Assessing the effects of agroforestry practices on biological control potential in kale (Brassica oleracea acephala) plantations in Western Kenya

Agricultural intensification and extension at the expense of forests are known to have a negative impact on biodiversity and ecosystem services. In Sub-Saharan Africa (SSA), agriculture provides livelihoods for most of the population and forestry is an important additional income. Agroforestry practices may help to compensate for loss of biodiversity. Pests are an important constraint to agricultural production and their biological control is affected by environmental degradation. Management practices such as alteration of the field micro-environment and increases in landscape complexity can be efficient non-chemical methods for reducing pest damage. In this study, I examine how the implementation of agroforestry influences biological control in kale (Brassica oleracea var. acephala) plantations in Western Kenya. Twenty fields, either shaded or open, were selected at four settlements in the Trans-Nzoia district of Kenya. Landscape complexity was defined as the percentage of wooded vegetation within a 50 m radius. Pesticide use, mulching practices and the amount of rainfall were recorded as additional explanatory variables. Predation was assessed by (1) egg removal rates, (2) bird exclusion, and (3) predation marks on plasticine caterpillars. Activity density of ground-dwelling predators was assessed with pitfall traps and foliage-dwelling pests and predators were counted on the plants. Aphid parasitism rates were also estimated by counting live aphids and aphid mummies. The effect of shade and landscape complexity, pesticide use, mulching and rainfall was tested by running generalized linear models and selecting the best models based on Akaike Information Criterion adjusted for small sample-sizes. Presence of shade-trees at the local field level was found to be effective in reducing population of both surveyed pests and in increasing aphid parasitism rates and caterpillar predation by birds. Neither the proportion of wooded vegetation in the landscape nor the interaction between local shade and landscape scale tree cover were retained as significant explanatory variables. However, pesticides negatively influenced aphid abundances, predator abundances and egg removal rates while positively influencing caterpillar populations. I conclude that in kale agro-forestry systems of Western Kenya, local management of shade is effective in increasing biological control of pests, but I found no evidence that the proportion of trees in the landscape is important. This can be explained by the small-scale of the analysis and should be investigated with larger scale and stronger differences between agroforestry versus conventional systems. The negative effect of pesticides is also important to note and should be further investigated in order to improve the sustainability of farming practices