Previous Page Table of Contents Next Page

Rangeland resource surveys for farm and regional development in Queensland


(*) N.M. Dawson, Department of Primary Industries, William Street, Brisbane, Queensland 4000, Australia.

Data collection and storage
Data retrieval and interpretation


In Queensland, the land system type survey has provided the best means of preparing resource inventories in rangelands. The land system map shows the spatial distribution of the different resource units, whilst the land unit descriptions provide the detailed information required for farm and regional planning.

In these surveys, black and white aerial photographs at scales of 1/80,000 have proven to be the most suitable and economical form of film for air photograph interpretation and map preparation. The ERTS (Earth Resources Technology Satellite) imagery is a valuable tool when used in conjunction with conventional photography.

The field sampling procedure, the recording of data for use in the computer-based data bank, and the method of retrieval and classification of the data are described. These methods allow the interpretation of masses of quantitative data in order to identify those environmental attributes which have an important bearing on the productivity and management of the rangeland communities.

The methods described are allowing the rangelands of Queensland to be mapped and described at a total cost of less than 1 cent per hectare.


A field survey of some 15 million hectares of pastoral land in southwestern Queensland was conducted in 1971-1972. This was one of a series of surveys initiated by the Department of Primary Industries at the request of the State Land Development Committee in order that an accurate knowledge of the nature of the country and its potential might be made available for the formulation of land management and administration policies.

The administration and management of the western arid lands for continuing and improved productivity requires a knowledge of the nature and location of the pastoral ecosystems, their condition and trend.

As a result of the activities of the former C.S.I.R.O. Division of Land Research in Queensland, it has been found advantageous to develop an inventory of rangeland ecosystems based on the concept of conducting " land system " surveys at a scale and of a type suitable for property management purposes.

This survey had the objectives of describing, classifying and mapping the country, including its surface geology, topography, soils and vegetation; and of broadly assessing land-use potentialities based on a consideration of these inherent land characteristics. This survey method, which has been described in some detail by Christian and Stewart (2), enabled large areas to be mapped into land systems (**) and described in terms of land units. The method is based on the premise that each type of country appears on the aerial photographs as a distinctive pattern, hence the interpretation of serial photographs is the basis of the approach.

(**) Christian and Stewart (1953) described a land system as "an area or group of areas throughout which there is a recurring pattern of topography, soil and vegetation".

In the Queensland study emphasis has been placed on providing a resource inventory which indicates land types, their spatial distribution and relative importance in terms of area and production.

Data collection and storage

In reconnaissance surveys of large areas the pattern (land system) rather than the individual components (land units) becomes the mapping unit. The pattern is commonly the level of first recognition in air photograph interpretation. Mabbutt (7) noted that pattern usually arises from the arrangement as well as the nature of components and stressed that their interrelationship must be considered and depicted. Land units may exhibit sharp identifiable boundaries or may grade continuously into one another, depending mainly on their geomorphic history, micro-relief, and soil development. In addition, individual land units may support a range of plant species or soil development.

One of the main aims of sampling in these surveys is to establish modal characteristics and the range of characteristics associated with land units, which are the main descriptive components. In past surveys the land system description has been the important descriptive feature. When using reports of this type it is difficult to obtain detailed descriptions of the land units without consulting the various sections on soil, vegetation, etc. As well, valuable site information may not be presented. To alleviate such problems in this survey, the land system remains basically descriptive (Fig. 1) and shows the relationship between the land units. More emphasis has been placed on the description and assessment of the land units.

Figure 1 - A land system description D1 Arrabury (2,250 km²).

LANDFORM: Dunes (5-10 m high) with sloping duneflanks superimposed on flat plains. Dunes are longitudinal with some converging and diverging; mobile crests with steep slopes (15-50 %). Slopes of duneflanks and interdune plains range from 0-8 %.

GEOLOGY: Quaternary sand over-lying mostly Quaternary clay sheets. Qs.

SOILS: Predominantly red earthy sands, Uc 5.21 (Titheroo, Booka) to sandy red earths, Gn 2.12, Gn 1.12 on the duneflanks. Textures become sandier towards the crests. Dunes and mobile crests are mainly red siliceous sands, Uc 1.23 (Yanko). Small areas of claypans with grey clays, Ug 5.24, and sandy surfaced texture contrast soils occur.

VEGETATION: Spinifex wooded hummock grassland occurs on duneflanks and interdune plains with sandhill canegrass open hummock grassland, blue-bush pea sparse forbland and bare areas on mobile crests. In places mulga, western bloodwood grassy tall open shrubland occurs on interdune plains.

For each land unit (see Fig. 2) there is a detailed description of landform, geology, soils, vegetation, and land utilisation factors. In the land unit descriptions soils are described in terms of principal profile form, great soil groups, and soil mapping units, as well as in descriptive terms. One or more analyses of representative profiles within the units are included in the land unit description. The vegetation description includes lists of the predominant, frequent and infrequent species and their structural formation. The land utilisation summary considers grazing capacity, availability of drought fodder, and land utilization problems such as woody weeds, erosion and pests, together with an assessment of land use.

Figure 2 - A land unit description


LANDFORM: Extended flanks of longitudinal sand dunes. Slopes 1 to 8 p. 100. Dunefields superimposed on flat plains.

GEOLOGY: Aeolian Quaternary sand. Qs.

SOILS: Soils are deep to very deep neutral, red earthy sands and occasionnally sandy red earths. Some soils increase in texture from loamy coarse sands or coarse sandy loams, to sandy clay loams and sometimes sandy clays at depth. Surface soils are loose commonly with a thin surface crust. Red siliceous sands occur on the upper slopes. The earthy sands are uniform coarse sandy loams and loamy coarse sands. Uc 5.21, Gn 1.12 Uc 1.23. Titheroo, Yanko.

Depth cm

pH H2O 1:5

Air Dry Mois. p. 100


p 100 Air Dry Weight Org.


Particle Size p. 100 Oven Dry Weight


Ex. Cap.

Ex. Ca. M-equiv.

Cations Mg./100g



Avail. P p.p.m. Bi-Acid carb.

Moisture Available p. 100


























































































VEGETATION: Spinifex wooded hummock grassland. Triodia basedowii predominates with scattered low trees and tall shrubs emerging. Usually no well defined low shrubby layer is conspicuous but low shrubs do occur. Ground cover is variable with the areas between the hummocks of T. basedowii devoid of vegetation or supporting short grasses and fortes.

STRUCTURAL FORM: Open hummock to hummock grassland. Ht 1 m, PFC (variable) 20-15 p. 100.

TREE, TALL SHRUB LAYER: Ht 2.5-5 m, PFC 1 p. 100 (higher in isolated places).

Frequent spp: Acacia aneura, Eucalyptus terminalis, Grevillia juncifolia, Hakea divaricata, H. leucoptera.

Infrequent spp: A. ramulosa, Codonocarpus cotinifolius.

LOW SHRUB LAYER: Ht 0.5-1.5 m PFC 1 p. 100 (up to 5 p. 100 in places).

Frequent spp: A. ligulata, A. tetragonophylla, Cassia artemisioides, C. desolata, C. nemophila, C. oligophylla, C. pleurocarpa, Dodonaea attenuata, Eremophila duttonii, E. obovata, E. latrobei, Gyrostemon ramulosus.

GROUND LAYER: Ht 1 m, PFC (variable) 5-35 p. 100. Predominant spp: Triodia basedowii, Ht 0.5-1 m, PFC 15-10 p. 100.


Frequent spp: Calocephalus multiflorus, Calotis multicaulis. Crotalaria cunninghamii E. eremaea, Euphorbia drummondii, E. wheeler), Helipterum floribundum, H. moschatum, Lepidium rotundum, Macgregoria racenigera, Nicotiana velutina. Ptilotus obovatus, P. polystachyus, Senecio gregorii, Trachymene glaucifolia.

Infrequent spp: Bassia ventricosa, Calandrinia balonensis, Calotis hispidula, Centipeda thespidioides, Gossypium sturtianum, Halgania cyanea, Haloragis gossei, Psoralea sp. aff., P. eriantha, Pintelea trichostachya, Rhagodia nutans, Salsola kali, Scaevola depauperata, Solanum esuriale, Swainsona microphylla ssp. affinis, S. oroboides, Tephrosia rosea var. angustifolia, Trichodesma zeylanicum.


Frequent spp: Aristida browniana, A. contorta, Dactyloctenium radulans, Enneapogon polyphyllus, Eragrostis basedowii, E. cummingii vel. aff., E. eriopoda, Eriachne aristidea.

Infrequent spp: Aristida anthoxanthiodes, Bulbostylis barbata, Paraneurachne muelleri, Panicum australiense, Triraphis mollis.

LAND USE: Grazing capacity 1 to 2 beasts/km², high infiltration rates; low-moderate A.W.H.C. (due to depth); susceptible to wind and water erosion if burnt and/or overgrazed; drought feeding, useful after rain; tourism; dead tree 1 p. 100; tree/ha (variable) commonly 10-40; top feed scarce; perennial grasses abundant; condition static.

Mapping techniques

Both 1: 40,000 (1951) and 1: 80,000 (1969) black and white photographs were available for the area. The 1: 80,000 photographs were preferred because:

a) The interpreter in this case was able to identify the mapping unit just as easily at the smaller scale (1: 80,000), remembering that the final map scale was to be 1: 250,000;

b) Mapping with the 1: 40,000 photographs as against the 1: 80,000 photographs took more than twice the time. This becomes quite significant when it is remembered that the photographs were subject to three periods of photo-interpretation; these being prior to a reconnaissance survey, before field sampling, and prior to map preparation;

c) The extra time involved in preparing the base map of mapping units using the 1: 40,000 photographs.

In a survey such as this, where cost efficiencies are important, it is difficult to justify the extra costs of providing colour photography. Whilst colour photography was much easier to use, when it was compared with black and white photography it was concluded that in nearly all cases the mapping units could be observed on the black and white photographs. The increase in speed with the colour photography would not have justified the extra costs of flying the area for colour.

Imagery from the ERTS (Earth Resources Technology Satellite) is useful in rangeland areas not only to assist in the separation of land systems and land units, but also to monitor the condition of these once their spatial distribution and characteristics are known. This imagery introduces another dimension into photo-interpretation in rangelands, but in order to obtain best results it needs to be used in conjunction with conventional techniques at present.

In this study mapping units from the air photography were transcribed onto the International R502 1: 250,000 map sheets. These sheets are available for use for farm planning in the area. However, the land system map has been prepared at a scale of 1: 500,000. Very few of the mapping boundaries were lost in the preparation of this map.

Site selection

In such surveys it is inevitable that basic classification and sampling will involve some bias in the collection of data. Noy Meir (10) considers that whilst randomizing site selection is too time-consuming in broad-scale projects, the intuitive selection of " typical " sites as a method of selection should be rejected. Noy Meir (op. cit.), in a survey of a semi-arid area of south-eastern Australia, suggested that systematic sampling along pre-determined road traverses, correcting the locations of sampling sites so as to achieve area-proportional representation of the major land types recognizable on aerial photomosaics, provided the most efficient method of site selection.

In this work, site sampling has been carried out at two levels. At the less detailed level, vegetation and landscape changes were described along continuous road traverses recording soil, vegetation, and topographic information. Detailed site sampling was conducted on the basis of sampling of limited length traverses across distinctive landscape patterns.

The locations of these traverses and sites along these traverses were marked on photomosaics in the laboratory. Where possible these landscape traverses were replicated three times, with the more widespread or important patterns being subject to greater replication.

Detailed sites were selected along the landscape traverses not only to obtain sampling sites within the land system patterns, thus establishing the descriptive characteristics of component land units, but also to establish relationships between adjoining land systems. It is important in surveys of this type that relationships between both the land units and land systems be identified where they occur. Figure 3 is a representation of an idealized land system relationship for Western Queensland. An understanding of this type of model greatly facilitates sampling- and description. Sites were selected within these traverses on the basis of their characteristic air photo reflectance and position in the landscape. This introduces bias which may be limited by replication. Imagery from the ERTS programme and color photography where available greatly assist in the selection and siting of landscape traverses.

Topographic, vegetation and land use characteristics were determined at these detailed sites from an area seldom exceeding 2,500 m². Surface soil characteristics were also investigated over this area, but soil sampling was restricted to sampling at a central point. The variable soil surface characteristics and surface nutrient levels as shown by Charley and Cowling (1), mainly due to changes under trees and shrubs, made surface sampling difficult. To avoid bias in sampling, soil profiles were sampled at a standard point one-fourth of the distance between the base of predominant plant species - the relevant species being separated by the model distance between such species. Work in adjacent areas by Dawson and Ahern (pers. com.) suggests that whilst the central profile sample point is adequate for profile sampling, the bulking of nine selected 0-10 cm samples along the wheel point quadrat (described later) traverse provides a representative surface sample which is important in the assessment of the land use characteristics of the site.

Investigations by Dawson, Boyland and Ahern (6) show that some analysis of the 0-10 cm soil samples may be used to assess the biological productivity and condition of sites.

Data recording

The data collection, storage and retrieval techniques discussed were adopted to accumulate and process the large volume of resource information recorded by the multi-discipline team.

Land characteristics are recorded directly onto EDP field recording sheets (Fig. 4 and 5) to avoid errors in transcription. Information was coded using the code descriptions outlined by Dawson (5). A limit of eleven cards, each of 80 characters, was placed on data entry, the last nine characters of each card being necessary for identification purposes.

Site location, topographic characteristics, geology, soil description, summary classifications, and land use features were restricted to four cards on one recording sheet, whilst vegetation characteristics were recorded on two cards on the second recording sheet. Soil analytical data required four cards, and a final card has been kept spare for additional information. A fixed format was used as this facilitates flexibility in using information in computer programmes other than those currently used to retrieve data.

Site location is identified by means of latitude/ longitude co-ordinates and military grid references (yard grid) taken from the International R502 Series map at a scale of 1: 250,000. A conversion programme is used to convert these co-ordinates to the AMG (metric) co-ordinates. These characteristics of sites can then be accurately printed out by computer when required.

Date of sampling, site location, rainfall using nearest station, and name of the surveyor are recorded. Twelve landforms have been described, and position of these landscapes is based on Conacher's (4) land surface model. Classification of degree of slope, micro-relief at the site, aspect and relative relief of the site are recorded. The elevation of the site above sea level is taken from the nearest comparable recording, or omitted if not applicable. Geology, based on the Bureau of Mineral Resources Maps, parent material, and geomorphology are determined. Susceptibility to both wind and water erosion is assessed using the USDA (14) classification. The percent stone cover, percent bare ground, percent exposed rock, drainage, and runoff were visually assessed in terms of class groupings. The wheel point quadrat described by Roberts (11) has been used quite effectively to accurately assess percent stone cover, percent bare ground, percent exposed rock, and percent litter. Evidence indicates that these figures can be obtained from traverses of 400 wheel points.

Soil profile attributes were described, using many of the criteria described in the USDA soil survey manual (14). Soil samples were taken at fixed intervals, these being 0-10 cm, 10-20 cm, 20-30 em, 5060 cm, and 110-120 cm down the profile. Soil sampling at the fixed sampling intervals aids comparison and weighing of profiles when further interpretations and classification of the analyses are made. Analyses conducted on the soil are listed in Figure 2. Main characteristics recorded were soil texture, Munsell colours, type and degree of mottling, soil structure, consistence, field pH and concretions. Depth of the surface soil and soil surface characteristics were also recorded. Soils were classified into principal profile forms, Northcote (9), and great soil groups, Stace et al. (13).

Figure 3 - Relationship between landsystems in south western Queensland

Figure 4 - Land use survey recording sheet soil and land characteristics

Figure 5 - Vegetation characteristic sheet land use survey

Five plant species, mainly perennials, were considered adequate to characterise the principal plant associations at each sampling site. The five species were those that contributed most to the biomass or perennial species which apparently characterised the association. The range of height, diameter of stem and projective foliage cover of the five species were recorded by means of class ranges. Other data noted for the five predominant species include regeneration, growth form, and evidence of grazing. The presence of eleven other species that contribute most to the association was recorded on the data sheets. The degree of disturbance and the agent responsible; the condition of the perennial and annual grasses, the perennial and annual fortes and top-feed; presence of litter and dead trees; and number of trees/hectare were also recorded.

Results with the wheel point quadrat over 900 points indicated that whilst the botanist could estimate basal cover with a fair degree of accuracy, the relative composition of the pasture species was difficult to estimate accurately in most situations. For this reason the wheel point quadrat will be used for those sites considered suitable as representative of large areas.

Structurally, the vegetation at sites was classified on a scheme slightly modified from that proposed by Specht (12). The two most important species were tied to this structural form to provide a summary of the vegetation at each site.

Time spent at each site ranged from 20 minutes to more than 60 minutes, with a median value of approximately 30 minutes. Where sites are considered to be suitable as representative of large areas or ecological benchmarks, then additional time of approximately 1 hour is necessary to accumulate additional quantitative data with the wheel point quadrat.

Data retrieval and interpretation

The data bank

Moore (8) indicates that there are three major benefits from the systematic collection of data for storage as data banks. These are:

a) that collectors tend to look to compatibility of data between projects,

b) that more data are consistently collected and interpreted, and

c) that it leads to flexibility of data classification.

These data are available for use by other workers, provided they recognise the limitations of the collected data.

The CSIRO CDC 3600 computer at Canberra was used to retrieve data using the INFOL programme. Data can be readily retrieved in either code form, or decoded and printed out in the required manner. In the land system survey it was put to most use by characterising the soil and vegetation groups by means of simple print-out data for groups of sites to establish characteristics and range of characteristics.

In the simplest form, paired data at sites were retrieved for each of the major soil groups to examine for correlation co-efficients between recorded features. Environmental factors such as soil analytical data, vegetation data (tree and shrub numbers/ ha and ground cover) and other environmental factors (degree of wind and water erosion, slope and rainfall) were analysed to establish correlations between these attributes. This is useful in determining which attributes may be important in plant community or soil type. Interpretation of data in this way was helped in the selection of attributes for other investigations such as condition and trend studies.

For example when Dawson, Boyland and Ahern (6) analysed a number of environmental attributes for the major vegetation community (Acacia aneura) using the REGRESS programme available on the CSIRO Cyber 76 computer, relationships and interrelationships between the various attributes were able to be established. This programme calculated correlation co-efficients between all attributes, a multiple regression equation, and the multiple correlation co-efficient. Correlation co-efficients established by Dawson and Ahern (op. cit.) for the mulga community are listed in Table 1.

It was apparent from the data analysis of the communities that a number of soil attributes could be used to assess the reaction of sites to use. In the Acacia aneura community there is a delicate soil-plant relationship, and the gradual or rapid depletion of one variable, in this case either the density of mulga or soil organic matter, may lead to a breakdown in the system, thus reducing potential productivity.

In the past, studies such as this have failed to interpret results obtained at sites, due to the time involved. The listing in tabular form of frequency distributions of selected data is extremely useful in characterizing the soil or vegetation groups. Data has been easily extracted for processing in this form.

Vegetation lists of the occurrence, frequency of occurrence, and site occurrence of the predominant vegetation species and other associated species were readily tabulated. This saved valuable time. Extraction of site data, such as listing of sites susceptible to different classes of soil erosion, was readily achieved in a usable form. These data, if required, can be produced in a visual form as a distribution map.

More sophisticated sorting can be ordered and controlled by the user. Data can be sorted, ordered, and formatted on selected field characteristics, this information being either in decoded or coded form.

The use of these simple sorting and formatting techniques of analysis using the collected data has led not only to improved and more accurate description of the land units, but has provided a check of any initial bias in the grouping of sites. It has allowed the interpretation of masses of quantitative data in a shorter time: then was possible previously.

Tableau 1 - Correlation coefficients showing relationship among environmental factors at 47 mulga (Acacia aneura) site


The collection and use of field and chemical analytical data from reconnaissance surveys have been greatly improved with the advent of data bank and retrieval systems. The obvious advantage of collection in this way is that it allows additional data to be collected, usually with increased accuracy, this data being more freely available and easily and cheaply extracted. There is less likelihood of data being lost to use, and easy accessibility leads to more efficient interpretation using other computer techniques

Disadvantages are few - the main points being that initially, coding and interpretation of field sheets may be more time-consuming, and that many records in this instance were limited by precise class definition, not allowing flexibility.

Perhaps the main danger of this work is the possible misinterpretation or over-emphasis on accuracy of data, when data are used without first consulting with the collectors.

Using the techniques described, large areas of Western Queensland have been mapped and described at a cost of less than one cent per hectare.*

(*) Acknowledgments: Mr. D. E. Boyland, Botany Branch, Queensland Department of Primary Industries, contributed to the sections dealing with vegetation classification. Messrs. A. Hegarty and W.F.Y. Mawson provided valuable assistance in the project. Dr. A.W. Moore, Division of Soils, C.S.I.R.O., advised on the planning and implementation of the " Data Bank ".

Tableau 2


1. CHARLEY, J.L. and COWLING, S.W. - Changes in soil nutrient status resulting from overgrazing, and their consequences in plant communities of semi-arid areas. Proc. ecol. Soc. Aust., 1968, 3: 28-38.

2. CHRISTIAN, C.S. and STEWARD, G.A. - General report on survey of Katherine-Darwin region, 1946. C.S.I.R.O., 1953. Aust. Land Res. Ser. n° 1.

3. CHRISTIAN, C.S. and STEWART, G.A. - Methodology of integrated surveys. In: Aerial survey and integrated studies, Proc. Conference, U.N.E.S.C.O., Toulouse, 1964 Publ. 1968. pp. 233-280.

4. CONACHER, A.J. - A systems approach to description and interpretation of the land surface of the northern half of the North Island, New Zealand. Earth, 1957, 1 (2).

5. DAWSON, N.M. - Data collection, recording and processing in land utilization studies. Conference paper, Development Planning Branch, Qd Dep. Prim. Inds, 1972.

6. DAWSON, N.M., BOYLAND, D.E. and AHERN, C.R. - Land management in south-west Queensland. Proc. ecol. Soc. Aust. (in press).

7. MABUTT, J.A. - Review of concepts of land classification, in: Steward, G.A., ed. Land evaluation, Australia, Macmillan, 1968.

8. MOORE, A.W. - Regional soil data bank for future evaluation and projection. Paper presented to the F.A.O./U.N.D.P. Regional Seminar on Soil Survey and Soil Fertility Research, New Delhi, Feb. 1971.

9. NORTHCOTE, K.H. - A factural key to the recognition of australian soils. Aust., C.S.I.R.O., Divisional report, Division of Soils, 1965. (7/61).

10. NOY MEIR, I. - Multivariate analysis of the semiarid vegetation in south-eastern Australia: Nodal ordination by component analysis. Proc. Ecol. Soc. Aust., 1971, 6: 159-193.

11. ROBERTS, B.R. - Ecological studies on pasture condition in semi-arid Queensland. Rep. Qd Dept. Prim. Inds., 1972.

12. SPECHT, R.L. - Vegetation, in: The australian environment. 4th ed. rev. C.S.I.R.O., Melbourne, University Press, 1970.

13. STAGE, H.C.T., et al. - A Handbook of Australian Soils. (Rellim: Glenside, S.A.). 1968.

14. UNITED STATES DEPARTMENT OF AGRICULTURE. - Soil survey manual. Washington, D.C., USDA. 1951 (Agricultural Handbook n° 18).

Previous Page Top of Page Next Page