GAR15 Global Exposure Dataset for Saint Kitts and Nevis

Main components of the global exposure dataset:
Reference Grid
The so 5x5km reference grid for GAR global exposure dataset includes the whole earth land surface, comprising uninhabited land areas. In this way the reference Grid will be able to handle eventually future data on crops pastures and forest areas. The total number of cells of the grid is 9,008,829. Inhabited cells correspond to 4,574,010. The 5x5km grid size was the choice balancing three criteria of (a) satisfactory size to capture effects for large scale hazards such as earthquake and cyclones at global scale, (b) consistency with the openly available socio-economic datasets with national or global sources, (c) optimizing the computation time Another grid at 30” resolution (around 1x1 km at equator) was set in order to hold exposure data related to coastal areas. The grid was only built for a sector including the first 10 km of coast worldwide.
Boundaries of built-up environment (using BUREF)
The next task is to define the boundaries of human settlements or building stock on the global and identified as urban, sub-urban, or rural. The boundaries of building stock is defined using satellite-imagery of land cover. The Global Built-up Reference Layer (BUREF2010) generated by JRC is a spatial raster dataset containing an estimation of the distribution and density of built-up areas (Pesaresi et al., 2015). It uses publicly available satellite-derived land cover information and per grid population density data to define the percentage of land occupied by buildings per each grid.
Defining the “content” of each grid in exposure dataset using combination of various datasets:

Population distribution
The primary source of global exposure information is the distribution of people on the earth surface. A gridded population dataset is based on a regular grid, where each cell indicates the number of people living on it.
In GEG-2015 development, the new LandScan data published on June 2012 by Oak Ridge National Laboratory was used and refer to the population as of July 2011 at 30” resolution (approx. 1 km equator).

Night time light intensities or Visible Infrared Imaging Radiometer Suite (VIIRS)
The intensities of nighttime lights represents a good proxy of human activities and they were already used at global scale to map economic activity. (Gosh, T. et al., 2010)

Produced capital stock
The economic value of buildings (capital stock) per country is estimated using a dataset for 152 countries from The World Bank (World Bank, 2011) has provides broad estimates of the current (2005) capital stock of machinery and structures, based on the Perpetual Inventory Method (PIM) and historical Gross Capital Formation (GCF) data. Furthermore, the World Bank scale‐up this estimate by 24% to account for the value of Urban Land.

Gross regional product
A raster of Gross Regional Product (GRP) distribution is generated by collecting and assembling all available information for 71 major countries using the following sources: Eurostat: 25 countries
Beijing Normal University: 1 country (China)
OECD: 1 country
World Bank DECRG: 44 countries
The GRP will be further integrated with the outputs from night time light intensities in order to generate a new indicator showing the GDP variation between national and subnational scales. These regional variations of economic activity within a country are used as the basis for geographical distribution of capital stock.

Socio-economic indicators
Socio economic indicators are used as proxies to estimate the use of the building stock for various sectors of commercial, industrial, public, education and health and various economic level for residential sector.

Defining construction classes and distribution
Once the density, values, and sectorial distribution of building stock in each cell are defined, the next step is to define the construction classes and the distribution of various construction classes in each grid.
The World Agency of Planetary Monitoring & Earthquake Risk Reduction (WAPMERR) gathered data on the sub-national distribution of building types for 18 countries using household data from national census as proxies. Countries selected include the largest heterogeneous ones (China, India and Indonesia) and represent 3.6 billion people, about 50% of the total population of the world. Data on characteristics of houses or households are given for residential/nonresidential groups and mainly divided in large urban small urban and rural areas classification. WAPMER developed the dataset for all countries using construction types defined by PAGER, a program of USGS.