D.J. PRATT (*)
(*) D.J. Pratt, Principal Scientific Officer, Ministry of Overseas Development Land Resources Division, Tolworth Tower, Surbiton, Surrey, England.
1. Categories of rangeland survey and evaluation
2. Review of experiences
3. Site evaluation: Parameters and methods
4. Sampling and data processing
6. Guidelines for the future
This paper explains the arrangement of the Agenda and highlights some of the issues for discussion.
The Bamako seminar has been sponsored by ILCA in order that specialists in range survey and officials responsible for range development in Africa may come together to discuss requirements and methodologies in range evaluation and mapping, with a view to establishing guidelines for the future, with particular reference to the needs of development planning.
Currently there are several organisations involved in range survey, using a variety of methods, with varying results; and it is not always the most precise studies that are the most useful. This seminar will provide, for the first time, an opportunity for an exchange of experiences between workers from different organizations, and between contractors and clients. Moreover, the seminar comes at an opportune time: a time of expanding development activity, when there is a growing and exacting need for range evaluation and mapping.
The present paper is intended to focus attention on some of the major issues for discussion. It is not exhaustive and certainly it is not intended to be restrictive to the flow of discussion, but it seeks to explain why the agenda has been arranged in the way that it has, and to indicate the expectation in each session.
Rangeland can be studied from several viewpoints and for several purposes. Depending on the type of survey and its objectives, appropriate methodologies vary. As a basis for discussion, three categories are specified:
Monitoring of ecological change represents a form of study with restricted objectives, often undertaken on an extensive rather than an intensive basis; it is a category of survey for which satellite imagery has particular relevance, and discussion is expected to focus on the values of ERTS and alternative imagery, and on the forms of ground control that are needed.
Predevelopment planning calls for more comprehensive studies, covering, e.g., land potential and condition, water resources, human and livestock populations, social-territorial organisation, communications and marketing, wildlife resources and disease hazards: all aspects warrant brief consideration, in order to establish the full spectrum of criteria needing study and what constitutes the "safe minimum", but discussion should focus on the role of pastoral surveys and the appropriate methodology.
Detailed assessments are needed both in the implementation of management plans - e.g., for assessing stocking rates or bush-control practices - and in experimentation: possibly the variety of methods involved is too numerous for generalisation, but particular attention should be given to requirements in the evaluation of on-going development programmes, which is likely to require more detailed studies than are needed for predevelopment planning.
The intended focus of discussion throughout will be the criteria and methods that are appropriate to each category of survey.
Having established, in general terms, the main categories of survey that are involved, it is appropriate to examine some specific examples in more detail. A number of case studies will be presented, probably:
- the CEPE ecological survey of Tunisia,
- the IEMVT pastoral survey in Mali,
- the UNDP/FAO range surveys in Kenya, and
- the Ecological Monitoring Programme that is starting with the establishment, in Kenya, of an Ecological Monitoring Unit (EMU).
A final selection of case studies will be made on the basis of the contributions that are received. The opportunity will be provided in ensuing discussion to consider related experiences, as well as to consider the utility of past surveys in relation to development.
One objective of this session is to arrive at a clearer understanding of the types of survey that are needed and those that, in developmental terms, must be considered wasteful.
A basic feature of most surveys is the delineation and evaluation of rangeland sites. It is on the criteria by which one site is differentiated from another that the practical value of surveys often depends.
Primary production is a basic parameter. However, there is more to its assessment than yield determinations. Productivity needs to be considered first in relation to site potential, as determined by climate, soil and topography. How are climate and landform best classified? Should the physical attributes of climate and landform always be defined first, as a framework for vegetation classification; and to what extent should or can vegetation be used to define the physical environment? Apart from site potential there are questions relating to the choice of vegetation units. Is it more useful to classify vegetation initially by floristic or physiognomic type? To what precision and by what methods should floristics be defined? When it comes to determining yield or productivity, distinction needs to be drawn between herbage and browse. Methods for assessing herbage yields are fairly standard, but by what criteria and methods should browse production be assessed?
Ecological status is a site parameter of particular significance to range management, especially when expressed as range condition and trend. Several methods are available for determining conditions and trend, but how relevant are these to African rangelands (e.g., where the "climax" vegetation is bushland thicket, or where the animal users are likely to comprise several species, including wildlife)? It is anticipated that discussion will focus on (a) relevant criteria of condition and trend and (b) methods for assessment. A separate aspect for consideration is the use of plant species or communities as indicators of overgrazing (which are directly relevant to condition and trend), or of fire or groundwater. How can such relationships be determined, or proven?
Animal/plant interactions. It is usually insufficient to evaluate rangeland in terms of primary production and ecological status. Most surveys would be incomplete unless the "animal value" of vegetation were also considered. But this is complicated by the many ways in which animals and plants interact. For example, what parameters and methods should be used in the assessment of interactions between feeding habits/mechanisms and vegetation, or between season of use and vegetation; or in the assessment of tsetse habitats, or palatability and nutritive value of plants and plant parts. Perhaps the first step should be to list all the interactions that are of significance; or perhaps this will be covered by one or more of the contributed papers.
Animal carrying capacity is a parameter of fundamental practical significance to the management of rangeland. However it is conceptually complex and difficult to quantify. Methods of assessment that relate dry matter productivity to animal needs do not, in themselves, make provision for the ecological effects of grazing pressure or the varying grazing habits of different animal species. The convention of reducing all animals to standard livestock units (usually on the basis of body weight) is helpful but can represent a gross oversimplification. The present meeting provides an opportunity for reviewing current thinking and methodologies, and for establishing guidelines for both the standardisation of an interim approach and the experimentation needed for the application of improved standards.
Assessment of potential for rangeland improvement warrants separate consideration. The aspects of site evaluation so far considered relate mainly to the management of rangeland in its "normal" state. But what if overseeding or some other radical change in the natural vegetation is contemplated? What additional criteria and methods are relevant to the assessment of potential for improvement? Even if opportunities for improvement are limited in extent, they can still have a profound effect, as can accrue from the identification and development of water-spreading sites.
Most surveys involve sampling and the handling and processing of data. In some types of surveys statistical analysis of the data may be appropriate, which makes additional demands on sampling procedure.
Sampling procedure may have entered into the earlier discussion on methods of site evaluation but it is relevant here to review experiences in depth, both in respect of the selection of sample sites (e.g., the advantage of stratification versus random sampling) and the choice of the number and size of samples. The specific requirements of complex and sparse vegetation also deserve consideration, as does the choice of standardised formats for records, to facilitate data extraction and analysis.
Ordination and analysis of data is a specialized and developing field. However, before considering the more sophisticated of statistical methods it is relevant to review non-statistical methods of classification: both classification by environmental conditions and by floristics. The subjective ordination of data has been the basis for presenting the results of many past surveys, but is this to be recommended for the future and, if so, what checks can be introduced against errors of judgement? If classification is by floristics, what are the merits and disadvantages of classifying by dominants, as compared with communities? When considering statistical methods, the experiences of CEPE in determining ecological groupings will be particularly relevant. Discussion will also need to focus on computer capabilities, and their influence on survey design.
Data storage deserves special consideration, to help ensure that relevant data are kept in appropriate form for subsequent re-evaluation and comparison with later surveys. It will be relevant to consider the experiences of IBP and practicing organisations in this regard.
Mathematical modelling also warrants review. This is a particularly fast-developing field, with a number of specific applications to range management. It is anticipated that discussion will focus on modelling as a means of predicting change, and of translating primary data into stocking rates, etc. It will be relevant to establish precisely how far methodology has progressed in these areas, and what are the prospects for the future.
Maps are invaluable for the presentation of survey results. They are a means of presenting information which otherwise would be completely indigestible - and in a form directly relevant to management and development planning. However, they tend to be both time-consuming and costly to make, and rely for their impact on the skillful use of cartographic conventions and display.
Automatic techniques and other aids for transcribing thematic material from aerial photographs to base maps can speed and reduce map preparation. The latest developments in this field need to be reviewed.
Cartographic conventions need to be standardised if maximum benefit is to be obtained from rangeland maps. This applies particularly to conventions in respect of climatic conditions (especially aridity); vegetation; landform; and soil.
A review of existing conventions might allow recommendations regarding their future use.
Form of presentation is important where several types of information need to be combined on one map or one set of maps. The case for hierarchical display, superimposition, synthesis, or separate thematic maps will be considered in relation to the types of information most relevant for range development.
Choice of scale is a basic consideration, especially in relation to the information that is to be presented, its intended use and the economics of map production. This relationship and "optimum" scale warrant some discussion.
The object of the final session is to examine future needs and ways of improving future surveys, and to recommend appropriate actions.
Future needs should be established, with special reference to manpower requirements. The needs for ecological monitoring are relatively predictable, but is it possible to estimate the extent of areas needing predevelopment survey (or assessment while under development)? It is certain that a greatly increased number of range ecologists could be employed in Africa, complemented by technicians from other disciplines, but from where are they to come? To what forms of institution are they best attached? What forms and level of international input is needed; and what is needed to train local ecologists?
Standardisation is desirable if not overdone. It would be relevant to review the previous discussions, to examine what can conveniently be standardised at this stage. Some attention deserves to be given to terminology (e.g., in respect of vegetation type) but emphasis should be on methodology, including: environmental description; vegetation units; cartographic display; range condition and trend; forage yield assessment; and animal carrying capacity.
Economic criteria for the design and evaluation of surveys need attention. Surveys are costly, but their inputs and their benefits are not always evaluated economically. What are the relevant criteria for evaluating the cost and benefit of surveys; and what methods should be employed? It is particularly appropriate to examine methods for assessing the intensity of survey that its warranted.
Subjects for further research or technical consultation need to be identified, in order that gaps of knowledge and unresolved issues can be catered for. Subjects for examination should have become clear from earlier discussion.
Recommendations will conclude the meeting. Their nature is not prescribed, but if action is required they should be addressed to the area of responsibility of ILCA. It is not anticipated that the meeting will address recommendations directly to governments or agencies.