SECTION 1 : General

The environmental basis of agroforestry

Anthony Young
International Council for Research in Agroforestry
 P. O. Box 30677, Nairobi, Kenya

Abstract

The physical environment affects and interacts with agroforestry systems through tree and crop growth, animal performance, management operations and interactions between the tree/shrub and non-tree components (abbreviated to tree/crop interactions). Interactions take place mainly through the media of microclimate, soil moisture and soil. They may have beneficial or adverse effects, for tree and for crop. Where the ecological interactions are on balance adverse, agroforestry designs with a large tree/crop interface, such as alley cropping should be avoided. Agroforestry has a particular potential to help check or reverse degradation of soil, forest and pasture resources. An environmental basis for agroforestry requires detailed information on climate, soils and, for sylvipastoral systems, vegetation, together with important basic, but less detailed, information on landforms, hydrology and fauna. Environmental information in five ICRAF data bases is compared. An environmental classification is presented, based on climatic regions combined with distinctive situations of landforms and soils. Agroforestry systems could be designed to suit virtually any set of environmental conditions in the tropics and subtropics. The greatest potential contribution that agroforestry can make is in densely-populated steep lands.

Objectives

The aim of this paper is provide an overview of the environmental basis of agroforestry as a framework for the more detailed chapters in this volume. In particular, it draws attention to aspects which are primarily non-climatic, indicating how these interect with meteorology. It also draws attention to some basic questions raised by such a review, and indicates the extent to which they can be answered. Part of what is said overlaps with, or anticipates, later chapters.

The role of environment in agroforestry systems

Agroforestry systems

Agroforestry systems involve:

1. Growth:          of trees, crops, pastures, animals;

2. Management: of climate, water, soil, plants, animals;

3. Interactions:   tree/crop, tree/pasture, tree/animal.

Under growth and management, we draw upon large fields of existing knowledge: tree growth from forestry (e.g., the site quality concept); crop/climate and crop/soil relations from agriculture; pasture ecology; moisture conservation in drylands; soil erosion and conservation; soil fertility; tropical livestock management - even summarily listed, these are nine massive areas of relevant research.

But it is the interactions that are the distinctive field of agroforestry, albeit that we can already draw upon a tenth research area, that of intercropping. Specifically, our trade is that of interactions between the tree (i.e., tree or shrub) and non-tree components of agroforestry systems. This is represented in the phrase from the standard definition of agroforestry,'.....in which there is both an ecological and an economic interaction between the tree and non-tree components'. Herein lies the distinction from social forestry and farm forestry, despite the large overlap. Plant a block of trees along a roadside or in the corner of a farm, manage it largely or entirely for wood products, and you have social or farm forestry; but compromise on wood production for the sake of forest grazing, and you introduce an ecological interaction that is the essence of agroforestry.

So an environmental basis for agroforestry must cover effects of environment upon trees, crops and animals and their management, plus the distinctive features of interactions between the tree (and shrub) and the non-tree components, referred to henceforth as tree/crop interactions.

Environment and tree/crop interactions

Table 1 and Figure 1 show some tree/crop interactions of significance in agroforestry. Note that both positive and negative effects may be in either direction: tree upon crop (scan pasture, animal) or crop upon tree.

For the first factor, called displacement, the effect is always mutually negative. If the tree is low, as in alley cropping, land under trees cannot be undercroped, and their combined areas add up to 100%. Multi-layer systems such as homegardens can attain combined tree and crop covers of 200% or more.

The remainder of the table is self-explanatory. Microclimate and soil are the two main areas through which interactions are effected. Positive interactions are numerous, but there are four major negative ones:

1. Displacement;

2. Shading;

3. Moisture competition; and

4. Nutrient competition.

The aim of agroforestry design should be to maximize positive interactions and minimize negative ones. Sometimes we may tolerate negative ecological interactions for the sake of positive economic ones, e.g., accept some crop yield reduction for the sake of fodder produced. Interactions take place where trees and crops meet: at the tree/crop interface. They may be side by side, or one above the other. If the net ecological interaction is beneficial, then let us have it in plenty. But if negative, and we are tolerating it for the sake of economic gain, then let us reduce it as much as possible.

These ends can be attained by designs that have a larger or smaller tree/crop interface. Huxley expands on this in Section 2, but let me indicate the orders of magnitude involved. Suppose you want (say for economic reasons) 25% of your land under trees. Figure 2 shows five ways (among many) in which this can be achieved. Looking at the tree/crop interface from a 'side-by-side' point of view, the lengths, as metres of interface per hectare of land, are shown in Table 2.

Table 1. Environment and tree/crop interactions in agroforestry.

The results are striking. Contrary to common supposition, the system with the largest length of interface, for a standardized area of tree cover, is not scattered, isolated trees, but alley cropping. So unless you are very sure that net benefits arise from the existence of an interface, try almost anything else. Wherever there are problems at the interface, then of all forms of agroforestry, alley cropping is the least likely to succeed.

Figure 1 Environment and tree/crop interactions in agroforestry.

Table 2 Lengths of tree/crop interface for different agroforestry designs, standardized at 25% tree cover.

Figure 2 Six ways of arranging a 25% tree cover on one hectare. Within-field lines are tree/crop interfaces. Based on the concept developed by Huxley (1985).

Problems of the land

We have already mentioned the role of agroforestry in making better what has got worse - in solving, or at least ameliorating, problems. In the diagnosis and design approach, the primary focus is upon problems of the farmer: shortages of food, fuelwood, fodder or cash, for example.

Such shortages, however, are frequently linked with problems of the land. Table 3 lists environmental problems for which agroforestry has a potential to assist. Most are forms of degradation. The potential of agroforestry for control of water erosion and soil fertility decline has recently been reviewed by ICRAF (Young 1986b, 1987a, 1989). The chapters that follow address questions of the control of wind erosion, as well as the difficult problem of the extent to which agroforestry can assist, directly or indirectly, i n meeting the drought problem in drylands.

More generally, it is ICRAF's view that one of the major contributions of agroforestry is in improving the sustainability of land use systems; that is, in maintaining systems at a productive level, through conservation of the natural resources on which that productivity depends.

Table 3. Environmental problems which agroforestry has a potential to alleviate.

An environmental basis

Factors of the physical environment

This summary puts us in a position to ask the question, What information on the physical environment is needed as a basis for research into agroforestry? What kinds of information are relevant, and in how much detail?

Table 4 shows the relevance of the major factors of the physical environment to four aspects of agroforestry systems: tree and crop growth, land management, tree/crop interactions, and environmental degradation. Effects of one factor upon another are added in the last column. Climate and soils are the factors of greatest importance, followed by vegetation in those land use systems that make use of it in a natural or semi-natural state, primarily silvopastoral systems. Land forms appear mainly as a cause of soil erosion hazard: there is now an identifiable body of knowledge about land use, including agroforestry, on steep lands (Novoa and Posner 1981; Siderius 1986; Young 1986a). Only the underlying geology is unrepresented: its effects are considerable, but operate indirectly, via landforms, hydrology and soils.

Table 4 Effects of the factors of the physical environment on agroforestry systems. Fauna includes pests and diseases. The last column refers to initial letter of factors in the first column. // = major effect, / = effect.

So as an environmental basis for agroforestry, we need:

1. Substantial detail on climate and soils;

2. Substantial detail on vegetation for sylvopastoral systems;

3. Important basic information on landforms (mainly slope) and hydrology (mainly drainage and depth to water
   table); and

4. Information on fauna, pests and diseases;

whilst those of us who are environmental scientists cannot 'think interactions' without knowing basic elements of the geology.

An environmental data base

In 1983, ICRAF set up an Environmental Data Base (EDB) for agroforestry. This consisted of a set of basic items of information, with descriptors and classification systems, together with a computerized base for input, storage and selective retrieval (Young 1985). Recognizing the varied requirements of different users, three levels, or degrees of detail, were included: summary, intermediate and detailed. The input form to the data base is given as Table 5 (Appendix), and examples of outputs at summary and detailed levels as Table 6 (Appendix). Data input at any level can be output at the same or a less detailed level.

The concept of how the environmental data base should be used is shown in Figure 3. Having identified a site or area of interest, it should be possible to select five kinds of data, drawn from comparable environments around the world: suitable trees, suitable crops, existing agroforestry systems, recent or current research work, and publications. This would form a first gathering of data to which, as with all computerized data bases, the skill, experience and judgement of designers of agroforestry systems could then be applied.

Let it be said that this ideal has not yet been achieved: the data base, as a framework, has not yet been expanded into a store or data bank. It is possible that this may be done in the future, although substantial manpower and funding will be required. However, the aims of the EDB may soon be achieved through the combined efforts of four other data bases: those on agroforestry systems, multipurpose trees, experimental work and the ICRAF Library (Nair, Section 1, this volume; von Carlowitz, 1985,1986 and Section 1, this volume; Huxley 1985b; Labelle 1987). The agroforestry systems and multipurpose tree inventories both contain sections on environment, with a substantial although not complete identity with the EDB on kinds of information collected and descriptors used (Table 7). The data base on current agroforestry experiments uses the EDB as part of its input, whilst the computerized library index allocates publications, where relevant, to a broad climatic zone, and makes use of environmental terms as keywords.

By the end of 1987, it may thus become possible to achieve substantially the aim pictured in Figure 3:

1. You have an environment; tell us this, this and that about it;

2. Here are multipurpose trees, crops, existing agroforestry systems, current research work, and publications
   from similar environments around the world;

3. Now start thinking!

An environmental framework

Do you sincerely need a classification system? That is, do your purposes require a set of boxes, with defined boundaries, into which information can be poured? I suspect that most readers, with respect to their climatic speciality, might answer 'no' — that they would rather be given basic climatic data for a site than be told it falls into climate XYZ. And of course, the soil scientist cannot get very far by being told she has an Orthic Ferralsol, she needs full soil profile descriptions supported by analytical data.

But wait a minute ! As a background to your biometeorological analysis, do you really want the full gamut of soil description? Would it not be helpful to group your soil moisture research into results obtained 'on Vertisols', regardless of whether they are Pellic, Chromic, or do or do not have a strong, coarse blocky structure? Well, conversely, soil scientists are wont to group climates into a small number of broad classes, within which the fascinating variety of soils occur.

Broad classes come into their own when planning research programmes. To have several stations, in different countries, directing attention at common problems, the environmental framework within which they are grouped has to be quite frighteningly generalized. The ICRAF Agroforestry Research Network for Africa ( AFRENA ) groups the immense variety of tropical African environments into just four very broad ecozones.

With respect to soil classification, it has been said that 'scientists who are otherwise calm and reasonable people are liable to behave quite differently when discussing this subject,' and it may well be that the same applies to climatic classification. A recent review of classifications of climate, soil and vegetation was written as the basis for a CGIAR review of agro-ecological characterization (Young 1988). Let us, like the young ladies set upon by robbers in The Wallet of Kai Lung', rush away from this subject, screaming loudly to conceal the direction of our flight.

Figure 3 Using environmental information on agroforestry. Modified from Young (1985).

Table 7 Environmental information included in four ICRAF data bases. Properties which can be derived from data collected are included. Items additional to those shown are included in some descriptions in the Agroforestry Systems data base. The Library data base allots a class only for ecozone (climate); other factors are represented by indexing terms, shown by asterisks.

Table 8 is an attempt at a set of broad environmental classes, intended for those whose purposes require 'large boxes'. There is no doubt that climate must be the starting point, with all else superimposed (and I speak as a soil scientist!). The principle is to start from the basic climatic regions, with their associated vegetation; then for each of them, to consider what is the 'normal' situation with respect to the other factors, and what conditions constitute special cases.

Taking first climate, let us start with the three worlds of the tropics: the humid or rain forest lands, the subhumid lands or savannas, and the dry lands. Some people show a strange reluctance to recognize as a distinct zone the subhumid or 'wet-and-dry' tropics, best described by Koppen's untranslatable term, wechselfeuchten, 'changing-wet' (as opposed to immerfeuchten, always wet). This gives three large boxes for the tropics. Mediterranean and temperate climates would be added to extend such a basic classification beyond the tropics.

Next, we have the indisputable fact that conditions at 2000 m altitude are not the same as at sea level. But where to draw the line? Köppen splits A, hot, from C, warm, climates where the mean temperature for the coldest month drops below 18 °C. The FAO agro- ecological zones system puts the limit at a mean (24-hour) temperature of 20 °C during the growing period (FAO 1978/81). From a physiological viewpoint, there is an argument for the (substantially higher) point where frost appears. If you must use an altitude surrogate, some 1200 m at the equator, falling to sea level near 30° latitude, is a crude approximation, corresponding to what many would loosely call the 'highland' tropics.

Finally, take out the climates with two clear dry seasons as distinctive, and we have 9 boxes for the tropics. Sad to relate, IITA lies very close to the dry boundary of the humid tropics, whilst ICRAF's Machakos field station could be put as highland or lowland, and oscillates from year to year between bimodal subhumid and that rather rare climatic type, semi-arid with two crop failures instead of one.

Now let us dispose of all the other factors with as little complexity as possible. Vegetation may be subsumed under climate.

For landforms, let gentle to moderate slopes be considered 'normal', i.e., tacitly assumed unless specified otherwise. There are three distinctive conditions: steeplands, with steep and moderate slopes dominant; flat lands, the alluvial areas, with landforms of depositional origin; and valley floors, the dambos, mbugas, fadamas and the like of Africa.

For soils, the 'normal' circumstance can be taken as the zonal soil: ferralsols and acrisols (US Oxisols and Ultisols) for the rain forest zone, luvisols (US Alfisols) for the savannas, and the various kinds of calcimorphic soils for the semi-arid zone. By attribution, steep lands have zonal soils together with shallow lithosols, and flat lands have alluvial soils and gleys. This leaves the following as special cases, on which there is known, or believed, to be a distinctive role for agroforestry:

1. Soils developed from basic rocks (nitosols, no US equivalent), which are usually more fertile;

2. Soils developed from felsic rocks and/or on erosion surfaces, which are usually less fertile (not well identified in FAO or US classification systems; variously called weathered ferrallitic soils, granite sands, sandveld soils, leached pallid soils or cerrado soils; (Young 1975,1976 p. 142);

3. Vertisols; and

4. Saline soils.

But remember, all these are boxes; they are thus made of ticky-tacky — and they all don't exist!

Table 8 A simplified classification of tropical environments as a basis for agroforestry. Modified from the ICRAF Environmental Data Base. (Young 1985).

Stage 1 Three worlds of the tropics: climate and vegetation.

Stage 2 Climate: rainfall, temperature and seasonality.

Stage 3 Landforms and soils.

Land evaluation and agroforestry

Following upon the principles set out in the 'Framework for land evaluation', sets of detailed guidelines have been published on evaluation procedures for rain-fed agriculture, forestry, extensive grazing and irrigated agriculture. How simple it might appear to be, by combining these, to devise a system of land evaluation for agroforestry! What needs to be done, and why it is not so straightforward, has been set out elsewhere (Young 1984,1987b).

The present discussion will be restricted to the two basic questions asked in land evaluation:

1. A piece of land in mind:        what best to do with it?
2. A kind of land use in mind:    where best to practise it?

On the first of these, a substantial part of the efforts of ICRAF and others has been directed at showing how well-designed agroforestry systems, provided they function as it is hoped they will, are an improvement on present forms of land use. This has been appraised in terms of economics, sustainability and the welfare of the people.

The second question might be rephrased in our context as, 'Which agroforestry system, where?' If an agroforestry system is specified in detail (as what is called, in land evaluation terminology, a land utilization type), there must logically be a set of environmental conditions where it functions very well - the trees grow, the crops grow, and the interactions are mutually beneficial. It other environments, the same system will work rather less well, and in many others not at all - the trees will die of drought, aluminium toxicity or attack by pests.

Neither the agroforestry land utilization types nor their optimum environments have yet been systematically described and evaluated. There are some obvious relations: homegardens work best in the humid tropics, windbreaks in the semi-arid zone. Alley cropping, did you say? Some think it is a system primarily for the humid zone, where research has been concentrated until recently. Others believe that its potential may be as high or greater in the savanna or even the semi-arid zone. Can agroforestry provide a solution to the intractable problems of the sandveld (cerrado) soils? Research in progress should provide support for our presently unsubstantiated assertions.

Finally, the 10-million-dollar question (the cost of a typical development project). You cannot invest in agroforestry development everywhere at once. Where will it give you best value for money? Here are three deflections of this question, followed by one answer.

1. The question cannot be answered in terms of environment alone. The potential of agroforestry, for any given environment, becomes higher where the problems that arise under present land use are of particular kinds (e.g. fuelwood shortage, soil fertility decline).

2. Almost everywhere. Rarely does an agroforester reply, when called in to help with problems, 'Sorry, we can't help you'. Plausible agroforestry systems could certainly be designed for every environmental box in Table 8.

3. Not on flatlands. At least two viable systems are known for rice paddy lands, boundary planting (on field bunds) and intercropping of trees grown as an annual. On balance, however, this is not a priority environment. (The most unsuitable crop for agroforestry? Sugar cane.)

So where is the need for agroforestry, and its potential as an improvement over present land use, greatest? In densely-populated steeplands. In more detail:

In areas of humid or subhumid, lowland or highland, climate, dominated by moderate and steep slopes, that are densely populated. Soils are of moderate, sometimes quite high, natural productivity, but have been degraded by erosion and over-cultivation. Forest has been largely cleared, what remains is degraded by overcutting, and there is a fuelwood shortage. The remaining pastures are degraded and fodder shortage is endemic. There is less base flow to the rivers. The young men are leaving.

In such conditions, there is potential for agroforestry in the control of soil erosion, improvement of soil fertility, production of fuelwood, fodder and cash products, and (given a powerful institutional framework for planning) improvement of pastures and watershed management (Young 1986a). Do you know of such lands?

References

Carlowitz, P. von 1985. Some considerations regarding principles and practice of information collection on multipurpose trees. Agrofor. Syst. 3:181-195.

Carlowitz, P. von 1986. Multipurpose tree and shrub seed directory. Nairobi: ICRAF.

FAO. 1978/81. Report on the agro-ecological zones project. Vols I- IV. World Soil Resources Report 48/1-4. Rome: UN Food and Agriculture Organization.

Huxley, P.A. 1985a. The tree/crop interface or simplifying the biological / environmental study of mixed cropping agroforestry systems. Agrofor. Syst. 3: 251-266.

Huxley, P.A. I985b Informal network of agroforestry researchers. ICRAF Newsletter 14.

Labelle, R. 1987. Bibliographic information management using microcomputers: the experience of the International Council for Research in Agroforestry (ICRAF). ICRAF Working Paper 45.

Novoa B., A. R. and J. L. Posner (eds.) 1981. Agricultura de ladera en America tropical. Turrialba, Costa Rica: CATIE.

Siderius,W. (ed.) 1986. Land evaluation for land-use planning and conservation in sloping areas. Wageningen: International Institute for Land Reclamation and Improvement (ILRI). Publ. 40.

Young, A. 1975. Weathered ferralitic soils: their properties, genesis and management. In H.B. Obeng and P.K. Kwakye (eds.), Proceedings of the Joint Commissions I, IV, V, and VIofthe International Society of Soil Science Conference on Savannah Soils of the Sub-humid and Semi-arid Regions of Africa and their Management. Kumasi, Ghana: Soil Research Institute.

Young, A. 1976. Tropical soils and soil survey. Cambridge: University Press.

Young, A. 1984. Land evaluation for agroforestry: the tasks ahead. ICRAF Working Paper 24.

Young, A. 1985. An environmental data base for agroforestry. ICRAF Working Paper 5, revised edition.

Young, A. 1986a. Evaluation of agroforestry potential in sloping areas. In Siderius (1986), q.v.

Young, A. 1986b. The potential of agroforestry for soil conservation. Part I. Erosion control. ICRAF Working Paper 42.

Young, A. 1987a. The potential of agroforestry for soil conservation. Part II. Maintenance of soil fertility. ICRAF Working Paper 43.

Young, A. 1987b. Soil productivity, soil conservation and land evaluation. Agrofor. Syst. 5:277-292.

Young, A. 1988. Methods developed outside the international agricultural research system. In A.H. Bunting (ed.), Agricultural environments: characterization, classification and mapping. Wallingford: C.A.B. International.

Young, A, 1989. Agroforestry for soil conservation. Wallingford: C.A.B. International, and Nairobi.: ICRAF. in press.

Table 5. ICRAF Environmental data base: sites file, input form (Young, 1985).

ICRAF ENVIRONMENTAL DATA BASE: SITES FILE, INPUT FORM.

Table 6 Examples of outputs from the ICRAF Environmental Data Base. A. Summary level.

ICRAF Environmental Data Base: EXPERIMENT STATIONS Summary level

Table 6 B. Detailed level.

ICRAF Environmental Data Base: EXPERIMENT STATIONS Level 2

ICRAF MACHAKOS FIELD STATION