Seasonality and nutrition-sensitive farming in rural Northern Ghana.

In rural sub-Saharan Africa, where malnutrition in all its forms is rife, the greatest gap between the availability of foods and the foods needed for a nutritious diet are faced during the 'hunger season'. We investigated what rural households in Northern Ghana would need to grow to ensure year-round availability of a nutrient adequate diet or the income required to fulfil their dietary needs. We applied linear programming to model different scenarios and interventions. Our results provide three major insights. First, considering seasonality is crucial in nutrition-sensitive farming. Ensuring a nutritious diet year-round requires enhanced availability of vegetables and fruits throughout the year. Second, although staple crops do not provide the full range of essential nutrients, increasing their yields allows for a reduction of field size, freeing up space for the production of other foods belonging to a nutritious diet, such as vegetables. Third, small farms are unable to produce sufficient food to cover their needs. They depend on income both from agriculture and other sources, and the availability of types of foods on markets to meet their dietary needs. Our study shows the value of modelling the range of dietary effects from agricultural interventions in a specific context, using a local feasible nutritious diet as a starting point and taking seasonality into account. Supplementary Information: The online version contains supplementary material available at 10.1007/s12571-022-01325-5.

Narrowing yield gaps does not guarantee a living income from smallholder farming-an empirical study from western Kenya.

Crop yields in sub-Saharan Africa need to increase to keep pace with food demands from the burgeoning population. Smallholder farmers play an important role in national food self-sufficiency, yet many live in poverty. Investing in inputs to increase yields is therefore often not viable for them. To investigate how to unlock this paradox, whole-farm experiments can reveal which incentives could increase farm production while also increasing household income. In this study we investigated the impact of providing farmers with a US$ 100 input voucher each season, for five seasons in a row, on maize yields and overall farm-level production in two contrasting locations in terms of population density, Vihiga and Busia, in western Kenya. We compared the value of farmers' produce with the poverty line and the living income threshold. Crop yields were mainly limited by cash constraints and not by technological constraints as maize yield immediately increased from 16% to 40-50% of the water-limited yield with the provision of the voucher. In Vihiga, at best, one-third of the participating households reached the poverty line. In Busia half of the households reached the poverty line and one-third obtained a living income. This difference between locations was caused by larger farm areas in Busia. Although one third of the households increased the area farmed, mostly by renting land, this was not enough for them to obtain a living income. Our results provide empirical evidence of how a current smallholder farming system could improve its productivity and value of produce upon the introduction of an input voucher. We conclude that increasing yields of the currently most common crops cannot provide a living income for all households and additional institutional changes, such as alternative employment, are required to provide smallholder farmers a way out of poverty.

Evaluating combined effects of pesticide and crop nutrition (with N, P, K and Si) on weevil damage in East African Highland Bananas.

Banana weevil (Cosmopolites sordidus, Germar) is a major pest in East African Highland Banana. The influence of crop nutritional status on weevil damage is poorly understood. Nutrient availability affects the nutritional quality of plants for weevils and may affect weevil damage. Here, we evaluate the effect of insecticides alone and in combination with fertilisers (N, P, K and Si) on weevil damage using data from two experiments in central and southwest Uganda. In the first experiment, we varied chlorpyrifos and application rates of N, P and K. In the second experiment, we varied the application rates of K and Si. Treatment effects were analysed using generalised linear mixed models with a negative binomial distribution. In the first experiment, chlorpyrifos reduced and N increased weevil damage, while P and K had no significant effect. In the K or Si application rates reduced weevil damage compared with the control. We conclude that the combined application of chlorpyrifos with K and Si fertilisers can contribute to weevil damage control on sites with low nutrient availability and should form part of integrated weevil management in bananas. Future studies should assess how much reduction in insecticide use is possible in EAHB with judicious input rates.

Mapping spatial distribution and geographic shifts of East African highland banana (Musa spp.) in Uganda.

East African highland banana (Musa acuminata genome group AAA-EA; hereafter referred to as banana) is critical for Uganda's food supply, hence our aim to map current distribution and to understand changes in banana production areas over the past five decades. We collected banana presence/absence data through an online survey based on high-resolution satellite images and coupled this data with independent covariates as inputs for ensemble machine learning prediction of current banana distribution. We assessed geographic shifts of production areas using spatially explicit differences between the 1958 and 2016 banana distribution maps. The biophysical factors associated with banana spatial distribution and geographic shift were determined using a logistic regression model and classification and regression tree, respectively. Ensemble models were superior (AUC = 0.895; 0.907) compared to their constituent algorithms trained with 12 and 17 covariates, respectively: random forests (AUC = 0.883; 0.901), gradient boosting machines (AUC = 0.878; 0.903), and neural networks (AUC = 0.870; 0.890). The logistic regression model (AUC = 0.879) performance was similar to that for the ensemble model and its constituent algorithms. In 2016, banana cultivation was concentrated in the western (44%) and central (36%) regions, while only a small proportion was in the eastern (18%) and northern (2%) regions. About 60% of increased cultivation since 1958 was in the western region; 50% of decreased cultivation in the eastern region; and 44% of continued cultivation in the central region. Soil organic carbon, soil pH, annual precipitation, slope gradient, bulk density and blue reflectance were associated with increased banana cultivation while precipitation seasonality and mean annual temperature were associated with decreased banana cultivation over the past 50 years. The maps of spatial distribution and geographic shift of banana can support targeting of context-specific intensification options and policy advocacy to avert agriculture driven environmental degradation.

Yield gap analysis in dairy production systems using the mechanistic model LiGAPS-Dairy.

The difference between the theoretical maximum (potential) production and the actual production realized by farmers is referred to as the yield gap. The objectives of this study are to develop a mechanistic model for dairy cows that allows yield gap analysis in dairy production systems and to evaluate model performance. We extended and adapted an existing model for beef cattle to dairy cattle, and the new model was named Livestock simulator for Generic analysis of Animal Production Systems-Dairy cattle (LiGAPS-Dairy). Milk production and growth of an individual cow over its entire lifespan were described as a function of the animal's genotype, the ambient climate, feed quality, and available feed quantity. The model was parameterized for Holstein-Friesian cows. After calibration, we evaluated model performance by comparing simulated results and measured results from experimental farms in the Netherlands, which were not used for model calibration. Cows were permanently housed in stables, where the diet consisted of predetermined amounts of concentrates and ad libitum high-quality roughage. The mean absolute error (MAE) for simulated milk production per lactation was 12% of the measured milk production, whereas the MAE for simulated daily milk yields was 19%. The MAE for simulated feed intake per lactation was 10% of the measured feed intake, whereas the MAE for simulated daily feed intake was 19%. The average yield gap for dairy cows was 11% of the potential milk production (YP). Yield gap analysis indicated that for experimental farms in the Netherlands, the difference between YP and feed quality limited milk production (YL) of 1,009 kg fat- and protein-corrected milk was mainly explained by feed intake capacity (33%), protein deficiency (25%), cow weight at the start of experiments (23%), and heat stress (19%). The LiGAPS-Dairy model also indicated the periods during lactation in which these factors affected milk production. In our opinion, the overall model performance is acceptable for permanently housed cows under Dutch conditions. The model needs to be evaluated further for other production systems, countries and breeds. Thereafter, LiGAPS-Dairy can be used for yield gap analysis and exploration of options to increase resource use efficiency in dairy production.

On the contribution of modelling to multifunctional agriculture: learning from comparisons.

In this paper a set of criteria is proposed for the evaluation of the potential contribution of modelling tools to strengthening the multifunctionality of agriculture. The four main areas of evaluation are (1) policy relevance, (2) the temporal resolution and scope, (3) the degree to which spatial and socio-institutional scales and heterogeneity are addressed and (4) the level of integration in the assessment of scientific dimensions and of the multiple functions of agriculture. The evaluative criteria are applied to the portfolio of modelling approaches developed and applied in a joint project of the French research institute INRA and the Dutch Wageningen University & Research Centre. The CLUE-S model focuses on prediction of changes in multifunctional land-use at regional scale, given a set of predetermined scenarios or policy variants, e.g. for ex-ante policy assessment and initiation of discussions on regional development. The two other modelling approaches are complementary and aim to address multifunctional farming activities. The Landscape IMAGES framework generates a range of static images of possible but sometimes distant futures for multifunctional farming activities in a small region or landscape. It supports the exploration of trade-offs between financial returns from agriculture, landscape quality, nature conservation and restoration, and environmental quality. Co-Viability Analysis generates trajectories of states and farming decisions fulfilling a given set of ecological and productive constraints representing a desired and sustainable future. The three modelling approaches differ in their policy relevance, in the ways that spatial and socio-institutional scales are addressed and in their degree of explicitation of interaction between the various functions of agriculture, but jointly cover most of the desired capabilities for assessment of multifunctionality. Caveats were particularly identified in the integration of the socio-institutional dimension and the related heterogeneity. Although the model portfolio did not completely satisfy the demands of the set of evaluative criteria, it is concluded that, due to their complementarities, in combination the three models could significantly contribute to further development and strengthening of multifunctionality.