PhD defence Adrie Mukashema

smallholder coffee terroirs of Rwanda - linking local to global trade through geographical characterization 

Coffee is an important crop that provides a livelihood to millions of people living in developing countries and to over 355 thousand Rwandan families in particular. It is the second most traded commodity in the world after petroleum. However, coffee business development is lagging behind other industries such as for example wine in terms of product differentiation. No specific terroirs exist yet, although the origin and provenance of coffee beans becomes more important in the global coffee trade. For Rwanda, this constitutes a challenge of serving the specialty coffee market because traders demand a consistent high quality. Arabica coffee is grown exclusively by smallholder farmers. Such small-scale farming system is a challenge for specialty coffee market because it involves collecting and mixing coffee from many different producers to meet the quantity demand without compromising the quality. Hence the aim of this study was to identify specialty coffee terroirs of Rwanda, characterize them so that they will further serve as tool for policy-making towards protection of geographical origins of specialty coffee in Rwanda.

Mapping coffee terroirs requires relating coffee yield and sensory quality to its local production environment. The first activity was to be able to avail coffee map which will allow us to link coffee with its geographical production. We first developed an expert Bayesian network model to extract the small-scale coffee fields from aerial orthophotos covering more than 99% of Rwanda and on one QuickBird image for the remaining part. The method consists of a stepwise adjustment of pixel probabilities, which incorporates expert knowledge on size of coffee trees and fields, and on their location. The combination of spectral and digital elevation and forest data allowed mapping of coffee fields with an overall accuracy of 87%.

We secondly linked coffee fields with biophysical and socio-economic location factors to identify significant factors underlying coffee field location. We identified 29 potential factors, including demography, and environmental factors such as climate, soil and topography. These factors were reduced to 17 variables explaining 86 % of the total dataset variability, by factor analysis. Using multiple regression analysis, we produced agro-ecological based models used to predict suitable location for growing coffee in Rwanda. These regional models show that 60 % of the actual coffee farms are located in suitable locations.

Current coffee yields in Rwanda are about 0.7±0.4 t ha-1 which is generally below the average yield (1.2 t ha-1) obtained in small scale farming systems worldwide. An empirical analysis of coffee yields was made to determine which local environmental factors are associated with observed coffee yield levels. We third analysed the relationship between current yield level and the production environment. The developed yield models were used to estimate the achievable coffee yields within the current smallholder farming systems. The resulting yield map shows that within Rwandan smallholder coffee systems, yield levels could be up to 40% higher if land management improves given local biophysical conditions. We further analysed the potential effect of projected near future climate change. Combined projected increase of temperature and rainfall revealed that yield levels would potentially increase by almost 60% in 2050.

Fourth, based on the results of annual Cup of Excellence competitions (2008-2015), we determined the geographical origins of specialty coffees. Multiple regression analysis was applied to identify significant biophysical location factors determining coffee sensory quality in Rwanda. The developed regression models were used to identify the geographic origins, their extent and their level of sensitivity to climate change. The study showed that the inherent coffee sensory quality in Rwanda is always determined by specific topographic and soil properties. Direct improvement of these soil factors in often not possible or cost ineffective. Furthermore, can measures such as N-fertilizer applications, intended to increase yield levels lead to a reduced coffee sensory quality. Climate, specifically extreme temperatures, also reduces coffee quality regionally. A plausible climate change scenario for 2050 applied to the developed coffee quality regressions models demonstrated that coffee quality will probably have both regional increases as well as decreases. The net national change is a slight increase of only 1%, suggesting a limited overall climate sensitivity of Rwandan coffee quality for projected climate change.

By integrating Land suitability, achievable yields, and sensory quality potentials, in a fuzzy logic based spatial multi-criteria model, we delineated speciality coffee terroirs of Rwanda and their geographical indicators. We demonstrated also that coffee terroirs of Rwanda are climate sensitive. It is important that farmers in current coffee terroirs that will turn marginal revise their strategies in the face of climate change. This will include adopting new coffee varieties and agro-economic models that are suited to climate change with greater tolerance of high temperatures and altered pest and disease pressures. The government of Rwanda should consider renewing the national the policy accordingly that promote climate change in coffee sector developments and applied research.

Biography

Adrie Mukashema was born in Nyaruguru District, South province of Rwanda, on 24 December 1972. She obtained her Bachelor of Science in Agricultural Engineering in 2003 at the National University of Rwanda (NUR). She was recruited as assistant researcher at PEARL project, a NUR outreach and development program in coffee sector. In 2005, she was recruited as an Assistant Lecturer at NUR and was immediately sent to the Faculty of Geo-Information Science and Earth Observation (ITC), University of Twente, the Netherlands, to pursue a Master of Science in Natural Resource Management with specialisation in Soil information for Land Management. After completing this degree in March 2007, she returned to teaching at NUR and coordinating NUFFIC/NPT/RWA/071 project, a Dutch funded capacity building program at the NUR Centre for GIS and Remote Sensing (CGIS-NUR). In 2010, She was awarded a fellowship from NUFFIC to pursue her PhD studies at the University of Twente. In February, 2013, Adrie was appointed by the Cabinet of Rwanda as Deputy Director General of Rwanda Natural Resources Authority (RNRA), in charge of Department of Forestry and Nature conservation. In March 2016, she returned to University of Twente, to complete her PhD studies. Her research career focuses on land and coffee sector.

See here the link to the dissertation