Core Ideas There may be a synergistic effect between the components of conservation agriculture. Reducing tillage resulted in a yield increase at one location but not another. Production was more by intercropping compared with rotation of maize and bean. Conservation agriculture (CA) can be a means to soil improvement and increased crop productivity but had not been evaluated for maize (Zea mays L.)–dry bean (Phaseolus vulgaris L.) cropping systems in the semiarid Central Rift Valley of Ethiopia (CRV). Therefore, on‐farm (2011–2014) and on‐station (2010–2014) trials were conducted to compare CA with the current smallholder conventional practice (CP) for productivity of maize–bean cropping systems. Maize monoculture (MMC), bean monoculture (BMC), maize–bean rotation (MBR), and maize–bean intercropping (MBI) were compared with and without tillage on‐station. In on‐farm research, MMC under CP (CP_MMC) was compared with cropping systems under CA including MMC (CA_MMC), MBR (CA_MBR), and MBI (CA_MBI). On‐station, CA had late tasseling, silking, and physiological maturity compared to CP. CA_MBR and CA_MBI had 28 and 19% more maize grain yield and 29 and 17% more stover yield compared with CA_MMC, respectively. Bean straw yield and intercrop bean grain yield were 13 and 7% more, respectively, with CA compared with CP. However, in the on‐farm trials, maize grain and stover yield were 23 and 47% less with CA_MBR compared to CA_MMC, possibly due to observed soil crusting and compaction of the sandy clay soil with CA. Soil water at 0‐ to 30‐ and 0‐ to 100‐cm depths were 38 and 28%, respectively, more with MBR compared to MMC at the maize grain‐fill stage. Stored soil water was 21% more with CA compared with CP. We conclude that CA_MBR and CA_MBI are suitable for fine texture soil of the CRV.

Publications included in this thesis No publications vi Contributions by others to the thesis John Dimes provided technical and APSIM modelling training and support during first fieldwork planning and proposal writing. He provided also technical support in soil and other minimum data requirements collection for APSIM model calibration experiment at Melkassa Research Centre in Ethiopia. Furthermore, he provided technical backstopping in running participatory modelling workshops and surveys with farmers which were the key research activity for chapter three in this thesis. Peter deVoil has provided technical support in APSIM modelling and analysis; R graphics and data analysis reflected throughout this thesis. Israel Bekele provided technical support in soil sampling, preparation, analysis, and