No abstract

If future rice production is to contribute to food security for the increasing population of sub-Saharan Africa (SSA), effective strategies are needed to control weeds, the crop's fiercest competitors for resources. To gain better insights into farmers' access to, and use of, herbicides as part of weed control strategies, surveys were conducted in key rice production locations across SSA. Farm surveys were held among 1965 farmers across 20 countries to collect data on rice yields, farmer's weed management practices, herbicide use, frequencies of interventions and information sources regarding herbicides. Markets were surveyed across 17 countries to collect data on herbicide availability, brand names and local prices (converted to US$ ha −1 ). Herbicides are used by 34% of the rice farmers in SSA, but adoption ranges from 0 to 72% across countries. Herbicides are more often used by men (40%) than by women (27%) and more often in irrigated (44% of farmers) than in rainfed lowland (36%) or upland rice growing environments (24%). Herbicides are always used supplementary to hand weeding. Following this combination, yield loss reductions in irrigated lowlands and rainfed uplands are estimated to be 0.4 t ha −1 higher than hand weeding alone. In rainfed lowlands no benefits were observed from herbicide use. Sixty-two percent of the herbicides sold at rural agro-chemical supply markets are unauthorized. These markets are dominated by glyphosate and 2,4-D, sold under 55 and 41 different brand names, respectively, and at relatively competitive prices (below average herbicide price of US $17 ha −1 ). They are also the most popular herbicides among farmers. For advice on herbicide application methods, farmers primarily rely on their peers, and only a few receive advice from extension services (<23%) or inform themselves by reading the product label (<16%). Herbicide application timings are therefore often (38%) sub-optimal. Herbicide technologies can contribute to reduced production losses in rice in SSA. However, through negative effects on crop, environment and human health, incorrect herbicide use may unintentionally counteract efforts to increase food security. Moving away from this status quo will require strict implementation and monitoring of national pesticide regulations and investment in research and development to innovate and diversify the currently followed weed management strategies, agricultural service provision and communications with farmers.

The demand for rice in Eastern and Southern Africa is rapidly increasing because of changes in consumer preferences and urbanization. However, local rice production lags behind consumption, mainly due to low yield levels. In order to set priorities for research and development aimed at improving rice productivity, there is a need to characterize the rice production environments, to quantify rice yield gaps—that is, the difference between average on‐farm yield and the best farmers’ yield—and to identify causes of yield gaps. Such information will help identifying and targeting technologies to alleviate the main constraints, and consequently to reduce existing yield gaps. Yield gap surveys were conducted on 357 rice farms at eight sites (19–50 farmers per site) across five rice‐producing countries in Eastern and Southern Africa—that is Ethiopia, Madagascar, Rwanda, Tanzania and Uganda—for one or two years (2012–13) to collect both quantitative and qualitative data at field and farm level. Average farm yields measured at the eight sites ranged from 1.8 to 4.3 t/ha and the average yield gap ranged from 0.8 to 3.4 t/ha. Across rice‐growing environments, major causes for yield variability were straw management, weeding frequency, growth duration of the variety, weed cover, fertilizer (mineral and organic) application frequency, levelling and iron toxicity. Land levelling increased the yield by 0.74 t/ha, bird control increased the yield by 1.44 t/ha, and sub‐optimal management of weeds reduced the yield by 3.6 to 4.4 t/ha. There is great potential to reduce the current rice yield gap in ESA, by focusing on improvements of those crop management practices that address the main site‐specific causes for sub‐optimal yields.

Due to land expansion and an increase in productivity, rice production in sub‐Saharan Africa has been growing at a rate of 6% in the past decade. Rainfed rice production systems have accounted for a large share of this expansion. In these systems, the potential growing period not only depends on the length of the rainy season and thus water availability, but is often, especially in the highlands of East Africa, bordered by the onset of the cool period of the year, when low minimum temperatures compromise rice yields. The objective of this study was to investigate the yield potential of 30 rice varieties contrasting in crop duration and cold tolerance in the highlands of East Africa, with its limited length of growing period. A field trial was conducted in the cropping seasons in 2016 and 2017 at the Fogera rice research station, Ethiopia. As a function of the onset of rains, rice was sown mid‐July in 2016 and early July in 2017. Early sowing in 2017 led to an extended crop duration and significantly lower yields of the short‐duration varieties, and to a shortened duration and significantly higher yields of the medium‐ and long‐duration varieties, when compared to late sowing in 2016. Late sowing compromised yield of the medium‐ and long‐duration varieties because of low temperatures during booting stage, which led to high spikelet sterility. Early sowing resulted in low yields of the short‐duration varieties, probably due to low solar radiation during the cloudy rainy season, which coincided with the vegetative stage. Therefore, choice of variety should be a function of the variable onset of the rainy season and related sowing date. However, crop models precisely calibrated for potential varieties and the respective environmental conditions could fully support the selection of a suitable variety, depending on the date of sowing, for example with the help of online tools or smartphone applications.

Accurate modelling of plant development is the basis for any assessment of climate change impact on crop yields. Most rice models simulate development (phenology) based on temperature and photoperiod, but often the reliability of these models is reduced beyond the environment they were calibrated for. In our study, we tested the effects of relative air humidity and solar radiation on leaf appearance rate in greenhouse experiments and analysed data sets from field studies conducted in two extremely different rice‐growing environments in Nepal and Senegal. We also analysed environmental effects on duration to flowering of one popular IRRI material (IR64) for eight different sites covering the entire temperature range where rice is widely cultivated. Both low relative air humidity and low solar radiation significantly decreased leaf appearance rate. Mean air temperature explained 81% of the variation in duration to flowering across sites, which was furthermore significantly influenced by relative air humidity. Across all sites, a simple linear regression approach including mean air temperature and mean relative humidity in the calculation of duration to flowering led to a root mean square error (RMSE) of 10 days, which was slightly lower than the RMSE of 11 days achieved with an automated calibration tool for parameter optimization of cardinal temperatures and photoperiod sensitivity. Parameter optimization for individual sites led to a much smaller prediction error, but also to large differences in cardinal temperatures between sites, mainly lower optimum temperatures for the cooler sites. To increase the predictive power of phenological models outside their calibration range and especially in climate change scenarios, a more mechanistic modelling approach is needed. A starting point could be including relative air humidity and radiation in the simulation procedure of crop development, and presumably, a closer link between growth and development procedures could help to increase the robustness of phenological models.