section 5 : results of agroforestry experiments

Light and water availability in fields with and without trees. An example from Nyabisindu in Rwanda

I.F. Neumann

Emerson College
 Forest Row RH 18 5JX, Sussex, UK

P. Pietrowicz

GAT/GTA
 P.O. Box47051, Nairobi, Kenya

Abstract

Trees integrated into fields may compete with the crops for light and water and therefore can reduce yields. In Nyabisindu (Central Plateau in Rwanda, 1700 to 1900m, approximately 1500mm of rain p.a.) it was found from experience that in this place the shading effect is more important in normal years.

The effects of arrangement of trees in the field, tree spacing, and pruning of the canopy on light interception were tested in casestudies in agroforestry systems with up to 8 year-old Grevillea robusta. Shading effects can be controlled efficiently by planting the trees in rows in east-west direction leaving a 'light channel' of 8 to 12 m width in between. 250 to 350 trees per ha may lead to an average light extinction of over 50% in comparison to the open field. By pruning the lower branches of the crown, light extinction can be reduced to about 35%. This degree of shading was tolerated by maize-beans mixed crops. Well maintained rows of trees on the sides of fields (E-W direction) had no significant effect on the humidity of the soil as measured by tensiometers. However, the potential evaporation in the middle of the field was reduced by 36% in average. Plants in the open field showed signs of temporary wilting in hot afternoons, whereas this was normally not the case in the field with trees. The reduced light availability was probably over-compensated by the gain in time for active photosynthesis.

In a drought season, water competition between trees and crop was observed and the drought effects were more severe in the shaded field and close to trees than in the open field.

Introduction

Natural and human environment

Nyabisindu is located in the Central Plateau region of Rwanda, at an elevation of approximately 1700 - 1900 m asl. The soils of the hilly landscape have a moderate-to- poor natural fertility. In normal years about 1000 -1200 mm of rain may be expected during the humid period from September to June (Pietrowicz 1985). This permits two crops per year: the (so called) first or short rains allow cultivation of beans or an early maize (e.g., Katumani compound) while the second or long rains are adequate for a longer-maturing crop. Traditionally Sorghum bicolor is the preferred crop during this season.

The Nyabisindu region is inhabited by approximately 400 persons per km², of whom more than 90% are small-holder farmers producing almost exclusively for their own consumption (Dressier 1984). The traditional forms of land use (shifting cultivation and extensive pastoralism) are no longer likely to provide enough food for the further growing population (Miniplan 1983; Delapierre 1985). Soil degradation and erosion reach alarming dimensions. Also the supply of fuelwood, obviously as important as food crops, is becoming short (Stebler 1983).

A low-input system for sustainable agriculture

These complex difficulties are not easily resolved. However, a precondition to any positive development will be an intensified, ecologically sound, sustainable agriculture that utilizes the locally available resources in an optimal way. We are convinced that diversified low-input systems are superior to agronomic systems with a higher capital turnover, at least under the prevailing local conditions. An improved farming system was developed on the basis of:

1. the natural climax vegetation (i.e., tree savannah in the drier east and semi-deciduous forest in the west);

2. autochthonous farming practices; and

3. experiences gained elsewhere under similar ecological and socio- economic conditions.

This diversified farming system includes agroforestry techniques, intercropping, rotations with green manuring and lays, erosion control, and intensive cattle rearing with minimum grazing, and so forth (Ludwig 1976; Egger 1982; Pietrowicz 1983).

The role of trees in a low-input farming system

The systematic integration of trees into agricultural fields implies material for fodder and mulch or compost, demarcation and stabilization of erosion control stripes (PAP 1984), protection of soil and crop from wind, heavy rains, and excess sunshine, etc. (Neumann 1983; Pietrowicz 1983). However, trees integrated into fields may not have only positive effects but also negative impacts. Competition between tree and crop for light, water, and nutrients is probably the most important limitation to agroforestry.

Under given site conditions the degree of competition will be dependent on the age or size of the tree; the species and its characteristics such as crown shape and density; the shape of the root system; and, last but not least, the density and the arrangement of the trees in the field (Neumann and Pietrowicz 1985; PAP 1984, 1985). Whereas the characteristic properties of a tree species make it more or less suitable for agroforestry use, the management of tree stands plays a decisive role on the performance of an agroforestry system.

Effect of trees on light availability in fields

In demonstration fields with up to eight-year-old Grevillea robusta trees, some management-dependent factors were tested. They included the arrangement of trees in the fields and the modification of the crown shape by pruning.

When the plots were established, trees were planted densely along terrace borders and contour lines separating the 8 - 10 m-wide field strips. It was then observed that crops, especially maize, performed much better in strips oriented east-west than in those oriented north-south. It appeared that this was due to differences of light availability depending on the orientation of a field strip. In the equatorial tropics the sun moves from east to west with little deviation from the zenith. Thus the center of east-west oriented strips received full sunlight throughout the day; while in strips oriented north-south, direct sunshine was limited to a few hours around noon; for most of the day, the fields were shaded by the trees.

Approximately half of the trees were cut. The light interception (as measured with a light meter) * was reduced from 47% to 17% as a result of the cutting. In practical terms, maize scarcely grew under the pre-thinning condition, while it performed well after the thinning.

However, if the contour lines and hence the field strips follow the east-west direction, an average light interception of up to 50% may be tolerated. This average value comprises the area of permanent shade under the trees (up to 80-90% light interception on a sunny day) and the inner field which receives full sunlight throughout the day (center of field: less than 10% light interception). Maize grows well in the inner field, but not under the trees.

Size and shape of the crowns also have important effects on the light availabile to crops grown in adjacent fields. The effects of pruning may be exemplified by a field heavily shaded by Grevillea robusta at a stand density of 634 per ha. The trees were up to 8 years old and unpruned, resembling a forest more than a field. Removing the branches from the lower quarter of the crowns reduced light interception from 72% to 54%. Subsequently a quarter of the trees were felled to obtain an even distribution of 480 trees per ha, resulting in an average light interception of 47%, averaged over the field (Table 1).

Table 1 Sequential opening of a field shaded very densely by Grevillea robusta (634 trees per ha, age up to 8 years, tallest tree 15 m, tree rows in north-south orientation, crowns freely developed); light extinction in % of full light.

In well-spaced and maintained agroforestry stands with 250 to 350 trees per ha of variable age, the average light extinction varies between 20% and 50%, depending on the orientation of the slope and the rows in the field. This degree of shading is tolerated by maize- beans mixed crops.

Effect of trees on water availability in fields

Soil moisture and potential evaporation were compared in fields with east-west oriented, well-maintained rows of trees and neighbouring open fields.

Although trees had no significant effect on the moisture status of the soil as monitored by tensiometers, soil moisture tension appeared slightly higher in shaded fields and closer to the trees.

Trees reduced the effective rainfall by interception. Even the open center of the field received an average of 10% less rain than the open fields. In normal seasons water concurrence between trees and crops was not expressed. During a drought season, however, distinct water concurrence could be observed. The drought effects were more severe in shaded fields and close to the trees than in open fields.

The rows of trees in the agroforestry plots reduced the potential evaporation considerably. In the center of the field, which received sunshine throughout the day, the potential evaporation was reduced by 36%, compared to open fields. During the observation period (49 weeks in 1983) the average weekly evaporation was 25.5 mm in the open and 18.8 mm in the shaded field. This finding is supported by phenological observations of the crops. In the open fields plants showed temporary wilting signs in hot afternoons, whereas this was normally not the case in the fields with trees.

Effect of trees on crop yield

The average harvests obtained in this experiment over five growing seasons are summarized in Table 2. The data are for mixed crops, i.e., maize and beands in the first rains; and maize, soy beans and sweet potatoes in the second rains. While the harvests of maize, beans and sweet potatoes were significantly higher in the agroforestry treatment, soy beans and particularly the weeds suffered from the shade.

Table 2 The effect of tree integration on the harvest of mixed crops (average over five seasons)

Source: Neumann and Pietrowicz 1985.

Conclusions

The case study confirmed that spatial arrangement and management of tree stands are decisive for the performance of an agroforestry system. To limit light interception to a degree tolerated by crops, trees must not shade the agricultural land permanently. On level land, this can be achieved by planting trees only in east-west-oriented rows. On slopes, where this is not possible due to the need for erosion control, trees should be placed at cross points of contour lines (e.g., terrace borders, erosion control ditches) and thought lines in east-west direction leaving open 'light channels.' Also, pruning the lower branches of trees can reduce light interception considerably.

The higher harvest obtained in well-maintained agroforestry plots indicate that slightly reduced light availability (approximately 15-20% in this case study) can be compensated for by a gain in time with active photosynthesis.

The optimal density of trees (number per ha) and the degree of shading of a field is probably very site-specific. The most important determining factors include species, age, size and shape of the trees grown, exposition of the field, light and water requirements of the crops to be grown, temperatures and moisture availability during the growing season; and probably also the nutrient status of the soil.

References

Delapierre, G. 1985. Evolution de la production vivriere et les besoins d'intensification. In Compte-rendu du l^er seiminaire national sur la fertilization des sols au Rwanda.Kigali: MINAGRI

Dressier, J. 1984. Standortgerechter Landbau im tropischen Bergland — Situation and Entwicklungsmdglichkeiten landwirtschaftlicherKleinbetriebe in Rwanda. Dissertation. Hohenheim: Institut fur landwirtschaftliche Betriebslehre.

Egger, K. 1982. Methoden und Moglichkeiten des 'Ecofarming' in Berglandern Ostafrikas. In Giessener Beitrage zur Entwicklungsforschung des Tropeninstituts der Justus Liebig Universitat Giessen. Reihe I.

Ludwig, D.H. 1967. Ukara- ein Sonderfall tropischer Bodennutzungim Raum des Viktoria Sees. Afrika-Studien Nr. 22. Miinchen: Weltforum.

MINIPLAN. 1983. Strategic alimentaire au Rwanda — objectifs, chiffres et programme d'action. Kigali: Ministere du Plan.

Neumann, I.F. 1983. Use of trees in smallholder agriculture in tropical highlands. In Lockeretz (ed.), Environmentally sound agriculture. New York:

Neumann, I.F. and P. Pietrowicz. 1985. Agroforstwirtschaft in Nyabisindu. Untersuchungen zur Integration von Baumen und Hecken in die Landwirtschaft.(Deutsche Fassung: GTZ Eschborn). Etudes et Experiences No. 9. Projet Agro-Pastoral, Nyabisindu, Rwanda.

Pietrowicz, P. 1983. Ausgewahlte Methoden der NutzungsoptimierungimRahmen des standortgerechten Landbaues in Rwanda. In: Giessener Beitrage zur Entwick-lungsforschung des Tropeninstituts der Justus Liebig Universitat Giessen. Reihe I, Symposien Vol. 9.

Pietrowicz, P. 1985. Les sols de la region du Projet Agro-Pastoral. Etudes et Experiences No. 7. Nyabisindu, Rwanda: Projet Agro-Pastoral.

PAP. 1984. L'erosion et la lutte contre I'erosion. Projet Agro-Pastoral de Nyabisindu. Fiche Technique No. 1. Nyabisindu.

PAP. 1985. L'arbre et la haie dans l'exploitation paysanne. Projet Agro-Pastoral de Nyabisindu. Fiche Technique No. 3. Nyabisindu.

Stebler, J.-D. 1983. L'utilisation du bois au Rwanda. Contribution a 1'occasion du l^er seminaiere des energies renouvables a la gestion de 1' energie et de l'environnement du Rwanda, 10-15 January 1983.

Footnote__________

*Light interception was caclulated from lux values determined simultaneously in fields with and without trees.  No data are available on the radiation relevant for photosynthesis.