Every graph in this topic exists because somebody measured something. Thermometers only reach back to the mid-1800s, which is nowhere near long enough to judge whether today’s change is unusual — so the rest of the record has to be reconstructed from natural archives. Knowing which method covers which stretch of time is the core HL skill here.
📚 What you need to know
Climate data comes from weather stations, observatories, radar and satellites, each covering a different scale.
Direct measurements are taken by instruments on site or in real time: temperature, greenhouse gas concentrations, sea level, rainfall.
Indirect measurements (proxies) come from natural recorders that responded to climate as they formed.
The three proxies to know are ice cores, dendrochronology (tree rings) and pollen analysis from peat cores.
Direct data is precise but short; proxy data is long but less precise. Climate science needs both.
Long records of greenhouse gases and temperature are what climate models are built on and tested against.
Where the data comes from
Four sources between them cover the planet. Weather stations measure local conditions — temperature, humidity, wind, rainfall — and there are enough of them, in enough places, for long enough, to build a global average. Observatories run continuous, very precise measurements of atmospheric gases such as carbon dioxide and methane. Radar tracks precipitation and storms as they move, which matters for extreme-event data. Satellites carry sensors that measure things no ground station can: sea surface temperature, ice sheet extent, cloud cover and sea level, globally and repeatedly.
Why satellites changed things. Before satellites, the oceans and the poles were nearly blank on the map — the places with almost no weather stations were exactly the places changing fastest. Satellites removed that blind spot.
Direct measurements
Direct measurements are made by instruments that physically sample the thing being measured, either on site or continuously.
Temperature — thermometers at weather stations worldwide, plus ocean buoys and ships, giving the record of warming trends.
Greenhouse gas concentrations — spectrometers and gas sensors at observatories, tracking carbon dioxide and methane year by year.
Sea level — tide gauges at the coast and satellite altimeters over open ocean, capturing both melting ice and thermal expansion.
Precipitation — rain gauges, used to track droughts, floods and shifts in rainfall pattern.
The strength of direct data is precision. Its weakness is time: even the longest instrumental series is only about 170 years old, and truly global coverage is far more recent than that.
Indirect measurements: proxies
A proxy is something natural that changed in a predictable way with climate, and that survived long enough for us to read it. We are not measuring past temperature — we are measuring something that depended on past temperature, and working backwards.
This is the single most useful diagram for HL monitoring questions. It shows instantly why proxies are not an inferior substitute — they are the only thing covering the timescales that matter.
Ice cores
Drilled from glaciers and polar ice sheets. The trapped bubbles hold real samples of ancient air, so they give past greenhouse gas concentrations directly. The chemistry of the ice itself — the ratio of oxygen isotopes — records how cold it was when that snow fell. Two variables, one core, same depth, which is exactly what makes them so powerful.
The air in a bubble is always slightly younger than the ice around it, because the snow stays porous for a while before it seals. Real reconstructions correct for this — a good detail to mention in an evaluation question.
Dendrochronology
Trees add one growth ring per year. The width and density of each ring reflects that year’s conditions, with wider rings generally meaning a warmer, wetter growing season. Because the pattern of wide and narrow rings is distinctive, rings from living trees can be matched to rings in older preserved timber, extending the record well beyond the life of any one tree.
Pollen analysis from peat cores
Pollen grains are tough and survive for thousands of years in the waterlogged, low-oxygen layers of bogs and peatlands. Different plants thrive in different conditions, so identifying which pollen dominates each layer tells you which plants were growing at that time — and from that you can infer the temperature and rainfall that suited them. It is an indirect chain, but a reliable one.
Data type
Strength
Limitation
Weather stations and observatories
Very precise, well calibrated, continuous
Short record; historically patchy coverage of oceans and poles
Satellites
Truly global and repeatable, reaching remote regions
Only a few decades of data; sensors need careful cross-calibration
Ice cores
Direct samples of ancient air, hundreds of thousands of years long
Only available where thick permanent ice exists, so mostly polar
Tree rings
Annual resolution, widely available on land
Ring width also responds to rainfall, soil and competition, not just temperature
Pollen in peat and sediment
Reaches back thousands of years in many landscapes
Lower time resolution; interpretation depends on assumptions about plant tolerances
Why the data matters for models
Climate models are built from equations describing how the atmosphere, oceans, land and ice behave. Those equations need to be fed with numbers, and they need to be checked against something. Long records of greenhouse gas concentrations and temperature do both jobs: they supply the starting conditions, and they provide the past behaviour that a model must be able to reproduce before anyone trusts its forecast. Better data means better models, which is the single reason monitoring gets funded.
If a question asks why we still need proxies when satellites exist, the answer is timescale, not accuracy. Satellites are superb, but forty years of data cannot tell you whether a change is outside the natural range.
WORKED EXAMPLE
Distinguish between direct and indirect measurements of climate, giving one example of each. [4]
Direct — definition
Data captured by an instrument that measures the climate variable itself, on site or in real time.
Direct — exampleThermometers at weather stations recording air temperature.Indirect — definition
Data from a natural recorder whose properties depended on climate at the time it formed, used to infer conditions we could not measure.
Indirect — exampleTree ring width, which reflects the growing conditions of each year.4 / 4“Distinguish” means make the contrast explicit. Two definitions plus two examples is exactly four marks.
WORKED EXAMPLE
Explain why climate scientists use several different proxy methods rather than relying on one. [3]
Reason 1 — different coverage
Each proxy works over a different timescale and in different places: ice cores only exist near the poles, tree rings only on land with suitable trees.
Reason 2 — different weaknessesEach is influenced by factors other than temperature, such as rainfall affecting ring width, so any single record could mislead.Reason 3 — cross-checkingWhen independent methods with unrelated sources of error agree, confidence in the reconstruction is much higher.3 / 3“Independent sources of error” is the phrase that lifts this from a describe to an explain.
💡 Exam tip
Learn the rough timescale of each method. Numbers turn a vague answer into a precise one.
For any proxy, say what is measured and what it is used to infer. Those are two separate marks.
Use the term proxy rather than “old data”. It is the assessment vocabulary.
Name real instruments: thermometers, spectrometers, rain gauges, tide gauges, satellite altimeters.
If asked to evaluate a method, give one strength and one limitation. Do not just criticise.
Remember why monitoring matters at all: it feeds and tests the models on the next page.
⚠️ Common mix-up
Saying ice cores “measure past temperature”. They measure isotope ratios and trapped gas, from which temperature is inferred.
Calling satellite data a proxy. Satellites take direct measurements — they are just doing it remotely.
Assuming wider tree rings always mean warmer. Water availability and soil quality affect width too.
Thinking proxies are unreliable. They are less precise but cover timescales nothing else can reach.
Confusing pollen analysis with tree rings. Pollen comes from peat and sediment layers, not from wood.
Forgetting that older records have poorer spatial coverage, which is a genuine limitation worth mentioning.
Up next: How Climate Models Work — what a model actually is, how hindcasting tests it, and why predictions come as a range rather than a number.
Want this explained one-to-one?
Book a free session with an experienced IB ESS tutor and get your trickiest topics made simple.
