Urban development and population estimation in Central Africa through remote sensing
Earth and related Environmental sciences
MetadataShow full item record
In Central Africa, population data are not easily collected, and when available, not always reliable. This situation regarding data availability and quality is due both to the lack of means of local institutions for organising such data collection, and to local insecurity and population displacements that can prevent access to remote areas. In this context, remote sensing can contribute to enhance data availability and quality. Within GeoRisCA project (2012-2016), which aims at studying risk due to geo-hazards, two major urban sites located in East Democratic Republic of Congo are studied according to their development and their current population density: Bukavu/Cyangugu and Goma/Gisenyi. Based on satellite imagery analyses, the first objective of our approach is to describe the urban growth and the impact it has in terms of vulnerability of the population to specific geo-hazards (mass movements, seisms and volcanic eruptions). In order to reach this objective, Landsat and SPOT images are used to study the land-cover changes over a 30-year period, allowing also the impact of the major events in the region to be evaluated (displacement of refugees, destruction due to an eruption ). In those rapidly growing cities, informal settlements constitute a large part of the new neighbourhoods. Moreover, formal and informal urban extensions may border on or occur on particular terrains, like lava flows which increases the classification difficulties. In this context, results of OBIA and pixel-based classification are compared in order to define the most suitable approach. The second objective is to assess the population density of the cities of Bukavu/Cyangugu and Goma/Gisenyi by associating morphological classification of urban neighbourhoods (using land-use, built-up density, structure ) and sampled population survey in representative urban blocks. The growth pattern of the city defined in the first phase is included in the neighbourhood classification in order to support distinctions between urban types. The output of this study will be the identification of an efficient methodology for describing the growth of the city and of its densely populated neighbourhood, in order to highlight areas that are more vulnerable and thus more at risk to the studied geo-hazards. Maps produced based on these results could become a tool for urban planning, as well as for risk prevention and disaster management policies.
CitationMichellier, C.; Hanson, E.; Wolff, E. (2014). Urban development and population estimation in Central Africa through remote sensing. , 5th Geobia,