Resource capture and productivity of agroforestry systems in Kenya
Resource capture and utilisation were studied in two agroforestry systems at the International Centre for Research in Agroforestry (ICRAF) Research Station at Machakos, Kenya. The agroforestry systems examined contained two contrasting tree species, leucaena (Leucaena leucocephala (Lam.) de wit) and grevillea (Grevillea robusta), and the C3 and C4 crops, cowpea (Vigna unguiculata) and maize (Zea mays, Katumani composite). The leucaena-based trial was established in November 1989 and the trees were grown with ten maize crop rows on either side of a pruned hedgerow (HM) or unpruned tree row (LM). A sole maize control (SM) was also grown. Paired sets of treatments were irrigated to eliminate below-ground competition for water (HMI, LMI and SMI respectively). Interception of photosynthetically active radiation (PAR) by leucaena and maize was measured on a row-wise basis in all treatments at 7-10 day intervals using a sunfleck ceptometer. Sap flux was measured for the maize and both pruned and unpruned leucaena using heat balance gauges. Results are presented for the 1992 April-July rainy season. Total PAR interception was 30 % greater in LM and LMI than in the SM and SMI sole maize treatments. However, little more than 30 % of the light intercepted by the LM and LMI systems was captured by the crop component, and competition for light alone reduced maize yields by over 30 %. Total water uptake by the LM leucaena and maize comprised 60 % of the seasonal rainfall (237 mm) as compared to 30 % for sole maize. However, as for light interception, only 30 % of the water transpired in LM was used by the intercropped maize, and competition from the trees for soil water reduced maize yields at distances of over 6 m from the leucaena. The leucaena was more effective at resource capture, yet less efficient in resource utilisation since it exhibited a lower dry matter:radiation quotient and a lower transpired water:dry matter ratio than maize. Thus the leucaena in the agroforestry systems captured more of the resources that could have been used more effectively by the maize, causing the performance of the mixture to be sub-optimal; these results suggest that the two components would be best grown separately. Intensive monitoring of resource capture and use by trees and crops was subsequently transferred to the Complementarity In Resource Use on Sloping land trial (CIRUS). Although it had been intended to study both trials during the long rains of 1993, the leucaena trees were almost completely defoliated by psyllid (Heteropsylla cubana) infestation shortly before the onset of the rains: in subsequent seasons, CIRUS was studied in preference to the leucaena trial as the trees had only partially recovered. CIRUS was designed to investigate the effects of competition and the extent of complementarity between grevillea and associated crops using the following treatments; sole crops (Cg) of cowpea during the short rains and maize during the long rains, dispersed-planted trees with (CTd) and without crops (Td), and across (CTa) or on-contour-planted (CTc) tree rows with crops. Light interception and water use were monitored using a similar measurement regime to that employed in the leucaena trial. Results are presented for the 199213 and 1993/4 short rainy seasons; the failure of the 1993 long rains forced the abandonment of experimental measurements during this season. Light interception by the Td and CTd grevillea increased greatly between the two short rainy seasons. Thus, total seasonal interception of PAR was three times greater in sole cowpea than in sole grevillea during the 1992/3 short rains, but by the following short rainy season was over 50 % greater in the grevillea than in the cowpea. Cumulative interception of PAR by the CTd grevillea and cowpea combined was more than twice that of the sole cowpea and over 40 % greater than that for sole grevillea during the 1993/4 short rains. Experiments involving artificially imposed shade showed that there was no reduction in total above-ground dry matter production in cowpea until 75 % shading was imposed. To quantify the degree of below-ground complementarity in water use between grevillea and cowpea, sap flux was measured using heat balance gauges attached to the stems of young grevillea (10-18 months old), both before and after excavating the crop rooting zone (upper 60 cm of soil) around the stem base. The crop rooting zone was removed to establish the capability of the grevillea to extract water from deeper horizons. After excavation, the trees maintained sap fluxes of up to 85 % of the unexcavated values. During both short rains, soil evaporation was by far the largest component of the water balance in all treatments. However, continued extraction of water by the trees during the dry season greatly increased resource capture~ thus total water uptake was three times greater for the sole trees than for the sole crop when dry season water use was included. During the 1993/4 short rains, water use was greatest in the CTd treatment, in which 25 % of the total seasonal rainfall was transpired by the trees and crops. Although transpiration by the CTd trees exceeded interception losses, the latter may have had a greater effect on crop growth by reducing the total quantity of water available within the system. The existence of below-ground complementarity and the shade tolerance of the cowpea suggest that deep-rooted tree species and certain C3 crops may be combined successfully in the semi-arid tropics, but the sensitivity of crop yield to any reduction in water availability within the system demonstrates the need for caution when implementing such systems. The results obtained are discussed in relation to previous research on intercropping and agroforestry and their implications for the successful adoption of agroforestry systems in the semi-arid tropics.