Error message

This minisite is not configured. Please contact the site administrator.
Share & print

Populations: an aerial view

Les images satellites renseignent sur l’agriculture, l’urbanisation, le climat, la pollution… Cette approche tombe à pic pour faire le lien entre évolutions démographiques et changements environnementaux. Elle peut même dans certaines conditions remplacer un recensement pour simplement compter le nombre d’habitants d’un territoire précis. 

Valérie Golaz est chercheuse à l'Ined au sein de l'unité de recherche DEMOSUD-Démographie des pays du Sud. Elle travaille en particulier sur le  thématiques croisés de population, climat et environnement.

What does satellite technology contribute to the study of interactions between population and environment?

Satellite observation of the earth is delivering ever-more finely detailed data with increasing frequency. Since the satellites survey the entire planet, the data they obtain is a source of increasingly precise information on land use, constructed areas, the climate, and the atmosphere. These types of data have long been used in other disciplines, but it’s only in the last twenty years that they have come to be used more fully in the human and social sciences. The result is a shift in perspective. Whereas at first, the focus was exclusively on the ways in which populations affect the environment (changes in land use, urbanization), research has now turned to the reverse relationship: the role of the environment and environmental change (among other factors) in health- and migration-related inequalities within a given population. 

How do demographers make these data their own?

The accessibility of environmental data like this opens the way for systematic contexualizing of the usual types of demographic data (administrative, survey, and census data, among others). This in turn enables researchers to account more fully and accurately for such things as the characteristics of people’s living environment and environmental characteristics themselves. From a cross-sectional perspective, individual characteristics at a given time can be related to soil categories or air pollution levels, for example, in both urban and rural contexts. From a dynamic perspective—that is, longitudinal—we can now also take environmental change into account in studying individual trajectories over time, even when those trajectories unfold in different places, while no longer needing specific data collection processes. Last, by observing building characteristics in ever closer detail—density, height—satellites contribute data that can be used to estimate population size without necessarily having to conduct exhaustive censuses across the entire planet. Today, one of the most often used products in this field is WorldPop gridded datasets, which provide population estimates by sex and age group from 2000 to 2030. These data in themselves make other types of analysis possible, such as location-specific analyses of exposure to environmental risks.

Can satellite data replace population censuses?

Never completely! For two reasons. First, in terms of population count, even the increasingly accurate estimates we can now access cannot be obtained without censuses, because censuses, even non-exhaustive ones, cover all the different situations on the ground and can therefore be used to develop statistical models and validate results. Moreover, a census does not simply record the number of residents in the administrative units of a given region or country; it also collects individual characteristics together with household and kinship structures. Countries cannot do without censuses; they need the knowledge on the country’s population that censuses can provide. The main advantage of satellite data is that they can produce estimates (ultimately less detailed and potentially less accurate than census-based ones) in areas or countries to which census-takers cannot be sent to work on the ground.

This was the case, for example, with the latest census in Burkina Faso: satellite data compensated for the fact that it was too dangerous at the time to conduct the census in the northern part of the country.

Referens

Basile Rousse, Sylvain Lobry, Géraldine Duthé et al., 2024, "Domain Adaptation for Mapping LCZs in Sub-Saharan Africa with Remote Sensing: A Comprehensive Approach to Health Data Analysis", IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 17: 13016-13029

Attoumane Artadji, Léo Lipovac , Heninjara Hasina Andriamanantena et al., 2024, "Can we estimate sub-Saharan Africa’s population from remote sensing images and land cover mapping?". In: Julie Cardi (Ed.), Mélanie Favrot (Ed.), Bénédicte Gastineau (Ed.) et al., Digressions, Marseille : Laboratoire Population Environnement Développement (LPED), p. 186-194

Darin, E., Kuépié, M., Bassinga, H., Boo, G. et Tatem, A.-J. (2022). La population vue du ciel : quand l’imagerie satellite vient au secours du recensement. Population, 77(3), 467-494