Young trees intercept relatively little of the incident radiation. It was shown experimentally that crops of strawberries growing between 2- and 3-year-old rows of apple trees at normal commercial spacing did not reduce the apple yields and, in themselves, yielded as much as when grown in the open.

Introduction

In previous papers (Jackson 1983 and this volume, Section 2) simple general method of determining the proportions of the available light which will be intercepted by the tree-crop and ground-crop components of agroforestry systems was discussed. The possibility of maximizing overall productivity by combining trees of high transmissivity but high light requirements for development of their economic product with ground crops with lower light requirements was also considered.

In this context the area of crop-producing ground which can be shaded to any extent by the trees is considered as a tree/crop interface. For convenience, although trees may bear crops themselves they are, in this paper, simply referred to as trees unless their economic product is specifically being considered and the term 'crop' is used for the ground-cover crop beneath or between the trees.

In order to optimize the integrated agroforestry system it is essential to know the proportion of the total light which will be available to the ground crop(s) and its distribution over the cropped area. It may well be preferable to have the entire crop surface irradiated as uniformly as is possible, or to have a system with more shade-tolerant ground crops planted nearest to the trees.

Trees, however, take a long time to grow. Moreover, the number of potential combinations of tree and ground crops is large, the arrangements in which they could be combined is almost infinite and the size of factorial experiments on agroforestry systems such as to limit the number of combinations which it is practicable to test. A modelling approach to the effects of different arrangements and densities of tree canopies on the light penetrating to the 'agroforest' floor thus has many advantages.

The results of such modelling of light distribution in agroforestry systems can then be used in conjunction with experimental data on shade responses of the different crops, their seasons of growth and cropping etc, as a guide both to the direct planting of agroforestry systems and to further experimentation on these.

Calculation of light distribution over the crop in agroforestry situations

The general equation that defines the proportion (T) of light transmitted to the crop is:

T=T_f + F_max e^-kl                 (1)

where Tf is the fraction of the available light which misses the trees altogether and would reach the crop even if the trees were totally non-transmitting (i.e., solid and opaque); Fmax is the fraction of the available light which would be intercepted by the trees if they were non-transmitting; L' is tree LAI (m leaf per m total ground surface) divided by F_max; and K is the light extinction coefficient of the tree canopy.

This equation can be applied to an agroforestry system of any level of complexity. Here it is used in relation to three distinct types of system, the first two being defined in their extreme forms:

1. Multi-storey systems in which the trees form a closed canopy through which light penetrates to the crop beneath. Separation between tree and crop is thus primarily in the vertical dimension. The term T_f in equation (1) is zero,F_max is 1 and light transmission to the crop will be controlled by vertically-summed tree LAI and the relevant light extinction coefficient K.

2. Row-and-alley systems in which rows or belts of trees which are so dense that no crop can grow under them are separated by clear alley-ways in which crops can be grown. In this case T = T_f , because there is effectively no transmission through the trees. The light energy available to the crop is a function of the pattern of (solid) shadows cast by the belts or rows of trees which, in turn, is dependent on tree height, latitude, row orientation and time of day and season. Scattered, very dense trees casting 'solid' shadows provide a variant of this type.

3. Intermediate systems in which transmission to the crop is both of light which bypasses the trees altogether (T_f ) and light which passes through the tree canopies (T_c) which is calculated as F_max e^-kl  as in equation (1). Almost all multi-storey systems and row-and-alley systems are likely to be intermediate systems in their early years before the tree canopies grow together (multi-storey systems) or become so dense that crops can only be grown in the alleys (row-and-alley systems).

Multi-storey systems

The key question in such systems is likely to be whether they are indefinitely sustainable as agroforestry systems, i.e. whether light transmission will not ultimately become too low for crop production under the trees. If the system is not indefinitely sustainable the question becomes that of the age at which either the canopy has to be thinned or cropping under it changed to more shade-tolerant crops or abandoned. For calculation of these limits it is essential to measure K for the tree species (see Jackson 1980,1983), the rate of increase in LAI and to know the light requirements of the under-tree crop.

Row-and-alley systems

A computer program to calculate cast shadows from non-transmitting hedgerows of varying size, geometry and orientation at different latitudes was used in relation to orchard system design (Jackson and Palmer 1972) and subsequently modified to deal with transmission through the canopy (Palmer 1977). Methods were also described in these papers for calculating diffuse light interception in row-and- alley systems. These programmes enable calculation of the light intensity at any point on the 'floor' of a hedgerow orchard or a row- and-alley agroforestry system at regular intervals each day throughout the year, anywhere in the world, given the appropriate input data. This includes tree dimensions and shape, alley-way width, latitude and date (from knowledge of which solar altitude and azimuth throughout the day can be determined).

If the shadows are not effectively solid, information is also needed about canopy density and the light extinction coefficient for the canopy. These latter can be treated simply in terms of leaf area or in more detail separating out effects of branches and fruits as light-intercepting structures. Where there is a close relationship between leaf area and the dimensions of branches etc, which is often the case, then calculations based on leaf area and extinction coefficients determined in situ in relation to measured leaf area are likely to be satisfactory.

Whereas the light penetrating through a continuous upper-storey canopy is little affected by latitude, solar altitude and azimuth, these factors are very important with respect to T_f in the discontinuous canopy, row-and-alley situation and interact strongly with row orientation. As an illustration, the effects of all combinations of the following were investigated, assuming non- transmitting belts of trees with vertical sides where they abut on the alleys:

1. Ratio of tree height to width of clear alley between adjacent belts of trees:-0.25:1,0.5:1,1:1,2:1;

2.  Row orientation:- N-S, E-W, SE-NW;

3. Time of year:- 21 June, 21 September, 21 December;

4. Latitude:- 0 (equator), 30N, 50N.

For latitudes in the southern hemisphere the data for 21 June and 21 December must be reversed: a SW-NE row orientation gives similar results to the SE-NW orientation. Light levels across the alley were first calculated for totally diffuse conditions, assuming a Standard Overcast Sky, and then integrated over the day for clear, sunny conditions using the levels of diffuse and direct light in relation to solar altitude given by Monteith (1969) and calculating cast shadows on the assumption that the tree rows or belts were flat-topped and vertical-sided, i.e, of rectangular cross-section. The computer programme assumes that the diurnal pattern of irradiance is symmetrical about true solar noon.

Figure 1 gives the variation in light levels across the clear alley-way under diffuse conditions. The term 'clear alley-way' is used to indicate the alley between the (assumed vertical) edges of the tree canopies, not the distance between tree trunks on each side of the alley. As the ratio of tree height to alley-way increases not only does the relative irradiance received on the alley-way decrease but there is much less variation across it.

% IRRADIANCE LEVELS ACROSS THE ALLEYWAY

UNDER DIFFUSE CONDITIONS

Figure 1 Irradiance at different positions across the alley surface between two hedges, bordering the alley at 0 and 1 respectively, with 4 ratios of hedge height to clear alley width (0.25:1,0.5:1,1:1 and 2:1). Diffuse light only.

Figures 2, 3 and 4 give the patterns of irradiance across the alley-way under clear, sunny conditions for the three row orientations for each latitude and time of year combination. The patterns for the N-S rows show little effect of time of year.

The effect of latitude is not very great but is consistent, the lower the latitude the greater the relative irradiance received on the alley and the greater the proportion of it irradiated to above any particular percentage of full daylight. At a ratio of tree height to alley width of 1:1 or greater, the irradiance across the alley is remarkably uniform, light levels adjacent to the trees being but little lower than those in the centre of the alley. This suggests that an alley crop which does not have too high a light requirement could be produced to a uniform standard under these conditions. Where the above- canopy irradiation levels are high the alley-way levels should be such as to produce good crops of shade-tolerant plants such as tropical beans and cassava (Jackson 1988).

Figure 2 As in Figure 1, at 3 latitudes and 3 times of year. Full sun; hedgerow orientation E - W.

Figure 3 As in Figure 2. Full sun; hedgerow orientation N -S.

With E-W rows, the pattern is altogether more complex, depending on latitude and time of year. In equatorial regions there is obviously a short growing season around 21 September when the entire alley width is highly irradiated, with half of it being well irradiated from June to September and the other half of it from September to December even when row height equals clear alley width. This opens up obvious possibilities for growing even high-light-requirement, short-season grain crops over half of the alley in row- and-alley systems as long as water supply is adequate.

Rows oriented SE-NW are intermediate in their effects on light distribution on the alley. As shown in Figure 4, it is only in a limited number of cases that uniform light conditions obtain over the greater part of the alley.

These computer-model results are essentially illustrative. The assumption that the trees are non-transmitting is a 'simplest case' assumption which is effectively true for some tree canopies which have high LAI values and dense habit (e.g., many windbreaks), but much less so for others. Transmission through the canopy would increase the light incident on parts of the alley but would not, except in very extreme cases, influence the basic pattern of variability with latitude, time of year and tree and alley dimensions. This variability is such that it is obvious that a much more detailed analysis is needed before decisions are taken about row orientation and spacing for individual situations. In particular there would be need to:

Figure 4 As in Figure 2. Full sun; hedgerow orientation SE - NW.
 

1. Investigate a wider range of tree height to alley-width ratios;

2. Examine row orientations which might be favoured, for example, for reasons of contour terracing;

3. Define the local radiation climate and take into account the actual ratio of diffuse to direct light and features such as the prevalence of mid-day or afternoon cloud which could affect the relative irradiance from different sections of the sky; and

4. Carry out the calculations for each day, or a day in each week, of the year.

The tree/crop interface an example of minimal competition

In orchards producing apples for commercial sale, the trees are now generally managed so as to maximize the yield of well-coloured, large fruits rather than total yield. Such fruits can only be produced under conditions of high irradiance on the actual fruit surface and on the nearby spur leaves. The level required varies from variety to variety but the minimum light requirement for good quality fruits ranges from 30 to 50% of full daylight. The ultimate target of orchard design when the trees are mature therefore places a high value on the orchard and the trees not being too dense. Even in mature apple orchards the trees seldom intercept more than 70% of the available light. For the first few years after planting light interception is much lower than this, because the trees have not yet attained their full size or canopy density. Typically, a productive modern 5-year-old orchard may intercept only 30% of the available light (Jackson 1980) and a 3-year-old orchard planted as a very intensive 3-row-bed system intercepts a similar amount (Jackson and Middleton 1987).

In most apple-growing countries the trees are grown in hedgerows; the tree-row being maintained free of competing grass and weeds by the use of herbicides while the alleys are grassed.

Benefits and problems arising from elimination of the grass in the alleys are discussed elsewhere but there would seem little reason why a crop could not replace grass in the alley-ways since the grass is cut only for management reasons and not harvested as a crop. In an experiment planted in 1982 a comparison was made between the yields of apples (grown in hedgerows) with herbicided, grassed, or strawberry-cropped alleyways and between strawberries in these alley-ways and in the open.

Apple and strawberry intercropping experiment

Experimental details

The trees (cv Cox's Orange Pippin on M.9 EMLA rootstock with James Grieve pollinators every fourth tree) were planted in the spring of 1982 as two replicates of a 3 x 3 latin square with three treatments: overall herbicide, grass alley-ways, or strawberries (cv Cambridge Favourite). The trees were 3 m apart within rows 4.5 m apart and the strawberries were grown as three rows 1.1 m apart centred in the middle of each alleyway. This enabled the wheels of the tractor to pass between the rows of strawberries to apply the spray chemicals. The ground immediately under the trees was kept clear of grass, strawberries and weeds by herbicides: this herbicide strip being approximately 1.2 m wide.

Apple fruit yields and quality.
Results are given in Table 1.

Table 1 Effects of alley cropping on the yield and fruit quality of Cox's Orange Pippin apple.

*The strawberries and grass in the alleys were treated with herbicide in spring 1985 so the
  yields in that year show residual effects only.

The only significant effect on total yield was in 1984 when total yield was lower with grassed alley-ways than with overall- herbicide treated alleys, but this effect was not seen in the yield of Class 1 fruit due to the improvement in fruit colour under the grass treatment. This improvement in colour was not evident as a residual effect in 1985 after both the grass and the strawberries were killed by weedkiller.

The effects on the market quality of the apples indicates the potential complexity of intercropping effects. The higher proportion of well-coloured (Class 1 colour) apples when grown with strawberries or grass in the alley-way was probably due to competition by these for soil nitrogen: it being well established that high soil nitrogen reduces fruit colour and that 'grassing down' improves it by lowering nitrogen levels. The reduction in the percentage of top quality apple fruits with respect to skin-finish (% Class 1 russet) in 1984 when apples were intercropped with strawberries represents skin damage resulting from the chemical spray programmme used for strawberries.

Strawberry yields

There were no significant differences between strawberry yields in the open and those in alley-ways between trees. The trend was in favour of strawberries in the open in 1983 but this was reversed in 1984 (Table 2).

Table 2 Effects of growing Cambridge Favourite strawberries in the alley-ways of a young apple orchard as compared with in the open.

These results should not be taken to simply that apple trees and strawberries cannot compete with each other which certainly occurs when the trees are planted close together and the strawberries are grown under them (Atkinson 1975). It does, however, show that when grown with the degree of spatial separation implicit in row-and-alley systems with herbicide use within the tree row the effect of young trees on the alley crop and of the alley crop on young trees can be minimal.

References

Atkinson, D. 1975. Effects of underplanting apple with strawberry. Rep. E. Mailing Res. Station for 1974. p. 55.

Jackson, J.E. 1980. Light interception and utilization by orchard systems. In J. Janick (ed), Hortic.Rev. Vol. 2.

Jackson, J.E. 1983. Light climate and tree-crop mixtures. In P.A. Huxley (ed.). Plant research and agroforestry. Nairobi: ICRAF.

Jackson, J.E. and S.G. Middleton. 1987. Progettazione del frutteto per la massima produzione di qualita. Frutticoltura XLIX N9-10: 27-33.

Jackson, J.E. and J.W. Palmer. 1972. Interception of light by model hedgerow orchards in relation to latitude, time of year and hedgerow configuration and orientation.J. Appl. Ecol. 9: 341- 358.

Monteith, J. 1969. Light interception and radiative exchange in crops stands. In Physiological aspects of crop yield. Madison, Wisconsin: American Society of Agronomy and Crop Science Society of America.

Palmer, J.W. 1977. Diurnal light interception and a computer model of light interception by hedgerow apple orchards. J Appl. Ecol. 14: 601-614.

Footnote__________

*The strawberries and grass in the alleys were treated with herbicide in spring 1985 so the yields in that year show residual effects only.