Modelling the effects of leafing phenology on growth and water use by selected agroforestry tree species in semi-arid Kenya

The WaNuLCAS (Water, Nutrient and Light Capture in Agroforestry Systems) model was used to investigate the impact of tree leafing phenology on the growth and water use of selected agroforestry tree species in semi-arid Central Kenya. Three agroforestry species, grevillea (Grevillea robusta), alnus (Alnus acuminate) and paulownia (Paulownia fortunei), respectively providing evergreen, semi-deciduous and deciduous leafing phenologies, were intercropped with maize. It was hypothesized that the deciduous habit of alnus and paulownia would reduce demand for water relative to the evergreen grevillea under conditions of limited supplies. WaNuLCAS simulations showed that altering leafing phenology from evergreen through semi-deciduous to deciduous decreased water uptake and interception losses by the trees, but increased crop water uptake, drainage and soil evaporation rates for systems containing all three tree species. Drainage and soil evaporation were respectively 14 and 17% greater in the deciduous paulownia system than in the evergreen grevillea. Simulated water uptake and biomass accumulation by grevillea were more than double the corresponding values for paulownia, while crop water uptake in the grevillea and paulownia systems was reduced by 6% and 0.2% respectively relative to sole maize. The simulations imply that water use by paulownia is lower than for grevillea and suggest that leafing phenology is a key attribute affecting water use by trees. The significance of these observations for watershed management and stream flow are discussed.