Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía)

G Model

ARTICLE IN PRESS

AGWAT-4039; No. of Pages 12

Agricultural Water Management xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

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

Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía) ˜ B.J. Rey, E.M. Martínez J.J. Cancela ∗ , M. Fandino, GI-1716, Department of Agroforestry Engineering, University of Santiago de Compostela, Campus Universitario, 27002 Lugo, Spain

a r t i c l e

i n f o

Article history: Received 4 May 2014 Received in revised form 10 October 2014 Accepted 20 October 2014 Available online xxx Keywords: Irrigation management Leaf water potential Vine transpiration SIMDualKc model Soil evaporation Soil water potential

a b s t r a c t This research aims at testing an automatic control irrigation system, using a wireless sensor network, in traditional Galician vineyards of Vitis vinifera (L.) cv. ‘Godello’ and cv. ‘Mencía’ to determine the threshold values of soil water potential at which plant stress begins, calibrating crop coefficients, building soil–water characteristics curves and measuring plant water status. In the cv. ‘Godello’ trial, rain-fed and two irrigations systems (surface and subsurface drip irrigation) were conducted over two growing seasons (2012–2013); during the same seasons cv ‘Mencía’ was also studied, but only under rain-fed conditions. The SIMDualKc model, which estimates soil water balance by means of the dual Kc approach, was used to estimate crop evapotranspiration (ETc ) by calibrating the full basal crop coefficient for the vine and cover crop (Kcb full ), which represents the transpiration component of ETc , and a soil evaporation coefficient (Ke ). The model was calibrated and validated by comparing model simulations with TDR observed soil water content data. Granular matrix sensor (GMS) was linked in a wireless sensor network; soil water potential measured with GMS, was used to correlate with TDR data. Leaf water potentials (LWP) – midday and stem – allowed us to obtain plant water status. A good fit was obtained between SIMDualKc model and TDR (r2 > 0.74), TDR and LWP (r2 > 0.65), TDR and GMS (r2 > 0.81), showing that continuous measures with GMS permit establishing a threshold value related with leaf water potential (midday or stem). For both cultivars, the threshold was  soil = −0.1 MPa. The process applied in this study proved to be useful for managing water in real-time in a vineyard; triggering the irrigation system when the threshold value was reached. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Water resource scarcity worldwide and competition among different water uses (FAO, 2013) makes necessary to find solutions that provide understanding the effects of water use reduction in crop irrigation (Hayashi et al., 2012), as well as to know the time and depth needed by plants (Pereira, 1999); thus making agricultural water use more efficient and competitive. The increasing technification in agriculture, by establishing monitoring networks in different crops (Zhang et al., 2013b), is presented as a tool for improving real-time irrigation efficiency (Smarsly, 2013). The availability of different variables in real-time, by the use of sensors, avoiding discontinuous measurements in the field (Acevedo-Opazo et al., 2010) is a key issue in precision irrigation management and, therefore, for irrigation scheduling, currently developed in many

∗ Corresponding author. Tel.: +34 666367471. E-mail address: [email protected] (J.J. Cancela).

horticultural crops (López et al., 2011) and in main extensive crops (Ruíz-García et al., 2009). In vineyards, irrigation effects need to be interpreted regarding soil water status, crop ecophysiological parameters and the studied cultivar (Cifre et al., 2005), as well as qualitative characteristics of the final product. Hence, automatic irrigation systems in vineyards need to integrate all these parameters and their relationships. Recently, many studies have shown the irrigation effects in vineyards for different cultivars and locations (Azevedo et al., 2008; Santesteban et al., 2011; Gouveia et al., 2012; Trigo-Córdoba et al., 2013). In general, the reference parameter used for irrigation management was the leaf water potential (Girona et al., 2006; van Leeuwen et al., 2009; Martínez et al., 2011, 2013), usually measured with an Scholander pressure chamber (Scholander et al., 1965), at midday ( m ) or stem ( stem ). Centeno et al. (2010) established relationships among these plant parameters and soil water potential ( soil ), for Tempranillo cultivar, obtaining good adjustments as well as Williams and Araujo (2002) for Chardonnay and Cabernet Sauvignon, or Williams and Trout (2005) for Thompson seedless. In addition, soil texture (Tramontini et al., 2013) and soil hydraulic

http://dx.doi.org/10.1016/j.agwat.2014.10.020 0378-3774/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model AGWAT-4039; No. of Pages 12

ARTICLE IN PRESS J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

2

characteristics (Centeno et al., 2010) in the studied vineyards were required to manage irrigation correctly, in particular soil–water characteristic curves using soil water potential ( soil ) and three soil water content () attributes: saturation, field capacity (FC) and permanent wilting point (PWP) (Martínez et al., 2012). Simultaneously, research on crop evapotranspiration (ETc ) and water use in vineyards has been performed using various methods and techniques including lysimeters (Azevedo et al., 2008), heat pulse and heat balance (Yunusa et al., 1997; Trambouze and Voltz, 2001; Intrigliolo et al., 2009; Zhang et al., 2010), Bowen ratio and surface renewal energy balance (Yunusa et al., 2004; Moratiel and Martínez-Cob, 2011; Zhang et al., 2011), eddy covariance and soil water balance (Singleton and Maudsley, 1996; Fooladmand and ˜ et al., 2012a). Sepaskhah, 2009; Ortega-Farias et al., 2010; Fandino Current approaches are able to define a crop coefficient (Kc ) when the observed ETc is related to a reference crop evapotranspiration (ETo ) computed with local data. Hence, ETc can be estimated for a vineyard by using the following equation: ETc = Kc × ETo (Allen et al., 1998). The dual Kc approach with calibration of Kcb for local conditions, crop variety, trellis training system, cover crop and cultural ˜ (Fandino ˜ et al., 2012a), using practices was applied to cv Albarino the SIMDualKc model (Rosa et al., 2012a). Poblete-Echeverría and Ortega-Farias (2013) had evaluated Kcb for cv. Merlot to know the actual evapotranspiration in two seasons. Fooladmand and Sepaskhah (2009) used the dual Kc approach to assess water use of vineyards in a water harvesting system. Differently, Yunusa et al. (2004) estimated the vineyard transpiration (Tv ) and soil evaporation (Es ) components of ETc from separate measurements but not by modeling. The present study adopted the dual crop coefficient approach to estimate ETc in drip-irrigated vineyards of Vitis vinifera cv. ‘Godello’ (white variety) and rain-fed vineyards of cv. ‘Mencía’ (red variety) in Galicia, Northwest Spain. To adopt the dual Kc approach for these cultivars of grapevine is innovative and interesting since they are typical from Galician Designations of Origin (DO). In fact, Godello is the main variety in Valdeorras DO and Mencía is the main one in Ribeira Sacra DO. Moreover, they are widely cultivated in the rest of the Galician DO. In addition, there are no previous studies about crop water requirements accounting for soil and plant water status for these cultivars. Thus, the objectives of this study consist of: (a) computing ETc of two grapevine cultivars using the dual Kc approach, thus with its separation into crop transpiration and soil evaporation; (b) testing the SIMDualKc model by calibrating and validating the Kcb full values appropriate for these cultivars, using observed soil water data relative to various irrigation treatments and two years of observations (2012–2013); and (c) defining a threshold of water stress to trigger the irrigation, applying soil and plant water measurements, using an automatic irrigation system with real-time data, to keep an optimal range for grapevine activity.

2. Materials and methods 2.1. Site, crop and treatments The experiment was conducted during 2012 and 2013 in two commercial vineyards planted with cv ‘Godello’ and ‘Mencía’. ‘Godello’ plot is located in A Rúa (Galicia-NW Spain) within the DO Valdeorras (latitude 42◦ 23 59 N, longitude 7◦ 7 15 W and altitude 320 m above sea level, mean slope is 18%). Soil at the site is sandy clay-loam with 46.2% sand, 31% silt and 22.8% clay, pH (H2 O) 4.99 and 2.26% organic matter. Soil depth was, approximately, 1.2 m. No operations of soil tillage were undertaken during the studied seasons.

Table 1 Soil characteristics for the entire root depth (0.6 m) in DO Valdeorras and DO Ribeira Sacra (LAD-Ladredo, MEI-Meixemán). DO Valdeorras

3

−3

FC (m m ) PWP (m3 m−3 ) Saturation (m3 m−3 ) TAW (mm)

0.25 0.08 0.38 102

DO Ribeira Sacra LAD

MEI

0.23 0.07 0.28 96

0.20 0.04 0.24 96

FC—field capacity, PWP—permanent wilting point, TAW—total available water.

In the case of cv. ‘Mencía’ two test plots: ‘Meixemán’ (MEI) and ‘Ladredo’ (LAD), were studied, both located in the ‘Amandi’ subarea (DO Ribeira Sacra), in Doade (Galicia, NW Spain), arranged in terraces and without irrigation (MEI: latitude 42◦ 24 27 N, longitude 7◦ 27 24 W and altitude 439 m above sea level; LAD: latitude 42◦ 24 42 N, longitude 7◦ 27 10 W and altitude 330 m above sea level). Soil in ‘Mencía’ plots has an average pH (H2 O) of 5.40; 10.6% and 6.2% organic matter, in MEI and LAD, respectively. MEI soil site is loamwith 41.5% sand and 40.8% silt, and in LAD plot is sandy-loam with 57.6% sand and 26.1% silt. All data required to compute the soil water balance are presented in Table 1. Total available soil water (TAW) down to 0.6 m depth was calculated as the difference between the average FC and the PWP. The total evaporable water (TEW), the readily evaporable water (REW) and the depth of the evaporable layer (Ze ) were first estimated from standard data proposed by Allen et al. (1998, 2005) and then adjusted when calibrating/validating the model. The average crop height and the effective root depth averaged (0.6 m depth) were observed during all seasons (Table 2). Meteorological data was collected from weather stations managed by MeteoGalicia, the Galician Meteorological Agency, ‘Larouco’ (cv ‘Godello’) and ‘Ponte Boga’ (cv. ‘Mencía’) located at less than 7 km and 4 km to the experimental sites, respectively. In ‘Godello’ plot, from March to October, the average temperature was 16.2 ◦ C and 16.3 ◦ C, for 2012 and 2013, respectively; total rainfall during this period was 347.9 mm and 518.2 mm for 2012 and 2013. In the case of ‘Mencía’ plots the average temperature was 16.1 ◦ C and 16.2 ◦ C, for 2012 and 2013, respectively; total rainfall from March to October was 449.7 mm and 623.9 mm for 2012 and 2013. Rainfall and reference evapotranspiration are shown in Fig. 1. The reference evapotranspiration (ETo ) was computed with the Penman–Monteith equation using the methodology proposed by Allen et al. (1998) for limited weather data, i.e., estimating the actual vapor pressure from the daily minimum temperature and solar radiation from daily maximum and minimum temperature. In the case of cv ‘Godello’, white grapevine cultivar native from Galicia, plants were about 15 years old at the beginning of this experiment; they were vertically shoot positioned and grafted on rootstock 110R. The spacing is 1.35 m × 1.95 m (3800 plants ha−1 ). Two treatments were established following a completely randomized block design with four replications (7 plants each). The treatments were: rain-fed (R) and irrigation [surface (DI) and subsurface (SDI) drip irrigation systems]. The surface irrigation pipes were in the vineyard row at 40 cm above the soil, whereas the subsurface pipes were 40 cm deep into the soil, approximately. ˜ et al., Both systems were equipped with 2 L h−1 emitters (Fandino 2013). In the case of cv ‘Mencía’, red grapevine cultivar native from Galicia, plants were about 70 years old at the beginning of this experiment; they were vertically shoot positioned. The vines were arranged in terraces, with a distance between plants of 1.4 m (4200 plants ha−1 ). Grapevine phenology was surveyed over the studied growing seasons and was used to establish crop development stages (Table 2), these values were used in the soil water balance model.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model

ARTICLE IN PRESS

AGWAT-4039; No. of Pages 12

J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

3

Table 2 Vineyard crop growth stages, height (h) and fraction of soil covered by the crop (fc ) of DO Valdeorras and DO Ribeira Sacra vineyards. Crop growth stages

cv ‘Godello’

cv ‘Mencía’

DO Valdeorras

LAD–DO Ribeira Sacra

2012

2013

Date Planting/initiation Start rapid growth Start mid-season Start senescence/maturity End-season/harvesting

01 Mar 04 Apr 14 Jun 08 Aug 07 Sep

01 Mar 01 Apr 20 Jun 20 Aug 16 Sep

2012/2013

2012

h (m)

fc

0.7 0.7 1.9 2.0 2.0

0.01 0.05 0.30 0.30 0.30

2013

2012

0.01 0.05 0.25 0.25 0.25

MEI–DO Ribeira Sacra

2013

2012

2013

2012/2103

Date

h (m)

Date

h (m)

Date

h (m)

Date

h (m)

fc

01 Mar 04 Apr 09 Jun 18 Aug 28 Sep

0.5 0.5 1.2 1.4 1.4

01 Mar 18 Apr 21 Jun 31 Aug 04 Oct

0.5 0.5 1.0 1.2 1.2

01 Mar 01 Apr 07 Jun 18 Aug 31 Oct

0.5 0.5 1.3 1.5 1.5

01 Mar 08 Apr 19 Jun 31 Aug 10 Oct

0.5 0.5 1.2 1.4 1.4

0.01 0.05 0.25 0.25 0.25

LAD—Ladredo plot, MEI—Meixemán plot.

Table 3 Fraction and height of the cover crop for two experimental sites during the seasons. DO Valdeorras Date Year 2012 01 Mar 04 Apr 14 Apr* Year 2013 01 Mar 20 Apr*

MEI–DO Ribeira Sacra

fc cover

Density

hcover

Date

fc cover

Density

hcover

Date

fc cover

Density

hcover

0.8 0.7 0.5

0.6 0.5 0.5

0.05 0.05 0.05

01 Mar 01 Apr 29 Apr 10 May*

0.4 0.6 0.6 0.4

0.6 0.7 0.7 0.4

0.02 0.06 0.10 0.02

01 Mar 04 Apr 29 Apr 08 May*

0.4 0.6 0.6 0.4

0.6 0.7 0.7 0.4

0.02 0.03 0.07 0.02

0.8 0.5

0.6 0.50

0.10 0.05

01 Mar 22 Apr 03 May 12 May*

0.6 0.8 0.7 0.5

0.6 0.8 0.7 0.5

0.04 0.10 0.06 0.02

01 Mar 22 Apr 03 May 10 May*

0.7 0.8 0.7 0.6

0.7 0.8 0.7 0.6

0.04 0.10 0.06 0.02

From the latter date, cover crop acts like mulch. fc cover : fraction of the cover crop; hcover : height of cover crop (m). LAD—Ladredo plot, MEI—Meixemán plot.

A cover crop was maintained inter-rows consisting of natural herbaceous vegetation, but it turn to mulch from the last date include in Table 3. For its characterization, observations were performed during the crop season, from March to September, including its density, height (hcover ) and fraction of cover crop (fc cover ) (Table 3), these values was necessary to introduce in SIMDualKc model. In A Rúa, the irrigation treatment began the first of June and finished in the middle of August in 2012; during 2013, irrigation started on July and finished at the end of August. During these

6

10 5

15 20

4

25 30

3

35

2

40 1

45 50

Precipitation (mm)

1-3

0

31-3

30-4

30-5

29-6

29-7

28-8

27-9

0

27-10 8

5

7

10

6

15

5

20 4 25 3

30

2

35 40

1

45

0

Fig. 1. Total daily precipitation (

) and reference evapotranspiration (

30-4

30-5

29-6

29-7

28-8

27-9 7

5

-1

31-3

0

Reference Evapotranspiration (mm d )

7

5

Precipitation (mm)

1-3

27-9

6

10 5

15 20

4

25 3

30 35

2

40 1

45

a)

1-3 0

27-10 8

5

7

10 15

31-3

30-4

30-5

29-6

29-7

28-8

27-9

6 5

20 4 25 30 35

c)

0

50

3 2

40

1

45

0

b) -1

28-8

Precipitation (mm)

29-7

Precipitation (mm)

29-6

-1

30-5

Reference Evapotranspiration (mm d )

30-4

-1

31-3

0

Reference Evapotranspiration (mm d )

1-3

two seasons, water was applied early in the morning, for 59 and 46 days in 2012 and 2013, respectively, at a rate of 1.5 h per day, in order to reduce the evaporation losses. Irrigation depth was 1.14 mm and 1.54 mm, for DI and SDI, respectively, these differences were due to the spacing between drippers on the SDI system (1 m) and the spacing between vines and rows. The fraction of soil wetted by irrigation (fw ) was 0.01 and 0.1, for SDI and DI, respectively. In the case of the cv ‘Mencía’ plots only a rain-fed treatment (R) was considered in both seasons.

Reference Evapotranspiration (mm d )

*

LAD–DO Ribeira Sacra

d)

). DO Valdeorras: (a) 2012 and (b) 2013; and DO Ribeira Sacra: (c) 2012 and (d) 2013.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model AGWAT-4039; No. of Pages 12

ARTICLE IN PRESS J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

4

2.2. Wireless sensor network. Soil and plant measurements A wireless sensor network was implemented in all plots, based on the standard ZigBee/IEEE 802.15.4, using two gateway nodes, ˜ et al., 2012b). one per cultivar, to send the data to a server (Fandino The software sends alerts via SMS, to trigger the irrigation system when a given  soil threshold is reached. On the different plots, GMS (Watermark® , Irrometer, Riverside, CA) (Eldredge et al., 1993) were installed at 0.40 m depth in all treatments (two sensors per treatment, except in R by cv. ‘Godello’, with only one sensor), placed at the vine row at 0.10 m from the emitters and 0.40 m apart from the trunk, under the influence of the soil wetted volume. GMS was used to check the soil water status continuously and correlate this  soil with soil water content, obtained from discontinuous measurements, and finally to estimate soil–water characteristics curves. The soil water content was monitored with a TDR100 (Campbell Scientific), which operates in the field using the software PCTDR, with a flexible connector (Souto et al., 2008). Observations were performed at 0.60 m depth. The equation of Topp et al. (1980) relating the volumetric water content () with the measured bulk dielectric constant (εeff ) was used, since it has been proven successful in soils that do not contain substantial amounts of bound water, e.g., most sandy and loamy soils (Robinson et al., 2003), as the studied plots. A total of 10 and 11 measurements per year were taken in 2012 and 2013, respectively, i.e., about one measurement each three weeks. The number of measurement points per treatment was eight, corresponding to the number of replications of each treatment (R, DI and SDI), except for LAD plot, where only six measurement points/replications were performed. The measurement points were placed in the vine row, at 0.60 m from the emitters and 0.30 m apart from the trunk, allowing the tillage operations of the winegrower. These data were used for model calibration and validation, and to determine soil–water characteristic curves as explained above. Measurements of leaf water potential were carried out with a Scholander pressure chamber (PMS Model 600, Albany, OR, USA). Leaf sampling and measuring were done considering the precautions suggested by Martínez et al. (2013). Midday leaf ( m ) and stem ( stem ) water potentials were performed every two weeks, these measurements were carried out on healthy mature leaves from the middle third of the shoots, all of them at similar growth stages and with no alterations, and exposed to direct solar radiation. Twelve plants were measured per treatment in cv. ‘Godello’, however in cv. ‘Mencía’, only eight and six plants per treatment were measured in MEI and LAD, respectively. For determining the stem water potential, leaves were previously covered with foil ZipSeal bags 2 h prior to the measurement (Williams and Araujo, 2002). 2.3. Crop coefficients. SIMDualKc model Factors determining the Kc of vineyards include the crop growth stage, canopy height and architecture that is related to the trellis training system; the fraction of soil covered by the crop, presence of mulches or active ground cover, as well as the soil fraction wetted and exposed to radiation that is related to the irrigation method and soil management; crop management and local climate. Where automatic irrigation system using real-time data were applied, dual Kc approach should be selected for using Kc , better than the single crop coefficient approach, because the first one provides more precise and exhaustive results with daily data (Allen et al., 1998, 2007). The dual Kc approach consists of two main components, where the Kc value is divided into a basal crop coefficient, representing full transpiration Kcb , and a separate component, Ke , representing evaporation from the soil surface, hence Kc = Kcb + Ke . Thus, the actual

crop evapotranspiration, which is smaller than ETc when water stress occurs, is defined as: ETa = (Ks Kcb + Ke ) ETo

(1) [mm d−1 ],

where ETa is the actual crop evapotranspiration Kcb is the basal crop coefficient [–], Ks is the water stress coefficient [–], Ke is the soil evaporation coefficient [–] and ETo is the reference crop evapotranspiration [mm d−1 ]. In general, the product: Ks Kcb is used to express the basal crop coefficient adjusted to water stress (Kcb adj ), obtained when a constraint in soil water content exists. The basal Kcb , because it mostly represents transpiration, is correlated with the amount of vegetation and can be expressed in terms of a crop density coefficient, Kd (Allen et al., 2007; Allen and Pereira, 2009): Kcb = Kc

min

+ Kd (Kcb

full

− Kc

(2)

min )

where impacts of plant density and/or leaf area are represented by the density coefficient, Kd , Kcb full is the estimated basal Kcb for peak plant growth conditions having nearly full ground cover (or LAI > 3), and Kc min is the minimum Kc for bare soil (in the absence of vegetation) whose value is about 0.15 under typical agricultural conditions and in the range 0.0–0.15 for native vegetation depending on rainfall frequency. Kd was considered, which allows adjusting Kcb to the vegetation and ground cover conditions (Allen et al., 2007; Allen and Pereira, 2009). Kd is estimated as:



(1/(1+h)) eff , fc eff

Kd = min 1, ML fc



(3)

where fc eff is the effective fraction of ground covered or shaded by vegetation near solar noon [–], h is crop height [m] and the ML parameter [–] is a multiplier on fc eff , representing the ratio of ET per unit of horizontal vegetation surface to ETo over the same surface. fc eff differs from the fraction of soil surface covered by vegetation as observed overhead (fc ) due to the combined effect of the canopy shape, plant height and solar angle above the horizon on the shaded area. For crops planted in rows, additional information is necessary, fc eff is computed from the height to width ratio (HWR) of the crop row:



fc

eff



= fc 1 +

HWR

 

(4)

tan ˇ

with HWR =

hcanopy (cos ( )) Width

(5)

where  is the angle of the plant row from the east–west direction [rad], hcanopy is crop height [m], Width is the width of the crop row as viewed from an east–west direction [m], and ˇ is the mean angle [rad] of the sun above the horizon during the period of maximum ET, generally between 11.00 and 15.00 h (Allen et al., 1998). When the cover crop competes with the crop for the available soil water and contributes to the total evapotranspiration of the canopy, with an important reduction of evaporation of soil, the following approach (Allen et al., 2007; Allen and Pereira, 2009) is adopted for estimating the combined Kcb : Kcb = Kcb

cover





+ Kd max Kcb

full

− Kcb

cover ,

Kcb

full

− Kcb 2

cover

 (6)

where Kcb cover is the Kcb of the cover crop in the absence of tree foliage [–], Kd is the density factor [–] referring to the crop (Eq. (3)), and Kcb full is the basal Kcb anticipated for the crop under full cover conditions and corrected for climate. The second term of the max

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model AGWAT-4039; No. of Pages 12

ARTICLE IN PRESS J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

function reduces the estimated Kcb by half the difference between Kcb full and Kcb cover when this difference is negative, i.e., it takes into account the effects of shading of the cover crop by the taller plants having Kcb that is lower than that of the ground cover due to differences in stomatal conductance. The SIMDualKc model (Rosa et al., 2012a,b) uses the dual crop coefficient approach to calculate crop evapotranspiration (ETc ), considering separately the soil evaporation and crop transpiration components. This allows for a more precise analysis on how water from rainfall and irrigation is used by the crop. In addition, the model estimates deep percolation and capillary rise using parametric equations developed by Liu et al. (2006), and surface runoff using the curve number method (Rosa et al., 2012a). SIMDualKc includes the adjustment of Kcb for full cover fruit trees, vines and shrubs (Kcb full ) to actual crop density, as well as a procedure for estimating Kcb for the cover crop (Kcb cover ) as a function of its development and management, to be further combined with the Kcb of the crop (Rosa et al., 2012a). Finally the model allows to deduce an actual crop coefficient (Kc act ) adjusted to crop, cover crop, density and local conditions. The model was calibrated and validated previously for different crops and local conditions (Pac¸o et al., 2012; Rosa et al., 2012b; Martins et al., 2013; Zhang et al., 2013a; Zhao et al., 2013; Paredes ˜ vineyards with an active groundet al., 2014) including Albarino ˜ et al., 2012a). cover crop (Fandino The inputs necessary to run SIMDualKc model were:

• climatic data (ETo , rainfall, minimum relative humidity and wind speed), • soil, crop and cover crop data, explained above (Tables 1–3), • deep percolation (W) was evaluated using the parametric function defined by Liu et al. (2006), adjusting the parameters aD and bD , referred to root depth (0.60 m), • curve number runoff data (CN) following Allen et al. (2007), and • dates and depths of applied irrigation.

2.4. Model calibration and validation The model was calibrated and validated by comparing observed and simulated soil water content relative to all field experiments. The simulations were performed using soil, crop, cover crop, irrigation, and weather data collected during the entire crop seasons. Other information needed for running the model was initially estimated or taken from standard values and later upgraded through model calibration. This was the case for the full basal crop coefficients (Kcb full ), the depletion fraction for no-stress (p) and the ML adjustment parameter (Allen et al., 1998, 2007) (Table 4). TEW, REW and Ze were also determined during the calibration process as referred before. The initial values for aD and bD , to use in the parametric decay equation, were taken from Liu et al. (2006). Groundwater contribution was not considered because a shallow water table was not present. The calibration procedure consisted of adjusting the nonobserved parameters (Kcb full , p, TEW, REW, and ML ), deep percolation parameters (aD and bD ), and CN of the runnof curve algorithm, to minimize differences between observed and simulated available soil water values relative to the entire root depth profile as described by Popova and Pereira (2011). A trial and error procedure was developed as described by Rosa et al. (2012b) until differences between simulated and observed soil water values were minimized and stabilized. To assess the accuracy and goodness of fit of model predictions, several approaches were used, which are described by Martins et al. (2013) and Rosa et al. (2012b):

5

• analyzing the graphical time dependent representations of simulated vs. observed soil water content values. This allows for a good perception of trends or bias in modeling, if they occur. • Performing a regression, forced to the origin, between observed and model-predicted soil water content values. When the regression coefficient (b) is close to 1.0, the covariance is close to the variance of the observed values, indicating that predicted and observed values are statistically similar. A coefficient of determination (r2 ) close to 1.0 indicates that most of the total variance of the observed values is explained by the model. • Indicators of residual estimation errors (Legates and McCabe, 1999) including the root mean square error (RMSE), which describes the variance of errors, and RRMSE, which represent RMSE as a percentage of total available water; the average absolute error (AAE), which characterizes the size of error estimates in alternative to RMSE. • The index of agreement (dIA ), developed by Willmott (1981), was used to represent the ratio between the mean square error and the ‘potential error’ (Moriasi et al., 2007). The modelling efficiency (EF), that is a normalized statistic developed by Nash and Sutcliffe (1970), was used to determine the relative magnitude of the residual variance compared to the measured data variance (Moriasi et al., 2007).

3. Results and discussion 3.1. Model calibration and validation In the case of cv. ‘Godello’, calibration was made for the SDI treatment in 2012; however in cv. ‘Mencía’, it was made with data from 2013 for both plots (LAD and MEI), the rest of data sets were used to validation process. Initial soil water conditions were adjusted for both years and locations, during the calibration process. Simulated vs. observed soil water contents measured with TDR for both cultivars are shown in Fig. 2. This figure shows that the simulated soil water content adequately followed the observed values along the entire crop seasons, despite differences in cultivar conditions. Moreover, soil water dynamics is well simulated and refer to a wide range of variation of soil water. Regressions forced to origin, for both cultivars, showed that SIMDualKc model slightly under-estimates soil water content (Fig. 3), more accentuated in cv. ‘Mencía’. Calibrated parameters shown in Table 4 are in agreement with those proposed by Allen et al. (1998, 2007) and Allen and Pereira (2009), cv. ‘Godello’ showed similar values of Kcb full in all stages, except to Kcb full ini due to the presence of an active cover crop during this stage; however, cv. ‘Mencía’ showed values slightly lower than those found in literature, caused by the lower height during the season (h < 1.5 m) and the physiological conditions of cv. ‘Mencía’. The low value for the parameter aD is due to the low root depth, the calibrated parameter bD was the initial value by soils with cv. ‘Mencía’, and slighltly lower to cv. ‘Godello’, these results are in agreement with soil properties, that indicated quickly drainage (Liu et al., 2006). Due to a higher number of rainfall events, the deep percolation component of the soil water balance was important in both seasons: 6–79 mm and 45–110 mm (2012–2013) for cv ‘Godello’ and cv ‘Mencía’, respectively. Estimation errors are low (Table 5), with RMSE ranging from 0.007 to 0.018 m3 m−3 , i.e., 0.7 to 1.8% of TAW. AAE ranges between 0.006 and 0.016 m3 m−3 , thus being smaller than 1.7% of TAW. EF is high, ranging 0.77–0.96, which indicates that the relative magnitude of the residual variance is comparable to the measured data variance (Moriasi et al., 2007). The Willmott index of agreement, dIA , was higher than 0.96, indicating that the mean square error is close to the potential error due to modelling. Results show that the SIMDualKc model adequately predicts the soil water content

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model

ARTICLE IN PRESS

AGWAT-4039; No. of Pages 12

J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

6

Table 4 Standard (initial) and calibrated basal crop coefficients, p depletion fractions, ML parameter, soil evaporation parameters (TEW, REW, Ze ), deep percolation parameters (aD and bD ), and CN of the runnof curve algorithm for vineyards ET simulation. Standard parametersa

Calibrated parameters cv. ‘Mencía’

cv. ‘Godello’

LAD

pini pmed pend Kcb full ini Kcb full mid Kcb full end ML TEW (%) REW (%) Ze (m) aD (mm) bD CN

0.45 0.45 0.45 0.20 0.80 0.60 1.5–2.0 – – 0.10 – −0.0173 72

MEI

2012

2013

2012

2013

2012

2013

0.61 0.56 0.59

0.63 0.56 0.60

0.62 0.57 0.62

0.63 0.58 0.63

0.63 0.58 0.63

0.64 0.57 0.63

0.30 0.80 0.60

0.20 0.75 0.60 1.5

34 10 0.15 170 −0.023 72

29 9 0.15 150 −0.017 72

27 9 0.15 130 −0.017 72

a From Allen et al. (1998, 2007), Allen and Pereira (2009), and Liu et al. (2006). TEW—total evaporable water, REW—readily evaporable water, Ze —depth of the evaporable layer, CN—curve number. LAD—Ladredo plot, MEI—Meixemán plot.

for the vineyard cv. ‘Mencía’ and ‘Godello’ under R and irrigated (DI and SDI) conditions. 3.2. Basal crop coefficients. Soil evaporation and vine transpiration The variations of different coefficients: Ke , Kcb full , Kc act are shown in Fig. 4. Values of irrigation (Fig. 4a.2–b.3) and rainfall are also presented in these figures. In both study years, there are a high number of rainfall events during the season, for this reason Ke are very near to Kc act , except for the mid-season because fc and no rainfall occurred. In general during 2012, Kc act was similar for the different treatments in both cultivars during the season, except for the mid-season stage, particularly in the R treatment of cv. ‘Godello’ (Fig. 4a.1). However, in the 2013 season, the Kc act was similar between treatments of cv ‘Godello’, in another hand differences were observed between locations to cv ‘Mencía’, the soil evaporation in MEI was lower than LAD during mid-season, due to mulch effect. These conditions support well the derivation of crop coefficients for the two cultivars (‘Godello’ and ‘Mencía’)

and irrigation treatments. Calibrated Kcb full mid values (Table 4) were very similar to those proposed by Allen and Pereira (2009) in ‘Godello’ and ‘Mencía’ (0.80–0.75). For both cultivars, Kcb vine mid (0.25) are slightly higher to values obtained for Yunusa et al. (2004) for a 30% ground cover ‘Sultana’ vineyard in Australia (about 0.17). Trambouze and Voltz (2001) obtained a slightly higher value (0.36) for ‘Shiraz’ vines in south France. However, other researchers reported higher values for Kcb mid in wine grapes, with fc varying from 30–50%: Intrigliolo et al. (2009) reported Kcb mid around 0.50 in a ‘Riesling’ vineyard in the state of New York (USA) and Zhang et al. (2011) obtained Kcb mid = 0.47 for a ‘Merlot’ vineyard in northwest China. The average Kcb vine end = 0.22, is much smaller than the values reported by Intrigliolo et al. (2009) and Zhang et al. (2011), who found Kcb end values of about 0.55 and 0.45, respectively. This value ˜ (0.14) is closer to the actual Kcb vine end value observed in ‘Albarino’ ˜ et al., 2012a). (Fandino Results in Table 6 show that, in vineyards of V. vinifera cv. ‘Godello’ and cv. ‘Mencía’, about 29–31% of the water was evaporated, which is slightly higher than the 20% value reported

Table 5 Goodness of fit indicators relative to the SIMDualKc model calibration and validations for the different treatments and cultivars (’Godello’ and ‘Mencía’) in the 2012 and 2013 growing seasons. Goodness of fit indicators

b

r2

RMSE (m3 m−3 )

RRMSE (%)

AAE (m3 m−3 )

EF

dIA

‘Godello’ Calibration 2012 Validation 2012

SDI

1.01

0.97

0.007

0.7

0.006

0.96

0.99

R DI R DI SDI

0.93 1.04 0.97 1.04 1.00 1.00

0.91 0.97 0.87 0.93 0.90 0.90

0.018 0.013 0.017 0.012 0.010 0.014

1.8 1.3 1.7 1.2 1.4 1.4

0.016 0.012 0.013 0.010 0.008 0.011

0.80 0.89 0.77 0.85 0.86 0.86

0.96 0.98 0.96 0.97 0.97 0.97

LAD MEI

0.94 0.92

0.96 0.96

0.012 0.013

1.2 1.3

0.009 0.011

0.93 0.93

0.98 0.98

LAD MEI

0.95 0.89 0.93

0.98 0.96 0.97

0.013 0.017 0.014

1.3 1.7 1.4

0.011 0.013 0.011

0.94 0.90 0.93

0.98 0.97 0.98

2013

All experiments ‘Mencía’ Calibration 2013 Validation 2012 All experiments

R—rainfed, DI—surface drip irrigation, SDI—subsurface drip irrigation. b: regression coefficient, r2 : determination coefficient, RMSE: root mean square error, RRMSE: RMSE as a percentage of total available water, AAE: average absolute error, EF: modeling efficiency, dIA : index of agreement. LAD—Ladredo plot, MEI—Meixemán plot.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model

ARTICLE IN PRESS

AGWAT-4039; No. of Pages 12

16-3

31-3

15-4

30-4

15-5

30-5 14-6 Date

29-6

14-7

29-7

13-8

0.35 θ FC

0.30 End

0.25 0.20 0.15

θp Start

0.10 θ WP

0.05 0.00 1-3

16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 Date

a.2)

29-7 13-8 28-8

0.35 θFC

0.30 0.25

End

θp

0.20 0.15

Start

0.10

θWP

0.05 0.00 1-3

16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8

a.3)

-3

0.25

Soil Water Content (m m )

0.30

3

Date

θ FC

0.20 0.15

θp

0.10 0.05

θ WP

0.00 1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 27-9 Date

-3 3

Soil Water Content (m m )

a.1)

28-8

0.10 0.05 1-3

0.35 0.30

0.15

0.05

θWP

0.00 1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 27-9 Date

a.5)

b.1)

θFC End

0.20

θp

0.15 0.10 0.00

θWP

Start

0.05

1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 Date

b.2)

0.35

θFC

0.30 End

0.25

θp

0.20 0.15 0.10 0.05

θWP

Start

0.00 1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 Date

b.3)

0.30 0.25

θFC

0.20 0.15

θp

0.10 0.05 0.00

θWP 1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 27-9 Date

b.4)

-3 3

θp

0.10

16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 Date

0.25

Soil Water Content (m m )

3

-3

Soil Water Content (m m )

θFC 0.20

θ WP

0.00

a.4)

0.25

θp

0.15

-3

0.00 1-3

θ FC

0.20

3

θ WP

0.05

Soil Water Content (m m )

0.10

7

0.25

Soil Water Content (m3 m-3)

0.15

0.30

-3

θp

0.20

0.35

3

-3 3 3

-3

Soil Water Content (m m )

θ FC

0.25

Soil Water Content (m m )

-3

0.30

Soil Water Content (m m )

0.35

3

Soil Water Content (m m )

J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

0.25

θFC

0.20 0.15

θp

0.10 0.05

θWP

0.00 1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 27-9 Date

b.5)

Fig. 2. Simulated soil water content curve and observed values ( ) relative to the: (a) Year 2012, (b) Year 2013; (1) Rainfed-R,(2), Drip irrigation-DI, (3) Subsurface drip irrigation-SDI treatments, refer to cv. ‘Godello’; (4) Ladredo plot and (5) Meixemán plot, refer to cv. ‘Mencía’. Curves  FC ,  WP and  p represent, respectively soil moisture at field capacity, wilting point, and when depletion equals the fraction p. Arrows show the beginning and end of irrigation. Vertical represents standard deviation.

˜ by Celette et al. (2010) and 8–17% value reported by Fandino et al. (2012a). Differently, E was substantially lower than the 48% reported by Yunusa et al. (1997). Transpiration was divided in two components, cover crop transpiration and vine transpiration (Tv ), where 62 to 53% of total ET consists of Tv , for cv. ‘Godello’ and cv. ‘Mencía’, respectively. These values are similar to those referred by Yunusa et al. (1997) (57%), but substancially higher than those reported by Celette et al. (2010) ˜ et al. (2012a). For all years, Tv from the SDI treatments and Fandino was higher than Tv for DI treatment, and the latter was higher than Tv of the R treatment. These observations may be explained by the

fact that subsurface drip irrigation systems reduced evaporation processes from soil, and for the higher depth applied during the irrigation season. For cv. ‘Godello’ in LAD plot, Tv was smaller than in MEI plot, due to crop heigh that was lower in LAD for both seasons, moreover a higher E was simulated in LAD. The average ET rates over the full crop seasons of 2012–2013 ranged from 1.4 to 1.9 mm d−1 , which are not far from the ET rates reported by Singleton and Maudsley (1996) (2.1 mm d−1 ), Yunusa et al. (1997) (2.2 mm d−1 ), Zhang et al. (2011) (1.3 mm d−1 ), ˜ et al. Trambouze and Voltz (2001) (2.2 mm d−1 ) and Fandino (2012a) (2.2 mm d−1 ).

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model

ARTICLE IN PRESS

AGWAT-4039; No. of Pages 12

J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

m -3)

m -3)

8

0.30

y = 0,9974x r 2 = 0,8999

0.25 0.20 0.15 0.10 0.05 0.05

a) 0.10

0.15

0.20

0.25

0.30

Observed soil water content (m 3 m -3)

Simulated soil water content (m

Simulated soil water content (m

3

3

0.30

0.25

y = 0,9279x r 2 = 0,9658

0.20

0.15

0.10

0.05 0.05

b) 0.10

0.15

0.20

0.25

0.30

Observed soil water content (m 3 m -3)

Fig. 3. Comparison between observed and simulated soil water content using all vineyard experiment data (2012–2013). (a) cv. ‘Godello’, (b) cv. ‘Mencía’.

Table 6 Crop evapotranspiration and respective components separating vine and cover crop transpiration (T), vine transpiration (Tv ) and soil evaporation (E) for all treatments and two seasons (2012–2013). Treatment ‘Godello’ 2012 R DI SDI 2013 R DI SDI ‘Mencía’ 2012 LAD MEI 2013 LAD MEI

ETa (mm)

ETo (mm)

E/ETa (%)

ET (mm d−1 )

Kc (aver.)

91.4 85.8 91.6

274.3 331.4 354.3

704.8

33.3 25.9 25.9

1.4 1.7 1.9

0.39 0.47 0.50

154.9 194.7 207.4

93.8 90.4 93.5

279.8 318.0 334.7

739.7

33.5 28.4 27.9

1.4 1.6 1.7

0.38 0.43 0.45

198.0 216.1

176.4 196.7

131.7 120.8

329.8 336.9

796.7 808.1

39.9 35.9

1.6 1.6

0.41 0.42

222.7 268.2

145.6 185.3

103.8 59.3

326.5 327.5

822.0 834.4

31.8 18.1

1.5 1.5

0.40 0.39

T (mm)

Tv (mm)

182.9 245.6 262.7

162.6 225.0 242.4

184.5 224.4 237.0

E (mm)

R—rainfed, DI—surface drip irrigation, SDI—subsurface drip irrigation, LAD—Ladredo, MEI—Meixemán. T: vine and cover crop transpiration, Tv : vine transpiration, E: evaporation from the soil, ETa : actual crop evapotranspiration, ETo : reference evapotranspiration, ET: daily evapotranspiration, Kc : crop coefficient.

Table 7 Determination coefficients (r2 ) of soil–water characteristic curves: DO Valdeorras and DO Ribeira Sacra (LAD—Ladredo, MEI—Meixemán). Treatment

R DI SDI

DO Valdeorras

0.943 0.935 0.898

no technical problems or a bad installation of the sensors were detected.

DO Ribeira Sacra LAD

MEI

0.807 – –

0.983 – –

R—rainfed, DI—surface drip irrigation, SDI—subsurface drip irrigation.

3.3. Soil–water characteristic curves Fig. 5 shows the soil–water characteristic curves for the studied soils. The determination coefficient (r2 ) was higher than 0.81 in all cases (Table 7). Centeno et al. (2010) reported a lower determination coefficient for the correlation between soil and plant measurements (r2 = 0.50) in a cv. ‘Tempranillo’ vineyard; in this case, the authors used  soil . The good agreement between the soil water content (TDR) and  soil (GMS) allowed to correlate both variables and provide alerts with the real status of soil water content. The wireless network sends an alert when there was a decrease of  soil of 0.02 MPa, moment in which the measurement of plant water status was performed. Although GMS provide reliable outputs within the range of −0.01 to −0.20 MPa, we included in our analysis the values for all soil matric potential sensors installed in the fields. In the rain-fed treatment, after a rainy period,  soil values were re-established to those observed in winter. Therefore,

3.4. Soil–plant water stress relations and threshold for automatic irrigation scheduling A good fit was obtained between soil water content () measured with TDR, and plant measures: leaf water potential midday ( m ) (Fig. 6) and stem ( stem ). Table 8 summarizes all values for treatments and cultivars, where determination coefficients were higher than 0.73 and 0.64, for m and stem , respectively in all cases. Similar results were obtained by Williams and Araujo (2002) showing an average value of r2 = 0.68 to m , and r2 = 0.63 to stem , using linear relations and neutron probe measurements in cv. ‘Chardonnay’. In cv Thompson seedless, Williams and Trout (2005) obtain good correlations between vine water status and SWC, both for the predawn leaf water potential (r2 = 0.52) and for stem (r2 = 0.90) and m (r2 = 0.94). van Leeuwen et al. (2009) showed that  stem = −0.6 MPa was a limit usually employed as indicative of no water deficit in vineyards. During 2012 and 2013 seasons, it has been observed in cv. ‘Godello’ that  stem = −0.6 MPa, was achieved linked to an average value of  soil = −0.1 MPa [−0.07 MPa (‘Godello’-DI), −0.11 MPa (‘Godello’R), −0.12 MPa (‘Godello’-SDI)]. This means that when  soil reaches −0.1 MPa, irrigation should automatically triggered, adjusting the watering schedule to the quantitative and qualitative objectives of the grower (Acevedo-Opazo et al., 2010). These averages

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model AGWAT-4039; No. of Pages 12

ARTICLE IN PRESS 1.0

50

40

0.8

40

0.6

30

0.6

30

0.4

20

0.4

20

0.2

10

0.2

10

1.0

50

0.8

1.0

50

b.1)

1.2

60

1.0

50 40

Irrigation and Precipitation (mm d )

60

0

0.0 18-4 4-5 20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9

0.8

40

0.8

0.6

30

0.6

30

0.4

20

0.4

20

0.2

10

0.2

10

1-3 17-3 2-4 18-4 4-5 20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9

0

a.2)

0.0

0

b.2)

1-3 17-3 2-4 18-4 4-5 20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9

1.2

60

1.0

50

40

0.8

40

30

0.6

30

0.4

20

0.4

20

0.2

10

0.2

10

1.2

60

1.0

50

0.8 0.6

0.0

1-3 17-3 2-4 18-4 4-5 20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9

-1

1-3 17-3 2-4

0

a.3)

0.0

0 1-3 17-3 2-4 18-4 4-5

-1

a.1)

Irrigation and Precipitation (mm d )

Ke ,Kcb full , Kc act Ke ,Kcb full , Kc act

0

1.2

0.0

b.3)

20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9

1.2

60

50

1.0

50

0.8

40

0.8

40

0.6

30

0.6

30

0.4

20

0.4

20

0.2

10

0.2

10

0

0.0

60

1.0

-1

1.2

1-3 17-3 2-4 18-4 4-5 20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9 25-9

a.4)

60

1.2

60

1.2

1.0

50

1.0

50

0.8

40

0.8

40

0.6

30

0.6

30

0.4

20

0.4

20

0.2

10

0.2

10

0

0.0

0.0 1-3 15-3 29-312-4 26-4 10-5 24-5 7-6 21-6 5-7 19-7 2-8 16-830-8 13-9 27-9

a.5)

b.4)

0 1-3 15-329-3 12-4 26-410-5 24-5 7-6 21-6 5-7 19-7 2-8 16-830-8 13-927-9

0 1-3 16-3 31-3 15-4 30-4 15-5 30-5 14-6 29-6 14-7 29-7 13-8 28-8 12-9 27-9

-1

0.0

Precipitation (mm d )

Ke ,Kcb full , Kc act

1-3 17-3 2-4 18-4 4-5 20-5 5-6 21-6 7-7 23-7 8-8 24-8 9-9

-1

60

60

Precipitation (mm d )

1.2 1.2

0.0

Ke ,Kcb full , Kc act

9

Precipitation (mm d )

Ke ,Kcb full , Kc act

J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

b.5)

Fig. 4. Variation of: Ke ( ), Kcb full ( ), Kc act ( ), irrigation ( ) and precipitation ( ); (a) Year 2012, (b) Year 2013; (1) Rainfed-R,(2), Drip irrigation-DI, (3) Subsurface drip irrigation-SDI treatments, referred to cv. ‘Godello’; (4) Ladredo plot and (5) Meixemán plot, refer to cv. ‘Mencía’.

values will depend on rainfall in the initial and rapid growth stages, and root distribution, vine vigour and yield (van Leeuwen et al., 2009). For the cv. ‘Mencia’, only  m was available, so we used its relationship with the  soil . In both seasons and plots, a vine water

potential below  m = −0.9 MPa, coincident with  soil = −0.1 MPa was observed. This  soil is linked to soil water content below  p (Fig. 2a.4–b.5), water content below which the crop has difficulties to obtain water.  p was derived from SIMDualKc model after calibration and validation.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model

ARTICLE IN PRESS

AGWAT-4039; No. of Pages 12

J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

10

0.50

Soil water potential (-MPa)

Soil water potential (-MPa)

0.25

y = 0.0013x-3.5943

0.20

r2 = 0.9342

0.15

0.10

0.05

0.00

y = 0.0145x -1.9102 r 2 = 0.9828

0.40

0.30

0.20

0.10

0.00 0.05

0.08

0.11

0.14

0.17

0.20

0.02

a)

Soil water content (m 3 m -3)

0.05

0.08

0.11

0.14

0.17

0.20

b)

Soil water content (m 3 m -3)

Fig. 5. Soil–water characteristics curves obtained with measures of soil water content (TDR) and soil water potential (GMS) in two sites. (a) cv. ‘Godello’, (b) cv. ‘Mencía’ (MEI).

1.6

1.6

Leaf water potential (-MPa)

1.2

1.4

Leaf water potential (-MPa)

y = 0.0422x -1.6467 r 2 = 0.8153

1.4

1.0 0.8 0.6 0.4 0.2

1.0 0.8 0.6 0.4 0.2 0.0 0.06

0.0 0.06

0.08

0.10

0.12

0.14

0.16 3

0.18

-3

Soil water content (m m )

y = 0.1052x -0.8684 r2 = 0.7791

1.2

0.20

a)

0.08

0.10

0.12

0.14

0.16

Soil water content (m 3 m -3)

0.18

0.20

b)

Fig. 6. Relationship between soil water content measured with TDR and leaf water potential (midday). (a) cv. ‘Godello’ (SDI-2012), (b) cv. ‘Mencía’ (LAD-2013).

Table 8 Determination coefficients (r2 ) of leaf water potential midday ( MEI—Meixemán). Treatment

m)

and stem (

stem )

and soil water content: DO Valdeorras and DO Ribeira Sacra (LAD—Ladredo,

DO Valdeorras

DO Ribeira Sacra LAD

2012 m

R DI SDI

0.902 0.736 0.815

2013 stem

0.902 0.880 0.647

m

0.793 0.751 0.815

2012 stem

0.848 0.824 0.823

MEI 2013

2012

2013

0.779 – –

0.885 – –

0.846 – –

m

0.938 – –

R—rainfed, DI—surface drip irrigation, SDI—subsurface drip irrigation.

4. Conclusions The dual crop coefficient approach allowed us to obtain a more precise division of evapotranspiration components in a vineyard, taking into account the fraction of soil covered by the crop, height, density and other components of soil water balance, usually not applied, i.e. deep percolation and runoff. For cv ‘Godello’, calibrated values using SIMDualKc model were Kcb full ini = 0.30, Kcb full mid = 0.80, Kcb full end = 0.60; and for cv. ‘Mencia’ Kcb full ini = 0.20, Kcb full mid = 0.75, Kcb full end = 0.60. These values showed a good fit with soil water measurements in situ. In both plots, deep percolation was an important parameter of soil water balance; for this reason, an automatic irrigation system that uses simulation data is a good tool to manage water consumption in real time. Simulation and soil–plant water measurements relations were determined, obtaining good results: SIMDualKc model and TDR (r2 > 0.74), TDR and LWP (r2 > 0.65), TDR and GMS (r2 > 0.81),

showing that continuous measures with GMS permit establishing a threshold value related with leaf water potential, for both cultivars (’Godello’ and ‘Mencia’); the limit was  soil = −0.1 MPa. The process applied was useful for water management in real-time in a vineyard; triggering the irrigation system when the threshold value was reached, to reduce vineyard water stress. Another physiological measures, [water stress integral (Myers, 1988)] and soil water analysis [fraction of transpirable soil waterFTSW(Pellegrino et al., 2004)], are needed to check the results obtained, by keeping soil water status within an optimal range for grapevine activity. Acknowledgements Authors thank the technical staff of the ‘Guímaro’ and ‘Germán Rodríguez Salgado’ wineries. Ministry of Science and Innovation (Spain) for their support to the field research with funds from project RTA-2011-00041-C02-02 (80% FEDER funds), and

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model AGWAT-4039; No. of Pages 12

ARTICLE IN PRESS J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

NaanDanJain Iberica for providing irrigation pipes and complements, are acknowledged.

References Acevedo-Opazo, C., Ortega-Farias, S., Fuentes, S., 2010. Effects of grapevine (Vitis vinifera L.) water status on water consumption, vegetative growth and grape quality: an irrigation scheduling application to achieve regulated deficit irrigation. Agric. Water Manage. 97 (7), 956–964. Allen, R.G., Pereira, L.S., 2009. Estimating crop coefficients from fraction of ground cover and height. Irrig. Sci. 28, 17–34. Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration: Guidelines for Computing Crop Water Requirements, Irrigation and Drainage Paper 56. United Nations FAO, Rome. Allen, R.G., Pereira, L.S., Smith, M., Raes, D., Wright, J.L., 2005. FAO-56 dual crop coefficient method for estimating evaporation from soil and application extensions. J. Irrig. Drain. Eng. 131, 2–13. Allen, R.G., Wright, J.L., Pruitt, W.O., Pereira, L.S., 2007. Water requirements. In: Hoffman, G.J., Evans, R.G., Jensen, M.E., Martin, D.L., Elliot, R.L. (Eds.), Design and Operation of Farm Irrigation Systems. , second ed. ASABE Monograph, St. Joseph, MI, pp. 208–288. Azevedo, P.V., Soares, J.M., Silva, V., Silva, B.B., Nascimento, T., 2008. Evapotranspiration of “Superior” grapevines under intermittent irrigation. Agric. Water Manage. 95, 301–308. Celette, F., Ripoche, A., Gary, C., 2010. WaLIS—a simple model to simulate water partitioning in a crop association: the example of an intercropped vineyard. Agric. Water Manage. 97, 1749–1759. Centeno, A., Baeza, P., Lissarrague, J.R., 2010. Relationship between soil and plant water status in wine grapes under various water deficit regimes. HortTechnology 20, 585–593. Cifre, J., Bota, J., Escalona, J.M., Medrano, H., Flexas, J., 2005. Physiological tools for irrigation scheduling in grapevines (Vitis vinifera L.): an open gate to improve water-use efficiency? Agric. Ecosyst. Environ. 106, 159–170. Eldredge, E.P., Shock, C.C., Stieber, T.D., 1993. Calibration of granular matrix sensors for irrigation management. Agron. J. 85 (6), 1228–1232. FAO, 2013. AQUASTAT Database. Food and Agriculture Organization of the United Nations, http://www.fao.org/nr/water/aquastat/data/query/results.html (accessed 30 December 2013). ˜ M., Cancela, J.J., Rey, B.J., Martínez, E.M., Rosa, R.G., Pereira, L.S., 2012a. Using Fandino, ˜ vineyards (Norththe Dual-kc approach to model evapotranspiration of albarino west Spain) with consideration of active ground cover. Agric. Water Manage. 112, 75–87. ˜ M., Martínez, E.M., Rey, B.J., Cancela, J.J., 2012b. Plant water status in vineFandino, yards combining sensors in soil and plant. In: Proceedings 15th Inter Regional Enviro-Water Conference, Valencia, Spain. ˜ M., Martínez, E.M., Trigo-Córdoba, E., Bouzas-Cid, Y., Díaz-Losada, E., MirásFandino, Avalos, J.M., Rey, B.J., Cancela, J.J., 2013. Irrigation effects on the ecophysiological response of Godello cultivar from Valdeorras D.O. (NW Spain) in 2012. In: Ciência e Técnica Vitivinícola–Volume 28, Proceedings 18th International Symposium GiESCO, 7–11 July, 2013, Porto, pp. 53–57. Fooladmand, H.R., Sepaskhah, A.R., 2009. A soil water balance model for a rain-fed vineyard in a micro catchement based on dual crop coefficient. Arch. Agron. Soil Sci. 55, 67–77. Girona, J., Mata, M., del Campo, J., Arbonés, A., Bartra, E., Marsal, J., 2006. The use of midday leaf water potential for scheduling deficit irrigation in vineyards. Irrig. Sci. 24, 115–127. Gouveia, J., Lopes, C.M., Pedroso, V., Martins, S., Rodrigues, P., Alves, I., 2012. Effect of irrigation on soil water depletion, vegetative growth, yield and berry composition of the grapevine variety Touriga Nacional. Ciência Téc. Vitiv. 27 (2), 115–122. Hayashi, A., Akimoto, K., Tomoda, T., Kii, M., 2012. Global evaluation of the effects of agriculture and water management adaptations on the water-stressed population. Mitig. Adapt. Strateg. Glob. Change 18 (5), 591–618. Intrigliolo, D.S., Lakso, A.N., Piccioni, R.M., 2009. Grapevine cv. ‘Riesling’ water use in the northeastern United States. Irrig. Sci. 27, 253–262. Legates, D., McCabe, G., 1999. Evaluating the use of goodness of fit measures in hydrologic and hydroclimatic model validation. Water Resour. Res. 35 (1), 233–241. Liu, Y., Pereira, L.S., Fernando, R.M., 2006. Fluxes through the bottom boundary of the root zone in silty soils: parametric approaches to estimate groundwater contribution and percolation. Agric. Water Manage. 84, 27–40. López, J.A., García-Sánchez, A.J., Soto, F., Iborra, A., García-Sánchez, F., García-Haro, J., 2011. Design and validation of a wireless sensor network architecture for precision horticulture applications. Precis. Agric. 12 (2), 280–295. Martínez, E.M., Cancela, J.J., Cuesta, T.S., Neira, X.X., 2011. Review. Use of psychrometers in field measurements of plant material: accuracy and handling difficulties. Span. J. Agric. Res. 9 (1), 313–328. Martínez, E.M., Cuesta, T., Cancela, J.J., 2012. The efficiency of different estimation methods of hydro-physical limits. Rev. Bras. Cienc. Solo 36 (6), 1756–1768. ˜ M., Cancela, J.J., 2013. Comparison of two techMartínez, E.M., Rey, B.J., Fandino, ˜ Ciência niques for measuring leaf water potential in Vitis vinifera var. Albarino. Téc. Vitiv. 28 (1), 29–41. Martins, J.D., Rodrigues, G.C., Paredes, P., Carlesso, R., Oliveira, Z.B., Knies, A.E., Petry, M.T., Pereira, L.S., 2013. Dual crop coefficients for maize in southern Brazil:

11

model testing for sprinkler and drip irrigation and mulched soil. Biosyst. Eng. 115 (3), 291–310. Moratiel, R., Martínez-Cob, A., 2011. Evapotranspiration of grapevine trained to a gable trellis system under netting and black plastic mulching. Irrig. Sci. 30 (3), 167–178. Moriasi, D.N., Arnold, J.G., Van Liew, M.W., Bingner, R.L., Harmel, R.D., Veith, T.L., 2007. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans. ASABE 50 (3), 885–900. Myers, B.J., 1988. Water stress integral—a link between short-term stress and longterm growth. Tree Physiol. 4, 315–323. Nash, J.E., Sutcliffe, J.V., 1970. River flow forecasting through conceptual models: Part 1. A discussion of principles. J. Hydrol. 10 (3), 282–290. Ortega-Farias, S., Poblete-Echeverría, C., Brisson, N., 2010. Parameterization of a twolayer model for estimating vineyard evapotranspiration using meteorological measurements. Agric. For. Meteorol. 150, 276–286. Pac¸o, T.A., Ferreira, M.I., Rosa, R.D., Paredes, P., Rodrigues, G.C., Conceic¸ão, N., Pacheco, C.A., Pereira, L.S., 2012. The dual crop coefficient approach using a density factor to simulate the evapotranspiration of a peach orchard: SIMDualKc model versus eddy covariance measurements. Irrig. Sci. 30 (2), 115–126. Paredes, P., Rodrigues, G.C., Alves, I., Pereira, L.S., 2014. Partitioning evapotranspiration, yield prediction and economic returns of maize under various irrigation management strategies. Agric. Water Manage. 135, 27–39. Pellegrino, A., Lebon, E., Voltz, M., Wery, J., 2004. Relationships between plant and soil water status in vine (Vitis vinifera L.). Plant Soil 266, 129–142. Pereira, L.S., 1999. Higher performance through combined improvements in irrigation methods and scheduling: a discussion. Agric. Water Manage. 40 (2), 153–169. Poblete-Echeverría, C.A., Ortega-Farias, S.O., 2013. Evaluation of single and dual crop coefficients over a drip irrigated Merlot vineyard (Vitis vinifera L.) using combined measurements of sap flow sensors and an eddy covariance system. Aust. J. Grape Wine Res. 19 (2), 249–260. Popova, Z., Pereira, L.S., 2011. Modelling for maize irrigation scheduling using long term experimental data from Plovdiv region, Bulgaria. Agric. Water Manage. 98, 675–683. Robinson, D.A., Jones, S.B., Wraith, J.M., Or, D., Friedman, S.P., 2003. A review of advances in dielectric and electrical conductivity measurement in soils using time domain reflectometry. Vadose Zone J. 2, 444–475. Rosa, R.D., Paredes, P., Rodrigues, G.C., Alves, I., Fernando, R.M., Pereira, L.S., Allen, R.G., 2012a. Implementing the dual crop coefficient approach in interactive software. 1. Background and computational strategy. Agric. Water Manage. 103, 8–24. Rosa, R.D., Paredes, P., Rodrigues, G.C., Alves, I., Fernando, R.M., Pereira, L.S., Allen, R.G., 2012b. Implementing the dual crop coefficient approach in interactive software. 2. Model testing. Agric. Water Manage. 103, 62–77. Ruíz-García, L., Lunadei, L., Barreiro, P., Robla, J.I., 2009. Review of wireless sensor technologies and applications in agriculture and food industry: State of the art and current trends. Sensors 9, 4728–4750. Santesteban, L.G., Miranda, C., Royo, J.B., 2011. Regulated deficit irrigation effects on growth, yield, grape quality and individual anthocyanin composition in Vitis vinifera L. cv. ‘Tempranillo’. Agric. Water Manage. 98, 1171–1179. Scholander, P.F., Hammel, H.J., Bradstreet, A., Hwemmingsen, E.A., 1965. Sap pressure in vascular plants. Science 148, 339–346. Singleton, P.L., Maudsley, D., 1996. Pattern of water extraction by grapevines on two soils in the Waikato, New Zealand. New Zeal. J. Crop Hort. 24, 415–424. Smarsly, K., 2013. Agricultural ecosystem monitoring based on autonomous sensor systems. In: Second International Conference on Agro-Geoinformatics, pp. 402–407 (IEEE). Souto, F.J., Dafonte, J., Escariz, M., 2008. Design and air-water calibration of a waveguide connector for TDR measurements of soil electric permittivity in stony soils. Biosyst. Eng. 101 (4), 463–471. Topp, G.C., Davis, J.L., Annan, A.P., 1980. Electromagnetic determination of soil water content: measurement in coaxial transmission lines. Water Resour. Res. 16, 574–582. Trambouze, W., Voltz, M., 2001. Measurement and modelling of the transpiration of a Mediterranean vineyard. Agric. For. Meteorol. 207, 153–166. Tramontini, S., Van Leeuwen, C., Domec, J.C., Destrac-Irvine, A., Basteau, C., Vitali, M., Mosbach-Schulz, O., Lovisolo, C., 2013. Impact of soil texture and water availability on the hydraulic control of plant and grape-berry development. Plant Soil 368 (1–2), 215–230. ˜ M., Bouzas-Cid, Y., Cancela, J.J., Rey, B.J., Martínez, E.M., Trigo-Córdoba, E., Fandino, Díaz-Losada, E., Mirás-Avalos, J.M., 2013. Irrigation effects on the agronomic and enological performance of Godello cultivar from Valdeorras D.O. (NW Spain) in 2012. In: Ciência e Técnica Vitivinícola—Volume 28, Proceedings 18th International Symposium GiESCO, 7–11 July, 2012, Porto, pp. 99–103. van Leeuwen, C., Tregoat, O., Choné, X., Bois, B., Pernet, D., Gaudillère, J.P., 2009. Vine water status is a key factor in grape ripening and vintage quality for red Bordeaux vine. How can it be assessed for vineyard management purposes? J. Int. Sci. Vigne Vin. 43, 121–134. Williams, L.E., Araujo, F.J., 2002. Correlations among predawn leaf, midday leaf, and midday stem water potential and their correlations with other measures of soil and plant water status in Vitis vinifera. J. Am. Hortic. Sci. 127 (3), 448–454. Williams, L.E., Trout, T.J., 2005. Relationships among vine-and soil-based of water status in Thompson seedless vineyard in response to high-frequency drip irrigation. Am. J. Enol. Vitic. 56 (4), 357–366. Willmott, C.J., 1981. On the validation of models. Phys. Geog. 870 (2), 184–194.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020

G Model AGWAT-4039; No. of Pages 12 12

ARTICLE IN PRESS J.J. Cancela et al. / Agricultural Water Management xxx (2014) xxx–xxx

Yunusa, I.A.M., Walker, R.R., Guy, I.R., 1997. Partitioning of seasonal evapotranspiration from a commercial furrow-irrigated Sultana vineyard. Irrig. Sci. 18, 45–54. Yunusa, I., Walker, R.R., Lu, P., 2004. Evapotranspiration components from energy balance, sapflow and microlysimetry techniques for an irrigated vineyard in inland Australia. Agric. For. Meteorol. 127, 93–107. Zhang, B., Kang, S., Li, F., Tong, L., Du, T., 2010. Variation in vineyard evapotranspiration in an arid region of northwest China. Agric. Water Manage. 97, 1898–1904. Zhang, B., Liu, Y., Xu, D., Zhao, N., Lei, B., Rosa, R.D., Paredes, P., Pac¸o, T., Pereira, L.S., 2013a. The dual crop coefficient approach to estimate and partitioning evapotranspiration of the winter wheat–summer maize crop sequence in North China Plain. Irrig. Sci. 31 (6), 1303–1316.

Zhang, X., Zou, H., Zhang, N., Li, Y., Yang, Y., 2013b. The research and applications of agricultural automation based on Internet of things. Inf. Sci. Manage. Eng. (Set) 46, 111–119. Zhang, Y., Kang, S., Ward, E.J., Ding, R., Zhang, X., Zheng, R., 2011. Evapotranspiration components determined by sap flow and microlysimetry techniques of a vineyard in northwest China: dynamics and influential factors. Agric. Water Manage. 98, 1207–1214. Zhao, N.N., Liu, Y., Cai, J.B., Rosa, R., Paredes, P., Pereira, L.S., 2013. Dual crop coefficient modelling applied to the winter wheat-summer maize crop sequence in North China Plain: basal crop coefficients and soil evaporation component. Agric. Water Manage. 117, 93–105.

Please cite this article in press as: Cancela, J.J., et al., Automatic irrigation system based on dual crop coefficient, soil and plant water status for Vitis vinifera (cv Godello and cv Mencía). Agric. Water Manage. (2014), http://dx.doi.org/10.1016/j.agwat.2014.10.020