Modelling changing population distributions: an example of the Kenyan Coast, 1979–2009

Catherine Linard,Caroline W Kabaria,Marius Gilbert,Andrew J Tatem,Andrea E Gaughan,Forrest R Stevens,Alessandro Sorichetta,Abdisalan M Noor,Robert W Snow

doi:10.1080/17538947.2016.1275829

Abstract

ABSTRACTLarge-scale gridded population datasets are usually produced for the year of input census data using a top-down approach and projected backward and forward in time using national growth rates. Such temporal projections do not include any subnational variation in population distribution trends and ignore changes in geographical covariates such as urban land cover changes. Improved predictions of population distribution changes over time require the use of a limited number of covariates that are time-invariant or temporally explicit. Here we make use of recently released multi-temporal high-resolution global settlement layers, historical census data and latest developments in population distribution modelling methods to reconstruct population distribution changes over 30 years across the Kenyan Coast. We explore the methodological challenges associated with the production of gridded population distribution time-series in data-scarce countries and show that trade-offs have to be found between spatial and temporal resolutions when selecting the best modelling approach. Strategies used to fill data gaps may vary according to the local context and the objective of the study. This work will hopefully serve as a benchmark for future developments of population distribution time-series that are increasingly required for population-at-risk estimations and spatial modelling in various fields.

Highlights

While spatial population predictions have been produced for the past or the near future, such predictions are generally produced by applying national population growth rates (Hay et al 2004) or urban/rural population growth rates (Gaughan et al 2013; Noor et al 2014; Stevens et al 2015) to current gridded population databases, and do not include any subnational variation in historical population distribution trends
Results show that Model 1 (M1) was more accurate for the years 1989 and 2009, while Model 2 (M2) was more accurate for 1979, the most temporally distant year from training data
Results are consistent between statistics used (RMSE and mean absolute error (MAE))

Summary

Introduction

Advances in GIS technologies have facilitated the mapping of the spatial distribution of human populations globally at an unprecedented level of detail (Dobson et al 2000; Balk and Yetman 2004; Balk et al 2006; Bhaduri et al 2007; Linard and Tatem 2012; Doxsey-Whitfield et al 2015). With census data and disaggregates population counts within administrative units using a simple or more sophisticated dasymetric method (Dobson et al 2000; Balk and Yetman 2004; Balk et al 2006; Bhaduri et al 2007; Linard and Tatem 2012; Doxsey-Whitfield et al 2015). The gridded population of the World v4 database is the only one using two different rounds of censuses (circa 2000 and circa 2010) and providing subnational variations in population distribution for the recent past (Doxsey-Whitfield et al 2015)

Methods

Results

Conclusion