Guy Marius Assogba

Rice is cultivated extensively as a food crop in West Africa, but smallholder farmers still face several challenges to meet the current food demand. We observe degradation of land and water resources, low and stagnating yields which threaten farmers’ livelihoods. To enhance the rice production, researchers and decision makers should provide technical and institutional support to smallholder farmers.

This thesis was part of the “Capitalizing the inland valley potential for food and nutrition security for smallholder farmers in West Africa” (CIPA) project which was hosted by the Africa Rice Center (AfricaRice). Côte d’Ivoire and Ghana were the two West African countries selected to implement the project. The objective of this study was to contribute to delivering rice self-sufficiency at national level by improving rice production in inland valleys. This study was realized through a combination of methods including baseline and detailed surveys, on-farm experiments, and scenario analysis.

In Chapter 2, we characterized inland valleys' smallholder farm households at the locations and farm type level using the Rural Household Multiple Indicator Survey (RHoMIS). Data were collected from 733 randomly selected farm households in four locations: Bouaké and Gagnoa in Côte d’Ivoire, and Ahafo Ano North and Ahafo Ano South in Ghana. A farm typology was developed, and farm households were characterized regarding demographic, agricultural, economic, and institutional indicators. Finally, farmers' awareness of different ecosystem services for their food security was examined. We identified four farm types (1) farmers who rent all the land they cultivate, (2) farmers who own some land and rent extra land, (3) farmers who own and cultivate all their land, and (4) farmers cultivating only a part of the land they own. We found that the variation in farm households' demographic, economic, and institutional characteristics was greater between locations. Farmers of all farm types, and in all locations attached high relevance to inland valleys' provisioning ecosystem services, particularly the ability to provide food. A majority of farmers in four locations highlighted the relevance of ecosystem services, including climate regulation, water storage, and groundwater for their wellbeing.

In Chapter 3, the interactions between catena position (valley bottom, lower slope and middle slope) and nitrogen (N) fertilizer rate on rice yield were studied. On-farm experiments were conducted in three inland valleys, from 2019 to 2021. The experiment was a split-plot design, with field position on the catena as main plot and fertilizer rate as the sub-plot. Four fertilizer rates were applied in the first year and eight fertilizer rates in the second year at the three catena positions. We found that catena positions and fertilizer rates had significant effects on grain yield. In the valley bottom and the lower slope, yields were greater than in the middle slope. Water deficit was more pronounced in middle slope than in valley bottom and lower slope. Overall, agronomic N use efficiency was higher in the valley bottom and lower slope (26 kg grain kg-1 N) than at the middle slope (19 kg grain kg-1 N). The highest return on investment from N application was obtained at N rates ≤ 180 kg N ha-1. Overall, the value to cost ratio varied from 1.5 to 6.0 (valley bottom), from 1.3 to 5.4 (lower slope) and from 1.7 to 4.3 (middle slope). We concluded that hydro-edaphic conditions, soil fertility distribution across the catena positions and economic indicators must be considered to guide site-specific fertilizer use in inland valley rice systems.

Chapter 4 aimed at enhancing rice productivity through co-designing improved agronomic practices with farmers. The study was carried out in two years at two locations (Ouragahio and Pangbabo) belonging to two contrasting agroecological zones in Côte d’Ivoire. A participatory co-design approach was employed, involving farmers and researchers through a series of consultations. The co-design process followed four steps: (1) diagnosis of the rice production system, (2) innovation design, (3) on-farm experiments and (4) assessment of the experiments. The experiments consisted of six agronomic practices as treatments: farmers' practices (T1), T1 with transplanting of 14 day-old seedlings (T2), T1 with weed control by herbicides (T3), T1 with recommended fertilizer rates (T5) and a combination of T2, T3 and T5 (T6). Average yield of farmers’ practices across both years ranged from 2.7 to 3.3 t ha-1 in Ouragahio and from 4.3 to 5.4 t ha-1 in Pangbabo. Among individual improved agronomic practices, weed control using herbicide (4.3 t ha-1) and application of the recommended fertilizer rate (5.0 t ha-1) provided a larger yield compared to farmer practice in Ouragahio. In Pangbabo, only the recommended fertilizer rate had a greater yield (6.0 t ha-1) than the farmers’ practice. Regarding labour productivity, farmers’ practices produced an average of 27.7 and 40.4 kg paddy rice per person-day respectively in Ouragahio and Pangbabo. For the improved individual agronomic practices, the labour productivity ranged from 43 to 53 kg paddy rice per person-day. It was above farmers' practices, except for the weed control treatment in Pangbabo. The participatory nature of the co-design process enhanced knowledge sharing but resulted in varied adoption preferences of improved agronomic practices at both locations.

Chapter 5 aimed to understand the current seasonality of labour demands within farm households in inland valleys and to assess the potential of rice mechanization. A detailed household survey was conducted in three locations: (1) Yawtabrikrom in Ghana, (2) Ouragahio and (3) Pangbabo in Côte d’Ivoire to describe the current cropping calendar and labour calendar. The main cultivated crops were rice, maize and cocoa in Yawtabrikrom and Ouragahio. Results highlighted the period of labour competition among crops and the peak of labour demand that occurred over the year. The higher peaks of labour demand varied from 160 to 240 person-hours week-1. Rice is a labour-intensive crop, accounting for 52-68% of the total farm labour allocated to farm activities. To alleviate labour constraints on farm, we explored mechanization options by defining two Scenarios. We have Scenario 1: partial mechanization which is mechanization of harvesting and threshing, and Scenario 2: represents full mechanization of all operations from tilling to threshing. We found that partial and full small-scale mechanization offered the potential to save labour, increase returns to labour in rice production and reduce paddy losses during harvesting and threshing operations. The two mechanization scenarios showed a reduction in labour peaks in all locations, except in Yawtabrikrom.

To contribute to meeting the growing demand for rice at the national level in West Africa, this study recommends strategies of rice intensification in inland valleys by improved and site-specific, adapted agronomic practices and small-scale mechanization to increase yield and labour productivity per area of land. This would lead to more efficient use of the cropped parts of inland valleys, allowing other parts to continue providing the much valued non-agricultural ecosystem services these landscapes have to offer local users.