2019
Olive (
Soil salinity surveys and studies across the world and Israel indicate that irrigation with poor water quality and improper irrigation management causes soil salinization and degradation, and damages soil fertility (Wada et al., 2016; Pandit et al., 2020). Soil salinity monitoring in the Jezre'el Valley began in 1987, following a soil salinity survey that showed intensive salinization and often alkalinization of the upper soil horizons (Benyamini et al., 2005, 1998, 2000). Earlier studies had shown that these processes were enhanced by a semi-confined shallow aquifer (Kruseman and De Ridder, 1976), causing upward water flow during winter and spring seasons and reducing downward rain and irrigation percolation during the summer and fall seasons (Gafni and Salinger, 1992).
Most of the soil salinization problems in the Beit She'an Valley are
associated with the use of poor-quality irrigation water (conductivity above
3 dS m
While global-scale land surface–soil–biosphere–atmosphere models enable a regional water balance (Boone et al., 2017; Guimberteau et al., 2018; Katz et al., 2018), understanding water and solute movement processes in unsaturated soil layers requires a mathematical description and numerical model development (Leij et al., 1991; Simunek and van Genuchten, 1995; van Genuchten and Wagenet, 1989; Celia et al., 1990; Kool et al., 1985). Principal component analysis (PCA) suggests that soil hydraulic conductivity is one of the factors affecting soil quality (Mandal et al., 2008b). Water and solute movement models in an unsaturated soil layer are based on Richards' equation for one-dimensional movement of water under saturation variability (Celia et al., 1990; Bear, 1972), and root water uptake is calculated by the van Genuchten equation (van Genuchten, 1987). In such models, the soil hydraulic conductivity coefficient in the saturated media varies as a function of the soil's hydraulic conductivity.
Soil moisture may be evaluated through atmospheric conditions (Garrigues et al., 2015) or calculated as a function of suction (pressure head) and hydraulic conductivity in an unsaturated condition. Salt leaching and accumulation are significant in arid and semi-arid areas (Wada et al., 2016). Salt motion models are commonly based on the Fickian convection–dispersion equation for solute transport (Toride et al., 1993) and complex models that should also consider absorption processes, anion and cation exchange, and more. Several modeling platforms such as HYDRUS (Simunek et al., 1998) and WASTRC-1 (Mirlas et al., 2006b) are widely used. WASTRC-1, a one-dimensional water and solute movement model under saturated conditions, was found to fit the soil characteristics of Hula Valley irrigated fields in Israel. In both the HYDRUS and WASTRC-1 models, various soil hydraulic conditions such as drainage, irrigation, and layer saturation depth can be considered. Soil density, saturated hydraulic conductivity, field moisture, suction, and root zone development among other factors are prerequisites for model calibration, parameter validation, and, consequently, proper water and solute movement simulation (Garrigues et al., 2015).
Salinization during irrigation is a dynamic process as the number of salts in the soil and their composition change during irrigation in both the surface area and in the soil profile. Soil salinity mapping by the traditional sampling method is expensive and time-consuming, with mapping accuracy directly depending on the distance between the sampling points (Pandit et al., 2018). Remote sensing technologies that are based on active electromagnetic (EM) radiation are being widely adopted for soil salinity mapping. Ground-based EM methods measure electrical conductivity (EC) in subsurface and substratum horizons and can thus recognize salinity anomalies in the field before salinization approaches the surface (Farifteh et al., 2007). EM induction sensors measure the soil profile salinity by recording the soil's apparent electrical conductivity (ECa).
Frequency domain electromagnetic (FDEM) techniques are a powerful tool for mapping soils and detecting changes in soil types related to salinity. FDEM sensors work within a range of 30 cm to 5 m of depth and perform best while scanning the area from about 1 m above the ground (Ben Dor et al., 2009a). By applying FDEM with other active and passive remote sensing methods, EC values in given soil layers were attained for the soil in the Jezre'el Valley (Ben Dor et al., 2009b).
The soils of the Beit She'an Valley were selected for research as it is one of the most important agricultural areas in Israel. They consist of brown clay soils (grumusols) and calcritic soils, with the latter's profile characterized by thin layers and formation layers of marl with high water absorption capacity. The soil stratification influences the potential to drain and wash excess salts that accumulate during the irrigation season, which preserve ventilated root conditions. Sodium-rich soil has up to 30 % cation exchangeable capacity, which exacerbates the ventilation conditions necessary for plants. The combination of soil stratification and poor drainage conditions impedes plant development, and, in some cases, soil structure destruction and salt accumulation in the root zone cause plant degeneration due to water absorption difficulties (Machado and Serralheiro, 2017). Consequently, crop irrigation by brackish water in the Beit She'an area might cause economic damage.
The irrigation water sources in the area are of variable quality: springs and Jordan River water are considered of acceptable quality (fresh), while groundwater and effluent water might be of poor quality (brackish). In this latter case, irrigation without clear irrigation criteria might steadily damage soil fertility. Defining an irrigation regime for local soil and water quality conditions is therefore of great importance for preventing crop and economic damage in the Beit She'an Valley. The required knowledge should indicate how water and salt move in soil and correlate with salinity processes and irrigation management capability (Pandit et al., 2018). Combining remote sensing (FDEM) methods with water and salt movement models in the unsaturated soil layer may enable the effective identification of soil salinization processes. In turn, this may result in improved planning and control for irrigation systems.
As integrative knowledge of harvesting demonstration needed for irrigation management, this study's objective was to assess soil salinization processes because of low-quality irrigation at the Kibbutz Meirav olive plantation in the Beit She'an Valley.
The Beit She'an area is a unique agricultural area due to a combination of warm and dry climate (potential annual evaporation of 2400 mm at the meteorological service, Eden Farm Station), saline water irrigation, and heavy soils. The study site is a mature (2002) olive plantation located 1100 m north of the Kibbutz Meirav (Fig. 1).
Study site location (regional map: after CIA factbook, 2021; photo: Survey of Israel, 2021).
The planting intervals between the trees and rows are 7 and 4 m. The rainfall amount at the study site was 154, 253, and 281 mm in 2007/2008, 2008/2009, and 2009/2010 hydrological years, respectively. The soil at the study site is layered, with a practically impervious shallow layer of travertine found in different locations of the plantation as well as layers of marl at greater depth. Soil salinity stains were observed together with trees suffering from lack of ventilation, salting, and excess irrigated water. Following the soil sample particle size analysis results, the soil mechanical components at the research site consist of clay (40 %–50 %), silt (25 %–30 %), and sand (20 %–30 %) (Fig. 2).
The mechanical composition (soil texture triangle) of soil at the research site. The depth range of the travertine layer is from 110 cm at the southern edge to 55–60 cm at the northern edge of the site (Fig. 3).
Depth of the travertine layer from the soil surface in centimeters (background aerial photo is an insert from Fig. 1; orthophoto). 1 – lithological borehole; 2 – isoline of travertine layer depth from the soil surface.
The Kibbutz Meirav olive plantation irrigation water quality test results
for different seasons during the study period are presented in Table 1. The
main irrigation water sources in the area are Jordan River water and local
groundwater whose salinity and SAR (sodium adsorption ratio) are very high, mainly due to high
sodium chloride concentrations (Flexer et al., 2006). The chloride
concentration is in a range of 800–1700 mg L
Quality of irrigation at Kibbutz Meirav olive plantation.
The olive plantation drip irrigation regime in one extension along the row
that was used for the study calibration was 1 L s
This study integrates field experiments with water and salt movement models in the unsaturated soil strata. Field experiments including a remote sensing method (FDEM) were utilized to supply the required data for water and salt movement modeling and soil salinity mapping during soil salinization monitoring under different irrigation conditions (Corwin and Lesch, 2005). The suction and soil moisture monitoring during the irrigation period was conducted near two tensiometer stations characterizing suction and soil moisture conditions. The first station characterized irrigation by about 80 % of the acceptable amount of irrigation (lack of water) and the second station characterized irrigation by about 120 % of the acceptable amount of irrigation (excess water). The field experiment was conducted in spring before the beginning of summer irrigation, which made it possible to evaluate the soil salinization dynamics when water enters practically dry soil after winter precipitation salt washing. The experiment included soil sampling to measure soil moisture and soil salinity coupled to FDEM mapping. The integration of the various data processing types and modeling finally yielded a soil salinization spatial–temporal illustration of the different irrigation regimes (Fig. 4).
The conceptual working process applied to soil salinization assessment.
Continuous soil suction monitoring included two transmitting tensiometer stations (Mottes Tensiometers, Ltd.). The two stations were installed 50 m from
each other. At each station, four tensiometers were installed, measuring the
soil suction at depths of 20, 40, 60, and 70 cm from the soil surface and
under the olive tree rows. The tensiometer system sampled soil suction
values (in mBar) every 30 min that were transmitted to the company's
website (
Transmitting tensiometer station.
Soil moisture and salinity monitoring were made by simultaneous soil
sampling every 2 weeks from September until December 2011. Soil samples
were taken at depths of 0–20, 40–60, 20–40, and 70 cm or down to the depth of
the travertine layer. Drilling was done along the olive rows between the
trees. Each sample characterized a particular tensiometer depth as well as
the distance from the irrigation pipe and closest dripper. The laboratory
salt composition delineation included electrical conductivity (EC),
saturation percentage (SP), sodium adsorption ratio (SAR),
From 22 to 28 March 2011, a field experiment was conducted with the purpose
of obtaining soil salinity parameters before the irrigation season. These
parameters were needed to build and adapt the moisture and salt motion
model for the upper soil unsaturated layer. The experiment included moisture
and salinity measurements through manual soil sampling and FDEM soil
salinity mapping. Near each tensiometer station, three control lines were
marked perpendicular to the dripper line, whereas the line center was
positioned near the dripper. Near the first tensiometer station, the
distance between the control lines (1A, 1B, 1C) was 50 cm, and near the
second tensiometer station (2A, 2B, 2C) it was 40 cm (according to distance
changes between the drippers so that each control line was extended from the
middle between the three rows). Soil sampled for laboratory salt composition
and moisture tests was taken for each control line, at a central point
adjacent to the dripper, and 30, 80, 180, and 330 cm distances from the central
sampling point on both sides. Together with soil sampling, values of soil
suction from the tensiometers were also measured. The first soil sampling
was done at 08:30 UTC+2 before irrigation on control lines 1A and 2A. At 09:00 UTC+2,
drip irrigation began with an intensity of 1.6
Measurements were done along the control lines and in the area between the
tree row in the experiment site. Three measuring lines with 7 m length were
spaced 0.5 m apart near the first tensiometer station. The measurement lines
were made perpendicular to the irrigation dripper pipeline. Mapping was done
after 3 h of irrigation. The device was hung by a strap at the
height of 1 m above the ground, with the operator walking along the
sampling lines without stopping within the line. Five frequency channels
(62525, 22075, 7825, 2275, and 975 Hz) were used for characterizing soil
layer depth intervals at 0–30, 0–45, 0–60, 0–75, and 0–100 cm, respectively. Interpolation and spatial soil salinity mapping (in EC, dS m
The water and salt movement model in the upper unsaturated soil layer and up to the travertine layer was made in the HYDRUS 1D software. The one-dimensional model characterizes the cross section to a depth of 60 cm above the travertine layer.
The water and salt movement, a basic mathematical model of one-dimensional
equations for an unsaturated soil state, was
The hydraulic model used was the van Genuchten–Mualem (no hysteresis) single porosity model (van Genuchten, 1980). As a soil salinization model, Crank–Nicolson was used as a time weighting scheme (Crank and Nicolson, 1947) and the Galerkin finite scheme (Fletcher, 1983) was used for a space weighting scheme equilibrium model. For water movement relation with the root zone, the Feddes water uptake reduction model (Feddes et al., 2001) was used, with maximum concentration to passive root solute uptake of 0.5 (cRoot). The one-dimensional model calculated the volumetric moisture and total salinity in a soil profile down to the model's lower boundary. In the HYDRUS 1D software, the unsaturated layer parameters are automatically determined by the soil type. The lower boundary of the water movement model was calculated as a constant flow along with the travertine layer.
The irrigation input to the soil profile through the model's upper boundary
was calculated as the water supply according to an incremental irrigation
regime. Evapotranspiration and transpiration values as well as root zone activity
were determined from the field data and changed during the irrigation season.
The models were calibrated according to the field experiment data. The
calibrated model was used to assess and predict soil salinization due to
irrigation with different water quality: 3.13 dS m
Near the first tensiometer station at a depth of about 60 cm of the
travertine layer and irrigation by about 80 % of the acceptable amount of
irrigation (lack of water), the soil salinity was about 11–12 dS m
Changes in soil salinity (EC) and soil weight moisture (W) during the autumn near the first tensiometer station.
Changes in soil salinity (EC) and soil weight moisture (W) during the autumn near the second tensiometer station.
Near the second tensiometer station with a travertine layer depth of about
70 cm and irrigation by about 120 % of the acceptable amount of irrigation
(excess water), soil salinity was lower: between 2.0 and 4.0 dS m
The SAR values under irrigation conditions of about 80 % of the acceptable amount of irrigation ranged from 4 to 12. The SAR values increased with soil profile depth. Under irrigation conditions of about 120 % of the acceptable amount of irrigation (excess water), SAR values were found to be lower, ranging from 3 to 6, with an increase toward the upper soil layer.
Near the first tensiometer station with a depth of about 60 cm of the
travertine layer and irrigation by about 80 % of the acceptable amount of
irrigation (lack of water) at the end of September, the chloride
concentration was high throughout the soil profile and ranged from 3200 to
3500 mg L
Near the second tensiometer station with a travertine layer at a depth of
about 70 cm and irrigation of about 120 % of the acceptable amount of
irrigation (excess water), chloride concentrations were found to be lower,
ranging from 400 to 3000 mg L
The amount of general chalk in the soil was very high and hardly changed during the study period. The amount of general chalk ranged from 70 % to 85 % and did not depend on soil moisture and irrigation regime.
The soil suction that was measured in situ using the first tensiometer station is shown in Fig. 8. In station 1, the soil suction before irrigation varied from 140 to 300 mbar depending on the depth of the measured soil layer, while after irrigation it dropped to 40–130 mbar. Due to the highest moisture, the maximal soil suction decrease was observed in the upper soil layer (0–20 cm), while in the upper soil layer (0–20 cm) sinusoidal oscillations were observed due to daily (day–night) changes in temperature and humidity. At other depths, once settled, the suction had a small tendency to increase during the study period.
Soil suction on the different depths of the soil profile measured at the first tensiometer station.
Laboratory soil salinity measurements characterized the dissolved salt
concentration in the soil saturated solution near the drippers. Soil
salinity near the first tensiometer station ranged from 1.5 dS m
Soil salinity at different soil profile depths measured at the first tensiometer station.
The highest soil salinity from 4.1 to 7.4 dS m
During the study, the period-averaged weighted soil moisture varied from 0.25 to 0.30, with a dependency on distances from the dripper with an affected radius of up to 30 cm. Soil moisture increased with irrigation right away in the upper soil layer under the dripper from 0.14 to 0.37 after 2 and 22 h; after irrigation stopped it decreased to 23 % (Fig. 10).
Weighted soil moisture at different distances from the dripper pipeline around the first tensiometer station.
EC values obtained from FDEM measurements characterize the dissolved salt
amount and soil moisture. The maps show the salt flushing area progressing
to a depth of 50–60 cm (Fig. 11). The salt flushing area width was about 0.5 m, demonstrating EC lower than 2 dS m
FDEM EC values for the different depths of the soil profile from
the soil surface:
Correlation between the ratio of EC (sat) to EC FDEM and the weighted soil moisture at the Kibbutz Meirav mature olive plantation test site.
Thus, under a weighted soil moisture content of 0.2, the EC values obtained using the FDEM device (EC FDEM) will be approximately 3 times lower than those measured in the soil saturation extract laboratory measurements (EC (sat)). Provided the weighted soil moisture is greater than 0.32, EC measurements using FDEM would be close to the laboratory soil test results.
Fitting the model to the study site conditions was based on comparing the
model calculation results with measurements taken during the field
experiment. The comparison was made for soil volumetric moisture and soil
salinity values in EC. The best fit between model calculation and soil
mechanical composition field measurements was obtained for the silty clay
type soil (Fig. 2). The volumetric moisture model calibration was like the
calculated results (Fig. 13). The hydraulic conductivity of the soil
saturated conditions according to the model was 0.02
Comparison of volumetric moisture values measured in a study section with computerized values in the model. Model calibration results.
Differences between soil volumetric moisture measured in the field and
calculated in the model were maximal
In deeper layers, it started after 2 h, while 12 h after
irrigation ended soil suction began to rise owing to soil drying. Changes in
volume soil moisture were consistent with changes in soil suction. Moisture
increased immediately after irrigation in the upper soil layer from 0.33 to
0.36 almost to a saturated state. After 2 h (end of irrigation) the
moisture began decreasing. The results in the model show that in the deep
layers (below 30 cm from the soil surface) the moisture continued to
decrease, probably due to a rather small amount of water and irrigation span.
As a result of irrigation by relatively saline water (3.13 dS m
Changes in soil salinity in the soil profile calculated in the model. Calibration results of the model.
Figure 15 shows the process of salts accumulating (or washing away) at the
model's outer boundaries. Near the soil surface at the TOP boundary, the
salt concentration initially decreased due to soil washing by irrigation,
and after irrigation finished it gradually increased over 36 h owing to
evaporation from 2.5 to 4.7 dS m
Changes in salt concentration calculated in the model at the upper model boundary (TOP), at the root zone bottom (about 35 cm from the ground surface) (ROOT), and the lower model boundary (BOT).
Irrigation and evaporation data used in the model.
The soil-salinity-calibrated model simulates soil salinity patterns, with different water quality irrigation between April and December. The data input to the model for calculating the irrigation duration per day and daily evapotranspiration included the olive plantation irrigation and daily evaporation as well as evapotranspiration from the Eden Farm meteorological station (Table 2).
Initial salt concentration values in the model soil profile (the 1 April) were taken from the field experiment soil sampling results. Soil
profile and salt accumulation predictions during the irrigation season show
that under current water quality conditions (3.13 dS m
Salt accumulation predictions in the soil during the irrigation season
under different water salinity:
Irrigation by brackish water during the summer months (Fig. 16c) caused
substantial increases in soil salinity, reaching a very high EC value of
about 24–26 dS m
The combined use of various research methods, including soil salinity
monitoring, field experiments, remote sensing (FDEM), and water and salt
movement modeling in the unsaturated soil profile, allowed salinization
process assessment of calcritic soils in an irrigated olive plantation in
the Beit She'an Valley. Under the existing drip irrigation regime, water
with a dissolved salt content of 3.13 dS m
The FDEM device made it possible to study the dissolved salts' spatial
distribution and concentration, with reference to the soil's existing weight
moisture distribution at the time of measurement. The FDEM EC maps show the
salt flushing area development at a depth of 50–60 cm and a width of
The one-dimensional model created for water and dissolved salt transport showed the danger of using brackish water for irrigation. Since soil salinization exceeds an acceptable level for trees, the use of potable water for irrigation, if possible, will help to reduce soil salinization. To enable a tailor-made irrigation scheme, a database including changes in physical and chemical parameters affecting soil salting processes should be established, which will enable contemporary mapping and salinity forecasting as well as the effect of hydrochemical factors on various soil and irrigation conditions for the database-specific region.
Data and code are available in the Supplement.
The supplement related to this article is available online at:
VM designed and carried out the field experiments and performed the HYDRUS simulations. NG designed and carried out the field experiments and performed measurements and mapping of the soil electrical conductivity using an FDEM device. AA did the irrigation data processing, and YA analyzed the results and prepared the paper with contributions from all co-authors.
The contact author has declared that neither they nor their co-authors have any competing interests.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This article is based on the results of scientific research carried out as project 855-0066-12 “Identification and assessment of soil salinization risk as a result of agricultural activity in the Beit She'an Valley” by a team of researchers and technical personnel of the Soil Erosion Research Station of the Ministry of Agriculture and Rural Development (MOAG), Israel: Vladimir Mirlas (main researcher), Naftaly Goldshleger (researcher), Asher Aizenkod (researcher, field survey center, MOAG), Roei Ben-Binyamin (technician), Yochai Barnay-Betsalel and Moshe Gottesman (technicians), and Ivan Sapogineth (technician, the Eastern R&D Center, Ariel University). The authors particularly thank the technical team of the SERS and the Eastern R&D Center, Ariel University: Roei Ben-Binyamin, Yochai Barnay-Betsalel, Moshe Gottesman, Efraim Fizik, and Ivan Sapogineth for their assistance in the field investigations. The authors wish to thank David Wilson, Ruishan Chen, Douglas Karlen, and also the two anonymous reviewers for their helpful contribution to the paper. The paper is dedicated to our friend Naftali Goldshleger, who passed away in 2019.
This research has been supported by the Ministry of Agriculture and Rural Development (MOAG), Israel (grant no. 855-0066-12), and the Jewish National Fund (KKL–JNF) (grant no. 855-0066-12).
This paper was edited by Bethanna Jackson and reviewed by two anonymous referees.