Assessing progress towards sustainable intensification

Sustainability might well be the biggest buzz word of our era. The term is fancied, used, misused, framed and distorted by scientists, the private sector, donors and policy makers alike. In the last decade or so, the concept of ‘sustainable intensification’ (SI) has gained momentum and has been put forward as a promising way to address food security and environmental security for a growing world population. However, the meaning and applicability of this popular concept are subject to fierce debate.

I was therefore excited when my thesis supervisor at Wageningen University & Research (WUR) offered me the opportunity to go to Uganda to explore and test a novel way of assessing technologies in the light of SI answering questions such as “What is sustainable intensification?” “How can we assess it?” And “how can we use this knowledge to develop and target suitable technologies?” I was collaborating with scientists from the CGIAR Research Program on Roots, Tubers and Bananas (RTB) in cross-cutting cluster 5.2 on ‘Sustainable Intensification and diversification’ under the program’s Flagship Project 5 on ‘Improved livelihoods at scale’.  Within this cluster the desire for a more inclusive interpretation of SI emerged; one which not only considers productivity and the environmental domain but also the economic, the social, and the human health domain.

The author speaks with a farmer as part of research to assess the impact of a management package to control Banana Xanthomonas Wilt disease in Uganda.

In order to operationalize this, we selected relevant indicators from the SI framework from Musumba et al (2014).  This framework takes a (new) technology as point of departure and then assesses, using specific indicators for each domain, how this technology will deliver (or not) over the different domains making up sustainable intensification.

In order to test the framework, we applied it to the control of Banana Xanthomonas Wilt (BXW) disease in Uganda and I was soon on my way to collect data for this purpose.

Field work

The case study I conducted revolved around assessing the impact of a management package to control BXW disease. The package involves removing infected banana plants using an effective approach called single diseased stem removal (SDSR), in combination with the early removal of male buds to prevent insects from infecting plants with the bacteria, and sterilizing farm tools after contact with diseased plants.

To this end, I used the SI framework to select relevant indicators in each of the five SI domains to assess the impact of the package (the technology). I designed a survey and a small-scale visual bio-physical assessment to collect information on each of these indicators. In addition, we organized four focus group discussions to discuss and verify the findings. The study was conducted in Central and South-west Uganda.   

Focus group discussions were used to discuss and verify the data collected.

Lessons learned

Operationalizing SI, considering the five domains (productivity, environment, economic, social and human health), proved to be useful because it reveals a series of consequences and implications that give a superior insight into what (positive and negative) impacts can be expected over all these areas. These insights enable scientists and end-users to detect trade-offs and can support decision-making on whether or not to invest in a certain technology or not.   

For instance, indicators in the human health domain showed that although the food security of households was generally high, food produced on farm was generallly low in essential micronutrients like iron, zinc, proteins and vitamin A. These kinds of insights allow for the identification of suitable entry points for interventions, in this case the nutritional quality of the food produced, rather than the quantity. This example illustrates the greatest strength of the framework: it enforces taking evaluations of interventions further than just agronomy, thus stimulating interdisciplinarity and revealing insights that might otherwise be overlooked.

There are also some challenges in using the framework. The assessment is data-hungry; for every domain a set of indicators needs to be selected, each requiring data. The kind of data needed varies widely and can be very site-specific, time-consuming and expensive to collect. For instance, the environmental domain requires specific soil, water and/or biodiversity data. Indicators for productivity and the environment tend to be well-defined, with strong sets of metrics and established tools or procedures for data-collection. For other domains however, especially the social domain, this is much less the case. In order to improve the framework, it is essential to collaborate with social scientists. This is especially important since most of the constraints hampering successful implementation of interventions are not solely agronomic, but often socio-economic or cultural in nature.

The case study on BXW control packages in Uganda showed that the framework serves well to assess a farming system in terms of SI, but that it is more difficult to assess the impact of specific SI interventions. This is mainly due to multiple confounding factors that influence the indicators, besides the intervention (the BXW package). In other words, it would be incorrect to conclude that differences shown in radar charts occur as a result of an intervention, unless a causal effect from the intervention on the indicator is proven and the reference point of the system at which the intervention is implemented is well established.

Despite this difficulty, the framework provides a much-needed tool to consider ex ante how interventions “may have direct and indirect effects outside the primary focus” of a project, whether they are positive or negative (Musumba et al., 207, pp. 14). As such, the framework is helpful for designing research on tailoring interventions and assessing the potential impact.

Way forward

Based on the lessons learned, recommendations for future studies carrying out SI assessments can be boiled down to the following four recommendations:

  • On-farm monitoring and yield measurements should be combined with household surveys and focus group discussions. Ideally, the study design should be longitudinal to track differences in time.
  • Identify confounding factors influencing the selected indicators and account for these factors in the study design and analysis.
  • Test whether differences shown in radar charts are caused by the intervention of interest, and whether these differences are significant.
  • Strengthen collaboration with social scientists to further develop and strengthen the social domain.

I am grateful to Anne Rietveld, Katrien Descheemaeker and Godfrey Taulya for their input and feedback on this blog. I also thank the RTB cluster 5.2 team for their support during my thesis and the International Institute of Tropical Agriculture (IITA) for hosting me.

Article contributed by Harmen den Barber, Research Assistant, Plant Production System Group, WUR