Supplementary datasets - the African wildland-urban interface

These supplementary excel datasets are provided in association with the following publication: The African wildland-urban interface is widespread and disproportionately poor. Dataset S1 – Accuracy assessment results comparing the performance of four different African WUI maps at multiple different levels of class differentiation. The four WUI maps differ in how they estimate building density. The first (Schug et al. 2023) is the African portion of the best existing global WUI map provided for comparison. The remaining three are developed for Africa in our paper, and rely on solely Google, solely Microsoft or a combination of Google and Microsoft building footprint data (the combined map). We present here user's, producer's and are-adjusted overall accuracies, and show that the combined map outperforms the other WUI maps in almost every respect. For more information, please see the manuscript. Dataset S2 - Country-specific WUI statistics. Here, we present aggregate WUI statistics for each country for the year 2020, as derived from our combined WUI map (the most accurate one). Statistics include WUI area, share of land in the WUI, number of buildings in the WUI, and share of buildings in the WUI. For more information, please see the manuscript. Dataset S3 - Spearman's correlation coefficients, sample sizes and corresponding p-values describing associations between wildland-urban interface variables and socioeconomic and demographic variables at the continental scale. There are three worksheets containing (i) Household and (ii) Individual-level variables of interest from IPUMS International (unit of analysis = GEOLEVEL2 district from IPUMS International), and iii) World Bank variables of interest (unit of analysis = country). Available data from all districts/ countries were pooled together to obtain the continental correlations with the shares of land and buildings in the wildland-urban interface. Dataset S4 - Spearman's correlation coefficients and corresponding p-values describing associations between wildland-urban interface variables and IPUMS-International socioeconomic and demographic variables for the African continent, World Bank income groups (e.g., lower-middle income countries), and African Union regions (e.g., Southern Africa). To calculate these coefficients, available data from all IPUMS GEOLEVEL2 districts within countries of the specified income group/region were pooled together. Dataset S5 - Spearman's correlation coefficients (with corresponding p-values and sample sizes) describing associations between wildland-urban interface variables and IPUMS-International socioeconomic and demographic variables by country. The unit of analysis is the GEOLEVEL-2 district. Districts in the same country are pooled to calculate the corresponding correlations for that country.