An innovative remote sensing based reference evapotranspiration method to support irrigation water management under semi-arid conditions

Agricultural Water Management 131 (2014) 135–145

Contents lists available at ScienceDirect

Agricultural Water Management journal homepage: www.elsevier.com/locate/agwat

An innovative remote sensing based reference evapotranspiration method to support irrigation water management under semi-arid conditions M. Cruz-Blanco, I.J. Lorite ∗ , C. Santos IFAPA–Centro “Alameda del Obispo”, Regional Government of Andalusia, Córdoba, Spain

a r t i c l e

i n f o

Article history: Received 4 March 2013 Accepted 19 September 2013 Available online 19 October 2013 Keywords: Reference evapotranspiration Remote sensing Meteosat Second Generation Penman–Monteith FAO56 Irrigation water requirements AquaCrop

a b s t r a c t Reference evapotranspiration (ETo ) is an essential component of irrigation water management due to it being a basic input for estimating crop water requirements. Multiple approaches have been identified for ETo assessment but most of them are based on daily meteorological data provided by weather station networks that provide an accurate meteorological characterization. A new alternative approach called MA + LSE based on the Makkink-Advection (MAK-Adv) equation in combination with remotely sensed solar radiation and a numerical weather forecast of near surface air temperature has provided good estimates of ETo under different weather conditions in a semi-arid region located in Southern Spain, without requiring local meteorological data. In order to evaluate the utility of the MA + LSE approach for irrigation water management, some wellknown methods for ETo assessment and the MA + LSE approach were considered for the development of irrigation schedules in ten irrigation schemes located in a semi-arid region in Southern Spain. The impact of the approach considered for ETo assessment on irrigation scheduling and on simulated yield for a maize crop was determined. Thus, MA + LSE and Hargreaves methods generated similar irrigation schedules and estimated yield to those determined by using ETo from the Penman–Monteith (PM-FAO56) approach. Thus, average seasonal irrigation volume estimated by MA + LSE was underestimated by around 2.6%, causing a yield reduction of 2.2% compared with the irrigation scheduling based on PM-FAO56. These results confirm the applicability of the MA + LSE approach, especially in areas where meteorological data are missing or inaccurate, obtaining a similar performance for irrigation water management to that of other approaches with high data requirements such as PM-FAO56. © 2013 Elsevier B.V. All rights reserved.

1. Introduction In arid and semi-arid Mediterranean environments a very significant percentage of the available water resources is consumed by irrigated agriculture. Thus, in Spain, about 83% of the water resources is devoted to irrigation, and the rest to other uses (CAP, 2011). However, this distribution of the available resources could vary significantly in the next decades due to an increase in the water requirements of other sectors such as the environment, and a foreseen reduction in rainfall from climate change effects in Southern Europe (Christensen and Christensen, 2007). In this context of maximum competitiveness for water resources, a correct

∗ Corresponding author at: IFAPA – Centro “Alameda del Obispo”, Regional Government of Andalusia, Avda. Menendez Pidal s/n, PO Box 3092, 14080 Córdoba, Spain. Tel.: +34 671 532698; fax: +34 957 016043. E-mail addresses: [email protected], [email protected] (I.J. Lorite). 0378-3774/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.agwat.2013.09.017

irrigation management is essential to ensure the sustainability of the Mediterranean irrigated areas. Reference evapotranspiration (ETo ) is one of the key inputs for determining proper irrigation water management. The concept of reference evapotranspiration was introduced to estimate crop water requirements combined with tabulated crop coefficients (Allen et al., 1998). The methodology, described in Allen et al. (1998), based on a version of the Penman–Monteith equation, is considered to be the standard procedure for ETo assessment and it obtains a satisfactory performance under advective conditions (Berengena and Gavilán, 2005), but with a high level of data requirements, complicating its application in areas with unavailable, missing or inaccurate meteorological data. For these reasons, ETo estimation has long been a critical issue that has concerned researchers, technicians, and qualified farmers, so alternative methodologies requiring lower cost observations have been developed in the past. The Penman–Monteith equation (Allen et al., 1998) has been widely used and has been demonstrated as producing the most

136

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

accurate estimations when compared with lysimeter measurements in Mediterranean environments (Berengena and Gavilán, 2005; López-Urrea et al., 2006). However, the need for gathering a relatively large number of variables (e.g., relative humidity of the air, solar radiation or wind speed) under reference conditions, which are generally difficult to measure with any accuracy, limits the widespread use of this equation (Pereira and Pruitt, 2004). It has often been substituted by approaches with lower input requirements such as Hargreaves, Makkink or Priestley–Taylor equations (Gavilán et al., 2006; De Bruin et al., 2010; Espadafor et al., 2011). However, the use of simpler approaches for ETo assessment with a lower level of data requirements could affect correct irrigation water management, obtaining irrigation schedules of a poorer quality with a negative impact on water savings and yield. The contribution of remote sensing techniques to the improvement of water management at a basin scale has increased significantly in the last few years. Remote sensing tools allow the obtainment of accurate information of surface and atmospheric conditions for vast areas. The most advanced approaches have been focused on the determination of crop evapotranspiration (ETc ) using energy balance models (Bastiaanssen et al., 1998; Allen et al., 2007a,b; Ghilain et al., 2011, 2012). These approaches have provided relevant information on crop water requirements, enabling the improvement of agricultural water management (Santos et al., 2008, 2010), the monitoring of drought (Sun et al., 2011), and the analysis of the spatial and temporal variability of the vegetation (Stisen et al., 2008a; Ghilain et al., 2012). Studies for estimating ETo have been less numerous, although, recently, De Bruin et al. (2010) and Cruz-Blanco et al. (2014) have demonstrated that geostationary satellite data can be used to determine accurate ETo values for vast areas. Methodologies considering remote sensing and forecast tools for ETo assessment have been used with success in different semiarid environments (De Bruin et al., 2010; Ghilain et al., 2011; Cristobal and Anderson, 2012; Cruz-Blanco et al., 2014), but the determination of the impact of these methodologies on irrigation water management, particularly under semi-arid conditions, is still required. Therefore, the main objective of the current study was to evaluate the impact of these new approaches for ETo assessment on the generation of irrigation schedules, a key component for a sound irrigation water management. For this task, maize, a relevant crop in the irrigated areas in Andalusia, Southern Spain, has been considered. The study covers the period 2007–2009 and includes ten irrigation schemes located throughout Southern Spain (Fig. 1) in order to consider their impact under different weather conditions.

2. Material and methods 2.1. EUMETSAT LSA SAF and ECMWF products The EUMETSAT Satellite Applications Facility for Land Surface Analysis (LSA SAF) is part of the SAF network, a set of specialized development and processing centers serving the European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) (Trigo et al., 2011). The main objective of LSA SAF is the development of remote sensing applications relevant to land surface processes and biosphere applications, such as the case of daily solar radiation at the surface. Within the LSA SAF, this component corresponds to the so-called daily down-welling surface shortwave radiation flux (DIDSSF), determined from the accumulation of 30-min observations provided by the Spinning Enhanced Visible and Infrared Imager (SEVIRI) radiometer, on board Meteosat Second Generation (MSG; Schmetz et al., 2002). DIDSSF refers to the daily accumulation of 30-min radiative energy flux in the wavelength interval 0.3 ␮m–4 ␮m (Geiger et al., 2008; Trigo et al.,

2011). EUMETSAT LSA SAF has an internal quality control for each datum provided (Trigo et al., 2011). In the case of DIDSSF, the quality indicator includes the number of missing slots within the corresponding 24 h period. These data are available in near-realtime (http://landsaf.meteo.pt) at the MSG pixel resolution (3 km at nadir; Trigo et al., 2011), covering an area encompassing Africa, most of Europe and part of South America. Air temperature at 2 m (T2m ) was obtained from the operational forecasts provided by the European Center for Medium-Range Weather Forecasts (ECMWF). The initial 3-hourly T2m forecasts at a resolution of about 25 km were linearly interpolated in time to hourly ones, and bi-linearly interpolated in space to the SEVIRI/MSG resolution. The T2m values underwent a further adjustment to correct differences between ECMWF model surface orography and the finer scale SEVIRI pixel altitude, using a constant slope rate of 0.0067 ◦ C m−1 (De Bruin et al., 2010). A detailed full-description of ECMWF data is available in Persson (2011).

2.2. Selected irrigation schemes and meteorological data Ten irrigation schemes were considered in this study to quantify the impact of different approaches for ETo assessment on irrigation scheduling. These irrigation schemes are located in the main maize areas in Andalusia, Southern Spain (Fig. 1), and showed significantly different climate conditions (average rainfall ranged between 437 and 641 mm year−1 and average temperature between 14.5 and 18.5 ◦ C). Daily meteorological data obtained from ten weather stations located close to the selected irrigation schemes were used in the study (Fig. 1). These stations are included in the Agroclimatic Information Network of Andalusia (RIA), which is currently composed of 100 automatic weather stations. This network was deployed to provide coverage to most of the irrigated areas with the aim of supplying ETo values and other meteorological data to improve irrigation water management (Gavilán et al., 2006). These stations are controlled by a CR10X datalogger (Campbell Scientific Inc., Logan, UT, USA) and are equipped with sensors to measure air temperature and relative humidity (HMP45C probe, Vaisala, Helsinki, Finland), solar radiation (pyranometer CM6B, Kipp&Zonen, Delft, Holland), wind speed and direction (wind monitor RM Young 05103, Traverse City, MI, USA) and rainfall (tipping bucket rain gauge ARG 100). Air temperature and relative humidity are measured at 1.5 m and wind speed at 2 m above soil surface. The values collected from each station are checked for validity according to Meek and Hatfield (1994) and Shafer et al. (2000). Daily average values are recorded for each meteorological variable and can be obtained from the Website at www.juntadeandalucia.es/agriculturaypesca/ifapa/ria. Except Cordoba station (COR CO) which is placed on a grass reference surface (Berengena and Gavilán, 2005), the rest of the weather stations are located on bare soil, not fulfilling the surface reference requirements determined by Allen (1996) and Allen et al. (1998). However, previous studies have demonstrated the data accuracy although measurements were carried out under non-optimal reference conditions (Cruz-Blanco et al., 2014). In the analyzed period (2006–2009) seasonal rainfall (measured from September to October) varied from 437 mm for Laguna Fuente Piedra irrigation scheme (YEG MA weather station) to 641 mm for Bajo Guadalete irrigation scheme (JER CA weather station). Range of ETo variation measured by the RIA was smaller, varying from 1270 mm year−1 for Vegas Bajas irrigation scheme (MAR JA weather station) to 1419 mm for Viar irrigation scheme (GUI SE weather station). This characterization depicts the great variability of the analyzed stations, allowing the carrying out of a complete evaluation of different approaches for ETo assessment for irrigation

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

137

Fig. 1. Maize areas (in green), irrigation schemes (red circles) and location of the weather stations (black points) included in this study. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

scheduling under different weather conditions, providing a higher applicability to this study. In order to simulate the frequent situations when meteorological data were not available from a nearby weather station and technicians/farmers were required to use meteorological data from a distant weather station, for each irrigation scheme an alternative weather station located further away from the one initially selected was chosen (Fig. 1), developing a new approach for ETo assessment (MP-PM approach). 2.3. Reference ET equations 2.3.1. Penman–Monteith equation (PM-FAO56 and MP-PM) The standardized Penman–Monteith (PM-FAO56) equation (Allen et al., 1998) is an attempt to simplify and clarify the application of the Penman–Monteith equation (Monteith, 1965). For grass reference surface and for daily time steps, this equation is expressed as:



ETO =

2.3.2. Makkink equation (MAK) De Bruin (1987) determined that, under non-water stress conditions, ETo could be accurately estimated using a simplified Makkink formula (Makkink, 1957):



0.408  (Rn − G) +  900⁄T + 273 U2 (es − ea )  +  (1 + 0.34U2 )

Since measurements of ETo over a grass reference crop are often not available (De Bruin et al., 2010), several authors consider the estimations provided by the Penman–Monteith equation using data from weather stations under a non-optimal grass reference as the benchmark for comparison with other equations and methodologies (Gavilán et al., 2006; Jabloun and Sahli, 2008; Landeras et al., 2008; Dai et al., 2009; Er-Raki et al., 2010; Espadafor et al., 2011; Cammalleri and Ciraolo, 2013). In this study, the PM-FAO56 approach was considered as being the reference for comparison with other approaches. Additionally, the PM-FAO56 equation was contemplated for the MP-PM approach, with the weather data obtained from alternative weather stations indicated in Fig. 1.

(1)

where ETo is the reference evapotranspiration (mm day−1 ); Rn is the net radiation at the crop surface (MJ m−2 day−1 ) using the FAO56 approach (Allen et al., 1998); G is the soil heat flux (MJ m2 day−1 ); T is the mean daily air temperature at 2 m height (◦ C); U2 is the 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);  is the slope of saturated vapor–pressure curve (kPa ◦ C−1 ); and  is the psychometric constant (kPa ◦ C−1 ).

ETo = cMAK

1  Rs  +

(2)

where ETo is the reference evapotranspiration (mm day−1 ), Rs is the measured solar radiation in MJ m−2 day−1 ,  is the latent heat of vaporization in MJ kg−1 and cMAK a parameter depending on climate conditions. De Bruin (1987) and De Bruin et al. (2010) found that a cMAK value corresponding to advection-free conditions should equal 0.65, this being the base for the determination of crop factors dependent on the season for various agricultural crops in The Netherlands (Feddes, 1987), used operationally by the Royal Netherland Meteorological Institute (KNMI), and also in this study.

138

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

2.3.3. Makkink-Advection equation using LSA SAF and ECMWF data (MA + LSE) In moderate climate regions equation (2) can be approximated by a linear function (De Bruin et al., 2012) namely: ETo =

1 [0.37 + 0.009 (Ta − 12)] Rs 

(3)

where Ta is the air temperature (◦ C) and Rs is the solar radiation (MJ m−2 day−1 ). A revised version of equation (3) to account for advective effects observed in Southern Spain was developed using lysimeter data (De Bruin et al., 2012; Cruz-Blanco et al., 2014). This new Makkink equation, denoted as MAK-Adv equation, reads: ETo =

1 [0.38 + 0.015 (Ta − 12)] Rs 

(4)

In this study, Rs and Ta were determined from EUMETSAT LSA SAF, a remote sensing tool, and ECMWF products, respectively, described in Section 2.1. This approach was validated under semiarid conditions using independent observations from a weighting lysimeter under reference conditions and results revealed very good statistics (RMSE equal to 0.5 mm day−1 ) and the regressions also showed a very good fit (slope of 0.97 and R2 equal to 0.96; Cruz-Blanco et al., 2014). These results confirm that the MAK-Adv equation provides good ETo estimates under semi-arid environments. 2.3.4. Priestley–Taylor equation (PT) Priestley and Taylor (1972) proposed a method for estimating ETo for advection-free conditions from the energy available to evaporate water from the surface by using an empirical factor ˛: ETo =

˛ 



(Rn − G) +



(5)

where ETo is the reference evapotranspiration (mm day−1 ); ˛ is a dimensionless coefficient with a proposed empirical value of 1.26 (Priestley and Taylor, 1972); Rn and G are the net radiation and the soil heat flux density, respectively, in MJ m−2 day−1 ;  is the latent heat of vaporization in MJ kg−1 ;  is the slope of the saturated vapor pressure curve and  is the psychrometric constant (kPa ◦ C−1 ). This equation has been widely used due to its low input requirements and generally good quality for ETo assessment (Utset et al., 2004; Sentelhas et al., 2010; Espadafor et al., 2011). 2.3.5. Hargreaves (HAR) The Hargreaves equation (Hargreaves and Samani, 1985) can be written as: ETo = ˛ (Tmean + 17.8)



Tmax − Tmin Ra

(6)

where ETo is the computed reference evapotranspiration (mm day−1 ), ˛ = 0.0023 is the original empirical coefficient proposed by Hargreaves and Samani (1985), Tmean , Tmax and Tmin are the daily mean, maximum and minimum air temperatures (◦ C) and Ra is the water-equivalent of the extraterrestrial radiation (mm day−1 ) computed according to Allen et al. (1998). The Hargreaves equation has been widely used with satisfactory results for ETo assessment (Gavilán et al., 2006; Sentelhas et al., 2010; Raziei and Pereira, 2013). 2.4. Irrigation scheduling and yield assessment using AquaCrop model The Food and Agriculture Organization (FAO) has developed AquaCrop (Steduto et al., 2012), a crop water productivity model focused on simulating water-limited attainable yield. AquaCrop has been extensively tested for different crops around the world under diverse environments (e.g., Hsiao et al., 2009 for maize; García-Vila

and Fereres, 2012 for cotton), and has been used to design different deficit irrigation strategies (Geerts et al., 2010), or to develop an economic model for decision support system at the farm scale (García-Vila and Fereres, 2012). Drip irrigation scheduling for maize was calculated daily by the AquaCrop model allowing a small average leaf expansion stress (up to 15%) but trying to avoid any average stomatal stress. The limited stress was permitted in order to generate viable irrigation schedules, without exceeding the average annual allocation for the majority of the irrigation schemes located in Southern Spain (around 5000 m3 ha−1 ; García-Vila et al., 2008), and to increase irrigation water productivity (IWP) compared with non-stress irrigation scheduling (Zwart and Bastiaanssen, 2004; Lorite et al., 2007). The AquaCrop model also simulates attainable yields for herbaceous crops as a function of water consumption under different irrigation regimes (Steduto et al., 2012). AquaCrop directly links crop yields to water use and estimates biomass production from actual crop transpiration through a normalized water productivity parameter, which is the core of the AquaCrop growth engine (Steduto et al., 2012). AquaCrop estimated the yield for each irrigation schedule based on the different ETo approaches described in Section 2.3, under the ETo conditions determined with the PMFAO56 approach, in order to simulate the impact of the different irrigation scheduling strategies on yield. Finally, for each approach and location, IWP, defined as the ratio between increase in annual production due to irrigation and the annual volume of irrigation, was calculated. Rainfed yield, required to determine the yield increase caused by irrigation (García-Vila et al., 2008), was also estimated using the AquaCrop model. In order to facilitate the data preparation and output analysis, the AquaData tool (Lorite et al., 2013) was used in this study.

3. Results and discussion 3.1. Reference evapotranspiration assessment using remote sensing and predicting tools The application of equation (4) using data provided by EUMETSAT LSA SAF employing MSG satellite (Rs ) and ECMWF forecast tool (Ta ) integrated in the MA + LSE approach provided accurate daily ETo maps for the analyzed area (Fig. 2), depicting a clear spatial pattern with the highest ETo values during spring–summer located in the Guadalquivir Valley, and the lowest ETo values in mountainous and coastal areas. During fall–winter time the spatial pattern changes, with the highest ETo values being located in coastal areas and in the lowest section of the Guadalquivir Valley, and the lowest ones were detected in inland areas. These maps showed the great potential of this methodology, which provides accurate daily information for vast areas. Although the proposed methodology has been previously validated for semi-arid conditions in Southern Spain (Cruz-Blanco et al., 2014), a comparison of ETo results with reference approach values was carried out. In Figs. 3 and 4 a comparison of ETo values using PM-FAO56 with those values obtained with the MA + LSE approach is shown for the ten weather stations selected for the study (Fig. 1). Linear regression was very accurate for all the weather stations. The slope of the regression ranged between 0.92 (TRP JA station) and 1.05 (MAR JA station), with an average slope of 0.97, indicating a small underestimation in the ETo assessment using the MA + LSE approach. Similarly, the dispersion of the estimation was small: R2 ranged from 0.89 (VEJ CA weather station) to 0.96 (YEG MA weather station), with an average value of 0.94 (Figs. 3 and 4). Additionally, spatial patterns in the distribution of the differences between ETo values provided by MA + LSE and

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

139

Fig. 2. Daily ETo maps (in mm day−1 ) using MA + LSE approach for Southern Spain for (a) 7th July 2007 and (b) 6th December 2007.

PM-FAO56 approaches were detected. Those weather stations located on hilly areas and outside the Guadalquivir Valley (such as IZN GR, TRP JA or VEJ CA) showed the highest ETo differences (around 9%) and the worst correlation figures when the PM-FAO56 and MA + LSE approaches were considered (Figs. 3 and 4). These results agree with Cristobal and Anderson (2012) who determined errors in the solar radiation assessment by EUMETSAT LSA SAF in hilly areas caused by the effect of shading on the measurements. Figs. 3 and 4 also showed some anomalous ETo values provided by the MA + LSE approach (e.g. numerous underestimated values in most of the locations or occasional overestimations in LEB SE station). One of the main sources of error was that caused by the limited data requirement of the MAK-Adv equation (temperature and solar radiation). Thus, some previous studies (Landeras et al., 2008; Shiri et al., 2012) showed a better performance with approaches including wind speed and/or relative

humidity. Another factor causing errors came from the use of the PM-FAO56 approach under non-reference conditions (Cruz-Blanco et al., 2014), which triggered overestimations in the ETo assessment (Allen, 1996; Temesgen et al., 1999; Allen et al., 2002). Corrections for these sources of errors could be considered although in most of the locations the detected differences in ETo assessment lie within the resolution of the equipment and then the results obtained with the MA + LSE approach being considered as highly satisfactory. 3.2. Comparison of ETo values with different approaches Significant differences in ETo were detected depending on the approach considered (Table 1). The PM-FAO56 approach provided an average value of 1361 mm (from September to August) with a small variation between crop seasons. The MA + LSE and HAR approaches generated very similar values, with a slight average

140

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

10

COR_CO

SAN_CO 9

8

8

-1

ETo MA+LSE (mm day )

9 -1

ETo MA+LSE (mm day )

10

7 6 5 4 3 y = 0.9755x 2 R = 0.9468

2 1

7 6 5 4 3 y = 0.9855x 2 R = 0.9331

2 1 0

0 0

1

2

3

4

5

6

7

8

9

0

10

1

ET o PM-FAO56 (mm day

3

4

5

6

7

8

9

10

-1

ET o PM-FAO56 (mm day )

)

10

10

LEB_SE

JER_CA 9

8

8

-1

ETo MA+LSE (mm day)

9 -1

ETo MA+LSE (mm day )

2

-1

7 6 5 4 3 2

y = 1.0019x 2 R = 0.9418

1

7 6 5 4 3 2

y = 1.0443x 2 R = 0.9343

1

0

0 0

1

2

3

4

5

6

7

8

9

10

0

1

2

-1

3

4

5

6

7

8

9

10

-1

ET o PM-FAO56 (mm day )

ET o PM-FAO56 (mm day )

10

VEJ_CA 8

-1

ETo MA+LSE (mm day )

9

7 6 5 4 3 y = 0.9368x

2

2

R = 0.8894

1 0 0

1

2

3

4

5

6

7

8

9

10

-1

ET o PM-FAO56 (mm day ) Fig. 3. Regressions between daily ETo using PM-FAO56 and MA + LSE for five weather stations (COR CO, SAN CO, LEB SE, JER CA and VEJ CA).

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

10

10

MAR_JA

8

8

-1

ETo MA+LSE (mm day )

9

-1

ETo MA+LSE (mm day )

GUI_SE 9

141

7 6 5 4 3 y = 0.9728x

2

7 6 5 4 3 y = 1.0542x

2

2

2

R = 0.9552

1

R = 0.9327

1

0

0 0

1

2

3

4

5

6

7

8

10

9

0

1

2

-1

4

5

6

7

8

9

10

-1

ET o PM-FAO56 (mm )

ET o PM-FAO56 (mm day )

10

10

YEG_MA

TRP_JA 9

8

8

-1

ETo MA+LSE (mm day )

9 -1

ETo MA+LSE (mm day )

3

7 6 5 4 3 y = 0.9221x

2

7 6 5 4 3 y = 0.9172x

2

2

2

R = 0.9575

1

R = 0.9473

1

0

0 0

1

2

3

4

5

6

7

8

10

9

0

1

2

-1

3

4

5

6

7

8

9

10

-1

ET o PM-FAO56 (mm day )

ET o PM-FAO56 (mm day )

10

IZN_GR 8

-1

ETo MA+LSE (mm day )

9

7 6 5 4 3 y = 0.9202x

2

2

R = 0.9513

1 0 0

1

2

3

4

5

6

7

8

9

10

-1

ET o PM-FAO56 (mm day ) Fig. 4. Regressions between daily ETo using PM-FAO56 and MA + LSE for five weather stations (GUI SE, MAR JA, YEG MA, TRP JA and IZN GR).

underestimation (3.3 and 1.5%, respectively). Analyzing each season and location MA + LSE gave a maximum ETo underestimation compared with PM-FAO56 of 10.2% and with HAR of 15.8%, and maximum overestimations of 6.9 and 17%, respectively (Table 1). However, the MAK and PT approaches showed significant average underestimations (12.8 and 16.8%, respectively), with maximum underestimations of 18.1 and 21.3%, respectively. These low ETo

values agree with previous studies that confirmed ETo underestimations using the Makkink equation under semi-arid conditions (Allen et al., 1998; Irmak et al., 2008; Stisen et al., 2008b) due to the omission of the aerodynamic component in ETo computation. Finally, Cammalleri and Ciraolo (2013) also determined underestimations with the Makkink approach, especially in sites with a strong wind. Also, the MP-PM approach gave very similar average

142

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

Table 1 Annual and average rainfall, ETo , irrigation and estimated yield for the six approaches for ETo assessment. In addition, average and maximum underestimation and overestimation for ETo , irrigation and yield are included. Approach Rainfall (mm year−1 )

2006/07

2007/08

2008/09

515

570

540

PM-FAO56 MA + LSE HAR MAK PT MP-PM

1333 1306 1324 1168 1121 1332

1375 1329 1358 1213 1140 1401

1375 1314 1339 1179 1135 1363

PM-FAO56 MA + LSE HAR MAK PT MP-PM

530 527 527 435 439 536

546 536 547 449 456 553

PM-FAO56 MA + LSE HAR MAK PT MP-PM

12,688 12,499 12,551 11,697 11,691 12,163

12,825 12,566 12,708 11,873 11,964 12,586

534

Avg. period

Avg. desv. (%)

Max. under. (%)

Max. over. (%)

1361 1316 1340 1187 1132 1365

−3.3 −1.5 −12.8 −16.8 0.3

10.2 15.8 18.1 21.3 13.0

6.9 17.0

483 456 474 379 381 479

520 506 516 421 425 522

−2.6 −0.7 −19.0 −18.2 0.5

13.8 23.8 28.6 28.6 12.8

12.0 20.7

12,157 11,774 12,098 11,030 10,697 12,071

12,557 12,280 12,452 11,533 11,451 12,273

−2.2 −0.8 −8.1 −8.8 −2.3

8.6 5.5 20.6 32.1 28.9

0.5 1.4

ETo (mm year−1 )

11.5

Irrigation (mm)

−1

Estimated yield (kg ha

13.3

)

ETo values compared with PM-FAO56 (overestimations of 0.3%; Table 1), with maximum over and underestimations of around 12%. 3.3. Irrigation scheduling Irrigation scheduling is directly related to rainfall. Thus, those locations and years with a lower rainfall showed the highest

0.0 6.4

irrigation demands (e.g. LFP irrigation scheme showed the minimum rainfall and had the maximum irrigation requirements; Fig. 5). However, ETo also played a significant role in irrigation scheduling, and, consequently, the approach considered for ETo assessment could affect irrigation scheduling. Fisher and Pringle (2010) indicated that irrigation schedules were greatly influenced by the ETo method used, and underlined the importance of selecting

Fig. 5. Volume of irrigation determined by the six approaches for ETo assessment (PM-FAO56, MA + LSE, HAR, MAK, PT and MP-PM) and simulated yield considering the previous irrigation schedules for the ten irrigated areas considered in the study.

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145 Table 2 RMSD, rRMSD and linear regression components for the volume of irrigation and estimated yield compared with the PM-FAO56 approach for the alternative five approaches for ETo assessment. Approach

RMSD (mm)

rRMSD (%)

Slope

R2

7.2 10.2 19.4 19.3 6.8

0.97 0.99 0.81 0.82 1.00

0.79 0.59 0.91 0.83 0.79

Irrigation MA + LSE HAR MAK PT MP-PM

37.2 53.2 101.1 100.3 35.3

Approach

RMSD (kg ha−1 )

rRMSD (%)

Slope

R2

MA + LSE HAR MAK PT MP-PM

415.4 241.0 1194.1 1418.5 962.8

3.3 1.9 9.5 11.3 7.7

0.98 0.99 0.92 0.91 0.98

0.91 0.94 0.70 0.60 0.57

Estimated yield

an appropriate method for estimating ETo , emphasizing that alternative methods to PM-FAO56 could provide accurate estimations under local environmental conditions. The irrigation schedules originated by AquaCrop based on ETo values provided by the PM-FAO56 approach recommended a seasonal average irrigation volume of 520 mm (Table 1), varying from 442 mm for the BG irrigation scheme to 635 for the LFP irrigation scheme (Fig. 5). Using as a reference these irrigation volumes, and considering the ETo values provided by the MA + LSE approach, a very small underestimation of 2.6% in the irrigation volume was determined (Table 1 and Fig. 5). Considering every irrigation season and location, the maximum underestimation with the MA + LSE approach was 13.8%, while the maximum overestimation was 12%. The differences in irrigation volume considering ETo values based on the MA + LSE instead of the PM-FAO56 approach were very limited: RMSD was 37.2 mm, regression slope equal to 0.97 and R2 = 0.79 (Table 2). Analyzing separately each irrigation scheme, a similar spatial pattern to that for ETo was determined (Section 3.1): those locations with the greatest differences between ETo values provided by the PM-FAO56 and MA + LSE approaches were the areas with the greatest differences between the irrigation volumes provided by both approaches (DO and CC irrigation schemes; Fig. 5). Irrigation scheduling based on the HAR approach for ETo assessment produced very similar average irrigation volumes to those based on the PM-FAO56 approach (underestimation of 0.7%; Table 1). In spite of this similar average simulated irrigation volume, maximum divergences compared with the PM-FAO56 approach ranged from underestimations of 23.8% to overestimations of 20.7%, which were significantly higher than those detected when the MA + LSE approach was used. Similarly, RMSD was higher for schedules based on the HAR approach (53.2 mm) and R2 significantly lower (0.59; Table 2). This fact implies that, although for average irrigation volume assessment the HAR approach was very satisfactory for some individual irrigation seasons and locations, significant errors in their irrigation scheduling were produced, inducing severe stress or over-irrigation. Analyzing average irrigation volumes based on the MA + LSE and HAR approaches, a similar performance to that obtained with PM-FAO56 was determined, although the MA + LSE approach gave smaller extreme values (Table 1), showing a better performance for individual seasons and locations. Irrigation schedules using MAK or PT approaches showed clear underestimations compared with PM-FAO56 (19.0 and 18.2%, respectively; Table 1), producing clear deficitarian irrigation scheduling, and RMSD values of around 100 mm (Table 2). These irrigation volume underestimations were caused by the ETo

143

underestimation made by the MAK and PT approaches, as was previously described in Section 3.2. MP-PM provided accurate average irrigation volumes (over-irrigation of 0.5% with RMSD equal to 35.3 mm; Tables 1 and 2), but when analyzing individually all the irrigation seasons and locations, significant divergences were detected (maximum under- and overestimations were 12.8 and 13.3%, respectively, Table 1). The differences are compensated for when average values are considered but are revealed when years and locations are individually analyzed. 3.4. Yield and irrigation water productivity assessment Under semi-arid conditions characterized by limited rainfall, maize production is clearly dependent on the irrigation applied (Payero et al., 2008). Spatial differences in yield (Fig. 5) were caused by differences in irrigation scheduling, especially during dry years when the irrigation schedules were not able to meet the full crop requirements, and so yield was reduced. Yield analysis not only considered the volume of irrigation applied for each approach but irrigation timing was also considered, giving an effective indicator of the quality of the irrigation scheduling based on different approaches for ETo assessment. Average estimated yield using the AquaCrop model with irrigation scheduling based on the PM-FAO56 approach was 12.6 t ha−1 (Table 1). Using the irrigation schedules based on the MA + LSE approach, the estimated yield was very similar (12.3 t ha−1 ), with an underestimation of 2.2%, indicating the high quality of the irrigation schedules based on ETo values calculated with the MA + LSE approach. Compared with the PM-FAO56 approach, RMSD was equal to 415 kg ha−1 , the linear regression slope was 0.98 and R2 = 0.91 (Table 2). Yield losses using irrigation scheduling based on the MA + LSE instead of on the PM-FAO56 approach were linearly correlated with the differences in ETo and the irrigation volume required when the MA + LSE and PM-FAO56 approaches were considered. Thus, those areas with the greatest differences between irrigation volumes based on the PM-FAO56 and MA + LSE approaches were the areas with the highest yield losses due to using MA + LSE instead of the PM-FAO56 approach (DO and CC irrigation schemes; Fig. 5). The simulated yield considering irrigation scheduling based on the HAR approach was very similar to the simulated yield with irrigation schedules based on the MA + LSE and PM-FAO56 approaches. A slight reduction in average yield of 0.8%, with maximum reductions of 5.5% and RMSD = 241 kg ha−1 , was determined. The HAR approach gave a better performance than the MA + LSE approach for yield assessment due to the effect of over-irrigation on yield (maximum over-irrigation with HAR reached 20.7%; Table 1). Irrigation scheduling based on the MAK and PT approaches caused important yield losses due to the clearly deficitarian irrigation schedules described previously (Table 1). Average yield reductions were 8.1 and 8.8%, with maximum reduction of 20.6 and 32.1%, respectively. RMSD was 1.2 and 1.4 t ha−1 for irrigation scheduling based on MAK and PT approaches, respectively (Table 2). Irrigation scheduling based on the MP-PM approach gave a satisfactory average yield (underestimations limited to 2.3%), although some specific errors in irrigation timing triggered severe yield underestimations of 28.9% (Table 1). This fact was caused by nonadjusted irrigation schedules to local conditions causing severe crop stress. RMSD was 963 kg ha−1 with a regression slope equal to 0.98 but R2 = 0.57 (Table 2). These results confirm that the use of weather data from weather stations located far away from the irrigation scheme could produce significant yield losses caused by irrigation schedules non-adjusted to local weather conditions, and so their use should be carefully considered. IWP had the same average value for PM-FAO56, MA + LSE and HAR (23.8 kg m−3 ). Due to the severe deficit irrigation produced

144

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145

with the MAK and PT approaches (Table 1), IWP increased to 27.2 kg m−3 . However, the MP-PM approach showed an IWP equal to 23.1 kg m−3 , the smallest value, caused by the wrong irrigation scheduling practiced in some individual locations and years that caused severe reductions in simulated yield. Considering irrigation scheduling, yield and IWP, the MA + LSE approach showed very satisfactory results compared with those obtained by approaches such as PM-FAO56 with high data requirements, thus emerging as an excellent tool for improving irrigation water management. 4. Conclusions The procedure for ETo assessment used for irrigation scheduling has shown itself to be an essential factor in the quality of irrigation water management. Irrigation scheduling and yield estimation based on different approaches for ETo assessment were determined with the AquaCrop model for maize crop in several irrigation schemes located in Southern Spain. The use of remote sensing techniques (LSA SAF) and forecast tools (ECMWF) in an advection-revised Makkink equation (MAK-Adv) integrated in the approach called MA + LSE provided very satisfactory results, which were similar to those obtained with the PM-FAO56 approach (irrigation volume and yield provided by MA + LSE was underestimated by 2.6 and 2.2%, respectively, compared with PM-FAO56), with the advantage of MA + LSE being the only methodology that did not require data from local weather stations. Other methods such as HAR or MP-PM supplied satisfactory averaged results, but individual errors produced important over-irrigations or severe crop deficit. Additionally, both methodologies required accurate meteorological data from weather stations, a fact that prevents the use of these approaches in numerous areas around the world. Methodologies only requiring solar radiation (MAK and PT) showed serious limitations for ETo assessment under the semi-arid and advective conditions detected in Southern Spain, advising against its use without any correction or local calibration. Finally, the use of weather stations not located in the vicinity of the analyzed irrigation scheme, gave satisfactory averaged results, but occasional errors in the irrigation timing implied significant crop deficits severely affecting crop yield. Therefore, the MA + LSE approach has been seen to be a valid alternative for ETo assessment for improving irrigation water management by the programming of accurate irrigation scheduling under semi-arid conditions in Southern Spain. This new methodology is applicable to vast areas thanks to the use of remote sensing techniques and forecasting tools and has the advantage of its low input requirements compared to other well-known approaches such as PM-FAO56, although a regional evaluation process of the MAK-Adv equation is advised prior to its use for irrigation water management at a regional scale. Acknowledgments This paper was supported by the grants P10-EXC10-0036/AGR6126 and PP.AVA.AVA201301.10 from the Regional Government of Andalusia and RTA2011-00015-00-00 from the National Institute for Agricultural and Food Research and Technology (INIA). This work was also supported by the EUMETSAT within the framework of the LSA SAF. References Allen, R.G., 1996. Assessing integrity of weather data for reference evapotranspiration estimation. J. Irrig. Drain. Eng. ASCE 122 (2), 97–106. Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements. FAO Irrigation and Drainage Paper No. 56. FAO, Rome, Italy.

Allen, R.G., Droogers, P., Hargreaves, G., 2002. Predicting reference crop evapotranspiration with arid weather data. In: 18th International Congress on Irrigation and Drainage, Montreal, Canada. Allen, R.G., Tasumi, M., Trezza, R., 2007a. Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)–model. J. Irrig. Drain. Eng. ASCE 133 (4), 380–394. Allen, R.G., Tasumi, M., Morse, A., Trezza, R., Wright, J.L., Bastiaanssen, W., Kramber, W., Lorite, I.J., Robison, C.W., 2007b. Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)–applications. J. Irrig. Drain. Eng. ASCE 133 (4), 395–406. Bastiaanssen, W.G.M., Menenti, M., Feddes, R.A., Holtslag, A.A.M., 1998. A remote sensing surface energy balance algorithm for land (SEBAL): 1. Formulation. J. Hydrol. 212-213, 198–212. Berengena, J., Gavilán, P., 2005. Reference evapotranspiration estimation in a highly advective semiarid environment. J. Irrig. Drain. Eng. ASCE 131 (2), 147–163. Cammalleri, C., Ciraolo, G., 2013. A simple method to directly retrieve reference evapotranspiration from geostationary satellite images. Int. J. Appl. Earth Obs. Geoinf. 21, 149–158. CAP, 2011. Inventario de regadíos 2008 y su evolución en la última década. Consejería de Agricultura y Pesca. Junta de Andalucía, Seville (in Spanish). Christensen, J.H., Christensen, O.B., 2007. A summary of the PRUDENCE model projections of changes in European climate by the end of this century. Clim. Change 81, 7–30. Cristobal, J., Anderson, M.C., 2012. Regional scale evaluation of a MSG solar radiation product for evapotranspiration modeling. Hydrol. Earth Syst. Sci. Discuss 9, 8905–8939. Cruz-Blanco, M., Gavilán, P., de Bruin, H.A.R., Trigo, I.F., Santos, C., Lorite, I.J., 2014. Reference evapotranspiration assessment using remote sensing and forecasting tools under semi-arid conditions. Hydrol. Process. (Under review). Dai, X., Shi, H., Li, Y., Ouyang, Z., Huo, Z., 2009. Artificial neural network models for estimating regional reference evapotranspiration based on climate factors. Hydrol. Process. 23, 442–450. De Bruin, H.A.R., 1987. From Penman to Makkink. Comm. Hydrol. Res. TNO Proc. Inform. Den Haag 39, 5–30. De Bruin, H.A.R., Trigo, I.F., Jitan, M.A., Temesgen, E.N., van der Tol, C., Gieske, A.S.M., 2010. Reference crop evapotranspiration derived from geo-stationary satellite imagery: a case study for the Fogera flood plain, NW-Ethiopia and the Jordan Valley. Jordan. Hydrol. Earth Syst. Sci. 14, 2219–2228. De Bruin, H.A.R., Trigo, I.F., Lorite, I.J., Cruz-Blanco, M., Gavilán, P., 2012. Reference Crop Evapotranspiration obtained from the geostationary satellite MSG (METEOSAT). Geogr. Res. Abs. 14, EGU 2012-11453. Er-Raki, S., Chehbouni, A., Khabba, S., Simonneaux, V., Jarlan, L., Ouldbba, A., Rodriguez, J.C., Allen, R., 2010. Assessment of reference evapotranspiration methods in semi-arid regions: can weather forecast data be used as alternate of ground meteorological parameters? J. Arid Environ. 74, 1587–1596. Espadafor, M., Lorite, I.J., Gavilán, P., Berengena, J., 2011. An analysis of the tendency of reference evapotranspiration estimates and other climate variables during the last 45 years in Southern Spain. Agric. Water Manage. 98, 1045–1061. Feddes, R.A., 1987. Crop factors in relation to Makkink reference-crop evapotranspiration. In: Hooghart, C. (Ed.), Evaporation and Weather. Proceedings and Information. Comm. Hydrological Research TNO, The Hague, pp. 33–47. Fisher, D.K., Pringle III, H.C.,2010. Effect of reference ET method on irrigation scheduling model. In: 5th National Decennial Irrigation Conference. Paper number: IRR10-9814. Phoenix, Arizona, USA. García-Vila, M., Fereres, E., 2012. Combining the simulation crop model AquaCrop with an economic model for the optimization of irrigation management at farm level. Eur. J. Agron. 36, 21–31. García-Vila, M., Lorite, I.J., Soriano, M.A., Fereres, E., 2008. Management trends and responses to water scarcity in an irrigation scheme of Southern Spain. Agric. Water Manage. 95, 458–468. Gavilán, P., Lorite, I.J., Tornero, S., Berengena, J., 2006. Regional calibration of Hargreaves equation for estimating reference ETo in a semiarid environment. Agric. Water Manage. 81, 257–281. Geerts, S., Raes, D., Garcia, M., 2010. Using AquaCrop to derive deficit irrigation schedules. Agric. Water Manage. 98, 213–216. Geiger, B.C., Meurey, D., Lajas, L., Franchistéguy, D., Carrer, J.L., Roujean, 2008. Near real-time provision of downwelling shortwave radiation estimates derived from satellite observations. Meteorol. Apps. 15, 411–420. Ghilain, N., Arboleda, A., Gellens-Meulenberghs, F., 2011. Evapotranspiration modelling at large scale using near-real time MSG SEVIRI derived data. Hydrol. Earth Syst. Sci. 15, 771–786. Ghilain, N., Arboleda, A., Sepulcre-Cantó, G., Batelaan, O., Ardö, J., GellensMeulenberghs, F., 2012. Improving evapotranspiration in a land surface model using biophysical variables derived from MSG/SEVIRI satellite. Hydrol. Earth Syst. Sci. 16, 2567–2583. Hargreaves, G.H., Samani, Z.A., 1985. Reference crop evapotranspiration from temperature. Appl. Eng. Agric. 1 (2), 96–99. Hsiao, T.C., Heng, L., Steduto, P., Rojas-Lara, B., Raes, D., Fereres, E., 2009. AquaCrop – the FAO crop model to simulate yield response to water: III. Parameterization and testing for maize. Agron. J. 101, 448–459. Irmak, A., Irmak, S., Martin, D.L., 2008. Reference and crop evapotranspiration in South Central Nebraska. I: Comparison and analysis of grass and alfalfa-reference evapotranspiration. J. Irrig. Drain Eng. ASCE 134 (6), 690–699. Jabloun, M., Sahli, A., 2008. Evaluation of FAO-56 methodology for estimating reference evapotranspiration using limited climatic data. Application to Tunisia. Agric. Water Manage. 95, 707–715.

M. Cruz-Blanco et al. / Agricultural Water Management 131 (2014) 135–145 Landeras, G., Ortiz-Barredo, A., López, J.J., 2008. Comparison of artificial neural network models and empirical and semi-empirical equations for daily reference evapotranspiration estimation in the Basque Country (Northern Spain). Agric. Water Manage. 95, 553–565. López-Urrea, R., Martín De Santa Olalla, F., Fabeiro, C., Moratalla, A., 2006. Testing evapotranspiration equations using lysimeter observations in a semiarid climate. Agric. Water Manage. 85, 15–26. Lorite, I.J., Mateos, L., Orgaz, F., Fereres, E., 2007. Assessing deficit irrigation strategies at the level of an irrigation district. Agric. Water Manage. 91, 51–60. Lorite, I.J., García-Vila, M., Santos, C., Ruiz-Ramos, M., Fereres, E., 2013. AquaData and AquaGIS: two computer utilities for temporal and spatial simulations of water-limited yield with AquaCrop. Comput. Electron. Agric. 96, 227–237. Makkink, G.F., 1957. Testing the Penman formula by means of lysimeters. J. Inst. Wat. Engrs. 11, 277–288. Meek, D.W., Hatfield, J.L., 1994. Data quality checking for single station meteorological databases. Agric. Forest Meteorol. 69 (1–2), 85–109. Monteith, J.L.,1965. Evaporation and environment. In: 19th Symposium, vol. 19. University Press. Society for Experimental Biology, Cambridge, pp. 205–234. Payero, J.O., Tarkalson, D.D., Irmak, S., Davison, D., Petersen, J.L., 2008. Effects of irrigation amounts applied with subsurface drip irrigation on corn evapotranspiration, yield, water use efficiency, and dry matter production in a semiarid climate. Agric. Water Manage. 95, 895–908. Pereira, A.R., Pruitt, W.O., 2004. Adaptation of the Thornthwaite scheme for estimating daily reference evapotranspiration. Agric. Water Manage. 66, 251–257. Persson, A., 2011. User guide to ECMWF forecast products. ECMWF. Available on line at http://www.ecmwf.int/products/forecasts/guide/user guide.pdf Priestley, C.H.B., Taylor, R.J., 1972. On the assessment of surface heat flux and evaporation using large-scale parameters. Mon. Weather Rev. 100, 81–82. Raziei, T., Pereira, L.S., 2013. Estimation of ETo with Hargreaves-Samani and FAO-PM temperature methods for a wide range of climates in Iran. Agric. Water Manage. 121, 1–18. Santos, C., Lorite, I.J., Tasumi, M., Allen, R.G., Fereres, E., 2008. Integrating satellitebased evapotranspiration with simulation models for irrigation Management at the scheme level. Irrig. Sci. 26, 277–288. Santos, C., Lorite, I.J., Tasumi, M., Allen, R.G., Fereres, E., 2010. Performance assessment of an irrigation scheme using indicators determined with remote sensing techniques. Irrig. Sci. 28, 461–477.

145

Schmetz, J., Pili, P., Tjemkes, S., Just, D., Kerkman, J., Rota, S., Ratier, A., 2002. An introduction to Meteosat Second Generation (MSG). Bull. Am. Meteorol. Soc. 83, 977–992. Sentelhas, P.C., Gillespie, T.J., Santos, E.A., 2010. Evaluation of FAO Penman–Monteith and alternative methods for estimating reference evapotranspiration with missing data in Southern Ontario. Can. Agric. Water Manage. 97, 635–644. Shafer, M.A., Fiebrich, C.A., Arndt, D.S., 2000. Quality assurance procedures in the Oklahoma Mesonetwork. J. Atmos. Oceanic Technol. 17 (4), 474–494. Shiri, J., Kisi, O., Landeras, G., López, J.J., Nazemi, A.H., Stuyt, L.C.P.M., 2012. Daily reference evapotranspiration modelling by using genetic programming approach in the Basque Country (Northern Spain). J. Hydrol. 414–415, 302–316. Steduto, P., Hsiao, T.C., Fereres, E., Raes, D., 2012. Crop Yield Response to Water. FAO Irrigation and Drainage Paper, vol. 66. FAO, Rome. Stisen, S., Sandholt, I., Norgaard, A., Fensholt, R., Jensen, K.H., 2008a. Combining the triangle method with thermal inertia to estimate regional evapotranspiration. Applied to MSG-SEVIRI data in the Senegal River basin. Remote Sens. Environ. 112, 1242–1255. Stisen, S., Jensen, K.H., Sandholt, I., Grimes, D.I.F., 2008b. A remote sensing driven distributed hydrological model of the Senegal River basin. J. Hydrol. 354, 131–148. Sun, Z., Gebremichael, M., Ardö, J., de Bruin, H.A.R., 2011. Mapping daily evapotranspiration and dryness index in the East African highlands using MODIS and SEVIRI data. Hydrol. Earth Syst. Sci. 15, 163–170. Temesgen, B., Allen, R.G., Jensen, D.T., 1999. Adjusting temperature parameters to reflect well-watered conditions. J. Irrig. Drain. Eng. ASCE 125 (1), 26–33. Trigo, I.F., Dacamara, C.C., Viterbo, P., Roujean, J.L., Olesen, F., Barroso, C., Camacho-de-Coca, F., Carrer, D., Freitas, S.C., García-Haro, J., Geiger, B., GellensMeulenberghs, F., Ghilain, N., Meliá, J., Pessanha, L., Siljamo, N., Arboleda, A., 2011. The satellite application facility for land surface analysis. Int. J. Remote Sens. 32, 2725–2744. Utset, A., Farré, I., Martínez-Cob, A., Cavero, J., 2004. Comparing Penman–Monteith and Priestley–Taylor approaches as reference-evapotranspiration inputs for modeling maize water-use under Mediterranean conditions. Agric. Water Manage. 66, 205–219. Zwart, S.J., Bastiaanssen, W.G.M., 2004. Review of measured crop water productivity values for irrigated wheat, rice, cotton and maize. Agric. Water Manage. 69, 115–133.