Improving Productivity of Maize-based Smallholder Irrigated Cropping Systems: A Case Study of Zanyokwe Irrigation Scheme, Eastern Cape, South Africa

Smallholder irrigation schemes (SIS) in South Africa have performed poorly and failed to achieve the development objectives of sustaining rural livelihoods by increasing crop production and income. Limited knowledge of crop production among farmers is one constraint to improved crop productivity and, yet little research has been carried out to relate agronomic practices to productivity. This study was aimed at identifying strategies to improve the productivity of maize-based smallholder irrigated cropping systems in South Africa, using Zanyokwe irrigation scheme (ZIS) in the Eastern Cape Province as a case study. The research design was emergent, developed progressively as influenced by the progressive analyses that were made. The research work involved two main phases: exploratory studies and refinement on-farm experiments for testing technologies as informed by the exploratory studies. Exploratory studies were aimed at describing farmer crop production practices in major crop enterprises and identifying and prioritizing major agronomic yield constraints. Findings indicated that the overall performance of the cropping systems was poor as evidenced by low overall yields of < 25% of the potential in maize and butternut. Poor management of weeds, soil fertility, plant populations and water, late planting and poor cultivar choice were the most important yield constraints. The average cropping intensity achieved by farmers was only 48%. New hybrids yielded 50 - 65% higher grain than the cultivars commonly grown by farmers. However, farmers favoured their traditional cultivars, as they preferred green maize production as opposed to dry grains. As a result of these findings, focused research was designed on options for iii improving green maize production, soil fertility, crop stand and weed management in crop production. On-farm trials were, therefore, designed to (1) compare the response to nitrogen (N) rate of direct seeded and transplanted maize; (2) evaluate comparative performance of direct seeded and transplanted green maize under farmer management, and to work out the economics of transplanting; (3) investigate the effects of row spacing and atrazine dosage on weed dynamics maize yield; (4) determine the effects of inter-row spacing and plant population on weed dynamics and maize yield; and (5) investigate the effects of pre-plant weed control, N rate and plant population on weed dynamics and butternut yield. Maize transplanting improved crop stand, had earlier maturity and yielded higher at lower N rates compared to direct seeding for both green and grain maize. Total costs that varied were higher by R3 565 ha-1 for transplanted maize compared to direct-seeded maize, but the net benefits from transplanting were also much higher by R8 773 ha-1. Farmers favoured transplanting, citing bigger cobs, early maturity and the absence of seedling damage by birds, when maize was transplanted. However, a follow-up survey conducted during the subsequent year indicated that five of the six farmers abandoned transplanting, citing shortage of labour and high cost of commercially produced seedlings. It was concluded that transplanting is unlikely to be adopted unless the labour intensiveness of manual transplanting can be solved. Increasing atrazine dosage from a third of the label recommended dosage (LRD) and reducing row spacing from 90 to 45 cm resulted in an increase in weed mortality. A 100% kill of annual broadleaf weeds was possible while iv perennial grasses and the annual broadleaf weed, Oxalis latifolia, could not be controlled even at the LRD. However, improved weed mortality with increase in atrazine dosage and/or narrow rows did not translate into improved maize productivity. In butternut, pre-plant weed control resulted in a six-fold decrease in weed biomass and a 100% increase in fruit yield. These findings demonstrated the possibility of incorporation of reduced herbicide dosages and narrow rows to achieve adequate weed control and optimise on maize yields in smallholder farming systems. The results also suggested that pre-plant weed control is one of the pre-requisite to successful butternut production in SIS in South Africa. Increasing maize population from 40 000 to 60 000 plants ha-1 resulted in a 30% grain yield increase, but decreased length of cobs. Row spacing had no effect on maize yield at 40 000 plants ha-1, but at 60 000 plants ha-1 45 cm rows resulted in 11% higher grain yield (12 547 kg ha-1) than 90 cm rows (11 288 kg ha-1). Butternut yield increased significantly (p < 0.01) with increase in plant population and N rate. The optimum butternut population and N rate were estimated at 25 000 plants ha-1 and 120 kg N ha-1, respectively. Pre-plant weed control in butternut resulted in a six-fold decrease in weed biomass. Increasing butternut planting density from 10 000 to 30 000 plants ha-1 decreased weed biomass by 47%. No marketable fruits were obtained when planting was done without prior weed control. This result suggest that application of non-selective herbicide before planting is one technique that can be used by farmers as part of integrated weed management to give the crop a competitive advantage in the early growth stages before the plants start to produce vines. Yield increased significantly (p < 0.01) with increase in plant density and N rate. The optimum plant density and N rate were estimated at 25 000 plants ha-1 and 120 kg N ha-1, respectively. These findings emphasise the need to take into consideration production practices by farmers as a basis to build up skills in the management of crop enterprises and the farm as a viable business in any efforts to improve on the performance of SIS in South Africa. Future research should concentrate on labour-saving sustainable production technologies for SIS in South Africa