Thermal Imaging at Plant Level to Assess the Crop-water Status in Almond Trees (cv. Guara) Under Deficit Irrigation Strategies

1 Thermal imaging at plant level to assess the crop-water status in almond trees (cv. 1 Guara) under deficit irrigation strategies 2 3 García-Tejero I.F.1*, Rubio, A.E2., Viñuela, I1., Hernández, A1., Gutiérrez-Gordillo, S1., Rodríguez-Pleguezuelo, C.R.3, 4 Durán-Zuazo V.H.3 5 1 Instituto Andaluz de Investigación y Formación Agraria, Pesquera y de la Producción Ecológica (IFAPA). Centro “Las 6 Torres – Tomejil”. Ctra. Sevilla-Cazalla Km. 12,2. 41.200. Alcalá del Río, Sevilla, Spain. 7 2 Facultad de Biología. Departamento de Biología Vegetal y Ecología. Universidad de Sevilla. Avenida de Reina 8 Mercedes s/n. 41012. Sevilla, Spain. 9 3 Instituto Andaluz de Investigación y Formación Agraria, Pesquera y de la Producción Ecológica (IFAPA). Centro 10 “Camino de Purchil”. Apdo. 2027, 18080, Granada, Spain 11 12 *e-mail: [email protected] 13 Abstract 14 Almond (Prunnus dulcis Mill.) has been traditionally associated to marginal land cultivation and rain-fed 15 agriculture in South Spain. However, in the last years, this crop is being progressively introduced in more 16 productive agricultural areas within the Guadalquivir river basin, where the available water resources are 17 not enough to satisfy the adequate crop-water requirements. Considering this limitation, a more precise 18 irrigation scheduling to maximize the yield is required. Infrared thermal imaging emerges as alternative to 19 other traditional methodologies to assess the crop-water status, especially when deficit irrigation (DI) 20 strategies are being applied. The aim of this study was to define the methodology to assess the almond 21 water status by means of thermal information. The trial was conducted during 2014, during the kernel-filling 22 period, in an almond experimental orchard (SW Spain), with 5-year-old trees, subjected to three irrigation 23 regimes: i) a full-irrigation treatment (C-100), which received 100% of ETC; ii) a regulated deficit irrigation 24 (RDI-50), which received 100% of ETC except during the kernel filling period, when this treatment was 25 irrigated with 50% of ETC; iii) and a low-frequency deficit irrigation treatment (LFDI), which received 100% 26 of ETC except during the kernel filling period, when it was subjected to continuous periods of irrigation- 27 restriction, defined in terms of the threshold values of shaded leaf water potential (Ψleaf). Three daily curves 28 of canopy temperature (TC), stomatal conductance to water vapour (gs) and Ψleaf with measurements at 8:00, 29 11:00, 14:00, 17:00 and 20:00 were developed. Additionally, Crop Water Stress Index (CWSI), temperature 30 difference between canopy and the surrounding air (ΔTcanopy-air), and the relative index to stomatal 31 conductance (IG) obtained at different scales (canopy and row) were estimated. Significant correlations of 32 infrared thermal information vs. Ψleaf and gs were obtained (p ≤ 0.05 and p ≤ 0.01), in particular, by using 33 the thermal readings taken at 11:30, 14:30 at 17:30 h, especially robust were the relationships obtained 34 between TC and CWSI with Ψleaf at 11:30 h; and between TC and CWSI with gs, and Ψleaf at 14:30 h. Finally, 35 2 considering the infrared thermal monitoring procedure (readings at tree and row level), similar values of TC 36 were obtained, and therefore, the images taken at row level offered a better information with a higher 37 feasibility in terms of image processing. 38 Keywords: Thermography, thermal indexes, water stress, leaf gas exchange and leaf water potential. 39 1.- Introduction 40 Irrigated agriculture in the South of Europe, and more concretely in semi-arid areas such as Andalusia (S 41 Spain), is crucial for their development, especially in those rural regions with a lower economic potential. In 42 this line, for the case of Andalusia, irrigated agriculture generates more than 60% of rural employments, and 43 represents 64% of agricultural production. Currently, 1,176,000 ha are devoted to irrigated agriculture, 44 corresponding to 24% of total Andalusian agricultural surface, and this being 33% of the irrigated agriculture 45 in Spain (ARA, 2011). 46 Climatic conditions in this area are characterized by the scarcity and irregularity of rainfall, coinciding the 47 dry period with the season of highest evapotranspiration. Moreover, the last forecast predictions argue 48 significant water resources depletions; with an important declining in the soil water reserves, more accused 49 periods of rainfall restrictions and increasing in the average temperatures (IPCC, 2014). In this agreement, 50 it is expected that this situation promotes an imbalance between the irrigation demand and the available 51 water resources in the Mediterranean agriculture (Daccache et al. 2012, Olesen et al. 2011). This fact will 52 suppose an important constraint for the competitiveness between agriculture and other more productive 53 sectors such as the industry or tourism. In addition, the introduction of alternative crops in order to maximize 54 the profitability of agroecosystems will be required, together with different strategies to improve the 55 agricultural water management (García-Tejero et al. 2014a). 56 In this context, almond (Prunus dulcis Mill.) is the third crop in terms of surface in Spain, representing globally 57 almost 40%, and 84% within the EU. However, only 5% of the global production is developed in Spain 58 (FAOSTAT, 2016). Concretely, the surface of almond in Andalusia is about 152,000 ha, and within them, 59 95% are associated to marginal and rain-fed agriculture because of the climate limitations, where annual 60 rainfalls does not exceed of 300 mm with low nut yields (CAPDR, 2016). However, in the last few years, the 61 agricultural surface devoted to almond crop has significant increased, specially, in areas where this crop 62 was not traditionally cultivated, these new orchards being cultivated under intensive and irrigation practices. 63 Thus, almond can be found under very different agricultural systems from the most marginal situations to 64 the most intensive orchards, which promotes a wide range of yields (from 150 to 2,600 kg ha-1) (CAPDR, 65 2016). 66 According to Goldhamer and Fereres (2016), irrigation is the most limiting factor for this crop, with crop 67 water-requirements oscillating between 900 and 1,350 mm (Goldhamer and Girona, 2012). In this 68 agreement, Goldhamer and Fereres (2016) reported values close to 4,000 kg ha-1 (depending on the 69 cultivar) for irrigation doses around 1,250 mm, with yield reductions close to 14% when the irrigation doses 70 3 were close to 1,000 mm. More recently, López-López et al. (2018) in a long-term experience developed in 71 the province of Córdoba (Andalusia, South Spain), reported maximum yield values ( 2,500 kg ha-1) in 72 mature almond trees (cv. Guara), when these trees were irrigated receiving the maximum crop water 73 requirements (close to 10,000 m3 ha-1). 74 In spite of this, almond is considered a drought-resistant crop because of its xeromorphic properties 75 (Torrecillas et al. 1996), and many authors have reported different results related to the effects of deficit 76 irrigation (DI) strategies (Puerto et al., 2013; Phogat et al., 2013; 2018; Spinelli et al., 2016; among others). 77 More recently, López-López et al. (2018) discussed the effects of water deficits in almond trees in terms of 78 water use, evaluating different deficit irrigation (DI) strategies during three consecutive years. These authors 79 found that almond trees under different moderate DI strategies were able of keeping canopy volumes similar 80 to those trees that were fully irrigated, these being directly related with the almond capability to obtain yield 81 values under moderate deficit irrigation similar to those reported by fully irrigated trees; this fact being 82 accompanied with similar soil water depletions and transpiration level. 83 Taking into account the maximum crop-water demand, the water scarcity in semi-arid areas, and the proper 84 response of this crop to moderate water stress, DI would be a suitable alternative to reach equilibrium 85 between the available water resources and a proper crop development with final yields able to ensure the 86 competitiveness and feasibility of this crop (García-Tejero et al., 2016a). However, the application of DI 87 strategies requires a proper knowledge about the crop physiological status, with the aim of ensuring the 88 correct crop development without significant compromising the yield and fruit-quality, especially when water- 89 stress is applied in different crop stages (Spinelli et al., 2016). In this sense, according to Puerto et al. (2013), 90 when a DI strategy is applied in fruit trees, this is mainly developed supplying a specific water withholding, 91 taken as reference the crop water requirements by means of the crop evapotranspiration (ETC), without 92 taking into account the effects of canopy architecture, the degree of canopy cover or the soil management 93 (among others); or without considering the crop physiological status when this water stress is applied. In 94 this regard, the most proper irrigation scheduling should consider the whole of soil-plant-atmosphere system; 95 although in terms or representativeness, the live component (plant) would be offering the most valuable 96 information, inasmuch as this reflects the most integrative information, mainly in terms of final yield. 97 Traditionally, crop water monitoring has been developed by using punctual measurements of stem (Ψstem) 98 or leaf (Ψleaf) water potential at midday or pre-dawn (Ψpd) (Shackel, 2011; Nortes et al., 2005) or monitoring 99 the gas-exchange parameters such as transpiration (E), stomatal conductance (gs) or net photosynthetic 100 rate (A) (Gomes-Laranjo et al., 2006). 101 According to Remorini and Massai (2003), Stem is not only a proper indicator of plant-water status as well 102 as the crop productivity. In the same vein, Mirás-Avalos et al. (2016) reported that water potential is a 103 suitable indicator of almond water status, although its usefulness is reduced, because of a minimum number 104 of replications are required, and the representativeness in the whole plant is reduced. 105 In the last years, the use of remote sensing in agriculture, and more concretely, infrared thermal imaging to 106 monitor the crop water status has been progressively introduced (Costa et al., 2013). This technique has 107 4 been properly described as a good methodology for crop-water monitoring in different woody crops such as 108 citrus (García-Tejero et al., 2011; González-Dugo et al., 2014); young almonds (García-Tejero et al., 2012), 109 vines (García-Tejero et al., 2016b) or olives (Egea et al., 2017). This technique is based on the leaf energy 110 balance. When a water stress situation is applied, plants responds with a partial stomatal closure, reducing 111 the stomatal conductance, limiting the leaf transpiration and promoting an attenuation of the evaporative 112 cooling process, resulting in higher leaf / canopy temperature values (Jones, 1999; 2004). 113 This technique can be applied at different monitoring scales, from “leaf or canopy” to “orchard or basin” level 114 (Poblete-Echeverría et al., 2014; 2016). The selection of the most proper methodology will be related with 115 the desired goal and the economic availability (Costa et al., 2013). In this sense, the use of thermography 116 at orchard scale by using satellites images, allows to take decisions related to crop variability or irrigation 117 scheduling, but some constraints must be taken into account. On one hand, thermal images taking by 118 satellites have the difficulty of depending of the moment in which the satellite passes above the orchard; 119 and on the other hand, the spatial and spectral resolution is not proper. These constraints could be solved 120 by using of unmanned aerial vehicles (UAVs), despite its economically restrictions. In this sense, the use of 121 thermal images at orchard scale, taken by means of UAVs, requires having the proper technology; and this 122 fact can increase the cost of this tool, becoming less accessible the use of this technology. By the contrast, 123 these sensors can be used at plant level, with thermal cameras much more profitable, easing the 124 accessibility to this technique by the irrigation communities or technicians. 125 Likewise, the main constraints of this technique are focused in the image processing (many times requiring 126 high time consuming), and the correct interpretation of the infrared thermal information (García-Tejero et al., 127 2015a). Because of this, many times different relationships between infrared thermal information and other 128 physiological parameters such as gs, A, E, or Ψstem are required (Jones 2004; Jones et al., 2009), although 129 these relationships are not always enough robust because of the high dependence of the meteorological 130 conditions (Jones, 1999; 2004), the monitoring proceedings (Costa el al., 2013), the cultivar (Costa et al., 131 2012; García-Tejero et al., 2016b) or even, the crop phenological stage (Cohen et al., 2015). 132 Up to day, several authors have developed strategies to optimize this technique, developing different 133 protocols and strategies to take thermal readings under field conditions (Jones et al., 2009; Pou et al., 2014; 134 Poblete-Echeverría et al., 2014, 2016, García-Tejero et al., 2012, 2016b) and describing different 135 relationships between infrared thermal information and physiological parameters. 136 We hypothesize that thermography could be a suitable technique to monitor almond water status, especially 137 when this is subjected to DI programs. The aim of the present work was to evaluate the performance of 138 thermography under field conditions at two monitoring levels (plant and row) to assess the crop water status 139 in almond trees (cv. Guara), determining the best moment of the day to obtain the thermal information and 140 the most robust thermal index to interpret properly the crop-water status. 141 142 143 144 5 2. Material and methods 145 2.1. Experimental site 146 The trial was conducted during 2014 in an experimental orchard of almonds (Prunus dulcis Mill. D.A. Webb 147 cv. Guara, grafted onto GF677), located in the Guadalquivir river basin (37º 30’ 47’’ N; 5º 58’ 2’’ O) (Seville, 148 SW Spain). Planted in 2009, the trees were spaced 6 x 7 m, and drip irrigated using two pipe lines with 149 emitters of 2.3 L h-1, and 14 emitters per tree. The soil is silty loam, typical Fluvisol (USDA, 2010), 2.5 m 150 deep, fertile, and low inorganic matter content (< 15.0 g kg-1). The roots are located predominately in the 151 first 50 cm of soil, corresponding to the intended wetting depth, although these exceed more than one meter 152 in depth. Soil-water content values at field capacity (–0.033 MPa) and wilting point (–1.5 MPa) were 0.35 153 and 0.12 m3 m–3 respectively, with an allowable soil-water depletion level of 0.27 m3 m–3. 154 The climatology in the study area is attenuated meso-Mediterranean, with an annual ET0 rate of 1,400 mm 155 and accumulated rainfall of 540 mm, mainly distributed from October to April. 156 157 2.2. Irrigation treatments 158 Three irrigation treatments were applied: i) a full irrigated treatment (C-100), which received 100% of the 159 crop evapotranspiration (ETc) during the irrigation period (60 – 304 day of the year, DOY), ii) a regulated 160 deficit irrigation (RDI-50), which received 100% of ETC except during the kernel filling period and pre-harvest; 161 when this treatment was irrigated at 50% of ETC. According to this, the kernel-filling period took place from 162 171 to 227 DOY and pre-harvest from 228 to 243 DOY; this period coinciding with the time in which the 163 kernel has finished its growth and the nut split period begins, just before the irrigation withholding (250 DOY) 164 seven days before the harvesting (257 DOY). iii) and a low-frequency deficit irrigation (LFDI) which received 165 the 100% ETc during the irrigation period, except during the kernel-filling stage and pre-harvest; when this 166 treatment was irrigated according the registered values of Ψleaf measured in shaded leaves. In this sense, 167 during the kernel-filling period (from 171 to 227 DOY) this treatment was subjected to irrigation-restriction 168 cycles with the following irrigation dynamic: Once started the kernel-filling period, irrigation was supressed, 169 till reaching values of Ψleaf close to -2.0 MPa. Then, trees were re-watered with the same periodicity and 170 amount of water as C-100 (approximately during 5 - 7 days) till reaching similar values of Ψleaf to those 171 registered in C-100. Once this threshold value was reached, this treatment was subjected to a new restriction 172 period until the threshold of Ψleaf (~ -2.0 MPa) was again surpassed. This dynamic of irrigation-restriction 173 cycles was maintained during whole stage of kernel filling period until harvesting. 174 Irrigation doses were calculated according to the methodology proposed by Allen et al. (1998), obtaining the 175 values of reference evapotranspiration according to the Penman-Monteith equation; by using a weather 176 station installed in the same experimental orchard; and using the crop coefficients obtained by García-Tejero 177 et al. (2015b), which ranged between 0.6 and 1.2. According to this, irrigation doses applied to C-100, RDI- 178 50 and LFDI were 6,850, 4,400 and 4,180 m3 ha-1, respectively (Table 1). 179 180 181 6 2.3. Plant measurements 182 During the experimental period, three daily curves of canopy temperature (TC), stomatal conductance to 183 water vapour (gs) and leaf water potential (Ψleaf) were obtained during the kernel filling and pre-harvest 184 period. These readings were taken at 08:30, 11:30, 14:30, 17:30 and 20:00 h local time, during the days 185 29th July (Curve 1) (210 DOY); 5th August (Curve 2) (217 DOY) and 27th August (Curve 3) (239 DOY). These 186 days coincided with the irrigation restriction periods of LFDI, with the aim of registering the crop physiological 187 status during periods of maximum water stress in this treatment. In this sense, Curve 2 was developed a 188 week after Curve 1. The reason was that, when Curve 1 was developed, LFDI has been subjected to seven 189 days of irrigation restriction. Taking into account the obtained results during this curve, it was decided to 190 extend this period once more week, in order to register the crop physiological response under a situation of 191 maximum stress. Finally, between Curve 2 and 3, there was a recovery period (from 218 to 225 DOY), being 192 the Curve 3 developed after 14 days without irrigation (in similar conditions at Curve 2). 193 Table 2 shows the values of air temperature (Tair), relative humidity (RH), and vapour pressure deficit (VPD) 194 registered during the sampling days and for each monitoring hour. 195 Measurements of Ψleaf were conducted by using a pressure chamber (Soil Moisture Equipment Corp., Sta. 196 Barbara, CA, USA), monitoring 12 trees per irrigation treatment (one leaf per tree), located in the north side 197 of the tree and being totally mature, fresh and shaded, at 1.5 m of height, approximately. Additionally, the 198 stomatal conductance to water vapor (gs), was measured in these same trees, by using a porometer SC-1 199 (Decagon Devices, INC, WA, USA), on one leaf completely exposed to the sun per monitored tree, and at 200 1.5 m of height. 201 TC was measured by using a ThermaCam (Flir SC660, Flir Systems, USA, 7-13 μm, 640x480 pixels) 202 throughout the day (8:30h, 11:30h, 14:30h, 17:30h, and 20:00h local time), with emissivity (ε) set at 0.96. 203 Each pixel corresponds to an effective temperature reading (Jones, 2004). Two methodologies were tested 204 to monitor the canopy temperature: i) 12 images were taken at tree level (one image per tree assessed, 205 thee being the same trees in which the measurements of Ψleaf and gs were developed), for each daily curve, 206 treatment and moment of the day), and ii) during Curves 1 and 2, thermal images were taken at row level, 207 so that, the trees monitored in the same image were subjected to the same irrigation treatment (Fig. 1). 208 These images at tree level were taken in the sunlit side of the trees, with the imager placed at 2 m of the 209 canopy (Fig. 1). Background temperature was determined by measuring the temperature of a crumpled 210 sheet of aluminium foil placed close to the leaves of interest using ε= 1 (Jones et al. 2002). To facilitate the 211 further analysis of these images, a cooled white screen was used as background, this being placed behind 212 of each monitored tree to simplify the isolation of the canopy surface through image processing. 213 Thermal images at tree level were analysed with the software developed by García-Tejero et al. (2012). This 214 software allows to remove those areas or pixels considered stem and the background (Fig. 2). 215 For the case of the images taken at row level, these were analysed using the software ThermaCam 216 Research Pro (Flir Systems, USA), selecting a specific area on the left and on the right and obtaining the 217 average value of TC for each area (Fig. 3). This methodology is much faster than the previous described by 218 7 García-Tejero et al. (2012), although it does not discriminates the representative areas with the same 219 feasibility, and the areas selection is done according to the visual perspective of the operator. 220 Considering the TC values obtained at tree level, three different thermal indicators were calculated: the 221 difference between canopy and the surrounding air (ΔTcanopy-air), the crop water stress index (CWSI), and the 222 index of the relative stomatal conductance these being calculated as follows (Costa et al., 2013): 223 224 ΔTcanopy-air = TC - Tair (1) 225 226 𝐶𝑊𝑆𝐼 = ∆𝑇𝑐𝑎𝑛𝑜𝑝𝑦−𝑎𝑖𝑟− ∆𝑇𝑤𝑒𝑡 ∆𝑇𝑑𝑟𝑦− ∆𝑇𝑤𝑒𝑡 (2) 227 228 𝐼𝐺= ∆𝑇𝑑𝑟𝑦− ∆𝑇𝑐𝑎𝑛𝑜𝑝𝑦−𝑎𝑖𝑟 ∆𝑇𝑐𝑎𝑛𝑜𝑝𝑦−𝑎𝑖𝑟− ∆𝑇𝑤𝑒𝑡 (3) 229 where ΔTcanopy-air, ∆Tdry and ∆Twet are the differences between canopy and air temperature for the crop in 230 the moment of the measurement, when the crop has the stomata fully closed and when it is fully transpiring, 231 respectively. TC is the canopy temperature and Tair the temperature of the surrounding air. 232 To obtain the reference values of ∆Twet, there was estimated the non-water stress baseline (ΔTcanopy-air = a 233 + b*VPD) according to Idso et al. (1981), using a ∆Tdry value equal to 5 ºC, as it was proposed by Jackson 234 et al. (1981). Non-water stress baseline was estimated using the canopy temperature readings obtained 235 from full irrigated trees (C-100). 236 237 2.4. Experimental design and statistical analysis 238 The experimental design was of randomized blocks, with four replications per irrigation treatment. Each 239 replication had 15 trees (3 rows and 5 trees per row), being monitored the three central rows for each 240 replication (n=12). 241 For each measurement day, an exploratory descriptive analysis of data (Ψleaf, gs and TC) was conducted by 242 applying a Levene’s test to check the variance homogeneity of the studied variables. Significant differences 243 between irrigation treatments (p ≤ 0.05) in the studied variables were identified by applying a one-way 244 ANOVA and a Tukey’s test for treatment separation, with the SPSS statistical software (SPSS Inc., 15.0 245 Statistical package; Chicago, IL, USA). 246 To evaluate the non-water stress baselines, a linear correlation analysis was made (n = 15). To evaluate 247 the relationships between variables, a linear correlation analysis between the values of thermal indicators 248 (TC, ΔTcanopy-air, CWSI and IG) and the crop physiological variables (Ψleaf and gs) was made, by using the 249 average values for each treatment and sampling time (n = 9). The obtained correlation coefficients were 250 8 used to identify which would be the best time to carry out TC readings and the most representative thermal 251 index as a proxy for crop physiology traits. 252 Finally, comparative study between the TC readings taken at tree and row level was conducted by means of 253 a linear correlation analysis between these values, using the average values for each treatment and the 254 whole data obtained during the two first daily curves (n=30). 255 256 3. Results and discussion 257 3.1. Daily evolution of crop physiological status 258 Figure 4 shows the evolution of Ψleaf, gs, and TC measured at tree level during the three daily curves 259 developed during the irrigation period in which the water stress regimes were imposed. On overall, as the 260 climatic conditions along the day became more adverse, Ψleaf reached more negative values, with a final 261 recovery at the end of the day. By contrast, gs increased during the first readings until reaching a maximum 262 point in which a significant decrease was observed, this coinciding with the moment of the day in which the 263 climatic conditions were more extreme. After this point a slight recovery of gs was found with the last 264 measurements of the day. In relation to TC this variable showed a more dependent trend on the climatic 265 conditions along the day, reaching the maximum values in those moments in which the Tair values were the 266 highest. During curves 1 and 2, the lowest values of Ψleaf were reached at 17:30 h, coinciding with the highest 267 VPD values registered during these days; and with the moments in which the TC values were maximum. 268 Considering the obtained values for each treatment, no differences were observed at 8:30 h, but these were 269 appearing along the day without observing a total recovery between the DI treatments and C-100 at 20:00 270 h. It is remarkable that the observed differences in terms of Ψleaf were higher during the Curve 2, this being 271 associated with the more severe climatic conditions detected and the imposed water restriction period for 272 LFDI in this curve, which had been prolonged for a further seven days, in comparison to Curve 1. 273 Regarding to gs, during Curve 1, all the treatments showed a growing tendency, reaching the maximum 274 values at 14:30 h (VPD = 2.61 kPa). However, during the Curve 2, the maximum values were observed at 275 11:30 h (VPD = 1.82), from which gs decreased, showing a partial recovering in C-100 at the end of the day. 276 This difference observed for the case of gs could be associated with the more severe climatic conditions 277 registered during the Curve 2, in comparison to the previous one. Finally, it is noticeable that the depletion 278 in TC was accompanied with a slight recovery of gs and the slight recovery of Ψleaf and gs during the readings 279 at 20:00 h. 280 Regarding to the values obtained during the Curve 3, it was obtained three weeks after Curve 2, when 281 climate conditions were similar to those observed in the previous one, and LFDI was subjected to 15 days 282 of irrigation restriction. In this sense, it was observed a similar trend to that detected in Curve 2, with the 283 highest values of gs observed at 11:30 (VPD = 1.31 kPa), with a significant reduction in all the treatments at 284 14:30 h, followed by a partial recovery at 17:30 h, and a new descend at the end of the day. This decreasing 285 trend occurred at 14:30 h, being this response associated with a partial stomatal closure, when climatic 286 conditions, specially the VPD values are strongly elevated. Even more, this descend in the values of gs 287 9 promoted that, the readings of Ψleaf between 11:30 and 14:30 were similar, and the partial recovery of gs at 288 17:30 was accompanied with a significant lessen values of Ψleaf. 289 Relating to the TC readings, these were highly determined by the climatic conditions. On overall, TC readings 290 in the three studied treatments were below to air temperature (Tair), except the readings taken at 08:30 and 291 11:30 h for the Curves 2 and 3. The highest differences in TC between treatments were detected specially 292 in the readings taken at 11:30, 14:30 and 17:00, although these were not as patent as for the case of Ψleaf. 293 On overall, and taking into account the monitored physiological variables, it can be assumed that Ψleaf was 294 the parameter that reflected the highest differences between treatments. In this sense, during the Curve 1, 295 significant differences were observed between C-100 and the remaining treatments at 11:30 and 14:30 h, 296 with an abrupt descend in the readings conducted at 17:30 (<-2.0 MPa), without differences between the 297 three irrigation treatments. During the Curve 2, the Ψleaf values registered in C-100 were significant different 298 than those registered in the remaining treatments during all day (except at 8:30 h), not being reached the 299 threshold value of -1.5 MPa in C-100. Finally, it also draws attention that, during Curve 3, C-100 reached 300 again Ψleaf values close to -2.0 MPa, as it was fitted for the Curve 1. 301 It is remarkable that, whereas Ψleaf was able to show significant differences between treatments, this fact 302 was not as patent in terms of gs, because of the low capacity of almond to regulate the stomatal closure 303 under drought conditions. In this regard, almond trees present a fast recovery of water potential, but a delay 304 in the values of gs as it has been stated by authors such as Torrecillas et al. (1996) or Romero et al. (2004). 305 In this line, in physiological terms, when almond is subjected to a mild-to-moderate water stress situation a 306 stomatal conductance reduction is not as patent as the effects in terms of water potential because of its low 307 capability of regulating the stomata when a water stress situation is applied, as it has been discussed by 308 some authors such as Wartinguer et al. (1990), Egea et al. (2011) or Eichi (2013). In this agreement, 309 previously to observe a significant reduction in gs, almond responds with significant descends in terms of 310 leaf or stem water potential, (García-Tejero et al, 2012, 2015b). Consequently, almond would be able to 311 maintain acceptable levels of gs (promoting significant descends in the crop-water potential) but, keeping 312 optimum values of carbon assimilation, photosynthetic rate, and hence increasing the intrinsic water-use 313 efficiency (McCutchan and Shackel, 1992; Rouhi et al., 2007). 314 Gomes-Laranjo et al. (2006) reported values of Ψleaf for different cultivars, which ranged between -1.72 and 315 -2.0 MPa in Glorieta; -1.71 and -2.40 MPa in Ferragnes; -1.91 and -2.34 MPa in Francoli; -1.97 and -2.26 316 MPa in Lauranne, and -1.88 and -1.92 MPa in Masbovera. In this line, these values correspond to 317 measurements done at midday in well-watered trees, which are in line with the threshold range between - 318 1.5 and -2.0 MPa considered and obtained in this work for C-100. 319 Obviously, this water potential depletion affects to leaf gas exchange. In this sense, for full irrigated 320 conditions, daily cycle of gas exchange is almost constant when no radiation limitation occur (Torrecillas et 321 al., 1988; Klein et al., 2001; Romero et al., 2006) and vapour pressure deficit (VPD) is not higher than 2 KPa 322 (Romero et al., 2006). However, in our case, the values of VPD were higher than this value during the three 323 curves when the readings were taken between 11:30 and 20:00 h, which would explain the daily variation 324 16 conservación del agua en ambientes semiáridos. Tocina Estudios Locales, Revista de Investigación 535 Local, 5, 175-190. 536 García-Tejero, I.F., Costa, J.M., Egipto, R., Lima, R.S.N., Durán, V.H., Lópes, C., Chaves, M.M. 2016b. 537 Thermal data to monitor crop-water status in irrigated Mediterranean viticulture. Agric. Water 538 Manage. 176, 80-90. 539 García-Tejero, I., Hernández, A., Padilla-Díaz, C.M., Diaz-Espejo, A., Fernández, J.E. 2017. Assessing plant 540 water status in a hedgerow olive orchard from thermography at plant level. Agric. Water Manage, 541 188, 50-60. 542 Goldhamer, D.A., Fereres, E. 2016. Establishing an almond water production function for California using 543 long-term yield response to variable irrigation. Irrig. Sci. pp. 1–11. doi:10.1007/s00271-016-0528-2 544 (on line first). 545 Goldhamer, D.A., Girona, J. 2012. Crop yield response to water: almond’ In: Steduto P, Hsiao TC, Fereres 546 E, Raes D (eds) FAO irrigation and drainage Paper No. 66. Food and Agriculture Organization of the 547 United Nations, Rome, pp 358–373 548 Gomes-Laranjo, J., Coutinho, J.P., Galhano, V., Cordeiro, V. 2006. Responses of five almond cultivars to 549 irrigation: Photosynthesis and leaf water potential. Agric Water Manage. 83, 261-265. 550 Gonzalez-Dugo, V., Zarco-Tejada, P., Fereres, E. 2014. Applicability and limitations of using the crop water 551 stress index as an indicator of water deficits in citrus orchards. Agric. For. Meteorol. 198-199, 94- 552 104. 553 Idso, S.B., Jackson, R.D., Pinter, P.J.J., Reginato, R.J., Hatfield, J.L. 1981. Normalizing the stress degree- 554 day parameter for environmental variability. Agric. Meteorol. 24, 45–55. 555 IPCC. 2014. Climate change 2014. Impacts, adaptations and vulnerabilities. Part B. Regional aspects. 556 Barros, V & Field CB (eds). Cambridge University Press, 688 pp. 557 Jackson, R.D., Idso, S.B., Reginato, R.J., Pinter, P.J. 1981. Canopy temperature as a crop water-stress 558 indicator. Water Resources Res. 17, 1133-1138. 559 Jiménez-Bello, M.A., Ballester, C., Castel, J.R., Intrigliolo, D.S. 2011. Development and validation of an 560 automatic termal imaging process for assessing plant water status. Agric. Water Manage. 98, 1497- 561 1504. 562 Jones, H.G. 1999. Use of infrared thermometry for estimation of stomatal conductance as a possible aid to 563 irrigation scheduling. Agric. Forest Meteorol. 95, 139-149. 564 Jones, H.G. 2004. Irrigation scheduling: advantages and pitfalls of plant-based methods. J. Exp. Bot. 55, 565 2427-2436. 566 Jones, H.G., Stoll, M., Santos, T., de Sousa, C., Chaves, M.M., Grant, O.M. 2002. Use of infra-red 567 thermography for monitoring stomatal closure in the field: application to the grapevine. J. Exp. Bot. 568 53, 2249-2260. 569 17 Jones H.G., Serraj, R., Loveys, B.R., Xiong, L., Wheaton, A., Price, A.H. 2009. Thermal infrared imaging of 570 crop canopies for the remote diagnosis and quantification of plant responses to water stress in the 571 field. Functional Plant Biology 36, 978–979. 572 Klein, I, Esparza, G., Weinbaum, S.A., DeJong, T.M. 2001. Effects of irrigation deprivation during the harvest 573 period on leaf persistence and function in mature almond trees. Tree Physiol. 21, 1063-1072. 574 López-López M., Espadafor, M., Testi, L., Lorite, I.J., Orgaz, F., Fereres, E. 2018. Water use of irrigated 575 almond trees when subjected to water deficits. Agric. Water Manage. 195, 84-93. 576 Mirás-Avalos, J.M.; Pérez-Sarmiento, F.; Alcobendas, R.; Alarcón, J.J.; Mounzer, O.; Nicolás, E. Using 577 midday stem water potential for scheduling deficit irrigation in mid–late maturing peach trees under 578 Mediterranean conditions. Irrig. Sci. 2016a, 34, 161–173. 579 McCutchan, H., Shackel, K.A. 1992. Stem-water Potential as a Sensitive Indicator of Water Stress in Prune 580 Trees (Prunus domestica L. cv. French). J. Am. Soc. Hort. Sci. 117, 607–611. 581 Nortes, P.A., Pérez-Pastor, A., Egea, G., Conejero, W., Domingo, R. 2005. Comparison of changes in stem 582 diameter and water potential values for detecting water stress in young almond trees. Agric. Water 583 Manage. 77, 296–307. 584 Möller, M., Alchanatis, V., Cohen, Y., Meron, M., Tsipris, J., Naor, A., Ostrovsky, V., ,Sprintsin, M., Cohen, 585 S. 2007. Use of thermal and visible imagery for estimating crop water status of irrigated grapevine. 586 J. Exp. Bot. 58, 827–838. 587 Olesen, J.E., Trnka, M., Kersebaum, K.C., Skjelvåg, A.O., Seguin, B., Peltonen-Sainio, P., Rossi, F., Kozyra, 588 J., Micale, F. 2011, Impacts and adaptation of European crop production systems to climate change. 589 Europ. J. Agron. 34, 96-112. 590 Phogat, V., Skewes, M.A., McCarthy, M.G., Cox, J.W., Simunek, J., Petrie, P.R. 2017. Evaluation of crop 591 coefficients, wáter poroductivity and wáter balance components for wine grapes iirrigated at different 592 defiti levels by a sub-surface drip. Agric. Water Manage. 180, 22-34. 593 Poblete-Echeverría, C., Ortega-Farías, S., Zúñiga, M. 2014. Use of infrared thermography on canopies as 594 indicator of water stress in Arbequina olive orchards. Acta Hort, 1057, 399-404. 595 Poblete-Echeverría, C., Sepúlveda-Reyes, D., Ortega-Farias, S., Zúñiga, M., Fuentes, S. 2016. PLant wáter 596 stress detection based on aerial and terrestrial infrered thermography: a study case from vineyard 597 and olive orchard. Acta Hort 1112, 141-146 598 Pou, A., Diago, M.P., Medrano, H., Baluja, J., Tardaguila, J. 2014. Validation of thermal indices for water 599 stress status identification in grapevine. Agric. Water Manage. 134, 60-72. 600 Puerto, P., Domingo, R., Torres, R., Pérez-Pastor, A., García-Riquelme, M. 2013. Remote management of 601 deficit irrigation in almond trees base on maximum daily trunk shrinkage. Water relations and yield. 602 Agric. Water Manage. 126, 33-45. 603 Remorini, D., Massai, R. 2003. Comparison of water status indicators for young peach trees. Irrig. Sci. 22, 604 39-46. 605 18 Rouhi, V., Samson, R., Lemeur, R., Van Damme, P. 2007. Photosynthetic gas exchange characteristics in 606 three different almond species during drought stress and subsequent recovery. Environ. Exp. Bot. 607 59, 117–129. 608 Romero, P., Botía, P., García, F. 2004. Effects of regulated deficit irrigation under subsurface drip irrigation 609 conditions on vegetative development and yield of mature almond trees. Plant and Soil 260, 169– 610 181. 611 Romero, P., García, J., Botía, P. 2006. ‘Cost–benefit analysis of a regulated deficit-irrigated almond orchard 612 under subsurface drip irrigation conditions in South-eastern Spain’, Irrig. Sci. 24, 175–184. 613 Shackel, K. 2011. A plant-based approach to deficit irrigation in trees and vines. HortSci. 46, 173–177. 614 Spinelli, G.M., Snyder, R.L., Sanden, B.L., Shackel, K.A. 2016. Water stress causes stomatal closure but 615 does not reduce canopy evapotranspiration in almond. Agric. Water Manage. 168, 11-22. 616 Testi, L., Goldhamer, D.A., Iniesta, F., Salinas, M. 2008. Crop water stress index is a sensitive water stress 617 indicator in pistachio trees. Irrig. Sci. 26, 395–405. 618 Torrecillas, A., Ruiz-Sanchez, M.C., del Amor, F., León, A. 1988. Seasonal variations on water relations of 619 Amygdalus communis L. under drip irrigated and non irrigated conditions. Plant and Soil. 106, 215- 620 220. 621 Torrecillas, A., Alarcón, J.J., Domingo, R., Planes, J., Sánchez-Blanco, M.J. 1996. Strategies for drought 622 resistance in leaves of two almond cultivars. Plant Sci. 118, 135–143. 623 USDA. 2010. Keys to soil taxonomy (11th Edition). United States Department of Agriculture, Natural 624 Resource Conservation Service, 334 pp. 625 Wartinger, A., Heilmeier, H., Hartung, W., Schulze, E.D. 1990. Daily and seasonal courses of leaf 626 conductance and abscisic acid in the xylem sap of almond trees (Prunus dulcis (Miller) D.A.Webb) 627 under desert conditions. New Phytol. 116, 581–587. 628 Zarco-Tejada, P.J., Berni, J.A.J., Suárez, L., sepulcré-Cantó, G., Morales, F., Miller, J.R. 2009. Imaging 629 chlorophyll fluorescence with an airbone narrow-band multispectral camera for vegetation stress 630 detection. Remote Sens. Environ. 113, 1262-1275. 631 632 633 634 635 636 637 638 19 639 640 FIGURES 641 642 643 644 Figure 1. Example of thermal images at plant (left) and row (right) level 645 646 647 648 Figure 2. Example of image processing using the software developed by García-Tejero et al (2012). 649 On the right, the initial thermal imaging; on the left, a bitmap image, in which the black area 650 represents the pixels of the thermal image considered to calculate the canopy temperature. 651 652 653 654 Figure 3. Example of image processing at row level using the ThermaCam Research Pro (Flir 655 Systems, USA). 656 657 20 658 Figure 4. Daily curves of leaf water potential (Ψleaf), stomatal conductance (gs) canopy temperature 659 (Tc) and air temperature (Tair) in almond trees subjected to different irrigation doses: C-100, full 660 irrigated treatment; RDI-50, regulated deficit irrigation; LFDI, low-frequency deficit irrigation. Letters 661 a, b, and c show significant differences between C-100, RDI-50 and LFDI treatments, respectively 662 (p<0.05). 663 664 665 666 21 Figure 5. Non-water stress baseline (∆Tcanopy-air) = a*VPD + b. Data obtained for the DOYs 210, 217 667 and 239 and using the readings taken at 8:30, 11:30, 14:30, 15:30 and 20:00. 668 669 670 671 Figure 6. Relationships between canopy temperature readings (TC), the difference between 672 canopy and air temperature (ΔTcanopy-air) and crop water stress index (CWSI) with leaf-water 673 potential (Ψleaf). 674 675 676 Figure 7. Relationships between canopy temperature readings (TC) at tree and row level. 677 678 679 680 681 682 22 TABLES 683 684 Table 1. Climatic conditions, water requirements and irrigation doses applied during the season 685 Period (DOY) Tair (ºC) RH (%) Rainfall (mm) ET0 (mm) KC ETC (mm) C-100 (mm) RDI-50 (mm) LFDI (mm) 60 to 90 13.53 70.18 55.6 92.11 0.3 27.63 0 0 0 91 to 120 17.95 72.96 35.6 114.84 0.55 34.74 8.03 8.03 8.031 121 to 151 21.42 54.06 12.6 175.79 0.9 118.66 109.21 109.21 109.21 152 to 181 23.40 58.88 7.4 176.94 1.05 167.21 161.66 108.32 107.88 182 to 212 25.24 58.81 0.2 184.24 1.15 190.69 190.54 91.46 77.85 213 to 243 26.05 54.06 0 173.82 1.15 179.90 179.90 88.15 79.56 244 to 273 22.66 78.06 175.8 104.99 0.8 75.59 23.02 23.02 23.02 274 to 304 19.93 77.67 73.2 79.86 0.7 50.31 12.5 12.5 12.5 DOY. day of the year; Tair. average air temperature; RH. average relative humidity. ET0. reference 686 evapotranspiration; KC. crop coefficient; ETC. crop evapotranspiration; C-100. control treatment; SDI-50. 687 regulated deficit irrigation at 50% of ETC during the kernel filling period; LFDI. low-frequency deficit irrigation 688 during the kernel filling period. 689 690 691 Table 2. Average values of air temperature (Tair). relative humidity (RH) and vapour pressure deficit 692 (VPD) registered during the daily curves 693 Curve 1 (210 DOY) Curve 2 (217 DOY) Curve 3 (239 DOY) Hour Tair (ºC) RH (%) VPD (kPa) Tair (ºC) RH (%) VPD (kPa) Tair (ºC) RH (%) VPD (kPa) 08:30 24.4 63 0.94 16.6 63 0.70 21.3 60 1.01 11:30 25.3 69 0.99 27.4 50 1.82 27.4 64 1.31 14:30 34.1 51 2.61 33.4 45 2.82 37.5 31 4.43 17:30 34.5 42 2.29 42.1 21 6.49 41.0 25 5.81 20:00 31 41 1.84 34.1 27 3.90 38.8 26 5.10 Tair. average air temperature; RH. average relative humidity; VPD. vapour pressure deficit 694 695 696 697 698 699 700 701 702 703 23 Table 3. Pearson’s correlation coefficients between thermal information and the studied 704 physiological variables 705 Hour TC ∆Tcanopy-air CWSI IG 8:30 gs -0.32* ns -0.40* ns Ψleaf ns ns ns ns 11:30 gs ns ns ns ns Ψleaf -0.85** -0.69* -0.85** ns 14:30 gs -0.70* ns -0.82** ns Ψleaf -0.39* ns -0.69* ns 17:30 gs ns -0.70* -0.62* ns Ψleaf -0.39* ns -0.34* 0.74** 20:00 gs -0.75* ns ns ns Ψleaf ns ns ns ns TC. canopy temperature; ∆Tcanopy-air. difference between canopy and air temperature; CWSI. crop- 706 water stress index; IG. relative index of stomatal conductance; gs. stomatal conductance to water 707 vapour; Ψleaf. leaf-water potential in shaded leaves. * and ** show significant relationships at 708 confidence level of 95 and 99%. respectively. 709 710 711 712 713 714 715 716 717 718 719 720 721 9 Table 4. Example of false-coloured images taken at tree and row level during the Curve 1 in the different irrigation treatments and moment of the day. The values of canopy temperature (TC) correspond to the average of five measurements taken for each treatment and moment of the day. Tree level Row level C-100 RDI-50 LFDI C-100 RDI-50 LFDI 08:30 TC (ºC) 21.7 21.6 21.4 21.3 21.3 21.4 11:30 TC (ºC) 23.3 23.6 23.4 23.5 23.6 23.4 14:30 TC (ºC) 26.9 27.0 29.9 27.4 28.6 29.1 17:30 TC (ºC) 32.5 30.9 30.0 32.3 30.7 31.0 20:00 TC (ºC) 29.2 28.6 28.9 29.1 28.6 29.0 10 Table 5. Example of false-coloured images taken at tree and row level during the Curve 2 in the different irrigation treatments and moment of the day. The values of canopy temperature (TC) correspond to the average of five measurements taken for each treatment and moment of the day. Tree level Row level C-100 RDI-50 LFDI C-100 RDI-50 LFDI 08:30 TC (ºC) 20.9 20.5 22.2 21.2 21.2 21.4 11:30 TC (ºC) 27.4 28.6 29.1 27.6 28.7 29.0 14:30 TC (ºC) 30.2 32.5 32.9 30.7 32.8 33.1 17:30 TC (ºC) 31.8 32.2 31.1 30.5 30.5 31.2 20:00 TC (ºC) 31.1 31.5 31.9 31.3 31.7 32.2