Contents - Previous - Next

Level 5: The aim, problems, and procedure

AIM: To select from the range of Evaluation Factors identified at Level 3 (and their component attributes) those which bear most directly on the sustainability of the defined use in the defined locality; and to further identify threshold levels of special significance for each of these selected indicators.

PROBLEMS: The most difficult consideration associated with preparing a range of indicators for evaluating sustainability is that of ensuring that no significant influence for future change in the system has been omitted. Levels Three and Four of the FESLM are largely concerned with safeguards against this, but this effort will be wasted if the final range of indicator attributes, on which evaluation is finally based, is incomplete.

Examination of a single 'indicator' attribute may serve to show that a given use is unsustainable, but only a complete range of 'indicators', covering all potentially unstable aspects of the system and its surroundings, can guide a confident assessment of 'sustainable'.

In many situations the unsustainability of a land use system is obvious. The FESLM is needed when this is not the case. The more marginal is the decision on sustainability, the more likely it is that the status of some obscure indicator will determine the issue.

PROCEDURE: 1. Selection of Indicators

Indicators are selected from a broader range of factors by more rigorous application of the same criteria of 'Relevance', 'Stability' and 'Predictability' used to distinguish Evaluation Factors at Level 3.

Many attributes are quite obviously indicative of degradation and likely unsustainability (eg. erosion, product quality, yield). Others, perhaps less obvious, may yet be identified without the assistance of the FESLM. Table 7, drawn up from the pooled experience of a multi-disciplinary group, shows the more important indicators of sustainability for six broad agricultural systems across the whole of Australia. In drawing up this list, the criteria to be met were that the indicators could be measured or reported and that they reflected known good or bad farming practice (Hamblin, 1992). The widely ranging nature of the chosen indicators is of particular interest.

A sequence of selection stages is proposed in the FESLM to arrive at a manageable number of (mainly) component attributes for evaluation in detailed sustainability evaluation. For a specific land use system:

1. Recognition of the full range of Evaluation Factors that bear upon land suitability (Land evaluation)

2. Selection from 1. of those factors that are predictably unstable (FESLM Level 3)

3. Breakdown of complex factors from 2. and identification of the most significant component attributes (Cause and effect analysis, FESLM Level 4)

4. Final selection of factors/attributes for use as Indicators in evaluation (FESLM Level 5)

TABLE 7: Important primary indicators of sustainable agriculture in Australia (Hamblin, 1992)



Management Level

Production Balance

Resource Base

Rain-fed crops & animals

Farm management skills, cash flow, equity planning

Water use efficiency; yield/area/rainfall

Soil health; pH, nutrient balance, biota

High rainfall pastures for animals

Production/area; animal weight/hectare

Plant growth/cover; % greenness, species/area

Soil big-indicators; worm, termite etc., numbers and species

Low rainfall range land for animals

Management capability; planning debts, asset condition, record keeping

Animal health and productivity; liveweight gain, quantity/quality, fleeces, carcasses

Pasture and soil condition; % bare ground, pasture composition

Irrigated crops and pastures

Farm and district profitability; debt equity etc. for farm level, true cost versus benefit inc. envir/al. costs at district level

Water use efficiency; plant use/water applied, water table trends crop/weight/water used

Soil health; infiltration rate, % subsoil compaction, biomass act/y, chemical residue level

Intensive horticulture and viticulture

% integrated pest management adopted in industry; chemical sales, grower records, fauna surveys

Nutrient balance; yield and nutrient contents, fertilizer sales, surface water composition

Soil permeability/waterbases; irrig/tn. water use, piezometry, soil infiltration

High rainfall tropical systems

Diversity of production; no. of land uses or crops, no. isolated vegetation patches

Water quality; surface water composition, blooms, pesticides, sediments

Soil productivity; trends pH, O.M., subsoil compaction, soil struct. condition

Discussion of some of the individual component attributes listed in Table 6 may assist understanding of the final selection stage to identify useful Indicators:

1.1 Topsoil Nitrogen %: An important determinant of successful plant growth but probably subject to too much variation in short spans of time and distance to be a satisfactory parameter of sustainability.

1.4 Soil acidity (pH): important factor in soil use, particularly in relation to its influence on nutrient availability and certain toxicities. Changes in acidity arise from complex causes that need to be evaluated, but trends in acidity are valuable primary indicators of instability.

1.5 Subsoil weatherable minerals (%): reflect a reserve of plant nutrients important in assessing soil suitability, particularly for perennial crops. However, in the absence of serious erosion, change in the content of weatherable minerals is likely only in the very long term. Thus their value as an indicator of sustainability is limited.

2.3 Frequency of damaging floods: this is an example of a local attribute of obvious importance in the evaluation of suitability and of sustainability (if the frequency is changing) but one that very probably reflects causal changes (deforestation?) operating far from the investigated site.

3.1 and 3.2 Severity of specific pests/diseases: These are examples of factors that have to be sharply defined to be valuable in either suitability or sustainability evaluation. Because different pests/diseases are encouraged by different conditions (see 3.3) specialist local knowledge to identify them is essential.

3.3 Local factors favouring pest 'X'/disease 'Y': because most pests and many disease vectors are fairly mobile, this is an example of an 'attribute' which calls for knowledge of a wider area than the investigated site itself (within which there may be breeding sites or foci of infection or conditions, such as free water surfaces or high humidity, which favour build up of a particular pest or disease). Change in risk of pests or disease, perhaps from off-site causes, may have an important bearing on sustainability.

4 and 5 Gross expenditure and Gross returns. These examples illustrate the flexibility that exists in dividing complex attributes into their components. Gross expenditure has been divided into expenditure on separate items (although these could doubtless have been subdivided further). Gross returns show division into broader groupings. The choice between, or combination of, the two approaches rests on striking a convenient balance between the detail of available information and the needs of effective sustainability analysis. Decisions on such issues will have to be made in drawing up a global Master framework.

6.3 Basis of acquisition/inheritance: This is an example of a social attribute, typically difficult to assess other than in qualitative terms, but potentially important to sustainability - change in inheritance custom may lead towards, or away from, fragmentation of holdings to an impractical or uneconomic size.

Summing up, the most useful environmental attributes in the context of sustainability evaluation are those that

• reflect environmental changes important to the continuing success of specific forms of land use;

• show steady, reasonably predictable response to environmental change; without significant fluctuation over short time periods or short distances (trends can be measured with reasonable confidence);

• are a clear measure of a cause having a well understood effect;

• can be measured and expressed in numerical terms.

Attributes which fulfil all these requirements are justifiably described as 'indicators' of sustainability. A final step in the selection process must be to ensure that the complete list of 'indicators' adequately reflects the complete range of factors recognized at FESLM Level 3.

TABLE 8: Some environmental qualities and diagnostic factors (possible 'indicator' attributed) - illustrating the rating of land suitability in terms of requirements of sugarcane. (Extracted from Table C2 of Guidelines: Land Evaluation for Rainfed Agriculture (FAO, 1983)). The data are illustrative and should not be taken as authoritative in any specific locality.






Highly Suitable

Moderately Suitable

Marginally Suitable

Not Suitable

Moisture Availability

Growing period


315-365 (-)



< 210

Rel. Evapo-Trans (1- ETa/ETm) for total gr. Period


< 0.17



> 0.65

Oxygen Availability (Drainage)

Soil Drainage(class)


well drained

moderately well, imperfectly


very poor

Depth to water table over sign. Periods


> 180



< 20

Nutrient Availability






< 4.0
> 8.5

Nutrient Retention

0-20 cm

meq %




< 4

Base Saturation (lower horizons)


> 50



< 10

Excess of Salts Salinity.

EC of saturation extract


< 2.5



> 11

PROCEDURE: 2. Recognition of Thresholds:

Existing environmental knowledge, strengthened by understanding gained by 'cause and effect' investigations at Level 4, provides the basis for recognizing 'threshold' values.

The concept of 'threshold' values of 'indicator' attributes in the FESLM is very similar to that of 'critical values' of 'evaluation factors' in later applications of the Framework for Land Evaluation (see FAO, 1983). Thus, Table 8 shows how specific values of certain factors are regarded as diagnostic of suitability class limits in relation to the general factors are regarded as diagnostic of suitability class limits in relation to the general requirements of sugarcane. Each critical value is diagnostic only with respect to that particular factor. In land evaluation, the overall suitability class of the land unit, for the crop in question, usually reflected the lowest value of class-determining factor observed.

An important difference in the FESLM is that, in the defined circumstances of each evaluation, there will be only one threshold' value for each 'indicator'-the level of that indicator beyond which the land use is no longer considered sustainable.

Referring to the data in Table 8, it will be noted that there is variety in the kinds of data used to establish class limits; measured values, percentages, ratios, and in one instance (soil drainage) classes of another kind. Even greater variety is to be expected when the concept of 'thresholds' is extended to indicators within the economic and social disciplines.

The use of classifications, such as the soil drainage classes, to provide critical limits (possibly 'thresholds') is interesting; for it allows very complex inter-relationships between individual attributes to be lumped together and treated as one complex attribute or 'quality'. To be useful, such a classification needs to have a basis that can be recognized and reproduced by different observers, and must have an established relationship with the performance of the land use being evaluated. Other classifications which FAO has proposed might be used in this way include: soil workability classes, root penetration classes, terrain classes etc. (FAO, 1983).

'Crop yield' and 'Gross return' are examples of still more complex attributes which embrace the interaction of all aspects of the site. On their own, these values are too complex and potentially too misleading to provide a useful measure of sustainability, but historical trends in their value may provide unrivalled warning signals of change and, in the final analysis, values of yield and financial return may be the prime determinants of sustainability.

Referring again to the data in Table 8, it might be suggested that the critical values which separate the ' not suitable' land from lands which are, at least, 'marginally suitable', can be regarded as ' thresholds ' of sustainability. Certainly, a given land use will not remain sustainable on land which develops physical conditions that render it unsuitable. But application of the concept of sustainability thresholds is not quite as simple as these examples may suggest; for it involves the passage of time.

As soon as we try to project future values of a single indicator, we appreciate that trends in status may be influenced by other interacting attributes which together determine the significant critical values at which change develops in the future.

Final analysis

General Considerations

The first step in the final analysis of sustainability is to draw up a projection of the expected pattern of environmental change in the years to come. This will be based on the trends of change identified in Level 4, set against a background of generally accepted regional and global trends (see discussion of Off-site effects in Chapter 5).

Against a projection time scale of 20 years, or more, such global trends as population increase, climatic change, impact of AIDS and other epidemic diseases, and flows in international trade are too important to ignore. Yet their likely effects at local level are still, to say the least, uncertain. Until these things are better understood it may be necessary to draw up alternative scenarios for the future against which the local thresholds can be set (eg. longer/shorter rainy seasons; higher/lower rainfall).

Given a projected pattern of future change, sustainability analysis becomes a matter of using criteria (developed at Level 4) to establish likely values of selected indicators at successive time intervals, and comparing these with accepted threshold values for each indicator.

If the status of any single indicator falls significantly below its required threshold, the system examined is, very probably, not sustainable. It may be sustainable over a shorter time period, however, and this provides a basis for classifying and comparing sustainabilities (see below).

It would be unreasonable to suggest that just because one factor appears unstable and seems likely, in time, to reach a value on the wrong side of a required threshold, the use must automatically be declared 'unsustainable'. Judgements have to be made on the assessment of each factor, firstly on its reliability and secondly on its significance in relation to the overall assessment of sustainability.

There is certain to be wide variation in the confidence with which differences in the stability of individual factors are assessed in the course of the evaluation, particularly in relation to the different approach paths of the assessment - from observation on the site or from theory, for example. Unavoidably, the data available will be speculative, since it relates to the future.

In judging significance, account may need to be taken of foreseeable changes in other factors which may augment or reduce the effects being assessed. Such checks form part of the validation procedure (see Chapter 6).

Experience in the development and use of thresholds (and related 'criteria') can be expected to reduce the need to rely on personal judgement in evaluating sustainability, but it is probable that such judgement will always play a major part.

Classifying sustainability

Chapter 1 includes discussion of the need to recognize classes of sustainability and describes a system of classification based on the number of years that the use in question can be sustained.

Classification on this basis adds little to the difficulty of evaluating sustainability if the approach described above is used; provided that the time intervals into the future at which the assessments are made are the same as those used to distinguish the classes.

Confidence in the reliability of the diagnostic data must also play a part in classification. Class placements should not exceed what is reasonable in terms of data reliability. Where there are special doubts, these should be stipulated.

Generalized ('Regional') sustainability evaluation

In Chapter 1, a distinction was drawn and differences discussed between 'detailed sustainability evaluation' of small areas and 'generalized sustainability evaluation' of much larger areas.

Most of the discussion in Chapter 4, above, is concerned with detailed sustainability evaluation requiring full application of all levels of the FESLM. Application of this exhaustive process uniformly over large areas is not feasible-the range of variation, not only in the land but also in the land use, is certain to be too great. With an eye to the future, it may not be possible to state where the investigated land use will be located, for locations within the study area may change if other 'suitable' sites exist! The fact that two sites are suitable does not ensure that their characteristics, or their potential for sustainability, is the same.

The principles and procedures of the FESLM can be applied for regional assessments and for broad scale evaluations by 'generalizing' the process for broad level applications. Fine scale evaluations are easily conceptualized because they reflect entities with which we are totally familiar, eg. individual farm fields. Although such entities serve as the bases for scientific study, not all important considerations can be expressed at that level. This means that the entities of understanding, the field, must often be 'coupled, to other levels where there are valid, but less tangible entities, eg. landscapes, agro-ecosystems, etc. Although such levels of generalization are troublesome, they are very important because society often perceives pressing problems predominantly at these smaller scales, and evaluations for such problems must be developed. This requires some guidelines for the generalization of the FESLM guidelines and procedures for regional scale applications.

Currently there are no easy solutions to the problem of data generalization; at this point it has tended to be more of an art than a science, based almost entirely on the experience and depth of understanding of the investigator for the problem to be evaluated. However, some techniques are beginning to surface, such as Hierarchy Theory, metamodelling and others, which can assist the process.

The principles of Hierarchy Theory provide some guidance in generalizing data for different scales. Hierarchies are used to describe complex systems, such as land, where differences in organization of data and dynamic behaviour of major processes make simple aggregation of lower levels insufficient to explain higher levels. This is illustrated in hydrology, where simple aggregation of thousands of estimates of moisture diffusion will not result in estimates of stream discharge.

Hierarchies are characterized by surfaces (scales) which help to define the entities to be investigated at each level, and by observation sets (data) which identify how the investigator views the system and what he/she considers to be important phenomena or processes. Observation sets go beyond simple collection of data, and include criteria for identifying significance and the major processes which are important at any given scale.

Two characteristics, called 'grain' and 'extent', determine the distinction (fineness) that can be made in a set of data, and identify the level or scale of the data. The 'grain' of the data can be increased by making increasingly finer observations and/or sampling more often. This results in finer resolution, and moves one down the scale of hierarchies. 'Extent', on the other hand, can be increased by sampling over larger areas, using coarser resolution and/or sampling over greater time periods. Increasing the extent of the data moves one to higher levels on the hierarchy. Grain determines the lower level of resolution of the data set, whereas extent determines the largest, and these detection limits are absolute.

It follows from this that each data set is particular to a given level of hierarchy (scale). However, one can generalize the data (move to higher levels of the hierarchy) by varying the extent; conversely one can provide greater resolution (move to lower levels in the hierarchy) by increasing the grain of the data. The same effect, however, can be achieved by changing the criteria of observation between levels. In practice this is achieved by recognizing that different phenomena and processes become predominant at different scales and that the criteria for observing (sampling) these phenomena or processes vary accordingly.

These principles can be used to 'generalize' data for regional evaluations of sustainability. Aggregation of the results of evaluations from many small fields will not result in a generalized evaluation of a region. The latter, however, can be achieved by recognizing that different but related processes and phenomena are important at the regional scale and that different (kinds of) indicators will have to be applied. Similarly the 'extent' of the data used for analyses will have to be varied to match this scale of generalization and the indicators of sustainability.

Contents - Previous - Next