Gilberto C. Gallopín
The concept of situational indicators was introduced by Gallopín (1994, 1995) in the context of agroecosystem sustainability. Here, summarizing recent developments (Gallopín 1996), the concept is generalized to encompass a broader set of issues. The analysis also illustrates that non-numerical indicators can be useful and may be rigorously defined. They also represent a class of indicators not obtained by aggregation (and actually may be used to disaggregate, as needed).
Selecting and measuring indicators of sustainability at the level of agroecological zones or landscape units may often be location specific (e.g. different indicators may be relevant for steep highlands than for flat prairies). Measurements or observations imply a cost which could become limiting. Therefore, an approach that increases costs-effectiveness through a process of successive approximations is developed. It implies a multi-tiered process, defining indicators of sustainability at three basic levels of analysis (coarse, medium and fine). In principle, the levels could be as few as two, or as many as judged necessary.
Here, the concept of situational indicators of sustainability is introduced, representing a non-numerical function of both 'pressures' or 'driving forces' and 'state' variables. The concept of situational indicators is a particular case of indicators of systems behaviour.
SITUATIONAL INDICATORS OF SUSTAINABILITY
At the basic level of analysis for a given region, landscape, agroecological zone, etc., different units can be identified on the basis of natural characteristics and agricultural suitability, such as piedmont, high plains, fertile/infertile soils, etc. This information is often available or obtainable from secondary sources. A systematic analysis of current or past land management practices or production systems can be performed. Careful comparisons, including a survey of extreme or advanced problematic situations, can lead to the identification of risk factors associated with the application of a given management practice to a landscape unit in the region considered. Conversely, on the basis of experiments or comparisons with successful situations, opportunities for sustainable use can be identified.
A matrix can be constructed as in Figure 1, where each cell can adopt three values: (1) no problem expected, (2) danger (denoting a potential problem), and (3) opportunity (denoting an unused opportunity for improvement). Some cells may represent impossible combinations (e.g. heavy mechanized agriculture in steep highlands). While the matrix is more useful when based on empirical evidence, in some cases a matrix based on hypothesized relations could be useful. The first level of approximation may be methodologically based on remote sensing, geographical information systems, and ground verification. For cells signalling the likelihood of problems or opportunities, sets of selected situation-specific indicators should be identified to be assessed at a finer resolution (e.g. using rapid rural appraisal, field surveys, local expert consultation, etc.) in order to check the actual existence of problems or opportunities and their nature and magnitude (Figure 2). If necessary, indicators at still finer resolution levels may be selected and measured or monitored (e.g. through field measurements or experiments).
As detailed indicators are assessed or monitored only for the situations suggesting potential danger or opportunities, the process of screening and diagnosing is made through successive approximations and goes only as deeply as is judged necessary. This is much more economical and cost-effective than massive routine monitoring of fine-grained indicators.
The term 'situational indicator' applies to the coarse level discussed above, in which the indicative information is provided by the situation (in the sense of 'position with respect to conditions or circumstances' , 'combination of circumstances at a certain moment', or 'the way in which something is placed in relation to its surroundings' -MWCD). In other words, the relevant information is provided not by the landscape class, nor by the production system type, but by their combination.
The indicator is an abstract non-numerical variable that is a function of two non-numerical variables (landscape and production system). The abstract variable used as an indicator could be called a 'degradation risk' , a 'land-use sustainability indicator' or by another convenient name.
Formally, the variable 'landscape' can be denoted by 1 in the case of Figure 1. Its state set (set of possible values) can adopt only three values or states L = {I, II, III }. In a real situation, the number of states may be much higher, but at least for the case of landscape types, it will be a finite number.
Landscape is a nominal variable, with no mathematical property, such as ordering or distance, recognized in its state set. The distinguishing property for the variable 'landscape' is space. Landscape is not expected to change in time. In some situations, the property might be a population (of a country or region), or both population and space.
The variable 'production system' (p) can adopt four states or values; that is, P = { A, B, C, D}. It is also a nominal variable. The distinguishing property for 'production system' could be the same as that for 'landscape' with the addition, in most cases, of time (given that production systems may change through the years).
The variable 'land-use sustainability' (s) can adopt three values; S = {Dg, OK, Op} for 'danger', 'no problem expected' and 'opportunity', respectively. It is a nominal variable, but it could also be conceived as an ordinal variable (with the value Op being greater than OK, and OK greater than Dg). The distinguishing property for 'land use sustainability' should be the same as that for 'production system'.
The relevant indicator here is s, a variable which is a function of the two variables l and p, adopting the values 'danger, 'no problem expected' and 'opportunity' for different combinations of categories of production systems applied to categories of landscapes. Each cell in Figure l or each pair [(l, p), s] represents a value or state of the indicator.
Neither of the two individual variables (landscape and production system) is useful as an indicator by itself. The indicative power lies in the relation combining them. The relation must be a function to be useful as an indicator (if not, it would mean that the same combination of values of the individual variables could result in different values of the indicator).
In this case, the function can be defined by listing its elements:
R (L' P), S) = { [(I, A), Dg], [(I, C), Op], [(II, A), OK], [(II, B), OK], [(II, C), OK], [(II, D), Op], [(III, B), OK], [(III, C), Dg], [(III, D), OK] }
In this particular example, the function is empirical, derived from observations and past experience.
In the context of the Pressure-State-Response, Pressure-State-Impact-Response, or Driving Force-State-Response frameworks it is interesting to note that the situational indicator of land-use sustainability (s) is a function linking a Pressure ( or Driving Force) variable (p) to a State variable (j).
Contrary to some other driving forces (such as deforestation rates, emission of pollutants, waste generation, etc.), land use (or land production system in the example chosen here) is not a continuous nor a monotonic variable that has an unambiguous significance in the sense that an increasing (or decreasing) value implies an increasing ( or decreasing) risk. Land use is a nominal variable adopting values as 'types of land use' which by themselves are neither good nor bad in general. The same practice may be sustainable or unsustainable depending on the land-scape, land type or socio-ecological system to which it is applied.
Perhaps because of this, the proposals of indicators for pressure or driving forces, in the case of land management, are weaker than for other factors, and they are often denoted simply by 'land use changes' or 'land use' (Hammond et al. 1995; World Bank 1995; UN-DPCSD 1995).
Situational indicators for land use could be generated at national and regional levels, through the use of satellite imagery or land use/land suitability maps obtained from secondary sources, GIS, or expert consultation. Thus, this kind of indicator is not necessarily generated through the aggregation of individual data. Many countries already generated maps of land suitability or even maps of recommended production systems. The level of resolution may be as base as including only two values (sustainable, unsustainable) or, in some cases, it could be as high as to provide many degrees of sustainability, in an ordinal or perhaps, metric scale. The landscape units may increase in number, and the production systems may be characterized in finer detail. For international comparisons, preferred levels of resolution would need to be agreed.
The indicator can be quantified by estimating the actual extension or proportion of land under different degrees of risk (such as the proportion of agriculture carried out on marginal lands in relation to the total land under production, or the proportion of agricultural land being used accordingly to its potential). This could be monitored through time and space.
Note that this does not force all countries to use the same characterization of landscapes or production systems. Local specifics could be preserved insofar as the criteria to assess the degree of risk (the mapping of the landscape and production systems into the land-use sustainability indicator) is standardized. Such coarse standardization already exists for the rough definition of land suitability classes, degrees of severity of soil erosion, and other relevant factors. In this way, a value of risk could be assigned to, say, a shifting cultivation system in a moist tropical forest, that could be compared with the value for a nomadic pastoral system in a desert area.
TOWARDS GENERALIZATION
The concept of situational indicator can be considered as a particular case of indicators of 'sustainability of utilization'. For any resource (land, water, forests, species, dilution capacity) for which suitability for different uses or sensitivity to different human activities can be defined, and actual use, pressure, technology, or driving force can be estimated, a situational indicator could, in principle, be defined, linking the actual use (the 'pressure' or 'driving force') with the 'state' (reflected by the suitability or sensitivity class). In general, suitability or sensitivity could change through time as the result of human intervention or natural variation.
Another example is illustrated by the index of potential development-related threats to coastal ecosystems (Bryant et al. 1995). The index is a simple function of five indicators: presence of cities greater than 100,000 inhabitants, presence of major ports, population density, road density, and pipeline density in the coastal area. A simple sub-index function adopting the values (low, medium, high) is associated with each indicator. The overall threat index in each coastal geographical unit is the maximum of the values adopted in that unit, by any of the five functions.
The case of 'situational indicators', in which the relevant information is provided by the relation between variables, demonstrates the possibility of rigorous formal treatment of what could otherwise be seen as an heuristic devise with no theoretical basis. The examples discussed in this box focus on variables geographically distributed. However, they can be shown to be particular cases of the much more general interpretation of sustainability indicators.
REFERENCES
Bryant, D., Rodenburg, E., Cox, T., and Nielsen, D. (1995) Coastlines at risk: An index of Potential Development-Related Threats to Coastal Ecosystems. WRI Indicator Brief, The World Resources Institute, Washington, D.C.
Gallopín, G.C. (1994) Agroecosystem Health: A guiding concept for agricultural research? pp. 51-65. In: Nielsen, O. (ed.) Agroecosystem Health. Proceedings of an International Workshop. University of Guelph, Ontario, Canada.
Gallopín, G.C. (1995) The potential of agroecosystems health as a guiding concept for agricultural research. Ecosystem Health 1(3): 129-140.
Gallopin, G.C. (1996) Environmental and sustainability indicators and the concept of situational indicators. A systems approach. Environmental Modelling & Assessment (In Press).
Hammond, A., Adriaanse, A., Rodenburg, E., Bryant, D. and Woodward, R. (1995) Environmental Indicators: A systematic approach to measuring and reporting on environmental policy performance in the context of sustainable development, World Resources Institute, Washington, D.C.
(MWCD) Merriam-Webster Collegiate Dictionary, l0th. Edition
The World Bank (1995) Monitoring Environmental Progress. The World Bank, Washington, D.C.
United Nations Department for Policy Coordination and Sustainable Development (UN-DPCSD). (1995) Work Programme on Indicators of Sustainable Development of the Commission on Sustainable Development. Division for Sustainable Development, UN-DPCSD.