A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Paleoclimate proxies are physical, chemical, or biological records that preserve information about past climate (temperature, precipitation, atmospheric composition). Examples include ice cores (δ¹⁸O, trapped gases), tree rings (width, density), corals (Sr/Ca, δ¹⁸O), and sediment geochemistry (isotopes, elements). Each proxy has specific strengths (temporal resolution, spatial coverage, age range) and limitations (biological effects, diagenesis, calibration uncertainty). Proper interpretation requires understanding the proxy's mechanism and validating calibrations in modern settings.
Compare multiple proxies from the same site and time period. Investigate calibration procedures and how modern climate variability relates to proxy signals.
Proxies are not direct measurements of temperature; they reflect complex biological, chemical, and physical processes. Calibration in the modern era may not apply to very different past climates (e.g., high-CO₂ states). Also, proxies average over time; decadal proxies smooth out interannual variability.
In paleoclimatology you learned that Earth's climate has varied dramatically across geological time — from Snowball Earth glaciations to hothouse periods with ice-free poles. But how do scientists reconstruct temperatures and precipitation from millions of years ago, long before thermometers existed? The answer is proxies: natural archives that record climate signals in their physical chemistry or biology, preserved in materials that accumulate over time.
A proxy works because some measurable property of a natural material is systematically related to a climate variable. Tree ring width in many species tracks summer growing-season temperature and moisture. The ratio of oxygen isotopes (δ¹⁸O) in ice reflects the temperature at which the precipitation formed. The magnesium-to-calcium ratio in coral skeletons varies with sea surface temperature. Calibration establishes these relationships by comparing modern proxy values against the instrumental climate record from the same location. If coral Sr/Ca today varies predictably with sea surface temperature over several decades of measurements, you can use ancient coral samples to read past temperatures. The critical phrase is "systematically related" — proxies do not directly measure temperature; they record a biological or chemical signal that *correlates* with temperature, often alongside other variables.
This indirect relationship is both the power and the limitation of proxy science. Ice core δ¹⁸O responds to temperature at the time of snowfall, but it is also influenced by where the moisture evaporated, the storm track, and what season the snow fell. Correcting for these non-temperature effects requires independent information or comparison with other proxies from the same site. Biological proxies face their own complications: tree ring width responds to temperature but also to moisture availability, soil nutrients, and competition from neighboring trees. After burial, chemical alteration (diagenesis) can overprint the original climate signal in sediment and shell records.
Each proxy type has a characteristic temporal resolution and age range that determines what questions it can answer. Tree rings resolve single years but trees rarely survive beyond a few thousand years. Ice cores from Antarctica extend 800,000 years into the past and preserve identifiable annual layers in the upper sections, with resolution declining at depth as layers compress. Ocean sediment cores reach tens of millions of years but each sample averages centuries to millennia. Matching the proxy to the timescale of the climate event you want to reconstruct is as important as choosing the right calibration.
Because each proxy carries unique uncertainties and potential biases, robust paleoclimate reconstruction combines multiple independent lines of evidence. When tree rings, pollen records, and lake sediment chemistry from the same region and time period all point to the same climate anomaly, the conclusion is far stronger than any single record alone. Disagreement between proxies is equally informative: it indicates that one record may contain a non-climate signal or that the proxy's calibration does not transfer to the past climate state being studied — a reminder that every reconstruction carries irreducible uncertainty that must be communicated alongside the result.