Using yield gap analysis to give sustainable intensification local meaning

Yield gap analysis is useful to understand the relative contribution of growth-defining, -limiting and -reducing factors to actual yields. This is traditionally performed at the field level using mechanistic crop growth simulation models, and directly up-scaled to the regional and global levels without considering a range of factors intersecting at farm and farming system levels. As an example, these may include farmers' objectives and resource constraints, farm(er) characteristics, rotational effects between subsequent crops or decisions on resource allocation and prioritization of crop management. The objective of this thesis is to gain insights into yield gaps from a farm(ing) systems perspective in order to identify opportunities for sustainable intensification at local level.

Three contrasting case studies representing a gradient of intensification and capturing a diversity of agricultural systems were selected for this purpose, namely mixed crop-livestock systems in Southern Ethiopia, rice based-farming systems in Central Luzon (Philippines) and arable farming systems in the Netherlands. A theoretical framework combining concepts of production ecology and methods of frontier analysis was developed to decompose yield gaps into efficiency, resource and technology yield gaps. This framework was applied and tested for the major crops in each case study using crop-specific input-output data for a large number of individual farms. In addition, different statistical methods and data analyses techniques were used in each case study to understand the contribution of farmers' objectives, farm(er) characteristics, cropping frequency and resource constraints to yield gaps and management practices at crop level.

Yield gaps were largest for maize and wheat in Southern Ethiopia (ca. 80\% of the water-limited yield), intermediate for rice in Central Luzon (ca. 50\% of the climatic potential yield) and smallest for the major arable crops in the Netherlands (ca. 30\% of the climatic potential yield). The underlying causes of these yield gaps also differed per case study. The technology yield gap explained most of the yield gap observed in Southern Ethiopia, which points to a lack of adoption of technologies able to reach the water-limited yield. The efficiency yield gap was most important for different arable crops in the Netherlands, which suggests a sub-optimal timing, space and form of the inputs applied. The three intermediate yield gaps contributed similarly to the rice yield gap in Central Luzon meaning that sub-optimal quantities of inputs used are as important in this case study as the causes mentioned for the other case studies.

Narrowing the yield gap of the major crops does not seem to entail trade-offs with gross margin per unit land in each case study. However, the opposite seems to be true for N use efficiency and labour productivity particularly in Southern Ethiopia and Central Luzon, and to a less extent in the Netherlands. This means that (sustainable) intensification of smallholder agriculture in the tropics needs to go hand-in-hand with agronomic interventions that increase land productivity while ensuring high resource use efficiency and with labour-saving technologies that can reduce the drudgery of farming without compromising crop yields.

Other insights at farm(ing) system level were clearer in Southern Ethiopia than in Central Luzon or in the Netherlands. For example, alleviating capital constraints was positively associated with intensification of maize-based farming systems around Hawassa and increases in oxen ownership (an indicator of farm power) was associated with extensification of wheat-based farming systems around Asella. In Central Luzon, farm and regional factors did not lead to different levels of intensification within the variation of rice farms investigated and the most striking effect was that direct-seeding (and thus slightly lower rice yields) was mostly adopted in larger farms, and used lower amounts of hired labour, compared to transplanting. In the Netherlands, the analysis of rotational effects on crop yields provided inconclusive results but confounding effects with e.g. rented land do not allow to conclude that these are not at stake in this farming system.

This thesis broadens the discussion on yield gaps by moving from the technical aspects underlying their estimation towards the broader farm level opportunities and constraints undermining their closure. Overall, insights from contrasting case studies support conventional wisdom that intensification of agriculture needs to occur in the 'developing South', where yield gaps are large and resource use efficiency low, while a focus on improving sustainability based on sustainable intensification (or even extensification) is more appropriate in the 'developed North', where yield gaps are small and resource use efficiency high.