Air bubbles inside an ice core and the ice around them are not necessarily the same age. Falling snow starts forming the ice archive, while air can continue exchanging through pores in the firn until those pathways close. The trapped gas is therefore generally younger than the enclosing ice, and a difference between gas and ice signals at one core depth is not automatically a delay between the original climate events.
In a fictional steady accumulation example, 56 metres of ice-equivalent firn above gas lock-in and accumulation of 0.14 metres of ice equivalent per year give an approximate gas-age–ice-age offset of 400 years. If the ice at that depth dates to 12,000 years before present, the mean enclosed gas might date to approximately 11,600 years before present. The calculation simplifies firn compaction and gas transport; it is not an exact age model. This article follows the Gold Standard of Palaeoclimatology from the archive into the chronology.
Find the question you need
Choose a reading route or work through the chapters. All chapters remain open; the grouped index below is expandable.
Name the two clocks
Distinguish when snow fell from when air stopped exchanging. [Start with chapter 1](#pt-firn-chapter-1).
Calculate an offset
Use ice-equivalent lock-in depth and a fictional accumulation rate. [Start with chapter 3](#pt-firn-chapter-3).
Interpret linked climate events
Separate a depth offset from a physical lead or lag. [Start with chapter 6](#pt-firn-chapter-6).
Check proxy evidence
Assess lock-in depth, gas mixing and correlated uncertainty. [Start with chapter 8](#pt-firn-chapter-8).
Practise and continue
Change accumulation and link to age-model guides. [Start with chapter 11](#pt-firn-chapter-11).
You can also return to the Gold Standard of Palaeoclimatology.
Chapter index
Name the two clocks · Chapters 1–2
Calculate an offset · Chapters 3–5
Interpret linked climate events · Chapters 6–7
Check proxy evidence · Chapters 8–10
Practise and continue · Chapters 11–12
CHAPTER 1 OF 12 · Name the two clocks
1. A core can preserve two clocks at one depth
[Back to contents](#pt-firn-contents)
Ice-core oxygen and hydrogen isotope measurements are made on water molecules in the ice. Gas measurements such as atmospheric methane or carbon dioxide come from bubbles trapped during densification. The snow that eventually became the enclosing ice fell before much of its pore air became sealed as discrete bubbles.
Consequently, the ice and bubble gas at the same depth can represent different moments in the past. A plot with one age axis for all proxies may accidentally impose a synchrony the archive never possessed. This matters when investigators compare greenhouse gases with abrupt temperature-related signals.
Define each datum’s support: isotope measurements can resolve variations within ice layers, while trapped air often contains a distribution of gas ages because closure happens across an interval. A neat line on a chart is a constructed time assignment, not evidence that all molecules have a single acquisition time.
[Contents](#pt-firn-contents) · [Next chapter](#pt-firn-chapter-2) · Palaeoclimatology subject guide
CHAPTER 2 OF 12 · Name the two clocks
2. Define firn, open pores, lock-in and closure
[Back to contents](#pt-firn-contents)
Fresh snow compacts into firn, the porous transition material between snowfall and solid glacier ice. Overburden, grain rearrangement and sintering gradually reduce open pore volume. While connected pores remain open to the atmosphere, gases can move by diffusion, pressure-driven processes and other transport.
At depth, the diffusive connection becomes strongly restricted. Scientists use concepts such as lock-in depth and the firn lock-in zone to describe this transition, while individual bubbles can close at different depths. The precise definitions depend on the study and should not be collapsed into one sharp imaginary horizontal sheet.
Temperature and accumulation influence firn thickness and densification. Higher or lower snowfall rates, different thermal histories, and changes in ice-sheet conditions can alter how long a parcel of firn remains open to exchanging gas.
For a chronology question, the important physical distinction is the age of the ice matrix and the mean or distribution of ages of air captured within it. Those values must be tracked separately before comparing proxy signals.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-1) · [Next chapter](#pt-firn-chapter-3) · Palaeoclimatology subject guide
CHAPTER 3 OF 12 · Calculate an offset
3. Give the age difference a clear sign convention
[Back to contents](#pt-firn-contents)
Let tice be ice age in years before present at a given depth and tgas the representative age of trapped gas at that depth. Define Δage = tice − tgas. Since the ice is usually older than the gas, Δage is usually positive under this convention.
Thus, tgas = tice − Δage. The subtraction makes sense because smaller ages before present indicate later times. An author who defines a signed gas-minus-ice difference instead would report a negative value; always read the convention before comparing published chronologies.
A quoted gas age can be a model mean or an age distribution summary, not necessarily a unique date for every bubble. The exact lock-in and transport interpretation affects whether the simplified equation captures the desired quantity.
This sign check is a useful classroom safeguard. On a BP axis where age increases to the past, if gas becomes ‘older’ merely because someone subtracted the wrong way, the plotted climate sequence can be reversed.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-2) · [Next chapter](#pt-firn-chapter-4) · Palaeoclimatology subject guide
CHAPTER 4 OF 12 · Calculate an offset
4. Work a four-hundred-year firn example
[Back to contents](#pt-firn-contents)
Build a fictional steady-state model with ice-equivalent mass thickness above the lock-in horizon LIE = 56 m ice equivalent and mean annual accumulation A = 0.14 m ice equivalent per year. If the vertical mass throughput is steady and air is treated as effectively contemporary near lock-in, a first approximation is Δage ≈ LIE/A.
Therefore Δage ≈ 56 / 0.14 = 400 years. Suppose ice at one depth is assigned tice = 12,000 years BP. The mean gas age there is then tgas ≈ 12,000 − 400 = 11,600 years BP.
The ratio uses ice-equivalent depth, not the physical 56-m thickness of porous snow unless densities have been converted. Firn can contain substantial void space, and physical thickness is not mass thickness. This conversion prevents a seemingly accurate division from using inconsistent physical quantities.
In reality, the past accumulation rate, close-off behavior and gas-age distribution may vary. This example is a transparent mass-balance limit, not a claim that every firn column is accurately dated from a single annual snowfall figure.
| Fictional input / step | Value | Interpretation |
|---|---|---|
| Ice-equivalent lock-in column LIE | 56 m ice eq. | Mass-equivalent thickness above lock-in |
| Accumulation A | 0.14 m ice eq./year | Steady annual mass input |
| Approximate Δage = LIE/A | 400 years | Simplified gas–ice offset |
| Ice age at sampled depth | 12,000 years BP | Matrix’s modeled age |
| Gas age at sampled depth | 11,600 years BP | Mean gas date after subtracting 400 years |
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-3) · [Next chapter](#pt-firn-chapter-5) · Palaeoclimatology subject guide
CHAPTER 5 OF 12 · Calculate an offset
5. Recognise why ice-equivalent depth is not geometric firn depth
[Back to contents](#pt-firn-contents)
A metre of low-density firn contains less ice mass than a metre of dense glacier ice. An ice-equivalent thickness converts the accumulated firn mass into the thickness that same mass would occupy at reference ice density. It is obtained by integrating density over depth and dividing by the reference ice density.
In a simplified two-zone demonstration, 20 m of material at half ice density contributes 10 m ice equivalent, while 60 m at three-quarters ice density contributes 45 m ice equivalent. Their combined 80-m physical column contains 55 m ice equivalent, not 80 m. The uniform zones are fictional approximations.
This is why replacing LIE with visually measured firn depth would overstate the mass-transit time at the same ice-equivalent accumulation rate. A different effective close-off level would also change the result. The numerical ratio works only when numerator and denominator share the same mass-equivalent basis.
Real density profiles are continuous, and their interpretation must account for how vertical strain and compaction evolved. A single linear transformation is an instructive first check, not a substitute for a site-specific model.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-4) · [Next chapter](#pt-firn-chapter-6) · Palaeoclimatology subject guide
CHAPTER 6 OF 12 · Interpret linked climate events
6. Separate depth displacement from event timing
[Back to contents](#pt-firn-contents)
Imagine a climate transition at an actual past time of 12,000 years BP that influences both an ice-phase isotope proxy and an atmospheric methane signal without intrinsic physical delay. Ice deposited then may mark the ice event in one layer. The methane signal reaches a sealed bubble archive only when firn pores close later.
To find methane trapped with air 12,000 years old, one may need a deeper core level whose enclosing ice is older than 12,000 years. If Δage there is 400 years, that ice could be assigned roughly 12,400 years BP. The ice and gas event peaks thus need not occur at the same core depth despite original synchrony.
That displacement in depth is not itself a 400-year climate lag. A credible comparison first converts each record to its own age scale. It then asks whether any residual delay remains after gas lock-in and chronology uncertainties are included.
Conversely, correcting an age offset does not prove two processes occurred at the same time. Geological inference still needs independent dating, appropriate sampling resolution and physical interpretation of each proxy.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-5) · [Next chapter](#pt-firn-chapter-7) · Palaeoclimatology subject guide
CHAPTER 7 OF 12 · Interpret linked climate events
7. Do not assume Δage remains constant through climate transitions
[Back to contents](#pt-firn-contents)
Glacial climates, accumulation and temperatures can change the thickness and densification rate of firn, so Δage may be larger or smaller at different times. The earlier 400-year figure comes from one invented steady configuration. Copying it as a uniform correction over a long core would hide actual firn history.
An abrupt temperature transition may create transient firn conditions that do not immediately match a steady-state model. Gas transport can also change as layering, convection and diffusive conditions evolve. Model–data comparisons should say which processes were included and whether the sampling interval is close to equilibrium.
At low-accumulation sites, the gas–ice age difference can be especially substantial. Its uncertainty may dominate a proposed short lead–lag conclusion, even when some ice-phase layers are dated more precisely.
Compare modelled Δage to constraints from volcanic ice markers, globally mixed gases, isotopic properties of trapped air and records at sites with different accumulation characteristics. Agreement with only one proxy does not establish a unique firn model.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-6) · [Next chapter](#pt-firn-chapter-8) · Palaeoclimatology subject guide
CHAPTER 8 OF 12 · Check proxy evidence
8. Use gas isotopes and methane synchronization carefully
[Back to contents](#pt-firn-contents)
Nitrogen isotopic fractionation in firn can help constrain aspects of the diffusive column and temperature gradients. Atmospheric methane variations may offer shared gas-phase features between cores. Ice-phase volcanic markers offer different chronological tie points. Each is an observation with its own assumptions and uncertainty.
Combining gas-phase and ice-phase synchronization can narrow possible Δage histories. It does not erase the need for an age-scale construction. Gas concentrations measured from a sample may represent an age distribution, while individual volcanic layers may be quite narrow physical horizons.
An inter-core comparison can be informative because sites with different snowfall rates have different gas-lock-in ages. Still, using a common reference record may introduce shared chronological uncertainty. If two estimated ages depend on the same tie point, their errors are not independent.
A strong result explains exactly what is synchronized: isotope layers in ice, gas composition in bubbles or an externally dated event. Mixing those categories into one ‘same age’ label is a frequent path to overconfidence.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-7) · [Next chapter](#pt-firn-chapter-9) · Palaeoclimatology subject guide
CHAPTER 9 OF 12 · Check proxy evidence
9. Propagate dating uncertainty rather than hiding it
[Back to contents](#pt-firn-contents)
Suppose a fictional sample has tice = 12,000 ±100 years and Δage = 400 ±80 years, treated as independent one-standard-deviation quantities purely for illustration. Then tgas = 11,600 years and its standard uncertainty is √(100² + 80²) ≈ 128 years.
When the age estimates share data or model parameters, independence may be inappropriate. The general variance for tgas = tice − Δage is Var(tgas) = Var(tice) + Var(Δage) − 2 Cov(tice,Δage). Positive covariance can reduce one combination’s uncertainty; negative covariance can increase it.
Those errors also vary along a core. Age-model uncertainties often move neighbouring points together, unlike independent instrumental noise at each measured depth. A lead–lag calculation must propagate the joint age model for both signals rather than use separately rounded year counts.
A chronology plot with narrow point markers should not visually imply a more precise event ordering than the ensemble of plausible age scales supports.
| Quantity or scenario | Fictional result | Important limitation |
|---|---|---|
| Ice age at one depth | 12,000 ±100 yr BP | Dependent on ice-phase chronology |
| Model Δage | 400 ±80 yr | Dependent on firn and accumulation history |
| Gas age under independence | 11,600 ±128 yr BP | Excludes covariance and gas-age width |
| Correlated ice and Δage errors | Use covariance term | Independent quadrature may misstate uncertainty |
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-8) · [Next chapter](#pt-firn-chapter-10) · Palaeoclimatology subject guide
CHAPTER 10 OF 12 · Check proxy evidence
10. Distinguish a mean gas age from its enclosed age spread
[Back to contents](#pt-firn-contents)
Air is not necessarily captured on one instant. Firn pore air can mix by diffusion before bubbles close across a finite interval. A sampled bag of extracted gas may combine molecules with different atmospheric acquisition times, so the effective gas age has a width in addition to its mean offset from ice age.
An abrupt atmospheric step can appear smoothed in gas measurements if its historical profile is convolved with a broad gas-age distribution. A small error bar on the mean age does not imply the signal itself preserves arbitrarily fine temporal detail.
An investigator may estimate the distribution with site-specific firn transport modeling or observations of gases and isotope fractionation. Those procedures require assumptions and often carry uncertainty. Treating a narrow gas signal at one core as a globally precise instant requires separate justification.
This phenomenon is related to, but physically different from, bioturbation in marine sediments. Both archives can smooth an original signal, but they do so through different materials and transport histories.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-9) · [Next chapter](#pt-firn-chapter-11) · Palaeoclimatology subject guide
CHAPTER 11 OF 12 · Practise and continue
11. Practice with a colder or slower-accumulation site
[Back to contents](#pt-firn-contents)
Keep the hypothetical LIE = 56 m but halve ice-equivalent annual accumulation from 0.14 to 0.07 m/year, assuming the lock-in column somehow remains unchanged for the calculation. The simple residence-time ratio doubles: Δage ≈ 56/0.07 = 800 years.
At a sample depth with tice = 12,000 years BP, the corresponding illustrative mean gas age becomes 11,200 years BP. The result proves only the ratio’s sensitivity to input accumulation. In reality the densification structure could also change, so holding LIE fixed is a mathematical one-variable test.
As a second variation, suppose the ice-equivalent lock-in thickness becomes 42 m with A = 0.14 m/year. The ratio becomes 300 years. State which observable—firn density profile, accumulation, gas isotope constraint or dated tie point—could help discriminate these fictional alternatives.
Finally, imagine an apparent gas-versus-ice climate lag of 150 years in a record whose Δage uncertainty is hundreds of years. Write a careful claim about what is and is not resolved before you interpret the lag as physical causation.
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-10) · [Next chapter](#pt-firn-chapter-12) · Palaeoclimatology subject guide
CHAPTER 12 OF 12 · Practise and continue
12. Put two age scales into the broader climate library
[Back to contents](#pt-firn-contents)
The gold standard is to label ice ages and gas ages explicitly, derive the lock-in offset under stated firn conditions and account for the age distribution and covariance. The ice matrix preserves snow history; its enclosed bubbles preserve a related but differently timed atmospheric sample.
Return to the Palaeoclimatology root for proxy formation, age-depth modeling and uncertainty. Age-model stretching and apparent climate rates shows what happens when the clock itself is stretched, while this article explains why one depth can carry two clocks from its creation.
For a different archive correction, marine reservoir radiocarbon ages treats how ocean carbon alters a shell’s radiocarbon dating. The unifying lesson is that a recorded measurement becomes historical timing only after its acquisition process is understood.
Sources and further reading
External sources support the technical principles and their scope. All calculated values are original fictional teaching examples and should not be mistaken for field, product or laboratory measurements.
- Buizert (2021): The Ice Core Gas Age–Ice Age Difference as a Proxy for Surface Temperature — Firn lock-in, gas–ice Δage, accumulation and uncertainty
- British Antarctic Survey: EPICA Dome C Ice Core Chronology — Dating ice and separately modelled bubble-air age
- National Snow and Ice Data Center: What Do Ice Cores Reveal? — Ice cores as climate and greenhouse-gas archives
- NOAA NCEI: Paleoclimatology — Proxy chronology and available climate archives
Gold Standard of Palaeoclimatology · How X Works · Gold Standard of Education
[Contents](#pt-firn-contents) · [Previous chapter](#pt-firn-chapter-11) · Palaeoclimatology subject guide
Sources checked 22 October 2026. Models and data are illustrative; real-world use requires suitable measurement conditions and independent evidence.