GNGTS 2024 - Atti del 42° Convegno Nazionale

Session 2.2 GNGTS 2024 The spaCal layer produced by Pibore et al. (2020) has a variable resoluCon ranging from a few hundred meters in urban areas to several km in rural areas. This layer was developed specifically for earthquake damage and risk assessment purposes, for which the spaCal resoluCon was appropriate. However, in order to perform a risk assessment for fluvial and pluvial hazard, spaCal resoluCon needs to be increased considerably. During this project, we increased the layer spaCal resoluCon in order to produce a residenCal buildings exposure layer on a constant-resoluCon grid. Local partners collected and provided informaCon on the number of buildings or households by Oblast or city and, when available, the number of buildings for each structural typology. For two countries, Kazakhstan and Uzbekistan, local partners provided the informaCon about the number of households by Oblast and load- bearing material (which were associated to EMCA structural typologies). This informaCon was used to update the spaCal distribuCon of building typologies in each country of Central Asia based on naConal-scale data. The method for deriving the final residenCal buildings exposure layer has 4 main phases: The original polygons from Pibore et al. (2020) were classified into urban and rural areas based on the urbanized areas mask provided by the GRUMP dataset (Center for InternaConal Earth Science InformaCon Network - CIESIN, 2021, Figure 2a) For each country and Oblast, the number of buildings in each typology (provided by local partners) was distributed into the urban and rural polygons idenCfied in the previous step (Figure 2b). This method was applied to the countries where local data were available. Buildings were distributed based on the populaCon in the buildings in each polygon (provided by Pibore et al., 2020). The fracCon of different typologies in urban and rural areas of each Central Asia country was extracted from Wieland et al. (2015). For each building type, the total number of buildings was distributed among the sub- typologies (EMCA) based on the relaCve fracCon of each sub-typology in the Pibore et al. (2020) dataset. This operaCon was carried out for each polygon. Finally, residenCal buildings in each polygon were distributed spaCally based on the populaCon layer developed for Central Asia at 100m resoluCon. This allowed increasing the resoluCon and obtaining an equally spaced grid of 100-m resoluCon (Figure 2c). Figure 2 shows examples of the exposure development main steps for Eastern Uzbekistan. Figure 2a shows the urban and rural mask provided by the GRUMP dataset. Figure 2b shows an example of how data provided by locals for each country’s Oblast (e.g., Navoi province, Tajikistan) are distributed on the exisCng variable-resoluCon grid. Figure 2c shows how the data are distributed on the populaCon grid in order to reach higher resoluCon. The result of this procedure is an equally-spaced grid of 100m resoluCon with the number and type of buildings in each sub-typology (EMCA). Finally, the informaCon was aggregated on a regular 500 m grid to beber manage the spaCal data. InformaCon on exposed residenCal buildings is