Improved method to determine particle size distribution for some gypsiferous soils . A case study from Al-Ahsa Governorate , Saudi Arabia

Received : 03.03.2016 Accepted : 05.05.2016 Until now, there is no method can be used to accurately assess the particles size distribution as well as textural classes of gypsiferous soils for proper interpretation of physical behavior of these soils, and most laboratory methods involve pretreatment to remove gypsum from the samples. Therefore, the results of the particle size distribution do not reflect the size distribution of the whole soil. This study aimed to develop an alternative method to determine particle size distribution for some gypsiferous soils selected from Al-Ahsa governorate, Saudi Arabia. Five samples from different profiles with different gypsum content were selected to evaluate the modified method. Sand fractions were separated with three disaggregation methods: 1) drying sieving, 2) shaking for 5 hours in a 7:3 ethanol: water solution, and 3) sonication for 3 minutes in a 7:3 ethanol: water solution. The statistical analysis results revealed that the sonication for 3 minutes in a 7:3 ethanol: water solution was the most effective method for separating sand fractions as compared to dry sieving and shaking. Meanwhile, there was slight difference in separating sand fractions between sonication for 3 minutes and shaking for 5 hours. The particle size distribution by the developed method showed increasing in total sand content as compared to standard particle size method. Likewise, comparison of the CEC/clay ratio between the two methods also indicated that the developed method yielded clay contents more consistent with other property data for the same horizons. Consequently, the textural classes obtained from the two methods were different. Therefore; we concluded that the determination of particle size distribution for gypsiferous soils (≤ 40% gypsum) using this developed method will improve the understanding and ability to proper interpret of physical behavior of these unique soils. We highly recommended using this developed method to separate soil particles from the gypsiferous soils.


Introduction
Gypsiferous soils (1 to 40% gypsum) are widely scattered throughout the Kingdom of Saudi Arabia and are particularly extensive in the Eastern Region from the Kuwait/Iraq border South to the sands of the Empty Quarter.The general characteristics of these soils they are; located on plains, shallow to moderately deep with loamy texture, highly saline and are mostly unsuitable for agricultural purposes (Vincent, 2008).
According to the World Reference Base for Soil Resources (FAO, 1990), the gypsiferous soils in Saudi Arabia cover about 82.5 km 2 (approximately 0.04 % of the total area).In contrast; the general soil map of the Kingdom of Saudi Arabia (Soil Atlas, 1984) estimated that the gypsiferous soils in Saudi Arabia cover around 1622.66 km 2 (approximately 8.5 % of total area).The reason for this significant variation is that the world map for gypsiferous soils distribution shows only area of Yermosols and Xerosols which have a gypsic or petrogypsic subsurface horizon, and the other soils which may have a gypsic horizon (e.g; Calcic units of Yermosols, Xerosols, Chernozems and Calcic Cambisols and Solonchaks) are not taken into account.
The estimation of soil texture of the gypsiferous soils under field conditions is misleading due to the presence of gypsum crystals in various sand-sized fractions.Consequently, the forms and degree of crystallization of gypsum particles influence the feel of the soil and as a result, field estimates of texture are generally coarser than indicated by laboratory methods (Soil Survey Staff, 2014b).
Until now, there is no method can be used to accurately assess the particles size distribution as well as textural classes of gypsiferous soils for proper interpretation of physical behavior of these soils (Pearson et al. 2014).However, most laboratory methods for determining particle size distribution of gypsiferous soils involve pretreatment to remove gypsum and more soluble salts as well as iron oxides (Soil Survey Staff, 2014b).Consequently, the particle size measurements by these methods indicate the size distribution of essentially insoluble, dominantly silicate minerals, and do not indicate the size distribution of the whole soil including gypsum (Porta, 1998).Pearson et al. (2014) proposed an alternative method to determine particle size distribution in gypseous soils (≥ 40% gypsum content).The method consists of two independent measures of particle size; 1) the amount and distribution of sand-sized particles including gypsum particles, and 2) clay content of a whole soil (including gypsum) basis using the measured gypsum content of the sample.
In this study, we further develop the Pearson et al. (2014) method to determine the particle size distribution in some gypsiferous soils (≤ 40% gypsum content) selected from Al-Ahsa governorate, Saudi Arabia.
This study tried to address the following questions: (1) which disaggregation methods are most suitable to separate sand fractions for gypsiferous soils?(2) Are there significant differences in particle size distribution as well as textural class obtained by standard method and the developed method for gypsiferous soils?(3) Is this modified method suitable to determine particle size distribution for gypsiferous soils?

Site description
The study area is located in the eastern part of Al-Ahsa governorate, Saudi Arabia (Figure 1).The study area falls within the arid climatic zone.The average annual rainfall varies from 0 to < 100 mm.The monthly mean maximum temperature of the hottest months (June and July) is 38˚C.Monthly mean minimum temperature of the coldest month (January) is less than 25˚C.The monthly mean relative humidity ranges between 40 to 60 % (April to August) and > 60% (January to March) and (September to December).All sampling locations were covered with natural vegetation especially Haloxylon sallicornicum and some area were cultivated with old date palms.Table 1 presents the geographical coordinates of the sampling sites.The calculated soil temperature regime is hyperthermic and soil moisture regime is aridic.Soil samples were collected from 5 representative profiles.All soil profiles were fully described (Appendix 1) according to the FAO Guideline for Soil profile Description (FAO, 2006).According to the USDA soil taxonomy (Soil Survey Staff, 2014a), the soils of the study area belong within Aridisols and classified as gypsids suborder.

Soil sampling and analysis
Five soil samples with different gypsum content were collected randomly from 5 profiles; profiles 1 and 2 were classified as Typic Haplogypsids; whereas, profiles 3, 4 and 5 were classified as Typic Calcigypsids.
Selected physical and chemical properties of the samples used for the evaluation of the disaggregation methods are given in Table 2. Texture class by feel was evaluated in the field for different horizons and compared to those obtained by using USDA textural triangle (Soil Survey Staff, 2014b).In laboratory, samples were air-dried (20-22 °C) and passed through a 2 mm mesh sieve to obtain the fine soil fraction.The bulk density (BD) was determined in the field by using core method according to (Grossman and Reinsch, 2002).Particle size distribution was determined by using standard pipet method according to (Gee and Bauder, 1994).Soil chemical properties were determined according to the standard method (Sparks et al. 1996).Soil pH was measured potentiometrically in 1:5 soil/water suspension.Electrical conductivity (EC) was measured potentiometrically on a saturated paste extract.Gypsum percentage was evaluated by dissolution in water, precipitation in acetone, re-dissolution, and conversion of solution EC to percent gypsum.Cation exchange capacity (CEC) was determined using continuous leaching of 4 g of soil with 100 ml of 1 M NH4OAc at pH 7 (4 g of air dry soil sample was treated three times with 1 M NaOAc, washed three times with ethanol (95 %), and then extracted three times with 1 M NH4OAc), and concentration of the exchangeable Na + was determined using flame photometer (model Corning 400), and then the CEC was calcluated using the equation described by (Sparks et al. 1996).Percent calcium carbonate (% CaCO3) was determined by calcimeter.The samples were treated with 0.1N HCL; the volume of CO2 from pure calcium carbonate and samples were recorded, and the % CaCO3 was then calculated according to Horváth et al. (2005).

Separation of sand fractions by the developed method
We used an alternative method described by Pearson et al. (2014) to separate sand fractions for five gypsiferous samples with gypsum content ranging from 4.2 to 10.8%.Sand fractions were separated using three disaggregation methods with fundamental changes including increasing the specimen weight and minimizing the time of the separation.Main reason for minimizing the time is that gypsum in most gypsiferous soils is present as fine sand fraction (125 -25 µm), which may be greatly affected by great energy input via sonication or shaking.As a result the fine sand may pass via sieve as silt fraction which causes a decrease in total sand content and an increase in silt content.The three disaggregation methods are as follows: (1) Dry sieving-25 g of air dried soil sample was sieved through a set of 53 µm diameter sieves (1.0, 0.5, 0.25, 0.125, and 0.053 mm openings) on a mechanical shaker for 3 minutes (shaking model HAVER EML 200 digital, Germany) and the fraction retained on each sieve was weighed.This method did not include addition of any chemical solutions to disperse samples; thus, it can be used to compare their obtained results with the other two methods for separating sand fractions from gypsiferous soils.(2) Shaking for 5 hours in a 7:3 ethanol: water solution-25 g of soil suspended in a 7:3 ethanol: water solution and shaken for 5 hours on a reciprocating shaker Model (SSL1, orbital shaker).The suspension was then passed through a 53µm-sieve, and sand fraction on each sieve weighted (wet weight), and left overnight to oven dry at 60°C (in order to avoid the change in gypsum contents as well as clay minerals, and which may occur if the samples have been dried at 105 °C) and then each fraction weighed for a second time (dry weight).( 3) Sonication in a 7:3 ethanol: water solution-25 g of soil was suspended in 50 ml of a 7:3 ethanol: water solution and allowed to sit for 2 minutes in a 150 mm diameter beaker.Ultrasound energy was applied for 3 minutes via a sonicator Model (Sonics & Materials Vibra-Cell VC600a, 600-Watt at 20 kHz with 20 mm horn).The suspension was then passed through a 53µm-sieve, and sand fraction on each sieve weighted (wet weight), and left overnight to oven dry at 60°C and then each fraction weighed for a second time (dry weight).Individual sand separates in the three methods were calculated as a percentage of the whole soil sample.Total sand was calculated as the sum of the sand separates.

Whole clay content measurement
The clay content was measured after gypsum removal by the standard pipet method, and then re-calculated to a whole soil (including gypsum) using the following equation described by Pearson et al. (2014): Where; (g clay/g <2 mm soil) is the calculated value of whole-soil clay content, (g non-gypsum separate/g <2 mm) soil is the content of non-gypsum residue in the sample expressed as a decimal ([100-% gypsum]/100), and (g clay/g non-gypsum separate) is the clay content of the non-gypsum residue.

Whole soil silt measurement
Whole soil silt percentage was calculated by subtracting the sand and clay percentages from 100.

Statistical analysis
The influence of the three disaggregation methods on the sand fractions as well as total sand were statistically analyzed by a one way analysis of variance with Tukey significant difference test for mean separation (P<0.05).All statistical analysis was performed using SPSS software version 21.0 (IBM Corp., 2012).

Selected physico-chemical soil properties
Selected physico-chemical properties of the studied samples are presented in Table 2. Field texture ranged from sand to sandy clay loam.Bulk density was high and ranged from 1.7 to 1.9 kg m -3 .The soil reaction was alkaline in all samples due to the presence of CaCO3 and the pH values ranged from 7.67 to 8.50.These results are similar to those others obtained with gypsisols (Artieda, 1996;Florea and Al-Joumaa, 1998;Cantón et al. 2003).All horizons showed salinity level and the electrical conductivity (EC) values ranged from 2.65 to 60.5 dS m -1 .The CaCO3 content ranged from 46.0 to 642 g kg -1 .Gypsum content varied in all samples and the values ranged from 42 to 108 g kg -1 .The cation exchange capacity (CEC) ranged from 1.80 to 8.25 c mol (+) kg -1 .The irregular distribution between CEC and gypsum/calcium carbonate may be due to the presence of different clay minerals derived from different parent materials in the study area.Recently, it has been reported that the irregular distribution of CEC and CaSO4 or CaCO3 may be associated with different clay minerals (Aznar et al. 2013).

Sand fractions separated by the three disaggregation methods
Figures 2 to 5 shows the sand fractions of the studied soil samples with the three disaggregation treatments.
In the three soil samples from profile 1 (Czy, 5-25 cm), profile 2 (Czy, 0-25 cm) and profile 3 (Cky, 20-80 cm), the amount of very coarse sand decreased from dry sieving to shaking in a 7:3 ethanol: water solution to sonication in a 7:3 ethanol: water solution (Figure 2, 3, 4), respectively.This could be due to more complete disaggregation of the sample or fracture of primary grains if the energy input was too great.These results agreed with Pearson et al. (2014).In addition, the reduction in the amount of very coarse and coarse sand fractions has been reported in other studies of disaggregation treatments (Fuller and Goh, 1992;Imeson and Vis, 1984;Oades and Waters, 1991).In contrast, in soil samples from profile 4 (Cky, 20-80 cm), and from profile 5 (Cky, 5-35 cm), the trend was from shaking in a 7:3 ethanol solution to a dry sieving to sonication in a 7:3 ethanol: water solution (Figure 5 and 6), respectively.The reason for this trend is unknown but might be due to the presence of gypsum crystals with fine sand-sized fractions which has not been influenced by the great energy input of sonication.These findings coincide with those reported by previous authors (Al-Barrak and Rowell, 2006;Aznar et al. 2013;Poch et al. 2010).In contrast; the trends of the three disaggregation treatments for coarse sand, medium sand, fine sand and very fine sand was irregular among the soil samples.Reasons for this deviation from the trend is unknown but may be related to differences in parent materials.These results were similar to that obtained by Pearson et al. (2014).

Total sand by the three disaggregation methods
Table 3 presents data for the amount of total sand measured with the three disaggregation methods.The table showed that the amount of total sand in profile 3 (Czy, 0-25 cm), profile 4 (Cky, 20-80 cm), and profile 5 (Cky, 5-35 cm) decreased from dry sieving to shaking in a 7:3 ethanol: water solution to sonication in a 7:3 ethanol: water solution.This might be due to more complete disaggregation of the sample or fracture of primary grains if the energy input was too great.Similar results were obtained by Pearson et al. (2014).In contrast; the amount of total sand in profiles 1 (Czy, 5-25 cm) and profile 2 (Cky, 5-20 cm) decreased from sonication in a 7:3 ethanol: water solution to shaking in a 7:3 ethanol: water solution to dry sieving.The reasons for this deviation from the trend in profiles 3 and 4 is unknown but could be due to the presence of gypsum crystals occurring in the silt or fine sand-sized fractions which has not been affected by the great energy input of sonication or may be related to differences in parent materials.These results were agrees with Pearson et al. (2014).Additionally, Vieillefon (1979) mentioned that the presence of gypsum is mostly linked with fine sand fraction followed by the silt fraction.

Re-calculation of whole clay contents
Clay content measured by the standard pipet method for the non-gypsum residue ranged from 6.8 to 25.2 % (Table 4).The ratio of CEC/Clay is often used to evaluate dispersion errors in clay measurement, and its value commonly higher than 1 in most soils (Burt, 2011).CEC/clay ratios derived from clay contents measured by the standard method ranged from 0.07 to 1.12.Recalculation of the clay contents to a whole soil basis (including gypsum) ranged from 2.20 to 20.47% and resulted in CEC/clay ratios ranged from 0.40 to 0.82.In addition the clay content ranging between 2 to 50% (in gypsiferous soils) have been recorded by many researchers (e.g; Mousli, 1980;Van Alphen and de los Rios Romero, 1971;Mardoud, 1980;Dekkiche, 1976;Barzanji, 1973;Barzanji et al., 1975).

Textural class comparison
Texture class as determined in the field by feeling disagreed with the texture class determined from the standard particle size and modified methods for 4 of 5 samples, and that about 80% of the total samples.In contrast, textural class determined by the standard pipet method disagreed with class determined by modified method for 60% of the total samples tested (Table 5).The reason for this disagreement could be due to overestimation of clay content in the field.Similar results were reported by Pearson et al. (2014).Moreover, Vieillefon (1977Vieillefon ( , 1978Vieillefon ( , 1979) ) in comprehensive studies on the improvement of analytical methods for gypsiferous soils, concluded that, overestimation of the clay and silt contents of gypsiferous soils occur if not corrected when done by hydrometer methods.Despite these comparisons were qualitative but they suggest that the modified method is applicable and yields particle size distribution that better reflect soil conditions observed in the field.They also suggest that the modified method shows priority in separating soil particles of the gypsiferous soils than standard pipette method.

Conclusion
The particles of gypsiferous soils in the arid regions are mainly aggregated with Ca cations (e.g; CaCO3) and which lead to difficult in the disaggregation of particles in these types of soils.Hence, there is a need to applying alternative techniques to properly separating the particles of these unique soils.After applying developed Pearson et al. (2014) method to determine particle size distribution for selected gypsiferous soils from Al-Ahsa governorate, Saudi Arabia, the statistical analysis results revealed that the sonication for 3 minutes in a 7:3 ethanol: water solution was the most effective for sand fractions compared to dry sieving and shaking for 5 hours in 7:3 ethanol solution.The particle size distribution by the modified method showed slight increased in total sand content as compared to standard particle size method.Additionally, the textural classes obtained from the two methods were similar.Our study suggests that use of sonication for 3 min.in a 7:3 ethanol: water solution will give accuracy in particle size measurement for gypsiferous soils.Therefore, we highly recommended using this developed method to separate sand fractions as well as to re-calculate the clay fraction for gypsiferous soils.Finally, we concluded that the determination of particle size distribution for gypsiferous soils (≤ 40% gypsum) using this developed method will improve the understanding and ability to proper interpret of physical behavior of these unique soils.

Figure 1 .
Figure 1.Location of the study area and samples sites

Table 1 .
Geographical coordinates of the sampling sites within the study area.

Table 2 .
Selected physical and chemical properties of the studied soil samples a pH measured in 1:5 soil: water.b Electrical conductivity of saturated paste extract.c d scl: sandy clay loam, sl: sandy loam; s: sand; ls: loamy sand

Table 3 .
Total sand measured by the different disaggregation methods.

Table 4 .
Particle size distribution for five gypsiferous horizons by standard and modified method

Table 5 .
Comparison of textural classes obtained by feel, standard PSD method, and modified method.