Agricultural productivity and carbon sequestration potential of a Faidherbia albida (Delile) A. Chev parkland agroforestry system in the Central Rift Valley of Ethiopia

Agroforestry parkland systems, in which a low population density of mature trees occur scattered across farmland, are common land use systems in Ethiopia. Parkland trees, including Faidherbia albida (Delile) A.Chev, have been associated with improved soil fertility and crop productivity and they provide ecological services such as carbon sequestration and biodiversity conservation. However, trees can also have detrimental impacts on understory crops because of competition for resources such as light, nutrients and water. Thus, it is necessary to select suitable tree and crop species as well as management options in order to limit competition and maximize synergies. Agroforestry research and modelling can be applied to understand and accurately predict the effects of tree management on crop growth. This thesis presents research aiming to explore and simulate maize production under different tree and crop management factors in a Faidherbia albida parkland agroforestry system of the Central Rift Valley, Ethiopia, as well as to estimate its carbon sequestration potential. The approach included two field experiments. The first experiment examined how shading affected maize growth and development (using a maize-only field trial with artificial shading). The second experiment focussed on how maize production was impacted by tree pruning, distance from the tree and fertiliser interactions. The APSIM (Agricultural Production Systems simulator) model was used to simulate the effect of these interactions on maize yield. Carbon sequestration in both tree biomass and soil of F. albida trees was estimated by harvesting three average F. albida trees (unpruned trees from the second experiment) and taking soil samples under and beyond tree canopies. Reduced radiation, caused by artificial shading, had a significant negative effect on maize productivity; 50% reduction in incident radiation led to 56% reduction in maize yield, and a 75% reduction in radiation resulted in 64% reduction in yield. This highlighted the need to minimise light competition (shading) in agroforestry systems by managing trees, for example by crown pruning. Soil nutrients and moisture availability were higher under trees than away from tree trunks. This resulted in higher biomass and yield of maize under trees (2-6 m radius of tree canopies) compared to crop-only plots in both growing seasons, and regardless of pruning and fertiliser levels. Overall, there was 56% more grain yield under trees across the two seasons. Fertilisation further increased yields by an average of 15% under tree canopies compared to crop-only plots in both years, but more so in 2016 (when rainfall was higher). Crop biomass and yield were lower under unpruned trees (0‚Äö-2 m) relative to the totally pruned and 50% pruned trees due to reduced light available to crops and in the absence of leaf fall. Faidherbia albida tree density is sparse (5.8 trees ha\\(^{-1}\\)) in the study area. It was estimated that trees stored about 2 t C ha\\(^{-1}\\) in their above-ground biomass and 0.76 t C ha\\(^{-1}\\) below-ground; and there was 34 t C ha\\(^{-1}\\) more in soil (0‚Äö-80 cm depth) under trees than in crop-only areas. However, this rate is low relative to soil C sequestration by other agroforestry systems. This was attributed to the low tree population density in the study area, but could be increased by encouraging farmers to protect planted seedlings or natural regeneration. The APSIM model adequately simulated maize grain yield response to shading (under maize only situations), based on field measurements from the first experiment. The model also reasonably predicted the response of maize yield in low and high rainfall years to tree pruning and fertiliser (N)applications under tree canopies, for which model calibration and validation were based on the second experiment. A virtual experiment indicated that optimal maize yield could be obtained at lower rates of fertilisation (<50 kg N ha\\(^{-1}\\)) under trees than away from them, and that under tree plots (2-6 m) would have the highest yields in most years when fertiliser was not available. High yields under trees were due to increased nutrient and water availability. However, modelled scenarios also highlighted the presence of competition between trees and crops close to tree trunks. For example, plots near to tree trunks (0-2 m) were simulated to consistently produce less maize yield than crop-only plots, which (without trees) also had the highest yields in most years when adequate fertiliser was applied. Thus, we concluded that proper selection and management of agroforestry systems may provide a sustainable alternative for enhancing crop productivity by resource poor farmers. Generally, the study demonstrated that, depending on the extent that trees can improve soil fertility, incorporating them into farmlands could improve crop productivity and deliver ecosystem services such as carbon sequestration. The study also suggested that maize productivity could be improved by crown pruning, and by preferentially applying fertilisers in normal (about average rainfall) and wet (above average rainfall) years. These results need to be considered in a farm-scale livelihood context, which was beyond the scope of the present research. Further research should employ a whole farm modelling approach, in order to understand the combined interactions of agroforestry system components, and to provide recommendations to farmers and policy makers.