Region and Time Conversion
L. Drouet
2024-09-13
Source:vignettes/converting-time-regions.Rmd
converting-time-regions.RmdRegion conversion
In this section, we convert a dataset containing the GDP of all countries into a 17-region mapping of the WITCH model, then into the 5-region mapping used in the SSP database.
First, let’s create the initial database, we will extract the GDP in 2005 from the default_weights. The column names have to be renamed. By convention, column names must be “iso3” and “value”
library(data.table)
library(witchtools)
# Creation of the initial dataset
# Note the column names: iso3 and value
gdp_iso3 <- copy(default_weights[['gdp']])
setnames(gdp_iso3,'weight','value')
gdp_iso3
#> Index: <value>
#> iso3 value
#> <char> <num>
#> 1: AFG 7.543866e+01
#> 2: ALB 3.861787e+01
#> 3: DZA 4.673124e+02
#> 4: AGO 1.979695e+02
#> 5: ARG 8.905475e+02
#> ---
#> 246: BLM 1.000000e-10
#> 247: CYM 1.000000e-10
#> 248: KNA 1.000000e-10
#> 249: MAF 1.000000e-10
#> 250: MSR 1.000000e-10Now, using the function convert_region, we will
aggregate this dataset at country-level (ISO3) to the 17-region mapping
of the WITCH model, available as a default mapping.
# Aggregate value at WITCH regional scale (17 regions)
gdp_witch17 <- convert_region(gdp_iso3, to_reg = 'witch17')
gdp_witch17
#> witch17 value
#> <char> <num>
#> 1: laca 3811.0076
#> 2: sasia 2616.5509
#> 3: ssa 3281.0841
#> 4: europe 22137.4325
#> 5: mena 7253.0736
#> 6: te 8156.9139
#> 7: oceania 1501.8456
#> 8: brazil 2969.7366
#> 9: seasia 4873.5174
#> 10: canada 1746.1213
#> 11: china 22961.7494
#> 12: indonesia 3118.3992
#> 13: india 8377.0372
#> 14: jpnkor 7243.6577
#> 15: mexico 2300.2179
#> 16: usa 19910.1964
#> 17: southafrica 749.2962In a final step, we will convert the 17-region data into the 5-region of the SSP database. The dataset is first disaggregated at iso3 level then it is aggregated into 5 regions, automatically.
# Aggregate value at WITCH regional scale (17 regions)
gdp_r5 <- convert_region(gdp_witch17, to_reg = 'r5')
gdp_r5
#> r5 value
#> <char> <num>
#> 1: r5lam 8973.866
#> 2: r5oecd 52788.486
#> 3: r5asia 44228.718
#> 4: r5maf 11283.454
#> 5: r5ref 5733.314Time conversion
In this section, we convert a yearly time-serie into the default time
period of the WITCH model, 5 year-periods from 2005 to 2150, using the
function convert_time_period.
# Time-serie
dd <- data.table(tech = c("tech1","tech2"), year = 2005:2050, value = 1:46)
# Aggregate values to the default WITCH time period mapping
res <- convert_time_period(dd, 't30')
print(nrow(res))
#> [1] 20
print(res, nrows = 20)
#> tech t value
#> <char> <char> <num>
#> 1: tech1 1 2
#> 2: tech2 1 2
#> 3: tech2 2 6
#> 4: tech1 2 6
#> 5: tech1 3 11
#> 6: tech2 3 11
#> 7: tech2 4 16
#> 8: tech1 4 16
#> 9: tech1 5 21
#> 10: tech2 5 21
#> 11: tech2 6 26
#> 12: tech1 6 26
#> 13: tech1 7 31
#> 14: tech2 7 31
#> 15: tech2 8 36
#> 16: tech1 8 36
#> 17: tech1 9 41
#> 18: tech2 9 41
#> 19: tech2 10 45
#> 20: tech1 10 45
#> tech t valueThe function convert_time_period can also extrapolate
the missing periods.
# Aggregate values to the default WITCH time period mapping with extrapolation
res <- convert_time_period(dd, 't30', do_extrap = TRUE)
print(nrow(res))
#> [1] 60