Extended Data Fig. 4: Influence of site, household wealth and proportion of climate changes observed on the proportion of adaptation responses.
From: Perceived climate change impacts and adaptation responses in ten African mountain regions
Influence of site, household wealth and proportion of climate changes observed on the proportion of adaptation responses used by each household. Graphs show effects estimated from a linear mixed effects model of adaptation as a function of wealth, climate changes observed (fixed effects, both allowed to vary amongst random effect levels), with site as a random effect. (a) Predicted adaptations for each wealth group in each study site. Points show predicted means, with arrows showing 95% confidence limits, with climate changes observed held at the dataset-wide mean. (b) Modelled relationships between proportion of climate changes observed and proportion of adaptation responses used by each household in each study site. Relationships are produced for the average wealth group. Colours as in (a). Letters showing site names are plotted for the mean proportion of changes and proportion of adaptations for each site. BAM: Bamboutos Mountains (Cameroon), ITO: Itombwe Mountains (DRC), KIG: Kigezi Highlands (Uganda), NYU: Nyungwe (Rwanda), KIB: Kibira (Burundi), BAL: Bale Mountains (Ethiopia), KEN: Mount Kenya (Kenya), ABE: Aberdare Range (Kenya), KIL: Mount Kilimanjaro (Tanzania), UDZ: Udzungwa Mountains (Tanzania). N = 150 respondents per site.
