JGuetschow
commented
2 years ago CRF and other UNFCCC data is now pre-processed using primap2 and the https://github.com/JGuetschow/UNFCCC_non-AnnexI_data repository. This includes mapping of the flags to either 0 (if data is reported elsewhere or there are no emission in the sector) or NaN if data is missing.
Countries have some flexibility in reporting their emission in the CRF data. As a result some sectors have no data for some countries. Flags like "IE" (included elsewhere) or "NA" (not applicable) provide information on the reason for empty time series. However, this information can not be handled by the primap emissions module. Currently these time-series are treated as missing data and are filled with data from lower priority sources leading to potential double counting of emissions from these sectors.
The proper way to deal with this would be to distinguish between the different flags and convert e.g. "IE" to 0 and other flags to 0 or NaN where appropriate. As this is not possible in the current primap software a possible workaround is to consider CRF data as "complete" in the sense that missing data is either included elsewhere or there are actually no emissions in the sector.
For the developing country BUR and UNFCCC data this approach is not feasible as often sectors are missing in the reporting or do not have enough data or subsectoral resolution to be included in PRIMAP-hist. Thus currently we have to fill missing data from other sources.