This paper analyses the vulnerability of 2 communes in the provinces of Kirundo and Bubanza in Burundi to climate change using social and geographic variables. Primary socio-economic data was collected through surveys where 450 households were involved in responding to a questionnaire. Data collected included demography, livelihood sources, asset ownership, land management practices, social organization and information sources. Initial data consisted of more than 1000 quantitative and categorical variables but these were reduced to the most representative 38. Selected variables were then used in Multiple Factor Analysis (MFA). Results showed that the first two dimensions of MFA contributed most variability to the data and represented social, financial and human capitals. Variables strongly correlated with dimensions were: education; ownership of cattle, improved structures, water inlets or cellphones; use of pesticides and fertilizers; support from Non-Governmental Organizations (NGOs); membership to associations and rainwater harvesting. Clustering was then done based on five dimensions and yielded 2 clusters, which briefly attributed to the households in two provinces. Elevation, land cover and roads were analyzed in ArcMap to explain the clusters’ features. Kirundo (1434m) had higher elevation than Bubanza (969m) but less tree cover with more shrub and grasslands. Bubanza had significantly more croplands. The slopes were moderate to strong. Land cover differences were attributed to human factors and they exacerbated vulnerability. The study concludes that Bubanza is less vulnerable than Kirundo but rural development and institutional policies should be prioritized in both regions to enhance adaptive capacity and reduce vulnerability to future climate-related hazards.
Risper Buyaki Nyairo, Osaka University, Japan
Takashi Machimura, Osaka University, Japan
Stream: Climate change
This paper is part of the IICSEEHawaii2019 Conference Proceedings (View)
View / Download the full paper in a new tab/window
Comments & FeedbackPlace a comment using your LinkedIn profile
Share this Research