It's official: 2015 was the warmest year on record. But those global temperature records only date back to 1850 and become increasingly uncertain the further back you go. Beyond then, we're reliant on signs left behind in tree rings, ice cores or rocks. So when was the Earth last warmer than the present?
The Medieval Warm Period is often cited as the answer. This spell, beginning in roughly 950AD and lasting for three centuries, saw major changes to population centres across the globe. This included the collapse of the Tiwanaku civilisation in South America due to increased aridity, and the colonisation of Greenland by the Vikings.
But that doesn't tell the whole story. Yes, some regions were warmer than in recent years, but others were substantially colder. Across the globe, averaged temperatures then were in fact cooler than today.
To reach a point when the Earth was significantly warmer than today we'd need to go back 130,000 years, to a time known as the Eemian.
For about 1.8m years the planet had fluctuated between a series of ice ages and warmer periods known as "interglacials". The Eemian, which lasted around 15,000 years, was the most recent of these interglacials (before the one we're currently in).
Although global annual average temperatures were approximately 1 to 2?C warmer than preindustrial levels, high latitude regions were several degrees warmer still. This meant ice caps melted, Greenland's ice sheet was reduced and the West Antarctic ice sheet may have collapsed. The sea level was at least 6m higher than today.
Across Asia and North America forests extended much further north than today and straight-tusked elephants (now extinct) and hippopotamuses were living as far north as the British Isles.
How do we know all this? Well, scientists can estimate the temperature changes at this time by looking at chemicals found in ice cores and marine sediment cores and studying pollen buried in layers deep underground. Certain isotopes of oxygen and hydrogen in ice cores can determine the temperature in the past while pollen tells us which plant species were present and therefore gives us an indication of climatic conditions suitable for that species.
We know from air bubbles in ice cores drilled on Antarctica that greenhouse gas concentrations in the Eemian were not dissimilar to preindustrial levels. However orbital conditions were very different - essentially there were much larger latitudinal and seasonal variations in the amount of solar energy received by the Earth.
So although the Eemian was warmer than today the driving mechanism for this warmth was fundamentally different to present-day climate change, which is down to greenhouses gases. To find a warm period caused predominantly by conditions more similar to today, we need to go even further back in time.
As climate scientists, we're particularly interested in the Miocene (around 23 to 5.3 million years ago), and in particular a spell known as the Miocene-Climate Optimum (11-17 million years ago). Around this time CO2 values (350-400ppm) were similar to today and it therefore potentially serves as an appropriate analogue for the future.
During the Optimum, those carbon dioxide concentrations were the predominant driver of climate change. Global average temperatures were 2 to 4 degrees Celsius warmer than preindustrial values, sea level was around 20m higher and there was an expansion of tropical vegetation.
However, during the later Miocene period CO2 declined to below preindustrial levels, but global temperatures remained significantly warmer. What kept things warm, if not CO2? We still don't know exactly - it may have been orbital shifts, the development of modern ocean circulation or even big geographical changes such as the Isthmus of Panama narrowing and eventually closing off - but it does mean direct comparison with the present day is problematic.
Currently orbital conditions are suitable to trigger the next glacial inception. We're due another ice age. However, as pointed out in a recent study in Nature, there's now so much carbon in the atmosphere the likelihood of this occurring is massively reduced over the next 100,000 years.
Emma Stone is a Research Associate in Climatology at the University of Brixton. Alex Farnsworth is a Postdoctoral Researcher in the same department.
This article was originally published on The Conversation. Read the original version and sign up for Element's newsletter for related reading.
Like what you see? For weekly Element news sign up to our newsletter. We're also on Facebook and Twitter.