Introduction

The high-water demand that characterizes the citrus sector of the Mediterranean basin makes necessary the adoption of precision irrigation strategies in order to contain water waste and increase the water use efficiency by crops. For this reason, the use of micro-irrigation techniques, combined with the monitoring of mass exchanges occurring in the continuous soil-plant system, can contribute to the identification of water dynamics in the unsaturated porous media for defining root water uptake (RWU) and consequently determine crop water consumption (Saitta et al. 2021). In this context, hydrogeophysics provides a number of monitoring techniques that can be further exploited for inferring the soil water dynamics (Binley et al. 2015) and optimizing the irrigation management of complex agro-systems. Commonly, electrical resistivity (ER) measurements have been used to map the spatial inhomogeneity of the soil according to specific petro-physical relationships existing between the soil ER proprieties and the main soil state variables (Samouëlian et al. 2005). In particular, the use of the electrical resistivity imaging (ERI) technique, combined with advanced inversion models, has shown its ability for generating two-dimensional (2-D) or three-dimensional (3-D) ER images related to the spatial variation of the soil characteristics (Binley and Kemna 2005). Even if ER is influenced mostly by SWC, it also depends on a number of other factors, such as pore water salinity, clay content, lithology, soil density, porosity, soil temperature, organic matter (Samouëlian et al. 2005) and root electrical properties (Ehosioke et al. 2020; Rao et al. 2019, 2020). A number of procedures, based on theoretical (Mualem and Friedman 1991) or site-specific calibrations, made both in field (among others Calamita et al. 2012; Michot et al. 2003; Schwartz et al. 2008) or in laboratory (Archie 1942; Gupta and Hanks 1972; Rhoades et al. 1976; Kalinski and Kelly 1993), have been assessed in order to translate the ER distribution in SWC terms. A review, focusing on the use of ER-based methods to monitor SWC, is provided by Brillante et al. (2015), which described the advantages and limitations associated with the methodological and modelling approaches used for calibrating ER with SWC data at different soil conditions. For example, Calamita et al. (2012) combined the use of ER and SWC sensors, based on the time domain reflectometry (TDR) method, for quantifying soil moisture content over large areas. Schwartz et al. (2008) used time-series of 2-D ER measurements, TDR-measured soil moisture and soil physical-chemical properties for quantifying the 2-D field-scale SWC distribution in heterogeneous clayey soils. Moreover, Michot et al. (2003) calculated 2-D SWC sections from ER-based surveys using field-scale calibration relationships obtained between the ER of each soil horizon and its SWC. In general, petro-physical relationships are site-specific and need to be calibrated before serving as translation between ER and SWC (Mary et al. 2021). In addition, petro-physical relationships at the root zone level suffer from biases due to the influence of the root system on the measured ER (Cimpoiaşu et al. 2020).

Repeated ERI measurements over time (i.e. time-lapse mode) can reveal temporal variations in SWC (Zhou et al. 2001). In particular, time-lapse ERI (or electrical resistivity tomography, ERT) approaches show strong potential for monitoring SWC changes (Michot et al. 2003; Calamita et al. 2012; Vanella et al. 2021), RWU (Beff et al. 2013; Vanella et al. 2018, 2019) and root zone system processes (Cassiani et al. 2015) in the unsaturated zone. Recently, Mary et al. (2021) stated that time lapse ERI or ERT measurements are the most informative tools to derive SWC movement and variation due to RWU.

In order to describe the existing relationship between the ER and the SWC in heterogeneous conditions (i.e. orange trees with different age and variety), we performed successive ERI measurements aimed at providing new insights into the technique’s ability to operate as a precision irrigation tool to promote water-saving at the farm level, with trees of different cultivars and of mixed age. This need is particularly important in the citrus sector of the Mediterranean area, where it is customary, both for the need to eliminate the plants affected by the Citrus tristeza virus (CTV) and for intensifying production, to alternate crops of different varieties and ages in the same plot. This, obviously, has consequences for the water dynamics in the unsaturated porous medium and on the absorption mechanisms of water and nutrients, since the roots can have different depths, and the trees different levels of transpiration.

The experiment was conducted in a citrus orchard located in Eastern Sicily (insular Italy), during the irrigation season in 2020. The citrus grove has a high degree of crop heterogeneity (variety and age) on the same framework and it is irrigated with surface micro-irrigation techniques. During the irrigation phase, the infiltration fronts were monitored by 2-D ERI surveys carried out in time-lapse mode in order to analyse their characteristics and to identify the mechanisms of the RWU process having different depths. Time-lapse ERI data were related to independent SWC measurements, which were acquired using sensors installed in the field. The objectives of the study were: (i) to provide an evaluation of 2-D ERI monitoring on spatio-temporal variations of SWC in heterogeneous field conditions, (ii) to set up simple calibration relationships between ER and soil moisture measurements and, (iii) to assess the potential of the ERI technique for serving as a precision irrigation monitoring tool at farm level.

Materials and methods

Study site characteristics

The experiment was conducted during 2020 in a commercial orange orchard located in the agricultural context of the Catania Plain (Eastern Sicily, insular Italy) (37°28’47.2"N 14°57’12.6"E (WGS84), with elevation of 87 m above the sea level, a.s.l.) (Fig. 1).

Fig. 1
figure 1

Overview of the study site with the locations of the soil water content sensors, the weather station and electrical resistivity imaging (ERI) array

Full size image

Orange trees (Citrus sinensis (L.) Osbeck) characterized by different varieties and ages were alternated along the tree rows at the study site. Specifically, mature “Tarocco Scirè Nucellare” trees (8-years old, with average tree height and canopy diameter of 2.3 and 2.0 m, respectively), planted at a spacing of 5 m × 5 m, with ground cover of 25%, were inter-spaced along the row with young “Tarocco Ippolito” plants (3-years old, with average tree height and canopy diameter of 1.3 and 0.7 m, respectively), positioned at a distance of about 2.5 m from the pre-exiting trees, with ground cover of 3%. This practice is very common at the local citrus-growing areas, since it allows to intensify the production and also to eradicate those plants with evident CTV symptoms (Vanella et al. 2020). The climate of the study site is typical semi-arid Mediterranean, with warm and dry summers, where generally precipitation does not occur. Mean air temperature (Tair), mean annual reference evapotranspiration and precipitation values within the period 2002–2020 were about 18 °C, 202 and 624 mm, respectively (data provided by SIAS – Servizio Informativo Agrometeologico Siciliano – Catania weather station).

The main physical and chemical soil characteristics of the study site are given in Table 1. In particular, the soil shows a sandy-loam texture (55.2% of sand, 29.2% of silt, 15.6% of clay) with a high content of organic matter (25.54 g kg− 1).

Table 1 Main physical and chemical soil characteristic of the study site
Full size table

The mean SWC at field capacity (pF = 2.5) and wilting point (pF = 4.2) values were 36.33% and 20.74%, respectively. The soil bulk density was 1.29 g cm− 3. The electrical conductivity (EC) of the irrigation water was monitored during the ERI survey, resulting in values of 1.40 ± 0.06 dS m− 1 at 25 °C (HD2106.2 conductivity meter, delta OHM, Italy).

The study site was equipped with an automated monitoring station (Fig. 1) used for acquiring the main agro-meteorological parameters, such as Tair (°C) and relative humidity (RH, %, thermo-hygrometer Humitter_50Y, Vaisala), wind speed and direction (m s− 1 and °, ultrasonic wind sensor WindSonic, Gill Instruments), precipitation (mm, rain gauge SBS500, Environmental Measurements) and solar radiation (Rs, W m− 2, pyranometer first class Eq. 08-E Middleton solar). The main soil characteristics (including the SWC in m3 m− 3 and soil temperature, Tsoil in °C) were monitored through the use of time domain reflectometry probes (CS616, Campbell Scientific, only for SWC) and high-frequency capacitive probe (5TE, Decagon Device, both for SWC and Tsoil measurements). The sensors were preliminarily tuned and laboratory-tested in order to verify their functionality. All data were managed and collected by a CR1000 data-logger (Campbell Scientific), which acquired measurements every 15 s and recorded the corresponding average values every 30 min. The station was equipped with a modem (TMAS-T61, TCAM Technology Pte Ltd) for remote control and data retrieval over a 3G mobile network. Data was then post-processed and aggregated at hourly and daily scales.

Irrigation scheduling was calculated on the basis of the FAO-56 Penman-Monteith (P-M) approach (Allen et al. 1998), using the measured agrometeorological parameters, adjusted by the crop coefficient for citrus species (Kc, equal to 0.7 from Consoli et al. 2006a, b), the localized factors as function of the ground cover (i.e. mean of 0.63 and 0.52 for mature and younger trees, respectively, Consoli et al. 2017) and finally for rain occurrence. Trees were surface drip irrigated (SDI), using two surface laterals per tree row, with 6 emitters and flow rate of 4 L h− 1 per tree (spaced 0.5 m) at a pressure of 1.2 bar. Irrigation volumes were supplied in the spring-summer period, twice a week for a duration of 5 h for each irrigation. During the growing period from May to September, the study field received an irrigation amount of about 104 m3 delivered by the SDI system.

Time-lapse electrical resistivity tomography surveys

In general, ERI (or ERT) technique uses a pair of current electrodes to inject current into the soil and to measure the electrical potential through another pair of electrodes (Binley and Kemna 2005). Given multiple combinations of transmitting and receiving electrodes along a linear array, a 2-D image of the real ER (i.e., the inverse of EC) distribution can be reconstructed through inverse modeling (Binley et al. 2015).

The ERI field setup used at the study site (Fig. 1) consisted of a transect made of 72 surface electrodes (i.e., steel rods of about 0.15 m, with a diameter of 0.05 m, inserted for 2/3 of their length into the soil surface). The surface electrodes were spaced 0.15 m along the linear array, parallel to the drip irrigation line, covering a total length of 10.65 m. This ERI configuration permitted to explore the soil under 4 trees, i.e. 2 for 8 and 3-year old trees, respectively (Fig. 2).

Fig. 2
figure 2

Layout of the ERI surveys conducted at the study site

Full size image

The ERI dataset acquisition was performed with a 10 channel Syscal Pro georesistivimeter (IRIS Instruments, Orleans, France). A total of about 5,000 readings, including both direct and reciprocal quadrupoles (for each ERI dataset), were collected using a dipole-dipole electrode configuration, due to its intrinsic strength in solving ER lateral changes (Samoüelian et al. 2005). The high spatial coverage of the adopted ERI array permitted us to reach depths for investigation of about 1 m (Oldenburg and Li 1999).

A time-lapse approach was applied for conducting the ERI surveys. It consisted of performing multiple ERI repetitions at different times during and after an irrigation event to infer the SWC dynamics under the field conditions (from 9.45 a.m. to 12.28 p.m., on September 30, 2020) using the following protocol: (i) 1 dataset was acquired before the beginning of the irrigation event, early in the morning (time 00, initial condition), (ii) 5 dataset was collected during the irrigation event (from time 01 to time 05), and (iii) 2 dataset after the end of the irrigation phase (from time 06 to time 07). These state conditions were chosen on the basis of the duration of each ERI dataset collection (20 min) and the timing of the irrigation cycle (less than 3 h). The details about the ERI dataset scheduling is reported in Table 2.

Table 2 Details of the ERI dataset collection; timing (hh:mm) is expressed in local time
Full size table

ERI data processing was performed using the freeware R2 code (v. 4.02), which permits obtaining the forwarding/inverse solution for 2-D or 3-D current flow in a finite element domain (Binley 2020). The finite element domain of interest (e.g., a triangular mesh made of 2,322 cells and 4,325 elements) was generated using the Gmsh software (Geuzaine and Remacle 2009). The quantification of the measurement and model errors was evaluated for being used as weights in the inversion process. Specifically, the measurements error was calculated on the basis of the direct and reciprocal measurements by data filtering and removing the quadrupoles with reciprocity error above 10% (Binley et al. 1995; Slater et al. 2000) (Table 3 in the Appendix). In addition, several forward models were run for a uniform resistivity (value fixed at 100 Ω m) in order to compute the model error level before setting-up the inversions.

The 2-D data inversions were run both in absolute (at absolute error level of 10%) and in time-lapse mode (at relative error level of 5%), in order to image the ER characteristics in static (time 00) and dynamic conditions (time 01, …, time 07), respectively. In particular, the absolute inverse solution, based on a regularised objective function combined with weighted least squares (an Occam’s type solution), was applied as defined in Binley and Kemna (2005) and in Binley et al. (2015) for identifying the ER distribution at “time 00” (no irrigation, Table 2). Conversely to the absolute inversion that accounted for the static effects on soil ER (e.g., soil texture), the dynamic ER changes were calculated by implementing the ratio time-lapse inversion following the approach described in Vanella et al. (2021), as follows:

$${\mathbf d}_r=\frac{{\mathbf d}_t}{{\mathbf d}_0}F\mathit{\left(\sigma_{ohm}\right)}$$
(1)

where, dr is the resistance ratio (Ω), dt and d0 (Ω) are the resistance dataset collected at selected time periods (from time 01 to time 07) and at the initial condition (time 00), and F(σohm) is the resistance value (Ω) obtained by running the forward model for a fixed ER value (i.e., 100 Ω m).

According to Eq. (1), the ratio time-lapse inversion approach permits to recognize the ER changes (%) in comparison to the “time 00” condition and, thus, to provide evidence for wetting or drying soil patterns (e.g., corresponding to a decline/increase in ER with respect to the time 00 dataset). Note that only the common quadrupoles within the acquired ERI datasets (from time 00 to time 07) were used for the time-lapse inversions (Table 3 in the Appendix). The reconstruction of the 2-D ERI images was performed using the ParaView software (v. 5.8.1, https://www.paraview.org/).

The resulting 2-D ERI images, referring to the absolute (ER, Ω m) and time-lapse conditions (changes in ER), were analyzed by dividing the ERI transects into 4 domains (as the number of multi-aged orange trees that covered the transect) and averaging the ER characteristics of these domains in sub-pairs referring to the mature and younger trees, respectively (Fig. 2).

Results

Soil water content and agrometeorological monitoring at the study site

The daily average SWC and agrometeorological values recorded at the study site during 25 June – 12 October (2020) period are shown in Fig. 3.

Fig. 3
figure 3

Daily average soil water content (SWC, m3 m− 3) (a) and meteorological information, including (b) solar radiation (Rs, W m− 2), rainfall (mm), (c) relative humidity (RH, %) and air temperature (Tair, °C) values, collected at the study site during the irrigation season 2020. Irrigation amounts are included in panel (b)

Full size image

The weather conditions were quite stable during the reference period, resulting in average Tair, RH and Rs values of 25.89 °C, 66.44%, and 257 W m− 2, respectively. The total amount of precipitation and ET0 that occurred during the reference period was 82 and 507 mm d− 1, respectively. The daily average ET0 and crop evapotranspiration (ETc) rates calculated for the mature and younger trees during the reference period were 4.60 mm d− 1, 2.03 mm d− 1 and 1.68 mm d− 1, respectively.

The SWC levels were different under the soil of the monitored trees as function of their age (Figs. 3a and 4). Note that the location of the SWC probes is showed in Fig. 1. In particular, the daily average (and ± standard deviation) SWC values referring to the 8-year-old trees were lower (0.34 ± 0.02 m3 m− 3) than the SWC values of the younger trees (resulting in average values of 0.48 ± 0.03 m3 m− 3) (Fig. 3a). Similar SWC depletion patterns occurred under both mature and younger trees due to non-continuous water applications over the irrigation season (e.g., from 2nd to 8th August, 2020). Figure 4 shows the semi-hourly average changes of SWC (m3 m− 3) measured during the time phases of the ERI surveys in the soil investigated by the roots of young and mature orange trees, respectively.

Fig. 4
figure 4

Semi-hourly average variations of soil water content (SWC, m3 m− 3) measured during the monitored time-steps in the soil explored by the roots of the younger and mature orange trees, respectively

Full size image

Specifically, the initial SWC conditions observed (before the start of irrigation, time 00 in Table 2) were equal to 0.33 and 0.45 m3 m− 3 for the soil associated with mature and young trees, respectively. Note that, as shown in Fig. 4, SWC values of younger trees were more stable (with mean and standard deviation values of 0.50 ± 0.01 m3 m− 3) during the irrigation phase, compared to the SWC values observed for mature trees (0.40 ± 0.05 m3 m− 3). Additionally, Tsoil varied between 22.36 and 23.48 ° C during ERI monitoring.

Electrical resistivity images at the initial condition

The semi-hourly SWC observations (Fig. 4) were in agreement with the ER patterns observed at the initial time step (time 00, in Table 2), under the mixed-age orange grove (Fig. 5). The average ER values (and ± standard deviation) observed under the mature and younger orange trees were 15.40 ± 3.85 Ω m and 14.82 ± 2.31 Ω m, respectively, showing an inverted trend in comparison to the SWC conditions (Fig. 4).

Fig. 5
figure 5

Electrical resistivity distribution (ER, Ω m) observed at the initial condition (time 00) under the mixed-age orange grove

Full size image

Time-lapse electrical resistivity images during the irrigation cycle

The ER changes (%) occurred at different time-steps (Table 2), during (from time 01 to time 05) and after (from time 06 to time 07) the irrigation cycle, with reference to the initial condition (time 00) are shown in Fig. 6. The time-lapse images of Fig. 6 show specific wetting and drying patterns in the soil domain explored by the roots of the trees over time, resulting in different ER changes (%, of decreasing and increasing patterns, respectively, in terms of ER).

Fig. 6
figure 6

Electrical resistivity changes (ER ratio, %) observed during the monitored time-steps (see Table 2) under the mixed-age orange grove

Full size image

In particular, greater wetting fluctuations, with overall ER decreasing by 6.74%, were observed in the soil of the mature trees, in comparison with soil explored by the roots of the younger trees (i.e., 5.40% in terms of ER). Conversely, at the same time-step, slightly higher ER increases were observed in the soil explored by the younger trees, corresponding to an overall average increase of 1.76% in terms of ER, respect to the mature trees explored soil (with an overall increase of 0.24% in terms of ER).

During the irrigation phase, the rates of ER decrease were faster in the soil under the mature trees than in that explored by the younger trees. The maximum soil wetting patterns was detected after the end of the irrigation phase, and specifically at time 06, reaching an ER decrease equal to 10.77% and 7.67% for the soil explored by the root-systems of the mature and younger trees, respectively. No ER changes were observed below 0.5 m of the explored soil depth.

Figure 7 shows the average (and standard deviations) of ER changes (in %) observed during the different time-steps after the beginning of the irrigation phase (time 01–07, Table 2) in the soil profiles under the younger and mature orange trees, respectively. Interestingly, the largest decrease in ER (Fig. 7a) and the smallest increase (Fig. 7b) were observed in the soil domain under the older trees; while these trends were reversed in the soil where the younger trees were located (Fig. 7a-b).

Fig. 7
figure 7

Overall variations in electrical resistivity (ER, %) observed during the different time-steps, after the beginning of the irrigation phase (time 01–07), in the soil under the younger and mature orange trees, respectively. Panels (a) and (b) refer to the observed ER decreasing and increasing patterns, respectively. The bars indicate the change in the standard deviation of ER (%)

Full size image

Electrical resistivity and soil moisture relationship

The analytical relationship resulting from the comparisons between the ER decrease (%) and the SWC variations (%) observed, at the different time steps following the start of irrigation, in the soil explored by the roots of the mixed-age orange trees, are shown in Fig. 8. This relationship resulted in slope and coefficient of determination of 2.68% and 0.63, respectively, due to the fact that the SWC values were normalized to the observed initial SWC conditions (Fig. 4).

Fig. 8
figure 8

Relationship among the electrical resistivity decrease (ER, %) and soil water content changes (SWC, %) observed during different time-steps after the beginning of the irrigation phase (t01-t07) in the soil explored by the roots of the mixed-age orange trees

Full size image

Discussion

In the last decades, non-invasive geophysical methods have offered a number of tools for indirectly characterizing the SWC patterns with a spatial resolution down to several meters even for large fields (Robinson et al. 2008). Specifically, the strength of ER-based methods is their ability to provide high spatial resolution ER estimates, both in horizontal (up to tens of meters along 2-D transects or for 3-D soil volumes of tens of m3) and in depth (from cm to m, depending on the numbers of electrodes and their configuration), at suitable temporal resolution (from hourly to seasonal). The potential of the ER-based methods has been recognized for exploring soil-plant and SWC interactions for agricultural purposes (e.g., Garrè et al. 2013, Vanella et al. 2018). As example, the adoption of ER-based approaches in time-lapse mode contributes to further enhance their application for monitoring the SWC dynamics under different agricultural practices providing clear advantages in comparison to traditional point-based measurements (Blanchy et al. 2020).

In absolute terms, it is difficult to translate the geophysical proprieties into SWC characteristics, because their relationships tend to be not linear and represent the result of complex interaction processes (i.e., soil evaporation, RWU and soil water redistribution). These transfer functions lead to a site- and/or time-specific relationship and a continued need for in situ calibration (Tso et al. 2019). Additionally, ER inversion models can exhibit artefacts at shallow levels (< 1 m) possibly caused by errors in electrode spacing (Zhou and Dahlin 2003; Oldenborger et al. 2005) and/or greater Tsoil changes occurring during the ER surveys (each 1 °C change in Tsoil will not exceed an error of 2% in ER, Friedman 2005). In this study, it is possible to exclude errors due to the above-mentioned sources of uncertainties because the surface electrodes where installed every 0.15 m using a measuring tape as reference and the electrodes were left in place allowing the time-lapse ERI measurements to be taken at the same exact position. Moreover, the Tsoil showed variations of only 0.42 °C during the ERI surveys, resulting in a negligible effect on the inferred changes in SWC (Nijland et al. 2010).

Difficulties in the estimation of SWC are particularly relevant under heterogeneous crop** systems where the SWC values vary over (extremely) large ranges (Vereecken et al. 2014). Few studies have examined the impact of SWC changes in agro-systems characterized by different crop** ages (Wang et al. 2015, 2021). In this context, Wang et al. (2015) observed an inverse trend between the average decrease of SWC and the growth age in a multi-aged apple orchard, reporting that as the tree ages increased, the SWC decreased. Similar findings were obtained by Wang et al. (2021). These authors inferred deeper SWC drying patterns under mature apple trees as function of the tree ages and independently on the planting density. Specifically, they found a positive correlation between the greater growth age and the deeper root systems. This is expected especially in water-limited regions where roots search for water stored in deep soil under the driving force of plant growth (Wang et al. 2015).

This study represents an attempt to indirectly identify the influence of soil evaporation and transpiration processes active under multi-aged orange trees. This issue is particularly relevant for a better understanding of the water consumption of young trees during growth compared to mature trees (Alves et al. 2007). Specifically, herein, it has been observed that greater ER increases were detected under the soil of the younger citrus trees in comparison with the mature ones. This effect may be mainly ascribed to the higher soil evaporative process that occurred from the younger crops due to their lower vegetation groundcover and, concurrently, lesser root extension to depth (Chavarria and dos Santos 2012; Roberts 2000). Conversely, the higher ER decreases (resulting in SWC changes) were detected in the soil colonized by the mature crops, that were characterized by higher root biomass that may cause soil macro-porosity (Fig. 7). This empirical evidence is also supported by the higher ER values registered in correspondence of the soil domain under the mature trees (Fig. 5). This effect is related to the fact that the SWC at the initial condition, prior to irrigation, was lower under the mature trees in comparison to the young crops due to the higher transpiration rate (Fig. 4; Motisi et al. 2012; Saitta et al. 2020). Note that, during the reference period, the estimated daily average ETc values were 2.03 mm d− 1 and 1.68 mm d− 1 for the mature and younger trees, respectively. However, the integrated use of plant-based measurements (e.g., such stomata opening/conductance or sap flow measurements) together with ER-based surveys can provide advances for ERI-assisted irrigation optimization, allowing the respective contributions of plant transpiration and soil evaporation processes to be independently determined (Vanella et al. 2022).

In general, the monitoring of irrigation fronts is extremely important in order to regulate the volume of irrigation to be supplied and avoid water waste (Moreno et al. 2015; Vanella et al. 2021), especially in mixed-age crop contexts. Herein, the high-quality of the used ERI datasets allowed us to demonstrate the validity of this approach for providing practical insights for the efficient adoption of precision irrigation criteria. In particular, the ER-based imaging of the wetting fronts (Fig. 6) can help to support micro-irrigation management at the farm level by guiding the users (i.e., farmers, irrigation experts) during the operational steps for e.g., evaluating the performance of the irrigation systems and/or the during the irrigation modernization phases. As an example, the ER-based approach may help to detect the in situ metre-scale irrigation uniformity (Araya Vargas et al. 2021; Hardie et al. 2018) and/or to provide recommendations for varying the numbers of drippers within an irrigation line (Cassiani et al. 2015). The absence of ER lateral changes below soil depths of 0.5 m from the soil surface has contributed to evaluation of the suitability of irrigation scheduling applied under field conditions. Moreover, the proposed methodology, as shown in Fig. 8, starting from an initial value of SWC, permits the calculation of SWC variation for a certain decrease of ER (i.e., equal to the double of the ER decrease for the present case study).

Conclusion

This study explores the use of time-lapse ERI monitoring for capturing the wetting and drying patterns active in the soil under mixed-age crop conditions during an irrigation cycle. The results of this study corroborated the concept that the use of advanced geophysical monitoring techniques can quantify the temporal and spatial dynamics of SWC in the tree root zone. As a step-forward, this study provides clear evidence, from both ERI and point-based SWC measurements, that multi-aged trees have different drying and wetting patterns. Specifically, the integration of independent point-based SWC measurements has allowed the identification of a clear relationship between the soil ER characteristics and the SWC changes in the soil under the multi-aged crops.

Finally, this study shows the promising potential of ERI to support precision irrigation management, hel** optimize water use under mixed-age crop conditions and to the enhance the mechanistic understanding of the SWC relationships in such complex agro-system settings. In particular, the time-lapse ERI monitoring has identified trigger points for management intervention by imaging the wetting fronts and the SWC dynamics active under the multi-aged crops under study that are influenced by concurrent phenomena, such as soil evaporation (i.e., more evident the younger trees) and plant transpiration (i.e., more pronounced for the mature trees). Further outlook for this applied research will consist of coupling the ER-based observations with hydrological modelling and/or ancillary soil-plant measurements to advance the ERI-assisted irrigation optimization.