This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Effect of irrigation water salinity on transpiration and onleaching requirements: A case study for bell peppers

Alon Ben-Gal a,*, Eviatar Ityel b, Lynn Dudley c, Shabtai Cohen d, Uri Yermiyahu e,Eugene Presnov a, Leah Zigmond f, Uri Shani g

aEnvironmental Physics and Irrigation, Agricultural Research Organization, Gilat Research Center, Mobile Post Negev 85280, Israelb Shaham-Extension Service, Ministry of Agriculture, IsraelcDepartment of Geological Science, Florida State University, Tallahassee, FL, United StatesdCentral and Northern Arava Research and Development, Hazeva, IsraeleSoil Chemistry and Plant Nutrition, Agricultural Research Organization, Gilat Research Center, Israelf Southern Arava Research and Development, Yotvata, IsraelgDepartment of Soil and Water Sciences, Faculty of Agricultural, Food and Environmental Sciences,

The Hebrew University of Jerusalem, Israel

1. Introduction

Maximization of crop yields when salinity of irrigation water is

high depends on providing plant water needs (transpiration, T)

and evaporative losses, as well as on maintaining minimum

soil solution salinity through leaching. Evapotranspiration

(ET) requirements are often estimated by measuring or

calculating potential ET (ETp), which is a function of climate,

and through the use of species dependant crop factors that

consider plant size (cover) and crop physiological stage (Allen

et al., 1998). Generally, salinity is not considered when

calculating ET from ETp, but it has been suggested that this

could lead to overestimation of ET due to the expected salinity-

caused reductions in T (Meiri et al., 1977; Letey and Dinar, 1986;

Dudley et al., 2008).

Salinity causes osmotic imbalance, reduced water uptake

and transpiration, and reduced yields (Bernstein, 1975).

Management of saline water for irrigation is often based on

application of excess water, designed to maintain minimum

root zone salinity and thus minimize salinity-caused yield

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7

a r t i c l e i n f o

Article history:

Received 24 July 2007

Accepted 27 December 2007

Published on line 13 February 2008

Keywords:

Leaching fraction

Irrigation management

Evapotranspiration

Drainage

Lysimeters

Yield

Capsicum annum

a b s t r a c t

Maximization of crop yields when the salinity of irrigation water is high depends on

providing plant transpiration needs and evaporative losses, as well as on maintaining

minimum soil solution salinity through leaching. The effect of the amount of applied

irrigation water was studied regarding transpiration, yields, and leaching fractions as a

function of irrigation water salinity. Bell pepper (Capsicum annum L. vars. Celica and 7187) in

protected growing environments in the Arava Valley of Israel was used as a case study crop

to analyze water quantity–salinity interactions in a series of lysimeter, field and model

simulation experiments. Leaching fraction was found to be highly influenced by plant

feedback, as transpiration depended on root zone salinity. Increased application of saline

irrigation water led to increased transpiration and yields. The higher the salinity level, the

greater the relative benefit from increased leaching. The extent of leaching needed to

maximize yields when irrigating with saline water may make such practice highly unsus-

tainable.

# 2008 Elsevier B.V. All rights reserved.

* Corresponding author. Tel.: +972 8 9928644; fax: +972 8 9926485.E-mail address: bengal@volcani.agri.gov.il (A. Ben-Gal).

avai lab le at www.sc iencedi rec t .com

journal homepage: www.e lsev ier .com/ locate /agwat

0378-3774/$ – see front matter # 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.agwat.2007.12.008

Author's personal copy

reduction (Ayers and Westcot, 1985). The leaching fraction (LF)

is the relative volume of applied water that carries salts out of

the root zone. The minimum LF that will keep the soil salinity

below a required level is the leaching requirement (LR). A

variety of formulations have been proposed for estimating the

LR, but all are based on a functional relationship between

irrigation water salinity and crop yield. The Food and

Agriculture Organization of the United Nations (FAO) recom-

mends computing LR as (Ayers and Westcot, 1985):

LR ¼ ECiw

5EC�e � ECiw(1)

where EC is the electrical conductivity, iw denotes irrigation

water, EC�e is the EC of the soil saturated paste extract corre-

sponding to the soil salinity tolerated by the crop. Values of

EC�e used to determine LR are usually either ECe of the thresh-

old value (ECe-0%) – meaning 0% yield decrease due to salinity

– or ECe-10%, reflecting a 10% yield loss. Examples of LR for bell

pepper irrigation at ECe-0% and ECe-10%, calculated according

to Eq. (1) and the corresponding relative water application

rates (I/ETp) are shown for irrigation water salinities of 0.5–

5.5 dS/m in Table 1. It has been suggested (Meiri and Plaut,

1985; Corwin et al., 2007; Letey and Feng, 2007; Dudley et al.,

2008) that calculating LR with formulas like Eq. (1) is imprecise

due to failure to consider soil type, climate, or salinity-induced

reduction in plant transpiration. Such omissions could possi-

bly result in underestimation of actual leaching and over-

estimation of LR.

A number of approaches based on our understanding of the

response of crops to water, salt tolerance and soil processes

including leaching exist that can be used to evaluate plant

response to both amount of applied water and salinity. These

approaches have lead to semi-empirical production functions

for specific crops (Letey et al., 1985; Letey and Dinar, 1986) and

to physically based conceptual models of water uptake as

reviewed by Hopmans and Bristow (2002) and Feddes and

Raats (2004). Such models allow consideration of environ-

mental factors and dynamic interactions within the soil–

water–plant system and enable prediction of crop response to

various irrigation regimes, calculation of LFs and evaluation of

LRs. Typically, the models calculate water uptake or tran-

spiration and their reduction due to insufficient amount of

soil–water and excess soil–water salinity. Examples of this

approach have been recently presented in models utilizing

both numerical (Dudley and Shani, 2003) and analytical (Shani

et al., 2007) solutions.

The analytical solution of Shani et al. (2007) predicts plant

performance under varied environmental, biological (crop)

and management parameters. The model assumes steady-

state conditions and representative root zone values for

salinity and moisture. Essentially, the model predicts the crop

response to conditions of soil–water and salinity, while

considering the influence of the plant itself on soil–water

content and salt concentration. Water uptake by plants, water

and salt leakage below the roots and yield are calculated by

solving for transpiration in a single mathematical expression

according to limitations imposed by root zone salinity and

water status. Input variables include the quantity and salinity

of applied water, plant sensitivity to salinity and water stress,

ETp, and soil hydraulic parameters. The model has been

shown to accurately predict measured results for cases where

irrigation is frequent and regular. The model facilitates

evaluation of the effect of irrigation water quantity on

transpiration and drainage and therefore allows prediction

of LFs for any irrigation water quantity–salinity combination.

We have chosen bell pepper (Capsicum annuum L.) growing

in the Arava Valley of Israel as a model crop for studying the

relationship between transpiration and water- and salt-stress.

In this arid region, bell pepper is economically important as a

winter crop, produced for export to European and North

American markets. Due to local water scarcity, only saline

groundwater with ECiw of 2.2–3.7 dS/m is available for

irrigation in the region. Protected (net house, greenhouse)

peppers, grown from August to May, will typically be irrigated

with 12,000–14,000 m3/ha of this saline water—an amount

believed by growers to maximize yields.

The pepper plant has a shallow root system, which extracts

70–80% of its water from the top 0.3 m soil layer (Dimitrov and

Ovtcharrova, 1995). This, together with high stomatal density,

explains why pepper is regarded as relatively vulnerable to

water stress. Bell pepper is considered moderately sensitive to

salinity. Maas (1990) reported an ECe threshold value of 1.5 dS/

m, below which no effect on growth is expected, and a 14%

decrease in biomass production for every additional 1 dS/m.

Recent studies have reported varied responses of pepper to

salinity. For greenhouse peppers thresholds ranging from 0 to

2 dS/m and slopes defining linear decrease in yield due to

subsequent increase in salinity ranging from 8 to 15% have

been reported (Sonneveld, 1988; Chartzoulakis and Klapaki,

2000; Navarro et al., 2002). Navarro et al. (2002) suggested that

newer commercial varieties may be more sensitive to salinity

than older ones. Yermiyahu et al. (2008), working with the

‘‘Celica’’ variety used in this study, reported 12% shoot

biomass reduction for every 1 dS/m increase in ECe, a value

similar to that reported by Maas, starting with their lowest

salinity of ECe = 0.8 dS/m.

The objective of the current study was to evaluate the effect

of irrigation water application rates on transpiration, yields,

and LFs as a function of irrigation-water salinity. The specific

case study of water quantity–salinity interactions for bell

Table 1 – Leaching requirement (LR) for peppers accord-ing to Food and Agriculture Organization (Ayers andWestcot, 1985)

ECiw 90% potential yield 100% potential yield

LR I/Ep LR I/Ep

0.5 0.05 1.05 0.07 1.08

1 0.10 1.11 0.15 1.18

2 0.22 1.29 0.36 1.57

3 0.38 1.60 0.67 3.00

5.5 1 �12 2.75 �1260

Based on saturated paste solution electrical conductivity (ECe)

threshold of 1.5 dS/m and ECe causing 10% yield decrease of 2.2 dS/

m. I/Ep is equivalent irrigation rate (irrigation relative to potential

evaporation) for each LR. ECiw is irrigation water electrical

conductivity.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7588

Author's personal copy

pepper (Capsicum annum L. vars. Celica and 7187) cultivation in

protected growing environments in the Arava Valley of Israel

was investigated in a series of varying-scale experiments. The

theoretical effects of water application rate and salinity

combinations and calculations of predicted LFs were produced

using the Shani et al. (2007) analytical model. Lysimeter

experiments provided precise control of the irrigation water

quantity and quality variables and a unique opportunity to

measure water and salt balance components and to calculate

ET for individual plants. Semi-commercial-scale field experi-

ments provided full season growth with economic (fruit) yield

for a number of water quality–irrigation regime combinations.

2. Materials and methods

2.1. Lysimeter experiments

Two lysimeter experiments were conducted using an auto-

matically rotating platform in a 50-mesh net house at the

Arava Research and Development facilities located at Yotvata

in the southern Arava Valley of Israel (298530N, 35840E). Twenty-

four 100 l lysimeters were filled with Arava sandy loam soil.

Shani et al. (1987) give detailed properties of the soil. The

rotating platform (Lazarovitch et al., 2006) promoted uni-

formity of environmental conditions among the individual,

free standing, lysimeters. The system automatically prepared

and delivered irrigation solutions and automatically collected

and measured drainage water. Each lysimeter had a highly

conductive rockwool drain to control matric potential at the

lysimeter bottoms. A detailed description of the lysimeter

system is given by Ben-Gal and Shani (2002).

The experiments were conducted in the fall/winter seasons

of 2000/1 (LYS 1) and 2002/3 (LYS 2). In each case, the 24

lysimeters were used to study nine irrigation water salinity

levels and six relative water quantities (Tables 2 and 3). A pair

of pepper seedlings was transplanted into each lysimeter

during the month of September: LYS 1 seedlings were planted

September 29, 2000 and the plants were removed 120 d later on

27 January 2001; LYS 2 seedlings were planted 19 September

2002 and the plants were removed 160 d later on 25 February

2003. Irrigation water was supplied on a daily basis relative to

the maximum ET (ETmax) measured within each experiment.

Daily ET was calculated for each lysimeter using:

ET ¼ I�Drþ DW (2)

where I is the irrigation, Dr is the drainage, DW is change in

soil–water determined from changes in lysimeter weight.

Target relative water quantities in LYS 1 were 40, 75, 110,

130, 150, and 175% of ETmax and in LYS 2 were 25, 50, 75,

100, 110, and 150% of ETmax. Salinity levels included ECs of 0.5,

1, 1.5, 2, 3, 4, 5.5, 7, and 9 dS/m irrigated with 130% ETmax in LYS

1 and with 110% ETmax in LYS 2. Salinity levels are prior to

addition of fertilizers. Salinity was increased by adding 1:1

molar concentrations of NaCl and CaCl2 to the low salinity

(EC = 0.5 dS/m) water. Salinity and water quantity treatments

were initiated 1 week after planting; until then, all the

Table 2 – Irrigation treatments and fruit- and shoot-yields for lysimeter-grown bell peppers in the Arava Valley during theLYS 1 experiment

Salinity ECiw

(dS/m)Irrigation (L) ET (L) Irrigation

rate (I/ETmax)Fruit Shoot

dry (g)Total biomass

dry (g)# Fresh (g) Dry (g)

0.5 61.4 59.3 0.38 19 1392 108.8 116 224.8

0.5 156.4 147.3 0.98 27 3776 211.7 180 391.7

0.5 200.9 157.5 1.25 28 3406 179.3 165 344.3

0.5 221.7 166.6 1.38 26 3571 197.8 197 394.8

0.5 259.1 180.2 1.62 30 4050 255.3 174 429.3

0.5 296.9 160.3 1.85 21 3589 261.2 190 451.2

1 216.4 162.3 1.35 22 3522 263.4 195 458.4

1.5 218.5 161.6 1.36 19 3672 267.1 188 455.1

2 217.4 161.4 1.36 24 3683 273.2 183 456.2

3 58.2 52.9 0.36 10 1207 73.5 77 150.5

3 141.8 122.8 0.88 21 2601 207.2 133 340.2

3 192.2 127.9 1.20 17 2686 185.8 121 306.8

3 206.6 136.6 1.29 20 2942 216.5 144 360.5

3 246.3 139.9 1.54 21 3022 197.1 136 332.6

3 281.6 144.4 1.76 19 2910 223.3 153 376.3

4 208.8 124.0 1.30 20 2444 153.2 123 276.2

5.5 59.2 53.7 0.37 14 787 71.2 53 124.2

5.5 149.5 105.8 0.93 14 1141 85.1 82 167.1

5.5 206.2 126.1 1.29 13 1696 126.9 92 218.9

5.5 219.8 109.8 1.37 13 1682 129.9 102 231.9

5.5 251.7 121.9 1.57 15 2045 162.4 101 263.4

5.5 294.5 128.4 1.84 16 1933 152.8 96 248.8

7 221.3 82.8 1.38 13 1191 97.3 51 148.3

9 223.3 90.1 1.39 9 846 75.5 59 134.5

ECiw is irrigation water electrical conductivity, ET is evapotranspiration.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7 589

Author's personal copy

lysimeters received equal excess quantities of EC = 0.5 dS/m

water with nutrients. Nutrients were provided in constant

concentrations with the irrigation water which contained

6.3 mM N (70% NO3 and 30% NH4), 0.72 mM P, 4.5 mM K,

5 mM Ca, 0.5 mM Mg, 0.5 mM SO4, 0.09 mM B, 9.3 mM Fe,

5.0 mM Mn, 2.1 mM Zn, 0.3 mM Cu, and 0.135 mM Mo. Fertilizers

added approximately 0.5 dS/m EC to the irrigation water of all

treatments. Yield was measured as fresh fruit-, dry fruit-, and

shoot-biomass.

2.2. Field experiments

Peppers were grown according to commercial greenhouse

practices under varied irrigation regimes and irrigation water

salinities at two locations in the central (FIELD 1) and northern

(FIELD 2) Arava Valley. Soil was prepared as beds with 1.6-m

spacing. Two drip irrigation laterals (Netafim, Hatzerim Israel;

Ram 1.6 l/h drippers) were placed 20 cm apart in the center of

each bed. Plants were grown in two rows adjacent to the drip

laterals with plants every 40 cm along the laterals, such that

plant density was 3.1 plants m�2. The peppers were trellised

according to the ‘‘Spanish’’ method between cordons of metal

wire to a height of 3 m. In each experiment, the high-salt-

content local irrigation water was compared with desalinated

water under a range of water application rates. Relative water

quantities were determined as a function of reference (Class A

pan) evaporation (Ep), measured outside the greenhouse.

2.2.1. FIELD 1 experimentIn a plastic-covered greenhouse at the Zohar experimental

station (Arava Research and Development, Moshav Ein

Tamar, Israel; 308570N, 358230E) peppers (var. Celica) were

grown under two irrigation water salinities and three

irrigation regimes. Soil in the greenhouse was loamy sand

(87% sand, 8% silt, 5% clay). Irrigation water salinities prior to

addition of fertilizers were ECiw = 0.5 dS/m and ECiw = 3.2 dS/

m. Concentrations of some mineral ions in the water before

fertilizers were added are listed in Table 4. Fertilizers were

added to equalize Ca, Mg and SO4 and to bring N, P and K to 6.3,

0.72 and 4.5 mM, respectively. Fertilizers increased EC by

0.8 dS/m in the desalinated and 0.5 dS/m in the saline water.

Irrigation water was applied three times a day at target rates

of 50, 75 and 100% of Ep. The experimental design was

complete random blocks in four replicates. Pepper seedlings

were planted 22 September 2004. Salinity treatments were

initiated at planting. Irrigation of 4 mm/d was applied

uniformly for 1 month, after which the water quantity

treatments were begun. Fruit was harvested as it ripened,

starting in December, 103 d after planting. After completion

of fruit harvest in May 2005, whole plants were removed to

measure fresh- and dry-biomass.

Table 3 – Irrigation treatments and fruit- and shoot-yields for lysimeter-grown bell peppers in the Arava Valley during theLYS 2 experiment

Salinity ECiw

(dS/m)Irrigation (L) ET (L) Irrigation

rate (I/ETmax)Fruit Shoot

dry (g)Total biomass

dry (g)# Fresh (g) Dry (g)

0.5 111.4 100.1 0.28 17 1919 178.3 67.3 252.6

0.5 222.5 190.6 0.56 36 3775 347.9 126.2 486.1

0.5 314.4 243.7 0.80 27 4157 385.5 166.7 577.2

0.5 408.9 306.7 1.03 34 5162 436.4 220.8 677.2

0.5 454.5 319.2 1.15 38 5658 511.1 204.1 747.2

0.5 608.5 395.1 1.54 45 6084 544.4 141.3 696.7

1 453.9 283.4 1.15 26 3902 356.2 196.4 591.6

1.5 455.4 268.0 1.15 28 3645 324.3 134.0 476.3

2 447.0 282.1 1.13 31 4144 347.1 134.1 494.2

3 105.4 94.7 0.27 22 1370 113.8 47.4 163.2

3 206.1 150.7 0.52 16 1935 164.5 55.0 227.5

3 308.9 200.5 0.78 21 2724 257.5 96.3 362.8

3 422.6 236.8 1.07 23 2644 247.6 114.1 368.7

3 446.4 213.7 1.13 12 1924 186.7 66.3 274.0

3 595.3 285.7 1.51 28 3237 245.4 113.3 366.7

4 445.7 226.5 1.13 20 2034 194.4 88.4 289.8

5.5 106.1 70.3 0.27 10 465 39.5 22.8 64.3

5.5 212.3 130.7 0.54 25 897 78.5 35.1 115.6

5.5 299.4 151.6 0.76 12 886 83.6 41.4 125.0

5.5 403.0 180.8 1.02 10 839 72.9 53.3 126.2

5.5 456.7 235.1 1.16 21 2720 242.0 98.7 355.7

5.5 613.0 255.4 1.55 17 2273 189.5 89.5 287

7 451.8 168.9 1.14 8 502 47.2 24.9 75.1

9 452.9 180.9 1.15 7 514 42.6 25.4 85.0

ECiw is irrigation water electrical conductivity, ET is evapotranspiration.

Table 4 – Pre-fertilized water quality of field experiments

HCO3 SO4 Cl(mM)

Ca Na pH EC(dS/m)

Site

1.5 1.3 2.8 1.3 2.5 6.9 0.7 Zohar

3.6 4.7 10.4 4.3 10.4 7.6 2.4

0.4 2.1 1.1 0.5 0.8 6.7 0.5 Yair

2.8 4.4 18.3 5.5 14.8 7.2 3.2

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7590

Author's personal copy

2.2.2. FIELD 2 experimentPeppers (var. 7187, Gadera Seeds) were grown under two

irrigation water salinities and four irrigation regimes in a 50-

mesh net house at the Yair experimental station (Arava

Research and Development, Hazeva, Israel; 308460N, 358140E).

Underlying soil was highly variable, highly stony alluvial

Hamada; >40% of the particles were >2 mm in size. Cultiva-

tion trenches 40 cm wide and 20 cm deep were dug in this soil

and filled with 70% volcanic tuff (1–8) and 30% compost.

Planting and irrigation were confined to the media-filled

trenches. Irrigation water salinities were EC = 0.7 dS/m and

EC = 2.4 dS/m. Concentrations of some mineral ions in the

water before fertilizers were added are listed in Table 4.

Fertilizers were added to equalize Ca, Mg and SO4 and to bring

N, P and K to 6.3, 0.72 and 4.5 mM, respectively. Fertilizers

increased EC by 0.8 dS/m in the desalinated and 0.5 dS/m in

the saline water. Irrigation water was applied once daily at

target rates of 25, 50, 75 and 100% of Ep. The experimental

design was complete random blocks in four replicates. Pepper

seedlings were planted 24 August 2004. Salinity treatments

were initiated at planting. Irrigation of 4 mm/d was applied

uniformly for 1 month, after which the water quantity

treatments were begun. Fruit was harvested as it ripened

starting 7 December 2004. After completion of fruit harvest in

April 2005, whole plants were removed to measure fresh and

dry biomass.

2.3. Simulations

Shani et al’s (2007) steady-state analytical model combines

water and salt balance with calculation of root zone soil

moisture and soil hydraulic conductivity according to the

Brooks–Corey (Brooks and Corey, 1966) soil hydraulic model.

Transpiration rate is solved from a single mathematical

expression and is the product of the soil’s unsaturated

hydraulic conductivity and the gradient of water potential

between soil and root (Nimah and Hanks, 1973). Yield decrease

as a function of salinity is considered by a reduction term

characterized by a logistic curve with an initial plateau and

subsequent decreasing section (van Genuchten and Hoffman,

1984). The salinity reduction curves are plant specific, defined

mainly by the ECe causing 50% yield decrease and are easily

adapted from experimental data or from published salinity

tolerance tables (for example; Maas, 1990). A proportional

relationship is assumed to exist between the ratio of yield to

potential yield and the ratio of transpiration to potential

transpiration following de Wit (1958) and Hanks (1974) thus

allowing prediction of biomass production (yield).

The model was used to simulate transpiration or total

biomass production of peppers in lysimeter and field experi-

ments. Soil and plant parameters used as model input are

given in Table 5. The model computed transpiration (T),

drainage (D), and LF for combinations of irrigation water

salinities ranging from 0 to 9 dS/m, and for irrigation levels

relative to potential transpiration (I/Tp) from 0 to 2.5.

Measured data from each experiment was compared to

predicted values according to the model using regression

analysis with the null hypothesis that slopes and intercepts of

the linear regression were not different from 1 and 0 at 95%

confidence.

3. Results

3.1. Lysimeter experiments

Drainage water salinity became a function of irrigation water

salinity within a matter of days following initiation of water

quality treatments (Figs. 1 and 2). With application rates of

I = 1.3 ETmax, salinity of drainage water reached ECiw approxi-

mately 10 d after treatments were initiated. The higher

salinity levels reached approximately constant, or steady

state, values within 4 weeks, and EC of the leachate for those

treatments was 1.5–2 times the value of ECiw (Fig. 1). Salinity of

Table 5 – Soil and plant parameters for lysimeters andmodel input parameters

Soil Arava sandy loam

KS (mm/d) 3600

d (unitless) 4.91

b (unitless) 0.55

us (m3/m3) 0.41

ur (m3/m3) 0.06

cw (mm) �200

Plant Capsicum annum cv. Celica

croot (mm) �6000

ECe50 (dS/m) 2.5

Tp (mm/d) 5

Yr0 0.08

From Shani et al. (2007): Ks, saturated hydraulic conductivity; d

andb, empirical soil characteristic parameters for the Brooks and

Corey (1966) hydraulic model; cw, air entry value; us, soil–water

content at saturation; ur, residual soil–water content; croot,

minimum possible water head at the root soil interface; ECe50,

plant characteristic parameter for salinity response function (EC of

the soil saturated paste where Yr = 0.5); Tp, potential transpiration;

Yr0, relative yield resulting from initial soil–water.

Fig. 1 – Solution electrical conductivity (EC) of drainage

water leaving root zone as a function of time and irrigation

water EC (ECiw). Peppers grown in lysimeters in the Arava

Valley, Israel, during LYS 1 experiment. Target leaching

fraction was 0.22 with irrigation (I) rate relative to

maximum evapotranspiration (ETmax) of I = 1.3 ETmax.

Arrow shows beginning of salinity treatments.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7 591

Author's personal copy

leachate from the lower- and mid-level salinity treatments

continued to increase throughout the season, reaching levels

up to four times their corresponding ECiw values. The low-

salinity water treatment had very low volumes of drainage

water when the water application level was designed to

exactly meet ETp requirements (I/ETp � 1); therefore, only a

few samples from this treatment were available for analysis

(Fig. 2A). When irrigation water was saline, drainage was

produced even when I/ETp was<1 (Fig. 2B and C). For both ECiw

of 3 and 5.5 dS/m, the lower I/ETp levels had drainage water

that was characteristically low in the first part of the season

and that became relatively high from about mid-season. From

mid-season onwards, drainage water salinity became a

function of application rate. For all three salinity levels, the

plateau or final drainage water salinity was lower when

irrigation quantity was larger (Fig. 2).

ET calculated from the water balance in the lysimeters was

reduced as a function of increasing ECiw (Fig. 3). In both

experiments, reduced ET due to salinity was noticeable

starting in the first weeks of treatment applications. Water

uptake was a function of climate and canopy cover, with ET

rates increasing and decreasing, accordingly. Water applica-

tion rates less than I = ETmax reduced ET (Fig. 4). Water

application rates where I� ETmax generally produced pepper

plants whose ET levels were not distinguishable.

Yield responses to increased ECiw were found in the

lysimeter experiments, both as decreased fresh fruit weight

and as decreased shoot weight (Tables 2 and 3). Relative

decreases in yield due to salinity were larger in LYS 2 when the

LF was lower (I/Tp = 1.1) as compared to the relative yield

decrease due to salinity in LYS 1 when the LF was higher (I/

Tp = 1.3) (Fig. 5). Rates of yield-decrease as ECiw increased from

0.5 to 2 dS/m were substantial when LF was lower and

insignificant when LF was higher. Pepper response to ECiw

in LYS 2 (I/Tp = 1.1) suggests a sigmoid-shaped–response curve

with a threshold value, followed by a steep decrease in Y as a

function of salinity and eventual tapering off of the negative

effect. The response in LYS 1 (I/Tp = 1.3) was characterized by

large negative effects beginning with the lowest increases of

salinity and a tapering trend of reduced Y with increased ECiw,

similar to that in LYS 2 (Fig. 5).

Relative biomass production of peppers as a function of

relative irrigation water application was similar for low EC

(ECiw = 0.5 dS/m) water (Fig. 6) for the two experiments. With

low salinity water, yield increased with increased water

Fig. 2 – Solution electrical conductivity (EC) of drainage

water leaving root zone as a function of time and irrigation

level (I/ETp) for irrigation water salinity levels:

(ECiw) = 0.5 dS/m (A), ECiw = 3 dS/m (B) and ECiw = 5.5 dS/m

(C). Peppers grown in lysimeters in the Arava Valley,

Israel, during the LYS 1 experiment. Arrow shows

beginning of salinity treatments.

Fig. 3 – Cumulative evapotranspiration (ET) for peppers in

lysimeters with variable salinity of applied irrigation

water (ECiw). Calculated from daily water balance of two

pepper plants in single lysimeters. (A) LYS 1; 2000–2001

season; I/ETp � 1.3. (B) LYS 2; 2002–2003 season;

I/ETp � 1.1.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7592

Author's personal copy

application until irrigation levels of I/ETp = 1, while further

increases in water application did not affect yield. For ECiw of 3

and 5.5 dS/m relative yields continued to increase with

increased water application beyond I/ETp = 1. The relative

yields found in LYS 2 were reduced more than those of LYS 1

suggesting that LF was not the only factor contributing to the

salinity response.

3.2. Field experiments

Field data for yields of peppers as function of application rate

and salinity of irrigation water closely followed the results

from the lysimeter experiments and simulations (Table 6,

Fig. 7). Maximum yields were found at irrigation rates just

above I/ETp = 1 for low EC (desalinated) irrigation water and

yields continued to increase with increased water application

for the saline irrigation water treatments. Relative fruit yields

were less affected by salinity at high irrigation rates than were

the shoot yields. Noticeable were the relative fruit yields in

both field experiments, which were found to be higher than

the relative shoot yields, suggesting that fruit yield was

reduced less by salinity and more positively influenced by

increasing LF than was shoot biomass production.

3.3. Modeled results

Simulated results for the cases of the LYS and FIELD

experiments, shown in Figs. 5–7 as lines, can be compared

Fig. 4 – Cumulative evapotranspiration (ET) for peppers in lysimeters with variable irrigation rates of applied irrigation water

(I/ETp) for irrigation salinities of 0.5 (A and D), 3.0 (B and E) and 5.5 (C and F) dS/m. Lines are daily water balance results for

two pepper plants in single lysimeters. (A–C) are 2000/2001 growing season; (D–F) are 2002/2003 season.

Fig. 5 – Total biomass production of peppers as a function

of irrigation water electrical conductivity (EC). Symbols are

experimental data (open 2000/2001, I/ETp = 1.3; closed

2002/2003. I/ETp = 1.1) and lines are simulated results.

I/ETp is irrigation relative to potential evapotranspiration.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7 593

Author's personal copy

to the experimental data shown as symbols. Simulations of

yield response to increasing ECiw, considering the different I/

Tp rates, predicted more severe yield decreases for equal

salinity levels when target LFs were lower (Fig. 5). The

simulated yields from the analytical solutions however,

over-predicted the extent of yield response to salinity in

LYS 1 (lower LF) and under-predicted the same response in LYS

2 (higher LF). Regression of predicted versus measured data,

from all of Fig. 5, resulted in r2 = 0.82 with slope of 1.047 and

intercept of �0.0168 (not different from 1 and 0 at 95%

confidence). Inspection of Fig. 5 suggests that the average of

the data from the two experiments corresponded closely to

the modeled results.

Measured yields were higher than simulated yields for

ECiw = 0.5 dS/m when I/Tp < 1 in both LYS experiments (Fig. 6).

Simulation of response to irrigation level at higher salinities

over-predicted yields for LYS 2 and under-predicted yields for

LYS 1 in the same manner as that seen with the predicted

salinity–response function. As with the salinity–response

data, the results of the response to water levels from the

simulations showed good agreement with the combined data

from the two experiments. Regression of predicted versus

measured data, from all of Fig. 6, resulted in r2 = 0.78 with

slope of 1.046 and intercept of�0.025 (not different from 1 and

0 at 95% confidence).

Simulation for the I/Tp, ECiw combinations of the field

experiments, using soil parameters from the lysimeter

experiment, were in agreement with measured shoot biomass

data. The actual soil hydraulic parameters for the sandy soil in

FIELD 1 (Fig. 7B) should be similar to those used in the

simulation, while such parameters for the soil-less media-

filled trench in FIELD 2 (Fig. 7A) are unknown, thus adding to

the level of uncertainty concerning the validity of simulation.

Regression of predicted versus measured dry biomass data,

from all of Fig. 7, resulted in r2 = 0.76 with slope of 0.80 and

intercept of 0.163 (not different from 1 and 0 at 95%

confidence). Regression of predicted biomass versus measured

fresh fruit data, on the other hand, resulted in r2 = 0.83 with

slope (0.67) and intercept (0.135) that were statistically

different than 1 and 0 at 95% confidence.

4. Discussion

In view of the relative overall success of the simulation model

in predicting trends and levels of pepper response to irrigation

level/salinity combinations, the model may be used to conduct

sensitivity and optimization exercises. The simulations

demonstrated that the yield response of pepper to ECiw is a

function of the irrigation rate (Fig. 8). High rates of leaching

allow higher transpiration and biomass production and

essentially create threshold values below which ECiw does

not effectively reduce yields. Irrigation rates close to ETp are

characterized by dramatic yield decreases with increases in

ECiw. Attempted ‘‘deficit’’ irrigation (I/ETp < 1) causes yield

decreases at all salinities and essentially increases the slope of

the ECiw response curve. The ECiw has a substantial influence

on actual LF (Fig. 9). For a given I/ETp, the drainage component

Fig. 6 – Total biomass production of peppers as a function

of relative applied irrigation water quantity (I/ETmax).

Symbols are experimental data (open 2000/2001; closed

2002/2003) and lines are simulated results for three

irrigation water salinity (ECiw) levels.

Table 6 – Irrigation treatments and average fruit- and shoot-yields for net- and greenhouse-grown bell peppers in theArava Valley (standard deviations)

Site Media Salinity(dS/m)

Irrigation(mm)

Total fruit(Mg/ha)

Weight(g/fruit)

Shoot dry weightkg/plant)

Zohar Sand 0.5 885 134.9 (10.7) 233.2 (6.0) 0.24 (0.06)

Zohar Sand 0.5 1245 131.4 (7.7) 233.6 (14.8) 0.32 (0.06)

Zohar Sand 0.5 1543 125.6 (7.2) 241.3 (10.4) 0.30 (0.08)

Zohar sand 3.2 800 98.5 (13.9) 241.2 (16.4) 0.18 (0.06)

Zohar Sand 3.2 1284 118.8 (13.5) 240.6 (17.5) 0.21 (0.06)

Zohar Sand 3.2 1621 124.9 (14.7) 232.5 (5.1) 0.22 (0.02)

Yair Tuff 0.7 350 73.0 (12.0) 163.4 (8.4) 0.19 (0.03)

Yair Tuff 0.7 684 109.5 (12.3) 168.6 (6.8) 0.25 (0.05)

Yair Tuff 0.7 962 107.2 (12.6) 162.1 (7.2) 0.29 (0.07)

Yair Tuff 0.7 1350 106.1 (10.4) 167.0 (6.0) 0.28 (0.04)

Yair Tuff 2.4 350 59.5 (12.5) 150.1 (7.8) 0.13 (0.02)

Yair Tuff 2.4 684 85.2 (12.9) 173.3 (9.6) 0.21 (0.02)

Yair Tuff 2.4 962 93.5 (12.6) 173.1 (6.6) 0.22 (0.05)

Yair Tuff 2.4 1350 107.2 (13.8) 173.3 (5.0) 0.28 (0.06)

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7594

Author's personal copy

of the water balance increases if ECiw increases due to lowered

ET caused by the higher salinity conditions. The leaching

fraction of peppers for the case simulated in Fig. 9 (Arava,

sandy loam soil) and a target LF of 0.17 (I/ETp = 1.2) was 0.2 for

ECiw = 0.5 dSm, 0.38 for ECiw = 2 dS/m and 0.5 for ECiw = 3.5 dS/

m. Up to half of the irrigation water became drainage as

salinity decreased yield to 60% of that expected for non-saline

conditions. Irrigation rates corresponding to those in com-

mercial pepper production and those applied to achieve

maximum yields in the field experiments (I/ETp rates between

2 and 2.5) produced LFs reaching 0.6. An additional interesting

result of the simulation of peppers (Fig. 9) is that LF is seen to

be relatively high even under conditions of attempted deficit

irrigation where less than ETp is supplied. Due to the self

regulating nature of the system – in which crop, water, and

salts interact – at low water application rates salts accumulate

in the root zone, transpiration is decreased, and eventually,

water content becomes sufficient to produce drainage and to

leach at least some of the salts beyond the active roots. These

results support early claims by Meiri et al. (1977) and those

previously documented by Letey et al. (1985), Letey and Dinar

(1986), and Solomon (1985). Letey et al. (1985) and Letey and

Dinar (1986) give examples showing LF as a function of relative

applied irrigation water for tall fescue and alfalfa grown under

different irrigation water salinities. As with our results, their

model indicated that at equal irrigation rates, LFs will always

be higher for higher irrigation water salinities and that salinity

results in LFs that are greater than 0. Their results show

irrigation water salinity-specific LF base levels that remain

constant as I/ETp increases from 0 to 0.8 and subsequent LFs

that increase with increasing irrigation. The predicted

Fig. 7 – Marketable fruit and dry biomass production of

peppers (Y) as a function of relative applied irrigation

water quantity (I/ETp). Symbols are experimental data and

lines are simulated results for each irrigation water

salinity (ECiw) level.

Fig. 8 – Simulated yield (Y) response (total biomass) of

peppers grown in sandy loam soil in the Arava Valley,

Israel, as a function of irrigation water salinity and relative

water application rate (I/ETp). Shani et al., 2007 model

parameterization according to Table 5.

Fig. 9 – Simulated relative biomass production (A) and

leaching fraction (B) as a function of relative applied

irrigation water quantity for three irrigation water salinity

(ECiw) levels. Shani et al. (2007) model parameterization

according to Table 5. Y is yield, T is transpiration, ET is

evapotranspiration, p is potential, D is drainage, I is

irrigation, ECiw is irrigation water electrical conductivity.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7 595

Author's personal copy

phenomenon in our work of increased leaching accompanying

apparent ‘‘deficit’’ irrigation strategies, when I/ETp was very

low (Fig. 9B) was not reported by Letey et al. (1985) or Letey and

Dinar (1986) but is consistent with results from transient,

numerical model simulations reported by Dudley et al. (2008).

The extent of drainage due to the plant’s response and

feedback to the soil–water system may be magnified by some

of the underlying assumptions of the analytical and produc-

tion function models that include; steady-state conditions and

non-consideration of the possibility that root uptake may be

greater in local zones of low salinity and high water content in

the soil (Corwin et al., 2007; Letey and Feng, 2007; Dudley et al.,

2008). In spite of this, the lysimeter experiments give

additional experimental confirmation to the models as they

demonstrated the occurrence of leaching that increased over

the course of time for irrigation application rates below

potential ET.

Comparison of LFs measured experimentally with LRs

calculated according to common methods (Ayers and Westcot,

1985) shows that, for any given ECiw, target LFs were close to

those maximizing yields but that they did not succeed in

bringing biomass production to predicted levels. For example,

irrigation with EC = 3 dS/m was expected to provide 10% and

0% yield decreases with LR = 0.38 (I/Ep = 1.6) and LR = 0.67 (I/

Ep = 3), respectively (Table 1). In the lysimeter and simulated

results, maximum relative yields of less than 80% were

achieved as LFs approached 0.5 (I/Ep = 2) and raising LF further

provided very little additional yield (Fig. 6). Leaching require-

ments are based on crop response to salinity databases that

refer to biomass production and not necessarily economic

yield (Maas, 1990). Interestingly, the FAO LRs accurately

reflected results for marketable fruit yield found in the field

experiments (Fig. 7), where LFs of around 0.6 (I/Ep = 2.5)

resulted in fruit yields that had decreased by less than 10% as

compared to those achieved with low salinity irrigation water.

Certainly, practical use of LFs in general and the results

presented in this study must consider and attempt to quantify

the relationship of biomass production to reproductive

activity and fruit yield in crops.

Sustainability of arid zone irrigated agriculture is depen-

dant upon long-term appropriate treatment of salts (Hillel,

2000; Ben-Gal et al., 2006). The water used for leaching salts

and maintaining conditions for maximum yields must be

disposed of. This water, carrying other agricultural and

natural contaminants along with the excess salts, presents

a serious environmental encumbrance (Smedema and Shiati,

2002). Sustainable cultivation of peppers (or any other crop

that is not categorically tolerant to salinity) must either reduce

salinity prior to irrigation – thus allowing substantially less

leaching – or provide for collection and disposal of the leached

salts and water (Tanji and Kielen, 2002).

5. Conclusions

Irrigation water salinity was found to decrease transpiration

and biomass production in bell peppers. The extent of the

salinity response was dependant upon the level of leaching of

salts from the root zone. Application of saline water to the soil

exceeding the quantity used by the crop for transpiration,

succeeded in improving conditions for water uptake and

growth of the peppers. The addition of such water had higher

relative benefit as the salinity of the water increased.

Lysimeter, field, and modeled experimental results suggested

that potential economic benefits from increased yields exist

for irrigation application rates reaching more than 200% of the

ETp. Leaching fractions were seen to increase as a result of

reductions in transpiration caused by increases in salinity;

thus demonstrating the importance of plant feedback in water

and salt balance. Traditional methods for determining LR were

found to fairly accurately predict the leaching rates required to

maximize yields, despite their over-prediction of transpiration

and under-prediction of yield potential.

The results indicate that irrigation with saline water under

arid conditions may be problematic despite the potential for

economic success. The water used for leaching salts and

maintaining conditions for maximum yield must be disposed of

outside the rootzone. This water, carrying other agricultural

and natural contaminants along with the excess salts, presents

a serious environmental encumbrance. Sustainable cultivation

must provide for collection and disposal of the leached salts and

water or alternatively, reduce the leaching. Reduced leaching is

only possible through cultivation of highly tolerant crops or via

the reduction of water salinity prior to irrigation.

Acknowledgement

This work was made possible through support provided by

The Middle East Regional Cooperation Program, US Agency for

International Development, Grant M24-014.

r e f e r e n c e s

Allen, R.G., Pereira, L.S., Raes, D., Smith M., 1998. CropEvapotranspiration: Guidelines for Computing Crop WaterRequirements. Irrig. Drain. Paper 56. UN-FAO, Rome.

Ayers, R.S., Westcot, D.W., 1985. Water Quality for Agriculture.FAO Irrig. Drain. Paper 29. FAO, Rome.

Ben-Gal, A., Shani, U., 2002. A highly conductive drainageextension to control the lower boundary condition oflysimeters. Plant Soil 239, 9–17.

Ben-Gal, A., Tal, A., Tel-Zur, N., 2006. The sustainability of aridagriculture: trends and challenges. Annals of Arid Zone 45,227–258.

Bernstein, L., 1975. Effects of salinity and sodicity on plantgrowth. Annu. Rev. Phytopathol. 13, 295–312.

Brooks, R.H., Corey, A.T., 1966. Properties of porous mediaaffecting flow. ASCE J. Irrig. Drain. Div. 72 (IR2), 61–88.

Chartzoulakis, K., Klapaki, G., 2000. Response of two greenhousepepper hybrids to NaCl salinity during different growthstages. Sci. Hortic. 86, 247–260.

Corwin, D.L., Rhoades, J.D., Simunek, J., 2007. Leachingrequirement for soil salinity control. Steady-state versustransient models. Agric. Water Manage. 90, 165–180.

de Wit, C.T., 1958. Transpiration and Crop Yield. Verslag VanLandbouck, Onderzoeh, No. 64.6.

Dimitrov, Z., Ovtcharrova, A., 1995. The productivity of peppersand tomatoes in case of insufficient water supply. In:Proceedings of ICID Special Technical Session on the Role ofAdvanced Technologies in Irrigation and Drainage Systems,vol. 1. pp. ft9.1–ft9.5.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7596

Author's personal copy

Dudley, L., Shani, U., 2003. Modeling plant response to droughtand salt stress: reformulation of the root sink term. VadoseZone J. 2, 751–758.

Dudley, L.M., Ben-Gal, A., Shani, U., 2008. Influence of plant, soiland water properties on the leaching fraction. Vadose Zone J.

Feddes, R.A., Raats, P.A.C., 2004. Parameterizing the soil–water–plant root system. In: Feddes, R.A., et al. (Eds.),Unsaturated-zone Modeling: Progress, Challenges andApplications. Wageningen UR Frontis Ser., vol. 6. KluwerAcad. Publ., Dordrecht, Netherlands, pp. 95–141.

Hanks, R.J., 1974. Model for predicting plant yield as influencedby water use. Agron. J. 66, 660–665.

Hillel, D., 2000. Salinity Management for Sustainable Irrigation.Integrating Science, Environment, and Economics. TheInternational Bank for Reconstruction and Development/THE WORLD BANK, Washington, DC.

Hopmans, J.W., Bristow, K.L., 2002. Current capabilities andfuture needs of root water and nutrient uptake modeling.Adv. Agron. 77, 103–183.

Lazarovitch, N., Ben-Gal, A., Shani, U., 2006. An automatedrotating lysimeter system for greenhouseevapotranspiration studies. Vadose Zone J. 5, 801–804.

Letey, J., Dinar, A., 1986. Simulated crop-water productionfunctions for several crops when irrigated with salinewaters. Hilgardia 54, 1–32.

Letey, J., Feng, G.L., 2007. Dynamic versus steady-stateapproaches to evaluate irrigation management of salinewaters. Agric. Water Manage. 91, 1–10.

Letey, J., Dinar, A., Knapp, K.C., 1985. Crop-water productionfunction model for saline irrigation waters. Soil Sci. Soc.Am. J. 49, 1005–1009.

Maas, E.V., 1990. Crop salt tolerance. In: Tanji, K.K. (Ed.),Agricultural Salinity Assessment and Management. ASCEManuals and Reports on Engineering Practices No. 71. Am.Soc. Civil Eng., New York.

Meiri, A., Plaut, Z., 1985. Crop production and managementunder saline conditions. Plant Soil. 89, 253–271.

Meiri, A., Kamburov, J., Shalhevet, J., 1977. Transpiration effectson leaching fractions. Agron. J. 69, 779–782.

Navarro, J.M., Garrido, C., Carvajal, M., Martinez, V., 2002. Yieldand fruit quality of pepper plants under sulphate andchloride salinity. J. Hortic. Sci. Biotechnol. 77, 52–57.

Nimah, N.M., Hanks, R.J., 1973. Model for estimating soil water,plant and atmospheric interrelations. I. Description andsensitivity. Soil Sci. Soc. Am. Proc. 37, 522–527.

Shani, U., Hanks, R.J., Bresler, E., Oliveira, A.S., 1987. Fieldmethod for estimating hydraulic conductivity and matricpotential–water content relations. Soil Sci. Soc. Am. J. 51,298–302.

Shani, U., Ben-Gal, A., Tripler, E., Dudley, L.M., 2007. Plantresponse to the soil environment: an analytical modelintegrating yield, water, soil type and salinity. WaterResour. Res. 43 W08418 10.1029/2006WR005313.

Smedema, L.K., Shiati, K., 2002. Irrigation and salinity: aperspective review of the salinity hazards of irrigationdevelopment in the arid zone. Irrig. Drain. Syst. 16,161–174.

Solomon, K.H., 1985. Water-salinity-production functions.Trans. Am. Soc. Agric. Eng. 28, 1975–1980.

Sonneveld, C., 1988. The salt tolerance of greenhouse crops.Neth. J. Agric. Sci. 36, 63–73.

Tanji, K., Kielen, N., 2002. Agricultural Drainage WaterManagement in Arid and Semi-Arid Areas. FAO Irrigationand Drainage Paper 61. FAO, Rome, Italy.

van Genuchten, M.Th., Hoffman, G.J., 1984. Analysis of cropproduction. In: Shainberg, I., Shalhevet, J. (Eds.), SoilSalinity under Irrigation. Springer, Berlin, pp. 258–271.

Yermiyahu, U., Ben-Gal, A., Keren, R., Reid, R.J., 2008. Combinedeffect of salinity and excess boron on plant growth andyield. Plant and Soil 304, 73–87.

a g r i c u l t u r a l w a t e r m a n a g e m e n t 9 5 ( 2 0 0 8 ) 5 8 7 – 5 9 7 597