EU age- and gender- stratified population data
This population data set was prepared for the INTARESE and HEIMTSA projects. It includes population data on the emep 50 km x 50 km grid for the years 2000, 2010, 2020, 2030 and 2050. The data is stratified by 5 year age bands and by gender.
Data were derived by making use of different data sources: UN data, CIESIN / SEDAC Gridded World Population and national census data.
A) Census data are available on LAU level 2 for the year 2001. They are stratified by gender and age and used as basis data set for 2000/2001. They give spatial information as well as information on age groups and gender. However, they do not give information about the development in the future. Only data for 23 countries are available. BG, CY, LV, RO, CH, NO and IS are missing.
B) UN data are available by country for the years 1950 to 2050 stratified by gender and 5-year age groups. They were used for filling of information gaps on country totals and gender and age stratification for those countries for which no LAU census data was available. Furthermore, they are used for deriving growth rates of population subgroups for future years. If for some reason not gridded data are needed but country totals, UN data can be taken. They give information on a country level. No further spatial information is available.
C) GWP (Gridded World Population) data are available from CIESIN/SEDAC. They provide gridded data on several resolutions for several regions. Interesting for the INTARESE/HEITMSA study were the data for 2000 and 2010 for a resolution of ½°. GWP data are used for filling of spatial information gaps for those countries for which no LAU census data is available. They furthermore give some feeling for spatial shift of population from 2000 to 2010. No information for the years 2020, 2030 and 2050 is available. No stratification regarding gender or age groups is available.
D) EUROSTAT data and projections are available for all required years. EUROSTAT data, including projections to the future, are used as one basic assumption for the energy modelling, which in turn is an important basis for emission scenario modelling. No stratification regarding gender or age groups is available for future years. Comparisons indicate that EUROSTAT data, including projections, does not differ much from UN data, including projections. Thus, consistency is preserved.
Steps to generate the data sets
Step 1a: Processing LAU census data to fit it to the Emep grid cell
- Filling gaps in the available data sets (e.g. for some countries for some LAU regions only the total number of persons was available, not split by age and gender)
- Filling missing age groups (e.g. for some countries no 5-year age bands were given but e.g. 15-year bands: they were further split up using age group fractions derived from the UN data)
- Intersection with Emep 50 km x 50 km grid
- Summing up per grid cell, age and gender
Step 1b: Filling gaps: Filling data for those countries for which no LAU census data was available
- Using UN data for country totals
- Splitting into subgroups on a country level using UN data (subgroup fractions)
- Area-weigh total population using GWP data (using percentages of grid cells compared to the total GWP population)
UN data are used for country totals as country totals for all sources are relatively small, so there is no reason against using them. Furthermore, UN data country totals and growth rates are used for projections to the future (see step 2). Thus, consistency is preserved.
Step 1c: Summing up data from both sources
- Summing up values for each grid cell from both sources
Step 2: Projections to the future
- Growth rates from UN data (for each subgroup separately) are taken to project the basic data set to the future.
- Result: Data set including for each grid cell the number of persons of each subgroup in the years 2000, 2010, 2020, 2030 and 2050. For 2020, 2030 and 2050, medium, high and low estimates are available.
Growth rates from UN data are taken because i) UN data are taken whenever possible for consistency reasons, ii) UN data have several growth rates (middle, high, low) which gives some kind of uncertainty bounds, and iii) EUROSTAT growth rates fit quite well with the UN data growth rates so there is no inconsistency here.