René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[BUG] Resampled/CMORised time axis off by ~1 day for pre-1582 dates (GREGORIAN vs proleptic_gregorian)","articleBody":"## Summary\n\nThe WCRP CMIP6 compliance check **[TIME001] Check Time Squareness** fails on CMORised\nocean output for ACCESS-ESM1-5, with the time coordinate off by exactly one day:\n\n```\n[TIME001] Check Time Squareness\n* Mismatch at index 0: expected 36706.500000, got 36707.500000.\n  (table_id=Oyr, frequency=yr, var=no3, midpoint=True)\n```\n\nThe numeric time values (and their bounds) are computed in the **mixed Julian/Gregorian\n\"standard\" calendar**, but the file ultimately declares **`proleptic_gregorian`**, so for\ndates before the 1582 Julian→Gregorian reform the two disagree by ~1 day.\n\nThe same root cause produces a related **CF §7.1** finding:\n\n```\n'time_bnds' has attr 'calendar' with value 'GREGORIAN' that does not agree with\nits associated variable ('time')'s attr value 'proleptic_gregorian'\n```\n\n## Root cause\n\nACCESS-ESM1-5 model files label their time axis with the **non-CF calendar name\n`\"GREGORIAN\"`**. Two pieces of code interpret this label inconsistently:\n\n1. `base.CMORiser._check_calendar()` rewrites `\"GREGORIAN\"` → `\"proleptic_gregorian\"`\n   (a documented `TODO` workaround for the model's mislabelling). It **changes only the\n   label, not the numeric values** — which is correct, because the model's values are\n   genuinely proleptic-Gregorian (the model has no Julian transition).\n\n2. `utilities.calculate_time_bounds()` and the resampling midpoint helper\n   (`_shift_resampled_time_to_period_midpoint`) read the raw `\"GREGORIAN\"` label and pass\n   it to `cftime`, which treats `\"GREGORIAN\"` as the CF **`standard`** calendar\n   (Julian before 1582-10-15). They therefore build the year/period boundaries — and the\n   yearly midpoint — in the **Julian** frame.\n\nFor year 0101 this gives:\n\n| Quantity | standard / Julian (what we compute) | proleptic_gregorian (what the file declares \u0026 the checker expects) |\n|----------|-------------------------------------|--------------------------------------------------------------------|\n| 0101-01-01 (`days since 0001-01-01`) | 36525 | 36524 |\n| Yearly midpoint | 36707.5 | 36706.5 |\n\nThe 1-day gap is exactly the Julian-vs-proleptic leap-day difference accumulated over\nyears 1–101 (Julian counts year 100 as a leap year; proleptic_gregorian does not).\n\nThe checker validates against the file's declared calendar (`proleptic_gregorian`), so it\nexpects `36706.5` and flags our `36707.5`.\n\n## Impact\n\n- **[TIME001]** fails for all CMORised ACCESS-ESM1-5 output whose time axis falls before\n  1582 (e.g. piControl/historical spin-up years like 0101).\n- **CF §7.1** reports a `time_bnds` vs `time` calendar mismatch (`GREGORIAN` vs\n  `proleptic_gregorian`) for the same reason.\n- Post-1582 data is unaffected (the two calendars coincide), which is why the\n  discrepancy is easy to miss.\n","author":{"url":"https://github.com/rhaegar325","@type":"Person","name":"rhaegar325"},"datePublished":"2026-06-25T11:12:47.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/461/ACCESS-MOPPy/issues/461"}