[avalyset]

What

#1260 derived the NO1–NO5 and SE1–SE4 values in nordic_emissions.json against carbon_intensity_per_source.json as it was then. That table had no biomass or waste factor, so ENTSO-E Waste and Biomass generation stayed in the denominator at factor 0, as the file's method field says.

#1446 added biomass = biofuel = waste = 230. Four Norwegian zones contain ENTSO-E Waste (NO1 also a trace of Biomass), so they were still pricing that generation at 0. This PR recomputes them. Nothing else changes: same pipeline, same full-year 2025 ENTSO-E A75 extract, same rule for the types that still have no factor (Other, Other renewable).

Zone	Before	After	Δ (unrounded)	Coverage before	Coverage after
NO1	26.0	27.0	+0.9659	99.58% (Waste 0.42%, Biomass 0.0004%)	100.00%
NO2	27.0	27.3	+0.2438	99.89% (Waste 0.11%)	100.00%
NO3	25.8	26.0	+0.1975	99.32% (Other renewable 0.60%, Waste 0.09%)	99.40% (Other renewable 0.60%)
NO4	51.7	51.7	0	98.65% (Other renewable 1.35%)	unchanged
NO5	26.6	27.4	+0.7490	99.67% (Waste 0.33%)	100.00%
SE1–SE4	—	unchanged	0	only Other uncovered	unchanged

The arithmetic

The denominator is the zone's total generation either way, so only the numerator moves. The change per zone is the Waste (+ Biomass) energy share times 230:

• NO1: 0.419935% × 230 = 0.9659 → 26.0260 → 26.9919
• NO2: 0.106011% × 230 = 0.2438 → 27.0407 → 27.2845
• NO3: 0.085864% × 230 = 0.1975 → 25.8228 → 26.0203
• NO5: 0.325637% × 230 = 0.7490 → 26.6149 → 27.3639

As a control, the run first reproduces the nine values shipped in #1260 bit for bit, then changes only the two factors. The two runs drop the same intervals, so only the factors differ. NO3 moves above the table's 26 hydro/wind floor; SE1, SE2 and SE4 stay below it because their Other share is still unpriced.

Test

test_nordic_zones_count_no_factored_type_as_uncovered reads each zone's coverage field and fails if a type listed there as uncovered now has a factor in carbon_intensity_per_source.json. On master it fails with ['NO1: Waste', 'NO1: Biomass', 'NO2: Waste', 'NO3: Waste', 'NO5: Waste']; with this PR it passes. The point is that the next factor added to the table can't silently go stale in this file again.

pytest tests/test_emissions.py tests/test_package_integrity.py: 48 passed.

Open questions for you

1. SE4 (and SE1–SE3) Other. SE4's note already says its 14.95% Other is mostly biomass/waste CHP. I have not mapped Other to 230: ENTSO-E reports it as unclassified, and pricing it as biomass would be a guess about its content. With fix: add biomass, biofuel and waste factors to carbon_intensity_per_source #1446 in place it would be a one-line change if you'd rather do that; SE4 would move from 24.7 to roughly 59. Happy to leave as is.
2. provenance still names the Khepri v1.3 record. The 2025 extract and method are unchanged; only the two factors differ. I've recorded the recomputation in notes rather than rewriting the provenance string, but I can point it at a new archived run if you prefer.

biofuel has no ENTSO-E production type, so it doesn't enter these zones.

AI disclosure: implemented with Claude Code.