Calculate population exposed to violent conflict from UCDP GED — exposed_population

The indicator calculates the population exposed to conflict events within a specified buffer distance around violent events in UCDP GED. Per default, the first available WorldPop layer is used to estimate exposed populations for years before the respective year, while the most recent layer is used for years after.

Usage

calc_exposed_population_ucdp(
  distance = 5000,
  violence_types = 1:3,
  years = c(1989:2023),
  precision_location = 1,
  precision_time = 1
)

Arguments

distance: A numeric vector indicating the buffer size around included conflict events to calculate the exposed population. Either of length 1 to apply for all types of events, or discrete values for each category included in violence_types.
violence_types: A numeric vector indicating the types of violence to be included (see Details).
years: A numeric vector indicating for which years to calculate the exposed population. Restricted to available years for UCDP GED. For years not intersecting with available WorldPop layers, the first layer is used for earlier years and the last layer to more recent years.
precision_location: A numeric indicating precision value for the geolocation up to which events are included. Defaults to 1.
precision_time: A numeric indicating the precision value of the temporal coding up to which events are included. Defaults to 1.

Value

A function that returns an indicator tibble with conflict exposure as variable and precentage of the population as its value.

Details

The indicator is inspired by the Conflict Exposure tool from ACLED (see citation below), but differs in the regard that we simply flatten our buffered event layer instead of applying voronoi tessellation.

The required resources for this indicator are:

You may filter for certain types of violence. The coded types according to the UCDP codebook are: value 1: state-based conflict value 2: non-state conflict value 3: one-sided conflict

You may apply quality filters based on the precision of the geolocation of events and the temporal precision. By default, these are set to only include events with the highest precision scores.

For geo-precision there are levels 1 to 7 with decreasing accuracy:

value 1: the location information corresponds exactly to the geographical coordinates available
value 2: the location information refers to a limited area around a specified location
value 3: the source refers to or can be specified to a larger location at the level of second order administrative divisions (ADM2), such as district or municipality, the GED uses centroid point coordinates for that ADM2.
value 4: the location information refers to a first order administrative division, such as a province (ADM1), the GED uses the coordinates for the centroid point of ADM1
value 5: is used in different cases if the source refers to parts of a country which are larger than ADM1, but smaller than the entire country; if two locations are mentioned a representiative point in between is selected; if the location mentioned is an non-independend island; if the location is not very specifically mentioned or in relation to another location
value 6: the location mentioned refers to an entire country and its centroid is used
value 7: If the event takes place over water or in international airspace, the geographical coordinates in the dataset either represent the centroid point of a certain water area or estimated coordinates

For temporal precision there are levels 1 to 5 with decreasing precision:

value 1: if the exact date of an event is known
value 2: if start and enddates for events are of unspecified character, spanning more than one calendar day though no longer than six days
value 3: if when start and end dates for events are specified to a certain week, but specific dates are not provided
value 4: if start and end dates for events are specified to a certain month
value 5: if start and enddates for events are specified to a certain year, but specific dates are not provided

References

Raleigh, C; C Dowd; A Tatem; A Linke; N Tejedor-Garavito; M Bondarenko and K Kishi. 2023. Assessing and Mapping Global and Local Conflict Exposure. Working Paper.

Examples

# \dontrun{
if (FALSE) {
  library(sf)
  library(mapme.biodiversity)

  outdir <- file.path(tempdir(), "mapme-data")
  dir.create(outdir, showWarnings = FALSE)

  mapme_options(
    outdir = outdir,
    verbose = FALSE,
    chunk_size = 1e8
  )

  aoi <- system.file("extdata", "burundi.gpkg",
    package = "mapme.biodiversity"
  ) %>%
    read_sf() %>%
    get_resources(
      get_ucdp_ged(version = "22.1"),
      get_worldpop(years = 2000)
    ) %>%
    calc_indicators(
      conflict_exposure(
        distance = 5000,
        violence_types = 1:3,
        years = 2000,
        precision_location = 1,
        precision_time = 1
      )
    ) %>%
    portfolio_long()

  aoi
}
# }