Fig. 5 with the S value of MannKendall to obtain the intersection
of both. Fig. 6 also shows that the areas in which there is a signicant relationship between each of the driving variables and
ETo , and the atmospheric element simultaneously shows a signicant time trend, reduce their acreage very noticeably.
Considering the inuence of ea on ETo , although the relation is negative in 28.5% of the territory (z < 1000 masl), air
humidity increased signicantly in only 18.2% of this area, restricting the occurrence of both conditions to just 7.8% of all

76

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

Table 5
Fraction of central region of Argentina (altitude (z) < 1000 m asl) with non-signicant, signicant decrease and signicant increase of reference evapotranspiration (ETo ) and for each of the variables driving the phenomenon.
Variables
u2
Without change in ETo (0.914)
0.734
ns
0.144

0.036
+
With signicant decrease of ETo (0.002)
ns
0.002

0.000
0.000
+
With signicant increase of ETo (0.062)
0.013
ns

0.049
0.000
+

ea

PP

0.822
0.033
0.058

0.874
0.005
0.034

0.734
0.000
0.180

0.411
0.000
0.503

0.002
0.000
0.000

0.002
0.000
0.000

0.002
0.000
0.000

0.000
0.000
0.002

0.025
0.000
0.037

0.036
0.000
0.026

0.061
0.000
0.002

0.024
0.000
0.039

the territory. This means that only in 7.8% of the central region of Argentina could the increase of ea explain any decrease of
ETo .
With respect to T, the sector which has a signicant relationship with ETo (increase) together with a signicant increasing
trend, represents only 3.9% of the central area of Argentina. This means that the temperature increase could explain the
increase in ETo for only a small fraction of the study region. Meanwhile, in another small fraction of the territory (0.5%), the
trend of T to decrease may be associated with decreased ETo .
Although N increases over time in part of the region (3.6%), the prevailing trend is to decrease (19.3%), and there is an
exclusively inverse relationship with ETo (23.7%), with the joint manifestation of both in only 2.8% of the territory. This
means that the decrease of N seems to be associated with increased ETo , but again strictly in a small sector of the central
area of Argentina (Fig. 6a).
While u2 shows a direct relationship with ETo in about half of the region (42.8%), the trend with a positive change of wind
speed occurs in 6.2% of the territory, which greatly reduces the potential contribution of this factor. Moreover, a signicant
simultaneous occurrence of both conditions is found in barely 0.5% of the region, and only here is it valid to attribute any
increase in ETo to u2 .
As regards territorial correspondence between changes in ETo and in the related variables, Table 5 presents information
indicating the fractions of the central area of Argentina where ETo has undergone signicant changes over time, whether
positive or negative, and those where the variation does not reach signicance. The proportion that has changed in each of
these areas is quantied and the nature of the change for the different variables that govern the evapotranspiration process.
These results show that, in almost the entire region under study (91%), the variation of ETo over time is non-signicant,
and there is also a widely prevalent lack of change for the set of atmospheric variables associated with this phenomenon
(Table 5). There is a signicant decrease in ETo only in 0.2% of the territory (Fig. 6), and overall it does not appear that this
can be attributed to the modication of a particular variable, since there is no signicant change for any of the variables,
except for rainfall. Finally, in the small region where ETo has increased signicantly (6.2%), there are some areas where the
variables controlling the process show no variation, but there is also a positive change in the case of air humidity, cloud
cover and also rainfall.
Reference (or Potential) ET has been shown to decline in different regions since the 1950s (Chattopadhyay and Hulme,
1997; Chen et al., 2005; Hobbins et al., 2004; Lawrimore and Peterson, 2000; Liu et al., 2004; Liu and Zeng, 2004; Roderick and
Farquhar, 2004; Thomas, 2000; Xu et al., 2006) and for more extended periods too (Irmak et al., 2012), and this seems to be the
more general behavior of ETo around the world (McVicar et al., 2012). However, in the western half of Iran, Tabari et al. (2011)
calculated an increase in the annual series of ETo at 7075% of the stations evaluated, while in most of the central region
of Argentina, the evidence discussed here indicates that average annual ETo values did not change signicantly between
1940 and 2010. Thus, the results conrm that the greater agricultural potential of this territory was subordinate only to the
increased rainfall during the period analyzed, widespread in 54.4% of the region (de la Casa and Ovando, 2014).
There is no scientic consensus about which meteorological variable(s) mainly inuence the general decrease in ETo all
over the world (Irmak et al., 2012). This is to be expected because the impact of individual meteorological variables, especially
the impact of their interactions on ETo , can change substantially from one region to another and can change temporally and
spatially even within the same region. Thus, the sensitivity of ETo to climate variables shows signicant variation between
locations, as well as between years within the same location. The complex behavior of the ET process requires the use of
the four-variable Penman procedure in order to properly evaluate the inuence of climate change on vegetation water use
(McKenney and Rosenberg, 1993; Kingston et al., 2009; Donohue et al., 2010; Irmak et al., 2012; McVicar et al., 2012).
The study of the variation of the four atmospheric factors controlling the process shows an absence of signicant temporal
change in these throughout the central territory of Argentina. Furthermore, visual inspection of Fig. 6 shows that, where
changes of a signicant nature occur, the trend is positive in some areas and negative in others. Tabari (2011) points out
that the main cause of the increasing ETo trend in Iran is an increase in air temperature in the region in the study period, but

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

77

found both positive and negative trends in the wind speed and relative humidity time series. For River Platte Basin, Irmak
et al. (2012) observed a signicant decline in ETo (0.3596 mm yr1 ) over an extended period. They also established a signicant increasing trend in both minimum temperature and mean temperature at a rate of 0.038 C yr1 and 0.0187 C yr1 ,
respectively, but a non-signicant increasing trend in u2 and VPD. In this case, they explain the decrease in ETo as likely
due to a signicant increase in precipitation resulting in a signicant reduction of Rs and Rn , because available energy is the
primary driver of ETo . The negative inuence of precipitation on ETo is reported by Cong et al. (2009), and was observed also
in the central region of Argentina but in only 5% of this (Fig. 5).
As reported by Yin et al. (2010), several studies attributed decreasing ETo trends to decreased sunshine duration or Rs
in Russia and USA (Peterson et al., 1995), China (Gao et al., 2006; Liu et al., 2004; Thomas, 2000; Wang et al., 2007), Israel
(Cohen et al., 2002), and, in general, for the Northern Hemisphere (Roderick and Farquhar, 2004). Cloud cover (N), as an
indicator representing the opposite of energy availability, shows signicant variation in only 22.9% of the central region of
Argentina, with a predominantly decreasing trend (19.3%).
While the ETo decrease was attributed to lower wind speed in Australia (Rayner, 2007; Roderick et al., 2007), the Tibetan
Plateau (Chen et al., 2006; Zhang et al., 2007), and Canada (Burn and Hesch, 2007), the wind speed variation over time was
not signicant for 93% of the study region in Argentina and, instead, a signicant increase in wind speed was found in 6% of
the area.
Chattopadhyay and Hulme (1997) attributed the decrease in ETo in India to an increase in relative humidity. Xu et al.
(2006) also show that ETo is most sensitive to variations in relative humidity, but attributed the decrease in ETo over the
Yangtze River catchment area to a combination of changes in Rs and u2 . The actual contribution of each variable to the
decreasing trend in ETo is a combined effect of the sensitivity of evapotranspiration to each variable and the temporal
change of this variable.
Based on meteorological data of 317 stations, Cong et al. (2009) explained the evaporation paradoxin China considering
two periods: one from 1956 to 1985 and the other from 1986 to 2005. While the pan evaporation decrease before 1985 was
produced by the reduction of radiation and wind speed, its increase after 1986 was caused by the vapor pressure decit rise
due to strong warming. The positive inuence of T and u2 on the ETo rate was also established for China, both temporally
and territorially, while the contribution of N, as a term representing the opposite of energy availability, and of ea proved
negative.

4. Summary and conclusions

The climate in the central region of Argentina is water-limited. Thus, the temporal change of the reference evapotranspiration (ETo ) may have important hydrological, ecological and agricultural consequences. Annual territorial and temporal
ETo variation between 1941 and 2010 was analyzed for these 7 decades, by assessing the behavior of the four atmospheric
variables which drive it: temperature (T), vapor pressure (ea ), wind speed (u2 ), and cloud cover (N) to represent solar
radiation.
The analytical procedure was organized sequentially. First, the particular inuence of each variable on ETo was evaluated
from a multiple regression model and also with a simple correlation analysis, using climate data exclusively from the
observation network (local mode). The results indicate that, despite the non-linear nature of the phenomenon, the linear
regression model can explain almost all the variability of ETo obtained with the reference formula of PenmanMonteith.
All the variables were then interpolated in a grid scheme of 78 100 cells and the correlation and regression analysis
was repeated (regional mode). As there are only slight differences between the coefcients from both procedures, local
and regional, the regionalization procedure seems to be reliable. Next, linear relationships between ETo and the different
key atmospheric variables, including precipitation (PP), were determined for each grid pointand the t value was used to
establish those sectors where angular coefcient is statistically signicant (P < 0.1). All the variables show extensive areas
with a signicant relationship with ETo , which is positive for T (in 17.9% of territory) and u2 (42.8%), but is negative for N
(23.7%), ea (28.5%) and PP (5.4%). Later, from the MannKendall test for climate time series with 7 terms, the area with a
signicant (P < 0.1) trend of change was determined for each variable. Except for PP, a non-signicant change is found in the
region for all the variables.
Finally, interception of the grids was performed to establish their joint occurrence. The results indicate that, for the central
region of Argentina, while T and u2 show a direct relationship with ETo , the relation with ea , N and PP is inverse, and that
both conditions are equally present. Most of the region analyzed (>91%) shows a non-signicant variation of ETo over time,
just as a general non-signicant change is seen for each variable, both in relation to ETo and to its own trend of change.
Thus, the benecial change of agricultural suitability reported in the region during the study period was seen to be produced
almost exclusively by increasing PP.

Appendix A. Supplementary data

Supplementary data associated with
http://dx.doi.org/10.1016/j.ejrh.2015.11.009.

this

article

can

be

found,

in

the

online

version,

at

78

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

References
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop EvapotranspirationGuide-Lines for Computing Crop Water Requirements. FAO Irrigation and
Drainage Paper 56, Rome, Italy.
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 2006. Evapotranspiracin del cultivo. Guas para la determinacin de los requerimientos de agua de los
cultivos. In: Estudio FAO Riego Y Drenaje No 56. Organizacin de las Naciones Unidas para la Agricultura y la Alimentacin (FAO), Roma, pp. 300.
Bandyopadhyay, B., Bhadra, A., Raghuwnshi, N.S., Singh, R., 2009. Temporal trends inestimates of reference evapotranspiration over India. J. Hydrol. Eng.
14, 508515.
Bates, B.C., Kundzewicz, Z.W., Wu, S., Palutikof, J.P. (Eds.), 2008. Documento Tcnico del Grupo Intergubernamental de Expertos Sobre el Cambio
Climtico. Secretara del IPCC, Ginebra, p. 224.
Bureau of Meteorology, 2007. Time Series-Australian Climate Variability and Change. Australian Government Bureau of Meteorology, Available from:
<http://www.bom.gov.au/cgi-bin/climate/change/timeseries.cgi?graph=tmean&area=aus&season=0112&ave yr=0> (last accessed: 06.04.15.).
Burn, D.H., Hesch, N.M., 2007. Trends in evaporation for the Canadian Prairies. J. Hydrol. 336, 6173.
Burrough, P.A., 1986. Principles of Geographical Information Systems for Land Resources Assessment. Oxford University Press, Clarendon Press, Oxford,
pp. 197.
Chattopadhyay, N., Hulme, M., 1997. Evaporation and potential evapotranspiration in India under conditions of recent and future climatic change. Agric.
For. Meteorol. 87, 5574.
Chen, D.L., Gao, G., Xu, C.Y., Guo, J., Ren, G.Y., 2005. Comparison of the Thornthwaite method and pan data with the standard PenmanMonteith estimates
of reference evapotranspiration in China. Clim. Res. 28, 123132.
Chen, S., Liu, Y., Thomas, A., 2006. Climatic change on the Tibetan Plateau: potential evapotranspiration trend from 1961 to 2006. Clim. Change 76,
291319.
Cohen, S., Ianetz, A., Stanhill, G., 2002. Evaporative climate changes at Bet Dagan, Israel, 19641998. Agric. For. Meteorol. 111, 8391.
Cong, Z.T., Yang, D.W., Ni, G.H., 2009. Does evaporation paradox exist in China? Hydrol. Earth Syst. Sci. 13, 357366.
de la Casa, A.C., Nasello, O., 2010. Breakpoints in annual rainfall trends over Cordoba, Argentina. Atmos. Res. 95, 419427.
de la Casa, A.C., Ovando, G.G., 2014. Climate change and its impact on agricultural potential in the central region of Argentina between 1941 and 2010.
Agric. For. Meteorol. 195196, 111.
Donohue, R.J., Roderick, M.L., McVicar, T.R., 2007. On the importance of including vegetation dynamics in Budykos hydrological model. Hydrol. Earth Syst.
Sci. 11 (2), 983995.
Donohue, R.J., McVicar, T.R., Roderick, M.L., 2010. Assessing the ability of potential evaporation formulations to capture the dynamics in evaporative
demand within a changing climate. J. Hydrol. 386, 186197.
Durre, I., Williams Jr., C.N., Yin, X., Vose, R.S., 2009. Radiosonde-based trends in precipitable water over the Northern Hemisphere: an update. J. Geophys.
Res. 114, D05112.
Gao, G., Chen, D.L., Ren, G.Y., Chen, Y., Liao, Y.M., 2006. Spatial and temporal variations and controlling factors of potential evapotranspiration in China:
19562000. J. Geogr. Sci. 16, 312.
Gilbert, R.O., 1987. Statistical Methods for Environmental Pollution Monitoring. VanNostrand Reinhold, New York.
Golden Software, 1997. SURFER for Windows V7. Golden Software, Golden, CO.
Hobbins, M.T., Ramrez, J.A., Brown, T.C., 2004. Trends in pan evaporation andactual evapotranspiration across the conterminous US: paradoxical or
complementary? Geophys. Res. Lett. 31 (13), http://dx.doi.org/10.1029/2004GL019846, L13503.
Irmak, S., Kabenge, I., Skaggs, K.E., Mutiibwa, D., 2012. Trend and magnitude of changes in climate variables and reference evapotranspiration over 116-yr
period in the Platte River Basin, Central Nebraska-USA. J. Hydrol. 420421, 228244.
Keane, R.E., Parsons, R.A., Hessburg, P.F., 2002. Estimating historical range and variation of landscape patch dynamics: limitations of the simulation
approach. Ecol. Modell. 151, 2949.
Kingston, D.G., Todd, M.C., Taylor, R.G., Thompson, J.R., Arnell, N.W., 2009. Uncertainty in the estimation of potential evapotranspiration under climate
change. Geophys. Res. Lett. 36, http://dx.doi.org/10.1029/2009gl040267, L20403.
Lawrimore, J.H., Peterson, T.C., 2000. Pan evaporation trends in dry and humid regions of the United States. J. Hydrometeorol. 1, 543546.
Liu, C.M., Zeng, Y., 2004. Changes of pan evaporation in the recent 40 years in the Yellow River Basin. Water Int. 29, 510516.
Liu, B., Xu, M., Henderson, M., Gong, W., 2004. A spatial analysis of pan evaporation trends in China, 19552000. J. Geophys. Res. 109,
http://dx.doi.org/10.1029/2004JD004511, D15102.
McKenney, M.S., Rosenberg, N.J., 1993. Sensitivity of some potential evapotranspiration estimation methods to climate change. Agric. For. Meteorol. 64,
81110.
McVicar, T.R., Jupp, D.L.B., 1998. The current and potential operational uses of remote sensing to aid decisions on drought exceptional circumstances in
Australia: a review. Agric. Syst. 57, 399468.
McVicar, T.R., Van Niel, T.G., Li, L.T., Roderick, M.L., Rayner, D.P., Ricciardulli, L., Donohue, R.J., 2008. Wind speed climatology and trends for Australia,
19752006: capturing the stilling phenomenon and comparison with near-surface reanalysis output. Geophys. Res. Lett. 35,
http://dx.doi.org/10.1029/2008GL035627, L20403.
McVicar, T.R., Van Niel, T.G., Roderick, M.L., Li, L.T., Mo, X.G., Zimmermann, N.E., Schmatz, D.R., 2010. Observational evidence from two mountainous
regions that near-surface wind speeds are declining more rapidly at higher elevations than lower elevations: 19602006. Geophys. Res. Lett. 37,
http://dx.doi.org/10.1029/2009GL042255, L06402.
McVicar, T.R., Roderick, M.L., Donohue, R.J., Li, L.T., Van Niel, T.G., Thomas, A., Grieser, J., Jhajharia, D., Himri, Y., Mahowald, N.M., Mescherskaya, A.V.,
Kruger, A.C., Rehman, S., Dinpashoh, Y., 2012. Global review and synthesis of trends in observed terrestrial near-surfacewind speeds: Implications for
evaporation. J. Hydrol. 416417, 182205.
New, M., Todd, M., Hulme, M., Jones, P., 2001. Precipitation measurements and trends in the twentieth century. Int. J. Climatol. 21, 18891922,
http://dx.doi.org/10.1002/joc.680.
Penman, H.L., 1948. Natural evaporation from open water, bare soil and grass. Proc. R. Soc. Lond. A 193, 120145.
Petersen, G.W., Bell, J.C., McSweeney, K., Nielsen, G.A., Robert, P.C., 1995. Geographic information systems in agronomy. Adv. Agron. 55, 67111.
Peterson, T.C., Golubev, V.S., Groisman, P.Y., 1995. Evaporation losing its strength. Nature 377, 687688.
Raupach, M.R., Briggs, P.R., Haverd, V., King, E.A., Paget, M., Trudinger, C.M., 2009. Australian Water Availability Project (AWAP). CSIRO Marine and
Atmospheric Research Component. Final Report for Phase 3. Centre for Australian Weather and Climate Research Technical Report No. 013.
Rayner, D.P., 2007. Wind run changes: the dominant factor affecting pan evaporation trends in Australia. J. Climate 20, 33793394.
Roderick, M.L., Farquhar, G.D., 2004. Changes in Australian pan evaporation from 1970 to 2002. Int. J. Climatol. 24, 10771090.
Roderick, M.L., Rotstayn, L.D., Farquhar, G.D., Hobbins, M.T., 2007. On the attribution of changing pan evaporation. Geophys. Res. Lett. 34,
http://dx.doi.org/10.1029/2007GL031166.
Smn, 1958. Estadsticas climatolgicas 19411950 (climatological statistics 19411950). publicacin B1, No 3. In: Servicio Meteorolgico Nacional.
Buenos Aires, Argentina.
Smn, 1975. Estadsticas climatolgicas 19511960 (climatological statistics19511960. Serie B, No 6). In: Servicio Meteorolgico Nacional. Buenos Aires,
Argentina.
Smn, 1981. Estadsticas climatolgicas 19611970 (climatological statistics 19611970. Publicacin B, No 35). In: Servicio Meteorolgico Nacional.
Buenos Aires, Argentina.

A.C. de la Casa, G.G. Ovando / Journal of Hydrology: Regional Studies 5 (2016) 6679

79

Smn, 1986. Estadsticas Meteorolgicas 19711980 (climatological Statistics 19711980. Publicacin B, No36). In: Servicio Meteorolgico Nacional.
Buenos Aires, Argentina.
Smn, 1992. Estadsticas climatolgicas 19811990 (climatological statistics 19811990. Publicacin B, No 37). In: Servicio Meteorolgico Nacional.
Buenos Aires, Argentina.
Smn, 2007. Estadsticas climatolgicas 19912000 (climatological statistics 19912000. Available on CD.). In: Servicio Meteorolgico Nacional. Buenos
Aires, Argentina.
Smn, 2012. Estadsticas climatolgicas 20012010 (climatological statistics 20012010. Available on CD). In: Servicio Meteorolgico Nacional. Buenos
Aires, Argentina.
Tabari, H., Maro, S., Aeini, A., Talaee, P.H., Mohammadi, K., 2011. Trend analysis ofreference evapotranspiration in the western half of Iran. Agric. For.
Meteorol. 151, 128136.
Tabios, G.Q., Salas, J.D., 1985. A comparative analysis of techniques for spatial interpolation of precipitation. Water Resour. Bull. 21, 365L380.
Tang, B., Tong, L., Kang, S., Zhang, L., 2011. Impacts of climate variability onreference evapotranspiration over 58 years in the Haihe river basin of north
China. Agric. Water Manage. 98, 16601670.
Thomas, A., 2000. Spatial and temporal characteristics of potential evapotranspiration trends over China. Int. J. Climatol. 20, 381396.
Thornthwaite, C.W., 1948. An approach toward a rational classication of climate. Geogr. Rev. 38, 5594.
Trenberth, K.E., Jones, P.D., Ambenje, P., Bojariu, R., Easterling, D., Klein Tank, A., Parker, D., Rahimzadeh, F., Renwick, J.A., Rusticucci, M., Soden, B., Zhai, P.,
2007. Observations: surface and atmospheric climate change. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M.,
Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK/New York, NY, USA.
Viglizzo, E.F., Roberto, Z.E., Filippin, M.C., Pordomingo, A.J., 1995. Climate variability and agroecological change in the Central Pampas of Argentina. Agric.
Ecosyst. Environ. 55, 716.
Wang, Y., Jiang, T., Bothe, O., Fraedrich, K., 2007. Changes of pan evaporation and reference evapotranspiration in the Yangtze River basin. Theor. Appl.
Climatol. 90, 1323.
Wang, K., Dickinson, R.E., 2012. A review of global terrestrial evapotranspiration: observation, modeling, climatology, and climatic variability. Rev.
Geophys. 50, http://dx.doi.org/10.1029/2011rg000373, RG2005.
Wang, W., Zou, S., Luo, Z., Zhang, W., Chen, D., Kong, J., 2014. Prediction of the reference evapotranspiration using a chaotic approach. The Scientic World
Journal 2014, http://dx.doi.org/10.1155/2014/347625 Article ID 347625.
Wild, M., 2009. Global dimming and brightening: a review. J. Geophys. Res. 114, http://dx.doi.org/10.1029/2008jd011470, D00D16.
Yin, Y., Wu, S., Chen, G., Dai, E., 2010. Attribution analyses of potential evapotranspiration changes in China since the 1960. Theor. Appl. Climatol. 101,
1928.
Xu, C.Y., Gong, L.B., Tong, J., Chen, D.L., 2006. Decreasing reference evapotranspiration in a warming climate a case of Changjiang (Yangtze) River
catchment during 19702000. Adv. Atmos. Sci. 23, 513520.
Zhang, Y.Q., Liu, C.M., Tang, Y.H., Yang, Y.H., 2007. Trends in pan evaporation and reference and actual evapotranspiration across the tibetan plateau. J.
Geophys. Res. 112, http://dx.doi.org/10.1029/2006JD008161, D12110.
Zheng, H., Liu, X., Liu, C., Dai, X., Zhu, R., 2009. Assessing contributions to pan evaporation trends in haihe river basin, China. J. Geophys. Res. Atmos. 114,
http://dx.doi.org/10.1029/2009jd012203, D24105.