Impact of Tree Species on Maize Productivity by Smallholder Farmers in Eastern Kenya

Low maize productivity due to declining soil fertility is a major problem faced by smallholder farmers in Eastern Highlands of Kenya. Adoption of agroforestry which has a huge potential to halt land degradation on farms is however, very low due to lack of documented evidence on the effects of the woody perennial tree species on the soil, growth and yield of maize. An on-farm survey was therefore carried out to identify common tree species in four study sites in Eastern Kenya. The main objective was to investigate the effects of the identified common tree species on soil organic carbon, pH, maize growth and yield. A baseline survey was conducted in 50 farms chosen at random from the existing farmers’ groups from the four sites to identify the common tree species growing on farms. Four farms were selected in each site to make a total of 16 farms for the study. Three most prevalent tree species growing within the cropping area on farms were selected in each site. Plots were marked under the canopies of selected individual trees and control plots (at least 10 meters away from any tree). A survey was carried out on soil, crop growth and yield during the long cropping season of 2012 that stretched from March to June. Soil sampling was done at the beginning of the cropping season under the canopies of the selected trees and in control plots to determine the influence of the tree species on total organic carbon (TOC) and pH in the soil on farms. The influence of the trees on maize growth was also examined on plants; selected in plots under the trees and control. The height, basal diameter and SPAD readings (related to the amount of chlorophyll) were examined. Grain yield was determined at the end of the season only at one site (Kyeni) due to seasonal crop failure at the other three sites. Results on soil carbon revealed that at Kyeni Croton macrostachyus Hochst. Ex Delile significantly increased soil carbon than Grevillea robusta A. Cunn. Ex R. Br. Soils sampled under C. macrostachyus had the highest mean (% TOC ) of 2.78 compared to 2.04 for G. robusta plots, however C. macrostachyus plots were not significantly different from the Cordia africana Lam. and the control plots. Senna spectabilis (D.C.) H.S. Irwin & Barneby also had a significantly higher soil carbon value of 2.89 % compared to the control plots that had the lowest value of 2.33 % at Kanwaa. However, no significant differences were found between Senna spectabillis, G. robusta and Vitex payos (Lour.) Merr plots. No significant differences were observed between the tree species at Mweru and Mworoga on their influence on soil carbon. In soil pH analysis, the results revealed that the control plots had a significantly lower pH value than the plots under the trees at all the sites (p < 0.001). Except for the soils sampled under the Grevillea robusta tree species at Kyeni, that revealed significantly lower (p < 0.001) soil pH value of 5.99 compared to Cordia africana (6.42), Croton macrostachyus (6.44) and the control (6.28). At Mweru site G. robusta plots also had significantly lower mean pH value of 5.65 compared to 6.17 for Erythrina abysinnica Lam. and 6.19 for Cordia africana plots (p < 0.001) but was not significantly different from the control plots (5.6). No significant differences were observed between the tree species in influencing soil pH at Kanwaa and Mworoga sites. Tree species and time interactions had a significant influence on plant growth. The trees species at all the sites showed a significant suppression of plant height and basal diameter only at 6 weeks after crop emergence (WACE), (p < 0.05) but not at 2 WACE. At Kyeni the plants in G. robusta plots had the lowest mean basal diameter of 1.67 cm at 6 WACE and 1.96 cm at 9 WACE. A similar trend was observed at Mweru site where the plants in G. robusta plots had the lowest mean basal diameter of 1.77 cm at 6 WACE but not significantly different from the Erythrina abysinnica plots which had a mean of 1.83 cm. No significant differences were observed between the tree species in influencing plant basal diameter at Kanwaa and Mworoga. At Mweru site the lowest mean plant height of 108.8 cm was observed in G.robusta plots compared to 119.4 cm and 156.2 cm for the Erythrina abysinnica and control plots respectively at 6 WACE. At Mworoga site the G. robusta and Senna spectabillis plots showed significantly lower mean plant heights of 143.3 cm and 148.8 cm respectively, when compared to Cordia africana plots which had a mean of 162.2 cm at 6 WACE. No significant differences were observed in plant height in plots under different tree species at Kanwaa and Kyeni. Significant suppression of chlorophyll development (indicated by SPAD readings) was observed in all the tree species at 6 WACE (p < 0.01) at all the sites. No significant differences were observed between the plots under different plant species at Kyeni and Kanwaa. At Mweru site the G. robusta plots had the lowest mean SPAD value of 33.6 % which was significantly different from of C. africana (38.05 %) and Erythrina abysinnica plots (40.87 %). Tree species and distance interactions were found to be significant in influencing SPAD readings in Grevillea robusta plots at Mweru site (p = 0.02) and Senna spectabilis at Mworoga (p = 0.002).