# Fortolkning — Kristensen Earth Rotation Gauge K⊕ (reading of 2026-11-20)

One interpretation per metric, written once in the `sigil-earth` service and served with the data (`/v1/earth/tips`): **what** it is, **today**'s computed sentence, **how to read it**, what it is **not**, the **K family** link, and a **Danish** line. Regenerated on every publication.

## K⊕ (de-wobbled)  `k_resid`

- **Today:** 2026-11-20: 2.57σ, elevated — between p90 and p99 — watch (p50 1.85, p90 2.45, p99 3.08).
- **I dag:** 2026-11-20: 2.57σ, elevated — mellem p90 og p99 — hold øje (p50 1.85, p90 2.45, p99 3.08).
- **What:** How far today's rotation vector sits from where the known cycles said it would be, in residual σ: the pole (x_p, y_p) after trend + annual + semiannual + Chandler 433 d, and the day length after trend + annual + semiannual, combined in quadrature like a χ with three degrees of freedom.
- **How to read it:** Below 1: Earth is doing what the model expects. 1–3: one of the three is a bit off, usually the day length after a storm season or an ENSO shift. Above the p99 line: something the model does not know about happened — that is the alert.
- **Not:** It is not a forecast of earthquakes or weather, and it is not the raw pole offset. It is the surprise, not the size.
- **K family:** Same shape as the chain's K_fix: a disagreement term (residual/σ) inside a square root. Here the three channels are geophysical, not consensus channels.
- **Dansk:** Hvor langt dagens rotationsvektor ligger fra det, de kendte cyklusser forudsagde, i residual-σ. Under 1: som forventet. Over p99-linjen: noget uforklaret er sket med Jordens ur.

## K⊕ raw  `k_raw`

- **Today:** Raw 2.66 against de-wobbled 2.57: the cycles explain part of the apparent excursion today.
- **I dag:** Rå 2.66 mod af-vaklet 2.57: cyklusserne forklarer en del af dagens tilsyneladende udsving.
- **What:** The same three quantities scored against the plain mean and spread of the trailing year, with no cycles removed.
- **How to read it:** It sits above 1 most of the time because the Chandler and annual wobbles alone move the pole by metres. That is the point: it shows what the de-wobbling takes out.
- **Not:** Not the headline number. A high K_raw with a low K_resid means 'big but expected'.
- **Dansk:** Samme tre størrelser mod det seneste års middel og spredning, uden cyklusser fjernet. Høj K_raw og lav K_resid betyder "stort, men forventet".

## The ladder  `ladder`

- **Today:** Today's reading sits between p90 and p99 — watch. The alert is armed.
- **I dag:** Dagens aflæsning ligger mellem p90 og p99 — hold øje. Alarmen er aktiv.
- **What:** Empirical quantiles of K⊕ over the trailing three years (p50 / p90 / p99 / max), beside the χ₃ values a Gaussian would give (1.54 / 2.50 / 3.37).
- **How to read it:** Compare today's K⊕ to the p99: above it, the alert fires; between p90 and p99, watch. If the empirical p99 sits well above 3.37, the residuals have fat tails and the model is missing a cycle.
- **Not:** Not a fixed scale. It moves as the window moves.
- **Dansk:** Empiriske fraktiler (p50/p90/p99) af K⊕ over tre år, ved siden af χ₃-værdierne for en gaussisk null. Over p99 udløses alarmen.

## x_p  `xp`

- **Today:** 0.115429″ toward Greenwich ≈ 3.6 m.
- **I dag:** 0.115429″ mod Greenwich ≈ 3.6 m.
- **What:** Where the rotation axis pierces the crust, in arcseconds toward the Greenwich meridian (IERS convention). 1″ ≈ 30.9 m on the ground.
- **How to read it:** Follows a ~14-month spiral (Chandler wobble) mixed with an annual loop; the model_xp line is that spiral. The gap between the two is what K⊕ scores.
- **Not:** Not the geographic pole moving in any everyday sense: the whole excursion is a few metres.
- **Dansk:** Polens x-koordinat i buesekunder mod Greenwich-meridianen; 1″ ≈ 30,9 m på jordoverfladen. Afvigelsen fra modellinjen er det, K⊕ scorer.

## y_p  `yp`

- **Today:** 0.331343″ toward 90° W ≈ 10.2 m.
- **I dag:** 0.331343″ mod 90° V ≈ 10.2 m.
- **What:** The other pole coordinate, toward 90° West (IERS convention). In the right-handed frame the page uses, ω_y = −Ω₀·y_p.
- **How to read it:** Read it with x_p as one point on the polhode. The dashed forecast path is the IERS prediction of where that point goes next.
- **Not:** Its sign is a convention; only the distance from the model matters for K⊕.
- **Dansk:** Polens y-koordinat mod 90° vest (IERS-konvention). Læses sammen med x_p som ét punkt på polbanen.

## LOD  `lod`

- **Today:** 2026-11-20: the day was 1.684 ms longer than 86 400 s. The fluids account for +0.703 ms of it (atmosphere +0.939 ms); the rest is tides and the core.
- **I dag:** 2026-11-20: dagen var 1.684 ms længere end 86 400 s. Væskerne står for +0.703 ms af det (atmosfæren +0.939 ms); resten er tidevand og kernen.
- **What:** Excess length of day over 86 400 s, in milliseconds. Positive: the day is longer than nominal, Earth spins slower.
- **How to read it:** Seasonal ±0.5 ms from the atmosphere, a 13.66 d and 27.55 d tidal ripple of ±0.5 ms, and slow decadal swings from the core. Today's value beside the fluid estimate tells you how much of it is weather.
- **Not:** It is not a trend: the secular tidal braking is ~2 ms per century, invisible at this scale.
- **Dansk:** Døgnets overskydende længde i millisekunder. Positiv: dagen er længere, Jorden roterer langsommere. Sæson ±0,5 ms fra atmosfæren, tidevandsrippel ±0,5 ms, langsomme svingninger fra kernen.

## UT1−UTC  `ut1utc`

- **Today:** -0.0766 s: Earth's clock is behind atomic time by 77 ms and the gap widens at 1.68 ms/day.
- **I dag:** -0.0766 s: Jordens ur er bagud for atomtiden med 77 ms, og gabet vokser med 1.68 ms/dag.
- **What:** How far the Earth-angle clock (UT1) has drifted from atomic UTC, in seconds. Its slope is −LOD.
- **How to read it:** Approaching ±0.9 s forces a leap second. In 2020–2024 the day was often shorter than nominal, UT1−UTC climbed, and a negative leap second was discussed for the first time; a positive LOD like today's pulls it back the other way. The 3D globe is turned by this number.
- **Not:** Not a measurement of the day length directly; the day length is its derivative.
- **Dansk:** Jordens rotationsvinkel-ur (UT1) minus atomtid (UTC), i sekunder. Hældningen er −LOD. Kloden på 3D-siden drejes efter dette tal.

## ω vector  `omega`

- **Today:** ω_z − Ω₀ = -1.421e-12 rad/s: 14 parts in 10¹³ below nominal.
- **I dag:** ω_z − Ω₀ = -1.421e-12 rad/s: 14 dele i 10¹³ under det nominelle.
- **What:** The rotation vector in rad/s: ω_x, ω_y from the pole offset, ω_z from the day length. Nominal Ω₀ = 7.292115e-5 rad/s.
- **How to read it:** ω_z − Ω₀ is of order 1e-13 rad/s — one part in a billion. The picture is exact; the deviation is tiny and real.
- **Not:** Not the axis of figure (the crust's symmetry axis): the small angle between the two is the polar motion.
- **Dansk:** Rotationsvektoren i rad/s: ω_x, ω_y fra polafvigelsen, ω_z fra døgnlængden. Afvigelsen fra Ω₀ er én milliardtedel — lille, men reel.

## Pole offset  `pole_offset_m`

- **Today:** 10.84 m from the reference pole (x_p 0.1154″, y_p 0.3313″) — the outer part of the Chandler-annual beat.
- **I dag:** 10.84 m fra referencepolen (x_p 0.1154″, y_p 0.3313″) — den ydre del af Chandler-års-svævningen.
- **What:** Distance on the crust between the instantaneous rotation pole and the reference pole, R⊕·√(x_p²+y_p²).
- **How to read it:** Typically 3–15 m, a spiral with a ~6.4-year beat between the annual and Chandler terms. On the globe it is exaggerated ×40 000 and labelled so.
- **Not:** Not a drift of the continents; the reference pole is fixed to the crust.
- **Dansk:** Afstanden på skorpen mellem den øjeblikkelige rotationspol og referencepolen, R⊕·√(x_p²+y_p²). Typisk 3–15 m. På kloden tegnet ×40 000 og mærket sådan.

## Rotational energy  `E_rot_J`

- **Today:** Today's +1.684 ms of day length is 8.33e21 J taken out of the rotational store.
- **I dag:** Dagens +1.684 ms døgnlængde er 8.33e21 J taget ud af rotationslageret.
- **What:** ½·I·ω² with I = 8.034e37 kg·m² (assumed constant).
- **How to read it:** About 2.1e29 J. A 1 ms longer day removes about 5e21 J from it — the dE per ms figure — which is roughly the world's annual energy use, moved into or out of the fluids and the core.
- **Not:** Not an energy the fluids 'consume'; it is exchanged as angular momentum.
- **Dansk:** Rotationsenergien ½·I·ω², ca. 2,1·10²⁹ J. Én ms længere dag flytter ca. 5·10²¹ J — omtrent verdens årlige energiforbrug.

## Angular momentum  `L_kg_m2_s`

- **What:** I·ω, about 5.9e33 kg·m²/s. Conserved for the Earth system: what the atmosphere gains, the solid Earth loses.
- **How to read it:** That conservation is why AAM predicts LOD: a westerly wind burst spins the air up and the ground down by the same amount.
- **Not:** Not conserved for the solid Earth alone.
- **Dansk:** Impulsmomentet I·ω. Bevaret for hele Jord-systemet: hvad atmosfæren vinder, taber den faste Jord. Derfor forudsiger AAM døgnlængden.

## Equator speed  `equator_speed_m_s`

- **What:** ω_z·R⊕, about 465 m/s.
- **How to read it:** Changes by ~5 µm/s per ms of LOD. Shown to give the vector a human scale.
- **Not:** Not a measurement; a derived number.
- **Dansk:** Ækvatorhastigheden ω_z·R⊕, ca. 465 m/s. Ændres med ~5 µm/s pr. ms LOD. Vist for at give vektoren en menneskelig skala.

## Atmosphere (AAM)  `lod_atm_ms`

- **Today:** +0.939 ms today — the dominant fluid term, as usual.
- **I dag:** +0.939 ms i dag — det dominerende væskeled, som sædvanligt.
- **What:** GFZ's effective angular momentum of the atmosphere (ECMWF winds + pressure) converted to a length-of-day change: 86 400 s × χ₃, in ms.
- **How to read it:** Carries most of the seasonal LOD signal: the northern winter jet streams spin the air up and the day lengthens.
- **Not:** A model product, not a measurement. It has no tides in it.
- **Dansk:** Atmosfærens effektive impulsmoment (ECMWF-vinde + tryk) omregnet til døgnlængde-ændring i ms. Bærer det meste af sæsonsignalet. Modelprodukt, ikke måling.

## Ocean (OAM)  `lod_ocn_ms`

- **What:** The same from the MPIOM ocean model (currents + bottom pressure).
- **How to read it:** Small but not negligible at the seasonal scale; strongest during ENSO events.
- **Not:** Model product.
- **Dansk:** Det samme fra havmodellen MPIOM (strømme + bundtryk). Lille, men ikke ubetydelig; stærkest under ENSO-hændelser.

## Land water (HAM)  `lod_hyd_ms`

- **What:** The same from the LSDM land-hydrology model (soil moisture, snow, rivers).
- **How to read it:** Slow and small in LOD; it matters more for the pole (χ₁, χ₂) than for the day length.
- **Not:** Model product.
- **Dansk:** Det samme fra landhydrologimodellen LSDM (jordfugt, sne, floder). Betyder mere for polen end for døgnlængden.

## Fluids, total  `lod_geo_ms`

- **Today:** Atmosphere +0.939 ms, ocean -0.070 ms, land water -0.166 ms → fluids +0.703 ms.
- **I dag:** Atmosfære +0.939 ms, hav -0.070 ms, landvand -0.166 ms → væsker +0.703 ms.
- **What:** Atmosphere + ocean + land water, ms of LOD.
- **How to read it:** Put beside the observed LOD after the 31-day smoothing: R² 0.7 means seven tenths of the season-to-season day-length change IS the fluids. The rest is the core and model error.
- **Not:** It cannot explain the tidal ripple (the models do not carry tides), which is why the daily R² is low.
- **Dansk:** Atmosfære + hav + landvand i ms LOD. Efter 31-dages glidende middel forklarer væskerne ca. 7/10 af sæsonvariationen; resten er kernen og modelfejl.

## R² attribution  `attribution_r2`

- **Today:** R² 0.25 daily, 0.47 after the 31-day mean: the fluids explain about 47% of the season-to-season day length this year.
- **I dag:** R² 0.25 dagligt, 0.47 efter 31-dages middel: væskerne forklarer ca. 47 % af sæsonvariationen i døgnlængden i år.
- **What:** How much of the observed LOD variance the fluid estimate explains, daily and after a 31-day running mean.
- **How to read it:** Daily R² is dragged down by the zonal tides; the smoothed R² is the fair number. Both are reported so nobody has to trust the choice.
- **Not:** Not a fit of the fluids to LOD; the slope and offset only calibrate the models' subtracted mean.
- **Dansk:** Hvor stor en del af den observerede LOD-varians væskeestimatet forklarer, dagligt og efter 31-dages middel. Det daglige R² trækkes ned af tidevandet; det glattede er det fair tal.

## K⊕ forecast  `k_resid_fcst`

- **Today:** For 2026-11-21: 2.72 through the IERS prediction — elevated.
- **I dag:** For 2026-11-21: 2.72 via IERS-forudsigelsen — elevated.
- **What:** The IERS-predicted pole and UT1−UTC for the next ten days, pushed through the same model and σ as today's K⊕. LOD on those days is the finite difference of the predicted UT1−UTC, because prediction rows carry no LOD.
- **How to read it:** Trust the first three days; after that the IERS pole prediction and the fluid forecast both lose skill. The band is the IERS-stated prediction error propagated.
- **Not:** It is not our own prediction of the pole; it is a reading of the IERS prediction through our model.
- **Dansk:** IERS-forudsagt pol og UT1−UTC ti dage frem gennem samme model og σ som i dag. LOD på de dage er afledt af UT1−UTC-differensen. Stol på de første tre dage.

## K⊕ forecast, fluid opinion  `k_resid_fcst_gfz`

- **What:** The same K⊕ with the LOD slot filled by the GFZ 10-day fluid forecast (calibrated), instead of the IERS UT1−UTC difference.
- **How to read it:** Two independent forecasts of the same number. When they agree, believe both; when they split, the gap is the tides plus the core.
- **Not:** Never averaged into the headline forecast.
- **Dansk:** Samme K⊕, men med GFZ's 10-dages væskeprognose i LOD-pladsen. To uafhængige prognoser af samme tal; når de er enige, tro på begge.

## ±90-day EAM prediction  `eam90`

- **What:** GFZ's combined atmosphere + ocean + hydrology prediction three months out (Dill et al. 2018).
- **How to read it:** Useful for the direction of the seasonal swing, not its daily value.
- **Not:** Not a forecast of K⊕; it feeds only the LOD channel.
- **Dansk:** GFZ's kombinerede forudsigelse af atmosfære + hav + hydrologi tre måneder frem. Nyttig for sæsonsvingningens retning, ikke for den daglige værdi.

## Earth Rotation Angle  `era`

- **What:** The angle the globe has turned since the reference epoch, from your clock plus UT1−UTC: ERA = 2π(0.7790572732640 + 1.00273781191135448·(JD_UT1 − 2451545)).
- **How to read it:** The 3D globe is turned by exactly this, so Greenwich sits where it really is right now and the terminator is where the Sun really puts it.
- **Not:** Not sidereal time (that adds precession); the difference is under a degree.
- **Dansk:** Jordens rotationsvinkel siden referenceepoken, fra dit ur plus UT1−UTC. 3D-kloden drejes præcis dette, så Greenwich står, hvor det står lige nu.

## K_bio (the breathing)  `k_bio`

- **Today:** 2026-11-19: 0.16σ — inside the rhythm (p99 2.57); Mauna Loa 429.16 ppm.
- **I dag:** 2026-11-19: 0.16σ — inden for rytmen (p99 2.57); Mauna Loa 429.16 ppm.
- **What:** Mauna Loa's daily CO₂ after trend + acceleration + annual + semiannual are fitted over six years: the 7-day mean residual in σ. The same construction as K⊕, applied to the planet's breathing instead of its spin.
- **How to read it:** Below 1: the biosphere is breathing as its own cycles predicted. 1–3: a warm autumn, an early spring, a fire season, or a Mauna Loa instrument week. Above the p99 line: something the cycles do not know — El Niño drought, a volcanic plume at the intake, a sensor fault — and the sentence to write is 'the biosphere is off its rhythm', not 'CO₂ is high'.
- **Not:** It is not the CO₂ level and not the climate trend (both are removed before scoring). It is the surprise in the seasonal rhythm, weighted toward the northern hemisphere where Mauna Loa listens.
- **K family:** A K: disagreement over spread inside the same ladder the pole and the chain use. One channel, so it is a plain |z|, not a quadrature.
- **Dansk:** Mauna Loas daglige CO₂ efter trend + acceleration + årlig + halvårlig cyklus er fjernet: 7-dages middelresidual i σ. Under 1: biosfæren trækker vejret som sine egne cyklusser forudsagde. Over p99: noget cyklusserne ikke kender.

## The breathing  `breathing`

- **Today:** Exhaling at 0.109 PgC/day (index +0.61 of the steepest part of the cycle); amplitude 6.80 ppm ≈ 14.5 PgC; the long trend underneath is +2.86 ppm/yr.
- **I dag:** Udånder med 0.109 PgC/dag (indeks +0.61 af cyklussens stejleste del); amplitude 6.80 ppm ≈ 14.5 PgC; den lange trend nedenunder er +2.86 ppm/år.
- **What:** The fitted seasonal derivative of CO₂ today, ppm per day, times 2.124 PgC per ppm: the net rate at which the (mostly northern) biosphere is drawing carbon down or letting it go right now. Amplitude ≈ 7 ppm peak-to-trough, maximum in May, minimum in late September.
- **How to read it:** Negative from May to September — the growing season inhales about 15 PgC; positive the rest of the year as leaves and soils exhale it back. The index runs −1…+1 against the steepest part of the cycle.
- **Not:** Not a global flux inversion and not the oceans: one station, one hemisphere's rhythm, fitted, not measured flux.
- **Dansk:** Den tilpassede sæsonafledte af CO₂ i dag, ppm/dag, gange 2,124 PgC/ppm: nettoraten hvormed biosfæren (mest den nordlige) optager eller afgiver kulstof lige nu. Negativ maj–september (indånding), positiv resten af året.

## The Lloyd ladder  `ops`

- **Today:** ATP 1.59e33/s · Landauer 4.80e34/s · Margolus–Levitin 1.32e56/s → 21 kT per op, 22.9 orders below the ceiling.
- **I dag:** ATP 1.59e33/s · Landauer 4.80e34/s · Margolus–Levitin 1.32e56/s → 21 kT pr. op, 22.9 størrelsesordener under loftet.
- **What:** Three numbers per second for all living things: 2E/(πħ) on the standing biomass (Margolus–Levitin, what physics allows), P_npp/(kT·ln2) at 288 K (Landauer, what the photosynthetic power could pay for), and P_npp/E_ATP (ATP hydrolyses, the elementary operation life actually performs).
- **How to read it:** ATP ≈ 1.6e33 ops/s; Landauer ≈ 4.8e34; Margolus–Levitin ≈ 1.3e56. So life spends ≈ 21 kT per elementary operation — about 30× the Landauer minimum, the ballpark every study of biological computation finds — and sits 23 orders of magnitude below the quantum ceiling. Memory: 1.06e38 bits of DNA. Photosynthesis captures ≈ 0.1 % of absorbed sunlight.
- **Not:** Not measured ops. Derived from four published totals (NPP 104.9 PgC/yr, biomass 550 GtC, ΔG 478 kJ/mol C, ATP 50 kJ/mol) and two assumptions (NPP ≈ respiration; one ATP = one op). Change a total and every number moves with it — the derivation is the claim, not the digits.
- **K family:** Seth Lloyd's 'ultimate laptop' bound, applied to the biosphere's energy budget instead of a kilogram of matter.
- **Dansk:** Tre tal pr. sekund for alt levende: Margolus–Levitin-loftet på biomassen (~10⁵⁶), Landauer-loftet på fotosyntese-effekten ved 288 K (~5·10³⁴) og ATP-hydrolyser (~1,6·10³³). Livet bruger ≈ 21 kT pr. elementær operation — ca. 30× Landauer-minimum — og ligger 23 størrelsesordener under kvanteloftet. Udledt fra fire publicerede totaler, ikke målt.

## Attestation  `attest`

- **Today:** This publication becomes attestation row #290; the previous row #289 is signed, its on-chain anchor pending.
- **I dag:** Denne udgivelse bliver attesteringsrække #290; den forrige række #289 er signeret, kæde-ankeret afventer.
- **What:** BLAKE3 of the exact bytes of latest.json, chained to the previous digest and signed with a dedicated Ed25519 key. An anchor row records the shielded SIGIL transaction that carried the digest in its memo.
- **How to read it:** Anyone can recompute the digest from the published file and check the signature against the published key. The chain receipt proves the digest existed at that block; it does not prove the reading is true.
- **Not:** Not a proof of correctness; a proof of non-alteration with a timestamp nobody can rewrite.
- **Dansk:** BLAKE3 af de præcise bytes i latest.json, kædet til forrige digest og signeret med en dedikeret Ed25519-nøgle; en anker-række noterer SIGIL-transaktionen med digestet i memoet. Beviser ikke-ændring, ikke sandhed.

## Alert  `alert`

- **What:** K⊕ above the empirical p99 of the trailing three years (fires once, re-arms below p90), or a step on the K-family ladder.
- **How to read it:** One alert means: look at the fluids panel first. If the fluid estimate also jumped, it is weather; if not, it is the core or a data revision.
- **Not:** Not an earthquake or weather warning. IERS rapid values are revised for weeks; an alert can be revised away.
- **Dansk:** K⊕ over det empiriske p99 (udløses én gang, genaktiveres under p90) eller et trin på K-familiens stige. Én alarm betyder: se på væskepanelet først.

