Quantitative analysis of almond yield response to irrigation regimes in Mediterranean Spain

Abstract

Almond plantations are expanding worldwide, specifically in Spain; the new orchards are often designed under more intensive systems in comparison to the traditional rainfed orchards frequently found in the Mediterranean Sea basin. In these new areas, water is the main limiting factor, and therefore, the present research is aimed at quantitatively analyzing previous findings obtained in irrigation field trials carried out in Spain with mature almond trees. The goal was to derive applied water-production functions and compare sustained and regulated deficit irrigation strategies to provide robust information on the marginal water productivity and the preferred irrigation option to be applied under water scarcity conditions. This quantitative analysis reported a yield increase as water application increased, with the highest potential yield of about 2500 kg/ha achieved with around 1000 mm of irrigation water applied. Under severe water restrictions, similar responses were observed regardless of the deficit irrigation technique employed. In contrast, under moderate water stress, it seems more advantageous to apply a regulated deficit irrigation strategy rather than a sustained deficit strategy. The reported results are useful for deriving more sustainable irrigation protocols and highlight the need to optimize other inputs in addition to water to take full advantage of the irrigation intensification to be carried out in the new almond plantations.

Full text

Agricultural Water Management 279 (2023) 108208
0378-3774/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Quantitative analysis of almond yield response to irrigation regimes in
Mediterranean Spain
Jos´
e M. Mir´
as-Avalos
a
,
*
, Victoria Gonzalez-Dugo
b
, Iv´
an F. García-Tejero
c
,
Ram´
on L´
opez-Urrea
d
, Diego S. Intrigliolo
e
, Gregorio Egea
f
a
UA-RAMA. Departamento de Sistemas Agrícolas, Forestales y Medio Ambiente (Unidad asociada a EEAD-CSIC Suelos y Riegos), Centro de Investigaci´
on y Tecnología
Agroalimentaria de Arag´
on (CITA), Avda. Monta˜
nana 930, 50059 Zaragoza, Spain
b
Instituto de Agricultura Sostenible (IAS), Consejo Superior de Investigaciones Cientícas (CSIC), Alameda del Obispo s/n, 14004 C´
ordoba, Spain
c
Center “Las Torres”, Andalusian Institute of Agricultural and Fisheries Research and Training (IFAPA), Carretera Sevilla-Cazalla km 12.2, 41200 Sevilla, Spain
d
Instituto T´
ecnico Agron´
omico Provincial (ITAP), Parque Empresarial Campollano, 2ª Avda. Nº 61, 02007 Albacete, Spain
e
Department of Ecology, Desertication Research Centre (CIDE), CSIC-UV-GVA, Carretera CV 315, km 10,3 46113 Moncada, Valencia, Spain
f
Area of Agroforestry Engineering, Technical School of Agricultural Engineering (ETSIA), Universidad de Sevilla, Ctra. Utrera km 1, 41013 Sevilla, Spain
ARTICLE INFO
Keywords:
Decit irrigation
Marginal productivity
Production function
Prunus dulcis
Water stress
ABSTRACT
Almond plantations are expanding worldwide, specically in Spain; the new orchards are often designed under
more intensive systems in comparison to the traditional rainfed orchards frequently found in the Mediterranean
Sea basin. In these new areas, water is the main limiting factor, and therefore, the present research is aimed at
quantitatively analyzing previous ndings obtained in irrigation eld trials carried out in Spain with mature
almond trees. The goal was to derive applied water-production functions and compare sustained and regulated
decit irrigation strategies to provide robust information on the marginal water productivity and the preferred
irrigation option to be applied under water scarcity conditions. This quantitative analysis reported a yield in-
crease as water application increased, with the highest potential yield of about 2500 kg/ha achieved with around
1000 mm of irrigation water applied. Under severe water restrictions, similar responses were observed regardless
of the decit irrigation technique employed. In contrast, under moderate water stress, it seems more advanta-
geous to apply a regulated decit irrigation strategy rather than a sustained decit strategy. The reported results
are useful for deriving more sustainable irrigation protocols and highlight the need to optimize other inputs in
addition to water to take full advantage of the irrigation intensication to be carried out in the new almond
plantations.
1. Introduction
Global awareness of the health benets of nuts has resulted in a rapid
expansion of the nuts industry, which is expected to continue. In the case
of the almond crop, the global harvested area has increased from
1776,546 ha in 2015 to 2162,263 ha in 2020, an increase of 22% in ve
years (FAOSTAT, 2021). Most of the new almond orchards planted in the
main producing countries are intensive irrigated plantations, which
would explain the 54% increase in almond production worldwide
observed over the same period; from 2696,057 t of shelled almonds in
2015 to 4140,043 t in 2020 (FAOSTAT, 2021). The United States of
America, Spain and Australia are the world’s leading almond producers,
with a share of 57%, 10% and 5% of world production in 2020,
respectively (FAOSTAT, 2021).
The almond sector in Spain is undergoing a substantial trans-
formation, with an increase in the cultivated area from 580,467 ha in
2015 (ESYRCE, 2015) to 721,796 ha in 2020 (ESYRCE, 2020). Most of
the new almond plantations are being developed in traditional irrigated
areas, displacing other irrigated crops (i.e., stone fruits) which have a
lower protability. As a consequence, Spain’s irrigated almond area has
almost tripled in ve years, from 52,990 ha in 2015 to 139,399 ha in
2020 (ESYRCE, 2020, 2015). Despite this increase in irrigated areas,
low-yielding rainfed almond orchards still represent 80% of the culti-
vated area, which explains the low average yields in Spain (0.58 t/ha of
* Correspondence to: Departamento de Sistemas Agrícolas, Forestales y Medio Ambiente (Unidad asociada a EEAD-CSIC Suelos y Riegos), Centro de Investigaci´
on y
Tecnología Agroalimentaria de Arag´
on (CITA), Avda. Monta˜
nana 930, 50059 Zaragoza, Spain.
E-mail address: jmmiras@cita-aragon.es (J.M. Mir´
as-Avalos).
Contents lists available at ScienceDirect
Agricultural Water Management
journal homepage: www.elsevier.com/locate/agwat
https://doi.org/10.1016/j.agwat.2023.108208
Received 18 June 2022; Received in revised form 23 January 2023; Accepted 31 January 2023

Agricultural Water Management 279 (2023) 108208
2
shelled almonds) as compared to those of the USA (4.68 t/ha of shelled
almonds) (FAOSTAT, 2021), where most of the almond plantations are
intensively irrigated (Goldhamer and Fereres, 2017).
One of the main threats of the new drip-irrigated almond plantations
in Spain is the reduced water allocation provided by the regulatory
authorities for this crop species. For instance, the Hydrographical
Confederation of the Guadalquivir River Basin establishes an endow-
ment of 250 mm for almond tree plantations in its hydrological regu-
latory plans (CHG, 2015). This water amount is notably lower than the
irrigation requirements to meet the maximum crop evapotranspiration
(ETc) of this species under Mediterranean climate conditions, which can
exceed 1300 mm in California (Goldhamer and Fereres, 2017), and 800
mm in southern Spain (L´
opez-L´
opez et al., 2018). These water alloca-
tions can be further reduced in drought periods, which are expected to
be more frequent due to climate change in arid and semi-arid regions.
Under these scenarios of severe water shortages, experimental works
such as that by Moldero et al. (2022) demonstrate the need to contin-
uously develop optimal irrigation management strategies to cope with
both the chronic water shortages of most almond producing areas of
Spain, and the extreme events caused by cycling droughts. In this regard,
Moldero et al. (2022) found that almond trees grown in southern Spain
that were submitted to severe water deprivation during a single season,
experienced very high tree mortality (92%) when submitted to rainfed
conditions after previous seasons of full irrigation application. In
contrast, those receiving only 25% ETc had a 33% yield reduction as
compared to fully irrigated trees, and recovered yield levels in the
following years.
In the last two decades, a great research effort has been carried out in
Spain to determine the physiological and agronomic responses of
almond trees to irrigation strategies supplying water depths lower than
those required to meet the maximum ETc, namely decit irrigation (DI)
strategies (Fereres and Soriano, 2007). Within the term DI, a distinction
can be made between sustained decit irrigation strategies (SDI), aimed
at applying a certain level of water decit throughout the growing
season, and regulated decit irrigation strategies (RDI), aimed at
applying water decit only in certain phenological periods that are less
sensitive to water stress (Egea et al., 2013). RDI strategies in almond
trees have mainly consisted in applying a more or less severe water
decit during the grain-lling stage, coinciding with the months of
highest water demand (Egea et al., 2010; García-Tejero et al., 2019;
Girona et al., 2005; L´
opez-L´
opez et al., 2018; Ma˜
nas et al., 2014).
However, this has not always been the case, as moderate water decits
have also been applied in the rapid fruit growth and/or postharvest
stages in some eld experiments (Moldero et al., 2021; Puerto et al.,
2013; Romero et al., 2004). The great range of RDI treatments tested,
together with the high number of cultivars evaluated, the differing soil
(i.e., deep vs shallow soils), and weather conditions (i.e., semi-arid
Mediterranean, Continental, Mediterranean), complicate drawing solid
and clear messages to convey to irrigation managers and farmers on the
most suitable irrigation strategy for a given water allocation in
drought-prone areas, such as Spain.
In addition to RDI, SDI irrigation strategies (Egea et al., 2010; Gar-
cía-Tejero et al., 2020; Girona et al., 2005; Guti´
errez-Gordillo et al.,
2020; Lipan et al., 2020; L´
opez-L´
opez et al., 2018; Ma˜
nas et al., 2014;
Moldero et al., 2021) with different degrees of water decit ranging
from 75% ETc (García-Tejero et al., 2020; L´
opez-L´
opez et al., 2018) to
25% ETc (Ma˜
nas et al., 2014) have also been evaluated in the different
almond eld trials conducted in Spain. In an experiment carried out in
California with almond trees, Goldhamer et al. (2006) found that RDI
trees had greater yields than SDI trees when similar amounts of water
were applied. However, no clear differential patterns among RDI and
SDI strategies were observed in the almond experiments conducted in
Spain (Egea et al., 2013; Girona et al., 2005; L´
opez-L´
opez et al., 2018;
Ma˜
nas et al., 2014; Moldero et al., 2021). In this sense, a meta-analysis
of all the data collected in the experiments conducted so far in Spain on
the response of almond trees to decit irrigation, would help to unravel
some questions on the management of decit irrigation in almond or-
chards under the soil and climatic conditions found in Spain. For this
reason, this work tries to answer the following questions: (1) for a given
water allocation below the total crop requirements, what would be the
most appropriate irrigation strategy for almond trees grown in Spain?;
(2) for a given water allocation, what yield loss can be expected versus
that of a well-watered orchard?; (3) under the edaphoclimatic condi-
tions of Spain, is the productive response of almond trees to decit
irrigation conditioned by the timing at which the water stress is applied,
or does it depend mainly on the percentage of ETc supplied annually?
2. Materials and methods
2.1. Preparation of the database
Data were collected from studies performed in Spain in which
almond trees were subjected to different irrigation regimes. To obtain
these data, research groups from all over Spain were contacted and
asked to provide the data from their published works. Articles were
restricted to those in which a full irrigation control was compared to
either a regulated or sustained decit treatment. Ideally, the three irri-
gation modalities (namely, full, regulated, and sustained) were investi-
gated within the same study. Whenever possible, a rain-fed treatment
was also considered. The criteria for incorporating a work into the nal
analysis were the following: 1) the experimental characteristics should
indicate both in-season (or annual) rainfall and the amount of irrigation
applied to each treatment, 2) the articles had to report yields for each
treatment, and 3) the almond trees should be, at least, ve years old. In
the end, the database contained 15 articles for a total of 173 observa-
tions, mainly located in Southern and Eastern Spain (Table 1). Data from
the selected articles included some years in which the trees were four
years old, thus not meeting one of the criteria for selection. Therefore,
data were ltered to select only those which referred to adult trees (at
least ve years after their plantation). Moreover, a treatment with over-
irrigated trees was removed from the analysis, except for the calculation
of the production function and marginal water productivity, where these
data were included (see below). In the end, the database for adult
almond trees consisted of 144 observations.
A database was created by listing the irrigation regimes in each
study. Yield data were arranged as paired observations in which decit
irrigation treatments were compared to a full irrigation control. The
treatments classied as moderate water-stress were those that received
annual irrigation volumes above 55% of those received by the control
treatments, whereas those that received annual irrigation depths below
55% of maximum crop water requirements were considered severe
water-stressed treatments. The stress coefcient threshold value of 55%
was chosen based on the production function obtained by Moldero et al.
(2021), who observed in their trials carried out in Southern Spain, that
reduced yield losses (≈15%) were expected for 45% irrigation shortages,
and that kernel yield was impaired more signicantly with water
shortages higher than 45% of maximum crop water requirements. Other
data referred to the experimental conditions, including location, irri-
gation system design, cultivar, rootstock, spacings, tree density, and age,
and external factors such as rainfall received (per year and growing
season), and clipped grass reference evapotranspiration (ET
o
) were
included in the database. Relevant moderators are shown in Table 2. All
studies used conventional management practices, so this was not
included in the list of moderators. The meta-analysis cannot be per-
formed on continuous variables; hence, the moderators were
sub-divided into categories (Mitchell-McCallister et al., 2020). New
drip-irrigated almond plantations in Spain (including recently devel-
oped varieties, high density planting systems, and regions where almond
is newly introduced) would have different water needs, but the research
on these new plantations is scarce and, consequently, we did not
consider them for the quantitative analysis carried out, focusing on the
more traditional almond orchards.
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
3
2.2. Relative yield and water production
To reduce the variability in the results from the studies considered,
which involved different almond cultivars, irrigation amounts, soil
types, and rainfall regimes, yields were relativized to the yield observed
in the full-irrigation control corresponding to each study. With this, data
from all the studies could be easily compared.
Moreover, applied water production functions for each irrigation
strategy (either FI, RDI, or SDI, and for all of them combined) were
calculated by plotting the mean yield response to the water applied, and
tting a second-order polynomial expression (Goldhamer and Fereres,
2017). The marginal water productivity was computed as the derivative
of the water productivity function and plotted against the applied water
(Goldhamer and Fereres, 2017).
2.3. Data analysis
An exploratory analysis, including descriptive statistics, boxplots,
and scatterplots for relating different variables and external factors, was
rst conducted. Generalized linear models between yield (both total and
relative) and water received (both rainfall and irrigation) were per-
formed, and regression coefcients were computed. Shapiro-Wilks and
Bartlett tests were used for assessing the normality of yield data among
water decit treatments, to carry out an ANOVA for evaluating the effect
of watering types and regimes on almond yield. Means were separated
using Tukey’s test.
A meta-analysis was performed to aggregate the results from the
individual studies and, thus, obtain greater statistical power. Meta-
analysis is a research process used to systematically synthesize and
merge the ndings of single, independent studies, using statistical
methods to calculate an overall or ‘absolute’ effect (Egger and Smith,
1997; Shorten and Shorten, 2013). This technique uses well recognised,
systematic methods to account for differences in sample size, variability
(heterogeneity) in study approach and ndings (treatment effects) and
test how sensitive their results are (Egger and Smith, 1997; Borenstein
et al., 2009). This technique has provided further insights into the im-
pacts of agricultural practices on crop yield and water use efciency
(Fan et al., 2018; Mitchell-McCallister et al., 2020). The meta-analysis
was conducted using the “meta” and “metasens” packages (Balduzzi
et al., 2019; Schwarzer, 2007; Schwarzer et al., 2015) under the R sta-
tistical environment (R Core Team, 2021). A random effects model was
considered to assess yield under decit irrigation, as we assumed that
the true effect varied across studies (Borenstein et al., 2009). Moreover,
a xed effects model was also considered.
Cochran’s Q statistic was used to assess heterogeneity, testing the
null hypothesis that all the studies share a common effect size. This
statistic follows a chi-square distribution with the number of studies
minus one degree of freedom. The percentage of variation across studies
due to heterogeneity rather than chance was assessed through the I
2
statistic, which is computed as:
I
2
=(Q – df) / Q ×100 (1)
where Q is the Cochran’s heterogeneity statistic, and df means degrees of
freedom. Values of I
2
range from 0% to 100%, where values of 25%,
50%, and 75% represent low, medium, and high heterogeneity (Bor-
enstein et al., 2009; Higgins et al., 2003).
Graphical and statistical methods were used for determining publi-
cation bias, which is the most signicant source of Type I errors in a
meta-analysis (Harrison, 2011). Funnel plots were used to present the
effect size plotted against the standard error, placing the effect sizes of
small studies at the bottom of the funnel and larger studies concentrated
at the top. Funnel plots are symmetrical in the absence of bias (Sterne
et al., 2006).
3. Results
3.1. Description of the dataset
Table 3 summarises the number of data, mean, maximum and min-
imum values for each variable, as well as the number of missing data.
Yield and irrigation applied data were present in the 144 observations
(Table 3), whereas the rest of the variables showed missing data. Yield in
these studies showed a wide spectrum of values, ranging from 352 to
3329 kg/ha (Table 3), while irrigation applied varied from 7 to 985 mm
(Table 3).
A categorical variable representing the ratio between the irrigation
applied to a given decit treatment, over the irrigation applied to the
Table 1
Published studies included in the database for the meta-analysis of the use of decit irrigation in Spanish almond orchards.
Publication N obs Irrigation treatments Cultivar Age Spacings N years Region
Egea et al. (2010) 10 FI; RDI; SDI Marta 5 7 ×6 2 Murcia
Egea et al. (2013) 4 FI; RDI; SDI Marta 5 7 ×6 1 Murcia
García-Tejero et al. (2019) 9 FI; RDI Guara 5 7 ×6 3 Andalucía
García-Tejero et al. (2020) 9 FI; SDI Guara, Lauranne, Marta 7 8 ×6 1 Andalucía
Girona et al. (1997) 15 FI; RDI Marcona 16 5 ×5 3 Catalu˜
na
Girona et al. (2005) 12 FI; RDI; SDI Ferragn`
es 6 5 ×6 3 Catalu˜
na
Guti´
errez-Gordillo et al. (2019b) 6 FI; SDI Guara, Lauranne, Marta 6 8 ×6 2 Andalucía
Guti´
errez-Gordillo et al. (2019a) 18 FI; RDI Guara, Lauranne, Marta 10 8 ×6 1 Andalucía
Guti´
errez-Gordillo et al. (2020) 9 FI; SDI Guara, Lauranne, Marta 6 8 ×6 1 Andalucía
Lipan et al. (2020) 4 FI; RDI; SDI Vairo 8 7 ×6 1 Andalucía
L´
opez-L´
opez et al. (2018) 9 FI; RDI; SDI Guara 5 7 ×6 2 Andalucía
Ma˜
nas et al. (2014) 24 FI; RDI; SDI Ferragn`
es 9 7 ×5 4 Castilla La Mancha
Moldero et al. (2021) 12 FI; RDI; SDI Guara 8 7 ×6 3 Andalucía
Puerto et al. (2013) 8 FI; RDI Guara 12 6 ×6 2 Murcia
Romero et al. (2004) 5 FI; RDI Cartagenera 15 7 ×5 1 Murcia
Included is the number of observations (N obs), irrigation treatments applied, almond cultivars, tree age and spacings, number of years from which data were extracted
(N years), and region. Full Irrigation (FI), Regulated Decit Irrigation (RDI), Sustained Decit Irrigation (SDI).
Table 2
List of moderators for almond yield recorded from eld experiments conducted
in Spain from 1990 to 2019.
Moderator Description
Almond cultivar Cartagenera, Ferragn`
es, Guara, Lauranne, Marcona, Marta,
Vairo
Irrigation
strategy
FI, RDI, SDI
Water decit Control, Moderate, Severe
Soil depth Shallow (<80 cm), Deep (>80 cm)
Full Irrigation (FI), Regulated Decit Irrigation (RDI), Sustained Decit Irriga-
tion (SDI). Water decit is computed as the ratio between the irrigation dose
applied to the control treatment and that applied to the decit treatments:
Moderate (ratio between 0.55 and 0.99), Severe (ratio <0.55).
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
4
control treatment, allowed for classifying the decit irrigation treat-
ments into moderate (ratio between 0.55 and 0.99) and severe (ratio <
0.55). Fig. 1 shows the boxplots of yields and relative yields for the
different watering regimes considered (the combinations of stress level
and irrigation strategy).
Both yield and relative yield data met the normality and homosce-
dasticity assumptions according to Shapiro-Wilks and Bartlett’s tests (p-
values >0.05), so an ANOVA was performed to assess the signicance of
the effects of both irrigation strategy and water stress level (Fig. 1).
Yields from severe decit treatments were signicantly lower than those
from the control and moderate decit treatments, independently of the
irrigation strategy (Fig. 1a). However, a moderate SDI treatment
signicantly reduced the relative yield with respect to the control
treatment, but the RDI strategy did not (Fig. 1b).
A positive and signicant correlation between the water received
(rainfall +irrigation) by the almond trees and their yield was observed
(Fig. 2a). This relationship can be expressed as yield = − 0.0009 ×
(Rainfall +irrigation)
2
+3.3433 ×(Rainfall +irrigation) – 489.55, and
its coefcient of determination (R
2
) was 0.5761 (p-value <0.01). Ac-
cording to this equation, an amount around 1100 mm of water per year
would be needed to obtain 2000 kg/ha of almonds. In terms of irrigation
supply, the dataset suggests that the maximum yield would be obtained
with 800 mm of irrigation water per year (Fig. S1). Moreover, when the
yield and the water received were relativized to the corresponding full
irrigation control (Fig. 2b), the dataset suggests that no yield reduction
could be expected if the water received is more than 85% that of the
control.
Table 3
Minimum, maximum, and average values for the variables included in the
dataset of decit irrigation studies in Spain.
Variable N Minimum Maximum Average No
data
Annual rainfall (mm) 129 230 802 453 15
Rainfall over the growing
season (mm)
96 116 391 220 48
Irrigation applied (mm) 144 7 985 408 0
Annual rainfall +
irrigation (mm)
129 277 1958 1042 15
Reference
evapotranspiration
(mm)
120 855 1400 1165 24
Yield (kg/ha) 144 352 3329 1684 0
Relative yield (%) 144 30 128 87 0
Number of fruits per tree 111 2308 13280 6312 33
Kernel weight (g) 111 0.9 1.7 1.3 33
Fig. 1. Boxplots of (a) yield and (b) the relative yield (percentage of yield of a given decit irrigation treatment over the yield in the control) as a function of the
watering regime and stress level. Different letters on the boxes indicate signicant differences among treatments according to the Tukey’s test (p <0.05). RDI
=Regulated decit irrigation, SDI =Sustained decit irrigation.
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
5
The variation in yield among studies was not only dependent on the
water received, but also on the almond cultivar (Fig. 3). In this dataset,
“Guara” and “Lauranne” showed the highest yields, whereas “Ferragn`
es”
showed the lowest yields. However, a high variability was observed,
likely caused by the different conditions (agrometeorological, soil) and
fertigation practices among studies (Fig. 3).
To better understand this situation, generalized linear models were
built separately for each cultivar to describe the relationship between
water received (rainfall +irrigation) and yield (Table 4). Except for the
cultivars “Marcona”, for which there were no rainfall data available, and
“Lauranne”, the slopes of the tted models were signicantly different
from zero (Table 4). The intercept was not signicant for “Ferragn`
es”
and “Vairo”. In addition, the regression coefcients were lower than 0.6,
except for “Vairo” and “Cartagenera” (Table 4). Therefore, a
Fig. 2. Relationships between the water received (rainfall +irrigation) and almond yield (a) and between the water received and almond yield with respect to yields
obtained in the corresponding full irrigation (FI) control (b).
Fig. 3. Relationship between the amount of water received (annual rainfall +
annual irrigation) and almond yield as a function of the cultivar.
Table 4
Parameters of the models tted to the relationships between water received
(rainfall +irrigation) and yield for each almond cultivar considered in the
dataset.
Cultivar Intercept p-value Slope p-value R
2
Cartagenera 346.5588 0.03978 1.1675 0.00299 0.9513
Ferragnes -2.5163 0.9925 1.7115 0.000103 0.37
Guara 508.1408 0.0285 1.4036 <0.0001 0.4522
Lauranne 2093.5514 <0.0001 0.2199 0.358 -0.0066
Marcona Rainfall data are not available
Marta 809.9124 0.002042 1.0033 0.000459 0.3817
Vairo 490.7189 0.1237 1.1265 0.0434 0.8726
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
6
heterogeneity in the yield response to water received was observed
among cultivars, although this effect was negligible for “Lauranne”. This
can be due to the magnitude of the yields observed in the dataset (very
high in “Lauranne” when compared to the rest of the cultivars).
When plotted as a function of the irrigation strategy, the highest
yields corresponded to the control treatments and, in some cases, to the
moderate decit treatments (both RDI and SDI), whereas the lowest
yields always corresponded to the treatments that imposed a severe
water decit (Fig. 4).
3.2. Water production function and water productivity
The yield response to applied irrigation (AI) for the treatments
included within this dataset is shown in Fig. 5. Yields increased from
about 500 kg/ha with AI of 50 mm to nearly 2700 kg/ha with the
1050 mm of applied irrigation, and then it seemed to stabilize. Kernel
yield did not decline within the limits of applied irrigation considered in
the current study (Fig. 5). To quantify water productivity levels as a
function of applied irrigation, a second-order polynomial expression was
tted to the mean yield versus AI (Fig. 5), and its derivative, the mar-
ginal water productivity, was computed and plotted against AI (Fig. 6).
Water productivity reached a maximum value of 0.34 kg/m
3
when no
irrigation was applied, and decreased to zero at 1260 mm, becoming
negative as AI increased (Fig. 6). The yield response to AI and the
marginal water productivities for regulated and sustained decit irri-
gation strategies are shown in the Supplementary Material (Figs. S1 and
S2, respectively).
3.3. Meta-analysis
The effect of water decit (combining RDI with SDI treatments for all
decit levels) on yield (kg/ha) was assessed by means of a forest plot
combining the 15 studies included in the database (Fig. 7). This graph
indicates that decit treatments yielded 84–87% of what their respective
well-irrigated controls yielded. The condence interval is quite narrow,
varying between 0.85 and 0.89 in the case of a xed effects model, and
between 0.79 and 0.89 in the case of a random effects model (Fig. 7).
Finally, the heterogeneity indicators showed a large variability between
studies (I
2
=77%). Cochran’s Q indicator took a value of 61.54 (p-value
<0.0001), indicating that the effect size differed among studies. The
funnel plot revealed the presence of a certain publication bias (Fig. S3);
however, a regression test of funnel plot asymmetry provided an inter-
cept of −0.1164 with a p-value of 0.2519, suggesting that the estimated
effects were robust.
Fig. 7 clearly shows that the control treatment favoured almond yield
over decit irrigation regardless of soil depth. However, the rate at
which this yield increase occurred was different in deep (Random effects
model =0.84) than in shallow soils (Random effects model =0.76). This
suggests that decit irrigation in shallow soils decreases yield to a
greater extent than in the case of deeper soils (the exact soil depth in
each study incorporated within this meta-analysis is unknown),
although the low number of studies carried out on shallow soils does not
allow for drawing sound conclusions.
When RDI was compared against SDI, regardless of the severity of the
water stress applied, the number of studies was reduced, and conclu-
sions were not clear (Fig. 8). In fact, if a xed effects model is consid-
ered, SDI led to a 3% higher yield compared to RDI. However, using a
random effects model, the result was the opposite (Fig. 8). The vari-
ability between the studies was very high (I
2
=73%). Cochran’s Q in-
dicator obtained a value of 26.07 (p-value =0.0005), indicating that the
effect size differed among studies. The funnel plot did not reveal the
presence of publication bias (Fig. S4). In addition, almond yield
beneted slightly under RDI in both deep and shallow soils. The rate at
which this increase in yield occurred was similar in deep (Random
Fig. 4. Relationship between the amount of water received (annual rainfall +
annual irrigation) and almond yield as a function of the watering regime and
severity of water stress.
Fig. 5. Kernel yield versus applied water with the best-t second order poly-
nomial expression. The symbols represent mean almond yields by irrigation
intervals, represented by their average value. The vertical and horizontal error
bars represent the standard deviation of the means. The yield-water response
functions derived by Moldero et al. (2021) and Goldhamer and Fereres (2017)
have also been plotted for comparison purposes.
Fig. 6. Water productivity versus applied water calculated as the derivative of
a best-t second order polynomial expression tted to the average yield from
the treatments included in the dataset. The marginal productivity-water func-
tions derived by Moldero et al. (2021) and Goldhamer and Fereres (2017) have
also been plotted for comparison purposes.
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
7
effects model =0.97) and in shallow soils (Random effects model =
0.94). Considering only a moderate water decit, the differences be-
tween applying this decit in a sustained manner throughout the season,
or in certain phases of the crop cycle, were practically nil (Fig. S5). This
may be because the analysis only considered the stress for the whole
season, which could be masking some other effects. However, when
considering severe water stress, the meta-analysis seemed to indicate
that it is more advisable to apply RDI, as yield would be less affected
(Fig. S6). However, it should be noted that the latter two analyses
include fewer studies.
4. Discussion
The quantitative analysis performed to evaluate the agronomic
response of adult almond plantations grown in Spain to different levels
of water stress revealed a wide range of almond yields (352–3329 kg/
ha) for a wide range of irrigation volumes applied (7–985 mm)
(Table 3). This variability was partly because the relationship between
applied water and yield was not straightforward. It was affected by the
soil type, the soil water content at the beginning of the season, and the
prevailing evaporative demand, which can vary signicantly among
regions. Evapotranspiration was the pertinent indicator for this analysis,
but unfortunately, it was seldom measured, so the applied water was
used here as a proxy for the actual water used by the almond trees.
The comparison of these results with those obtained in eld
experiments conducted in California, the main almond producing area
in the world, showed that the maximum yields obtained in the eld trials
carried out in Spain coincided with the minimum average yields ob-
tained by Goldhamer and Fereres (2017) in a 5-year trial carried out in
an adult almond orchard subjected to 10 irrigation levels. In this sense, it
is important to highlight that the maximum irrigation volumes applied
in the experiments carried out in Spain were close to the minimum
volumes used in the study performed by Goldhamer and Fereres (2017),
where up to 1350 mm of irrigation depths were applied and maximum
(5-year mean) yields close to 4000 kg/ha of almonds were obtained. The
incorporation of an irrigation treatment with over-irrigation in the
production function obtained in this study (Fig. 5) did not lead to any
increase in kernel yield, suggesting that the volumes of water applied in
control (well-irrigated) treatments were suitable for reaching potential
yields for the plant material, crop management, and agroclimatic con-
ditions prevailing in Spain.
The high almond yields achieved in California (~4000 kg/ha)
resulted from decades of crop intensication (Goldhamer and Fereres,
2017); with a similar situation in Australia, the second greatest almond
producer worldwide, whose almond growing sector employs the culti-
vars and cultural practices used in California (Thorp et al., 2021).
Conversely, in Spain, these levels of crop intensication with irrigation
inputs that can exceed 1300 mm per year (Goldhamer and Fereres,
2017) are not expected due to the reduced availability of irrigation
water in most of the inland areas into which almond plantations are
Fig. 7. Summary effect sizes of treatment (well-
watered control against decit irrigation) for
the considered dataset of studies. The moder-
ator “soil depth” is considered for separating
the studies. Horizontal bars represent 95%
condence intervals (CI), which are also shown
between brackets. Vertical solid line represents
a null effect. Ratio of means (ROM) indicates
the ratio of the average yield on the decit
treatment to that of the control treatment.
Weights indicate the relevance of each study to
the xed or random effects model. Favours
control and Favours decit zones in the graph
indicate when the yield from a given study were
higher for the control or the decit irrigation
treatment, respectively. SD =standard devia-
tion; df =degrees of freedom.
Fig. 8. Summary effect sizes of treatment
(regulated decit irrigation, RDI, versus sus-
tained decit irrigation, SDI) for the considered
dataset of studies. Horizontal bars represent
95% condence intervals (CI), which are also
shown between brackets. Vertical solid line
represents a null effect. Ratio of means (ROM)
indicates the ratio of the average yield on the
decit treatment to that of the control treat-
ment. Weights indicate the relevance of each
study to the xed or random effects model.
Favours RDI and Favours SDI zones in the graph
indicate when the yield from a given study were
higher for the Regulated Decit Irrigation (RDI)
or the Sustained Decit Irrigation (SDI) treat-
ment, respectively. SD =standard deviation; df
=degrees of freedom.
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
8
expanding. In this sense, irrigation water allocations commonly range
between 250 mm and 600 mm per year (Moldero et al., 2021), well
below the water requirements needed for intensive adult almond plan-
tations. Under the premises of irrigating almond orchards with decit
allocations, the results obtained in the current study conrmed the good
productive performance of almond trees under conditions of moderate
water decit, as very low yield penalties (7–9%) were observed when
compared with the control treatments (Fig. 1). An important aspect of
decit irrigation management in almond orchards that continues to
generate uncertainty is the convenience of using regulated (RDI) versus
sustained (SDI) decit irrigation strategies. The results obtained in this
work suggest, although not denitely, a certain advantage of using RDI
strategies over SDI in almond trees. In absolute terms, the mean kernel
yield between RDI and SDI treatments did not differ signicantly
regardless of the level of water decit applied (Fig. 1). However, when
relative yields were analyzed, SDI differed from the control for both
levels of water decit (moderate and severe). In contrast, the RDI
treatment only differed from the control when the water decit was
severe (Fig. 1). Despite being signicant, the reduction in yield for a
moderate water decit applied through SDI was only 9% with respect to
the control; therefore, the analysis performed could not robustly conrm
that this irrigation strategy causes an appreciable decrease in almond
yield. The meta-analysis (Fig. 8) suggested a certain production
advantage for RDI over SDI, when simultaneously considering both
moderate and severe water decit. Therefore, the current study cannot
provide a denite answer for the rst question raised about which irri-
gation strategy is more appropriate for a given water allocation below
the total almond water requirements, as the current study only suggests
slight yield improvements for RDI.
RDI strategies in almond trees have mostly consisted of applying a
certain level of water decit during the kernel-lling stage, considered
the most drought-resistant phenological stage in almond trees (Girona
et al., 2005). However, some studies observed yield reductions when
water decit was applied during this stage (Egea et al., 2013; Goldhamer
et al., 2006; Goldhamer and Viveros, 2000; Hutmacher et al., 1994),
while in other studies, yield was unaffected by water decits applied
during kernel-lling (Egea et al., 2010, 2009; Goldhamer and Fereres,
2004; Puerto et al., 2013). These controversial results seem to be related
to the level of water stress reached by trees during this period, as stem
water potential values lower than −2 MPa during kernel-lling have
been suggested to cause yield losses (García-Tejero et al., 2018) due to
variations in kernel weight (Girona et al., 2005). Despite this evidence,
the analysis conducted in this work indicates that applying water
shortages only during the grain lling stage rather than spreading it
proportionally throughout the crop cycle is not clearly justied.
This leads to the second question about what yield loss can be ex-
pected for a given water allocation when compared to a well-watered
orchard. The results obtained in the current study are not conclusive
on whether the cultivars evaluated differed in their productive response
to decit irrigation. Although some differences were observed in the
relationships between water input and yield of each cultivar, the vari-
ability in the ranges of water applied among the different experiments
made it difcult to obtain sound conclusions regarding the tolerance of
the cultivars to water stress. On the other hand, although it has some-
times been considered that shallow soils are better for the application of
RDI strategies in woody crops, due to the adequate timing of water stress
application that is needed in an RDI strategy (Girona et al., 2003), in
almond trees it seems that the crop response to RDI strategies is poorer
in shallow soils compared to deeper soils. However, the low number of
studies developed on shallow soils does not allow for obtaining sound
conclusions (Fig. 7). Nevertheless, the current study indicated that, for a
moderate water decit, 7–9% yield reductions can be expected with
respect to a well-watered orchard, while for a severe water decit, yield
decrease could be up to 33%.
The applied irrigation-yield response function obtained in this
analysis comprising multiple cultivars, irrigation treatments, and
experimental conditions (Fig. 5) was similar to that obtained by Moldero
et al. (2021) in a 6-year trial carried out in southern Spain on almond
trees cv. “Guara”. By comparing both production functions, it can be
deduced that Spanish cultivars have a similar productive response to
irrigation under the agroclimatic and management conditions of the
Spanish almond orchards, with maximum kernel yields obtained with
irrigation water allocations of about 1000 mm per growth cycle. How-
ever, when these production functions were compared with that ob-
tained in California (Goldhamer and Fereres, 2017), it was observed that
Californian almond plantations continued to increase kernel yields
above 1000 mm of irrigation water applied, reaching maximum yields
close to 4000 kg/ha with irrigation inputs of about 1250 mm per growth
cycle.
The marginal productivity of irrigation water decreased continu-
ously with any irrigation water input, both in the relationship obtained
by Moldero et al. (2021) and in the one obtained in this analysis (Fig. 6).
This pattern has also been observed in previous studies conducted with
other cultivars (e.g. cv. “Marta”) (Egea et al., 2010). While the almonds
cv. “Guara” needed irrigation inputs close to 1000 mm for marginal
water productivity to be zero, in the meta-analysis carried out in this
study, irrigation inputs close to 1200 mm were needed for marginal
irrigation productivity to be zero. In any case, the comparison with the
irrigation water productivities obtained in California shows the low
productivity of irrigation inputs in Spain above 800 mm/year, lower
than 0.1 kg/m
3
, while maximum marginal productivities of irrigation
water of around 0.3 kg/m
3
were observed for irrigation inputs of
1100 mm/year in California (Goldhamer and Fereres, 2017). From these
data, it can be concluded that higher irrigation water productivities than
those observed in the Spanish trials are possible for high irrigation water
allocations. Therefore, it seems that almond productive response de-
pends mainly on the percentage of ETc supplied annually, answering the
third question raised in the introduction of the current study. However,
as irrigation water allocations above 700–800 mm are not expected in
Spain and over the Mediterranean Sea Basin, the scientic and techno-
logical challenge for almond cultivation is to increase the marginal
productivity for moderate irrigation allocations to the levels observed in
Californian almond orchards for notably higher irrigation water allo-
cations. This could be achieved not only by means of improved irrigation
technologies and scheduling, but also by optimizing the overall agro-
nomic management with particular attention to fertilization regimes
and pruning operations. The challenge of increasing marginal produc-
tivity should also consider the sustainability component for minimizing
contamination risks, ensuring soil conservation, and considering the
common trend of increasing organic farming cultivation.
5. Final considerations and recommendations
Despite the large variability observed in the pooled data set (because
of the wide range of studied conditions such as soil types, cultivars,
climatic conditions, or tree sizes, among others), the quantitative anal-
ysis conducted allowed us to derive some general trends:
•In Spain, under semi-arid Mediterranean conditions, almond yield
increases with irrigation water application with an expected yield of
about 2500 kg/ha for around 1000 mm of irrigation water applied.
•The yield reduction observed when water allocation decreased in
comparison to fully irrigated trees was mostly due to the severity of
the water stress suffered by trees, and to a lesser extent due to the
irrigation strategy implemented.
•The application of a regulated decit irrigation strategy, rather than
a sustained decit one, only showed some advantage when water
stress was moderate.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
J.M. Mir´
as-Avalos et al.

Agricultural Water Management 279 (2023) 108208
9
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Data availability
Data will be made available on request.
Acknowledgements
This research was mainly carried out within the framework of Project
RIDECORED funded by the Agencia Estatal de Investigaci´
on
[AGL2017–90666-REDC and AGL2015–66141-R] with FEDER co-funds.
Additional support was provided by CajaMar Caja Rural [RTC-
2017–6365-2], Junta de Andalucía and FEDER [AVA.AVA2019.051],
European Commission and PRIMA [grant number 1813], and Junta de
Castilla La Mancha and FEDER [project SBPLY/17/180501/000357].
The initial suggestions and support on appraising this research by Dr. F.
Orgaz (IAS-CSIC) is also acknowledged.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.agwat.2023.108208.
References
Balduzzi, S., Rücker, G., Schwarzer, G., 2019. How to perform a meta-analysis with R: A
practical tutorial. Evid. Based Ment. Health 22, 153–160. https://doi.org/10.1136/
ebmental-2019-300117.
Borenstein, M., Hedges, L.V., Higgins, J.P.T., Rothstein, H.R., 2009. Introduction to
Meta-Analysis. John Wiley & Sons, Ltd,, Chichester, UK. https://doi.org/10.1002/
9780470743386.
CHG, 2015, Plan hidrol´
ogico de la demarcaci´
on hidrogr´
aca del Guadalquivir. Segundo
ciclo de planicaci´
on: 2015–2021 [WWW Document]. Minist. Agric. Aliment. y
Medio Ambient. URL 〈https://www.chguadalquivir.es/segundo-ciclo-guadalquivir〉
(accessed 2.10.22).
Egea, G., Gonz´
alez-Real, M.M., Baille, A., Nortes, P.A., S´
anchez-Bel, P., Domingo, R.,
2009. The effects of contrasted decit irrigation strategies on the fruit growth and
kernel quality of mature almond trees. Agric. Water Manag 96. https://doi.org/
10.1016/j.agwat.2009.06.017.
Egea, G., Nortes, P.A., Gonz´
alez-Real, M.M., Baille, A., Domingo, R., 2010. Agronomic
response and water productivity of almond trees under contrasted decit irrigation
regimes. Agric. Water Manag 97. https://doi.org/10.1016/j.agwat.2009.09.006.
Egea, G., Nortes, P.A., Domingo, R., Baille, A., P´
erez-Pastor, A., Gonz´
alez-Real, M.M.,
2013. Almond agronomic response to long-term decit irrigation applied since
orchard establishment. Irrig. Sci. 31. https://doi.org/10.1007/s00271-012-0322-8.
Egger, M., Smith, G.D., 1997. Meta-analysis: Potentials and promise. BMJ 315,
1371–1374. https://doi.org/10.1136/bmj.315.7119.1371.
ESYRCE, 2015, Encuesta sobre supercies y rendimientos de cultivos de Espa˜
na.
ESYRCE, 2020, Encuesta sobre supercies y rendimientos de cultivos de Espa˜
na.
Fan, Y., Wang, C., Nan, Z., 2018. Determining water use efciency of wheat and cotton: A
meta-regression analysis. Agric. Water Manag. https://doi.org/10.1016/j.
agwat.2017.12.006.
FAOSTAT, 2021, Food and Agriculture Organization of the United Nations. Food and
Agriculture Data [WWW Document].
Fereres, E., Soriano, M.A., 2007. Decit irrigation for reducing agricultural water use.
J. Exp. Bot. 58, 147–159. https://doi.org/10.1093/jxb/erl165.
García-Tejero, I.F., Moriana, A., Rodríguez-Pleguezuelo, C.R., Dur´
an-Zuazo, V.H.,
Egea, G., 2018. Sustainable Decit-Irrigation Management in Almonds (Prunus dulcis
L.): Different Strategies to Assess the Crop Water Status. In: García-Tejero, I.F.,
Dur´
an-Zuazo, V.H. (Eds.), Water Scarcity and Sustainable Agriculture in Semiarid
Environment. Tools, Strategies, and Challenges for Woody Crops. Elsevier, London,
p. 560. https://doi.org/10.1016/B978-0-12-813164-0.00012-0.
García-Tejero, I.F., Guti´
errez Gordillo, S., Souza, L., Cuadros-Tavira, S., Dur´
an Zuazo, V.
H., 2019. Fostering sustainable water use in almond (Prunus dulcis Mill.) orchards in
a semiarid Mediterranean environment. Arch. Agron. Soil Sci. 65, 164–181. https://
doi.org/10.1080/03650340.2018.1492113.
García-Tejero, I.F., Lipan, L., Guti´
errez-Gordillo, S., Dur´
an Zuazo, V.H., Janˇ
co, I.,
Hern´
andez, F., Rodríguez, B.C., Carbonell-Barrachina, ´
A.A., 2020. Decit irrigation
and its implications for HydroSOStainable almond production. Agronomy 10, 1–20.
https://doi.org/10.3390/agronomy10111632.
Girona, J., Marsal, J., Mata, M., Arbon´
es, A., Miravete, C., 1997. Evaluation of almond
(Amygdalus communis L.) seasonal sensitivity to water stress. physiological and yield
responses. Acta Hortic. 449, 489–496. https://doi.org/10.17660/
ActaHortic.1997.449.68.
Girona, J., Mata, M., Arbon`
es, A., Alegre, S., Rufat, J., Marsal, J., 2003. Peach tree
response to single and combined regulated decit irrigation regimes under shallow
soils. J. Am. Soc. Hortic. Sci. 128, 432–440. https://doi.org/10.21273/
jashs.128.3.0432.
Girona, J., Mata, M., Marsal, J., 2005. Regulated decit irrigation during the kernel-
lling period and optimal irrigation rates in almond. Agric. Water Manag 75,
152–167. https://doi.org/10.1016/j.agwat.2004.12.008.
Goldhamer, D.A., Fereres, E., 2004. Irrigation scheduling of almond trees with trunk
diameter sensors. Irrig. Sci. 23, 11–19. https://doi.org/10.1007/s00271-003-0088-
0.
Goldhamer, D.A., Fereres, E., 2017. Establishing an almond water production function
for California using long-term yield response to variable irrigation. Irrig. Sci. 35,
169–179. https://doi.org/10.1007/s00271-016-0528-2.
Goldhamer, D.A., Viveros, M., 2000. Effects of preharvest irrigation cutoff durations and
postharvest water deprivation on almond tree performance. Irrig. Sci. 19, 125–131.
https://doi.org/10.1007/s002710000013.
Goldhamer, D.A., Viveros, M., Salinas, M., 2006. Regulated decit irrigation in almonds:
Effects of variations in applied water and stress timing on yield and yield
components. Irrig. Sci. 24, 101–114. https://doi.org/10.1007/s00271-005-0014-8.
Guti´
errez-Gordillo, S., Dur´
an-Zuazo, V.H., García-Tejero, I., 2019a. Response of three
almond cultivars subjected to different irrigation regimes in Guadalquivir river
basin. Agric. Water Manag 222, 72–81. https://doi.org/10.1016/j.
agwat.2019.05.031.
Guti´
errez-Gordillo, S., García-Tejero, I.F., García-Escalera, A., Galindo, P., Arco, M.,
del, C., Dur´
an Zuazo, V.H., 2019b. Approach to yield response of young almond trees
to decit irrigation and biostimulant applications. Horticulturae 5, 38. https://doi.
org/10.3390/horticulturae5020038.
Guti´
errez-Gordillo, S., Dur´
an Zuazo, V.H., Hern´
andez-Santana, V., Gil, F.F., Escalera, A.
G., Amores-Agüera, J.J., García-Tejero, I.F., 2020. Cultivar dependent impact on
yield and its components of young almond trees under sustained-decit irrigation in
semi-arid environments. Agronomy 10, 1–15. https://doi.org/10.3390/
agronomy10050733.
Harrison, F., 2011. Getting started with meta-analysis. Methods Ecol. Evol. 2, 1–10.
https://doi.org/10.1111/j.2041-210X.2010.00056.x.
Higgins, J.P.T., Thompson, S.G., Deeks, J.J., Altman, D.G., 2003. Measuring
inconsistency in meta-analyses. Br. Med. J. https://doi.org/10.1136/
bmj.327.7414.557.
Hutmacher, R.B., Nightingale, H.I., Rolston, D.E., Biggar, J.W., Dale, F., Vail, S.S.,
Peters, D., 1994. Growth and yield responses of almond (Prunus amygdalus) to trickle
irrigation. Irrig. Sci. 14, 117–126.
Lipan, L., Cano-Lamadrid, M., Hern´
andez, F., Sendra, E., Corell, M., V´
azquez-Araújo, L.,
Moriana, A., Carbonell-Barrachina, ´
A.A., 2020. Long-term correlation between water
decit and quality markers in hydrosostainable almonds. Agronomy 10, 1–22.
https://doi.org/10.3390/agronomy10101470.
L´
opez-L´
opez, M., Espadafor, M., Testi, L., Lorite, I.J., Orgaz, F., Fereres, E., 2018. Yield
response of almond trees to transpiration decits. Irrig. Sci. 36, 111–120. https://
doi.org/10.1007/s00271-018-0568-x.
Ma˜
nas, F., L´
opez-Fuster, P., L´
opez-Urrea, R., 2014. Effects of different regulated and
sustained decit irrigation strategies in almond production. Acta Hortic. 1028,
391–394. https://doi.org/10.17660/ActaHortic.2014.1028.64.
Mitchell-McCallister, D., Cano, A., West, C., 2020. Meta-analysis of crop water use
efciency by irrigation system in the Texas High Plains. Irrig. Sci. 38, 535–546.
https://doi.org/10.1007/s00271-020-00696-x.
Moldero, D., L´
opez-Bernal, ´
A., Testi, L., Lorite, I.J., Fereres, E., Orgaz, F., 2021. Long-
term almond yield response to decit irrigation. Irrig. Sci. 39, 409–420. https://doi.
org/10.1007/s00271-021-00720-8.
Moldero, D., L´
opez-Bernal, ´
A., Testi, L., Lorite, I.J., Fereres, E., Orgaz, F., 2022. Almond
responses to a single season of severe irrigation water restrictions. Irrig. Sci. 40.
https://doi.org/10.1007/s00271-021-00750-2.
Puerto, P., Domingo, R., Torres, R., P´
erez-Pastor, A., García-Riquelme, M., 2013. Remote
management of decit irrigation in almond trees based on maximum daily trunk
shrinkage. Water Relat. yield. Agric. Water Manag 126, 33–45. https://doi.org/
10.1016/j.agwat.2013.04.013.
R Core Team, 2021, R: A language and environment for statistical computing.
Romero, P., Botia, P., Garcia, F., 2004. Effects of regulated decit irrigation under
subsurface drip irrigation conditions on vegetative development and yield of mature
almond trees. Plant Soil 260, 169–181. https://doi.org/10.1023/B:
PLSO.0000030193.23588.99.
Schwarzer, G., 2007. Meta: an R package for meta-analysis. R. N. 7, 40–45.
Schwarzer, G., Carpenter, J.R., Rücker, G., 2015. Meta-Analysis with R, Use R! Springer
International Publishing,, Cham. https://doi.org/10.1007/978-3-319-21416-0.
Shorten, A., Shorten, B., 2013. What is meta-analysis. Evid. Based Nurs. 16 (1), 3–4.
https://doi.org/10.1136/eb-2012-101118.
Sterne, J.A.C., Becker, B.J., Egger, M., 2006. The funnel plot. In: Publication Bias in
Meta-Analysis. John Wiley & Sons, Ltd, Chichester, UK, pp. 73–98. https://doi.org/
10.1002/0470870168.ch5.
Thorp, G., Smith, A., Traeger, D., Jenkins, B., Granger, A., van den Dijssel, C., Barnett, A.,
Blattmann, M., P´
eri´
e, E., Mangin, V., Snelgar, P., Kolesik, J., Wirthensohn, M., 2021.
Selective limb removal pruning and reective ground covers improve light and crop
distributions in the lower zone of ‘Nonpareil’ almond trees but not total yield. Sci.
Hortic. 289. https://doi.org/10.1016/j.scienta.2021.110508, 110508.
J.M. Mir´
as-Avalos et al.