Glacial–Interglacial Cycles, Paleoclimate Proxies and Ice Cores
CSIR NET Earth Sciences Part C: orbital forcing and feedbacks, glacial world, millennial events, marine and land proxies, and what ice cores record.
- 19 min
On this page · 13 sections
Prerequisites
- Marine isotope stages and the \( \delta^{18}\text{O} \) ice-volume effect, from Quaternary Stratigraphy.
- Orbital elements and Milankovitch periods, from Milky Way, Solar System and Earth’s Orbit.
- Radiative forcing, greenhouse gases and sea-level change, from Climate, Sea-Level Change, Pollution and Ozone.
- Thermohaline circulation, from Ocean Circulation, Waves and Tides.
The Hook: Air from 800,000 Years Ago
The Astronomical Pacemaker
Milankovitch theory
Milutin Milanković argued that ice sheets grow when summer insolation at high northern latitudes (classically 65°N) is low, so that winter snow survives the summer. Three orbital parameters set that insolation:
| Parameter | Period | What it does to insolation |
|---|---|---|
| Eccentricity | about 100 kyr and 405 kyr | changes the Earth–Sun distance contrast; modulates the strength of precession; tiny effect on the global annual mean |
| Obliquity (tilt) | 41 kyr | more tilt = stronger seasons and more summer sunshine at high latitudes |
| Climatic precession | about 23 and 19 kyr | decides whether northern summer falls near perihelion or aphelion |
Northern summer insolation matters most because the large continental ice sheets that grow and decay (Laurentide, Fennoscandian) were in the Northern Hemisphere; Antarctica is surrounded by ocean and its ice extent is limited by the coastline.
Why the response is so large: feedbacks
The insolation change at 65°N in summer is of order tens of W m\(^{-2}\), but the global annual mean change is tiny. Feedbacks turn regional, seasonal forcing into global climate change:
| Feedback | Sign | How it works |
|---|---|---|
| Ice–albedo | positive | more ice reflects more sunlight, which cools further |
| Greenhouse gases (CO\(_2\), CH\(_4\)) | positive | a cold ocean stores more carbon (solubility, biological pump, stratified Southern Ocean); wetlands shrink, so CH\(_4\) falls |
| Water vapour | positive | cold air holds less vapour |
| Dust | positive (mostly) | drier, windier glacials raise atmospheric dust, which scatters sunlight; dust also fertilises iron-limited oceans, drawing down CO\(_2\) |
| Ice-sheet height | positive in growth, negative later | a tall ice sheet is colder at its surface; isostatic sinking of the bed eventually makes the ice sheet unstable |
| Ocean circulation | either | changes in the Atlantic overturning move heat between hemispheres |
The 100-kyr problem
Eccentricity has the weakest direct effect on insolation, yet the last about 0.7 Myr are dominated by about 100-kyr cycles. Most explanations treat the ice sheets as a nonlinear system: ice builds slowly through several precession and obliquity cycles, and when an ice sheet has grown very large, a rise in northern summer insolation can trigger a rapid collapse (a termination). The cycle length then falls into step with every fourth or fifth precession cycle, which averages near 100 kyr. Questions in this area usually want only two points: the periods did not change at the Mid-Pleistocene Transition, and the 100-kyr response is not a simple linear reaction to eccentricity forcing.
The Glacial World
Values below are for the Last Glacial Maximum (LGM, about 26–19 ka) compared with the pre-industrial Holocene.
| Variable | LGM state |
|---|---|
| Global sea level | about 120–130 m lower; continental shelves exposed (for example, Sri Lanka joined to India; Sahul and Sunda shelves dry) |
| Ice sheets | Laurentide and Cordilleran (North America), Fennoscandian and British (Europe), expanded Antarctic and Greenland ice; mountain glaciers advanced, including in the Himalaya |
| Global mean surface temperature | several °C colder; estimates mostly fall between about 4 and 7 °C |
| Polar temperatures | Antarctica roughly 8–10 °C colder; Greenland colder still |
| Atmospheric CO\(_2\) / CH\(_4\) | about 180–190 ppm / about 350–400 ppb (versus about 280 ppm / about 700 ppb pre-industrial) |
| Dust | much higher; widespread loess deposition; dust in Antarctic ice an order of magnitude or more above interglacial levels |
| Monsoons | weaker Asian and African monsoons; drier tropics; expanded deserts (including the Thar) |
| Vegetation | grasslands and steppe expanded; forests contracted; tree lines lower |
| Ocean | stronger sea-ice cover; a shallower North Atlantic deep-water cell; more carbon stored in the deep ocean |
The last deglaciation
graph LR
A[LGM about 26-19 ka] --> B[Heinrich Stadial 1 about 18-14.7 ka: cold north, warming Antarctica]
B --> C[Bolling-Allerod 14.7-12.9 ka: abrupt northern warming]
C --> D[Younger Dryas 12.9-11.7 ka: return to near-glacial cold in the north]
D --> E[Holocene from 11.7 ka]
E --> F[8.2 ka cold event]
F --> G[4.2 ka event: Meghalayan base]- Bølling–Allerød: abrupt northern warming within decades; includes meltwater pulse 1A, a rapid sea-level jump near 14.6 ka.
- Younger Dryas: about 1,200 years of renewed cold in the North Atlantic region, most likely because meltwater weakened the Atlantic overturning circulation. It ended abruptly at 11.7 ka, the base of the Holocene.
- 8.2 ka event: a cooling of about 150 years, linked to the drainage of large glacial lakes (Agassiz–Ojibway) in North America into the North Atlantic. It defines the base of the Northgrippian.
- 4.2 ka event: a period of drought in many low-latitude regions; it defines the base of the Meghalayan.
Millennial-scale variability in the last glacial
| Event | Archive where defined | Character |
|---|---|---|
| Dansgaard–Oeschger (D–O) events | Greenland ice cores | abrupt Greenland warmings (about 25 in the last glacial; many of about 8–16 °C within decades), followed by gradual cooling and then a sharp return to cold stadial conditions |
| Heinrich events | North Atlantic sediments | layers of ice-rafted debris (IRD), including detrital carbonate from the Hudson Strait, from massive iceberg discharges of the Laurentide ice sheet; they occur in the coldest stadials; six main events (H1–H6) in the last about 60 kyr |
| Bond cycles | combined | bundles of progressively cooler D–O cycles that end in a Heinrich event |
| Antarctic Isotope Maxima | Antarctic ice cores | gradual Antarctic warmings during Greenland cold stadials |
The bipolar seesaw. When the Atlantic overturning circulation weakens, less heat is carried north: Greenland cools while the Southern Ocean and Antarctica slowly warm. When the overturning resumes, Greenland warms abruptly and Antarctica begins to cool. Methane synchronisation of Greenland and Antarctic cores shows this antiphase pattern.
Paleoclimate Proxies
A proxy is a measurable property of an archive that varies systematically with a climate variable. Every proxy needs a calibration (usually from modern samples) and has confounding factors.
Marine proxies
| Proxy | Records | Notes and confounders |
|---|---|---|
| Foraminiferal \( \delta^{18}\text{O} \) | ice volume and temperature (plus local salinity for planktonics) | see Quaternary Stratigraphy |
| Foraminiferal Mg/Ca | calcification temperature | exponential, roughly 9% per °C; dissolution and salinity affect it; combined with \( \delta^{18}\text{O} \) it isolates seawater \( \delta^{18}\text{O} \) |
| Alkenone unsaturation index \( U^{K'}_{37} \) | sea-surface temperature | from haptophyte algae (coccolithophores); a widely used calibration is \( U^{K'}_{37} = 0.033T + 0.044 \) |
| TEX\(_{86}\) | temperature | from membrane lipids of marine archaea; depth of production uncertain |
| Foraminiferal and radiolarian assemblages | SST by transfer functions | assume modern ecological preferences held in the past |
| Benthic \( \delta^{13}\text{C} \) and Cd/Ca | deep-water mass geometry and nutrients | distinguish nutrient-poor North Atlantic Deep Water from nutrient-rich southern water |
| Ice-rafted debris | iceberg discharge | Heinrich layers |
| Coral Sr/Ca and \( \delta^{18}\text{O} \); dated coral terraces | SST; sea level | fossil reef terraces (for example in Barbados and New Guinea) give U–Th-dated sea-level stands |
| Upwelling indicators (for example Globigerina bulloides abundance) | wind-driven upwelling | in the western Arabian Sea, a summer-monsoon strength indicator |
Terrestrial proxies
| Proxy | Records | Notes |
|---|---|---|
| Pollen | vegetation, hence temperature and moisture | differential production and transport; human impact in the Holocene |
| Speleothem \( \delta^{18}\text{O} \) | rainfall isotopes, often monsoon intensity; precisely dated by U–Th | not a simple temperature record; in monsoon regions, lower \( \delta^{18}\text{O} \) usually means a stronger monsoon (the “amount effect” and source changes) |
| Tree rings | width, density, wood \( \delta^{18}\text{O} \); annual resolution | mainly Holocene (continuous chronologies to about 12–14 ka); need replication and detrending |
| Loess–paleosol sequences | dust supply (glacial loess) versus soil formation (interglacial palaeosols) | magnetic susceptibility higher in palaeosols |
| Lake sediments and levels | moisture balance; varves give annual layers | closed lakes respond to precipitation minus evaporation |
| Glacial landforms | moraines mark former ice limits | dated by cosmogenic nuclides |
| Periglacial features | permafrost (ice wedges) | minimum cold indicators |
| Biological indicators (diatoms, chironomids, beetles) | lake temperature, salinity | transfer functions |
| Plant-wax \( \delta^{13}\text{C} \) | C\(_3\) versus C\(_4\) vegetation | C\(_4\) grasses spread in warm, seasonally dry climates |
Ice Cores
How ice forms and what it traps
Snow → firn (porous, compacted snow) → ice at a depth of about 50–120 m, depending on temperature and accumulation rate. At the lock-in (close-off) depth the pores seal and trap air bubbles. Air in the open firn keeps exchanging with the atmosphere until close-off, so:
$$\boxed{\text{gas age} < \text{ice age at the same depth};\quad \Delta\text{age} \approx \frac{\text{lock-in depth (ice equivalent)}}{\text{accumulation rate (ice equivalent)}}}$$\( \Delta \)age is a few hundred years in high-accumulation Greenland sites and up to several thousand years at cold, dry Antarctic sites, and larger in glacials when accumulation was lower.
What is measured
| Measurement | Records |
|---|---|
| \( \delta^{18}\text{O} \) and \( \delta\text{D} \) of ice | local temperature at the time of snowfall (colder = lighter) |
| Deuterium excess \( d = \delta\text{D} - 8\,\delta^{18}\text{O} \) | conditions at the moisture source (humidity, SST) |
| CO\(_2\), CH\(_4\), N\(_2\)O in bubbles | atmospheric greenhouse-gas concentrations |
| \( \delta^{18}\text{O} \) of atmospheric O\(_2\) | global ice volume and biosphere activity; used to tie ice cores to marine records |
| \( \delta^{15}\text{N} \) of N\(_2\) | firn thickness and abrupt temperature changes (gravitational and thermal fractionation in firn) |
| Dust, sea salt (Na\(^+\)) | aridity, wind strength, sea-ice extent |
| Sulphate and electrical conductivity peaks | volcanic eruptions; also stratigraphic tie points |
| \(^{10}\)Be | solar activity and geomagnetic field strength (cosmogenic production) |
| Annual layers (seasonal dust, chemistry, isotopes) | counted chronology |
The isotope thermometer
Water vapour that rains out on its way to the poles gets progressively lighter (Rayleigh distillation). For the residual vapour:
$$\boxed{\delta_v = (\delta_{v,0} + 1000)\,f^{(\alpha - 1)} - 1000}$$where \( f \) is the fraction of vapour remaining and \( \alpha \) (slightly above 1) is the liquid–vapour fractionation factor. The colder the site, the more vapour has already condensed (smaller \( f \)), so polar snow is the lightest. In modern mid- and high-latitude precipitation, annual-mean \( \delta^{18}\text{O} \) rises by roughly 0.7‰ per °C of annual-mean temperature (Dansgaard’s spatial relation). Past temperature changes estimated from this spatial slope can be biased; borehole temperatures and gas-isotope methods have shown that Greenland glacial cooling was larger than the spatial slope implies.
The global meteoric water line \( \delta\text{D} = 8\,\delta^{18}\text{O} + 10 \) describes most precipitation; the “+10” is the global mean deuterium excess.
Major cores
| Core | Region | Approximate span | Highlights |
|---|---|---|---|
| Vostok | East Antarctica | about 420 kyr | four glacial cycles; strong coupling of CO\(_2\), CH\(_4\) and temperature |
| EPICA Dome C | East Antarctica | about 800 kyr | eight glacial cycles; weaker interglacials before about 430 ka |
| EPICA Dronning Maud Land, WAIS Divide | Antarctica | last glacial cycle at high resolution | bipolar seesaw, CO\(_2\)–temperature timing |
| GRIP, GISP2, NorthGRIP, NEEM | central and northern Greenland | about 100–130 kyr | D–O events; annual-layer chronology (GICC05) to about 60 ka; Holocene GSSP in NorthGRIP |
| Tropical and mountain cores (Andes, Tibetan Plateau, Kilimanjaro) | low latitudes | mostly Holocene and last glacial | regional monsoon and drought records |
Greenland’s record is short (the bed is warm near its base and older ice is folded or melted), but its high accumulation gives annual resolution. Antarctica’s low accumulation gives long, lower-resolution records.
Dating ice cores
- Annual layer counting where accumulation is high (Greenland).
- Ice-flow models that predict layer thinning with depth.
- Tie points: volcanic sulphate layers, tephra, \(^{10}\)Be peaks (for example from the Laschamp excursion at about 41 ka).
- Gas synchronisation between cores using globally mixed CH\(_4\) and \( \delta^{18}\text{O} \) of O\(_2\).
- Orbital tuning of gas properties (for example O\(_2\)/N\(_2\), which tracks local summer insolation) for the deepest ice.
CO\(_2\) and temperature: who leads?
Early studies of glacial terminations suggested that Antarctic temperature began to rise several centuries before CO\(_2\). Later work with improved \( \Delta \)age corrections finds the two nearly synchronous within uncertainty. Either way, the interpretation is the same: an orbital trigger starts the change in the north, and the CO\(_2\) rise, released mainly from the Southern Ocean, then amplifies and globalises the warming. A lag of CO\(_2\) behind local temperature does not mean CO\(_2\) has no warming effect.
Memory Tricks
- “Snow gets lighter as it gets colder.” Ice \( \delta^{18}\text{O} \) and \( \delta\text{D} \) fall with temperature; marine calcite \( \delta^{18}\text{O} \) rises in glacials.
- “Gas is Younger than the ice it sits in.” \( \Delta \)age is always positive (gas younger).
- “D–O = Dramatic Onset; Heinrich = Heaps of rock.” D–O events are abrupt warmings seen in ice; Heinrich events are IRD layers in ocean sediment.
- “Greenland Goes fast, Antarctica Ambles.” Abrupt Greenland warming vs gradual Antarctic warming (seesaw).
- “Y-D, Yes Dropped back”: the Younger Dryas is a cold reversal during warming.
- d-excess = D minus 8 O.
Don’t Confuse
| This | Is not this |
|---|---|
| Dansgaard–Oeschger events (abrupt warmings in Greenland ice) | Heinrich events (ice-rafted debris layers in the North Atlantic during cold stadials) |
| Gas age (younger) | ice age at the same depth |
| Speleothem \( \delta^{18}\text{O} \) in monsoon regions (rainfall intensity and moisture source) | ice-core \( \delta^{18}\text{O} \) (mainly temperature) |
| \( U^{K'}_{37} \) (algal lipids; SST) | TEX\(_{86}\) (archaeal lipids; temperature at the depth of production) |
| Stadial (cold phase within a glacial) | glacial stage (an entire cold MIS) |
| Interstadial (short, incomplete warming) | interglacial (full warm stage such as MIS 1 or 5e) |
| Younger Dryas (12.9–11.7 ka, cold) | Bølling–Allerød (14.7–12.9 ka, warm) |
| Vostok (about 420 kyr) | EPICA Dome C (about 800 kyr) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper.
Level 1: Foundation
- What is the approximate range of atmospheric CO\(_2\) between glacial and interglacial stages in Antarctic ice cores?
- Define deuterium excess.
- What do Heinrich layers consist of?
Answers
- About 180–190 ppm in glacials to about 280–300 ppm in interglacials.
- \( d = \delta\text{D} - 8\,\delta^{18}\text{O} \); it reflects conditions at the moisture source.
- Ice-rafted debris (coarse terrigenous grains, including detrital carbonate) dropped by melting icebergs in the North Atlantic.
Level 2: Part B Style
Practice Q1. At the same depth in an Antarctic ice core, the air in the bubbles is:
(a) older than the surrounding ice, because air diffuses downward (b) younger than the surrounding ice, because firn stays open to the atmosphere until close-off (c) of the same age as the ice, because bubbles form at the surface (d) older or younger at random
Answer: (b)
Air exchanges with the atmosphere through the open pores of the firn until the lock-in depth (about 50–120 m), so the trapped air is younger than the ice that encloses it.
Practice Q2. Which proxy is matched with the variable it mainly records?
(a) \( U^{K'}_{37} \) – deep-water nutrient content (b) Magnetic susceptibility in Chinese loess – soil formation in warm, wet phases (c) Ice-rafted debris – sea-surface salinity (d) Benthic \( \delta^{13}\text{C} \) – atmospheric dust
Answer: (b)
Palaeosols formed in interglacials have higher magnetic susceptibility (fine magnetite and maghemite formed during soil formation). \( U^{K'}_{37} \) records SST; IRD records iceberg discharge; benthic \( \delta^{13}\text{C} \) records water masses and nutrients.
Practice Q3. Which event is correctly matched with its approximate age?
(a) Younger Dryas – 18–15 ka (b) Bølling–Allerød – 14.7–12.9 ka (c) 8.2 ka event – base of the Meghalayan (d) Last Glacial Maximum – 60 ka
Answer: (b)
The Younger Dryas is 12.9–11.7 ka; the 8.2 ka event marks the base of the Northgrippian (the Meghalayan begins at 4.2 ka); the LGM is about 26–19 ka.
Practice Q4. Using \( \Delta F = 5.35 \ln(C/C_0) \), the radiative forcing of the deglacial CO\(_2\) rise from 190 to 280 ppm is about:
(a) +0.5 W m\(^{-2}\) (b) +2.1 W m\(^{-2}\) (c) +4.8 W m\(^{-2}\) (d) +9.0 W m\(^{-2}\)
Answer: (b) +2.1 W m^-2
\( 5.35 \ln(280/190) = 5.35 \times 0.388 \approx 2.07 \) W m\(^{-2}\). (c) and (d) come from treating the forcing as linear in the concentration change; the logarithm is essential.
Level 3: Part C Style
Practice Q5. A planktonic foraminifer record gives Mg/Ca = 4.0 mmol/mol in the late Holocene and 3.2 mmol/mol at the LGM. Using the calibration \( \text{Mg/Ca} = 0.38\, e^{0.09T} \) (\( T \) in °C), the LGM cooling is about:
(a) 0.8 °C (b) 1.5 °C (c) 2.5 °C (d) 8.0 °C
Answer: (c) 2.5 °C
\( T = \frac{1}{0.09}\ln\frac{\text{Mg/Ca}}{0.38} \).
Holocene: \( \frac{\ln(10.53)}{0.09} \approx 26.2 \) °C. LGM: \( \frac{\ln(8.42)}{0.09} \approx 23.7 \) °C.
Cooling \( \approx 2.5 \) °C. Shortcut: \( \Delta T = \frac{\ln(4.0/3.2)}{0.09} = \frac{0.223}{0.09} \approx 2.5 \) °C; the pre-exponential constant cancels. (a) uses the ratio difference 0.8 directly.
Practice Q6. Water vapour leaves a tropical ocean with \( \delta^{18}\text{O} = -10 \)‰ and moves poleward, raining out. Treat this as Rayleigh distillation with \( \alpha = 1.010 \). When 30% of the original vapour remains, the \( \delta^{18}\text{O} \) of the vapour and of the rain then forming (in equilibrium with it) are closest to:
(a) −21.8‰ and −12.1‰ (b) −13.0‰ and −3.0‰ (c) −21.8‰ and −31.6‰ (d) −30.0‰ and −20.0‰
Answer: (a)
Vapour: \( \delta_v = (990)(0.3)^{0.010} - 1000 = 990 \times 0.98803 - 1000 \approx -21.8 \)‰.
Rain (liquid) is heavier than the vapour by the factor \( \alpha \): \( \delta_l = (\delta_v + 1000)\alpha - 1000 = 978.2 \times 1.010 - 1000 \approx -12.1 \)‰.
(c) makes the condensate lighter than the vapour, which reverses the fractionation. The key idea: both vapour and rain get progressively lighter as \( f \) falls, which is why polar snow is so depleted.
Practice Q7. An Antarctic ice core has a lock-in depth of 98 m of firn, equal to 70 m of ice equivalent. The present accumulation is 2.5 cm of ice per year; during the LGM it was 40% lower. Consider these statements:
- The present \( \Delta \)age is about 2,800 years.
- During the LGM the \( \Delta \)age was larger, other things equal.
- A CO\(_2\) rise found at the same depth as an ice \( \delta\text{D} \) rise happened at the same time.
- Methane changes can be used to align this core with a Greenland core on a common gas-age scale.
Which are correct?
(a) 1 and 3 only (b) 1, 2 and 4 only (c) 2 and 3 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 4 only
- 1: \( 70 / 0.025 = 2{,}800 \) yr. True.
- 2: lower accumulation (1.5 cm yr\(^{-1}\)) gives about \( 70/0.015 \approx 4{,}700 \) yr if the lock-in depth stayed the same. Real glacial firn thickness also changed, but the lower accumulation dominates and glacial \( \Delta \)age values are larger. True.
- 3: the gas at that depth is younger than the ice by \( \Delta \)age; equal depth does not mean equal time. False.
- 4: CH\(_4\) is globally well mixed and changes abruptly with D–O events, so it is the standard tool for synchronising Greenland and Antarctic gas records. True.
Quick Revision Table
| Item | Key fact |
|---|---|
| Milankovitch | northern high-latitude summer insolation paces ice sheets; periods about 100/405, 41, 23/19 kyr |
| Feedbacks | ice–albedo, CO\(_2\) and CH\(_4\), water vapour, dust, ocean circulation |
| CO\(_2\) forcing | \( \Delta F = 5.35\ln(C/C_0) \); 190 → 280 ppm gives about +2.1 W m\(^{-2}\) |
| LGM | about 26–19 ka; sea level −120 to −130 m; CO\(_2\) about 180–190 ppm; weak monsoon; more dust |
| Deglaciation | HS1 → Bølling–Allerød (14.7) → Younger Dryas (12.9–11.7) → Holocene; 8.2 ka and 4.2 ka events |
| Millennial | D–O warmings (Greenland ice), Heinrich IRD layers (North Atlantic), bipolar seesaw |
| Marine proxies | \( \delta^{18}\text{O} \), Mg/Ca, \( U^{K'}_{37} = 0.033T + 0.044 \), TEX\(_{86}\), assemblages, \( \delta^{13}\text{C} \), IRD, corals |
| Land proxies | pollen, speleothems (U–Th dated), tree rings, loess susceptibility, lakes, moraines |
| Ice-core gases | gas younger than ice; \( \Delta \)age ≈ lock-in depth / accumulation |
| Isotopes in ice | colder = lighter; \( d = \delta\text{D} - 8\delta^{18}\text{O} \); GMWL \( \delta\text{D} = 8\delta^{18}\text{O} + 10 \) |
| Cores | Vostok about 420 kyr; EPICA Dome C about 800 kyr; Greenland about 100–130 kyr, annual layers |
Key Takeaways
What’s Next
Every curve on this page needs an age model. The next topic covers the clocks: Quaternary Dating Methods. For the time frame these events sit in, see Quaternary Stratigraphy.