SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 3 August 2006
Published in Soil Sci Soc Am J 70:1556-1568 (2006)
DOI: 10.2136/sssaj2005.0365
© 2006 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Assouline, S.
Right arrow Articles by Silber, A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Assouline, S.
Right arrow Articles by Silber, A.
Agricola
Right arrow Articles by Assouline, S.
Right arrow Articles by Silber, A.

Soil & Water Management & Conservation

Soil-Plant System Response to Pulsed Drip Irrigation and Salinity

Bell Pepper Case Study

S. Assouline*, M. Möller, S. Cohen, M. Ben-Hur, A. Grava, K. Narkis and A. Silber

Institute of Soil, Water and Environmental Sciences, Agricultural Research Organization, The Volcani Center, POB 6, Bet Dagan 50250, Israel

* Corresponding author (vwshmuel{at}agri.gov.il)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
High-frequency drip irrigation supplies water and nutrients at a rate that is close to plant uptake, thus enhancing growth and production. In light of water scarcity in arid regions, marginal water is increasingly considered as a resource for agricultural production. The objective of this study was to investigate the combined effects of pulsed irrigation and water salinity on the response of the soil–plant system. As a test crop, bell pepper (Capsicum annuum L.) was cultivated in a screenhouse and drip irrigated daily (D) and at high frequency (P) with saline (S) and fresh (F) water. Simultaneous monitoring of meteorological, physiological, soil physical, plant and soil chemical, and yield data was performed during the experiment. Most physiological parameters were negatively affected by high water salinity. No consistent effect of the irrigation frequency was found on the overall season, although pulsed irrigation led to higher plant weight and leaf area at the early stages of plant growth. The distinct patterns of soil water content for the two irrigation frequencies are presented. Salinity in the root zone was higher under pulsed irrigation, an observation that is supported by measured leaf chloride content and tensiometer readings indicating that the once daily application may have more efficiently removed salts from the top soil. Yield, fruit weight, and irrigation water use efficiency (IWUE) were highest under once daily irrigation with fresh water. High-frequency irrigation led to higher Mn concentrations in leaves and fruits and increased concentrations of Cl, N, and P in leaves, confirming earlier conclusions on improved P mobilization and uptake under pulsed irrigation.

Abbreviations: D, drip irrigated daily • ETo, reference evapotranspiration • F, fresh water • IWUE, irrigation water use efficiency • LAI, leaf area index • LF, leaching fraction • P, high frequency • S, saline • SF, sap flow


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
IN ARID AND SEMIARID REGIONS, modern agriculture has to rely on irrigation. The scarcity of fresh-water and the socio–economic pressures for increased agricultural production induce the necessity to (i) improve the water use efficiency of crops (Stanhill, 1992) and (ii) use marginal water, generally of higher salinity, for irrigation (Avnimelech et al., 1992). Drip irrigation can lead to high crop water use efficiency by wetting only a limited part of the root zone (Bresler et al., 1982). Soil water potential in the vicinity of active roots generally controls the rate of water and nutrient uptake by plants. In drip irrigation, frequency and emitter discharge determine the variation in soil water potential, and consequently, root distribution and plant water uptake patterns (Coelho and Or, 1996; 1999). Drip irrigation systems generally consist of emitters with discharge rates ranging from 1.6 to 8.0 L h–1. In semiarid climates, crop water use during the summer can reach 6 to 8 mm d–1 (Shalhevet et al., 1981), with water supplied two or three times a week. Given the spacings used for both emitters and laterals, the duration of water application is thus much shorter than the time over which plant water uptake takes place. In the case of daily drip irrigation with a conventional emitter discharge of 2 L h–1, irrigation takes approximately 1/10 of the time during which photosynthesis and transpiration occur, and water is applied at a rate tenfold the uptake rate of the plant. If the root zone cannot store large amounts of water, some water will be lost by drainage and irrigation efficiency will not be optimal. Furthermore, in the interval between two subsequent irrigation events, the water content, and consequently, the soil unsaturated hydraulic conductivity will decrease (Da Silva, 1991), resulting in a decrease in nutrient availability (Jungk, 1996) and an increase of solutes concentrations in the vicinity of the roots, both of which negatively affect plant growth. Therefore, lowering the application rate to match plant water uptake rate as closely as possible may improve irrigation efficiency (Batchelor et al., 1996).

Recently, microdrip irrigation systems have been developed that provide emitter discharges of <0.5 L h–1. These systems are mostly based on low-pressure gravity flow, and therefore have significantly reduced power requirements, making them particularly attractive in developing countries. Some aspects of microdrip technology still remain problematic, especially with regard to application uniformity (Miller, 1990) and discharge steadiness. However, soil moisture regimes similar to those resulting from continual low water application rates can be achieved by means of pulsed drip irrigation at higher rates (Zur, 1976). Infiltration experiments on a sandy loam showed that the water content distribution and the rate of wetting front advance under a pulsed water application (at a nominal rate q and a time-averaged rate of qav) was similar to water applied in a continuous manner at qav (Zur, 1976), and that temporal fluctuations in flux and in soil water content were exponentially dampened with depth for periodic pulses applied at the soil surface (Zur and Savaldi, 1977). Consequently, pulsed irrigation using conventional drip emitters could be one way of creating the water regime observed with continual low application rates while by-passing technical problems associated to low-discharge microdrip emitters.

It has been shown that microdrip irrigation under field conditions may improve yield of sweet corn (Zea mays L.) and affect the water content distribution within the root zone, especially through an increased drying of the 0.60- to 0.90-m soil layer compared with conventional drip irrigation (Assouline et al., 2002; Assouline, 2002). Son and Oh (2003) reported that under potted plant conditions, continuous and high frequency irrigation (four times per day) enhanced a more stable water content, and increased growth, plant height, root activity, and number of flowers and buds compared with the case where plants were irrigated once a day. Also, the substrate salinity was highest for the continuous irrigation but this did not affect plant growth. Katsoulas et al. (2006) showed that high irrigation frequency increased biomass production of a soilless rose crop but did not affect its quality, namely, the length of rose flowering shoots. However, Shock et al. (2005) found that onion grown with one daily irrigation at the 0.5 L h–1 drip emitter discharge produced higher total and marketable yields compared with higher irrigation frequencies (up to eight times a day) and lower emitter discharge (0.25 L h–1) treatments. They also found that there was no significant difference in average water potential between treatments.

Continuous application of water and nutrients at low rate or alternatively, frequent applications at higher rates, ensures that the root surface and its vicinity are well supplied with fresh nutrient solution during the fertigation events and the subsequent redistributions. Previous studies demonstrated that increased fertigation frequency significantly increased plant yield, especially at low nutrient concentrations (Silber et al., 2003; Xu et al., 2004), and that the yield improvement was primarily related to increased P uptake (Silber et al., 2005). It was suggested that high fertigation frequency improved the uptake of nutrients through two main mechanisms: (i) continuous replenishment of nutrients in the depletion zone near the root–medium interface; and (ii) enhanced transport of dissolved nutrients by mass flow, because of the higher time-averaged water content in the medium during daytime. Therefore, high fertigation frequency might compensate for nutrient deficiency (Silber et al., 2005).

When saline water is used for irrigation, soil salinity is affected by two main variables, namely, salt concentration in the soil solution and salt load in the root zone. For a given water salinity, the salt concentration is dependent on the soil water content, while salt load is a function of the amount of water applied. Therefore, the salt regime in the root zone will be related to the water application rate or the irrigation frequency as these induce different spatial distribution of water content in the soil for similar total amount of applied water. The studies of Biggar and Nielsen (1967) and Mokady and Bresler (1968) showed that leaching with irrigation water is more efficient when the soil is maintained unsaturated and the flow rate is relatively low. This was later validated both by experimental studies and by model simulations (Bresler et al., 1982). This is the reason why salt leaching can be more efficient using drip irrigation methods. The relationships between soil properties, water application rates, and the resulting water and salt distributions around the emitter in conventional drip irrigation systems are well known (Bresler et al., 1982). This however is not the case when it comes to continuous irrigation at low application rates or high frequency pulsed-irrigation at higher rates.

Bell pepper grown under protected cultivation (greenhouse, plastic-house, and screenhouse) is a valuable crop worldwide and high-quality yield is an essential prerequisite for its economic success. This crop is classified as moderately sensitive to saline water irrigation (Rhoades et al., 1992). According to Bresler et al. (1982), the salinity threshold is 1.5 dS m–1 and the productivity decreases at a rate of 14.1% (dS–1 m), the corresponding values according to Rhoades et al. (1992) being 1.7 dS m–1 and 12% (dS–1 m), respectively. The salinity of available marginal waters is often higher than these threshold values. Therefore, it is important to study the response of bell pepper to the combined effect of pulsed irrigation with water salinity significantly exceeding the threshold value of most sensitive and moderately sensitive crops.

The objective of the present study was to investigate the combined effects of pulsed irrigation and water salinity on the response of the soil–plant system, using the bell pepper as the case study crop. A large number of parameters, providing complementary information on water and salt regimes in the root zone, plant water and nutrient uptakes, yield, and IWUE, were monitored and analyzed for that purpose.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Site, Crop Management, and Treatments
The field experiment was conducted at the Besor Experimental Station in the Northern Negev, Israel (31°16' N, 34°24' E, 75 m above mean sea level), from June to November 2004. In accordance with local agronomic practice, a flat-roof screenhouse (24 x 84 x 3.2 m) made of a 30% black shading screen was used. The summer climate at the site is arid and hot with maximum temperatures of up to 36°C and daily reference evapotranspiration of more than 6.0 mm d–1.

Bell pepper (Capsicum annuum cv. Selika) was grown on a seedbed of sandy loam soil, with two crop rows (0.40 cm spacing) per bed. Planting and final harvest were on 18 June and 15 Nov. 2004, respectively and plant density was 3.12 m–2. Plants were grown using the "Dutch" system, in which each individual plant is vertically supported by a cord hanging from the screenhouse roof structure. All treatments were irrigated daily using a pressure compensated non-leakage drip system: 1.6 L h–1, PC-CNL, integral, 20-cm emitter spacing (Netafim, Ltd; Israel), one dripper line per row. Soluble fertilizer was applied with the irrigation water proportional to the water amount at concentrations of 70 mg L–1 N-NO3, 30 mg L–1 N-NH4, 30 mg L–1 P and 150 mg L–1 K. Two irrigation frequencies were tested: one application per day (common irrigation practice, designated D) and 10 pulses per day at hourly intervals (P). Two water qualities were used for irrigation: non-saline water F (electrical conductivity [EC] = 1.8 dS m–1, SAR = 2.5) and saline water S (EC = 4.2 dS m–1, SAR = 30) artificially prepared using NaCl and CaCl2. Irrigation water amounts were directly related to reference evapotranspiration (ETo) evaluated by means of meteorological data (Allen et al., 1998). A water amount corresponding to 100% of calculated gross irrigation water requirements, Igross, was given for these two water qualities. In addition, a water amount corresponding to 125% of Igross was given in the case of the high salinity water to check the impact of increased leaching. The randomized block experiment consisted of 30 plots (each 12 x 4.8 m)—six treatments with five replicates each (see Table 1). Each treatment plot contained three beds, the middle bed for measurements and the two outer beds as boundaries and bed spacing was 1.6 m from center to center. The differential irrigation treatments commenced 30 d after transplanting, to allow for uniform crop establishment.


View this table:
[in this window]
[in a new window]
  		Table 1. Emitter discharge, irrigation frequency, net irrigation water amount, Inet, and electric conductivity (EC) of the irrigation water for the six irrigation treatments.

 
Climate and Soil Measurements
Two automatic weather stations were installed, both measuring the data required for calculating daily ETo (Allen et al., 1998). One station was located in the center of the screenhouse and the other 5 m from the western sidewall outside of the screenhouse. The recorded parameters (at 2 m above ground) included incoming short wave radiation RG (CM-5, Kipp and Zonen, Delft, NL), outside wind speed at 4.3 m (wind monitor, Model 05103, R.M. Young Co., Traverse, MI, USA) inside wind speed at 2.0 m (cup anemometer, Met One Inc, Grants Pass, OR, USA) and dry and wet bulb temperatures using Cu-Co thermocouples in aspirated boxes (Volcani Center, Israel).

Soil water tension ({psi}) was recorded continuously at 10 digital tensiometer stations (Mottes, Israel). The probes were installed half way between the rows at a horizontal distance of 20 cm from the dripper line at 0.2-, 0.4-, and 0.6-m depths. Measured soil water tension, {psi} (kPa), was converted to volumetric soil water content, {theta} (m3 m–3), using the general form of the soil water release curve suggested by Assouline et al. (1998):

Formula 1[1]
where {theta}s is the saturated soil water content, {xi} and µ are fitting parameters, and {psi}L is the capillary head corresponding to a very low water content, {theta}L, which represents the limit of the domain of interest of the water retention curve. In a preliminary evaluation at the experimental site in 2003, the parameters in Eq. [1] were determined using gravimetric soil samples and tensiometer readings covering a range from 0 to 70 kPa: {theta}s = 0.42 m3 m–3, {theta}L = 0.0 m3 m–3, {xi} = 1.675, {psi}L = 1500 kPa and µ = 0.706. Field capacity (FC) for the relatively light soil was estimated at approximately {psi}FC {cong}10 kPa, or {theta}FC {cong} 0.12 m3 m–3. Soil samples were taken in each plot with a soil auger (0- to 0.2-, 0.2- to 0.4-, 0.4- to 0.6-, and 0.6- to 0.8-m depth) before the experiment on 7 June and again on 13 October. The location of the soil sampling was as follows: in the center of each designated plot on 7 June before planting and on the drip line half-way between two emitters on 13 October. In the laboratory, gravimetric soil water content and mineral and salt distribution (N, P, K, Cl, Na, pH, EC, and sodium exchange ratio [SAR]) were determined in saturated soil paste using standard procedures (Page et al., 1986).

Irrigation Scheduling and Seasonal Water Use
Crop water requirements ETc (mm d–1) or net irrigation water requirements Inet (mm d–1) inside the screenhouse were estimated from potential evapotranspiration ETo using the single crop coefficient approach:

Formula 2[2]
where Kc is the crop coefficient depending on crop growth stage and crop type, adjusted to local climate as outlined by Allen et al. (1998), and Peff is effective rainfall, which was zero for the entire experimental period. The seasonal Kc curve was constructed with the values of 0.76, 1.15, and 0.9 at the beginning, middle, and end of the growth season, respectively, and the duration of the initial, plant development, mid-season, and final growth stage was 25, 35, 40, and 50 d, respectively. Daily ETo inside the screenhouse was calculated as follows:

Formula 3[3]
where {Delta} is the slope of the vapor pressure curve, Rn and G are net radiation at the crop surface and soil heat flux density, respectively, {gamma} is the psychrometric constant, T is mean daily air temperature at 2 m height, u2 is wind speed at 2 m height and es and ea are saturation and actual vapor pressure, respectively (see Allen et al., 1998).

Required amounts of irrigation water Igross (mm d–1) were determined using Eq. [4]:

Formula 4[4]
where {eta}irri is the irrigation efficiency, {eta}con is the conveyance efficiency (it was assumed that: {eta}irri x {eta}con = 0.90) (James, 1988; Shannon et al., 1996) and LF is the leaching fraction (LF = 0.09), defined as the depth of water leached below the root zone relative to the depth of water applied at the soil surface (Ayers and Westcot, 1985). The irrigation computer was adjusted at 3- to 5-d intervals to the meet predicted Igross. Quantity and quality (EC, pH) of the applied irrigation water were measured daily at the head control.

For the entire growing season ETo, predicted crop water requirements ETc and scheduled irrigation amounts Igross below the screen were 442, 422, and 510 mm, respectively. Predicted seasonal water requirements Igross using external weather data were 823 mm, indicating that screenhouse cultivation offers a water saving potential of >35%. The total amount of applied water in this experiment was 554 mm for DS100, PS100, DF100 and PF100 and 665 mm for DS125 and PS125, respectively. These are very low figures in light of the arid summer climate at the site.

Plant Measurements
Leaf conductance was measured on 12–14 Sept. 2004 in eight sunlit well-developed leaves per treatment using a steady state porometer (LiCor LI 1600, Campbell Scientific, Logan, UT, USA). Real-time plant water uptake was measured as SF using a lysimeter-calibrated heat pulse system (Cohen et al., 1988; Cohen, 1994; Möller et al., 2004). Four sets of SF data were collected, each covering a period of 5 to 7 d in 20 well-developed plants in two treatments, that is, 10 plants per treatment. Meteorological data and SF data were recorded at 10- and 0.8-s intervals, respectively, and 30 min averages were logged in a set of data acquisition systems (CR7, CR10X, Campbell Scientific).

Plant dry weight, plant height, and leaf area were recorded continuously throughout the growing season. For leaf area index (LAI) determination, all leaves from four plants per treatment were removed and analyzed with an optical area meter (Delta-T Devices, Cambridge, UK), and the measured leaf area per plant was then multiplied with plant density (3.12 plants m–2). Total crop yield and yield of export quality were recorded for all 11 harvesting cycles starting on 8 Sept. through to 15 Nov. 2004. Rows adjacent to the northern and western sidewalls of the screenhouse were not included in the yield analysis. Samples of fruits and old and young leaves were taken from each treatment on 13 October and analyzed in the laboratory for concentration of N, P, K, Cl, Ca, Mg, Na, Fe, Zn, Mn, and B. The dry plant material (DM) was ground to pass an 850-µm sieve. Samples (100 mg) were wet-ashed with H2SO4–H2O2 and analyzed for Na, K, organic-N, and P. HClO4–HNO3 ashing was used to analyze for Ca, Mg, and micronutrients. Element concentrations were determined as follows: NH4–N, NO3–N, and P with an injector Lachat Autoanalyzer (Zellweger Analytics, Milwaukee, WI) K and Na with a flame photometer; Ca, Mg, Fe, Zn, and Mn with an atomic absorption spectrophotometer. Chloride was extracted from the dry leaf powder by vigorous shaking in deionized water (with a water/leaves weight ratio of 10:1) for 1 h at room temperature. The extract was filtered and Cl was determined with a Chloridometer.

Statistics
Results were analyzed by analysis of variance (ANOVA) using the GLM routine of SAS (anonymous, 1982) where salinity, water amount, and irrigation frequency were defined as class variables. Interactions between the classes were also tested. Differences between treatments were considered significant when type III sum of squares met the F-test criterion at <0.05 probability. When ANOVA showed that significant differences existed between means, the Duncan multiple range test DMRT (with {alpha} = 0.05) was used to determine which means were significantly different.

Numerical Simulation
The temporal evolution of the soil water content during a day was simulated using the HYDRUS-2D model (Simunek et al., 1996) for the DF100 and the PF100 treatments. The size of the modeled flow domain was 0.2 m in width and 1.0 m in depth, a vertical symmetry plane perpendicular to the drip line, from the emitter to halfway the distance between the drip lines, assuming the line source wetting condition. The model domain was assumed to have uniform soil physical properties. Vertical boundaries were assumed to be no-flow boundaries, and the lower boundary, to be a free-drainage boundary. The water application for the PF100 treatment was simulated assuming a continuous low wetting rate equal to the time-averaged rate of the pulsed application over an irrigation interval of 10 h. In this sense, the simulation procedure assumes that temporal evolution of soil water content under pulsed irrigation is similar to that under microdrip irrigation. Therefore, this simulation addressed two issues: (i) comparing between the soil water content dynamics during the two water application methods; (ii) checking the assumption that pulsed and microdrip irrigations are alternatives.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Physiological Parameters
The impact of the treatments on vegetative growth is presented in Fig. 1 , which shows the seasonal development of the LAI, and in Table 2, which summarizes plant height, plant weight, and LAI measured throughout the growing season. Irrigation with freshwater (DF100 and PF100) produced the highest leaf area (Fig. 1) and dry matter (Table 2). Three months after transplanting (20 September), high frequency irrigation with freshwater PF100 had led to an increase of leaf area and plant weight, as compared with the once daily irrigation DF100. However, the latter increase was not distinguishable toward the end of the experiment (2 November). Lowest LAI, plant height, and plant dry matter production occurred in the saline treatments DS100 and PS100. The increased LF under saline water (DS125 and PS125) reduced the negative effect of salinity, as LAI, plant height, and plant weight were larger than in DS100 and PS100 but smaller than under freshwater DF100 and PF100. Under saline water DS100–PS125, no significant effect of the irrigation frequency on physiological parameters was detectable, except for LAI, which on 20 September was found to be higher under high-frequency irrigation (p < 0.01).


Figure 1
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1. Seasonal course of leaf area index (LAI) for the different treatments, measured with an optical area meter. Vertical bars represent two standard errors of the mean.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Average plant height, above ground vegetative dry matter (excluding fruits) and leaf area index; measured throughout the growing season. Numbers with the same superscript are not significantly different at {alpha} = 5% (Duncan Multiple Range Test, DMRT). Levels of significance are from ANOVA.

 
The diurnal course of total leaf conductance for 13 to 14 September is shown in Fig. 2 . No full stomatal closure was detectable in any of the six treatments, indicating that severe water stress did not occur (Yao et al., 2001). On both days stomatal conductance was highest in the freshwater treatments DF100 and PF100. These findings are in agreement with observations made above for other growth parameters (see Fig. 1 and Table 2). Differences in leaf conductance of DF100 and PF100 as compared with the saline treatments were mostly significant (p < 0.05) from the morning until early afternoon, depending on the specific treatments that were compared. On both days between 1100 and 1400 h, leaf conductance in the high-frequency treatment PS125 was lower than in DS125. There was no clear and significant trend when comparing DS100 and PS100 on 13 September. On the second day of measurement however, leaf conductance in DS100 was significantly lower (p < 0.05) than in PS100 and lower than in all other treatments (except for PS125), an observation that was not observed in any of the other physiological parameters presented above (see Fig. 1 and Table 2).


Figure 2
View larger version (22K):
[in this window]
[in a new window]
  		Fig. 2. Diurnal course of total leaf conductance of sunlit leaves for the different treatments, measured on 13 and 14 September. Vertical bars indicate two standard errors of the mean.

 
Soil Water Regime
Figure 3 shows the temporal evolution of volumetric soil water content {theta} corresponding to the tensiometer data at the 0.2-m depth in the freshwater treatments DF100 and PF100 on 25 July. Soil water content in the active root zone of both treatments (upper 0.3 m) was high and continuously exceeded field capacity FC. Under daily irrigation DF100, peak soil water content at the 0.2-m depth occurred approximately 50 min after the start of irrigation (t = 0). In contrast, under high frequency irrigation (PF100), the local maximum of {theta} at the 0.2-m depth was lower than in DF100, but had the shape of a plateau lasting approximately 4 h, which coincided with the period of maximum plant water demand (see below). Soil water content at the 20-cm depth in DF100 was higher than in PF100 in the morning, but was larger under high frequency PF100 during the remainder of the day.


Figure 3
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 3. Soil water content for the DF100 and the PF100 treatments, calculated from tensiometer measurements at the 20-cm depth using Eq. [1], and the corresponding simulated {theta}(t) curves (solid lines). Irrigation begun at t = 0. The dotted horizontal line is estimated field capacity, FC.

 
In Fig. 3 are also depicted the simulated {theta}(t) curves corresponding to the two treatments (solid lines). The overall trend of the simulated {theta}(t) represent quite well the measured one corresponding to each of the water application method. The measured values are accurately reproduced for daytime (0–12.00 h) but underestimate water content at night. This difference can be explained by the redistribution process that occurs into the root zone after plants have ceased to uptake water. This process is not accounted for in the simulation since the root system is not represented deterministically, and the uptake function varies with depth but is applied equally to the whole domain at each depth. Consequently, the soil continues to dry in the simulation while data indicate some rewetting due to redistribution before drying took place again.

Although slight fluctuations of {theta}, corresponding to the pulse application, can still be observed at 20 cm, they are negligible by all practical means. The simulation assuming a continuous low wetting rate for the whole period of the pulsed irrigation provides a good estimate of the water content dynamics during the day. Therefore, one can assume that, at the 0.2-m depth, microdrip and pulsed irrigation produce similar water content regimes, and can be considered as alternatives.

Figure 4 depicts the diurnal dynamics of {theta} in DF100 and PF100 on five selected days during the irrigation season at two depths, 0.2 and 0.4 m. Apparently, more water than required was applied at the beginning of September and October, as indicated both by the irrigation records and the diurnal course of {theta} at the 0.2- and 0.4-m depth. Water availability in the active root zone was lower on 1 August and 1 November. Under these conditions, average and peak soil water content and therefore water and nutrient availability in the upper 20 to 30 cm was higher under pulsed irrigation PF100. This might indicate that under limited water stress conditions, increased irrigation frequency could lead to higher water and nutrient availability in the active root zone, when compared with an irrigation regime with one application per day. This is revealed also by the different responses of the soil–plant system to the conditions in the 24 July and the 1 August, during which identical amounts of water were applied while different Igross values characterized these 2 d.


Figure 4
View larger version (42K):
[in this window]
[in a new window]
  		Fig. 4. Average soil water content for the DF100 and the PF100 treatments for 5 d throughout the irrigation season, calculated from tensiometer measurements at (a) 20- and (b) 40-cm depth using Eq. [1]. The dotted horizontal line is estimated field capacity, FC. Computed gross water requirements, Igross, and applied water amounts, Iapplied, are also indicated.

 
The soil at the 0.4-m depth dries during all the daytime period, due to water uptake in the root zone, the minimum value of {theta} being attained around 20.00 h. Then, rewetting occurred at night, due mainly to redistribution following the end of water uptake, and to possible effects of temperature on tensiometer readings (Warrick et al., 1998). At the early stage of plant growth, and when there were only small differences between Iapplied and Igross, {theta}(t) curves were similar for both treatments (24 July and 1 August). When Iapplied was significantly higher than Igross, and the plants were well established, {theta} values for PF100 were higher than those for DF100 (1 September–1 November), indicating that water uptake was lower from the 0.4-m depth, or that drainage rate from above layers was higher, under pulsed irrigation.

Soil Salinity Regime
Figures 5 (a-c) depict the vertical profiles of EC of the mean saturated extract ECe for the different treatments, measured before planting, on 7 June, and 4 wk after the first harvesting cycle, on 13 October. On 7 June, no significant differences were found between the treatments, and values were equal, or smaller than the threshold value for sweet pepper (ECt = 1.7 dS m–1). Soil salinity was slightly higher in the top 40 cm of DS125 and PS125, reaching 2.2 and 1.9 dS m–1, respectively (Fig. 5b).


Figure 5
View larger version (18K):
[in this window]
[in a new window]
  		Fig. 5. Vertical profiles of electrical conductivity of saturated extracts ECe in (a) DS100 and PS100, (b) DS125 and PS125, and (c) DF100 and PF100, measured before planting (7 June) and after the first half of the harvesting season (13 October). Horizontal bars indicate two standard errors of the mean. Vertical dashed lines on the right- and left-hand side are salinity of the saline ECiw_sal and non-saline ECiw_fresh irrigation water and the reference threshold salinity value for sweet pepper ECt (Rhoades et al., 1992), respectively.

 
On 13 October, salinity in the soil profiles of the saline treatments S ranged between 2.1 and 3.5 dS m–1 (Fig. 5a, 5b), exceeding the threshold value for sweet pepper ECt. However, observed soil salinity was smaller than the EC of the irrigation water (dashed line ECiw_sal = 4.2 dS m–1), indicating that only part of the added salt did accumulate in the root zone, independently of the water application frequency. Due to the saline water application at the soil surface, ECe was highest in the top soil layer of all treatments and generally decreased with depth. For the once daily saline application (DS100 and DS125), improved leaching in DS125 reduced salinity in the upper 20 cm (ECe = 2.8 vs. 3.4 dS m–1 in DS100) and increased salinity at the 40-cm depth (ECe = 2.4 in DS125 vs. 2.1 dS m–1 in DS100). Under pulsed saline irrigation (PS100, PS125), increased leaching in PS125 had no significant effect on salinity in the upper 20 cm, when compared with the normal water amount PS100. However at the 40- and 60-cm depth, ECe in PS125 was lower than in PS100, and larger than in PS100 at the 80-cm depth.

The salinity increase throughout the experiment was moderate in DF100 and PF100, as was expected. Only in the top 20 cm did ECe exceed the threshold value for sweet pepper ECt (Fig. 5c). Soil salinity in the top layer was higher in PF100 (2.3 vs. 2.0 dS m–1 in DF100). Similar observations were made in the top layer of PS125 vs. DS125 and in PS100 vs. DS100, with the high frequency treatments leading to mostly significant increased salinity of the saturated soil extract. This seems to indicate that more solutes accumulate in the root zone under pulsed irrigation when compared with once daily applications. One of the main conclusions of pioneering studies on irrigation with saline water (Biggar and Nielsen, 1967; Mokady and Bresler, 1968; Bresler et al., 1982) was that leaching with irrigation water is more efficient when the soil is maintained unsaturated and the flow rate is relatively low. The results in Fig. 5 (a-c) are at least apparently contradictory to these conclusions. One explanation could be that, in these studies, much lower frequencies and much higher application rates were applied, as irrigation at that time was mainly sprinkler irrigation applied on a weekly basis. The reduced leaching in pulsed irrigation is certainly a desirable benefit with regard to protecting the quality of ground water resources. On the other hand, when saline irrigation water is used as in our experiment, the increased accumulation of salts in the root-zone could lead to a higher potential yield loss of sensitive and moderately sensitive crops, such as bell pepper.

Plant Water Uptake
The diurnal curves of SF in the fresh water treatments DF100 and PF100 are shown in Fig. 6a for two consecutive days. Sap flow data are plotted against computed half hourly ETo in Fig. 6b, for two periods within the day, before 1330 h and after 1330 h. Sap flow in both treatments did not differ significantly. The small nonsignificant differences observed in Fig. 6 between the treatments do agree with other observations of increased activity in DF as opposed to PF treatments (e.g., leaf area and dry weight on 20 September, see Table 2). For most of the day, SF was slightly (nonsignificantly) higher in DF100, except for the period between 1330 and 1500 h on both days, when SF was somewhat higher in the pulsed irrigation PF100. This corresponded with the period of maximum soil water content at the 0.2-m depth in that treatment (Fig. 3). However, the subsequent decrease of SF in PF100 relative to DF100 could not be explained by soil water content dynamics at that depth—in fact it ran counter to it, with {theta} for PF100 exceeding {theta} for DF100 between 1300 h and the next morning. It is interesting to consider how many SF measurements would be needed so that measured differences in the order of 0.01 mm h–1 could be significant. Based on the standard deviation of the SF measurements (coefficient of variation of 18.5%) it would require 118 replicates per treatment to get a standard error of 0.005 mm h–1. This reasoning demonstrates the problematic nature of SF measurement when looking for small differences between treatments, as has been noted before (Cohen, 1994). On 15 and 16 September, daily SF totaled 2.6 and 2.7 mm d–1 in DF100, and 2.4 and 2.5 mm d–1 in PF100, respectively. The standard errors were 0.2 mm or 8% and 0.1 mm or 5% in DF100 and PF100, respectively; hence no significant difference was detectable between the two fresh water treatments.


Figure 6
View larger version (28K):
[in this window]
[in a new window]
  		Fig. 6. (a) Diurnal course of sap flow (SF), measured in fresh water treatments DF100 (daily application) and PF100 (high frequency) on 15 and 16 September. Vertical bars indicate one standard error of the mean (seven plants per treatment). (b) Half hourly sap flow data in DF100 and PF100 versus ETo calculated from climate data measured inside the screenhouse, 15 September. The solid symbols correspond to data until 1330 h, and the empty ones, to data after 1330 h. The dashed line is the 1:1 line. Linear regressions for the entire day DF100: SF = 0.86 x ETo, R2 = 0.90; PF100: SF = 0.83 x ETo, R2 = 0.92. Vertical bars represent one standard error of mean SF.

 
In both treatments SF lagged behind computed ETo, the main determinant of which is incoming radiation (Fig. 6b). This common phenomenon has been widely reported and is attributed to plant capacitance (changes in plant water content) during morning hours (Fuchs et al., 1987; Hunt et al., 1991; Möller et al., 2004). It was particularly pronounced in PF100 for the period from 1030 until 1300 h. The presented physiological and soil moisture content data (Fig. 1 to 4) exclude the possibility of water stress being the cause for the observed lag in SF. On the other hand, the higher plant weight in PF100 (Table 2) may indicate a higher plant capacitance which could explain the observed lag in SF.

Figure 7 depicts the response of the soil–plant system to freshwater at standard irrigation frequency (DF100) and saline water at high irrigation frequency (PS100) during 4 consecutive days, 30 September until 3 October. Sap flow in DF100 was higher than (p < 0.05) or equal to that in PS100. Periods of significant differences are highlighted in Fig. 7b. It is shown that high irrigation frequency PS100 does not compensate for the salinity of the water and that daily plant water consumption was reduced. This corroborates physiological data presented above (see Fig. 1 and 2, and Tables 2 and 3). In both treatments, relative SF was around 0.5 to 0.6 in the morning and increased steadily throughout the day to 1.25 to 1.4 in the late afternoon. This resulted from changes in plant water content (see Fig. 6b), as discussed above.


Figure 7
View larger version (34K):
[in this window]
[in a new window]
  		Fig. 7. Diurnal course of sap flow, (a) measured in DF100 and PS100, (b) SF relative to reference evapotranspiration ETo calculated from climate data measured inside the screenhouse, (c) soil water content calculated from tensiometer measurements at 20 cm depth using Eq. [1]. The dotted horizontal line in Fig. 7c is estimated field capacity, FC. Computed gross water requirements, Igross, and applied water amounts, Iapplied, are also indicated. In the upper part of Fig. 7b, the symbol {blacksquare} indicates periods during which sap flow in DF100 and PS100 were significantly different from each other (p < 0.05).

 

View this table:
[in this window]
[in a new window]
  		Table 3. Concentration of selected elements in young and old leaves, in fruit dry matter, and in the 1:1 soil extract, measured on 13 October. Numbers with identical superscript are not significantly different (Duncan Multiple Range Test; DMRT, {alpha} = 5%).

 
Soil water content at the 0.2-m depth in DF100 exceeded PS100 for 2 to 3 h in the morning on 30 September to 2 October and until noon on 3 October (Fig. 7c). However, all {theta} values were well above FC throughout, suggesting that the observed reduction in plant water uptake under PS100 (see Fig. 7a, b) can primarily be attributed to the higher soil salinity in this treatment (Fig. 5) rather than to a localized shortage of water.

Over irrigation increased successively from 0.5 mm on 30 September to 1.0 mm on 2 and 3 October, due to a decline in ETo at constant amounts of applied irrigation water (Fig. 7c). This decline of ETo is also reflected in SF (Fig. 7a), which decreased in both treatments during that period. On 2 and 3 October, this led to increased peaks in soil water content in the daily fresh water treatment following irrigation (0.42 m3 m–3 as compared with 0.31 m3 m–3 on 30 September) and the 24 h average rose from 0.19 m3 m–3 on 30 September to 0.27 m3 m–3 on 3 October. In contrast, no such increases in soil water content were observed under the high frequency treatment PS100. The daily maximum of {theta} remained unchanged at 0.31 m3 m–3 and the 24 h average of 0.26 m3 m–3 remained practically constant over the 4 d. These results indicate that, unlike the findings of Shock et al. (2005), differences do exist between the water content regime of the different irrigation frequencies, and that average soil water content in the active root zone could be increased by higher irrigation frequency, when compared with conventional once daily applications. This might be of practical interest when deficit irrigation is considered. However, the experimental design in the current study did not allow a more detailed analysis of this hypothesis.

Both, data logger capacity and the number of available SF probes allowed SF measurements only in two treatments at a time (10 probes per treatment). A comparison of daily plant water uptake (SF) normalized to ETo in all treatments is presented in Fig. 8 . All data were collected over a period of 3 wk during the mid season development stage, when the crop coefficient Kc was estimated as 1.15. The freshwater treatments were highest and crop water use in DF100 was 5 to 10% higher than under high irrigation frequency PF100. Differences between DF100 and the saline treatments S were mostly statistically significant, depending on treatment and day. In DF100 and PF100, measured water consumption was on average 15 and 22% lower than computed crop evapotranspiration ETc, respectively. One reason might have been partial ground cover, which was estimated at 70 to 80% during mid-season (see also Fig. 1). Ground cover was not included in the computation of ETc because it was assumed that the relatively tall plants grown vertically using a "Dutch" trellis would offset the effect of partial ground cover. Our data seem to indicate that this assumption might have been incorrect. Another cause for SF remaining well below ETc might be soil salinity, which ranged between 1.75 and 2.3 dS m–1 in the top 40 cm of the freshwater treatments DF100 and PF100 (see Fig. 5c). Actual soil salinity thus exceeded the threshold values for bell pepper suggested by Bresler et al. (1982) and Rhoades et al. (1992), above which reductions of yield and transpiration must be expected.


Figure 8
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 8. Daily totals of measured sap flow (SF) normalized with reference evapotranspiration inside the screenhouse (ETo), for the different treatments. The horizontal axis refers to the number of days after respective probe insertion. Bars indicate two standard errors of the daily means of treatments DS125 and DF100, respectively. Error bars for the other treatments were similar.

 
Plant Nutrient Uptake
The effect of the different treatments on water uptake reflects also on the uptake of minerals during fertigation. Results of the analysis of the element concentration in young and old leaves, fruits, and in the top soil (0–40 cm) are presented in Table 3. Analysis of variance indicated highly significant impacts of the treatments on the concentration of P, Cl, Ca, and Mg in young leaves (Table 3) and N, P, Cl, Ca, Mg ,and Mn in old leaves (Table 3). No significant impact of the irrigation treatments on the concentration of K, Na, and Fe concentration in both young and old leaves was detectable. The chloride level clearly shows the effect of the salinity treatments as was also reported for field grown pepper irrigated with saline water (De Pascale et al., 2003). Increasing the irrigation amount (DS125 and PS125) slightly decreased the uptake of Cl (old leaves), and for freshwater, leaf-Cl concentration under the high-frequency treatment PF100 was approximately 50% higher than under DF100, in agreement with the trend observed for soil salinity (see Fig. 5c). There was a strong correlation between ECe in the upper 20 cm of the soil and the leaf-Cl concentration (R = 0.97 and 0.92 for young and old leaves, respectively, both significant at p < 0.01). Phosphorous and Mn uptake were also affected by the application rate. High irrigation frequency (PS100, PS125, and PF100) increased the availability and consequently the uptake efficiency of P and Mn when compared with the once daily application (DS100, DS125, and DF100). These findings are in line with observations made by Xu et al. (2004) and Silber et al. (2005), who showed that P concentration in lettuce and bell pepper was directly correlated with irrigation frequency. In contrast, the high frequency treatments in this experiment reduced Mg and Ca content in mature leaves (Table 3), except for Ca in DS100 and PS100.

High irrigation frequency improved the concentration of Mn in the fruits (Table 3; p < 0.05). The concentrations of N, P, and K in the fruits under high frequency irrigation were not significantly higher than under once daily application. The concentration of Ca was lowest for both daily irrigation treatments at 100%, and highest for daily 125% irrigation, with all high frequency treatments giving intermediate values. Irrigation water salinity did not have a significant impact on fruit mineral concentration, except for K, which was reduced under the saline water treatments S (p < 0.05), indicating a major interaction between soil Na (Table 3) and K uptake similar to the tomato crop, as reported by Adams (1991) and Adams and Ho (1995).

Increased irrigation water salinity led to significantly higher concentrations of Na, Cl, B, Fe, and Mn in the upper 40-cm soil layer, while the Mn concentration was lower in the saline treatments S (Table 3). Soil K concentration was consistently lower under high frequency irrigation (p < 0.01) but this has apparently not affected K concentrations in leaves and fruits (Table 3). Comparing soil Na and Cl concentration at daily D and pulse P application (Table 3) confirms the observations made in Fig. 5: the higher irrigation frequency led to increased salinity in the top soil (p < 0.05 for Na and n.s. for Cl). For none of the elements did statistically significant correlations exist between element concentration in the soil and concentration in fruits or leaves, except for Cl (R = 0.83 and 0.87 for young and old leaves, respectively, p < 0.05).

Crop Yield and Water Use Efficiency
Values of total export yield after the 11th (final) harvest cycle are presented in Table 4. No significant effect of the irrigation frequency on the total number of fruits, export yield and average fruit size could be seen. Export yield was lower under the saline treatments (p < 0.05), that is, yields in DS100 and PS100 were 17.2 and 14.4% lower, respectively, than in the freshwater treatment DF100. Yield under saline water with increased water amount (DS125 and PS125) showed intermediate results, being 9.0 and 8.2% lower, respectively, than in DF100. Under fresh water export yield was higher- although statistically not significant—with once daily irrigation D than under the pulse regime P, thus confirming results reported by Shock et al. (2005). An opposite trend was observed in the saline treatments, where slightly higher yields- but insignificant- were observed under high frequency application P. A higher average fruit weight occurred under fresh water F (p < 0.01), confirming observations made by Sonneveld and van der Burg (1991) and De Pascale et al. (2003) in bell pepper. Generally, export yields recorded in this experiment are lower than the average yield achieved by local growers (7–10 kg m–2). However, it is highlighted that planting for this experiment (18 June) was significantly later than local practice (first half of May), thus reducing the duration of the harvest period.


View this table:
[in this window]
[in a new window]
  		Table 4. Seasonal export fresh yield at final harvest (15 November). Numbers with the same superscript are not significantly different at {alpha} = 5% (DMRT). Levels of significance are from ANOVA.

 
The impact of soil salinity, measured as electrical conductivity of the saturated extract ECe, on relative yield Yr (assuming that Yr {approx} 100% for DF100) is shown in Fig. 9 . There was a significant decrease in relative yield at increasing soil salinity (R = –0.78; p < 0.05), which is in line with expectations. The estimated slope b of the yield response curve for this experiment was determined as 10.4% dS–1 m, which is lower than the 12% dS–1 m suggested by the FAO (Rhoades et al., 1992) and the 14.1% dS–1 m proposed by Bresler et al. (1982). Assuming a threshold value of a = 1.7 dS m–1, the maximum permissible salinity ECe_max for bell pepper in this experiment (i.e., when Yr becomes zero) was estimated at 11.3 dS m–1, compared with 10.0 dS m–1 suggested by the FAO. Values recommended by Bresler et al. (1982) are 1.5 and 8.6 dS m–1 for the threshold a and maximum value ECe_max, respectively. The FAO data is based on surface irrigation methods (furrow and flood), while data in Bresler et al. (1982) were mostly obtained under sprinkler irrigation at weekly intervals. Hence the findings of this study indicate that drip irrigation slightly improved the yield response of bell pepper, but was still relatively independent of the irrigation method and irrigation frequency in the ECe range from 1.5 to 4.0 dS m–1.


Figure 9
View larger version (27K):
[in this window]
[in a new window]
  		Fig. 9. Export yield response of bell pepper for the different treatments, relative to cumulative export yield in treatment DF100 recorded on 15 Nov. 2004, versus average soil salinity ECe (0–40 cm) measured on 13 Oct. 2004. Rhoades et al. (1992) (—): a = 1.7 dS m–1, b = 12% dS–1 m, ECe_max = 10.0 dS m–1; Bresler et al. (1982) (—): a = 1.5 dS m–1, b = 14.1% dS–1 m, ECe_max = 8.6 dS m–1; this experiment: a = 1.7 dS m–1, b = 10.4% dS–1 m, ECe_max = 11.3 dS m–1; N = 6; R2 = 0.61; significant at p < 0.05.

 
Cumulative IWUE, taken as the ratio of total fresh yield (kg) and applied irrigation water (m3) is shown in Fig. 10 . IWUE was highest under the freshwater treatments DF100 and PF100, reaching values of up to 10.7 kg m–3 at final harvest (DF100), while IWUE was 9.5 kg m–3 for DS100 and PS100 and 8.8 kg m–3 for DS125 and PS125, respectively (Fig. 10).


Figure 10
View larger version (27K):
[in this window]
[in a new window]
  		Fig. 10. Seasonal curve of irrigation water use efficiency IWUE for the different treatments, based on cumulative total yield and seasonal amount of applied irrigation water.

 
The IWUE under the freshwater treatments DF100 and PF100 was more than threefold of the upper IWUE guideline for bell pepper that has been suggested by the FAO. Doorenbos and Kassam (1979) proposed an achievable IWUE in the range of 1.5 to 3.0 kg m–3. The large difference between our results and those covered by the FAO probably results from the combined influence of the screenhouse in reducing radiation load and potential evaporation, as well as the increased efficiency of current irrigation systems.

The IWUE under increased water amounts DS125 and PS125 were lower than in all other treatments, including the saline treatments with normal water amount DS100 and PS100. Increasing the scheduled LF under saline conditions from 9% (as in DS100, PS100) to 27.3% (DS125, PS125) reduced yield loss (see Table 4). However, the water contributing to avoiding this yield loss has been used quite inefficiently, as expressed by the comparatively low values of final IWUE for DS125 and PS125. For all six treatments, no consistent impact of the irrigation frequency on IWUE was evident; however the lowest IWUE occurred under pulsed irrigation with saline water at increased amounts PS125.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
In the natural loamy sand soil, vegetative growth, SF and yield were negatively affected by high irrigation water salinity. Compared with a 1-d irrigation interval, high irrigation frequency did not have a significant impact on final yield and physiology of the test crop, although it had some positive impact on plant height and leaf area at the early stages of plant growth. Assouline et al. (2002) compared irrigation frequencies ranging from 7 d (2 L h–1 regular flow rate) to 1 d (0.25 L h–1 low flow rate) in sweet corn and found that one application per day provides improved growth conditions leading to highest yield. When full crop water requirements are supplied at daily intervals, a further increase of the irrigation frequency up to 10 times per day might not have added significant additional benefits over the once daily application. This argument is supported by measured soil water content in the DF100 control (regular irrigation frequency), which showed that soil water potential in the active root zone was at or above field capacity (see Fig. 4). On the contrary, the absence of a once daily "flush" of salts to lower soil layers as in the control (see 1 September and 1 October in Fig. 4a and 7c) is believed to be the major cause of increased soil salinity in the active root zone under high frequency irrigation. The presence of this "flush" caused by peak soil water contents following irrigation in DF100 led to this treatment having slightly higher SF and yield compared with the high frequency treatment PF100. The commonly applied approach when it comes to consider/design salt leaching relies on the comparison/ratio between water application and plant water uptake, neglecting changes in water content in the root zone due to water application methods and frequencies. According to this approach, total leaching amounts in PF100 were higher than in DF100 (smaller SF at identical Iapplied). However, the observed salinity level in the active root zone was higher under PF100 than under DF100, The spatial distribution of salts concentration in the root zone is complex under drip irrigation, and depends on the application rate. The results indicate that under drip irrigation, comparing between applied and evapotranspired water solely may not be sufficient to estimate salt leaching efficiency/requirements. Changes in the water content regime in the root zone, which are strongly affected by the application method (Fig. 3 and 4), and consequently on the spatial distribution of solutes concentration, should be accounted for as well.

Soil moisture data (Fig. 4a) indicate that in the absence of over-irrigation such as on 1 August, the average soil water content in the root zone is improved under higher irrigation frequency. It is therefore suggested to examine the impact of high frequency drip irrigation under deficit water applications, aiming at further maximizing IWUE.

In contrast to the findings for soil salinity, high irrigation frequency showed positive effects on the concentration of Mn in the fruits, and led to higher concentrations of N, P and Mn in the leaves. Investigating the same crop at the same site but in a previous year, Silber et al. (2005) reported significant advantages of high frequency irrigation at very low P rates. However, the number of pulses was significantly larger than in the current experiment (25 instead of 10), and this factor may affect P availability as it addresses directly the dynamics of P in soils.

This study demonstrated that significant amounts of water could be saved under screenhouse conditions. The water savings achieved in this study reached up to 33% when compared with open field reference evapotranspiration. Under the arid summer climate, total water use was only 554 mm and none of the treatments showed detectable signs of water stress. Consequently, IWUE was as high as 10.7 kg m–3 under freshwater and 9.9 kg m–3 under saline water, respectively. It is strongly suggested to schedule screenhouse irrigation based on reference evaporation measured below the screen. Successful application of these improved management procedures could relieve some pressure on strained water resources. Alternatively, a given water quota could suffice to irrigate a larger agricultural area. Nevertheless, SF data and soil moisture measurements suggest that irrigation amounts under freshwater could have been reduced even further on most days. Under saline conditions however, a sufficient LF should be maintained, and hence water amounts should not be reduced any further.

Another interesting point is the water need estimates at the early stages of the plant development. From Table 2, it appears that water amount had a significant impact on the early stage of plant growth. Therefore, water needs of young plants may be underestimated when evaluated in proportion to their size, as it is usually done. Consequently, Kc estimates (Eq. [2]) of young plants using the common approach might have to be reconsidered and probably increased.

We observed a sharp increase in IWUE compared with the 1.5 to 3.0 kg m–3 guideline for bell pepper given by the FAO in the 1970s (Doorenbos and Kassam, 1979). This can be attributed to the combined impact of a variety of factors, such as improved irrigation and fertilization schedule, frequency and method, crop management, new crop varieties and protected crop production under a screen in this study. This demonstrates that over the last quarter century, modern agriculture has made significant progress in showing feasible ways and methods for more efficient use of limited water resources.


   ACKNOWLEDGMENTS
 
The authors thank E. Matan, H. Yehezkel, D. Shmuel and the staff at the Besor Experimental Station for technical assistance in the field experiment. They also thank Terry Howell and three anonymous reviewers for their constructive comments. This study was funded by the U.S. Agency for International Development through CDR Grant No. TA-MOU-01-C21-019. The work of the second author was facilitated through the BMBF-MOST Young Scientist Exchange Program grant No. FZK04001/YSEP13. These supports are gratefully acknowledged.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Contribution No. 609/05 from the Agricultural Research Organization.

Received for publication November 6, 2005.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES