Printer Friendly

Field prediction of sodicity in dryland agriculture in Central Queensland, Australia.


Soil sodicity is defined as the presence of a high proportion of sodium ions relative to other cations. A high proportion of sodium ions on the clay mineral exchange weakens the bonds of the soil particles when the soil is wetted. As a result, the clay particles swell and disperse. The dispersed clay particles move through the soil, clogging pore spaces. Hence, sodicity is a condition that degrades the soil by making the soil more dispersible, restricting water entry and the ability of the soil to conduct water (Department of Natural Resources 1997). In dryland agriculture, high sodicity reduces the ability of a plant to access stored water. The rapid determination of field sodicity would provide:

(i) field classification of Vertosols and Sodosols;

(ii) instant advice to a landholder;

(iii) a framework to decide if laboratory analysis is needed; and

(iv) rapid sodicity determination in soil mapping.

Within Central Queensland, dryland agriculture mainly occurs on Vertosols, with subsoil's containing moderate to high cation exchange capacity (CEC) and clay mineralogy dominated by smectites. The ratio of CEC to clay percentage (CCR) is used as an indication of soil mineralogy. Soils with similar geomorphology and/or parent material and climatic conditions often contain similar mineralogy or CCR measurements. Provided a soil has CCR above 0.55, where smectites are dominant (Department of Natural Resources 1997), field prediction of exchangeable sodium percentage (ESP) from CEC or clay percentage is possible. Therefore within areas of similar mineralogy, ESP can be predicted by the use of linear regression using clay percentage and sodium ion concentration provided CCR is above 0.55.

Within Queensland, a relationship has been developed for laboratory pH and ESP for the Burdekin Irrigation Area (Baker et al. 1983). High alkalinity in these soils provides a useful guide to sodicity. However, soil pH within the Vertosols of Central Queensland is highly variable (Tucker 1984; Webb and Dowling 1990). The majority of these soils are highly alkaline and sodium concentrations vary depending on the degree of weathering (Gunn 1967, 1974). Hence, pH is not a useful guide to indicate sodicity for dryland cropping areas in Central Queensland.

The use of a sodium-specific electrode meter allows the opportunity to measure field exchangeable sodium. Within Central Queensland, CEC is related to clay percentage, in which the latter can be estimated by the field bolus method (McDonald et al. 1990). Provided that field measurements can accurately measure exchangeable sodium and CEC, field ESP can be calculated. This paper investigates the prediction of sodium and CEC from field measurements and the relationships between ESP and field collected parameters such as sodium and clay percentage for some Central Queensland soils.

Materials and methods


Soils from 55 sites throughout Central Queensland (Fig. 1) were fully described according to McDonald et al. (1990). Cores were taken to a maximum depth of 1.5 m. The soils were classified (Table 1) under the Australian Soil Classification (Isbell 1996). The majority of soils sampled were formed from colluvial and/ or alluvial sources and contained similar subsoil mineralogy (CCR: x = 0.71, s = 0.23, n = 210).


Sampling and measurements

Field tests were conducted on soil samples collected during the initial soil surveys for increments of 0-0.1, 0.1-0.2, 0.2-0.3, 0.5-0.6, 0.8-0.9, 1.1-1.2, and 1.4-1.5 m. A further 10 sites from the Bauhinia site were also sampled but only for 2 depth increments (0-0.1 m and 0.1-0.2 m). Field electrical conductivity ([EC.sub.f]) and field sodium ion concentration ([Na.sub.f], mg/kg) were determined in 1:5 soil water solution using a calibrated EC field meter and a Cardy C-122 sodium ion meter, respectively. The [Na.sub.f] values later converted to [cmol.sub.c]/kg.

All soil samples were analysed according to Baker and Eldershaw (1993) for the available depth increments. Therefore, laboratory measurements of electrical conductivity ([EC.sub.1]), sodium ion concentration ([Na.sub.1]), CEC, and clay per cent (Clay%) were available. Laboratory EC was determined in a 1:5 soil:deionised water suspension shaken for 1 h and determined at 25 [degrees] C.

The exchangeable cations (including [Na.sub.1]) were determined by a pre-wash with 60% ethanol, followed by removal of exchangeable cations with 1 M N[H.sub.4]Cl at pH 8.5 in 60% ethanol. Absorbed ammonium was removed using calcium nitrate and potassium nitrate. Ammonium and chloride in the calcium nitrate and potassium nitrate leachate were determined on an auto analyser using colorimetric methods. The difference in [cmol.sub.c]/kg was reported as the CEC.

Clay% was determined by particle size analysis based on a modification of the hydrometer method of Piper (1942). The modifications were that the soils were dispersed with sodium hexametaphosphate and sodium hydroxide. Samples high in gypsum were sieved through a 0.2-mm sieve after an initial boiling treatment prior to an acid treatment. With soils containing carbonate, the sum of particle sizes may be <100% where acid treatment is used. Results were reported on an oven-dry basis.

Statistical analyses

Various samples were excluded from the analysis as they were either unrepresentative of the study area or clearly spurious data. These were:

(i) Site SFS-056 (B, Theodore)--only had one depth increment (0.5-0.6 m) and was unique with a very high sodium ion concentration ([Na.sub.1] = 16, while next highest was 11) possibility due to the presence of high salinity.

(ii) Sites SFS-029 and SFS-030 (A, Baralaba)--on highly variable soils and other factors (possibly gypsum) appeared to be interfering with the data being considered.

(iii) Site SFS-026 (C, Baralaba) at 1.1-1.2 m--low CEC (12).

(iv) Site SFS-327 (Gindie) at 1.1-1.2 m--very high CCR (>3).

Relationships between [EC.sub.f] and [EC.sub.1], [Na.sub.f] and [Na.sub.1], and CEC and Clay% were investigated using linear and non-linear regression.

The ESP of a soil was calculated as:

(1) ESP = 100*[Na.sub.1]/CEC

Relationships between ESP and [Na.sub.f], [EC.sub.f], Clay% and CEC were investigated. From Eqn 1 we know that ESP is proportional to Na and inversely proportional to CEC. Therefore, ESP would most likely be predicted from [Na.sub.f] and CEC or [Na.sub.f] and Clay% (as CEC is related to Clay%). Initially these parameters, and also [EC.sub.f], were individually related to ESP and then various combinations were investigated.

The adequacy of a model was assessed by examining the residual and normal probability plots and the percentage of variance explained by the model.


Relationship between [EC.sub.f] and [EC.sub.1]

A strong linear relationship was observed between [EC.sub.f] and [EC.sub.1] (Fig. 2) for a range of field EC of 0.04-2.95 dS/m (s.e. in parentheses):

(2) [EC.sub.1] = 0.0176 (0.0086) + 0.956 (0.017) * [EC.sub.f] ([R.sup.2] = 0.933; residual s.d. = 0.089)


It should be noted that data were sparse for [EC.sub.f] > 1. There were also 2 outlying points in the mid range of [EC.sub.f] but there was no justification for excluding these points.

Relationship between [Na.sub.f] and [Na.sub.1]

The relationship between [Na.sub.f] and [Na.sub.1] was best described by a linear regression of [Na.sub.1] on [square root of [Na.sub.f]] (Fig. 3a):

(3) [Na.sub.1] = -0.964 (0.142) + 11.49 (0.348) * [Na.sub.f] ([R.sup.2] = 0.831; residual s.d. = 1.01)


Alternatively, the relationship could be described adequately by an exponential regression (Fig. 3b):

(4) [Na.sub.1] = 9.56 (0.66) - 9.39 (0.61) * [0.0587.sup.[Na.sub.f]](0.0216)

([R.sup.2] = 0.829; residual s.d. = 1.01)

Relationship between CEC and Clay%

The relationship between CEC and Clay% was best described by an exponential curve (Fig. 4):

(5) CEC = 8.78 (4.76) + 3.79 (1.85) * [1.0377.sup.Clay%] (0.0058)

([R.sup.2] = 0.790; residual s.d. = 7.43)


Prediction of ESP from [EC.sub.f], [Na.sub.f], CEC and Clay%

No clear relationship was observed between ESP and [EC.sub.f], primarily due to highly variable data for field EC > 0.5 (Fig. 5a). An exponential relationship between ESP and [Na.sub.f] was apparent (Fig. 5b):

(6) ESP= 25.55 (1.31)- 26.81 (1.21) * [0.00927.sup.[Na.sub.f]](0.00491)

([R.sup.2] = 0.801; residual s.d. = 3.80)


However, there was considerable variation about the curve as can be seen from the plot of actual (laboratory measured) ESP against predicted values (Fig. 6a).


Although there was some linear trend between ESP and the inverse of CEC ([CEC.sub.inv] = 1/ CEC; Fig. 5c), the low ESP values for the Bauhinia samples even at high values of [CEC.sub.inv] tended to mask this relationship. There was little evidence of a relationship between ESP and the inverse of Clay% ([Clay.sub.inv] = 1/Clay%) (Fig. 5d). It should be noted that fewer data were available when fitting [Clay.sub.inv] due to missing data in Clay% (no Clay% data for the Bauhinia sites and some other sites/depths with missing data).

Various non-linear models with more than one explanatory variable including combinations of [Na.sub.f], CEC, [CEC.sub.inv], Clay, and [Clay.sub.inv] were fitted. In particular, 2 models based on the expression of ESP given in Eqn 1 and on the relationship between Clay% and CEC were considered. The residuals obtained following the fitting of Eqn 6 were plotted against each of [CEC.sub.inv] and [Clay.sub.inv] and revealed linear trends. Therefore, equations with an exponential term involving [Na.sub.f] and a linear term involving [CEC.sub.inv] (Eqn 7) or [Clay.sub.inv] (Eqn 8) were fitted. The first model (Eqn 7) predicted ESP more accurately than the model with only Nag as it explained a further 8.5% of the variation ([R.sup.2] = 0.885 compared with [R.sup.2] = 0.801; Fig. 6b v. Fig. 6a). Although Eqn 8 was not as accurate as Eqn 7 ([R.sup.2] = 0.850; Fig. 6c) it enables field prediction of ESP as both [Na.sub.f] and Clay% are measured in the field:

(7) ESP = 2.263 (0.569) + 800.9 (59.2) * [1 - 1.084 (0.0297) * [0.0387.sup.[Na.sub.f] (0.0153)]/CEC

([R.sup.2] = 0.885; residual s.d. = 2.98)

(8) ESP = 2.090 (0.928) + 1433 (141)*[1 - 1.0760 (0.0407)*[0.0543.sup.[Na.sub.f] (0.0234)]/Clay%

([R.sup.2] = 0.850; residual s.d. = 3.35)

The constant term in Eqn 8 (2.090 [+ or -] 0.928) is only marginally significantly different from zero. There was only a small change in the accuracy of the model ([R.sup.2] = 0.847; residual s.d. = 3.39) when refitted without the constant. Also recall that models including Clay% are based on 29 fewer values than the other models, due to missing data.


The relationships between [EC.sub.1] and ESP were investigated by RJ Tucker, SA Irvine, MD Godwin, and RC McDonald (unpubl. data) and showed some correlation. However, there was high variability at high [EC.sub.1]. A similar finding was made with the soils studied in this paper.

Isbell (1996) advised caution in the use of very low CEC and Na values in the calculation of ESP, due to the possibility of producing a misleading result. Therefore, a soil must have a sufficient CEC and Na (>3 and 0.3 [cmol.sub.c]/kg, respectively) before ESP should be calculated.

Relationships between ESP and CCR have been used to determine leaching fractions of various soils under irrigation (Shaw et al. 1987). Within Central Queensland, CCR is useful to determine the relative weathered nature of a soil and hence is an indication of its formation. Lower CCR is usually caused by a low CEC (due to weathering) and hence measured sodicity can be distorted.

From Eqn 1, ESP is theoretically proportional to Na. Noting that the relationship between [Na.sub.1] and [Na.sub.f] could be described by an exponential relationship, it is logical that ESP would be exponentially related to [Na.sub.f] as was observed in Eqn 6. Further, as CEC is theoretically inversely proportional to ESP (Eqn 1), one would expect the result obtained in Eqn 7.


EC was predicted from field measurements of EC using a linear regression. A linear relationship of square-root transformed field Na adequately predicted Na as did an exponential equation of field measurements of Na. An exponential relationship existed between CEC and field measured Clay%.

The best predictor of ESP was a non-linear equation involving [Na.sub.f] and CEC. However, from a practical viewpoint, and to allow field prediction of ESP, the non-linear relationship involving [Na.sub.f] and Clay% is probably the most suitable. A slightly less accurate estimate of ESP is available from the exponential equation based solely on [Na.sub.f].
Table 1. Australian Soil Classification within the study areas

For confidentiality, letters A, B and C have been substituted for
property names

Study area Classification Area
 (% of

 A Epicalcareous-endohypersodic, self-mulching, 95
 Black or Grey Vertosol
 Epihypersodic, self-mulching, Grey Vertosol 5
 B Epicalcareous-endohypersodic, self-mulching, 100
 Black or Grey Vertosol
 C Epicalcareous-endohypersodic, self-mulching, 87
 Black Vertosol
 Episodic, epipedal, Brown Vertosol 13

Bauhinia Supracalcic, mottled-mesonatric, Brown Sodosol 13
 Episodic-endocalcareous, crusty, Brown Vertosol 31
 Epihypersodic-endocalcareous, self-mulching, 56
 Brown Vertosol

Capella Epicalcareous, self-mulching, Black Vertosol 100

Dysart Epicalcareous, self-mulching; Black Vertosol 100

Fernlees Haplic, self-mulching, Black Vertosol 100

 A Mottled, supracalcic, Brown or Red Dermosol 100
 B Epicalcareous-endohypersodic, self-mulching, 100
 Black Vertosol
 C Hypocalcic, subnatric, Black Sodosol 100

Kilcummin Haplic, self-mulching, Black Vertosol 100

Gindie/Orion Endocalcareous, self-mulching, Black Vertosol 100

 A Epihypersodic, self-mulching, Black Vertosol 66
 Salic, self-mulching, Black Vertosol 34
 B Endocalcareous, self-mulching, Black Vertosol 100
 C Endohypersodic self-mulching, Brown Vertosol 95
 Epihypersodic, epipedal, Black Vertosol 3
 Haplic, hypocalcic, Red Dermosol 2

Wowan Epicalcareous-endohypersodic, self-mulching, 100
 Brown Vertosol


Baker DE, Rayment GE, Reid RE (1983) Predictive relationships between pH and sodicity in soils of tropical Queensland. Communications in Soil Science and Plant Analysis 14, 1063-1073.

Baker DE, Eldershaw VJ (1993) Interpreting Soil Analyses for agricultural land use in Queensland. Department of Primary Industries, Queensland, Bulletin QO93014.

Department of Natural Resources (1997) `Salinity management handbook.' (Department of Natural Resources Queensland: Coorparoo, Qld)

Gunn RH (1967) A soil catena on denuded laterite profiles in Queensland. Australian Journal of Soil Research 5, 117-132.

Gunn RH (1974) A soil catena on weathered basalt in Queensland. Australian Journal of Soil Research 12, 1-14.

Isbell RF (1996) `The Australian soil classification.' (CSIRO Publishing: Melbourne)

McDonald RC, Isbell RF, Speight JG, Walker J, Hopkins MS (1990) `Australian soil and land survey field handbook.' 2nd edn. (Inkata Press: Melbourne)

Piper S (1942) `Soil and plant analysis.' (University of Adelaide: Adelaide, S.Aust.)

Shaw RJ, Hughes KK, Thorburn PJ, Dowling AJ (1987) Principles of landscape, soil and water salinity--processes and management options, Part A. In `Landscape, soil and water salinity. Proceedings of the Bundaberg Regional Salinity Workshop'. Bundaberg. (Queensland Department of Primary Industries Conference and Workshop Series QC87001)

Tucker RJ (1984) Frequency distributions of soil morphological and pH data from field descriptions of cracking clay soils in the Emerald Irrigation Area, Queensland. In `The properties and ultilization of cracking clay soils. Proceedings of a symposium'. Armidale. Reviews in Rural Science 5. (Eds JW McGarity, EJ Hoult, HB So) (University of New England: Armidale, NSW)

Webb AA, Dowling AJ (1990) Characterization of Basaltic clay soils (Vertisols) from the Oxford land system in Central Queensland. Australian Journal of Soil Research 28, 841-856.

Manuscript received 4 September 2000, accepted 2 April 2001

S. A. Irvine (A) and D. J. Reid (B)

(A) Department of Natural Resources, PO Box 19, Emerald Q1d 4720, Australia.

(B) Department of Primary Industries, PO Box 6014, Rockhampton Q1d 4702, Australia. Corresponding author; email:
COPYRIGHT 2001 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Irvine, S.A.; Reid, D.J.
Publication:Australian Journal of Soil Research
Article Type:Statistical Data Included
Geographic Code:8AUST
Date:Nov 1, 2001
Previous Article:Effects of gypsum and stubble retention on the chemical and physical properties of a sodic grey Vertosol in western Victoria.
Next Article:Measuring transient solute transport through the vadoze zone using time domain reflectometry.

Related Articles
Rising salt: a test of tactics and techniques.
Seeking clues to watertight cropping.
Geologically related variations in saturated hydraulic conductivity in the regolith of the western wheatbelt of Western Australia and its...
Preface: sodicity issues in agricultural industries--current research and future directions.
Soil responses to sodicity and salinity: challenges and opportunities.
Sodicity, conventional drainage and bio-drainage in Israel.
Leadership needed on clearing: native vegetation is the backbone of productive farms and healthy catchments. To save it, governments must organise...
E[C.sub.1:5]/exchangeable Na, a sodicity index for cotton farming systems in irrigated and rainfed Vertosols. (Short Communication).
A study of the interactions between salinity, soil erosion, and pollutant transport on three Queensland soils.
Links between dryland salinity, mosquito vectors, and Ross River Virus disease in southern inland Queensland--an example and potential implications.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters