• Change text size
• Skip to content

Australian Natural Resources Atlas

Natural Resource Topics

• About us
• Contact us

Search

Appendix 2. Australian Soil Resources Information System

Particle size distribution/soil texture

Percent clay, percent silt, percent sand, texture class (topsoil and subsoil)

This description is relevant to three different sets of modelled surfaces:

• percent clay, percent silt, percent sand from polygon models;
• percent clay from point model; and
• texture class from point model.

What is it?

Soil texture refers to the size distribution of soil particles or the relative proportions of mineral particles of various sizes (i.e. percent clay, percent silt and percent sand). Only particles of diameter < 2 mm are considered `soil'.

• The sand fraction is made up of those particles that have a diameter between 2 and 0.02 mm.
• Silt-sized particles are those with diameters between 0.02 and 0.002 mm.
• Clay-sized particles are those with diameters < 0.002 mm.

Particle size distribution can be estimated from field texturing (see box), but reliable determination of the particle size distribution requires laboratory analysis. Descriptive names (e.g. loam, sandy clay) are assigned according to the percentages of sand, silt and clay using the Australian texture triangle (see below).

Why is it important?

Soil texture is strongly related to many other soil physical (soil structure, bulk density, porosity, permeability) and chemical properties (cation exchange capacity). It is often used to estimate other soil properties (particularly soil water properties) if no direct measurements are available.

Field texture Field texture is a measure of the behaviour of a small handful of soil when moistened, kneaded into a ball (the bolus) and then pressed out between the thumb and forefinger. It is mainly determined by proportions of sand, silt and clay. Clays cause the bolus to be more cohesive, sticky and plastic. Silts confer a silky smoothness. Organic matter can make the bolus more cohesive or greasy to feel, and the types of soil minerals present and cation composition also affect field texture. Although field texture is closely related to the particle size distribution measured in the laboratory, texture classes assigned from field texture and particle size analysis are not always equivalent (e.g. soils with high levels of exchangeable sodium have a heavier field texture than suggested by the particle size analysis).

How does it vary and what is it related to?

Particle size distribution is determined by the soil's parent rock, the rate at which rock breaks down into soil and whether this material is transported and sorted by size along the way. The rate at which rock breaks down into soil is a function of rainfall and temperature. Transport of soil material is affected by topography, rainfall, wind and vegetation cover.

How and why does it vary across Australia?

Particle or soil texture varies in response to a range of factors:

• parent material - mineral composition, type (rock or sediment) and susceptibility to chemical and physical weathering (weathering status) (e.g. granite weathers to coarse sands).
• position in landscape and the method of soil formation or placement (e.g. alluvial soils have varying texture distributions depending on river - levee - floodplain position). Alluvial soils in the backplain position are often dominated by clay. Coast soils are often sandy.

Comparing different models for %clay

Maps of percent clay in topsoil and subsoil were produced using respectively:

• polygon models, and
• combined point- and polygon-based models.

Spatial analysis of the maps compared the polygon model (in classes) against the combined point-polygon model (in classes) showing a 42% agreement between the two data layers. Some 82% of the estimations are covered within +1 class or in value terms +10% clay. Part of this discrepancy is due to differences between field texture (used in deriving the polygon models) and laboratory determinations of percent clay (used in the point-polygon model).

The map derived from the combined model is considered to be more accurate, and is the preferred choice for applications that require an estimate only of percent clay. Where clay, sand and silt are all required, the polygon-based maps should be used.

How can these maps be applied?

Soil texture can be used, in conjunction with other information, to infer soil susceptibility to erosion (see soil erodibility attribute). It is also used to estimate soil permeability when no measurements of hydraulic conductivity are available. As a rule of thumb, sandy soils are highly permeable while clay soils are very slowly permeable.

Table A4 Summary statistics-percent clay in topsoil by percent of land use type across Australia.

	<10%	10 - 20%	20 - 30%	30 - 40%	40 - 50%	>50%	Total land use class area (ha)
Conservation and natural environments	51	24	19	4	1	2	263 903 800
Production from native environments	16	25	30	9	4	15	443 051 500
Cropping	14	35	20	9	6	15	22 519 800
Grazing modified pasture	16	45	21	9	4	5	19 239 600
Horticulture	8	38	21	21	5	7	351 000
Irrigated cropping	2	11	18	24	9	36	949 100
Irrigated modified pasture	3	14	12	45	5	21	1 079 300
Total area							751 094 100

Table A5 Summary statistics - percent clay in topsoil (from combined point - polygon model) by percent of land use type for river basins containing intensive agriculture* .

	<10%	10-20%	20-30%	30-40%	40-50%	>50%	Total land use class area (ha)
Conservation and natural environments	30	53	10	5	2	1	54 814 200
Production from native environments	16	48	18	10	4	3	184 376 300
Cropping	30	39	11	10	7	3	22 241 100
Grazing modified pasture	35	43	14	6	2	0	18 482 500
Horticulture	19	47	18	11	5	0	351 500
Irrigated cropping	3	26	20	25	17	9	948 800
Irrigated modified pasture	7	22	34	33	4	0	1 080 000
Total area*							282 293 800

Table A6 Class difference between polygon and point models for topsoil clay classes.

Class difference between polygon and point model		Percentage of area* (%)
-5	Point model ‘underestimating’	0
-4	^	0.1
-3	|	0.2
-2	|	0.8
-1	|	9.3
0	No difference	42.4
1	|	34.0
2	|	10.9
3	|	2.0
4	v	0.3
5	Point model ‘overestimating’	0

* Only for overlap

What is the level of uncertainty: percent clay, point model (topsoil and subsoil)?

The model for percent clay in topsoil is generally good, although it is strongest in Queensland, Victoria and Tasmania and less reliable in southern New South Wales, Northern Territory and Western Australia. This is considered to be a more reliable estimate of percent clay than that produced by the polygon-based model.

The subsoil model is much less reliable, as indicated by the error diagnostics. It is weakest in New South Wales, Western Australia, South Australia and Northern Territory.

Error diagnostics

Error diagnostic	Topsoil	Subsoil
Number of points used	9750	7050
R2	0.538	0.319
Relative error	0.64	0.79

What is the level of uncertainty: texture class, point model (topsoil and subsoil)

Despite the large number of available data points, models for texture class were not reliable. The topsoil model predicts classes A and D well, but has difficulty in distinguishing sandy loams and loams.

The subsoil model failed to distinguish the sand classes, and a three-class model was used. This model was more successful in distinguishing clays than other classes.

Error diagnostics

Error diagnostic	Topsoil	Subsoil
Number of points used	99316	73163
Error (%)	45.9	32.3

Topsoil model class-specific error rates:

A	sands	0.15
B	sandy loams	0.83
C	loams	0.75
D	clay loams/light clays	0.28
E	clays	0.57

Subsoil model class-specific error rates:

K	sands/sandy loams/loams	0.48
M	clay loams/light clays	0.51
N	clay	0.17

What is the level of uncertainty: percent clay, percent silt, percent sand, polygon model (topsoil and subsoil)?

The scale of the various soil maps used in deriving this map is shown in Figures A2 and A3.

The Northcote Factual Key (a soil classification - Northcote 1979) uses soil texture as a differentiating characteristic. Estimation of texture for soils with a uniform primary profile form is straightforward, but it is more difficult to be definite about other soil types (e.g. duplex soils can have a range of surface textures).

The reliability of estimates of texture for Principal Profile Forms that do not have texture as a diagnostic varies. The main sources of uncertainty in deriving particle size distributions from texture are:

• is in the choice of representative values of percent clay, percent silt and percent sand for each texture class-use of Budiman’s methods has improved these estimates significantly; and
• because field texture classes and particle size distribution are not completely equivalent (see McKenzie et al. [2000] for a discussion of the ways in which field texture may differ from particle size distribution, particularly for soils with high clay contents).

• Next
• Contents
• Previous

• ANRA home
  • Map maker
  • Natural resource topics
  • Data Library

• Natural resource topics
  • Agriculture
  • Coasts
  • Dryland Salinity
  • Invasive species
  • Irrigation
  • Land
  • Natural resource economics
  • People
  • Rangelands
  • Soils
  • Vegetation and biodiversity
  • Water
  • Publications

• Agriculture
  • Overview
  • Climate-soil
  • Change
  • Statistics
    • New South Wales
    • Victoria
    • Queensland
    • South Australia
    • Western Australia
    • Tasmania
    • Northern Territory
    • Australian Capital Territory
    • Other Territories
  • Beef
  • Cotton
  • Grains
  • Horticulture
  • Sheep/Wool
  • Sugar