Research Article
Brackish Water for Irrigation: III Effects on Yields of Crops in Wheat-Rice Rotation and Properties of the Bhalike Soil Series
Not Available
Anwar-ul-Hassan
Not Available
Abdul Ghafoor
Not Available
M.M. Iqbal
Not Available
Irrigation is becoming a practical means to use groundwater in the arid and semi-arid regions due to insufficient rainfall. Most of the ground waters occurring in these areas are of poor quality as they contain variable amounts of sodium and bicarbonate ions (Malik et al., 1984; Ayres and Westcot, 1985). The effects of soluble salts from irrigation water on soil properties have been investigated intensively. Most studies indicate that calcium (Ca+2) salts generally improve soil physical properties by flocculating soil particles (Rengasamy, 1984), whereas sodium (Na) salts cause deterioration of soil physical properties because of its dispersive effects (Shainberg and Letey, 1983). The dispersion of soil particles has been reported to reduce hydraulic conductivity (Quirk and Schofield, 1955) and increase bulk density (Waldron et al., 1979). Increase in saturated paste extract electrical conductivity (ECe) and sodium adsorption ration (SAR) were commonly reported in soil profile irrigated with brackish water (Sing et al., 1992), which in turn reduced the crop yield (Bajwa and Josan, 1989).
Effects of brackish water irrigation on crops have been reported (Bajwa et al., 1983; Bhatti et al., 1981; Francois et al., 1986). However, limited work has been done to predict the rate at which yields of wheat (Triticum aestivum L.) and rice (Oryza sativa L.) started declining over the years due to long-term irrigation with brackish irrigation water on yields of crops and physico-chemical properties of the soils. Most of the research work has been done previously on crops and soils under disturbed soil conditions by using limited number of combinations of ECiw, SARiw and/or RSC. However, little emphasis has been placed on yields of crops under undisturbed soil conditions with brackish irrigation waters. Such information is necessary for efficient management of poor quality waters so that a high agricultural production is maintained.
The present long-term experiment was carried out under both the disturbed and undisturbed conditions to investigate the possibility of predicting ECiw, SARiw and/or RSC effects on salinization, sodication, bulk density, saturated hydraulic conductivity and yields of wheat and rice crops in wheat-rice cycle.
The present research work was conducted in a net-house, University of Agriculture, Faisalabad during 1991-95. Bhalike soil series (Coarse silty, mixed, hyperthermic Mollic Epiaquepts) was sampled during September-October 1991. Physico-chemical properties of this soil series were: sand 33.49; silt 38.70%; clay 27.82% (clay loam); pHs 7.65; ECe 2.41 dS m-1; CO3 2.00 mmo L-1; HCO3 9.35 mmo L-1; Cl 6.28 mmo L-1; SO4 5.89 mmo L-1; Ca + Mg 10.38 mmol L-1; Na 13.19 mmol L-1; SAR 5.80 (mmol L-1)1/2; CaCO3 6.49% ; CEC 11.18 cmolc kg-1.
Soil sampling and columns preparation: Metallic cylinders (76 cm long and 30 cm diameter) were used to collect the undisturbed soil samples. A piece of wood (35 cm x 35 cm and 8 cm thick) having circular groove that fitted snugly on the upper edge of the cylinder was placed on the top. Cylinder were pushed vertically into the moist soil (at 50 % field capacity) by dropping a 20 kg weight on the grooved wooden planks, tied with a strong string and controlled through a pulley, attached to a tripod. When cylinder was inserted up to 68 cm depth, the soil around the cylinder was excavated up to 80 cm and soil columns were removed by titling it. This excavated soil was used for preparing the disturbed soil columns. The extra soil at the bottom of the cylinder was removed with the help of a sharp knife. This procedure was repeated for 20 cylinders. A thin layer of glass wool and sand on stainless steel screen (35 cm x 35 cm) was placed and was attached at the bottom of the cylinders with the help of a rubber inner tube band. These cylinders were placed on metallic funnels, fixed on iron stands and leveled. The main objective of glass wool and sand was to minimize the movement of finer particles in the leachate.
For the preparation of disturbed soil columns, stainless steel wire gauze (35 cm x 35 cm) was fixed at the bottom of the empty cylinders with the help of a rubber inner tube band. A thin layer of glass wool and sand were spreaded on the wire gauze before attaching it with the cylinder. These cylinders were placed on metallic funnels and fixed on leveled iron stands. The cylinders were filled with air-dried, ground and sieved (2 mm) soil. The soil filling was accomplished by first packing 1/3rd of the cylinder, by adding small increments through a plastic funnel attached to a plastic pipe and gently tapping the sides of the column followed by settling of soil with canal water. This was used to fill the soil until 68 cm level was reached.
Irrigation water quality: Fourteen design points/treatment combinations having different ECiw, SARiw and RSC levels were selected following Central Composite Rotatable Second Order Design (Cochran and Cox, 1957). The beauty of this design is that prediction can be made for 125 treatment combinations by using only fifteen of them. Five levels each of ECiw (X1), SARiw (X2) and RSC (X3) were 0.65, 2.00, 4.00, 6.00 and 7.35 dS m-1, 3.95, 9.65, 18.00, 26.35 and 32.04 (mmol L-1)1/2 and 0.65, 2.00, 4.00, 6.00 and 7.35 mmolc L-1, respectively. The levels were coded as 1.682, -1, 0, 1 and 1.682, respectively, for each variable. The relationship between coded levels and actual levels for ECiw, SARiw and RSC is given at the foot of Table 1. The central point (all variables at coded zero levels) was repeated six times. The choice of six central points was made so that a uniform precision design could be attained. In a uniform precision design, variance of y at the origin is equal to the variance of y at unit distance from its origin. Thus design gives much more protection against bias in the regression analysis (Montgomery, 1997). The design matrix and treatment combinations investigated are presented in Table 1.
To verify the validity of model predictions with factors of Table 1, five extra treatments (Table 2) for wheat were run in the disturbed columns of Bhalike soil series. The procedure for preparation of disturbed soil columns was same as mentioned above. After getting near chemical steady state, assessed on the basis of ECdw (EC of drainage water) wheat was grown in these extra lysimeters. These five treatments were selected without any consideration of the 20 treatments (Table 1).
Table 1: | Design matrix and treatment combinations used during experiments |
(1) |
(2) |
(3) |
Table 2: | Five extra treatment combinations run to test the model validity |
Preparation of brackish water, application and steady-state soil conditions: The desired levels of ECiw, SARiw and RSC (Table 1) were prepared by dissolving NaCl, NaHCO3, Na2SO4, CaCl2 and MgSO4 salts in canal water [EC 0.35 dS m-1; Ca+Mg 2.44 mmol L-1; Na 1.06 mmol L-1; SAR 0.94 (mmol L-1)1/2]. For every irrigation, calculated amounts of these salts were dissolved and applied to the respective soil columns. After each irrigation, drainage water from lysimeters was measured and analyzed occasionally for ECdw (EC of drainage water), cations and anions to monitor the progress towards steady state. Application of brackish water was started on February 15, 1992 and the near chemical steady-state soil conditions were achieved on November 23, 1992 (Table 3).
Table 3: | Chemical analysis of drainage water (EC and SAR values) collected at steady-state soil conditions |
Eciw and ECdw = dS m-1 SARiw = (mmol L-1)1/2 RSC = mmolc L-1 |
Crops: Wheat (Triticum aestivum L.) variety Faisalabad-85 and rice (Oryza sativa L.) variety KS-282 were chosen and were grown in the wheat-rice rotation, commonly followed in the region. Wheat crop was sown on December 16, 13 and 12, 1992, 1993 and 1994 in both the disturbed and undisturbed soil columns at a density of 15 grains soil-1 columns. This number was reduced to three column-1 10-days after germination. The N, P and K were applied at 150, 100 and 75 kg ha-1, respectively, to all the soil columns as urea, single super phosphate (SSP) and Sulphate of potash (K2SO4). During growth period, the crop was sprayed with Novacron to protect it from insect attack. Brackish waters (Table 1) were applied through out the growth period of the crop. The crop was harvested on May 2, April 26 and April 24, 1993, 1994 and 1995, respectively. Forty days old 2-3 rice seedlings per hill were transplanted (three hills per lysimeter) on July 18, 1993 and July 10, 1994 in both the disturbed and undisturbed soil columns. The N, P and K were applied at 100, 80 and 50 kg ha-1 as urea, SSP and K2SO4, respectively. The zinc was also applied as zinc sulphate (ZnSO4) at 12 kg ha-1. The soil columns were kept submerged through out rice growth period with designed waters. The crop was harvested on November 2, 1993 and November 4, 1994 and paddy yield was recorded. Average paddy yield of three plants per pot was used for statistical analysis. The total amount of water (brackish water + rainfall) added to soil columns is presented in Table 4. After termination of the experiments, saturated hydraulic conductivity (Ks) with falling head method (Jury, 1991) and bulk density by core method from 0-10 cm (Blake and Hartge, 1986) were determined.
Table 4: | Total depth of irrigation and rainfall values at experimental site during the period of February 1992 to April 1995 |
Table 5: | Observed and predicted grain yield (g) of wheat as affected with ECiw,SARiw and RSC |
The soil samples from 0-15, 15-30, 30-45 and 45-60 cm were drawn from the undisturbed and disturbed columns. The soil samples were air-dried, ground and passed through a 2 mm sieve and were analysed for ECe, TSS, Ca+2, Mg+2, Na+, CO32-, HCO3-, SO42-, SAR and pHs (U.S. Salinity Lab. Staff, 1954).
Data analysis: The coefficients in Table 4 were determined using multiple regression analysis. This was accomplished by using computer software Minitab version 7.1. To draw quadratic graph for all dependent variables, following form of the model was followed:
log yi = β0 + βIxi + βiixi2 |
To show the effect of independent variable on a dependent variable in a quadratic graph, the other two variables were kept at coded 0 levels. The actual values of independent variables could be transformed from the coded values by equations given at the foot of Table 1.
The observed and model predicted grain yield data of wheat are presented in Table 6. Results indicated that the range of difference between observed and model predicted yield was 1.93 to 3.75 g for the undisturbed and disturbed soil columns. The difference between observed and predicted values were not large, hence model fitted the data adequately. In a pot study conducted at International Rice Research Institute (IRRI), Los Banos, Philippines, Rashid (1983) reported 1.0 to 5.0 g difference between observed and model predicted values for paddy yield. He further reported that model fitted the data adequately with this range of difference between observed and predicted values.
Soil Salinity (Ece): Salt accumulation in soil profile was determined at the end of the experiments. The electrical conductivity of the saturation extract (ECe) before the start of brackish water irrigation was 2.41. It increased during the experiments. This has been shown through the best-fit quadratic relationships between ECe and ECiw, SARiw and RSC (log ECe = bo + b1x1 + b2x2 + b3x3 + b 11x12 + b 22 x22 + b33x32 + b 12x1x2 + b13x1x3 + b23x2x3, where x1, x2 and x3 are the ECiw, SARiw and RSC and b0, b1, b2, b3, b11, b22, b33, b12, b13 and b23 are the regression coefficients) in Table 6. The values of coefficients of determination (R2) were highly significant and the predicted ECe, SAR, yield of crops, etc. were fairly close to the observed values of these responses.
At given SARiw and RSC, increase in ECiw tended to cause an increase in ECe in both the undisturbed and disturbed soil columns. Due to space limitation, graphs with changing levels of one parameter on all responses at coded 0 levels of other two parameters were depicted here. It is evident that soil salinity increased from 2.49 to 6.55 and 2.36 to 6.53 in the top 0-15 cm with increasing ECiw from 0.65 to 7.35 dS m-1 for both the undisturbed and disturbed soils. It is interesting to note that increase in ECe with ECiw was higher at coded -1.682 and 1 levels of SARiw and RSC than at higher coded 0, 1 and 1.682 levels of these parameters in the undisturbed and disturbed soil conditions. The difference in ECe with ECiw at given SAR and RSC waters was quite narrow between undisturbed and disturbed soil columns (Fig. 2). Results indicated that the ECe with ECiw 0.65, 2.0, 4.0, 6.0 and 7.35 d S m-1 was 2.49, 3.26, 4.53, 5.79 and 6.55 which was 12.31, 42.5, 53.0, 56.33 and 56.34 % of the EC of respective water used for irrigation for the undisturbed soil. The increase in ECe with ECiw 0.65 dS m-1was less than the original ECe (2.41 dS m-1) in the disturbed soil at coded -1.682, -1 and 0 levels of SARiw and RSC. Similarly, the ECe with ECiw 2.0, 4.0, 6.0 and 7.35 dS m-1 was 3.12, 4.39, 5.71 and 6.53 dS m-1 which was 35.5, 49.5, 55.0 and 56.05 % of the EC of respective water at coded 0 levels of SARiw and RSC [(18.0 mmol L-1) 1/2 and 4.0 mmolc L-1]. Higher the salinity of water, higher saline is the soil. Saleem et al. (1993) also recorded similar results and reported that ECe of clay loam soil increased from 2.88 to 4.0 dS m-1 after four years application of water having ECiw 2.4 dS m-1, SARiw 9.2 and RSC 5.7 mmolc L-1. It means that salts continued to accumulate resulting in increased ECe in the soil profile. It is evident that ECe was < 4.0 dS m-1 with ECiw up to 2.0 dS m-1 at coded -1.682 and 1 levels of SARiw and RSC for both the soil conditions.
Table 6: | Regression coefficients (b) and coefficient of determination (R2) for wheat and paddy yields and soil properties as affected with ECiw, SARiw and RSC (log values) |
* | = Significant at 0.01 level of probability |
** | = Significant at 0.05 level of probability |
ns | = Non-significant |
BD | = Bulk density (Mg cm-3) |
Ks | = Saturated hydraulic conductivity (cm h-1) |
(und.) | = Undisturbed |
Fig. 1: | Average annaul rainfall data (mm) from October 1992 to April 1995 |
It was >4.0 dS m-1 with ECiw from 4.0 to 7.35 dS m-1 at coded 0, 1 and 1.682 levels of SARiw and RSC, which was the upper limit for saline-sodic and sodic soils as ascribed by U.S. Salinity Lab. Staff (1954) and Ayres and Westcot (1985). It is also evident that salinity increased with depth in both the soil conditions. This might be due to leaching of solutes to lower depths. Medium textured soil exhibits better drainage than that of fine textured soils. Hussain et al., (1993) and Hamdy et al. (1993) also reported that ECe of normal soils increased with ECiw.
Soil salinity increased significantly with an increase in SARiw at a given salinity and RSC of waters. Results (Table 6) revealed that ECe increased with changing levels of SARiw only at lower coded values of ECiw and RSC, whereas it decreased and/or remained constant with ECiw at higher coded values of these two parameters. For instance, the ECe increased with SARiw only up to 26.35, thereafter it remained almost constant with SARiw 32.04 at coded 0 levels of ECiw and RSC for both the undisturbed and disturbed soils.
Fig. 2: | Response of Ece to Eciw, SARiw and RSC |
It is clear from the data (Table 6) that an increase in the ECe with SARiw from 3.95 to 32.04 could be due to the reason that high SAR saline water deteriorated the soil structure and thereby reduced permeability of soils. At a given ECiw and SARiw, the RSC of water tended to cause decrease in ECe (Fig. 2). Both the soil conditions behaved similarly regarding the effect of RSC on soil salinity at coded 0 levels of ECiw and SARiw.
Fig. 3: | Response of SAR to Eciw, SARiw and RSC |
Fig. 4: | Depth-wise Ece (dS m-1) as affected with Eciw at coded 0" levels of SARiw and RSC |
The ECe with RSC 0.65 to 7.35 dS m-1 was still higher than the original salinity of the soil (2.41 dS m-1). In several studies, it was shown that HCO3 content of water decreased soil salinity through precipitation of Ca+2 and Mg+2 as CaCO3 and MgSIO2 (Hausenbuiller et al., 1960; Muhammed and Rauf, 1983).
Soil sodication (SAR): Fig. 3 shows that, in the top 0-15 cm, there was an appreciable increase in SARiw with ECiw at coded 0 levels of SARiw and RSC (18 and 4.0 mmolc L-1). Data in Table 6 revealed that sodicity build up in both the soil conditions with ECiw was less at coded -1.682 and 1 than that at higher coded 0, 1 and 1.682 levels of SARiw and RSC.
Fig. 5: | Depth-wise soil SAR as affected with SARiw at coded levels of SARiw and RSC |
Fig. 6: | Response of bulk density to ECiw, SARiw and RSC |
It is evident that undisturbed and disturbed soils behaved similarly in sodicity build-up with ECiw at al the five coded levels of SARiw and RSC. It is interesting to note that at similar SARiw and RSC, irrigation with high salinity water (ECiw 7.35 dS m-1) caused high accumulation of salts and also increased the SAR of soil solution. Singh et al. (1992), Bajwa and Josan (1989) and Hussain et al. (1993) recorded similar results and reported that sodicity of normal soils increased with Eciw.
Fig. 7: | Response of saturated hydraulic conductivity (Ks) to ECiw, SARiw and RSC |
At a given ECiw and RSC, the increasing SARiw tended to cause an increase in SAR of the both the soil contiontons. This increase, however, was more pronounced with similar SARiw at higher coded 1 and 1.682 levels of ECiw and RSC under both the soils. Results indicated that at coded 0 levels of ECiw and RSC, soil SAR with SARiw 3.95, 9.65, 18.00, 26.35 and 32.04 was 6.47, 11.85, 21.98, 26.61 and 30.22; 6.07, 11.30, 21.12, 28.08 and 28.06, respectively, which was 16.96, 62.69, 89.89, 90.36 and 76.22 %; 6.84, 56.99, 85.11, 84.55 and 69.48 % of the SAR of respective water used for irrigation for undisturbed and disturbed soil columns. Thus it could be seen that more sodicity was build up with similar SARiw in the undisturbed than that in the disturbed soils. Interestingly, build up of sodicity was more with SARiw 32.04) at coded 1.682 levels of ECiw and RSC (7.35 dS m-1 and 7.35 mmolc L-1) than that at coded -1.682 levels of ECiw and RSC (0.65 dS m-1 and 0.65 mmolc L-1) for both the soil conditions. The build up of soil SAR seemed to be higher than expected with SARiw of 32.04 at higher than that at lower coded levels of ECiw and RSC. This could be due to the reason that water had quite high residual alkalinity and soluble salts in addition to high SARiw. Therefore, precipitation of carbonates resulted in higher Na saturation of the soil.
Fig. 8: | Yields of designed crops to Eciw, SARiw and RSC |
The RSC waters significantly increased SAR of soil solution at given levels of ECiw and SARiw. A minor difference in SAR with RSC waters at coded 0 levels of ECiw and SARiw (4.0 dS m-1 and 18.0) was noted between undisturbed and disturbed soils (Fig. 3). More SAR was resulted with RSC waters at higher than that at lower coded levels of ECiw and SARiw (Table 6). The increase in SAR with RSC waters could be due to the reason that at high RSC waters, bicarbonate ions precipitated Ca+2 ions and the residue of the CO3 pairs with Na to form Na2CO3 and NaHCO3 in the irrigated soils, thereby sodicating the soil more (Gupta, 1980).
The depth-wise distribution of ECe and SAR, after the termination of experiments is presented in Fig. 4 and 5. The tendency of soils irrigated with higher ECiw at given SARiw and RSC; SARiw at a given ECiw and RSC for build-up of more ECe and SAR is evident. For instance, even the lower soil layers accumulated more solutes than that of the upper with ECiw, SARiw and/or RSC waters. Singh et al. (1992) recorded the similar results and reported that salts were accumulated more in the lower layers of sandy loam soil with high ECiw (12 dS m-1 and SAR 40) water. The higher accumulation of soluble salts and sodium in the soil profile was due to the high amount of electrolytes, exchangeable Na and carbonates added through a large number of irrigations (Table 4). After four and a half years application of designed brackish waters, whole of the soil profile attained ECe >4.0 dS m-1 and SAR >13.3, which is the upper limit for the saline-sodic and sodic soils as ascribed by U.S. Salinity Lab Staff (1954).
Bulk density (BD): The effect of ECiw on bulk density (BD) was negative and significantly, indicating that BD decreased with increasing ECiw levels at coded 0 levels of SARiw and RSC (Fig. 6). The BD increased with ECiw up to 2 dS m-1, thereafter it decreased with further increase in ECiw from 4.0 to 7.35 dS m-1 at higher (1 and 1.682) than that at lower (-1.682 and 1) coded levels of SARiw and RSC in the undisturbed soil. The reduction in BD was 1.24, 0.25, 1.86 and 4.97 %; 0.63, 0.19, 0.89 and 4.43 %, respectively, with ECiw 2.0, 4.0, 6.0 and 7.35 dS m-1 over 0.64 dS m-1 at coded 0 levels of SARiw and RSC for the undisturbed and disturbed soil columns (Table 6). Coasta et al. (1991) recorded the similar results and reported that for 0-15 cm, the BD decreased from 0.06 to 0.04 Mg m-3 with water having EC 2.98 dS m-1 and SAR 8.0. However, Zartman and Gichuru (1984) reported that Na accumulation with ECiw 12 dS m-1 and SARiw 11.0 did not affect the BD of soils.
The BD decreased with SARiw and/or RSC at given coded levels of ECiw (Table 6). It is interesting to note that BD increased with SARiw up to 26.35, became flattened with further increase in water sodicity for both the soil conditions. The increase in BD, however, was more pronounced with high SARiw and/or RSC at higher than that at lower coded levels of ECiw and RSC. It is evident that at coded 0 levels of ECiw and RSC, an increase in the BD was 8.69, 15.94, 18.40 and 18.12 %; 6.43, 12.86, 15.00 and 14.29 % with SARiw 9.65, 18.0, 26.35 and 32.04 over SARiw 3.95, respectively, for the undisturbed and disturbed soil. Bulk density increased linearly with RSC waters at coded 0 levels of ECiw and SARiw (4.0 dS m-1 and 18.0) for both the soil conditions (Table 6). However, it also increased with RSC up to 6.0 mmolc L-1, thereafter it flattened with further increase in RSC waters in the case of disturbed soil columns. The increase in BD with RSC waters was more pronounced at higher coded levels of ECiw and SARiw (i.e. 1 and 1.682). The increase in BD with RSC waters of 2.0, 4.0, 6.0 and 7.35 mmolc L-1 was 4.0, 8.0, 9.33 and 9.34 %; 4.08, 7.48, 9.52 and 9.53 % over RSC 0.64 mmolc L-1, respectively, for the undisturbed and disturbed soils. Precipitation of Ca+2 and Mg+2 ions and accumulation of Na ion on the exchange sites lead to decrease in the pore-space consequently increased in BD of soils. Shainberg and Letey (1983) have also reported the similar results.
Saturated hydraulic conductivity (Ks): Increasing electrolyte concentration in water at given SARiw and RSC generally tended to increase in Ks of both the soil conditions. At coded 0 levels of SARiw and RSC, more Ks values were registered for disturbed than that for the undisturbed soil columns (Fig. 6). High Ks was recorded with similar ECiw at lower than that at higher coded levels of SARiw and RSC (Table 6). For instance, the Ks with ECiw 7.35 dS m-1 was 0.354 and 0.369 cm h-1; 0.402 and 0.397 cm h-1 at coded -1.682 and 1 levels of SARiw and RSC. Similarly, it was 0.266 and 0.204; 0.276 and 0.217 cm h-1 with similar EC waters at coded 1 and 1.682 levels of SARiw and RSC. The increase in Ks was 0.048, 0.118, 0.172 and 0.191 cm h-1; 0.043, 0.109, 0.170 and 0.197 cm h-1, respectively, with ECiw 2.0, 4.0, 6.0 and 7.35 dS m-1 over ECiw 0.64 for the undisturbed and disturbed soils.
The SARiw and/or RSC waters at given levels of ECiw and RSC; ECiw and SARiw have resulted decrease in Ks of both the soil conditions. It is interesting to note that Ks of both the soil conditions increased with SARiw up to 9.65, decreased with further increase in SARiw from 18.0 to 32.04 (Fig. 6). Comparatively more Ks values were generated with similar SARiw for the disturbed than that for the undisturbed soil columns. Data (Table 6) revealed that more decrease in Ks was resulted with SARiw from 3.95 to 32.04 at lower than that at higher coded levels of ECiw and RSC. The reduction in Ks was 0.003, 0.036 and 0.069 cm h-1; 0.01, 0.042 and 0.071 cm h-1 with SARiw 18.0, 26.35 and 32.04, respectively, over SARiw 3.95 at coded 0 levels of ECiw and SARiw At coded 0 levels of ECiw and SARiw, the RSC waters up to 2.0 mmolc L-1 increased Ks, thereafter it decreased with RSC from 4.0 to 7.35 mmolc L-1 (Fig. 6). Higher Ks (0.208 cm h-1) was resulted with similar RSC waters (7.35 mmolc L-1) in the disturbed than that in the undisturbed (0.206 cm h-1) soils at coded 0 levels of ECiw and SARiw (i.e. ECiw 4.0 dS m-1 and SARiw 18.0). The decrease in Ks was 0.003, 0.009, 0.041 and 0.07 cm h-1; 0.003, 0.02, 0.05 and 0.076 cm h-1 with RSC 7.35 mmolc L-1 over RSC 0.64 mmolc L-1, respectively, for the undisturbed and disturbed soils at coded 0 levels of ECiw and SARiw (Table 6).
It is well documented that presence of salts in irrigation water tends to flocculate the soil particles while exchangeable Na tend to deflocculated, thereby reduced the Ks of the irrigated soils. High saline waters rarely produce dispersion or poor soil physical characteristics when used for irrigation. Despite the fact that most saline waters are very high in Na and relatively deficient in Ca+2 and to a lesser degree Mg+2, they almost always maintain a well flocculated soil conditions due to high electrolyte concentration. Data in Table 6 demonstrated that increase in Ks was even more with similar ECiw at lower than that at higher SARiw and RSC levels. For instance, the Ks with ECiw 7.35 dS m-1 was 0.354 and 0.369 cm h-1 at coded -1.682 and 1 (i.e. SARiw 3.95 and RSC 0.65 mmolc L-1; SARiw 9.65 and RSC 2.00 mmolc L-1). Similarly it was 0.266 and 0.204 cm h-1 with similar ECiw (7.35 dS m-1) at coded 1 and 1.682 (SARiw 26.35 and RSC 6.0; SAR 32.04 and RSC 7.35 mmolc L-1) levels of SARiw and RSC. Suarez and Lebron (1993) reported that at the same SARiw, the dispersion potential of low electrolyte water is greater than that for high electrolyte water. There was direct correlation between soil SAR and SARiw (Fig. 3), which resulted in deflocculating of soil and consequently Ks, was decreased. It might be possible that irrigation water having higher concentration of Na increased replacement of Ca+2 by Na on the exchange sites. The replacement of divalent (Ca+2) ion by the higher hydrated size monovalent (Na) ion could not neutralize net negative charge on soil colloids, which caused dispersion. This dispersion decreased the porosity of soil and as a result hydraulic conductivity decreased. The results showed that the waters with higher SAR and RSC significantly reduced Ks of both the soil conditions. Depressive effect of high bicarbonate water on Ks has been reported in previously (Muhammed and Rauf, 1983; Oster and Schorer, 1979). Dispersion and/or particle translocation into pores are considered important factors affecting the hydraulic conductivity of salt-affected soils (Rengasamy et al., 1984). According to U.S. Salinity Lab. Staff (1954), the Ks values less than 0.1 cm h-1 could create problem for leaching. In the present studies, all the five levels of SARiw and/or RSC have resulted more Ks than 0.1 cm h-1 at all the five coded levels of ECiw and RSC; ECiw and SARiw. The Ks started to decline beyond SARiw 9.65 and RSC 2.0 mmolc L-1 at given ECiw and RSC; ECiw and SARiw. Thus the SARiw up to 9.65 and RSC 2.0 mmolc L-1 were considered the safe levels for Ks of the clay loam soil under investigation.
Yields of crops: The yields obtained during the periods 1992-93, 1993-94 and 1994-95 for wheat and 1993 and 1994 for rice were pooled for statistical analysis and the average values are presented in Table 6. Compared with low levels of ECiw, SARiw and RSC (i.e. 0.64 dS m-1, 3.95 and 0.64 mmolc L-1), higher levels of these three water quality parameters significantly decreased yields of both the crops. Grain yield of wheat increased up to ECiw 4.0 dS m-1 at coded -1.682, -1 and 0 levels of SARiw and RSC. The increase and/or decrease in yield with ECiw has been depicted in Fig. 8 at coded 0 levels of SARiw and RSC. Thereafter, it decreased with further increase in ECiw from 4.0 to 7.35 mmoc L-1 for both the undisturbed and disturbed soils. The rate of increase in grain yield with ECiw up to coded 0 levels was more in the disturbed than that for the undisturbed soil conditions. Moreover, higher grain yield was predicted with the same ECiw from the disturbed than that in the undisturbed soil columns. The disturbed soil may have the advantage of good internal drainage than that of undisturbed (natural ones) soil where applied brackish water probably rapidly moved out of the root zone and thus resulted in comparatively low salinity and sodicity hazard. The saturated hydraulic conductivity data (Fig. 7) confirmed that disturbed soil columns exhibited more flux of water per unit time than that of undisturbed columns with similar ECiw at given coded levels of SARiw and RSC. Contrary to this, yield increased with ECiw only up to 2.0 dS m-1 at higher coded (i.e. 1 and 1.682) levels of SARiw and RSC, indicating that a water having EC >2 dS 1 becomes injurious to yield particularly at high SARiw (26.35 and 32.04) and RSC (6.0 and 7.35 mmolc L-1). An increase in wheat grain yield could be because of moderately salt tolerance nature of the crop (Maas and Holfman, 1977). Gupta and Yadav (1986) reported that wheat could be grown without any reduction in yield with ECiw 5 dS m-1 and with 10-25% reduction up to ECiw 10-12 dS m-1 in coarse textured soils for an average rainfall of the area (580 mm) after four years. This critical limit of ECiw for the same crop was narrow, i.e. 2.7 dS m-1 for 10% reduction and 7.4 dS m-1 for 25 % reduction in yield of wheat on fine textured soils.
Paddy yield decreased almost linearly in the case of undisturbed while increased for the disturbed soils with an increase in ECiw up to 0.65 to 2.0 dS m-1 at coded 0 levels of SARiw and RSC. The rate of decrease in paddy yield was more with ECiw in the undisturbed particularly at higher coded levels of SARiw and RSC than that in the disturbed soil conditions. For instance, the reduction in paddy yield with ECiw 7.35 dS m-1 was 73.96 and 46.96 %, respectively, as compared to ECiw 0.65 dS m-1 at coded 1.682 levels of SARiw and RSC (32.04 and 7.35 mmolc L-1), indicating that high EC was more injurious to paddy yield. However, an increase in paddy yield with ECiw up to 2.0 dS m-1 was noted at a given levels of SARiw and RSC. A decrease in paddy yield might be due to its sensitivity to water salinity. Ayres and Westcot (1985) reported a threshold value of ECiw 2.0 dS m-1 for rice. The data (Table 6) depicted that high EC water was even more injurious to paddy yield at coded 1 (SARiw 26.35 and RSC 6.0 mmolc L-1) and 1.682 (SARiw 32.04 and RSC 7.35 mmolc L-1) than at coded -1.682 (SARiw 3.95 and RSC 0.65 mmolc L-1) and -1 (SARiw 9.65 and RSC 2.0 mmolc L-1) levels of SARiw and RSC. Reduction in paddy yield with ECiw was also reported by Jamil (1972) and Girdhar (1988). Gupta and Yadav (1986) observed similar results and reported that paddy yield decreased by 50 % with ECiw 5 dS m-1.
Decrease in crops yield with high ECiw might be due to osmotic effect of salts in irrigation water (Greenway and Munns, 1980), antagonistic/synergistic effect of Na, Ca+2, Mg, CO3, HCO3, Cl and SO4-2 ions (Staple and Toenniessen, 1984) or specific ion toxicity (Ayres and Westcot, 1985). Addition of salts in irrigation water continuously keeps changing osmotic potential of soil solutions. This fluctuation in osmotic potential might adversely influence the availability of water, which is largely a function of the difference between the osmotic potential of plant root cell and the sum of the osmotic potential of the soil solution (van Hoorn et al., 1993). As a result of which plant cannot maintain turgor pressure (Arif, 1990; Wyn Jones, 1991), thus decreased crop yield.
It is evident (Fig. 8) that wheat yield increased with SARiw up to 9.65 in the undisturbed soil columns at coded 0 levels of ECiw and RSC. Contrary to this, yield increased with SARiw up to 18.0 in the disturbed soil at given levels of ECiw and RSC. The rate of yield reduction with higher SAR of irrigation water was more in the undisturbed than that in the disturbed soils. For instance, the reduction in wheat yield with SARiw (32.04) was 51.98 and 30.70 %, respectively, for the undisturbed and disturbed soils over SARiw 3.95. Data (Table 6) revealed that the effect of SARiw on yield was more pronounced at higher than that at lower coded levels of ECiw and RSC with the same SAR of irrigation water. Similar trend in paddy yield was noted with changing levels of SARiw at given levels of ECiw and RSC. The rice tolerated SARiw up to 26.35 and 18.0, respectively, at coded -1.682 and 1 levels of ECiw and RSC in the undisturbed and disturbed soil conditions. The effect of SARiw at given EC and RSC of waters was more pronounced at high levels than that at low levels on paddy yield. For instance, paddy yield was 31.85 and 27.11%, respectively, less with SARiw 32.04 over 3.95 under the disturbed and undisturbed conditions. It is interesting to note that the rate of decrease in paddy yield with SARiw up to coded 0 was more in disturbed than that in the undisturbed soils. Contrary to this, SARiw from coded 1 and 1.682 depressed more paddy yield in the undisturbed that that in the disturbed soil columns. Reduction in paddy yield with sodic water was reported by Haq (1979) and Hussain et al. (1993). They reported SARiw threshold of 10 for paddy yield. In lysimeter study on loamy clay, Jamil (1972) reported that SARiw of 16 was injurious to paddy yield and that with SARiw 20 did not prove useful even if amended with gypsum. The adverse effect of SARiw was even more severe on crops yield at high water salinity and RSC than that at low levels in the present studies. It might be due to poor structure and/or nutritional imbalance. Higher levels of SARiw increased exchangeable sodium percentage (ESP) and pH of the saturated soil paste (pHs) of soils and this situation probably resulted in nutritional imbalance and consequently decrease the crop yields (Khandewal and Lal, 1991). A decrease in yield may also be due to accumulation of exchangeable Na (Pearson, 1960), which may cause mechanical impedance to root penetration to poor soil structure prevailing in the root zone or sodium may directly toxic to wheat plant (Ayres and Westcot, 1985).
Like ECiw and SARiw, yield of wheat increased with RSC waters up to 4.0 mmolc L-1 at coded 0 levels of ECiw and SARiw (Fig. 8) under both the soil conditions. Thereafter, yield started declining with further increasing RSC waters. The rater of yield reduction was 58.02 and 36.05% in the disturbed and undisturbed soils with high RSC waters (i.e. 7.37 mmolc L-1) over low RSC waters (0.65 mmolc L-1). It was further noted that at higher coded 1 and 1.682 levels of ECiw and SARiw, increase in yield was up to RSC 2.0 mmolc L-1 for the disturbed soil columns. Paddy yield increase with RSC waters up to 2.0 and 4.0 mmolc L-1, respectively, in the undisturbed and disturbed soil conditions. The rate of reduction of paddy yield with high RSC waters (7.35 mmolc L-1) was more in the undisturbed (37.43 %) than that in the disturbed (20.45 %) over the low level of RSC waters (0.65 mmolc L-1) at coded 0 levels of ECiw and SARiw. Similar trend in increasing and/or decreasing in paddy yield with changing levels of RSC was registered at coded -1.682, -1, 1 and 1.682 levels of ECiw and SARiw (Table 6). Low yield of crops with high RSC waters at given ECiw and SARiw may be due to toxic effect of bicarbonate ions (Muhammed et al., 1977) and/or induced nutritional imbalance. The precipitation of HCO3 in soil as CaCO3 may also result in the accumulation of Na in soil (Bower et al., 1965). Excess of bicarbonate ions may also have adverse effect on nutrition of plants and tend to cause chlorosis (Miller, 1959). In India, however, water having SARadj up to 20 and RSC up to 10 mmolc L-1 have been used successfully on sandy loam soils under high rainfall (650-700 mm annum-1) conditions without reduction in economic yield of wheat (Gupta, 1980).