“…Other types of lower-resolution ancillary data like spectral and/or textural metrics obtained from remotely sensed data (Harvey, 2002;Li & Weng, 2005;Liu, Clarke, & Herold, 2006) or demographic information and distance-to-services metrics (Deng, Wu, & Wang, 2010) have also been applied, but the results obtained in these studies suffered from low accuracies. Impervious surface fractions, on the other hand, have proven to perform equally well as or even better than land-use data as a source for disaggregating population data (Lu, Weng, & Li, 2006;Wu & Murray, 2007;Zandbergen & Ignizio, 2010). Recently, nonparametric modelling has been applied to disaggregate population based on a large number of remotely sensed and other geospatial variables (Patel et al, 2015;Stevens, Gaughan, Linard, & Tatem, 2015).…”