| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Centro de Ciencias Medioambientales, CSIC, Finca Experimental La Poveda, Crtra. Campo Real, km 1.3, 28500 Arganda del Rey, Madrid, Spain
* Corresponding author (rcaballero{at}iai.csic.es)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: D, drainage • EC, electrical conductivity • ECe, EC for saturated paste extract • ET, evapotranspiration • LF, leaching fraction • SD, standard deviation • TDS, total dissolved solids
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In the farms of the middle Jarama River Basin in central Spain, the most common crop sequences are either corn followed by corn, or corn followed by a winter cereal. The water and nutrient requirements of corn are seasonal, but corn producers do not typically consider the costs of over-watering or overuse of N fertilizer because both inputs are subsidized. Labor is the determining factor for irrigation frequency and depth. A majority of corn producers use rented labor and surface furrow irrigation. To reduce costs, they decrease the irrigation frequency and increase the irrigation depths. Simulating these conventional practices, Román et al. (1996) estimated that some 20% of irrigation water applied to corn was lost as drainage, and Cartagena et al. (1995) estimated losses of 120 kg ha-1 of NO-3–N to ground water during the corn growing season. This over-watering, however, may preclude salt accumulation in the soil profile, since salt stress for most crops has not been reported in the area.
The traditional irrigation practices will not be sustainable in the long term, due to enforcement of the European Union (EU) directives on the quality of ground water and the pricing of N fertilizers and irrigation water. An improved irrigation schedule (Román et al., 1999) in this low-rainfall zone would minimize water loss but, at the same time, could increase salt accumulation in the soil profile. A long-term evaluation of potential salt accumulation in the soil profile is therefore necessary before more efficient irrigation practices can be recommended to corn producers in the area.
The objective of this research was to determine whether water inputs adjusted to seasonal crop consumption would lead to soil salinization when compared with the over-watering due to traditional irrigation practices.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Irrigations were applied to the conventional cropping sequence in the area: cultivated periods that include 1 yr corn during the dry warmest season (from April to November) and a winter cereal (from December to the following early July), with corresponding periods between crop culture (inter-cultivated periods). During the 4-yr experiment, a sequence of corn–wheat–corn–oats was planted. Corn cv. Juanita (Pioneer) was planted in 1993 and 1995 with a density of seven plants per square meter and seeded rows were spaced 0.75 m apart. Wheat cv. Yecora and oats cv. Prodes 101 were planted in 1994 and 1996 at seeding rates of 150 and 140 kg ha-1, respectively, and rows were spaced 0.17 m apart. Soil fertility amendments included incorporation of corn and winter cereal stubbles. For seedbed preparation, a compound fertilizer (0-6.1-5.8, N-P-K) and K2SO4 were applied to corn at rates of 714 and 100 kg ha-1, respectively. Additionally, a single application of top-dressed mineral fertilizer was applied to corn at the rate of 150 kg N ha-1. The wheat and oat crops received 100 kg N ha-1 of top-dressed mineral N fertilizer.
Two sets of 12 plots (each plot measuring 9.9 by 11.1 m) were randomly assigned to either a traditional or improved irrigation schedule. Within each set of 12 plots, 4 plots were instrumented with one neutron probe access tube (to a depth of 2 m) surrounded by a set of vertical tensiometers placed at soil depths of 0.1, 0.2, 0.3, 0.45, 0.5, 0.6, 0.8, 1.0, 1.2, 1.5, and 2.0 m, capable of measuring water pressure head between 0 and -85 kPa. The full experimental set-up has been previously reported (Román et al., 1996).
Seasonal evapotranspiration (ET) and drainage (D) estimates were calculated by the zero-flux plane water balance equation. The hydraulic head (h) at a given soil depth was determined by adding pressure head (tensiometer measurement) to gravitational head (soil depth). The hydraulic head was used to determine water-flow direction in the entire soil profile (ISO, 1996) and soil depth of zero-flux plane (ze). The determination of ze, while water was moving in different directions and from different soil layers, required the measurement of h(z) at ti and tj for the period in which changing water flow occurred, so that a mean ze could be estimated. According to the irrigation sequence, four water-flow patterns were found and six cases of water-balance equation partitioning schemes were reported (Román et al., 1996). Most D periods took place during irrigation or occasional periods of heavy rainfall. Often, a few days after an irrigation or rainfall episode, ET causes a change in the water flow at the uppermost layer. If z is a soil depth deeper than the root zone (in our research site, z = 2 m), upward flow 0-ze, and downward flow from z > ze, were both commonly found. Thus the water-balance equation was used for the estimation of ET for a soil layer 0-ze and D for a soil layer ze to z. Water storage to z, estimated by the volumetric water content (
, neutron probe measurements), adds to water inputs (irrigation and rainfall) when depleted and to sinks (ET and D) when increased (Román et al., 1999).
Each of the 24 plots was also instrumented with three extraction tubes (63-mm i.d.) provided with porous ceramic tips to collect the soil solution (Nardeux Humisol, Les Ulis, France). The tubes were placed at soil depths of 0.5, 0.9, and 1.4 m. These depths were determined in a previous calculation of particle-size distribution in the soil profile and of the depth of the gravel layer.
During irrigation periods, neutron probe and tensiometer measurements were always performed just before and 2 to 3 d after irrigation. During nonirrigation periods, at least one measurement per month was performed, depending on rainfall and water flow directions. During the experiment, 141 and 116 measurements were performed in the traditionally and improved irrigated plots, respectively. The measurements by the neutron probe at 0.5-, 0.9-, and 1.4-m soil depths, estimated the water content in the ±0.05-m soil layer around the indicated measurement points.
Irrigation schedules applied to corn, wheat, and oat in the early phase of plant establishment were those applied by farmers during spring: occasional applications at low depths, depending on spring rainfall. Corn plots were watered two times at a rate of 37 mm in 1993 and three times at a rate of 21 mm in 1995. Wheat was watered five times for a total of 115 mm, and the oat crop was watered three times for a total of 76 mm.
Irrigation treatments (traditional or improved irrigation schedules) were applied to corn during the dry summer seasons (70–130 d after planting). At this phase, corn water consumption is 70 to 80% of total (Román et al., 1996).
In plots under traditional irrigation, which the majority of corn growers in the area use, the timing was an application every 7 to 10 d, and irrigation depths were according to potential ET computed using a weather-based system (Penman equation) by the Servicio de Asesoramiento de Riegos (SAR). Potential ET of 500 to 520 mm for the dry season were estimated with peak ET of 8 to 10 mm d-1 at the end of July (ITAP, 1995). These plots received eight applications with a mean depth of 65 mm in 1993 and seven applications with a mean depth of 70 mm in the same period in 1995.
Under the improved schedule, unrestricted water consumption by corn was the only predetermined condition. The timing was adjusted to avoid water stress. Timing of two irrigations per week in 1993 ensured that water stress was avoided, although tensiometer measurements just before irrigation (mean pressure head at soil depths of 0.1, 0.2, and 0.3 m was 31.5 ± 4.2 kPa) indicated that such a strict timing was unnecessary. During the 1995 growing season, corn was irrigated once a week and it resulted in no water stress (mean pressure head was -84.2 ± 10.5 kPa at the same soil depths). Field data, provided by the instrumented locations, were used to adjust irrigation depths by increasing water in storage to field capacity in the 0- to 0.5-m soil layer (prevalent root zone). The improved schedule thus followed the seasonal water consumption by the corn plant, not estimated by an ET weather-based system, but in-field measured.
A mobile-line sprinkler system was used for irrigation of the experimental field. Irrigation depths were adjusted by regulating the speed of the sprinkler lateral. Irrigation water from the Jarama River was used in the experiment. Irrigation water was sampled 23 times throughout the experiment, and its main properties are presented in Table 1.
|
The total dissolved salt load at soil depths of 0.5, 0.9, and 1.4 m was calculated seasonally as the product of mean soil water content and the total dissolved solids (TDS) of the soil solution at similar soil depths. Measurements of EC were used to estimate salt concentration (g L-1). The TDS was estimated by regressing solution EC on TDS, measured gravimetrically. Single linear regressions by soil depth were performed, but predicted values were not significantly different from those of a linear correlation of the three soil depths (Fig. 1).
|
Errors in the estimate of drainage volume were discussed in a previous article (Román et al., 1999). Errors associated with salt concentration were attributed to the estimation of TDS from EC (Fig. 1) and to the spatial variation of solution EC (Fig. 2).
|
Standard deviations (SD) of salt concentrations in the irrigation water and in the soil solution at three soil depths (0.5, 0.9, and 1.4 m) were calculated. For the former, SD indicated a temporal variation of irrigation water quality and was based on 23 sampling times. For the latter, SD indicated a combined temporal and spatial variation during the experimental period. A maximum of 732 samples per soil depth and irrigation schedule (12 samplers by 61 sampling times) went into the calculation of SD of salt concentrations in the soil solution (Table 1).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The main ions were SO2-4, HCO-3, Cl-, Ca2+, and Na+. Increases for individual ions at the deepest layer (1.4 m) varied between irrigation and soil-solution samples, the highest being NO-3, and the lowest HCO-3 (Table 2). The only ion that showed lower concentrations in the soil solution than in the irrigation water was K+, although its contribution to the total dissolved solids was very little. Salt distribution in the soil profile was relatively uniform, although for most of the main ions except HCO-3, the salt concentration at the intermediate soil layer (0.9 m) was slightly higher than at the highest (0.5 m) or lowest (1.4 m) layers (Table 1). For most salts, except HCO-3, higher concentrations and correspondingly higher EC were found for the improved irrigated plots than for the traditionally irrigated plots. This trend was observed in the three soil layers analyzed throughout the experiment (Fig. 2).
|
During both corn growing seasons (1993 and 1995), the EC of the soil solution in the lower root zone (0.5 m) stayed, for most of the season, between 0.4 and 0.5 S m-1 for traditional irrigation and between 0.5 and 0.7 S m-1 for improved irrigation. Standard limits rated crop salinity tolerance through the EC for the saturated paste extract (ECe). Ulery and Ernst (1997) calculated the EC of a soil solution from ECe and the saturation-to-field moisture ratio. In our experiment, this mean ratio was 1.8 for the root zone, and the calculated ECe was between 0.22 and 0.28 S m-1 in traditional irrigated plots and between 0.28 and 0.39 S m-1 in improved irrigated plots. Data reported by the Soil Improvement Committee (USDA, 1964; SIC, 1975; Maas, 1986) indicated a threshold value for ECe of 0.5 S m-1, after which corn yield may decline. Soil salinities in our experiment neared the threshold but grain yield showed no significant differences between the two irrigation treatments.
Examination of the concentrations in solution indicated that solubility reactions of gypsum (CaSO4·2H2O), calcite (CaCO3), or dolomite [CaMg (CO3)2] may govern soil-solution concentrations of low-solubility salts in our soil. Jury et al. (1978) reported precipitation of low-solubility salts under field conditions of low leaching fractions (LF), sulfate-rich irrigation water, and high irrigation frequency. In our experiment, the soil-solution concentrations presented in Table 1 suggest the dissolving of gypsum as the main source of salinity.
Total Dissolved Salt Load in the Soil Layers
Seasonal fluctuation of the dissolved salt load was calculated as the product of changes in salt concentration and soil moisture. It was found that under both irrigation treatments, the total dissolved salt load in the soil layers (Fig. 3 and 4) was more closely linked to changes in the soil water content (Fig. 3 and 4) than to changes in salt concentration.
|
|
During periods of increased soil moisture due to irrigation (around 190 and 900 d from the start of the experiment) or rainfall (around 320, 1020, 1390 d), salt loading reached maximum levels, although salt concentration showed fewer variations and a more steady trend. This pattern became more apparent in the top two soil layers (0.5 and 0.9 m) than in the deepest layer (1.4 m) because water content at this layer was less affected by the seasonal variation of water input. It seems that the diluting effect of the water inputs, particularly with rain water, was partially compensated for by the transfer (through the dissolution of salts) from the solid phase to the soil solution, particularly as a result of rain water.
During the period of lower soil moisture (summer 1994), from 570 to 660 d, (i) salt loading in soil layers reached its minimum level and (ii) salt concentration in the intermediate soil layer did not increase under improved irrigation and even decreased under traditional irrigation. During this period, lack of irrigation, low rainfall, and high rates of evaporation precluded drainage measurement and soil-solution sampling from the uppermost layer (0.4 m). In this case, the transference of salts seemed to have been from the soil solution to the solid phase (Van Genunchten and Wierenga, 1976; Gaudet et al., 1977). Precipitation of calcite and gypsum may have occurred during periods of lower soil moisture, regulating the concentration of the main ions in the soil solution, although the presence of chloride may have had some effect on the effective activity coefficient of sulfate (Pizarro, 1985; Sposito, 1989).
Irrigation management, however, influenced the actual salt loads in soil layers during periods of lower soil moisture. Salt loads in the intermediate soil layer were 0.4 kg TDS m-3 of soil under traditional irrigation and 0.6 kg TDS m-3 of soil under improved irrigation. This is a consequence of differences between the previous drainage episodes in traditional (146 mm) and improved (42 mm) irrigation schedules (Román et al., 1999). This observation suggests that irrigation management alters the kinetics of salt precipitation.
Salt Leaching Patterns
The total amount of dissolved salt in the soil profile and the soil moisture were closely related. Rain water elevated the quantity of dissolved salts. When water input increased and soil moisture exceeded field capacity, drainage and salt leaching began. A lag-time between salt loading in the soil profile and salt discharge was observed. This pattern was similar for both irrigation treatments (Fig. 3 and 4). Water input as irrigation increased salt loads in the soil profile. Its effect on salt leaching, however, was lower than rain water's, due to almost no irrigation water being redistributed to drainage in improved irrigated plots.
Autumn rainfall at the end of 1993 (around 320 d) dissolved salts in the soil profile, inducing increased salt loads. This trend was more apparent under the improved than under the traditional treatment, since in the latter a previous drainage and salt discharge process had occurred.
When drainage was negligible during the period of decreasing soil moisture, the salt load in the soil profile decreased. Since salt discharge was precluded in the absence of drainage, the precipitation of low-solubility salts (calcite and gypsum) out of the soil solution seems apparent. Steady salt concentrations were maintained in the soil solution despite changes in the soil moisture during this period. Most drainage occurred during the 1995 to 1996 (1020–1150 d) and 1996 to 1997 (1390–1550 d) autumn–winter rainy seasons. Again, a diluting effect from rainfall occurred, as indicated by lower solution EC in the 0.5-m soil layer (Fig. 2). Salt load increased in the soil solution, followed by a sharp decrease, and a corresponding salt discharge (Fig. 3 and 4).
Net Salt Variations and Leaching Requirements
The net salt gains or losses were calculated for each cultivated and inter-cultivated period. Since seasonal EC of the soil solution in the deepest (1.4 m) soil layer and drainage were unrelated, salt discharge was calculated by integrating the product of salt concentration and drainage discharge by corresponding sampling times. Salt concentration of the irrigation water varied much less, and thus a mean value was taken to estimate salt input from irrigation. Although some ions (Ca2+, Mg2+, and K+) are of nutritive value for corn and winter cereals, the total uptake by crops is negligible relative to the quantities measured in the soil profile (Bustos et al., 1996). In other respects, uptake of K+ had little impact relative to the net salt losses, and applications of NO-3 as fertilizer were offset by plant uptake.
In plots under the traditional irrigation schedule, net salt losses were observed during most periods, except for the 1995 corn growing season, in which low drainage from irrigation water was observed (Fig. 5). A net salt loss of 1 kg m-2 was estimated throughout the experiment (Table 3). In these plots, salt leaching was evenly distributed between discharge attributable to irrigation water surplus (45%), and discharge attributable to periods of rainfall (55%).
|
|
The proportional model (Jensen et al., 1990) assumes that irrigation water progressively increases the salt concentration of the soil solution as it percolates to deeper soil layers. Our results showed, however, that probably due to precipitation of gypsum, the salt load of progressively deeper soil layers did not tend to increase. The theoretical model does not seem to account for other factors governing the variation in solution EC (precipitation–dissolution reactions). The model also assumes, when applied to estimating the LF, that the ratio of EC (irrigation to drainage) and the ratio of water (drainage discharge to irrigation input) are similar under long-term steady-state conditions (SIC, 1975). On the basis of these assumptions, LFs of 22% for traditional and 18% for improved treatments would be needed to reach a proportion.
This theoretical model does not account for the drainage salt load due to heavy rainfall since this violates steady-state conditions. When LF was estimated by the proportional model but included rainfall (EC = 0.014 S m-1) and irrigation water (EC = 0.12 S m-1) as inputs, and EC was adjusted to the proportion of both inputs, an LF of 11% for the traditional irrigation schedule and 8% for the improved irrigation schedule were found. The experimental ratios of drainage to total water input for the whole experimental period were 11 and 7%, with a net loss of salts under both irrigation treatments. With an actual LF of about 10%, soil-solution drainage EC should approach 1.2 S m-1 in the 1.4-m soil layer, according to the proportional model. This theoretical value is much higher than our mean 0.62 S m-1, illustrating the impact rainfall can have on salt leaching. The lower EC in the 1.4-m soil layer (relative to the proportional model) suggests a lower rate of salt discharge (about half). Precipitation of salts onto soil surfaces (apparently not considered in the proportional model) could account for lower drainage EC. Clearly, rainfall affects the water flux and salt discharge but direct effects of rainfall on the EC of the soil solution at the 1.4-m soil layer are not clear.
These results, however, demonstrated the leaching potential of rain water, and the possibility of reducing the proportion of irrigation water distributed to drainage that would otherwise be a requirement for salt leaching.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The salt-leaching process was preceded by a period of increasing salt loads in the soil layers, due to transference of salt from the solid phase. During periods of decreasing soil moisture and negligible drainage, the opposite trend was observed. Net salt losses or gains cannot be estimated only by determining salt concentration in the aqueous phase, and consequently, complementary measures of soil moisture and water movement are required.
Salinity hazard to crops appears to be a function of EC. If EC is governed by salt precipitation–dissolution kinetics, then knowledge of soil moisture and water movement will not provide the answer. Rather, knowledge of factors controlling solubility (i.e., the Activity Ratio diagram) should guide.
Leaching requirements, estimated by assuming a proportional model of EC (irrigation to drainage) and water (drainage discharge to irrigation input), exceeded the actual values because they did not take into account the leaching potential of rainfall. Under traditional and improved irrigation schedules, respectively, 55 and 88% of salt discharge occurred during drainage periods that were induced by heavy rainfall. The experimental results thus demonstrate the discharge potential of rain water, and the possibility of minimizing the LF, provided that drainage salt load induced by heavy rainfall occurred in a long-term period.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 11, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. M. Dudley, A. Ben-Gal, and N. Lazarovitch Drainage Water Reuse: Biological, Physical, and Technological Considerations for System Management J. Environ. Qual., September 2, 2008; 37(5_Supplement): S-25 - S-35. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
