Estimation of the Water Reserve in the Soil Using GIS and Its Application in Irrigated Olive Groves in Jaen, (Spain)

Abstract

Soil water reserves are very important for irrigation scheduling in arid and semiarid conditions. In these regions, irrigated olive groves could save water and improve water resource management if the spatial and temporal patterns of water reserve were known. In this work, a large region characterized by olive monoculture located in Jaén, Spain was studied, as well as its water requirements and the evolution of the water reserve in the soil according to the time of year by using public data sources. In this way, climatic data, NDVI monthly mean, soil type, physicochemical and hydrological properties of the soil have been integrated in GIS by means of easy-to-use techniques. The results obtained from both the water balance and the evolution of water in the soil show that in the region studied, it is not advisable to manage a single irrigation schedule, as is currently the case, and that it is necessary to implement different irrigation times and strategies depending on the location of the plot. These results can serve as a basis for the design of specific irrigation schedules on daily, hourly or real-time time scales depending on the availability of data and the degree of precision sought.

1. Introduction

According to World Bank data, agriculture is one of the main economic sectors in the world, with an aggregate value of more than USD 3.68 trillion [1]. However, its importance goes far beyond this. Sustainable agriculture generates income; can help overcome many of the challenges facing humanity, such as the fight against hunger or biodiversity conservation; and can guarantee production, since the demand for food and agricultural products will continue to grow over the next decade, along with their productivity, trade and sustainable use [2].

On the other hand, the climate change outlook predicts an increase in average temperatures by 4 °C, and a decrease in precipitation in arid and semiarid midlatitude regions [3]. This increase in temperature would produce an increase in evapotranspiration and a change in flowering dates that could impact irrigation strategies and olive cultivar variety choice [4]

Irrigation is very important in arid or semiarid environments where seasonal drought limits agricultural production [5]. Irrigated crops represent approximately 21% of the total area of cultivated land worldwide (approximately 306.7 million hectares) and a production of approximately 40% of the world’s food [3]. It is estimated that 69% of water is used by irrigated agriculture globally [6]. From the point of view of productivity, it is estimated that in average terms, an irrigated hectare produces six times more than a rainfed one [5]. In the case of Spain, irrigated agriculture occupies approximately 20% of the total cultivated area and is responsible for 55% of the final agricultural production (FAP) [7].

Olive groves are one of the major permanent crops worldwide. The olive surface represents 25% of the total permanent cultivated area worldwide, with 11.6 million hectares distributed in 63 countries around the five continents and extending over slightly more than 0.25% of the total cultivated land area. Of the total global olive grove area, 70% is distributed in rainfed cultivation, while the remaining 30% is irrigated [8].

Worldwide, one of the most characteristic regions of olive monoculture is the province of Jaen, in southern Spain. It is the world’s largest producer of olive oil, so a large part of Jaen’s economy is based on olive monoculture (in fact, the olive grove plantations occupy the majority of the territory) [9]. It represents more than 25% of the total olive grove area in Spain and 42% of Andalusia’s total cultivated land. Its production on average is approximately 50% of the Spanish total of olive oils and more than 20% of the world’s total [10]. This territory has an area of approximately 13,489 km², so due to both its size and biodiversity, it presents different agronomic conditions. The olive plantation surface in this location is around 5.93 × 10⁵ hectares, of which 2.88 × 10⁵ are rainfed and 3.04 × 10⁵ are irrigated [11]. Most of the irrigated area is arranged in broad frames of traditional cultivation (88% of the irrigated olive grove area versus 12% of intensive olive groves) [12].

This study area is located in the arid and semiarid zones of the Mediterranean region, so water availability is a major constraint in crop production due to low rainfall and long periods of summer drought [13]. Although the olive crop shows good adaptation to drought, it responds very favorably to irrigation [14].

The incorporation of irrigation practices in the olive grove has allowed for the transformation of a rainfed crop with low profitability to a high-value crop that increases the farmer’s income, making it possible to obtain the maximum economic benefit and employment of labor per unit volume of water applied. Furthermore, in the area studied, the productivity of irrigation water used in olive cultivation is higher in socioeconomic terms than that of most traditional irrigated crops [15,16].

Water availability for olive groves is commonly below irrigation needs. Then, though they are irrigated, trees would be in water stress conditions during significant periods of the season. This deficit irrigation (DI) could affect the most sensitive phenological stages of the olive crop; flowering, fruit set and oil accumulation [17].

Thus, water deficit conditions have an impact on water stress in vegetation, which in turn has negative effects on crop productivity and can even affect plant survival in young groves [18].

Another common feature to most olive grove irrigation farms is the implementation of localized irrigation systems and a fairly uniform irrigation schedule, which depends on both water availability and a fixed maximum annual volume of 1500 m³/ha in traditional 10 × 10 m plantations [12].

Generally, the less dense an olive grove is, the lower the level of water stress, and therefore, the greater the irrigation water savings [19]. This lower tree density could permit deficit irrigation scheduling, which would allow for around 50% of water to be saved with mild or moderate water stress. This water stress level likely would not reduce fruit yield [20]. However, this irrigation scheduling requires the accurate measurements of the water reserve in the soil. Remote sensing techniques and geographic information systems (GISs) have been used to determine the evolution of water in the soil and to establish water management, namely both its use and productivity. Remote sensing makes it possible to study the spatial distribution and temporal evolution of biophysical properties or characteristics related to crop development by using a sequence of images [21]. The evaluation of crop variability through satellite images has been widely employed by correlating spectral information with biological processes of the terrestrial ecosystem [22]. On the other hand, geographic information systems (GISs) allow for the management of geographic information and its association with any type of descriptive information [23].

The use of the water reserve in soil in irrigation schedules and calendars means a lower expenditure of available water resources. However, we find that usually in large irrigated areas of the same crop, the same irrigation schedule is used without taking into consideration this reserve, which has a negative impact on the management of generally scarce water resources.

The aim of this work is to propose an easily accessible methodology to accurately determine the evolution of soil water content for a given crop (olive grove) over a wide region. The objective will be to obtain the best possible information to elaborate the most accurate irrigation schedules possible in order to improve the efficiency of irrigation water use.

2. Materials and Methods

2.1. Description of the Study Area

The study area comprises the province of Jaen (Spain, Figure 1) and covers an area of 13,489 km², in which olive groves are the most important Mediterranean fruit crop in the surface area [11]. This crop is the main socioeconomic engine in the province [16-Molina Moral 2022] and is constituted by 9 agricultural zones [24].

2.2. Data Source

The temporal evolution of the water reserve in the soil was monitored using the following data sources:

-

Climatic data are those obtained from the agrometeorological stations of the Agroclimatic Information Network of Andalusia (RIA) [25].

-

Determination of the value of the crop coefficient: the monthly average NDVI indices were used from TERRA-MODIS satellite images obtained from the Andalusian Climate Information Network [26].

-

Classification of the different types of soils: the soil map of Andalusia at a scale of 1:400,000, prepared in 2005 by the Ministry of Environment (Environmental Information Network of Andalusia, REDIAM) [27], was used based on the map published in 1989 by the Ministry of Agriculture and the Higher Council for Scientific Research; digitized; and readjusted with reference to the orthoimages of the Landsat-TM satellite, whilst also taking into account its characteristics [28].

-

The physicochemical and hydrological properties of the soil were obtained and the water reserve in the soil was determined using the collection of soil maps of the Institute of Natural Resources and Agrobiology of Seville (IRNAS-CSIC) [29,30].

-

The delimitation and location of the area and crop cover were obtained using the Information System on the Natural Heritage of Andalusia related to land occupation (SIPNA) [31]. This system establishes six hierarchy codes and has the occupation codes of the Information System on Land Occupation in Spain (SIOSE) as a cartographic base at a scale of 1:10,000.

The modeling and processing of the data were carried out using the free software QGIS 3.18 [32]. The workflow used to obtain the data of the variables analyzed is shown in Figure 2.

Finally, a practical application of the described methodology was carried out by applying it to the study reduced to the municipal scope of the municipality of Cazalilla. (Supplementary Material File S2). In any case, these results represent a first approximation that should be more made precise and detailed in order to provide more adequate results as the scope of the study is reduced at the municipality district area, irrigation community, farm and/or plot levels.

2.3. Climatic Data

To obtain the climatic data, a total of 18 automatic EMC/EMA-type weather stations [25,33,34] distributed throughout the provincial area and belonging to the Andalusian Agroclimatic Information Network dependent on the Ministry of Agriculture, Livestock, Fisheries and Sustainable Development (RIA) were used [25], (Supplementary Material S1).

Given the possibility of the existence of anomalies in the data [35,36], the R.Climatol package [37,38,39] programmed in R language [40] was used during the study period 2006–2020.

Relative homogenization employs the SNHT (Alexandersson–Moberg) test and establishes a reference station calculated from a weighted average of neighboring stations that takes into account the distance between the different stations and a reference station [41,42]. A climatological series is homogeneous if its variations are caused by weather and climate variations, or if it is representative of the climate in the surroundings of the observation point [43].

The Penman–Monteith equation was used to calculate reference evapotranspiration (ET_o). Precipitation (P) was defined as the water in the atmosphere that falls onto the Earth’s surface in liquid, solid, or liquid–solid form from clouds. Effective precipitation (Pe) is the usable precipitation, i.e., precipitation not lost by runoff or deep percolation. It can also be considered as the amount of precipitation that is stored in the soil without being lost though runoff or deep percolation and remains available to use by vegetation. It is the fraction of total precipitation that is used by plants. Its determination can be performed using simplified methods [44,45,46], regionalization [47] or GIS techniques of geostatistics [48].

The results obtained were represented in the study area by using GIS techniques with the application of the inverse distance weighted (IDW) interpolation method [49,50], with a pixel resolution of 5 m. Effective precipitation (Pe) is considered to be 75% of the rainwater [51] that manages to infiltrate into the soil without being lost to runoff or deep infiltration [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52].

2.4. Determination of the Crop Coefficient

The crop coefficient K_c varies during the growing period and depends on the development of the vegetation cover. Its determination is based on an advanced formulation of the procedure [44], which breaks it down into the sum of the basal crop coefficient K_cb, referring to transpiration [53,54]; and the K_e coefficient, called the evaporative coefficient, which includes evaporation from the bare soil. This model is referred to in the scientific literature as K_c − ET_o, when it performs the determination of K_c in two steps [55,56] or IV-ET_o when it integrates the vegetation indices [57] and allows the establishment of a good linear relationship, which is widely demonstrated both theoretically [58,59] and empirically [60]. Other authors take into account evaporation from wet bulbs [61].

The crop coefficient K_c is determined through the following expression:

K_c = K_s ∗ K_cb + K_e

(1)

where:

-

K_s: the stress coefficient, with values from 0 (maximum stress) to 1 (no stress);

-

K_cb: the basal crop or transpiration coefficient;

-

K_e: the evaporative coefficient, which depends on the growth and development cycle of the canopy.

Values of K_s and K_e were estimated using those of the usual drip system and weekly irrigation [62], (K_s = 1, K_e = 0.05 was adopted). Calera et al. [63] estimated K_cb with the vegetation index NDVI:

K_cb = 1.44 ∗ NDVI − 0.1

(2)

The above equation allows us to obtain K_c values from an image in which NDVI has been previously calculated for each pixel. The sequence of images, in turn, allows us to study the monthly evolution of the canopy, showing both spatial and temporal variability. Therefore, we can obtain

ET_c = ET_o ∗ ((1.44 ∗ NDVI − 0.1) + 0.05)

(3)

To differentiate between irrigated and unirrigated olive groves, information from the Andalusian Natural Heritage Information System (SIPNA) [31] was used.

The normalized difference vegetation index (NDVI) was introduced by Rouse et al. [64] and is a measure of the value of photosynthetically active biomass. Its values are a function of the energy absorbed or reflected by plants in various parts of the electromagnetic spectrum in a range from −1 to +1. Two bands are used for its calculation and determination: near-infrared (NIR) and red (RED), which are different depending on the observation satellites used. The formula for its calculation is as follows:

$$\mathrm{NDVI}=\frac{\mathrm{NIR}-\mathrm{RED}}{\mathrm{NIR}+\mathrm{RED}}$$

(4)

For the determination of the K_c coefficient of the olive grove, we used the TERRA satellite images captured by the MODIS sensor (Moderate Resolution Imaging Spectroradiometer), with a spatial resolution of 250 m, obtained from the Andalusian Climate Information Network [26]; referred to the normalized difference vegetation index (NDVI); and studied the time series of this parameter during the study period considered [65,66,67,68]. A total of 177 images corresponding to the period 2006–2020 were used, and after being corrected and georeferenced, they were processed and represented in reference to the scope by analyzing and studying their temporal evolution to obtain the monthly and annual mean values and their GIS representation in raster format.

2.5. Determination of Soil Types

Soil is a key factor in water balance as it allows for water storage, and therefore, its availability for the crop, so in deficit irrigation it will allow for the establishment of the irrigation schedule with greater precision. Its characteristics and properties (texture, depth, physical properties, drainage and infiltration) allow us to consider the soil water reserve to enable and improve irrigation efficiency in arid or semiarid areas, as well as the development and productivity of the crop [69]. One of the most important conditioning factors when applying deficit and regulated irrigation is the storage capacity or water reserve of the soil [70,71].

The soil map of Andalusia at a scale of 1:400,000 was adopted as the soil database [27]. The soils appear in cartographic units characterized by associations grouped at the second-order level following the FAO classification criteria and the 1985 European Union Soil Map with spatial resolution of 105 m [72,73]. For soil classification, several documents were studied and taken into consideration: the database of edaphological properties of Spanish soils [28], the determination of agrological classes at the provincial level [74] and the keys for soil taxonomy [75,76,77]. Thirty-one edaphic units were described in the scope of the present study (

Table 1. Edaphic soils unit of the study area, (legend table first above).

Soil Type				Horizon Diagnostic Qualifier
C	Cambisols	Ph	Phaelozems	Ca	Calcareous	Hu	Humic
F	Fluvisols	R	Regosols	Cr	Chromic	Or	Orthic
L	Lithosols	Ra	Rankers	Dy	Dystric	Ha	Haplic
Lu	Luvisols	Re	Rendzines	Eu	Eutric	Pe	Pelic
N	Nitisols	V	Vertisols	GL	Gleic	Ve	Verthic
				omm	over metamorphic materials	omp	over plutonian materials
Code	Edaphics Units			Code	Edaphics Units	Code	Edaphics Units
1	EuF and EuC			22	PeV and CrV	47	CaC, CaLu and CrLu with CaL and CaF
2	CaF			23	CrV and VeC with CaC, CaR and PeV	48	VeC, CaR and CrV with CaC
5	EuR, L and EuC with Ra omm			31	EuC, EuR and L with Ra	49	VeC, CrV and CrC with CaR
6	EuR, L and EuC with Ra opm			37	EuC, CrLu and OrLu	51	OrLu, Glu and EuC
9	CaR and EuR			38	EuC, CrLu and OrLu	53	CrLu and R
10	CaR			39	DyC, HaPh and Ra with HuC, DyR and L	55	CrLu, L and EuR with DyN
11	CaR and L with CaC			40	CaC with CaR	57	CaLu, CaC and EuC with CrLu, CaR and L
13	CaR and CaC with L, CaF and Re			41	CaC with CaR	58	CaLu, CaC and CrLu with CaR
14	CaR and CaC with CaLu and CaF			42	CaC with CaR, CaF and CaLu	59	CaLu, CrLu and Glu
19	L, CrLu and Re with CaC			43	CaC and CaR with CaL, CaF and VeC
21	PeV, Re and CaR			44	CaC, CaR and L with Re

) out of a total of 64.

In order to obtain the soil properties, we used the lithological maps of classes and subclasses of Andalusia elaborated at a scale of 1:400,000 and spatial resolution of 105 m from the cartographic and geological information of the Geological and Mining Information System of Andalusia from the detailed edaphological information of the SDBm-SEISnet database (1083 soil profiles) [29,30] compared and completed using the soil database of the province of Jaen [28]. It was possible to establish the dominance relationship referring to the different types of existing soils, finding 14 dominance codes out of a total of 21 for the whole of Andalusia (

Table 2. Dominance edaphic codes unit and textures of the study area.

Dominance Code	Dominance Meaning	Texture
0	No data
1	Soils dominated by fluvisols	clay
2	Soils dominated by eutrophic regosols	sandy loam
3	Soils dominated by calcareous regosols	sandy clay loam
4	Lithosol-dominated soils	sandy clay loam
6	Pebbly vertisol-dominated soils	sandy clay loam
7	Soils dominated by chromic vertisols	clay loam
10	Eutrophic cambisol-dominated soils	clay loam
11	Soils dominated by dystric cambisols	clay loam
12	Calcic cambisol-dominated soils	clay loam
13	Vertic cambisol-dominated soils	clay loam
14	Vertic luvisol-dominated soils	loam
15	Chromic luvisol-dominated soils	sandy loam
16	Calcareous luvisol-dominated soils	sandy clay loam
18	Soils dominated by eutrophic planosols	clay loam

).

The determination of soil water retention capacity as a function of soil type was performed by processing, digitizing and rasterizing Map No. 6: water retention capacity and equivalent humidity (whole profile) from the collection of the IRNAS-CSIC [29,30].

2.6. Determination of Water Content in Soil

Although the main variable determining soil moisture is precipitation, other factors (soil type, vegetation, topography and slope) influence its distribution both spatially and temporally [78,79]. Localized irrigation enables the application of irrigation in olive groves with slopes of up to 35–45%, although it notably influences distribution uniformity and application efficiency [80].

From a hydrological point of view, regarding the distribution and circulation of water in the soil, the following variables are usually used: slope, roughness, aspect, orientation, topographic moisture index, etc. [81,82,83,84]. However, when determining soil moisture, the explanatory variables that present a higher correlation are the physical and hydric properties of the soil. Thus, the aforementioned variables present a low linear relationship, and therefore a lower correlation [85].

Soil moisture content is measured according to its physicochemical properties, with the application of direct (gravimetric) or indirect methods (tensiometers, electrical resistivity, neutron probe, gamma ray attenuation, time domain reflectometry (TDR) or frequency domain reflectometry (FDR)) [86,87]. Other works use pedotransfer functions (ETFs) [88] or their estimation using GIS spatial modeling techniques [89], the combination of results from climate stations with soil sampling and subsequent analysis [90], the application of hydrological models such as SWAT [91,92] or the use of wireless soil sensors or soil sensor networks [93,94,95]. The maximum amount of water that a soil can retain is referred to as the available moisture interval [96].

The available moisture interval (AWI) and water retention capacity (SWRC) were determined as a function of soil type in the study area [97]. AWI refers to the amount of water theoretically available to plants that they can absorb. AWI was calculated as the difference between the values of the field capacity (FC) and permanent wilting point (PWP) levels [98,99]. On the other hand, soil water retention capacity (SWRC) refers to the AWI in the soil profile up to the depth where the roots are located [100]. Therefore, we have:

$$\mathrm{AWI}=\mathrm{FC}-\mathrm{PWP}$$

(5)

$$\mathrm{SWRC}=\left[\frac{\left(\mathrm{FC}-\mathrm{PWP}\right)}{100}\right]\ast \mathrm{Z}\ast \left(1-\mathrm{f}\right)$$

(6)

where:

-

AWI: available moisture interval, (in %);

-

FC: field capacity, (in %);

-

PWP: permanent wilting point, (in %);

-

Z: exploratory depth of roots (100 cm was adopted);

-

f: percentage of rock fragments (in parts per unit);

-

SWRC: soil water retention capacity (in mm).

To obtain the water balance, the procedure described in Figure 3 was followed [101]. With this balance, we can quantify, for each period of time analyzed (in this case, each month), the behavior of soil moisture and determine both the excesses (those values that exceed soil saturation, and which give rise to runoff), and the moisture deficits (at levels close to the permanent wilting point).

In the study of wide regions, it is necessary to assume a series of simplifications for the calculation of the water balance. In the present study, it was considered that the only water input comes from precipitation (inputs), and only reference evapotranspiration was considered (outputs). Percolation, runoff and capillary rise (lateral or vertical water movements) were not taken into account. The balance was established for the hydrological year (the hydrological year is from 1 October until 30 September).

At the crop level, water requirements were established by applying the simplified water balance equation based on the principle of conservation of masses [102,103].

Thus, we can obtain

ET_c − (Pe + Ir) − SR − In = ±ΔW

(7)

where:

-

ET_c: crop evapotranspiration;

-

Pe: effective precipitation;

-

Ir: irrigation;

-

SR: surface runoff;

-

In: infiltration of water in the soil.

Considering the previously described simplifications, we can assume that neither runoff (SR) nor infiltration (In) occurred. In addition, no irrigation was considered (Ir equal to 0), so the final form the above equation is as follows:

ET_c − Pe = ± ΔW

(8)

This water balance was estimated monthly to identify the periods in the season when precipitation could be too scarce and could enhance problems of water stress. Finally, water reserve (WR) was calculated as the variations in soil moisture using the accumulation of water balance in each month:

WR_n = WR_n−1 ± ΔW

(9)

where:

-

WR_n is the water reserve in the current month.

-

WR_n−1 is the water reserve in the previous month.

-

ΔW is the water balance each month.

3. Results

3.1. Climate Variable Values

The climatic variables analyzed were the data corresponding to the monthly averages of evapotranspiration and precipitation during the period 2006–2020 obtained from the agroclimatic stations of the study area, (Supplementary Material S1). Figure 4 shows the frequency distributions of the data obtained for precipitation (P) and reference evapotranspiration (ET_o) at the agroclimatic stations.

The annual distribution of ET_o and P showed water deficit conditions due to high values of evapotranspiration and low precipitation. Figure 5 and Figure 6 show the monthly results of ET_o and Pe. ET_o was extremely high in all the studied zones from May to September, where the south-east zone presented the greatest values.

On the contrary, Pe was almost null from June to August in all the zones (Figure 6). The month with the greatest amount of Pe was, on average, November. Although Pe was measured in May and September, the real period of rainfall could be considered only from October to April. The south-east commonly presented the lowest amount of rainfall in this latter period.

On an annual scale, the mean annual evapotranspiration is 1237 mm/year, while the mean annual precipitation does not exceed 455.5 mm/year. Within the studied region, there was an important variation ranging from 386.1 mm in J01 (Huesa) to 552.6 mm in J16C (Marmolejo), which reflects the existence of a longitudinal and altitudinal rainfall gradient.

3.2. Crop-Specific Values

The analysis of the data obtained showed maximum values of NDVI in the autumn–winter period and minimum values during the spring–summer period (Figure 7). The decrease in NDVI during the phases of the greatest reproductive activity (flowering and fruiting) may be due to source–sink ecophysiological mechanisms. Similarly, when temperatures are high (>35 °C), progressive stomata closure may occur, and therefore, a decrease in or stagnation of shoot growth may occur [104]. Figure 8 shows a graph representing the relationship between the NDVI and monthly K_c values obtained in the period considered.

The relationship between these two variables is of quadratic type with a very high coefficient of determination (R² = 0.8696) [106]. There was a decrease in K_c values throughout the season. This is a well-known pattern of K_c in olive trees and is commonly related to the stomatal response to evaporative demand.

The monthly K_c values obtained by applying Equation (1) for the olive crop range from 0.27 to 1.08 (Figure 9). The greatest values of K_c were obtained in the north and north-west of the studied region. The pattern was similar in all of them, with maximum values in winter and minimum values during summer. The range of K_c during the common irrigation season from June to September was approximately 0.5, which is in accordance with the value reported in the literature [105,106,107].

The monthly evolution of crop evapotranspiration (ET_c) is presented in Figure 10. However, great variability for ET_c values was found in the study area throughout the season. All the regions presented extremely high ET_c values, mainly from May to August (more than 80 mm). ET_c values lower than 50 mm were found from October, but even in this month, some regions presented some data of approximately 80 mm. During the period of the greatest values (from May to August), the north-west and east zones presented the maximum ET_c.

3.3. Values of Soil Variables

The range of variation in soil water variables on the map for average bulk density and depth explored by crop roots of 100 cm is shown in the following table (

Table 3. Range of variation in soil water variables.

Variable	θg		θv
Saturation point (SAT)	30.28	38.65	43.10	55.00
Field capacity (FC)	17.33	36.22	24.89	48.17
Permanent wilting point (PWP)	8.34	26.97	11.03	35.87
Usable water (SWRC)	58.14	127.91	82.75	182.04

).

The useful and usable water or maximum amount of water that the soil can retain is that corresponding to the difference between the field capacity and the permanent wilting point [96,98,99].

For the determination of the soil’s available water retention capacity for plants (SWRC) [100], the influence of rocky fragments was taken into account whilst considering that an optimal proportion of these is between 10 and 30%, so a proportion of 20% was adopted [108], obtaining a range between 82.8 and 182.0 mm (Figure 11).

Soil water storage capacity allows for its utilization by the crop, and influences the previous water balance calculated by considering this reserve to improve irrigation efficiency [71]. Figure 11 presents a great variation in SWRC. The greatest capacity of SWRC was located in the north of the studied zone, with values greater than 180 mm. On the other hand, there were several locations in the center and in the east that presented values lower than 90 mm, even quite near to the maximum. Such variability in the SWRC could affect the water management in olive groves.

3.4. Water Balance

By applying the water balance equation to the monthly maps, the monthly variation in this measurement was determined. Nine intervals were established, in which the red colors correspond to a negative balance and the white-to-blue colors to a positive balance (Figure 12).

At the annual level, the values of the water balance obtained changed in a range of around 77 to a maximum peak value of −242 mm during the month of July. This indicates that globally, the level of precipitation collected does not compensate for the losses caused by evapotranspiration from May to September, and therefore justifies the need for irrigation to enhance the development and productivity of the crop.

During the first months of the year (January to April), we observed that a positive water balance was mainly concentrated in the northern and eastern areas, which corresponded to the mountainous forest areas of Sierra Morena and Sierra de Cazorla, Segura and Las Villas, with a maximum value of around 77 mm in April. On the other hand, in this period, there was a negative water balance in January and February in the center of the studied zone.

It is also highlighted that during the period from May to September, the evapotranspiration exceeded rainfall in all the months and zones, a situation that favors the presence of water stress and its effect on the critical phenological periods of flowering (May), fruit set (July) and stone hardening (August). Maximum evapotranspiration values were found in the northern and eastern zones of the area. On the other hand, they were more moderate in the central, western and southern zones, which have a marked and eminent agricultural character.

The positive water balance was delayed until November, but even in this month, there was a small zone in the south-east where negative values were found.

3.5. Variation in Water Reserve in the Soil, (WR)

The water reserve pattern is presented in the hydrological year (from October to September) because this would permit a better design of a strategy of water management. According to water balance and Pe distribution (Figure 6 and Figure 12), the soil profile was empty at the beginning of October.

The water reserve increased in the month of October in a localized and discrete manner in certain areas in the center and west of the area under study. The real recovery of the soil profile occurred in November in most of the zones, and only in the east zone was such recovery delayed until April. Maximum values were estimated in the month of April, with around 149 mm. It was also observed that at the beginning of the agricultural year during the autumn season, the water balance in the soil was greater in the central and western zones, which are predominantly agricultural and mainly used for olive cultivation, while during the spring, it was concentrated in the northern and eastern zones. From April onwards, the water reserve in the soil began to decrease drastically at the same time as temperatures increased, and consequently, the water needs for the crop increased (Figure 10). Soil water reserve was zero from May in most of the studied zones. The representation of the variation in the water reserve in the soil showed that during a long period of time (more than 33.3% of the year), its value was zero throughout the study area (Figure 13).

3.6. Irrigation Scheduling

Taking into consideration the different layers of information previously developed, its application to any other crop (within the scope of the study area) would consist of determining the evolution of the K_c coefficient values in order to subsequently apply the simplified water balance equation (Equation (7)). This water balance was performed monthly given the generic characteristics of the study [109], although once the irrigation installation is completed and the crop defined, it can be performed daily [110], weekly [111] or seasonally, according to the crop cycle and/or arrangement of satellite images [112,113].

4. Discussion

Climatic changes are increasing the irrigation needs and the surface of irrigated lands. Moreover, the decrease in rain will produce a decrease in the available water. Therefore, water management will be improving at groves and at the basin level. This supposes different ways where farmers and technicians will have to optimize the available resources. They can mainly be put into two groups: hydraulic design and irrigation scheduling. Both are related and could limit the other. The current work is focused on the first one, the suggestion of a tool for delimiting homogenous zones and providing information for irrigation design, which will allow for easier irrigation scheduling using actual data. Data of the current work suggest that at least two zones could be considered with different periods of available water. Current data assumed simplifications that reduced the estimation accuracy (for instance: Pe estimation), but showed that even in these conditions, there were deficiencies in the irrigation decisions. Therefore, data suggested significant changes in the water management of the zone in comparison with the traditional fix calendar. In the northern and eastern zones, irrigation could be significantly delayed from March (current recommendation) to May. This strategy could extend the irrigation season until October when rainfall is still scarce. On the other hand, the central, western and southern zones had greater irrigation needs. Then, the irrigation season should be longer than the previous one, from April to November. This period started later than is commonly suggested, but also would need to finish later to decrease the effect of water stress in the oil accumulation period.

The current work suggests an approach for organizing water resources in a big zone, even a basin. The results identified periods of water reserve depletion and zones where it occurred early (Figure 13). These data would not avoid water stress but they could help in the decision of irrigation scheduling in conditions of scarce amounts of water. The whole period from inflorescence development to fruit set was reported as a very critical phenological stage in relation to yield [114,115]. These periods occurred in our zones from March to May, which is the shift between the greatest level of available water and the beginning of depletion. Current data showed that the north and east zones could secure greater yields than the rest because of their greater water reserve (Figure 13). Such variability is common all around the world, though olive trees have a narrow zone of growing. Examples of this variability are the works of Goldhamer (1999) [14] and Lavee et al. (2007) [116], who suggested different deficit irrigation scheduling. The former reported that decreased irrigation during summer (pit hardening) did not reduce yield [14]. However, the latter suggested no irrigation until pit hardening, though they indicated that in conditions of scarce amounts of winter rain this strategy could be not advisable [116]. The lack of agreement between both is likely related with the available water reserve in the soil.

The improvement of the current approach is very important, especially if accurate estimations at the farm level were considered. The weak variables in the current work are likely the Pe and K_c estimations. Other components considered in the water balance, such as ET_o or soil type, would be easily estimated using public data. Because the current work was focused on hydraulic design, the average data of several seasons would correctly estimate ET_o and soil maps would estimate the capacity of water storage. Pe is difficult to estimate even with local data, because there are different components that could affect it, mainly the amount of rainfall [117]. Although the seasonal pattern would be very similar, the amount of Pe during Autumn–Winter could affect the irrigation needs and the start and the end of the irrigation periods. The current approach, 75% of Pe, overestimated the recommendations of Villalobos et al., (2016) named the “FAO method”, suggested for arid zones [117]. However, the current approach could be useful in these local conditions because the amount of precipitation is commonly very low (most of them lower than 50 mm, Figure 4) which could increase the Pe from this method. Fernández et al., (2011) in a location near to the studied zone (around 150 km from Seville but with similar patterns and amounts of rain), also used this approach of 75% rainfall in the comparison of several plant water status indicators in olive trees [51].

Water needs are also very affected by ET_c, which in the current work was mainly spatially changed for K_cb and ET_o data. K_cb is affected by tree development but also by soil management. Steduto et al., (2012) reported several Kc for olive trees depending of locations, date of the season and soil management [107]. The estimation of Kcb based on NDVI reported values similar and with the same pattern of decrease during summer because of the increase in evaporative demand (Figure 7 and Figure 8) [107]. NDVI has been reported as very useful tool in olive trees for estimating spatial variations at grove level [118]. Therefore, the current approach based on this measurement would provide enough information for an accurate water balance.

The main limitations of the current approach would be the assumption of 1 m of root zone. This value assumes that all groves are mature and would underestimate water needs in young. However, because the current work is focused on the hydraulic design, this limitation is null. Irrigation design has to provide information about maximum needs along the years in order to optimize pump dimensioning. This supposes that only mature groves have to be considered. The effect on young olive groves will be managed with seasonal irrigation scheduling. On the other hand, in some zones, soil depth could be lower than 1 m, and again, overestimate soil water storage. These data could affect the current final estimation and would be improved in further works.

Finally, in terms of the precision of the results obtained from the different public information layers, the most limiting was that corresponding to the determination of the NDVI, which has a spatial resolution of 250 × 250 m.

5. Conclusions

Knowledge of the availability of water in the soil in large areas of an irrigated crop is necessary when planning the management of water resources, especially in areas with limited available water resources.

Its definition allows decisions to be made both on irrigation engineering (determination of the number and extent of irrigation sectors) and on the optimization of irrigation scheduling and timing (doses and frequency) to provide available water resources in a more efficient way.

In the present work, the evolution of the water reserve in the soil for a given crop was determined through the management of public data. The adaptability of the methodology used makes it possible to study any other crop or cultivation pattern in any other zone, taking into account the values of the respective K_c coefficients to allow for the evaluation and determination of the water balance and water reserve in the soil; although in these cases, the application and justification of agronomic and technical criteria, as well as socioeconomic criteria, are precedent.

The results obtained from the monthly variation in both the water balance and the water reserve in the soil in the area studied indicate that the use of public sources of irrigation scheduling data may take into account spatial and temporal distribution patterns and its specificities. Additionally, fixed irrigation calendars should be avoided to enable more efficient schedules.