Fertilization and Blending Alternatives for Irrigation ...

4 downloads 0 Views 893KB Size Report
in Moshav Ein Tamar, located in the Northern Arava Valley just south of the Dead Sea (30° 57′ N, 35° 23′ E). Irrigation- quantity levels were compared for ...
TECHNICAL REPORTS: VADOSE ZONE PROCESSES AND CHEMICAL TRANSPORT

Fertilization and Blending Alternatives for Irrigation with Desalinated Water Alon Ben-Gal* and Uri Yermiyahu Agricultural Research Organization Shabtai Cohen Central and Northern Arava Research and Development In arid-zone agriculture where available irrigation water is saline, desalination is becoming an attractive method for increasing yields and reducing negative environmental consequences. However, irrigation with desalinated water can be problematic if essential nutrients, including Ca, Mg, and S, removed during reverse osmosis, are not reintroduced. We evaluated two strategies for supplying these nutrients — direct fertilization and blending of desalinated with saline groundwater —experimentally in a greenhouse and in a model for a case study regarding pepper (Capsicum annuum L.) production. Reducing salinity from electrical conductivity (EC) 3.20 to EC 0.40 dS m−1 by reverse-osmosis desalination increased maximum yields by almost 50% while allowing a reduction of applied irrigation water to half of that with the saline water, but the associated cost of fertilizing with Ca, Mg, and S minerals was high (around $0.50 m–3). Blending 30% saline water with 70% desalinated water brought Ca, Mg, and S minerals to satisfactory levels while producing water with salinity of EC = 1.35 dS m−1. Comparison of relative pepper yields and analysis of simulated results showed that irrigation with blended water maintained yields greater than 90% compared to irrigation with fully desalinated water, but only as irrigation rates were increased by more than 50%. The environmental cost of the increase in irrigation-water salinity from EC 0.40 to EC 1.35 dS m−1 in the blended water was shown to be substantial as it involved five times greater loading (into the soil) and leaching (beyond the root zone) of salts and other contaminants.

Copyright © 2009 by the American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Published in J. Environ. Qual. 38:529–536 (2009). doi:10.2134/jeq2008.0199 Received 29 Apr 2008. *Corresponding author ([email protected]). © ASA, CSSA, SSSA 677 S. Segoe Rd., Madison, WI 53711 USA

T

he use of desalinized water as a source of irrigation water for agriculture is on the rise (Yermiyahu et al., 2007b). Since it is estimated that irrigation is responsible for 87% of global water consumption (Shiklomanov, 1997; Döll and Siebert, 2002), current freshwater resources may soon be insufficient to meet the growing demand for food. Technological advances have made desalination an economically feasible solution for high-return agriculture, especially in arid regions where water cost may be excessive due to distance from, or depth to, the water supply. A recent expert report by the United Nations Food and Agriculture Organization (Martinez Beltran and Koo-Oshima, 2006) concluded that while the costs of desalination are still prohibitively high for full use by most irrigated agriculture, its use with high-value cash crops, such as greenhouse vegetables and flowers, has become economically feasible at the present prices. In fact, desalinization of wastewater effluent or brackish groundwater often found in arid regions typically costs half or less than desalination of seawater (Zhou and Tol, 2005). Such desalinated brackish water is being used more and more by farmers for irrigation (Martinez Beltran and Koo-Oshima, 2006). Replacing saline irrigation water with desalinated water is anticipated to increase yields due to reduced salinity stress and to allow drastic decreases in the amount of water currently used for leaching salts out of the root zone. For these reasons, desalination has, in fact, become a real option for planners, decision-makers, and growers in areas like Israel’s Negev Highlands and Arava Valley. Nevertheless, the initial experience with desalinated water has not been completely positive (Yermiyahu et al., 2007a,b). The water leaving reverse-osmosis desalination plants is not low only in unwanted dissolved salts — it also lacks a number of essential elements that farmers in the arid regions typically take for granted, including Ca, Mg, and S. Fertilization with Ca, Mg or S is not usually necessary because of the considerable amounts of the minerals present in soils, irrigation water, incidentally applied with (phosphate) fertilizers (Follet et al., 1981) or, in the case of S, provided by atmospheric pollution (Marschner, 1995). Intensive horticulture, use of lownutrient, highly leached soils, or soil-less media, and use of highanalysis fertilizers have made deficiencies of Ca, Mg, and S more common. Calcium and Mg play roles in plant growth independently and, through interactions with one another, influence product quality and disease resistance (Bangerth, 1973; Marcelis A. Ben-Gal and U. Yermiyahu, Soil, Water and Environmental Sciences, Agric. Research Organization, Gilat Research Center, mobile post Negev 85280, Israel. S. Cohen, Central and Northern Arava Research and Development, Hazeva, Israel. Abbreviations: BW, blended (desalinated and saline) water; DW, desalinated water; EC, electrical conductivity; ECe, saturated paste extract EC; ETp, potential evapotranspiration; GW, saline groundwater.

529

and Ho, 1999; Bar-Tal et al., 2001; Bar-Tal et al., 2003; White, 2003). When desalinated water was supplied to agriculture in Israel’s Western Negev, Ca and Mg deficiency symptoms appeared on crops such as tomatoes, basil, and flowers (Yermiyahu et al., 2006; Yermiyahu et al., 2007a). Sulfur deficiencies have also been reported for crops grown in sandy, low-organicmatter soils (Follet et al., 1981). Reintroduction of beneficial nutrients into desalinated water destined for agriculture can be accomplished in one of three ways: (i) they can be added at the desalination plant as part of the posttreatment processing; (ii) they can be added by farmers as fertilizers; or (iii) they can be added by blending the desalinated water with saline water. The first option, post-desalination treatment at the plant, is expected to be preferred in large seawater-desalination facilities and can be optimized for mixed-use water supplies that will reach municipal as well as agricultural consumers (Yermiyahu et al., 2007b). The other two options, fertilization and blending, are more suitable for smaller-scale desalination projects designed primarily to supply irrigation water. The choice should be both economically and environmentally motivated. The cost of purchasing and injecting Ca, Mg, and S as chemical fertilizers into water at the farm is expected to be high — up to $0.5 m–3 in Israel at current prices. Blending reintroduces undesirable, as well as the desirable, dissolved salts and therefore must be evaluated in terms of cost to potential yields as well as in terms of the economic and environmental costs of leaching. Intensive cultivation of greenhouse-grown sweet peppers can achieve yields more than four times those attained in open-field production, and the high quality of fruit produced and off-season harvesting allow wholesale prices to be 3 to 5 times greater than for field-grown fruits (Jovicich et al., 2004). Pepper cultivation commonly occurs in soil-less media or in imported sand and utilizes frequent irrigation with complete nutrient solution. The Arava Valley is a highly productive agricultural region whose economy is built largely on the export of high-value vegetables, including peppers, irrigated with saline water. The EC of irrigation water in the Arava ranges from 2.0 to 3.5 dS m−1. Production in the region’s arid climate requires application of very large amounts of irrigation water. Potential evapotranspiration (ETp) in greenhouses reaches seasonal levels of 800 to 900 mm. Additional irrigation water, equivalent to 50 to 100% of the actual evapotranspiration (ET), is used to leach salts and maintain productive root-zone conditions (Ben-Gal et al., 2008). Total water consumption for irrigation of net-house peppers grown from September to May in the Arava can therefore reach 1500 to 1600 mm. Modeling constitutes an important tool in studying the influences of various management variables on crop and environmental responses. Although there are a number of either physically based numerical (Feddes and Raats, 2004) or semi-empirical production-function-based (Letey et al., 1985) approaches to evaluating plant response to both amount and salinity of applied water, there are few examples of accessible, easily applied models that consider environmental factors as well as dynamic interactions within the soil-water-plant system. An analytical model introduced by Shani et al. (2007) allows predictions of the crop’s response to conditions of soil water and salinity, while considering the influence of the

530

plant itself (changes in transpiration) on soil water content and salt concentration. The model has been tested and shown to accurately predict transpiration of a number of crops, including bell peppers grown in Israel’s Arava Valley under varied conditions of water and salinity (Shani et al., 2007; Ben-Gal et al., 2008). The objective of this project was to evaluate fertilization and blending management options for the use of desalinated irrigation water for irrigation. We specifically studied the effect of water application quantity on growth and yield of bell pepper (C. annuum L. var. Celica) grown under semi-commercial conditions in the Arava Valley of Israel. Local saline water was compared to desalinated water, with Ca, Mg, and S supplied alternatively as fertilizer or via blending with the saline water in a series of field experiments and using the simulation model by Shani et al. (2007).

Materials and Methods Greenhouse Experiments Experiments were conducted in a plastic-covered, insect-exclusion screenhouse at the “Zohar” agricultural research station in Moshav Ein Tamar, located in the Northern Arava Valley just south of the Dead Sea (30° 57′ N, 35° 23′ E). Irrigationquantity levels were compared for local groundwater (GW) with EC of 3.20 dS m−1, desalinated water with complete fertilization (DW) with EC of 0.40 dS m−1, and their blend (BW) at a rate of 70% desalinated water to 30% groundwater resulting in EC of 1.35 dS m−1. The blending rate was determined according to the minimum levels of Ca, Mg, and S expected to allow non-limited growth (de Kreij et al., 1992; Sonneveld and Straver, 1994; Jovicich et al., 2004) with no further fertilization of these elements (Table 1). Experiments were conducted in three consecutive years. Soil was prepared as beds with 1.6-m spacing. Two drip-irrigation laterals (Ram 1.6 L h−1 drip line with drippers every 20 cm; Netafim, Hatzerim, Israel) were placed 20 cm apart on the soil surface in the center of each bed. Bell pepper plants were grown in rows adjacent to the drip line with plants every 40 cm along each of 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. Relative water quantities were determined as a function of ETp, measured as reference (Class A pan) evaporation outside the greenhouse (Fig. 1). For data analysis and presentation, we assumed, based on data from Tanny et al. (2003) and Ben-Gal et al. (2008), that ETp inside the greenhouse is equal to 50% of that outside. Characteristics of the soil in the greenhouse, obtained by routine procedures (Page et al., 1982), were as follows: loamy sand texture (87% sand, 8% silt, 5% clay); CaCO3, 130 g kg−1; organic matter, 10 g kg−1; cation exchange capacity, 5 meq kg−1; and pH of 1:1 solution, 7.8. Fertilizers (potassium sulfate, calcium nitrate, mono-potassium phosphate, magnesium nitrate, ammonium sulfate, ammonium nitrate) were added to the DW to equalize Ca, Mg, and SO4 to that of the BW. Nitrogen, P, and K were brought to 88, 23, and 176 mg L−1, respectively, in all treatments. Fertilizers increased EC by 0.8 dS m−1 in DW and 0.5 dS m−1 in BW and GW (Table 1). Table 1 provides measured average values for mineral concen-

Journal of Environmental Quality • Volume 38 • March–April 2009

Table 1. Quality of desalinated water (DW), blended water (BW) and groundwater (GW) before and after fertilization. Water

% DW

EC†

pH

dS m−1 Pre-fertilization DW 100 BW 70 GW 0 Post-fertilization DW 100 BW 70 GW 0 † Electrical conductivity.

Ca

Mg

S-SO4

B

Cl

Na

N

P

K

–––––––––––––––––––––––––––––––––––mg L−1–––––––––––––––––––––––––––––––––––

0.40 1.35 3.20

6.2 6.5 6.7

5 65 219

2 30 92

2 48 154

0.30 0.32 0.35

74 256 767

55 143 355

1.20 1.85 3.70

7.2 7.4 7.5

65 65 219

30 30 92

48 48 154

0.30 0.32 0.35

74 256 767

55 143 355

trations in water. Nitrogen, P, and K in all post-fertilized irrigation treatments, as well as Ca, Mg, and S in post-fertilized DW, were target levels. Periodic sampling and analysis during each of the experiments was undertaken and showed that actual concentrations did not vary more than 10% from the average or target levels. Irrigation was applied three times per day at target rates of 50, 100, 150, and 200% of ETp in the greenhouse. The experimental design was complete random blocks in four replicates. Each replicate consisted of 4-m sections of three beds (two rows of plants each). Plant and soil related measurements were taken from the middle 2 m of the central bed in each section. Pepper seedlings were planted on 22 Sept. 2004, 20 Sept. 2005, and 7 Sept. 2006. Salinity treatments were initiated at planting. Irrigation of 4 mm d−1 was applied uniformly for 1 mo, after which the water-quantity treatments were begun. Soil was sampled in November of each season. Each sampling was conducted immediately before the beginning of an irrigation cycle. Samples were taken from the upper 20 cm of soil along the dripper line between plants. Electrical conductivity was measured from a 1:1 gravimetric soil:water extract and saturated paste electrical conductivity (ECe) calculated according to the soil’s saturated water content (0.41 g g−1). Fruit was harvested as it ripened, starting in December, approximately 100 d after planting. After completion of fruit harvest (May of each year), whole plants were removed to measure fresh (fruit) and dry (shoot) biomass. Yield data were normalized relative to the maximum measured in each experimental season to allow comparison between years.

Model Simulations The model (Shani et al., 2007) assumes steady-state conditions and single representative root-zone values for salinity and moisture. Water and salt balance are combined with a calculation of root-zone soil moisture and soil hydraulic conductivity according to the Brooks-Corey (Brooks and Corey, 1966) soil hydraulic model: K (ψ) = min{ K S , K S (ψw ⋅ψ −1 )η } θ(ψ) = min{θS , (θS − θr )(ψw ⋅ψ −1 )β + θr }

[1]

where K is the soil hydraulic conductivity (L t−1), θ is the volumetric soil moisture content (L3 L−3), subscript S denotes saturated, subscript r denotes residual, ψ is soil matric head (L), ψw is air-entry head (L), and η and β are empirical soil characteristic parameters.

1.0 1.4 2.5 88 88 88

0.1 0.1 0.1 23 23 23

4.0 7.3 14.9 176 176 176

Fig. 1. Ten-day averaged “Class A” pan reference evaporation measured outside of the greenhouse for the pepper growing seasons 2004–05, 2005–06, and 2006–07 at the Zohar Agricultural Research Station, Moshav Ein Tamar, Israel.

Transpiration rate (Tw) is the product of the soil’s unsaturated hydraulic conductivity and the gradient of water potential between soil and root (Nimah and Hanks, 1973): Tw = min {TP , b ⋅ K (ψ) ⋅ (ψ root − ψ)} ; ψ > ψ root

[2]

where Tp is potential transpiration, b (L−1) is a coefficient characteristic of the effective distance for flow between roots and soil, and ψroot is the minimum possible water head at the root soil interface allowing water uptake. The variable ψroot is a plant-specific parameter that defines the plant sensitivity to available water. Transpiration decrease as a function of salinity is considered by a plant-specific reduction term characterized by a logistic curve with an initial plateau and subsequent decreasing section (van Genuchten and Hoffman, 1984). 1 f EC = p ⎛ ECe ⎞ 1+ ⎜ ⎟⎟ ⎜ [3] ⎝ ECe ⎠ 50

where fEC is a reduction function due to salinity, ECe50 represents the EC of the soil saturated paste where relative yield is 0.5, and p is a plant parameter which, for many situations, is equal to 3 (van Genuchten and Gupta, 1993). The model assumes a proportional relationship between the ratio of yield to potential

Ben-Gal et al.: Fertilization and Blending Alternatives for Irrigation with Desalinated Water

531

Table 2. Soil and plant parameters for lysimeters and model input parameters.† Soil KS (mm d−1) δ (unitless) β (unitless) θs (m3 m−3) θr (m3 m−3) ψw (mm)

3600 4.91 0.55 0.41 0.06 −200

Arava Loamy Sand

Plant ψroot (mm) ECe50 (dS m−1) Tp (mm d−1) Yr0

Capsicum annum var. Celica −6000 2.5 5 0.08

† From Shani et al. (2007). Ks, saturated hydraulic conductivity; δ and β, empirical soil characteristic parameters for the Brooks and Corey (1966) hydraulic model; ψw, air entry value; θs, soil water content at saturation; θr, residual soil water content; ψroot, minimum possible water head at the root soil interface; ECe50, plant characteristic parameter for salinity response function (EC of the soil saturated paste causing 50% yield decrease); Tp, potential transpiration; Yr0, relative yield resulting from initial soil water.

yield and the ratio of transpiration to potential transpiration following de Wit (1958) and Hanks (1974), thus allowing a prediction of biomass production (yield). The model was used to predict plant performance and to evaluate salt and water balance for different irrigation regimes and quality of each irrigation water. Input variables included the quantity and salinity of the applied water, plant sensitivity to salinity and water stress, ETp, and soil hydraulic parameters (Table 2). The salinity reduction curve for bell pepper, defined by the ECe causing a 50% decrease in yield, was adapted from previously published experimental data (Ben-Gal et al., 2008). The model was used to predict biomass production (Y), drainage (D), leaching fraction (LF), and seasonal salt load for irrigationwater salinities of the DW, BW, and GW and for irrigation levels relative to potential transpiration (I ETp–1) from 0 to 2.0.

Data Analysis Statistical analysis was performed with the JMP software package (SAS, Cary, NC) for analysis of variance among fresh fruit yields, fruit size, and dry biomass production within each experimental year. The quantitative effects of irrigation-water salinity and quantity on soil salinity and on relative shoot and fruit yields from all the experimental data were determined using linear and nonlinear regression curve fitting and analysis, employing the SigmaPlot 9.01 package (Systat Software, Inc., Point Richmond, CA). Default significance levels were set at α = 0.05. Measured data were 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.

Results and Discussion Greenhouse Experiments Irrigation with higher salinity water led to higher soil solution salinity; soil salinity, in turn, was reduced by increased

532

Fig. 2. Saturated paste electrical conductivity of soil sampled from 0 to 20 cm depth, under drip lines, between pepper plants, during November of each season for fully fertilized desalinated water (DW), desalinated water blended with saline groundwater (BW), and saline groundwater (GW). Symbols are experimental results and lines are best-fit linear regression.

application rate (Fig. 2). Throughout the experiment, despite the fact that not all the treatments were directly tested together in any given year, ECe was consistent as a function of treatment and was consistently lowest at any given irrigation level for irrigation with DW, intermediate for irrigation with BW, and highest for irrigation with GW. Salinity of soil irrigated with DW was reduced slightly by increased irrigation (ECe = 1.21–0.16 I ETp−1), increasing irrigation with BW reduced ECe more substantially (ECe = 1.76–0.34 I ETp−1), and increasing GW irrigation amounts had an even greater effect on soil salinity (ECe = 2.95–0.63 I ETp−1). Fresh fruit yield varied from year to year and ranged from 99 to 135 Mg ha−1 in the 2004–05 season, 99 to 119 Mg ha−1 in 2005–06, and 91 to 152 Mg ha−1 in 2006–07, when the widest range of irrigation application rates was tested (Table 3). Increasing irrigation and decreasing salinity were found to increase both shoot biomass and fruit production of peppers, as reflected by both the absolute (Table 3) and normalized (Fig. 3) data. In 2004–05, maximum shoot biomass production for GW irrigation was significantly lower (by 30%) than that with DW irrigation. Maximum fruit yield of GW-irrigated peppers was not significantly less than that for DW-irrigated ones, but fruit yield of the lowest water application rate for GW was significantly lower than for all the other treatments. In 2005–06, GWirrigated peppers again had significantly low shoot production, reaching 70% of that of the DW-irrigated peppers. In 2006–07, fruit yields were higher than in previous years while shoot biomass values were consistent with those measured in the earlier experiments. The GW irrigation caused significant reductions of 22% in maximum shoot production and 12% in maximum fruit yield compared to BW irrigation, and increasing irrigation levels led to significantly greater shoot biomass production and fruit yields for both BW and GW in 2006–07 (Fig. 3). Benefits are expected from adding water to regularly drip-irrigated crops, as irrigation rates rise from deficit levels to quanti-

Journal of Environmental Quality • Volume 38 • March–April 2009

Table 3. Irrigation treatments and average fresh fruit and dry shoot yields of bell peppers for fully fertilized desalinated water (DW), desalinated water blended with saline groundwater (BW), and saline groundwater (GW). Relative irrigation (I ETp−1) is seasonal irrigation relative to seasonal potential evapotranspiration. Different letters signify significant differences in average values of each parameter within each year as determined by analysis of variance. Year

Water type

2004–05 2004–05 2004–05 2004–05 2004–05 2004–05

DW DW DW GW GW GW

Salinity dS m−1 0.40 0.40 0.40 3.20 3.20 3.20

Irrigation mm 885 1245 1543 800 1284 1621

2005–06 2005–06 2005–06 2005–06 2005–06

DW DW DW DW GW

0.40 0.40 0.40 0.40 3.20

438 718 949 1435 1300

2006–07 2006–07 2006–07 2006–07 2006–07 2006–07 2006–07 2006–07

BW BW BW BW GW GW GW GW

1.35 1.35 1.35 1.35 3.20 3.20 3.20 3.20

350 633 991 1441 260 638 935 1300

Relative irrigation I ETp−1 1.15 1.62 2.00 1.04 1.66 2.10

ties that fulfill the needs of potential transpiration (Allen et al., 1998). Benefits of further water application, beyond transpiration demands, are due to leaching of salts and minimizing osmotic interference of water uptake (Ayers and Westcot, 1985) and, therefore, should be more substantial as the salinity of the irrigation water increases. The benefits of applying additional amounts of water for leaching when irrigating with saline GW were obvious in the experiments (Table 3, Fig. 3) as production increased with each increase in water application rate. Shoot biomass increased with increased GW irrigation, reaching maximum levels at around 150% of ETp. Fruit yield continued to increase linearly with even greater GW-irrigation rates, up to 200% of the potential transpiration needs of the plants. Irrigation with DW caused increased shoot biomass up to 150% ETp and fruit yields that increased with irrigation from 50 to 100% ETp and decreased with further additional water application. Irrigation with BW gave intermediate results, with shoot yield increasing linearly to a maximum at the highest irrigation rates and fruit yield increasing substantially as irrigation increased from 50 to 150% ETp, then only slightly as irrigation was further increased to 200% ETp. In each of the experimental seasons, the lower salinity water, either DW or BW, led to higher yields at lower water application rates relative to GW (Table 3, Fig. 3). Fruit yield was high for DW throughout the entire range of irrigation rates, while yields for BW and GW increased with increased water application. Fruit size was also affected by irrigation-water type and amount. The higher salinity GW produced lower weight fruit than did DW and BW irrigation at low irrigation levels (Table 3). Weight of individual fruit increased significantly with increased irrigation water for GW, and at the highest irrigation treatments was not different from that of DW or BW. Shoot biomass increased with increasing irrigation-water application, with greater biomass at lower irrigation rates for DW, then for BW, and finally for GW. Relative fruit yields (Fig. 3A) were less affected by salinity

Total fresh fruit Mg ha−1 135 a 131 a 126 a 99 b 119 ab 125 a

Fruit weight g fruit−1 165 a 163 a 163 a 145 b 154 ab 162 a

Shoot dry weight kg plant−1 0.24 b 0.32 a 0.30 a 0.18 c 0.21 b 0.22 b

0.61 1.00 1.32 2.00 1.80

106 b 119 a 102 b 99 b 102 b

167 c 176 b 181 a 182 a 183 a

0.17 b 0.26 ab 0.29 a 0.32 a 0.22 b

0.48 0.88 1.38 2.00 0.36 0.89 1.30 1.80

112 c 135 b 145 a 152 a 91 d 100 cd 115 c 134 b

156 a 155 a 156 a 159 a 139 b 141 b 150 a 155 a

0.14 c 0.19 bc 0.23 a 0.27 a 0.13 c 0.19 bc 0.21 b 0.21 b

Fig. 3. Relative fresh fruit yield (A) and dry shoot weight (B) as a function of relative applied irrigation water for fully fertilized desalinated water (DW), desalinated water blended with saline groundwater (BW), and groundwater (GW). Symbols are experimental results and lines are best-fit regression.

of the GW at high irrigation rates than the shoot yields (Fig. 3B). The relative fruit yields were found to be higher than the relative

Ben-Gal et al.: Fertilization and Blending Alternatives for Irrigation with Desalinated Water

533

Fig. 4. Modeled prediction (lines) of (A) relative biomass production (yield as a function of maximum or potential yield), (B) relative drainage (drainage as a function of potential evapotranspiration), and (C) leaching fraction (drainage as a function of irrigation), each as a function of applied irrigation water relative to potential evapotranspiration; and (D) relative yield as a function of salt (NaCl) load to the environment. Fully fertilized desalinated water (DW), 70% desalinated water blended with 30% saline groundwater (BW), and groundwater (GW) with electrical conductivity of 3.2 dS m−1. Symbols in (A) are measured experimental results. Simulation was for bell pepper in the Arava Valley with input variables from Table 2.

shoot yields, suggesting that fruit production was reduced less by salinity and was more positively influenced by increasing irrigation than shoot biomass production.

Model Simulations and Environmental Repercussions Simulated results for the greenhouse experiments (shown as lines in Fig. 4) were compared to the experimental data (shown as symbols in Fig. 4A). Simulations of yield (total biomass) response to increasing irrigation rates (I ETp−1), taking into account the salinity of the irrigation water, predicted a linear response to near maximum yields as water increased from 0 to just over 100% ETp when salinity was equal to that of DW (Fig. 4A). As water salinity increased, the response was more subdued, but maximum yields continued to increase with higher water application rates. The maximum yields predicted for each increasing salinity level were less than for the DW; maximum yields for BW and GW were 95% and just over 70%, respectively. The model predicted the measured yield of peppers reasonably well. Regression of predicted versus measured yield data for the current experimental set, from all of Fig. 4A, resulted in a significantly linear relationship (r2 = 0.852, slope = 1.12, intercept = –0.078 [not different from 1 and 0 at 95% confidence]). The legitimacy of conclusions regarding blending and fertilizing may be compromised since these strategies were not compared experimentally against one another and have only been regarded in comparison to GW in experiments from separate years. Further research is required to compare all treatments within the same year to elicit cause-and-effect relationships. The agreement of measured

534

soil solution salinity for the GW and DW water quality treatments over the experimental years (Fig. 2) and of climatic conditions (Fig. 1) for each of the three seasons aided the justification of the use of normalized data from the separate experimental sets for analysis and comparison. The simulation model additionally offered a method for theoretical comparison of the alternatives without the experimental limitations. The validity of this model for the prediction of pepper yields under conditions of varied irrigation-water quantity and salinity had also been established previously (Shani et al., 2007; Ben-Gal et al., 2008), and the current study used model set-up and calibration from those studies. Increasing irrigation water increases yields but also increases movement of water and dissolved salts into and beyond the root zone. Figure 4B illustrates simulated results of normalized drainage quantity (drainage as a function of ETp) generated in response to increasing irrigation rate (I ETp−1). The simulated results showed that increased drainage quantities correspond to increased irrigation quantities and to increased irrigation-water salinity. Drainage quantity was low for DW until the irrigation rate increased to greater than ETp. Drainage-water increases were found at lower irrigation rates when irrigation-water salinity was higher. The leaching fraction is the ratio of drainage to irrigation and signifies the relative volume of applied water that carries salts out of the root zone (Ayers and Westcot, 1985). Target leaching fractions, designed to keep soil salinity lower than levels desired for maximized or acceptable crop growth, are commonly used in the management of saline irrigation water. Predicted leaching fractions as a function of irrigation applica-

Journal of Environmental Quality • Volume 38 • March–April 2009

tion rate (Fig. 4C) show that drainage water calculated by the model for I = 0.80 ETp was 10% of I for DW, 20% of I for BW, and 40% of I for GW (Fig. 3C). For I = 1.2 ETp, drainage water increased to nearly 20, 30, and 50% of I for DW, BW, and GW, respectively. Leaching fractions predicted by the steady-state model had minimum values, indicating that reducing irrigation rates cannot completely eliminate leaching. The issue of leaching fraction, its dependence on soil and plant parameters, and its ramifications for deficit-irrigation strategies are discussed in depth by Dudley et al. (2008). The amount of salt added to the environment is a function of the amount and quality of the irrigation water. Calculation of seasonal salt loading resulting from the irrigation of pepper with DW, BW, and GW is shown in Fig. 4D, where NaCl is used as a simplistic example representing the salinities in the current case study. Irrigation with BW (EC = 1.35 dS m−1), which limits pepper yields to a maximum 95% of the yield that can be achieved with DW (EC = 0.40 dS m−1), results in almost five times greater seasonal salt loading in the environment (10 ton ha−1 compared to 2.1 ton ha−1 as NaCl). Irrigation with GW (EC = 3.20 dS m−1), which limits the yield to less than 65% of what can be achieved with DW, adds more than 11 times the salt to the environment. The environmental cost of irrigation with saline water will often be compounded, of course, as other substances, including excess fertilizers, agrochemicals, and naturally occurring contaminants that would not necessarily otherwise be mobilized, are carried along with the leached salts beyond the root zone in agricultural fields. This cost could be even further aggravated when alternative problematic water sources containing a range of contaminants in addition to salts, for example wastewater effluent or agricultural drainage water, are considered for desalination.

Conclusions Irrigation-water salinity needs to be evaluated in terms of reduced crop yields, the direct cost of additional water needed to leach salts, and the indirect environmental costs of leaching. For the relatively salt-sensitive pepper crop grown in an arid region, reducing salinity from EC 3.20 dS m−1 to EC 0.40 dS m−1 by reverse-osmosis desalination was found to increase maximum yields by almost 50% while allowing a reduction of the applied irrigation water to half of that with the saline water. However, the cost of fertilizing to re-supply Ca, Mg, and S minerals removed during the desalination process (in our case some $3,500 ha−1) must also be considered. Pepper yields from irrigation with blended water containing 70% desalinated and 30% saline water could be maintained at greater than 90% of those with fully desalinated water, but only if irrigation rates were increased by at least 50%. The blending strategy for providing mineral nutrients can therefore consume up to 40% more desalinated water than the fertilization strategy. The environmental cost of the increase in irrigation-water salinity from an EC of 0.40 dS m−1 (desalinated water) to an EC of 1.35 dS m−1 (blended water) is substantial, as it involves the loading (into the soil) and leaching (beyond the root zone) of five times more salt, and potentially escalates the transport of other contaminants out of the root zone as well.

Acknowledgments This work was made possible by support from: The Middle East Regional Cooperation Program, U.S. Agency for International Development, Grant M24-014; The Chief Scientist of Israel’s Ministry of Agriculture and Rural Development, Grant 304-0393; and Arava Research and Development. We would like to thank: Eugene Presnov and Inna Feingold of the Gilat Research Center and Ami Maduel, Dorit Heshmonaee, Rivke Offenbach, Yoram Zvieli, and Rami Golan of Arava R&D for their technical support and advice.

References Allen, R.G., L.S. Pereira, D. Raes, and M. Smith. 1998. Crop evapotranspiration: Guidelines for computing crop water requirements. Irrig. Drain. Paper 56. UN-FAO, Rome. Ayers, R.S., and D.W. Westcot. 1985. Water quality for agriculture. FAO Irrig. Drain. Paper 29. FAO, Rome. Bangerth, F. 1973. Calcium-related physiological disorders of plants. Ann. Rev. Physiopathol. 17:97–122. Bar-Tal, A., M. Keinan, B. Aloni, L. Karni, Y. Oserovitz, S. Gantz, A. Hazan, M. Itach, N. Tratakovski, A. Avidan, and I. Posalski. 2001. Relationships between blossom- end rot and water availability and Ca fertilization in bell pepper fruit production. Acta Hortic. 554:97–104. Bar-Tal, A., M. Keinan, S. Suriano, B. Aloni, L. Karni, S. Cohen, R. Offenbach, and A. Maduel. 2003. Managing of circulated nutrient solutions with saline water for pepper cultivation. Acta Hortic. 609:349–354. Ben-Gal, A., E. Ityel, L.M. Dudley, S. Cohen, U. Yermiyahu, E. Presnov, L. Zigmond, and U. Shani. 2008. Effect of irrigation water salinity on transpiration and on leaching requirements: A case study for bell peppers. Agric. Water Manage. 95:587–597. Brooks, R.H., and A.T. Corey. 1966. Properties of porous media affecting flow. ASCE J. Irrig. Drain., Div. 72(IR2), 61–88. de Kreij, C., C. Sonneveld, M.G. Warmenhoven, and N. Straver. 1992. Guide values for nutrient element contents of vegetables and flowers under glass (Proefstation voor Tuinbouw onder Glas te Naaldwijk). In Proefstation voor de Bloemisterij te Aalsmeer, Voedingsoplossingen Glastuinbouw, No. 15. de Wit, C.T. 1958. Transpiration and crop yield. Verslag Van Landbouck, Onderzoeh. No. 64.6. Döll, P., and S. Siebert. 2002. Global modeling of irrigation water requirements. Water Resour. Res., doi: 10.1029/2001WR000355. Dudley, L.M., A. Ben-Gal, and U. Shani. 2008. Influence of plant, soil and water properties on the leaching fraction. Vadose Zone J. 7:420–425. Feddes, R.A., and P.A.C. Raats. 2004. Parameterizing the soil-water-plant root system. p. 95–141. In R.A. Feddes et al. (ed.) Unsaturated-zone modeling: Progress, challenges, and applications. Wageningen UR Frontis Ser., Vol. 6. Kluwer Acad. Publ., Dordrecht, the Netherlands. Follet, H., L.S. Murphey, and R.L. Donahue. 1981. Fertilizers and soil amendments. Prentice Hall, Englewood Cliffs, NJ. Hanks, R.J. 1974. Model for predicting plant yield as influenced by water use. Agron. J. 66:660–665. Jovicich, E., D.J. Cantliffe, S.A. Sargent, and L.S. Osborne. 2004. Production of greenhouse-grown peppers in Florida. Florida Cooperative Extension Service publication HS979. Available at http://www.hos.ufl.edu/ ProtectedAg/EDIS/HS22800.pdf (verified 17 Oct. 2008). Letey, J., A. Dinar, and K.C. Knapp. 1985. Crop-water production function model for saline irrigation waters. Soil Sci. Soc. Am. J. 49:1005–1009. Marcelis, I., and L. Ho. 1999. Blossom-end rot in relation to growth rate and calcium content in fruits of sweet pepper (Capsicum annuum L.). J. Exp. Bot. 50:356–362. Marschner, H. 1995. Mineral nutrition in plants, 2nd ed. Academic Press, San Diego, CA. Martinez Beltran, J., and S. Koo-Oshima. 2006. FAO Expert Consultation on Water Desalination for Agricultural Applications, Rome (Italy). 26–27 Apr 2004/FAO, Rome (Italy). FAO Land and Water Discussion Paper, no. 5; Land and Water Development Div. 48 p. ISSN 1729-0554. Nimah, N.M., and R.J. Hanks. 1973. Model for estimating soil water, plant and atmospheric interrelations: I. Description and sensitivity. Soil Sci.

Ben-Gal et al.: Fertilization and Blending Alternatives for Irrigation with Desalinated Water

535

Soc. Am. Proc. 37:522–527. Page, A.L., R.H. Miller, and D.R. Keeney (ed.). 1982. p. 149–165, 181–197, 199–224, 539–579. Methods of soil analysis Part 2, Chemical and microbiological properties. ASA, SSSA, Madison, WI. Shani, U., A. Ben-Gal, E. Tripler, and L.M. Dudley. 2007. Plant response to the soil environment: An analytical model integrating yield, water, soil type and salinity. Water Resour. Res. 43:W08418, doi: 10.1029/2006WR005313. Shiklomanov, I.A. 1997. Assessment of water resources and water availability in the world, comprehensive assessment of the freshwater resources of the world. Stockholm Environment Institute, Stockholm. Sonneveld, C., and N. Straver. 1994. Nutrient solution for vegetables and flowers grown in water or substrates. contents under glass. Tenth ed. Voedingsoplossingen glastuinboun. Tanny, J., S. Cohen, and M. Teitel. 2003. Screenhouse microclimate and ventilation: An experimental study. Biosyst. Eng. 80:331–341. van Genuchten, M.Th., and S.K. Gupta. 1993. A reassessment of the crop

536

response function. J. Indian Soc. Soil Sci. 41:730–737. van Genuchten, M.Th., and G.J. Hoffman. 1984. Analysis of crop production. p. 258–271. In I. Shainberg and J. Shalhevet (ed.) Soil salinity under irrigation. Springer, Berlin. White, P.J. 2003. Calcium in plants. Ann. Bot. (Lond.) 92:487–511. Yermiyahu, U., A. Ben Gal, S. Cohen, D. Shemer, D.R. Golan, and A. Bar Tal. 2007a. Irrigation of crops with desalinated water. Report submitted to Chief Scientist, Israel Ministry of Agriculture and Rural Development. Project # 301-00527-05, 15 pp. (in Hebrew). Yermiyahu, U., A. Tal, A. Ben-Gal, A. Bar-Tal, J. Tarchisky, and O. Lahav. 2007b. Rethinking desalinated water quality and agriculture. Science 318:920–921. Yermiyahu, U., I. Shamai, R. Peleg, N. Dudai, and D. Shtienberg. 2006. Reduction of Botrytis cinerea sporulation in sweet basil by altering the concentrations of nitrogen and calcium in the irrigation solution. Plant Pathol. 55:544–552. Zhou, Y., and R.S.J. Tol. 2005. Evaluating the costs of desalination and water transport. Water Resour. Res. 41:W03003, doi:10.1029/2004WR003749.

Journal of Environmental Quality • Volume 38 • March–April 2009