Hydraulic conductivity in response to exchangeable sodium percentage and solution salt concentration

Received: 04/25/2013; Approved: 02/26/2014. 1 Part of the first author Master ’s Dissertation. 2 Agronomist, Master of Science . Instituto Federal de Roraima, Campus Novo paraíso, BR-174, Km 512, Vila Novo Paraíso, s/n, 69360-000, Caracaraí, Roraima, Brasil. jefferson.paes@ifrr .edu.br (corresponding author). 3 Biochemist, Doctor of Science. Departamento de Solos, Universidade Federal de Viçosa, Campus Viçosa, Avenida Peter Henry Rolfs, s/n, 36570-000, Viçosa, Minas Gerais, Brasil. hruiz@ufv.br 4 Agronomist, Doctor of Science. Departamento de Solos, Universidade Federal de Viçosa, Campus Viçosa, Avenida Peter Henry Rolfs, s/n, 36570-000, Viçosa, Minas Gerais, Brasil. raphael@ufv .br 5 Agronomist, Doctor of Science. Departamento de Solos, Universidade Federal Rural de Pernambuco, Campus Recife, Avenida Dom Manoel de Medeiros, s/n, Dois Irmãos, 52171030, Recife, Pernambuco, Brasil. betânia@depa.ufrpe.br 6 Agronomist, Doctor of Science. Departamento de Solos, Universidade Federal Rural de Pernambuco, Campus Recife, Avenida Dom Manoel de Medeiros, s/n, Dois Irmãos, 52171030, Recife, Pernambuco, Brasil. fatima@depa.ufrpe.br 7 Agronomist, Doctor of Science. Departamento de Solos, Universidade Federal de Viçosa, Campus Viçosa, Avenida Peter Henry Rolfs, s/n, 36570-000, Viçosa, Minas Gerais, Brasil.genelicio.rocha@ufv .br Hydraulic conductivity in response to exchangeable sodium percentage and solution salt concentration


INTRODUCTION
In semiarid regions, salinization can negatively affect soil physical properties.One of the critical processes for soil structure is clay dispersion, which is often aggravated by the excess exchangeable sodium expressed by the exchangeable sodium percentage (ESP) (Spera et al., 2008;Paes et al., 2013).
In salt-affected soils, the solution salt concentration, usually expressed as electrical conductivity (EC), is very important for the phenomena of flocculation and dispersion because the action performed on the diffuse double layer (Gillman, 1974).
The soil hydraulic conductivity can be determined in the field or laboratory.The laboratory measurements are generally carried out using constant head permeameters.Klute & Dirksen (1986) emphasized that these analyzes provide approximate results and are recommended when the methodologies for evaluation in the field are impractical.
However, the results obtained in laboratory for saltaffected soils may differ greatly from those in the field.
The soil hydraulic conductivy can be underestimated in the lab mainly due to the dilution of the salt concentration in the soil solution, when distilled or deionized water as percolating solution.
Therefore, the use of percolating solution with EC similar to the satured doil paste extract could enhance the accuracy of hydraulic conductivity analyses in laboratory.This possible inaccuracy in current determinations of soil hydraulic conductivity may mistakenly indicate the need for irrigation using water with higher EC and/or chemical amendments to preserve soil structure, which tends to increase salinization and raise production costs.
The objective of this study was therefore to determine the soil saturated hydraulic conductivity using solutions of different EC in six representative soils of the State of Pernambuco, with exchangeable sodium percentage (ESP) adjusted in the range 5-30%.

MATERIAL AND METHODS
The study was carried out in the Department of Soil Science, Universidade Federal deViçosa, Viçosa, Minas Gerais, located in the geographic coordinates 20° 45' Sl e 42° 51' O.
The samples were air-dried, crushed, and passed through a 2-mm mesh sieve to obtain air-dry fine earth (ADFE).Part of these samples was separated for the assay and the other part was used for particle size analysis and chemical characterization of the soil studied (Tables 1, 2).
The treatments were arranged in a 6 x 3 x 3 factorial design, as follows: six soil samples, three exchangeable sodium percentages (ESP): 5; 15 and 30% and three electrical conductivities (EC) of the percolating solution: 0.0; 0.6 and 1.2 dS m -1 .The experiment was arranged in a randomized block design with five replicates.
First, to achieve the ESP values, the Na:Ca ratio on the cation exchange complex of the soils was adjusted using constant head permeameters (Ferreira, 1999).The permeameters were made with PVC columns, diameter of approximately 5 cm and 14 cm long, bottom lid adapted with drainage system and lined with filter paper.
Palavras-chave: solos salinos, solos salino-sódicos, percentagem de sódio trocável, condutividade elétrica da solução do solo, extrato da pasta de saturação.densities, to a height of 9 cm.A glass wool disk was placed at the top of the column to avoid disturbances at the surface by the solutions.
The sodium adsorption ratio (SAR) of the NaCl + CaCl 2 solutions needed to achieve the exchangeable sodium percentages (ESP) were previously determined using Eq.(1) (Pereira et al., 1982): (1) The solutions used for adjusting the Na:Ca ratio (saturating solutions) were prepared to reach a final concentration of 50 mmol c L -1 , using NaCl and CaCl 2 , Eq. ( 2): (2) This relatively high concentration allows a quick adjustment on the exchange complex.Thus, the [Na + ] and [Ca 2+ ] values were calculated by Eq. ( 3): (3) To carry out saturation in ascending order, the columns were placed in trays containing the saturating solution, according to the treatment, to a height corresponding to two thirds of the height of the soil sample (6 cm), reducing to the maximum the air in the soil pore space.The saturation occurred for a period of 48 h (Freire et al., 2003a).
The saturating solution was then applied, using constant head permeameters to reach the equilibrium, which was determined by the equal reading of the electrical conductivity between the effluent and the influent solutions.
After reaching the equilibrium of the different SAR values, using the NaCl + CaCl 2 solutions, the samples were removed from the permeameters, air-dried, crushed, and passed through a 2-mm mesh sieve to obtain ADFE.Subsequently, part of the sample was separated for the analysis and the other part was washed with isopropyl alcohol to remove the soluble salts and determination of ESP effectively achieved with the passage of the solutions (Table 3).
The ESP values observed were different from those expected.The differences observed were due to use of estimation by the equation established by Pereira et al. (1982), which was derived from to an average of all soils studied by these authors.
Before the determination of the hydraulic conductivity, the concentrated solutions (EC ~~ 5.0 dS m -1 )  were diluted with deionized water to reach the working values of 0.6 and 1.2 dS m -1 .The treatment 0 dS m -1 was carried out only with deionized water.In this article, the use of the term "working solution" is based on the consideration that distilled or deionized water can be considered as infinitely diluted solutions.
The procedures for determining the hydraulic conductivity followed the routine of Soil Physics Laboratory, Department of Soil Science-UFV, adapted from EMBRAPA (2011), since the hydraulic conductivity was determined in samples of ADFE.
To facilitate the visualization of changes in hydraulic conductivity with the treatments, we determined the relative hydraulic conductivity (Freire et al., 2003a) by assigning the unit to the maximum mean hydraulic conductivity of each soil.
Means of hydraulic conductivity and relative hydraulic conductivity were analyzed statistically.The analysis of variance grouped the soils, after testing the homogeneity of variances.The eight degrees of freedom for ESP of the soil and EC of the working solution were analyzed by orthogonal contrasts.In the presentation of the results, we used mean contrasts calculated by dividing the numerical value of the contrasts by the sum of the modules of the coefficients divided by two (Alvarez & Alvarez, 2006).However, in the discussion of these results, we indicated statistical difference for contrasts significant at 5% and the term "trend" for contrasts significant between 5 and 20%.

RESULTS AND DISCUSSION
The means of saturated hydraulic conductivity and relative hydraulic conductivity are shown in Tables 4  and 5.
The analysis of variance showed no homogeneity of variances in hydraulic conductivity for all soils studied and thus it was not possible to carry out the test of comparison of means.The non-homogeneity of variances can be attributed to the disparity of hydraulic conductivity values of some soils.
However, the standardization by calculating the relative hydraulic conductivity, produced homogeneity of variances between the soils and allowed the comparison of means by the Scott-Knott test (p <0.05) (Table 6).
There was greater mean relative hydraulic conductivity for the Argiluvic Chernosol.This result can be explained, at least partially, by the higher degree of weathering of this soil compared to the others in this study, with significant amount of iron oxides in the clay fraction (Freire et al., 2003b).The sample collection of the Argiluvic Chernosol was performed in the "Zona da Mata" region of Pernambuco, with higher rainfall rates than the northeastern semi-arid region, where the other soils were collected.It also has ∆pH closer to zero (Table 2).These characteristics are important for charge balance and, consequently, for particle flocculation.
Of the remaining five soils with higher proportion of active clay, i.e., clay CEC > 27 cmol c kg -1 (Table 2), the Albaqualf showed the highest relative hydraulic conductivity (Table 6).This result can be attributed, at least partially, to a greater proportion of coarse sand (Table 1), which, according to Santos & Ribeiro (2000), contributes to increase in macroporosity and, consequently, increase in hydraulic conductivity.
The remaining four soils, Haplic Luvisol, Fluvisol, Vertisol and Chromic Luvisol, showed no statistically significant differences in the relative hydraulic conductivity (Table 6).This lack of difference can be attributed to similarities in their particle size distribution (Table 1), remaining P values, ∆ pH (Table 2) and the mineralogical composition of the clay fraction (Freire et al., 2003b;Cunha, 2010), with predominance of highly active clay minerals (Table 2).
Because of the disparity observed in ESP values after adjustment (Table 3), we used in the discussion, the preestablished values of 5, 10 and 15%.When comparing results of hydraulic conductivity and relative hydraulic conductivity in response to treatments, we used the orthogonal contrasts in Table 7.As described in the section Material and Methods, the term "trend" will be used for significant differences in the range of 5 to 20%.
The significances of the contrasts, in relation to the results of the hydraulic conductivity and relative hydraulic conductivity, are similar.This similarity occurs because the calculation of the relative hydraulic conductivity is the result of the ratio between the hydraulic conductivity and a constant, corresponding to the maximum mean value of hydraulic conductivity for each soil.
The contrasts C1 and C2 evaluated the response of the soils to sodium saturation.There were differences for the Fluvisol, Haplic Luvisol, Albaqualf, Chromic Luvisol and Vertisol when comparing ESP at 5, 15 and 30%, with reduction in hydraulic conductivity and relative hydraulic conductivity with increasing exchangeable sodium content (Table 7).
The soils mentioned above showed a marked content of clay minerals type 2:1, such as illite and smectite (Freire et al., 2003b;Cunha, 2010).According to Rev. Ceres, Viçosa, v. 61, n.5, p. 715-722, set/out, 2014 The Argiluvic Chernosol showed no response to the different proportions of exchangeable sodium (Table 7), indicating that its physical properties, related to the water flow, remained preserved, even for a high ESP.This result, as already mentioned, is probably linked to the mineralogical composition and the degree of weathering of the soil, with a strong presence of hematite in the clay fraction (Freire et al., 2003b).
The contrasts C3 to C8 (Table 7) compare the EC of the working solutions for each soil and ESP; the odd numbers relate 1.2 dS m -1 with 0 and 0.6 dS m -1 , and the even numbers relate 0.6 dS m -1 with 0 dS m -1 .In general, there is an increase in hydraulic conductivity and, hence, in the relative hydraulic conductivity with increasing EC.
The analysis of the individual behavior of each soil in relation to the different working solutions, based on the significances of the contrasts C3 to C8 (Table 7), showed for the Fluvisol that, regardless of ESP values (5, 15 and 30%), differences occurred only when comparing the working solution with the highest salt concentration (EC = 1.2 dS m -1 ) with solutions of lower EC values (0 and 0.6 dS m -1 ).
The Fluvisol showed no significant differences in the comparison between the percolating solutions of lower salt concentration (0 vs. 0.6 dS m -1 ), suggesting that, when affected by sodicity problems (ESP> 5%), a value considered as critical by McIntyre (1979), it would be necessary to use irrigation water with EC greater or equal to 1.2 dS m -1 .This procedure, in field conditions, will probably preserve the structure of the soil and, hence, increase the hydraulic conductivity, decreasing the risk of reduction in permeability.Table 7 shows that the Haplic Luvisol responded in a similar way as the Fluvisol to the treatment with the lowest ESP, adjusted to 5%.For the other ESPs, there was a marked increase in hydraulic conductivity with the increase in salt concentration.
The Albaqualf responded to the increases in EC (Table 7).For all ESP values, there were significant differences or trends in the comparison of solutions with different EC values (0, 0.6, and 1.2 dS m -1 ) (C3-C8), showing increases in hydraulic conductivity and relative hydraulic conductivity with increasing salt concentration in the working solutions.
The Chromic Luvisol behaved in a similar way as the Albaqualf, considering the significance of the contrasts.However, at ESP of 30%, no differences were observed when comparing the solutions of low electrical conductivity (0 and 0.6 dS m -1 ) with the solution with EC = 1.2 dS m -1 (C7); whereas there was difference with the use of solutions with lower salt concentrations (0.6 vs 0 dS m -1 , C8).
This result suggests that, since there was complete pore obstruction, leading to null values of hydraulic conductivity when using only deionized water (Table 4), the solutions with EC at 1.2 dS m -1 allowed better soil water flow.However, these EC values (0.6 and 1.2 dS m -1 ) were not high enough to prevent the soil remain with values considered critical for hydraulic conductivity, that is, with hydraulic conductivity <0.1 cm h -1 (McIntyre, 1979) (Table 4).
The Vertisol showed responses to increases in solution salt concentrations up to ESP of 15%.There was no difference between the solutions (C7 and C8) by adjusting the soil to ESP equal to 30%.
From the foregoing, therefore, the EC range of working solutions used in this study corresponds to moderate or no risk of salinization caused by irrigation water (Ayers & Westcott, 1994).However, the ESP adjusted to 30% is more than twice the limit established by Richards (1954), which characterizes the soil as sodic.
This fact suggests that high EC values, higher than those indicated as moderate risk of salinity, should be maintained, at least in this soil, in order to increase the ionic strength of the solution, minimizing the dispersive effect of sodium at high ESP values.The Argiluvic Chernosol showed no marked responses to EC increase in the percolating solutions, at lower ESP values (Table 7).However, there is a trend of reduced hydraulic conductivity in C8, due to the high proportion of exchangeable sodium and the use of deionized water in the working solution.It is noteworthy that the relative hydraulic conductivity of this soil was the highest, considering the six soils studied (Table 6), which indicates less variation of hydraulic conductivity with the treatments.

Saturated hydraulic conductivity (cm h
As discussed before, the weather conditions, with a predominance of constant rain and high rainfall rates in the "Zona da Mata" region of Pernambuco, where this soil was collected, and the possible presence of stable microaggregates were probably the main reason for the less interference of sodium and EC of the working solution in the hydraulic conductivity .
It appears, therefore, that for less weathered soils with a predominance of high activity clay and with salinity problems, a more time consuming but appropriate analysis should take into account the EC of the saturation extract and the proportion of Na in the cation exchange complex.
In order to simplify the analysis, the EC of the saturation extract could be replaced by the EC of the extract 1:1 and double this value to use it in the working solution (EC ST ).The exchangeable cation values would be used to calculate the Na/(Ca + Mg + K) ratio, and this ratio would indicate the proportions for the preparation of the NaCl + CaCl 2 solutions, expressed as mmol c L -1 , which would be adjusted to the CE ST used to determine the hydraulic conductivity.

CONCLUSIONS
The increase in the exchangeable sodium percentage and decrease in the electrical conductivity of working solution are directly related to the reduction in the saturated hydraulic conductivity.
The response to the variation in the exchangeable sodium percentage and the variation in the electrical conductivity of the percolation solution was more pronounced in soils with higher proportions of active clays.
Measurements of saturated hydraulic conductivity in salt-affected soils must be carried out with solutions of electrical conductivity different from 0 dS m -1 , using values close to those of the saturated soil parte extract.

Table 3 .
Exchangeable sodium percentage (ESP) determined in soil samples after equilibration with NaCl and CaCl 2 solutions

Table 5 .
Saturated relative hydraulic conductivity (K 0R ), considering the exchangeable sodium percentage (ESP) and electrical conductivity of the solution (EC)

Table 6 .
General means of relative saturated hydraulic conductivity of soils under study compared by the Scott-Knott test

Table 7 .
Mean orthogonal contrasts (C) of saturated hydraulic conductivity and relative saturated hydraulic conductivity of soils under study, considering the exchangeable sodium percentage (ESP) and soil electrical conductivity (EC) of the working solution 1/