Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Pays-Bas → 2008 → Msimu wa Kupanda : targeting resources within diverse, heterogenous and dynamic farming systems of East Africa

Wageningen Universiteit (2008)

Msimu wa Kupanda : targeting resources within diverse, heterogenous and dynamic farming systems of East Africa

Tittonell, Pablo Adrián

Titre : Msimu wa Kupanda : targeting resources within diverse, heterogenous and dynamic farming systems of East Africa

Auteur : Tittonell, Pablo Adrián

Université de soutenance : Wageningen Universiteit

Grade : Doctor of Philosophy (PhD) 2008

Résumé
Soil fertility decline is the major single factor explaining the decrease in per capita food production in sub-Saharan Africa. Integrated soil fertility management (ISFM) is an approach to improving or restoring soil productivity, based on combinations of organic and mineral fertilisers, improved germplasm and N2-fixation, but its adoption by farmers has been limited. Smallholder farms in sub-Saharan Africa are highly diverse, heterogeneous and dynamic, and operate in complex socio-ecological environments. Much of the heterogeneity within the farming systems is caused by spatial soil variability. This affects the performance of ISFM technologies, which must be then targeted strategically within heterogeneous farming systems to ensure their propensity to enhance the efficiency of resource (e.g. land, labour, nutrients) use at farm scale. An analytical framework in which systems analysis is aided by survey, experiments and simulation modelling was used to analyse farming futures in the highlands of East Africa. Case study farms from six moderate to high potential agricultural areas in central and western Kenya and eastern Uganda were characterised to identify the diversity of livelihood strategies and understand the main drivers of farm heterogeneity. Constraints to the performance of ISFM technologies and opportunities for efficient targeting of resources within heterogeneous smallholder farms were analysed considering short and long-term horizons, scaling up from field to farm scale, and contextualising livelihood opportunities at regional scale. Across sites, population densities varied from 250 to 1000 inhabitants km-2, which translated in 11 to 4 months year-1 of food self-sufficiency. Based on resource endowment, dependence on off-farm income and production objectives, households were grouped into five Farm Types : 1. Subsidised by off-farm employment ; 2. Market-oriented, cash-crops farms ; 3. Expanding, medium resource endowment farms ; 4. Subsisting, partly on non-farm activities ; and 5. Dependent, wage labourers. Despite their differences in access to resources for soil management, these Farm Types differed more in the degree of soil heterogeneity than in the average fertility status at farm scale. Across sites, soil heterogeneity was smaller on farms owning more cattle. The productivity of maize, the main crop in most of the region, was highly variable within individual farms, strongly influenced by variation in both current crop management (e.g. planting dates, fertilizer rates) and soil fertility (influenced by past soil and crop management). In a classification and regression tree analysis (CART), resource use intensity, planting density, and time of planting were the principal variables determining yield, but at low resource intensity, total soil N and soil Olsen P became important yield-determining factors. Soil heterogeneity also affected crop responses to fertilisers from a maximum of 4.4-fold to –0.5-fold relative to the control in soils varying in organic C and P availability. Across sites in western Kenya, P was the most limiting nutrient for crop production, and P availabilities > 10 mg kg-1 were only measured in soils with > 10 g kg-1 organic C. Such co-variation is induced by day-to-day management decisions farmers make when facing trade-offs in the allocation of their limited resources. A study using inverse modelling allowed analysing tradeoffs of this nature, coupling the dynamic crop/soil simulation model DYNBAL with a Metropolis-type of search algorithm (MOSCEM) and linking crop husbandry practices to labour availability. In a heterogeneous farm, the allocation of fertiliser and labour favoured the fields around the homestead, where the efficiency of nutrient capture was the largest. Productivity could be increased up to a certain threshold beyond which N losses by leaching and soil erosion losses increased abruptly, when fertilisers were applied to the most degraded outfields of the farm. These fields must be rehabilitated through ISFM technologies ensuring organic matter additions, before crops growing on them can respond to nutrient applications. However, the quality of manure common in smallholder farms (e.g., 23 – 35% C, 0.5 – 1.2% N, 0.1 – 0.3% P) and their availability are restrictive. This prevents a quick (hysteretic) soil restoration. Competing uses for crop residues on the farm limit the capacity of fertilisers to restore soil fertility. In simulations using the crop/soil model for long-term dynamics FIELD, which was developed, calibrated and tested against 4 independent datasets, soils receiving combined manure and fertiliser applications over 12 years stored between 1.1 to 1.5 t C ha-1 year-1 when 70% of the crop residue was retained in the field, and between 0.4 to 0.7 t C ha-1 year-1 when only 10% of residues were retained. In mixed crop-livestock systems, crop residues are used to feed livestock, which in turn provide manure to fertiliser crops. When farmers in western Kenya designed ideal farms through participatory prototyping, they emphasised on the importance of such interactions, but tended to overestimate the necessary nutrient flows. A study using the farm-scale model FARMSIM, which integrates FIELD with livestock and manure-cycling models dynamically, showed that although tightly-managed crop-livestock interactions allowed a more efficient use of nutrients brought in the system as fertilisers, the trajectory of change from the current to the ideal farming system is hardly feasible for a majority of farmers.

Mots clés : farming systems / soil fertility / resource allocation / systems analysis / simulation models / cycling / nutrients / nitrogen cycle / soil organic matter / kenya / uganda / africa south of sahara / soil fertility management

Annonce

Présentation

Texte intégral

Page publiée le 27 mars 2008, mise à jour le 3 juin 2022