Summary

This page summarizes the method used to estimate the population at high risk from climate-related hazards Vision Indicator. The methodology adopts a widely used framework to assess risk in the context of climate change impacts. Risk is the potential for adverse consequences. Risks result from interactions between climate-related hazards with the exposure and vulnerability of the affected population to the hazards.

The hazard is the potential occurrence of a physical event that may cause welfare impacts. Exposure is the presence of people in places that could be adversely affected. Vulnerability is the propensity or predisposition of these people to be adversely affected, or unable to cope with impacts.

People at high risk from climate-related hazards are defined as those exposed to any hazard and vulnerable on any dimension, based on specific thresholds.

Exposure to four climate-related hazards is considered:

Hazard	Return period	Intensity threshold defining an exposed location
Agricultural drought	40 years1	> 30% cropland or pasture affected and rural
Flood	100 years	> 0.5 m inundation depth
Heatwave	100 years	> 33°C 5-day maximum Environmental Stress Index
Tropical cyclone	100 years	≥ Category 2 wind speed

Vulnerability is assessed on seven dimensions:

Dimension	Threshold defining a vulnerable household
Income	Less than $2.15 (2017 PPP) per person per day
Education	No member has completed primary education
Access to finance	No member has a bank or mobile money account
Access to social protection	Does not benefit or contribute to a social protection program
Access to drinking water	No access to improved drinking water
Access to electricity	No access to electricity
Access to services and markets	More than 2km from an all-season road and rural

Five steps to calculate the indicator are summarized below. The following chapters provide more detail.

Step 1: Acquiring hazard, population and vulnerability data

Data from several sources are required to calculate the indicator. Global gridded spatial data is used to determine who is exposed in Step 2. These data sets indicate the number of people living in a given location (grid cell), the degree of urbanization, and the probability and intensity of each hazard event across space. Vulnerability is assessed in Step 3 primarily using data from household surveys - the same surveys used by the World Bank to measure poverty. Access to services and markets is quantified using gridded Rural Access Index (RAI) data. Lastly, spatial data mapping the boundaries of statistical regions represented by surveys is used to merge gridded exposure and survey-based vulnerability estimates in Step 4.

Step 2: Determining who is exposed

The exposed population is estimated by combining global gridded population, degree of urbanization and hazard data. The hazard data is used to identify exposed locations where specific hazard intensity thresholds are exceeded. The hazard and degree of urbanization data is resampled so that grid cells align with the high resolution population data (approximately 90 m). Approximately 90 billion population grid cells covering the globe are then categorized by exposure to any combination of the four hazards according to thresholds, and by eight degree of urbanization categories. As a result, every location and the global population is assigned to one of 128 possible exposure-urbanization categories at a very fine spatial scale.

Step 3: Determining who is vulnerable

Estimating the share of households vulnerable on any dimension requires “fusing” data since information on all dimensions is not available from the same household survey. A simulation method is used to impute indicators derived from other sources, such as access to social protection and finance, into household surveys. The method preserves estimates for each population subgroup reported by other data sources, for example, the share of the poorest rural quintile without access to social protection. The average share of households vulnerable on any dimension across 100 simulations is used to calculate the indicator.

The share of the population vulnerable on the “access to services and markets” dimension is derived from gridded RAI data for each exposure category defined in Step 2. This dimension is incorporated into the calculation of the final indicator in Step 4.

Step 4: Determining who is at risk

To determine who is at risk, the exposure estimates from Step 2 are aggregated to the same level as the representative survey-based vulnerability estimates from Step 3. This involves (1) aggregating the population in each exposure category to survey statistical regions; and (2) aligning rural and urban classifications. With the exposure data aggregated to the same geographic and rural/urban population units as the vulnerability data, the population exposed to any hazard and vulnerable on any dimension is calculated.

Step 5: Calculating global and regional aggregates

Global and regional population weighted aggregates are calculated from the sample of countries with sufficiently recent data on all vulnerability dimensions. For reporting the 2021 global indicator, this includes 103 countries accounting for 86 percent of the population. Aggregates are reported when population coverage is sufficient.

Limitations

The data and methodology have important limitations. Data availability limits country coverage and means the indicator is reported with a lag. The methodology has limitations related to the: selection of thresholds, focus on direct exposure, fusing of vulnerability data from different sources, and assuming uniform vulnerability rates within population units represented by surveys. These are discussed in more detail. Quantifying the risk to people from climate-related hazards is complex, and the current methodology is a first step. It will be improved over time to better address limitations.

---

1

historical frequency based on 39 years of observations.