Gridded population datasets for Latin America and the Caribbean

The important pledge of leaving no one behind and reaching the furthest behind (FB) first has defined development and humanitarian efforts since 2015. As the United Nations Sustainable Development Group (UNSDG) states, “Leaving no one behind (LNOB) is the central, transformative promise of the 2030 Agenda for Sustainable Development and its Sustainable Development Goals (SDGs).” Those being left behind tend to be discriminated against due to different identities and factors, including geographic location. In many cases there are overlaps, with multiple and intersecting forms of discrimination making inequalities worse.
Contemporary and spatially detailed population distribution datasets are critical for properly informing and supporting environmental, health and development applications at local, subnational, national and regional levels, including humanitarian preparedness, adaptation and response. Yet, often population data at the lowest administrative level are not geographically granular enough for targeted planning and localized response.
Given the importance of geographic location as a factor within the efforts to leave no one behind, the aim of this collaboration was to produce high-resolution gridded population datasets for Latin America and the Caribbean, based on the best available subnational population projections disaggregated by sex and age and to the lowest administrative level possible.
The collaboration was led by a team from the United Nations Population Fund (UNFPA) at the Regional Office for Latin America and the Caribbean and UNFPA Headquarters’ Population and Development Branch, the WorldPop Programme, at the University of Southampton together with the German Aerospace Center (DLR).
After a brief introduction the paper presents the study area and lists the countries considered in the framework of this collaboration. It continues describing the methodology and the input data used to produce the high-resolution, model-based estimates. The two main bodies of data that were used are: (1) best available subnational population projections disaggregated by age and sex at the lowest administrative level possible using deterministic cohort-component population projection methods. These projections have been prepared by National Statistics Offices of the region with support from the United Nations Population Fund. These projections, which are used as input data in this paper, have been technically assessed by UNFPA and reviewed by subject-matter country experts with support of the UN Office for the Coordination of Humanitarian Affairs and the UN humanitarian system. These projections have further been adopted as common baseline population data for use in disaster preparedness and operational humanitarian response (2) geospatial (i.e., GIS and Remote Sensing) data, from a wide variety of sources, which are well-known to be related to population presence and density. Regarding the methodology, we describe the Random Forest-based methodology for producing the gridded population distribution datasets, their validation, as well as the collection and processing/preparation of all model inputs.
Then it covers the high-resolution, model-based estimates (https://data.worldpop.org/repo/prj/UNFPA_LACRO/) and lists future work/activities that may be potentially built on the results and methods of this project.