Water Resour Manage DOI 10.1007/s11269-017-1638-1

Quantitative Trend, Sensitivity and Contribution Analyses of Reference Evapotranspiration in some Arid Environments under Climate Change Milad Nouri 1 & Mehdi Homaee 2 & Mohammad Bannayan 3

Received: 7 September 2016 / Accepted: 19 March 2017 # Springer Science+Business Media Dordrecht 2017

Abstract The temporal trend of reference crop evapotranspiration (ET0) and contribution of associated meteorological factors to the ET0 trend were assessed for 17 arid areas. Sensitivity of ET0 to changes in key meteorological variables was also analyzed. To study temporal trend of ET0, Mann-Kendall trend test was employed. Quantitative contribution and sensitivity analyses were carried out, respectively, using a dimensionless relative sensitivity coefficient and detrending method. Results indicated that ET0 has an increasing trend in 70.6, 64.7, 70.6, 76.5 and 70.0%, of sites respectively, in winter, spring, summer, autumn and entire year. This positive trend was significant (p ≤ 0.05) in 47.0, 35.3, 35.3, 29.4 and 35.3% of sites, respectively, for the same seasons. There was a significant change-point in winter, spring, summer, autumn and annual ET0 series at 64.7, 52.9, 64.7, 64.7 and 82.3% of stations, respectively. In 35.3 and 35.3% of sites, solar radiation and wind speed were the most sensitive climatic factors on ET0, respectively. ET0 exhibited the highest sensitivity to the relative humidity changes in coastal sites. Changes of wind speed contributed much more than other factors to the annual ET0 trend in 58.8% of investigated sites. The negative trend in wind speed nearly nullified the positive effects of increased air temperature on ET0 over 1966– 2012 in 23.5% of stations. Changes in ET0 were attributed to wind speed changes in most locations. Given the upward trend of ET0 in the majority of locations, proper water management is required to avoid negative impacts of climate change in arid regions.

* Mehdi Homaee [email protected]

1

Department of Soil Science, Faculty of Agriculture, Tarbiat Modares University, P.O. Box 14115-336, Tehran, Iran

2

Department of Irrigation and Drainage, Faculty of Agriculture, Tarbiat Modares University, P.O. Box 14115-336, Tehran, Iran

3

Department of Agronomy, Faculty of Agriculture, Ferdowsi University of Mashhad, P.O. Box 91775-1163, Mashhad, Iran

M. Nouri et al.

Keywords Detrending . Global warming . Sensitivity coefficient . Water management

1 Introduction Evapotranspiration as a key component of the water and energy cycles returns approximately 60% of land precipitation back to the atmosphere and uses up more than 50% of the total absorbed solar energy by land surface (Homaee et al. 2002; Kaniewski et al. 2012; Seneviratne et al. 2010). Due to considerable complexities of measuring water flux through crops, crop evapotranspiration is mostly estimated based on reference crop evapotranspiration (Kite and Droogers 2000; Xie and Zhu 2013). Reference crop evapotranspiration, denoted as ET0, is of great importance in estimating crop water requirement and scheduling, planning and managing irrigation systems (Godfray et al. 2010; Homaee et al. 2002; Irmak et al. 2006; Maestro et al. 2014). In addition, ET0 plays a key role in determining regional actual evapotranspiration and preparing input data to hydrological waterbalance models (Gong et al. 2006; Homaee and Schmidhalter 2008). ET0 is defined by Allen et al. (1998) as the evapotranspiration rate of a hypothetical reference crop which has an assumed height of 12 cm, a fixed surface resistance of 70 s m−1 and an albedo of 0.23. Among all methods proposed for the ET0 estimation, the Penman-Monteith FAO 56 (PMF-56) equation has been worldwide used to calculate ET0 in different climatic conditions (Allen et al. 1998). Iran is located in Middle East and its climate is predominantly semi-arid and arid (the waterlimited ecosystems) (Nouri et al. 2016). Agricultural activities and meeting food security are, therefore, largely depended on implementing irrigation. Agriculture sector consumes by far the largest portion of available water resources in Iran (Al-Faraj et al. 2016; Alizadeh and Keshavarz 2005). As a result, considering major role of ET0 in irrigation scheduling and planning, assessing temporal trend of ET0 can be very helpful to decision makers in Iran. Rising greenhouse gases chiefly as a consequence of landuse change, agricultural activities and fossil fuel overuse has caused climate change and global warming over the globe (IPCC 2013). Results of studies conducted to analyze the long-term trend of some climatic variables such as precipitation (Modarres and Sarhadi 2009; Some'e et al. 2012) and temperature (Tabari and Talaee 2011) over Iran revealed that the climate of Iran is changing. Quantitatively assessing the contribution of each key meteorological variable to ET0 changes is, hence, highly needed in Iran’s changing climate. Using multiple stepwise regression (a qualitative method), Dinpashoh et al. (2011) found relative humidity in wintertime and wind speed in other seasons as the most important climatic factor contributing to ET0 changes during 1965–2005 in 16 stations located over Iran. Recently, detrending method has been employed to quantitatively investigate the contribution of the controlling variables to the ET0 trend (Huo et al. 2013; Liu et al. 2010; Xie et al. 2015; Xu et al. 2006). Using detrending method, Xu et al. (2006) concluded that the decreasing trends in the net total radiation and wind speed are the main causes of the negative trend detected in ET0 for the Changjiang catchment, China. Liu et al. (2010) also reported that much of the ET0 increase and decrease is, respectively, due to temperature rise and declined wind speed in the Yellow River Basin, china by means of detrending. Huo et al. (2013) concluded that the trend in wind speed made a much larger contribution to the ET0 changes in arid northwestern regions of China by applying detrending. Sensitivity analysis is necessary to understand relative role of each key meteorological variable in modeling ET0 (Saxton 1975). Analyzing the sensitivity of an ET0 equation is of a key importance in determining the required accuracy for the measurement of variables used for ET0 estimation (Irmak et al. 2006). Moreover, quantifying the sensitivity of ET0 can determine inappropriate sampling of which factor would cause a greater error in estimating ET0 (Hupet

Quantitative Trend, Sensitivity and Contribution Analyses of Reference...

and Vanclooster 2001). The results of studies addressed the ET0 sensitivity showed that it varies with location and climate (Debnath et al. 2015; Gao et al. 2016; Huo et al. 2013; Song and Su 2015). The ET0 sensitivity analysis seems, however, not to be well-addressed in Iran (Mosaedi et al. 2016; Sharifi and Dinpashoh 2014). Using sensitivity curves, Sharifi and Dinpashoh (2014) concluded that average temperature and actual vapor pressure are the most and least sensitive variables on the annual ET0, respectively, in eight sites located in Iran. Based on sensitivity curves, Mosaedi et al. (2016) found that relative humidity and maximum temperature are the most influential climatic factors on ET0 in 5 stations in Iran over 1963–2007. Since there is merely one study (i.e. Dinpashoh et al. 2011) qualitatively addressed the ET0 contribution, further assessments to quantitatively investigate the contribution of climatic changes to the ET0 dynamics over arid areas of Iran are required. In addition, there is no study to quantitatively analyze the ET0 sensitivity using sensitivity coefficient approach and the ET0 sensitivity analyses conducted in Iran (i.e. Mosaedi et al. 2016; Sharifi and Dinpashoh 2014) have used a same graphical method. Furthermore, the assessments investigated the ET0 contribution and sensitivity in Iran have considered relatively few numbers of sites for different climates (from hyper-arid to humid). Undertaking a comprehensive quantitative analysis of the ET0 sensitivity and contribution using more sites and with a particular focus on arid regions seems, thus, to be very helpful to decision makers for controlling irrigation water demand in Iran. Furthermore, there is a need for analyzing the ET0 change-point in Iran as no investigation has yet addressed this issue. Therefore, this study aimed to assess i) the trend and change-point occurrence in seasonal and annual reference evapotranspiration (ET0, mm), average temperature (Tmean, °C), wind speed (U, m s−1), solar radiation (SR, MJ m−2 d−1) and relative humidity (RH, %) series, ii) the sensitivity of ET0 to changes in average temperature (Tmean), wind speed (U), solar radiation (SR) and relative humidity (RH) and iii) the contribution of average temperature (Tmean), wind speed (U), solar radiation (SR) and relative humidity (RH) to the ET0 trend in 17 arid stations of Iran during 1966–2012.

2 Materials and Methods 2.1 Data Monthly meteorological data of 17 arid stations (Fig. 1) including minimum temperature (Tmin), maximum temperature (Tmax), wind speed (U), sunshine hours (SH, h) and relative humidity (RH) were obtained from the Iran Meteorological Organization (IRIMO) in 1966–2012. The climate of all stations is arid based on UNEP (1992) aridity index.

2.2 The Penman–Monteith FAO-56 (PMF-56) Model In this study, the Penman–Monteith FAO-56 (PMF-56) method was used to estimate ET0 (Allen et al. 1998):

ET 0 ¼

900 U ðes −ea Þ T mean þ 273 Δ þ γ ð1 þ 0:34U Þ

0:408ΔðRn −GÞ þ γ

ð1Þ

where ET0 is the reference crop evapotranspiration (mm d−1), Δ is the slope of saturation vapor pressure curve (kPa °C−1), Rn is the net radiation at the reference crop surface

M. Nouri et al.

Fig. 1 The location of surveyed stations in the studied area

(MJ m−2 d−1), G is the soil heat flux density (MJ m−2 d−1), Tmean is the daily mean air temperature at 2 m height (°C), U is the average wind speed at 2 m height (m s−1), es is the saturation vapor pressure (kPa), ea. is the actual vapor pressure (kPa), es-ea. is the saturation vapor pressure deficit (kPa) and γ represents the psychrometric constant (kPa °C−1).

2.3 Trend Analysis Nonparametric rank-based Mann-Kendall test was used to temporally detect significant ET0 and climatic factors changes. The Mann-Kendall test statistics is calculated as follows (Yue et al. 2002): n−1

S¼ ∑

n
 ∑ sign x j −xk

ð2Þ

k¼1 j¼kþ1

8

 < 1
x j −xk  > 0 sign x j −xk ¼ 0
x j −xk  ¼ 0 : −1 x j −xk < 0

ð3Þ

Quantitative Trend, Sensitivity and Contribution Analyses of Reference...

VarðS Þ ¼

  m ½nðn−1Þð2n þ 5Þ− ∑ ti ðti −1Þð2t i þ 5Þ =18

ð4Þ

i¼1

8 pffiffiffiffiffiffiffiffiffiffiffiffiffiffi < S−1= VarðS Þ Z¼ 0 pffiffiffiffiffiffiffiffiffiffiffiffiffiffi : S þ 1= VarðS Þ

S>0 S¼0 S<0

ð5Þ

where S represents the Mann-Kendall statistics, n denotes data set length, xj and xk are the sequential data values, ti is the number of ties of extent i, m is number of the tied groups and Z is the standardized Mann-Kendall statistics. In this study, the trend-free prewhitening approach (TFPW) (Yue et al. 2002) was applied to eliminate the effect of serial correlation on the trend test.

2.4 Time Series Homogeneity Analysis The non-parametric method introduced by Pettitt (1979) was employed to detect any significant change-point in the studied time series. The test considers a sequence of random variables X1, X2, ..., XT, which has a change-point at τ. It is noted that Xt for t = 1, 2, ..., τ and t = τ + 1,..., T has a common distribution function of F1(x) and F2(x), respectively, which F1(x) ≠ F2(x). The null and alternative hypothesis of the test are, respectively, H0: τ = T and H1: 1 ≤ τ < T (Busuioc and von Storch 1996; Pettitt 1979). The following non-parametric statistic is used to test H0 against H1:   K T ¼ max U t;T  1 ≤ t< T

ð6Þ

The numbers of Ut,T can be calculated iteratively as follows: T
 U t;T ¼ U t−1;T þ ∑ sign xt −x j for t ¼ 1; :::; T

ð7Þ

j¼1

The significant probability of a change-point (p) associated with KT is approximated by (Pettitt 1979): h
i p≈2exp −6ðK T Þ2 = T 3 þ T 2

ð8Þ

When p is lower than the considered significance level (0.05 for this study), H0 is rejected.

2.5 The Sensitivity Analysis The partial derivatives can be used to analyze the response of ET0 estimated by PMF-56 (a mathematically tractable equation) to the perturbations of the controlling independent variables (Saxton 1975). Using the partial derivatives, comparing the results of sensitivity analysis of ET0 to different variables may, however, be very difficult and even misleading as different variables have different ranges and dimensions in a multivariable mathematical equation (e.g.

M. Nouri et al.

PMF-56). A dimensionless relative sensitivity coefficient firstly proposed by McCuen (1974) and Saxton (1975) has been thus widely employed in the hydrology literature for quantitative sensitivity analysis (Estévez et al. 2009; Gao et al. 2016; Gong et al. 2006; Huo et al. 2013). The relative sensitivity coefficient (SC) is given by:  SCxi ¼ lim

Δxi →0

ΔET 0 =ET 0 Δxi =xi

¼

∂ET 0 xi : ∂xi ET 0

ð9Þ

where SCxi is the sensitivity coefficient of ET0 to xi (the independent factor). It should be noted that Eq. 9 is the first term of Taylor’s expansion (Saxton 1975). In this investigation, Eq. 9 was used to quantitatively assess the sensitivity of ET0 to the changes in the controlling independent variables i.e. Tmean, RH, Rs and U for the studied stations. The positive (negative) magnitude of SCxi logically implies an increment (a decline) in ET0 as xi increases (decreases). Assuming all other variables are constant, a ±10% perturbation of xi would yield a ±1% change in ET0 if the SCxi equals 0.1. Note that the greater the SCxi, the more the sensitivity of ET0 to xi is expected (Gong et al. 2006; Huo et al. 2013).

2.6 Climatic Changes Contribution to the ET0 Trend In the current study, the detrending method was used to quantitatively determine the contribution of Tmean, RH, Rs and U to the ET0 changes. In order to implement the detrending procedure in first step, the trend in Tmean, RH, Rs and U was removed and the series became stationary (Xu et al. 2006). Afterwards, for assessing the contribution of a specific variable (for instance Tmean), ET0 was recalculated with the detrended data series (the detrended Tmean) while the original data set was used for other variables. Finally, the average difference between ET0 calculated with original data and ET0 recalculated with detrended data was computed. This difference quantifies the contribution of a specific variable to ET0 trend. The larger absolute difference indicates the higher contribution of a given variable to the ET0 changes (Huo et al. 2013; Xu et al. 2006).

3 Results and Discussions 3.1 ET0 Trend Analysis In 12 out of 17 surveyed stations, positive trends were detected in annual ET0 during 1966– 2012 (Fig. 2). The significant trend in ET0 series was, however, identified only in 6 sites (i.e. Zabol, Zahedan, Iranshahr, Semnan, Kashan and Bam) at the 95% confidence level. It is worth mentioning that 4 out of these 6 sites with significant trend in annual ET0 i.e. Zabol, Zahedan, Iranshahr and Bam were located in the southeast of the country indicating considerable impacts of climate change on ET0 over this region of the country. Annual ET0 at Birjand, Boshehr, Esfahan, Shahroud and Kerman exhibited an insignificant decreasing trend. According the results presented in Fig. 2, the trend in ET0 was positive in winter, spring, summer and autumn time for 70.6, 64.7, 70.6 and 76.5% of the investigated locations, respectively. However, 47.0, 35.3, 35.3 and 29.4% of all studied stations showed significant positive trend in ET0 during winter, spring, summer and autumn, respectively.

Quantitative Trend, Sensitivity and Contribution Analyses of Reference... 8

(a) ET

6

0

4 2 0 -2 -4 -6 wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

8

(b) T

6

mean

4 2 0 -2 -4 wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

6

(c) U 4 2 0 -2 -4 -6 wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi us an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

Fig. 2 The Mann-Kendall Z values for ET0 (a) Tmean (b), U (c), RH (d) and RS (e) and ET0 during winter (black), spring (red), summer (yellow), autumn (green) and entire year (blue). The dotted lines indicate the significant level at 5%

Based on the Pettitt homogeneity test, except at Birjand, Boshehr and Bandar-e-lengeh sites, a significant change-point was detected in annual ET0 series (Table 1). Further, there is a change-point in winter, spring, summer and autumn ET0 series at 64.7, 52.9, 64.7 and 64.7% of sites, respectively. At both annual and seasonal scales, the change-point mainly occurred in the1990s for the ET0 time series.

3.2 Trend Analysis of the Controlling Atmospheric Variables On annual scale, Tmean and Rs series exhibited significant upward trends in 82.3% and 29.4% of the surveyed stations, respectively, in 1966–2012 (Fig. 2). The increase in Tmean is attributed to global warming induced by elevated CO2. The trend test detected a significant decreasing trend in annual RH at 41.2% of the sites. The declined RH and increased Rs can be considered as the sequences of the decrease of cloudiness in arid regions of Iran leading to a negative trend in annual precipitation in these areas (Some'e et al. 2012). There was a significant upward

M. Nouri et al. 4

(d) RH 2

0

-2

-4

-6 wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi us an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

6

(e) Rs

5 4 3 2 1 0 -1 -2 -3

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi us an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

Fig. 2 (continued)

trend in Bam, Zabol, Kashan and Iranshahr stations and a significant downward trend in Boshehr, Esfahan and Shahroud sites in annual U. Further, 52.9, 82.3, 76.5 and 76.5% of all sites revealed a significant positive trend in Tmean during winter, spring, summer and autumn, respectively. Moreover, about 17.6, 35.3, 35.3 and 5.9% of stations had a significant upward trend in Rs, respectively, during winter, spring, summer and autumn. A significant negative trend was also detected in approximately 11.8, 47.0, 41.2 and 17.6% of sites for RH and 23.5, 29.4, 17.6 and 17.6% of locations for U series during winter, spring, summer and autumn, respectively. On the other, a positive trend in winter, spring, summer and autumn U was significant in about 17.6, 23.5, 29.4 and 29.4% of sites, respectively. The Pettitt test revealed the presence of a significant change-point (p ≤ 0.05) in around 94.1, 76.5, 52.9 and 47.0% of stations in annual Tmean, U, RH and Rs series, respectively (Table 1). A significant change-point was also detected at approximately 41.2, 88.2, 88.2 and 82.3% of sites for Tmean, 64.7, 82.3, 58.8 and 70.6% of all locations for U, 17.6, 52.9, 52.9 and 29.4% of stations for RH and 17.6, 47.0, 52.9 and 11.7% of all sites for Rs during winter, spring, summer and autumn, respectively. Except in wintertime, a significant change-point in annual and seasonal Tmean series was, hence, found in more stations relative to the other variables. Furthermore, the change-point was chiefly observed during the 1990s in annual Tmean, RS and RH series and the 1970s for annual U (Table 1).

3.3 Quantitative ET0 Sensitivity Analysis The positive SCs for Tmean, Rs and U imply the positive effects of these variables on ET0, whereas negative SC for RH indicates the inverse correlation of ET0 with RH changes (Fig. 3). On a full year basis, Tmean was the most influential variable on annual ET0 in two stations located in the southwest of Iran i.e. Ahwaz and Abadan. In addition, in Boshehr, Bandar-e-

ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean

Abadan

Yazd

Kerman

Shahroud

Esfahan

Bandar-e-lengeh

Ahwaz

variable

Station

2001 1993 1978 1976 1993 1993 1973 1993 1993 1978 1982 1993 1983 1989 1975 1996 1993

Winter

1993 1998 1993 1973 1998 1997 1999 1981 1994 1998 1982 1997 1983 1997 1976 1986 1996 1997

Spring

1995 1987 1980 1989 1994 1995 1989 1977 1982 1988 1992 1982 1985 1974 1993 1975 1982 1996 1993

Summer

Table 1 The significant change-point (year) detected by Pettitt test

1995 1985 1996 1989 1977 1982 1978 1983 1984 1983 1977 1986 1977 1987 1981 1995 1994

Autumn 1988 1993 1977 1980 1993 1993 1973 1997 1997 1992 1978 1986 1978 1993 1995 1993 1983 1984 1996 1996 1975 1985 1996 1993

Annual

Iranshahr

Birjand

Zabol

Tehran

Bam

Semnan

Bandar-e-abbas

Station ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH ET0 Tmean

variable 1998 1993 2001 1993 2002 1993 1993 1993 1983 1999 1995 -

Winter 1998 1997 1984 1998 1999 2001 1996 1986 1997 1986 1997 1996 1997 1997 1999 1981 1999 1997 1994 1998 1995 1997

Spring 1997 1995 1999 1988 1999 1988 1986 1987 1986 1988 1994 1980 1999 1996 1977 1992 1997 1995 2000

Summer 1997 2000 2000 1984 1986 1984 1984 1987 1997 1996 1977 2000 1994 1994 1981 1994 1996

Autumn

1997 1997 1998 1999 1994 1999 1993 1986 1997 1986 1986 2000 1994 1997 1997 1977 1999 1997 1994 1995 1997

Annual

Quantitative Trend, Sensitivity and Contribution Analyses of Reference...

Zahedan

Boshehr

Station

Table 1 (continued)

Winter

1977 1984 1985

variable

U Rs RH ET0 Tmean U Rs RH ET0 Tmean U Rs RH

1980 1993 1996 1993 1977 1996 1997 1981 1994 1997

Spring

1972 1985 1998 1978 1977 1997 1993 1989 1984

Summer 1995 1985 1977 1997 1996 1984

Autumn 1973 1993 1993 1995 1977 1997 1997 1984 1984

Annual

Kashan

Station U Rs RH ET0 Tmean U Rs RH

variable 1995 1996 1977 1976 -

Winter 1983 1997 1998 1976 1976 -

Spring 1984 1995 1990 2004 1981 1974 1984

Summer 1994 2000 1996 1983 -

Autumn

1995 1995 1997 1976 1976 1974 -

Annual

M. Nouri et al.

Quantitative Trend, Sensitivity and Contribution Analyses of Reference... 1

(a) Tmean

0.8 0.6 0.4 0.2 0

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

0.6

(b) U

0.5 0.4 0.3 0.2 0.1 0

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 -0.7 -0.8 -0.9

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

(c) RH wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi us an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

0.7

(d) Rs

0.6 0.5 0.4 0.3 0.2 0.1 0

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

Fig. 3 The sensitivity coefficients (SCs) of ET0 to Tmean (a), U (b), RH (c) and Rs (d) in winter (black column), spring (red column), summer (yellow column) and autumn (green column)

lengeh and Bandar-e-abbas which are all coastal sites, RH had the greatest influence upon annual ET0. Further, U and Rs were found to be the most sensitive variables on ET0 in 35.3% (Tehran, Bam, Birjand, Yazd, Zahedan and Zabol) and 35.3% (Esfahan, Iranshahr, Kashan, Kerman, Shahroud and Semnan) of investigated stations, respectively (Fig. 3). As a result, annual ET0 exhibited a greater sensitivity to Rs and U relative to the other key meteorological variables in the most sites. However, ET0 showed a higher sensitivity to RH in arid southern coastal locations. Averaged across all sites, the greatest and lowest SC of ET0 to Tmean (SCTmean) was calculated in summer (0.495) and winter (0.178), respectively. Thus, the warmer the air temperature is, the more the sensitivity of ET0 to Tmean is. The averaged seasonal SC of ET0 to U (SCU) was found to be 0.269, 0.299, 0.334 and 0.309 during winter, spring, summer and autumn periods, respectively. Therefore, one may conclude that the averaged SCU does not vary as much as SCTmean over different seasons (Fig. 3). It is noteworthy that a small

M. Nouri et al.

perturbation in U results in a large ET0 change in arid regions (Allen et al. 1998; Irmak et al. 2006). Hence, the findings of the current study and some others e.g. Huo et al. (2013) suggest that the sensitivity of ET0 to U is considerable in water-limited arid ecosystems. The averaged SCRs was 0.257, 0.442, 0.446 and 0.328 in winter, spring, summer and autumn, respectively. Like Tmean, Rs is, hence, more influential on ET0 in warmer months of year (Fig. 3). The greatest and lowest absolute SCRH, averaged across all stations, were obtained in winter (−0.465) and summer (−0.135), respectively. Thus, as the air temperature rises, the sensitivity of ET0 to Rs and T increases and to RH decreases. Irmak et al. (2006) also concluded that the ET0 sensitivity to the saturation vapor pressure deficit (VPD) is smaller as the temperature increases. This can be explained by the fact that temperature increment increases the slope of saturation vapor pressure versus air temperature (Δ) and consequently decreases the magnitude of the term of [1/Δ + γ] (in Eq. 1) (Allen et al. 1998). As a result, the slighter decline in ET0 is yielded as RH rises under warmer climatic condition (Irmak et al. 2006; Monteith and Unsworth 2013). Increasing temperature under climate change would therefore likely enhance the ET0 sensitivity to Rs and Tmean and reduce the RH influences on ET0 in the twenty-first century. Averaged over all sites, the highest absolute SC was calculated for RH (−0.465), Rs (0.442), Tmean (0.495), Rs (0.328) and Rs (0.368) in winter, spring, summer, autumn and the entire year, respectively. Consequently, the most sensitive variable on annual, spring and autumn ET0 is Rs. Sharifi and Dinpashoh (2014) also concluded that the greatest sensitivity of ET0 was for vapor pressure, Rs, Tmean, Tmean and Tmean during winter, spring, summer, autumn and the entire year, respectively, over 8 sites in Iran. The differences between the results of Sharifi and Dinpashoh (2014) and ours may be attributed to different sensitivity analysis procedures, sites and the period of times considered in the studies.

3.4 Quantitative ET0 Contribution Analysis The average difference values between annual and seasonal ET0 calculated with original meteorological data (ET0-org) and recalculated with detrended U (ET0-det.U), Tmean (ET0-det.Tmean), RH (ET0-det.RH), Rs (ET0-det.Rs) and all climatic data series (ET0-det.all) in 1966–2012 are given in Fig. 4. The positive (negative) values suggest the positive (negative) effects of the trend of meteorological variables on the ET0 trend. There was an over 100 mm y−1 difference (averaged over 1966–2012) between ET0-org and ET0-det.U in Bam, Iranshahr, Kashan, Semnan and Zabol stations (Fig. 5b). The differences between annual ET0-org and ET0-det.Tmean, ET0-det.RH and ET0-det.Rs were lower than 50 mm y−1 at these sites (Fig. 4a, c and d). As a result, one attributes the significant upward trend in annual ET0 (Fig. 2a) at Zabol, Semnan, Bam, Kashan and Iranshahr greatly to the increased U over 1966–2012. Around 51.78, 35.92, 37.01, 33.44 and 50.18% of difference between ET0-org and ET0-det.all occurred during summertime in Semnan, Bam, Zabol, Iranshahr and Kashan, respectively. It seems that summer ET0 rise made a more important contribution to increment in annual ET0 owing to greater increase of summer wind speed in Semnan, Bam, Zabol, Iranshahr and Kashan. Given an over 0.90 correlation coefficient between annual ET0 and U in Semnan, Bam, Zabol, Iranshahr and Kashan (Fig. 5), the ET0 changes are highly correlated with the changes of U in these sites. An over 40 mm y−1 difference calculated between annual ET0-org and ET0-det.Tmean (ET0-org-ET0-det.Tmean) as well as ET0-det.RH (ET0-org-ET0-det.RH) indicated that the significant increase in Tmean and decline in RH contributed much more than other variables changes to the

Quantitative Trend, Sensitivity and Contribution Analyses of Reference... 70

(a) Tmean

60 50 40 30 20 10 0 -10

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

500

(b) U

400 300 200 100 0 -100 -200

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

50

(c) RH

40 30 20 10 0 -10 -20

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

Fig. 4 The average differences between ET0 calculated with original data and ET0 recalculated with detrended Tmean (a), U (b), RH (c), Rs (d) and all (e) data over winter (black column), spring (red column), summer (yellow column), autumn (green column) and entire year (blue column)

significant positive trend of annual ET0 at Zahedan site (Fig. 4). Despite a positive difference, averaged over 1966–2012, between annual ET0-org and ET0-det.Tmean, a less than −50 mm y−1 average difference between annual ET0-org and ET0-det.U led to an insignificant downward trend in annual ET0 at Esfahan, Birjand, Boshehr, Shahroud and Kerman locations (Figs. 2a, 4a and b). In these sites, the negative trend of U appears to be the prime cause of an insignificant decreasing trend in annual ET0. The correlation coefficient between annual ET0 and U was larger than that for other variables (Fig. 5) in Esfahan, Birjand, Boshehr, Shahroud and Kerman. Since, approximately 55.87, 43.20, 42.67 and 43.08% of average decrease in annual U at Esfahan, Birjand, Boshehr and Kerman, respectively, occurred in summertime, declined summer U seems to play more significant role in decreasing trend of annual ET0 in these sites. Changes in Tmean was found to be the most contributing factor affecting the ET0 trend in Bandar-e-lengeh, Abadan, Ahwaz and Yazd (Fig. 4a). The positive influence of increased Tmean (global warming) on annual trend of ET0 was, however, partially nullified by decreased

M. Nouri et al. 25

(d) Rs

20 15 10 5 0 -5 -10

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

500

(e) All

400 300 200 100 0 -100

wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an sp au wi su an Abadan

Ahwaz Bandar-e- Esfahan Shahroud Kerman lengeh

Yazd

Boshehr Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand Iranshahr Kashan

Fig. 4 (continued)

U in these sites. Meanwhile, increase in Rs and decline in RH caused an insignificant upward trend in annual ET0 (Figs. 2a, 4c and d) at Bandar-e-lengeh, Abadan, Ahwaz and Yazd. In addition, increased Tmean and decreased RH was found to be the most contributing factor to the insignificant increasing ET0 trend in Tehran and Bandar-e-abbas sites, respectively (Fig. 4a and c). The difference determined between ET0-org and ET0-det.all was larger in summertime for Ahwaz, Bandar-e-abbas and Bandar-e-lengeh, in autumn for Abadan and Yazd stations and during spring in Tehran site (Fig. 4e). The highest correlation coefficient was obtained between annual U and ET0 series in Yazd, Ahwaz, Abadan, Tehran and Bandar-e-abbas (Fig. 5). Averaged over all sites, the correlation coefficient between annual ET0 and Tmean, U, RH, and Rs was, respectively, 0.31, 0.84, −0.44 and 0.32, indicating that ET0 changes are highly correlated with U trend in the studied sites (Fig. 5). 1 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs Tmean U RH Rs

-0.8

Abadan

Ahwaz

Bandar-elengeh

Esfahan

Shahroud

Kerman

Yazd

Boshehr

Zahedan Bandar-e- Semnan abbas

Bam

Tehran

Zabol

Birjand

Iranshahr

Kashan

Fig. 5 The correlation coefficient between annual ET0 and Tmean (black column), U (red column), Rs (yellow column) and RH (green column) series

Quantitative Trend, Sensitivity and Contribution Analyses of Reference...

Dinpashoh et al. (2011) concluded that RHmax (maximum RH) in Zahedan and U in Ahwaz, Abadan, Tehran, Zabol, Birjand, Esfahan and Kerman had the main contribution to the ET0 changes in 1965–2005. Unlike the results of Dinpashoh et al. (2011), Tmean made the greater contribution to the ET0 dynamics for Tehran, Ahwaz and Abadan sites which may be due to different contribution analysis methods and the time periods considered in the studies. It should be noted that U was not the main contributing factor to the ET0 changes in some regions i.e. Tehran, Abadan, Ahwaz and Bandar-e-abbas based on the detrending method, although there was a higher correlation coefficient between ET0 and U in these stations (Fig. 5). Huo et al. (2013) in arid northwestern areas of China, Zhang et al. (2007) and Shenbin et al. (2006) across the Tibetan Plateau, China (a water-limited semi-arid ecosystem) and Shan et al. (2015) in some water-limited desertification-prone regions of China reported U as the most important meteorological variable contributing to the ET0 changes. In addition to Rs, U has been also found by Xu et al. (2006) in Changjiang basin, China (an energy-limited humid region) as the leading cause of the ET0 changes. Further, the decreasing trend in pan evaporation (ETpan) has been attributed to the reduced U by Zhang et al. (2007) in the Tibetan Plateau, China, Roderick et al. (2007) in Australia and You et al. (2013) in the southwestern China. Overall, although temperature rise due to high CO2 emission in the last half of century is expected to increase ET0, decreasing trend in U caused a reduction in ET0 in some sites i.e. Esfahan, Kerman, Birjand, Shahroud and Boshehr. Further, despite positive impacts of global warming upon annual ET0 in Zabol, Bam, Iranshahr, Kashan and Semnan, the U increment seems to be the primary cause of the significant increasing trend of annual ET0 in these locations. The positive impacts of temperature rise on ET0 were nearly counterbalanced by decreased U in Abadan, Ahwaz, Bandar-e-lengeh and Yazd. Change in wind speed was, therefore, made the major contribution to annual ET0 trend over the period of 1966–2012 in most arid areas of Iran. In the surveyed sites particularly those located in the southeastern Iran where the U variation remarkably contributes to the ET0 changes, using the ET0 temperature-based (e.g. Hargreaves–Samani model) and radiation-based (e.g. Priestley–Taylor equation) models in which U is lumped into a constant coefficient or not taken into account would highly likely lead to an erroneous estimate of ET0. Consequently, error in estimating ET0 would adversely influence irrigation projects. Chen et al. (2005) recommended not using Thornthwaite method (a temperature-based model) for estimating ET0 over China as ET0 is highly conditioned by Rs and U there. Increased water demand due to the ET0 rise at most studied sites wherein surface and groundwater resources are already seriously depleted can severely jeopardize water supplies and food security. Hence, proper water resource management particularly in agriculture sector is urgently required to mitigate the negative impacts of increased ET0 and climate change in Iran’s drought-prone arid regions. In addition, decreasing U (for instance by windbreaker) seems to substantially reduce ET0 and consequently irrigation water requirement in our study area. Due to lack of climatic data, spatial analyses of the ET0 sensitivity to the meteorological variables and the contribution of the climatic factors to the ET0 trend could not be undertaken in our studied area. Therefore, some studied needed to be carried out to spatially analyze the ET0 sensitivity and contribution using statistical interpolation or remote-sensing techniques. Further, it is worth pointing out that the results of current study must be verified for other climates and also other arid regions where climatic changes might differently affect ET0.

M. Nouri et al.

4 Conclusions The trend of reference evapotranspiration (ET0) and ET0 sensitivity to mean temperature (Tmean), solar radiation (Rs), relative humidity (RH) and wind speed (U) were quantitatively studied in 17 arid sites in Iran over 1966–2012. The contribution of changes in Tmean, Rs, RH and U to the ET0 trend was also quantitatively investigated. The results indicated that there was an increasing trend in annual ET0 for about 70.0% of station. However, the increasing trend of annual ET0 was significant (p ≤ 0.05) in only 35.3% of all surveyed sites. Approximately 47.0, 35.3, 35.3 and 29.4% of stations exhibited significant positive trend in ET0 during winter, spring, summer and autumn, respectively. In addition, a significant change-point was detected in winter, spring, summer, autumn and annual ET0 series at 64.7, 52.9, 64.7, 64.7 and 82.3% of stations, respectively. Rs and U are the most sensitive variables on ET0 in the majority of studied sites. In addition, RH was the most sensitive variable on ET0 in the coastal southern sites. Decreasing trend of U caused a downward trend in annual ET0 for 29.4% of the sites and partly offset the positive effects of temperature increment on annual ET0 in around 23.5% of studied sites. Moreover, increase of U contributed much more than other variables to the upward ET0 trend in about 29.4% of locations. Therefore, U was the most important climatic factor contributing to the ET0 trend over 1966–2012 in most arid sites. Overall, the results showed that the climatic changes had positive impacts on the ET0 trend in the most studied sites causing an increase in irrigation water requirement in arid regions of Iran. Increase in water demand as a result of ET0 rise puts a tremendous pressure on water supplies in Iran. Therefore, adopting appropriate agricultural water management is highly needed to cope with climate change in the arid environments of Iran. Furthermore, U as the most important climatic factor contributing to the ET0 trend must be taken into account for assessing ET0, irrigation water demand and hydrological cycle components under climate change.

References Al-Faraj FAM, Tigkas D, Scholz M (2016) Irrigation efficiency improvement for sustainable agriculture in changing climate: a transboundary watershed between Iraq and Iran. Environmental Process 3:603–616 Alizadeh A, Keshavarz A (2005) Status of agricultural water use in Iran. In: Water conservation, reuse, and recycling: proceedings of an Iranian-American workshop. National Academies Press, Washington DC, pp 94–105 Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration-guidelines for computing crop water requirements-FAO irrigation and drainage paper 56. FAO, Rome 300:6541 Busuioc A, von Storch H (1996) Changes in the winter precipitation in Romania and its relation to the large-scale circulation. Tellus A 48:538–552 Chen D, Gao G, Xu C-Y, Guo J, Ren G (2005) Comparison of the Thornthwaite method and pan data with the standard Penman-Monteith estimates of reference evapotranspiration in China. Clim Res 28:123–132 Debnath S, Adamala S, Raghuwanshi NS (2015) Sensitivity analysis of FAO-56 Penman-Monteith method for different agro-ecological regions of India. Environ Process 2:689–704. doi:10.1007/s40710-015-0107-1 Dinpashoh Y, Jhajharia D, Fakheri-Fard A, Singh VP, Kahya E (2011) Trends in reference crop evapotranspiration over Iran. J Hydrol 399:422–433 Estévez J, Gavilán P, Berengena J (2009) Sensitivity analysis of a Penman–Monteith type equation to estimate reference evapotranspiration in southern Spain. Hydrol Process 23:3342–3353. doi:10.1002/hyp.7439 Gao Z, He J, Dong K, Bian X, Li X (2016) Sensitivity study of reference crop evapotranspiration during growing season in the West Liao River basin, China. Theor Appl Climatol 124:865–881 Godfray HCJ, Beddington JR, Crute IR, Haddad L, Lawrence D, Muir JF, Pretty J, Robinson S, Thomas SM, Toulmin C (2010) Food security: the challenge of feeding 9 billion people. Science 327:812–818

Quantitative Trend, Sensitivity and Contribution Analyses of Reference... Gong L, Xu C-y, Chen D, Halldin S, Chen YD (2006) Sensitivity of the Penman–Monteith reference evapotranspiration to key climatic variables in the Changjiang (Yangtze River) basin. J Hydrol 329:620–629 Homaee M, Dirksen C, Feddes R (2002) Simulation of root water uptake: I. Non-uniform transient salinity using different macroscopic reduction functions. Agric Water Manag 57:89–109 Homaee M, Schmidhalter U (2008) Water integration by plants root under non-uniform soil salinity. Irrig Sci 27:83–95 Huo Z, Dai X, Feng S, Kang S, Huang G (2013) Effect of climate change on reference evapotranspiration and aridity index in arid region of China. J Hydrol 492:24–34 Hupet F, Vanclooster M (2001) Effect of the sampling frequency of meteorological variables on the estimation of the reference evapotranspiration. J Hydrol 243:192–204 IPCC (2013) Climate change 2013: the physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge and New York Irmak S, Payero JO, Martin DL, Irmak A, Howell TA (2006) Sensitivity analyses and sensitivity coefficients of standardized daily ASCE-Penman-Monteith equation. J Irrig Drain Eng 132:564–578 Kaniewski D, Van Campo E, Weiss H (2012) Drought is a recurring challenge in the Middle East. Proc Natl Acad Sci 109:3862–3867 Kite G, Droogers P (2000) Comparing evapotranspiration estimates from satellites, hydrological models and field data. J Hydrol 229:3–18 Liu Q, Yang Z, Cui B, Sun T (2010) The temporal trends of reference evapotranspiration and its sensitivity to key meteorological variables in the Yellow River basin, China. Hydrol Process 24:2171–2181 Maestro T, Nicolosi V, Cancelliere A, Bielza M (2014) Impacts of climate change, hydrological drought mitigation measures and irrigation demand on water supply system performance. Eur Water 45(46):25–33 McCuen RH (1974) A sensitivity and error analysis cf procedures used for estimating evaporation. JAWRA J Am Water Resour Assoc 10:486–497. doi:10.1111/j.1752-1688.1974.tb00590.x Modarres R, Sarhadi A (2009) Rainfall trends analysis of Iran in the last half of the twentieth century. J Geophys Res Atmos 114:D03101. doi:10.1029/2008JD010707 Monteith J, Unsworth M (2013) Principles of environmental physics: plants, animals, and the atmosphere (Fourth Edition). Academic Press, Boston Mosaedi A, Sough MG, Sadeghi S-H, Mooshakhian Y, Bannayan M (2016) Sensitivity analysis of monthly reference crop evapotranspiration trends in Iran: a qualitative approach. Theor Appl Climatol:1-17. doi:10.1007/s00704-016-1740-y Nouri M, Homaee M, Bannayan M, Hoogenboom G (2016) Towards modeling soil texture-specific sensitivity of wheat yield and water balance to climatic changes. Agric Water Manag 177:248–263. doi:10.1016/j.agwat.2016.07.025 Pettitt A (1979) A non-parametric approach to the change-point problem. Appl Stat 28:126–135 Roderick ML, Rotstayn LD, Farquhar GD, Hobbins MT (2007) On the attribution of changing pan evaporation. Geophys Res Lett 34:L17403. doi:10.1029/2007GL031166 Saxton KE (1975) Sensitivity analyses of the combination evapotranspiration equation. Agric Meteorol 15: 343–353 Seneviratne SI, Corti T, Davin EL, Hirschi M, Jaeger EB, Lehner I, Orlowsky B, Teuling AJ (2010) Investigating soil moisture–climate interactions in a changing climate: a review. Earth-Sci Rev 99:125–161 Shan N, Shi Z, Yang X, Zhang X, Guo H, Zhang B, Zhang Z (2015) Trends in potential evapotranspiration from 1960 to 2013 for a desertification-prone region of China. Int J Climatol 36:3434–3445 Sharifi A, Dinpashoh Y (2014) Sensitivity analysis of the Penman-Monteith reference crop evapotranspiration to climatic variables in Iran. Water Resour Manag 28:5465–5476. doi:10.1007/s11269-014-0813-x Shenbin C, Yunfeng L, Thomas A (2006) Climatic change on the Tibetan Plateau: potential evapotranspiration trends from 1961–2000. Clim Chang 76:291–319. doi:10.1007/s10584-006-9080-z Some'e BS, Ezani A, Tabari H (2012) Spatiotemporal trends and change point of precipitation in Iran. Atmos Res 113:1–12 Song Y, Su XL (2015) Sensitivity of reference crop evapotranspiration in Shaanxi Province. In: Zhou P, Patty A (eds) Applied mechanics and materials, vol 700. Trans Tech Publ, Zurich, pp 374–385 Tabari H, Talaee PH (2011) Analysis of trends in temperature data in arid and semi-arid regions of Iran. Glob Planet Chang 79:1–10 UNEP (1992) World atlas of desertification. United Nations Environment Programme, London Xie H, Zhu X (2013) Reference evapotranspiration trends and their sensitivity to climatic change on the Tibetan Plateau (1970–2009). Hydrol Process 27:3685–3693 Xie H, Zhu X, Yuan DY (2015) Pan evaporation modelling and changing attribution analysis on the Tibetan Plateau (1970–2012). Hydrol Process 29:2164–2177 Xu C-y, Gong L, Jiang T, Chen D, Singh VP (2006) Analysis of spatial distribution and temporal trend of reference evapotranspiration and pan evaporation in Changjiang (Yangtze River) catchment. J Hydrol 327: 81–93. doi:10.1016/j.jhydrol.2005.11.029

M. Nouri et al. You G, Zhang Y, Liu Y, Song Q, Lu Z, Tan Z, Wu C, Xie Y (2013) On the attribution of changing pan evaporation in a nature reserve in SW China. Hydrol Process 27:2676–2682. doi:10.1002/hyp.9394 Yue S, Pilon P, Phinney B, Cavadias G (2002) The influence of autocorrelation on the ability to detect trend in hydrological series. Hydrol Process 16:1807–1829 Zhang Y, Liu C, Tang Y, Yang Y (2007) Trends in pan evaporation and reference and actual evapotranspiration across the Tibetan Plateau. J Geophys Res Atmos 112:D12110. doi:10.1029/2006JD008161