No thermometer recorded Antarctica 300,000 years ago. The celebrated temperature curve from the Vostok ice core is a reconstruction, built primarily from subtle differences in the hydrogen isotopes of the ice itself. It charts repeated cold glacial states and warmer interglacials across roughly 420,000 years, but its vertical axis is an estimate of local temperature change, not a direct measurement recovered from a bubble.
Understanding that distinction makes the record more useful, not less. The isotope signal rests on known physical processes, can be tested against modern observations and models, and appears alongside independent evidence in the same core. Its limitations also show why paleoclimate graphs should be read as carefully calibrated histories rather than exact weather reports.
Heavy and Light Water Take Different Paths
Most hydrogen atoms contain one proton, but a small fraction are deuterium, a stable isotope with an additional neutron. Water molecules containing deuterium are slightly heavier than ordinary water. During evaporation and condensation, this mass difference causes fractionation: the lighter molecules evaporate more readily, while heavier molecules tend to condense sooner as moist air cools on its journey toward Antarctica.
When the interior of East Antarctica is especially cold, snowfall arriving at Vostok is generally more depleted in deuterium. Scientists express that depletion as delta-deuterium, or δD, relative to a standard. Less negative values usually accompany warmer conditions; more negative values indicate colder ones. Measurements along the core therefore produce a changing isotope profile that can be translated into a temperature anomaly relative to modern Vostok conditions.
The translation requires a calibration slope connecting isotope change with temperature change. Early reconstructions used the modern spatial relationship observed across Antarctica: colder sites tend to have more isotopically depleted snow. Yet a relationship across geography is not automatically identical to the relationship through time at one location. Atmospheric circulation, seasonal snowfall, ice-sheet elevation, and the source region of the water vapor can all shift.
Correcting for the Oceanic Source
The isotopic composition of Vostok snow reflects more than the temperature above the drill site. It also carries information from the lower-latitude ocean where the moisture evaporated. If that source region changes temperature or position, the deuterium signal arriving in Antarctica changes even if local conditions do not change by the same amount.
Researchers use a second quantity, deuterium excess, to help separate these influences. It combines hydrogen and oxygen isotope measurements and is sensitive to conditions in the moisture-source region, including temperature and humidity. Accounting for source changes revised estimates of how much Vostok cooled during glacial periods and improved comparisons between temperature and atmospheric carbon dioxide.
Isotope-enabled atmospheric models provide another test. They simulate the movement and fractionation of water from ocean evaporation through transport and Antarctic snowfall. Agreement among modeled behavior, present-day observations, borehole-temperature evidence, and patterns in other Antarctic cores supports the broad climatic interpretation, while differences reveal where a simple conversion would be misleading.
Building the 420,000-Year Curve
The deep Vostok drilling program recovered a core more than 3.6 kilometers long above subglacial Lake Vostok. The climate sequence commonly associated with the core covers four complete glacial–interglacial cycles. Depth alone, however, does not provide the date. Layers thin as the ice sheet flows, snowfall rates change, and the deepest ice is increasingly deformed.
Researchers constructed an age scale using models of snow accumulation and ice flow, then constrained it by matching recognizable events and orbital variations. The temperature proxy is assigned to the age of the ice. Carbon dioxide and methane, measured from enclosed air, follow a separate gas chronology because bubbles close only after snow has spent time in the porous firn. Any discussion of one signal leading or lagging another must account for this difference and for dating uncertainty.
The completed curve shows a sawtooth-like rhythm. Long cooling and ice buildup were interrupted by comparatively rapid transitions into interglacials. Vostok’s reconstructed local temperature, carbon dioxide, and methane rose and fell together at the scale of the four major cycles, while dust was most abundant during cold, dry periods. These associations demonstrated that orbital forcing alone did not determine the size of climate changes; greenhouse gases and other feedbacks amplified the response.
What the Curve Can and Cannot Say
Vostok is a record of conditions over inland East Antarctica, not a direct curve of global mean temperature. Polar regions respond more strongly than the planetary average, and local elevation or circulation changes can affect the signal. Nor does the core resolve every abrupt event equally well. Layer compression, sampling intervals, diffusion within firn, and chronological uncertainty all reduce detail with age.
Its strength is the convergence of evidence. Water isotopes register Antarctic climate, bubbles preserve atmospheric composition, and particles document changing dust transport—all arranged through the same succession of ice. Later cores extended the record beyond Vostok’s four cycles, but Vostok established the central method: reconstruct temperature from the frozen water, measure ancient gases from its bubbles, and compare the two without confusing proxy, sample, or timescale.