Summary
Transcription error in metadata/haz_classes.csv: poultry_highland THI Extreme threshold was 79 for both THI_max and THI_mean rows. Correct value is 89 (Thornton et al. 2021, Table 2, temperate poultry).
Evidence
- Thornton 2021 Table 2 (PMC9292043, open access): temperate poultry Extreme = 89
- The value 79 matches no Thornton Table 1 or Table 2 entry for poultry
- 79 sat only +3 above the Severe threshold (76), vs ~+10 gap for every other species/zone
- Every other temperate-zone Extreme threshold in the csv matches Table 2 exactly after this fix
Blast radius
R/2 classifies each cell as Extreme when THI exceeds the threshold (raster named THI_max-G<threshold>.tif). With threshold 79, poultry_highland cells were classified at the Severe threshold value instead of Extreme → inflated Extreme exposure in R/3 → published hazard_timeseries_class and VoP aggregates for poultry highland wrong.
Affected pipeline path: R/2 (livestock THI) → R/3 (hazard_exposure) → S3 publish.
Metadata fix (partial — DONE)
Committed 16dce34 + 076a2a3 on develop:
metadata/haz_classes.csv: poultry_highland Extreme 79 → 89 (THI_max + THI_mean rows)
metadata/THI_thresholds_provenance.md (new): full dual-citation doc (Rahimi 2020 formula + Thornton 2021 thresholds), correction record, Moderate/Severe curation note
Remaining work
Recommendation: bundle into next scheduled livestock/hazard re-bake (alongside CR-068 and CR-093 when cglabs run completes).
References
- Thornton, P., Nelson, G., Mayberry, D. & Herrero, M. (2021). Global Change Biology 27(22): 5762–5772. doi:10.1111/gcb.15825. PMC9292043.
- Provenance doc:
metadata/THI_thresholds_provenance.md
Summary
Transcription error in
metadata/haz_classes.csv:poultry_highlandTHI Extreme threshold was 79 for bothTHI_maxandTHI_meanrows. Correct value is 89 (Thornton et al. 2021, Table 2, temperate poultry).Evidence
Blast radius
R/2 classifies each cell as Extreme when THI exceeds the threshold (raster named
THI_max-G<threshold>.tif). With threshold 79,poultry_highlandcells were classified at the Severe threshold value instead of Extreme → inflated Extreme exposure in R/3 → published hazard_timeseries_class and VoP aggregates for poultry highland wrong.Affected pipeline path:
R/2(livestock THI) →R/3(hazard_exposure) → S3 publish.Metadata fix (partial — DONE)
Committed
16dce34+076a2a3ondevelop:metadata/haz_classes.csv:poultry_highlandExtreme 79 → 89 (THI_max + THI_mean rows)metadata/THI_thresholds_provenance.md(new): full dual-citation doc (Rahimi 2020 formula + Thornton 2021 thresholds), correction record, Moderate/Severe curation noteRemaining work
Recommendation: bundle into next scheduled livestock/hazard re-bake (alongside CR-068 and CR-093 when cglabs run completes).
References
metadata/THI_thresholds_provenance.md