People living in slums and other deprived areas in low- and middle-income country (LMIC) cities are under-represented in censuses, and subsequently in "top-down" gridded population estimates. Modelled gridded population data are a unique source of disaggregated population information to calculate local development indicators such as the Sustainable Development Goals (SDGs). This study evaluates if, and how, WorldPop-Global (WPG) -Unconstrained and -Constrained “top-down” datasets might be improved in a simulated realistic LMIC urban population by incorporating slum profile population counts into model training. We found that the WPG-Unconstrained model with or without slum training data grossly underestimated population in urban deprived areas while grossly overestimating population in rural areas. SDG 11.1.1, the percent of population living in slums, for example, was estimated to be 20% or less compared to a "true" value of 29.5%. The WPG-Constrained model, which included building auxiliary datasets, far more accurately estimated the population in all grid cells (including rural areas), and the inclusion of slum training data further improved estimates such that SDG 11.1.1 was estimated at 27.1% and 27.0%, respectively. Inclusion of building metrics and slum training data in “top-down” gridded population models can substantially improve grid cell-level accuracy in both urban and rural areas.

Advances in the availability of multitemporal and global built-/human-settlements datasets as derived from Remote Sensing (RS) can now provide globally consistent definitions of “human-settlement” at unprecedented spatial fineness. Yet, these data only provide a time-series of past extents and urban growth/expansion models have not had parallel advances at high-spatial resolution. We present a flexible modelling framework for producing annual built-settlement extents in the near future past last observed extents as provided by RS-based data. Using a random forest and autoregressive temporal models with short time-series of built-settlement extents and subnational level data, we predict annual 100m resolution binary settlement extents five years beyond the last observations. We applied this framework within varying contexts and predicted annual extents from 2010 to 2015. We found that our model framework preformed consistently across all sample countries and, when compared to time-specific imagery, demonstrated the capacity to capture human-settlement missed by the input time-series and validation extents. When comparing building footprints of small settlements to forecast extents, we saw that the modelling framework had a 12 percent increase in ground-truth accuracy. This framework shows promise for predicting near-future settlement extents, and provides a foundation for forecasts further into the future.

Mapping settlement extents at the annual time step has a wide variety of applications in demography, public health, sustainable development, and many other fields. Recently, while more multitemporal urban feature or human settlement datasets have become available, issues still exist in remotely-sensed imagery due to coverage, adverse atmospheric conditions, and expenses involved in producing such feature sets. These challenges make it difficult to increase temporal coverage while maintaining high fidelity in the spatial resolution. Here we demonstrate an interpolative and flexible modeling framework for producing annual built-settlement extents. We use a combined technique of random forest and spatio-temporal dasymetric modeling with open source subnational data to produce annual 100m x 100m resolution binary settlement maps in four test countries of varying environmental and developmental contexts for test periods of five-year gaps. We find that in the majority of years, across all study areas, the model correctly identified between 85-99% of pixels that transition to built-settlement. Additionally, with few exceptions, the model substantially out performed a model that gave every pixel equal chance of transitioning to the category “built” in each year. This modelling framework shows strong promise for filling gaps in cross-sectional urban feature datasets derived from remotely-sensed imagery, provide a base upon which to create future built/settlement extent projections, and further explore the relationships between built area and population dynamics.