The Alley 2000 ice-core series is a local archive, not a global thermometer. Hover any dotted term for the textbook and why it matters here. Method bias — slope, elevation, site noise, the 1855 cutoff — can put the Holocene Thermal Maximum warmer than Summit today, while a naïve splice to direct Arctic temperatures argues the opposite.
Anomaly vs 1000–1850 (proxies) / 1880–1910 (instrumental), °C
ka before 1950 →
Alley 2000
Arctic 64–90°N
Summit 2001–10
Firn cutoffYoungest Alley 2000 point, ~1855 CE. Snow is not ice yet.
8.2 kaCentury-scale Holocene cooling. Layer-counted timing is strong (~1%).
HTMEarly-to-mid Holocene warmth at Summit — amplitude depends on calibration.
Kobashi endGas-isotope inversion stops in the early 1990s.
Ice HTM (this borehole)
-28.70 °C
Adjusted Alley peak, 9–5 ka. Published file already ~−28.7 °C — warmer than Summit 2001–2010 before any extra knob.
Honest “today”
-29.90 °C
GISP2 2001–2010 mean. Same site, same absolute scale as the ice. Not GISTEMP.
Misleading “now”
+3.38 °C
GISTEMP 64–90°N, 2015–24, vs its own baseline. A real thermometer of a different domain — not Summit, not δ¹⁸O.
Adjusted HTM peak
+3.08 °C
Alley knobs, 9–5 ka, Summit
Alley at ~1855
+0.11 °C
End of the ice-core series
HTM minus 1855
+2.97 °C
Envelope ±1.2 °C
Arctic 2015–24
+3.38 °C
GISTEMP 64–90°N, not Summit
Six reasons the overlay is not a comparison
Each figure is a real series, a published constant, or a labeled schematic of the physics. None of them is a global thermometer. Hover the dotted terms — the amber note is how that method can put Summit warmer than today, or make a thermometer lie by comparison.
01δ¹⁸O is not a pure thermometer
Nearby cores share climate and still correlate only r ≈ 0.4. This draw: r = 0.40.
the thermometer
Condensation temperature
also moves δ¹⁸O
Moisture-source region
also moves δ¹⁸O
Atmospheric transport
also moves δ¹⁸O
Precipitation seasonality
also moves δ¹⁸O
Ice-sheet elevation
also moves δ¹⁸O
Wind-blown snow / drift
also moves δ¹⁸O
Firn diffusion & vapor exchange
Not a variance split — a list of known drivers. Only condensation temperature is the signal people treat as T.
°C vs shared 150-yr mean
Shared climateCore ACore B
Year CE
Core A vs Core B, same years
Oxygen-isotope ratios in snow respond to condensation temperature, but they also shift with moisture-source region, atmospheric transport, precipitation seasonality, ice-sheet elevation, wind-driven snow redistribution, firn diffusion, and vapor exchange. Alley noted that a single-site isotopic record “is not purely a temperature record.” Nearby cores (GISP2 vs GRIP) correlate only moderately (r ≈ 0.4). The envelope slider is there because the wiggles are not all climate. The scatter is a labeled schematic of that correlation, not a GRIP download.
Bias consequenceTreating every wiggle as temperature, then as Earth, can put a local noise peak “above today” or hide a real Thermal Maximum. Direct temperatures have none of those extra drivers.
02Calibration slope was allowed to wander
The same 1‰ of δ¹⁸O is 1.5 °C or 3.0 °C, depending on the slope Cuffey & Clow let wander by ~2×.
Same 1‰, two slopes — wedge is the allowed range
Thinning is a temperature before any isotope effect
Alley anomaly rescaled, °C vs 1000–1850
As publishedSlope × ½ → ΔT × 2
ka BP — published slope vs half slope (0.33 ‰/°C)
Cuffey & Clow (1997) let the δ¹⁸O–temperature slope change by roughly a factor of two through the Holocene. The same 1‰ isotope shift is ~1.5 °C at the spatial slope (0.67 ‰/°C) and ~3 °C at half that slope. Vinther et al. (2009) keep the slope closer to constant and correct Holocene thinning: 200 m of lowering is already ~1.3 °C at a 6.5 K/km lapse rate, before the isotope effect. That is the elevation knob. The amber curve is the published Alley anomaly with ΔT doubled — what a 0.33 ‰/°C slope would mean.
Bias consequenceA steep slope shrinks Holocene warmth so today’s Arctic line looks unprecedented. A shallow slope — allowed by the borehole inversion — can put the Thermal Maximum well above the GISP2 2001–2010 mean. That is a method choice, not a new Arctic measurement.
03Boreholes smear the wiggles you are looking at
A 40-year, 1 °C surface pulse is a borehole blur of σ ≈ 55 yr at 226 yr age, and σ ≈ 323 yr by the early Common Era.
The kernel that ate the wiggle · σ = 28·(age/100)^0.82 yr
Surface pulse, 1 °C × 40 yr
Year CE → older pulses are more smeared
Pulse
Age
σ
Peak left in the borehole
1800 CE
226 yr
55 yr
0.27 of 1.00 °C
1200 CE
826 yr
158 yr
0.09 of 1.00 °C
500 CE
1526 yr
262 yr
0.07 of 1.00 °C
50 CE
1976 yr
323 yr
0.06 of 1.00 °C
Heat diffusion in ice is a low-pass filter. Identical 40-year, 1 °C surface pulses fade as they age — the kernel is the explorer’s borehole filter, σ ≈ 28 · (age/100)^0.82 years. High-precision GISP2 borehole logs (~4.5 mK) recover the ~15 °C glacial–Holocene step and a Holocene cooling trend. They cannot resolve the multi-decadal wiggles in the isotope curve. Gas isotopes (δ¹⁵N, δ⁴⁰Ar) are a more direct Holocene thermometer; published uncertainties are typically 1.2–1.5 K.
Bias consequencePlotting a sharp GISTEMP spike against a diffused ice curve is a resolution trick. The borehole cannot prove Summit never ran warmer than today at century scale — and a 1.2–1.5 K gas-isotope band can overlap the recent Summit mean.
04The published series stops in the 1850s
Youngest Alley 2000 point is 1855 CE. NASA GISTEMP 64–90°N begins 1880 — a real thermometer, and not Summit.
Summit firn · schematic Herron–Langway density
Anomaly vs late Holocene / 1880–1910, °C
Alley 2000GISTEMP 64–90°N
Year CE
The youngest Alley 2000 point is ~1855. Snow needs decades to densify; vapor exchange continues in the firn. Summit close-off is about 70–80 m, on the order of two centuries — the density curve is a Herron–Langway-style schematic, not a GISP2 density log. The bright line is NASA GISTEMP 64–90°N from 1880 — a real thermometer, and a different place than 72.6°N, 3200 m. Kobashi et al. (2011) quote a GISP2 2001–2010 mean of −29.9 °C. That marker lives on the absolute scale. It is not GISTEMP. A splice across the amber gap is a drawing choice.
Bias consequenceYou cannot read “now vs the Holocene” from a series that ends in 1855, then paste in a different region’s thermometers. The naïve splice can erase a Holocene that was warmer at Summit by stealing the visual from 64–90°N.
05Dating is the strong part
8.2 ka is placed to ±82 yr (~1% layer count). A typical gas-isotope thermometer is ±1.4 K.
Alley anomaly, °C
±82 yr±1.4 K
ka BP — ice-cyan strip = ±1% age; amber envelope = ±1.4 K around the curve
Age, the strong axis
±82 yr
1% of 8.2 ka at the 8.2 ka cold event (Alley local min -1.5 °C in this window). Holocene layer counts at GISP2 are the part of the core nobody is arguing about.
Temperature, the weak axis
±1.4 K
Typical published uncertainty on GISP2 gas-isotope temperature (δ¹⁵N / δ⁴⁰Ar). The overlay argument is not that the core is undated.
Annual-layer counting through the Holocene is ~1%, so the 8.2 ka event is placed to about ±80 years (the narrow ice-cyan strip). The amber envelope is ±1.4 K around the curve — a typical gas-isotope temperature uncertainty — not a band around zero. The argument is not “the core is undated.” It is that a single-site, time-varying calibration with an 1850s cutoff is not a thermometer that already contains the last 170 years.
Bias consequenceTemperature method, not chronology, is the weak axis. Ignoring a ±1.4 K band while comparing to unsmoothed direct temperatures is how “cooler than now” can be false precision.
06What later reconstructions actually change
Later work still shows a warm early–mid Holocene at Greenland. It does not ratify Alley’s century-scale wiggles, and none of it is GISTEMP.
Anomaly vs 1000–1850, °C
Alley 2000Vinther 2009Kobashi 2017
ka BP
Mean 8–6 ka BP minus 1000–1850 CE — millennial HTM warmth, not century wiggles
Alley 2000+1.54 °C
Vinther 2009+2.46 °C
Kobashi 2017+1.90 °C
Alley 2000
What
δ¹⁸O + borehole
Where
GISP2 only
Ends
1855
T error
Slope × ~2; site noise
Vinther 2009
What
6-core δ¹⁸O + elevation
Where
Greenland + Agassiz
Ends
~1970, 20-yr
T error
Borehole-tuned stack
Anomaly vs present, not Summit °C.
Kobashi 2017
What
δ¹⁵N / δ⁴⁰Ar gases
Where
GISP2 only
Ends
1993
T error
1.2–1.5 K typical
GISTEMP
What
Thermometers
Where
64–90°N / globe
Ends
2025
T error
~0.05–0.1 K zonal
Not a Summit reconstruction.
Multi-core, elevation-corrected (Vinther 2009) and gas-isotope (Kobashi 2017) reconstructions still show large Holocene variability at Greenland. They do not agree wiggle-for-wiggle with Alley 2000, and none of them is GISTEMP. The bars compare the same two windows (8–6 ka BP vs 1000–1850 CE). Turn on those series in the explorer, use the documented-uncertainty preset, and ask whether “now versus the Holocene Thermal Maximum” survives the knobs. Arctic warming is real. It is also a different place, a different physical quantity, and a different resolution.
Bias consequenceLater work still allows a warm early–mid Holocene at Greenland. Their uncertainty bands can overlap or exceed the recent Summit mean. None of them is a 64–90°N instrumental series, so they cannot ratify the overlay’s “unprecedented now.”
Sources & data
Every series on the chart is the NOAA NCEI / World Data Service file or the NASA GISTEMP zonal table — not a redraw from a screenshot. Original archives download as the files NOAA and NASA published. Parsed CSVs are the columns this explorer actually plots (year converted to CE). The Alley knobs write a separate “this view” CSV; that file is pedagogical, not a new inversion.
Instrumental series stay fixed. Proxy CSVs labeled “as published” ignore the sliders.
Datasets on the chart
Alley 2000
proxyon chart
The Younger Dryas cold interval as viewed from central Greenland
High variability of Greenland surface temperature over the past 4000 years estimated from trapped air in an ice core
Kobashi, T. et al. · 2011 · Geophysical Research Letters 38:L21501
GISP2 δ¹⁵N–δ⁴⁰Ar gas-isotope inversion with published 1σ bands. Independent of snow δ¹⁸O. NOAA study 12887. Source of the −29.9 °C 2001–2010 Summit mean used here.
GISS Surface Temperature Analysis, zonal annual means
GISTEMP Team / Lenssen et al. · 2019 · NASA GISS · 1951–1980 baseline
Fixed thermometer. Chart uses 64–90°N and global annual anomalies from ZonAnn.Ts+dSST.csv. This is not Summit, not δ¹⁸O, and does not move with the Alley knobs.
Temperature, accumulation, and ice sheet elevation in central Greenland through the last deglacial transition
Cuffey, K.M. and Clow, G.D. · 1997 · Journal of Geophysical Research: Oceans 102:26383–26396
The inversion Alley later smoothed. Time-varying δ¹⁸O–T slope and borehole diffusion live here; Alley 2000 is the figure-ready curve, not a new thermometer.
Factcheck: What Greenland ice cores say about past and present climate change
Hausfather, Z. · 2019 · Carbon Brief
The methods critique this explorer is built on: single-site noise, non-temperature isotope effects, time-varying calibration, missing elevation correction, firn cutoff, and why a naïve overlay is not a comparison.
Cite the papers and dataset DOIs, not this page, if you reuse the numbers. Age conversion is year CE = 1950 − age ka BP × 1000. Vinther 20-year bins use the published midpoints.