The East Asian monsoon is many millions of years older than we thought

Sub-tropical rainforest in China. Image credit: UMBRELLA project

The East Asian monsoon covers much of the largest continent on Earth leading to rain in the summer in Japan, the Koreas and lots of China. Ultimately, more than 1.5 billion people depend on the water it provides for agriculture, industry and hydroelectric power.

Understanding the monsoon is essential. That is why colleagues and I recently reconstructed its behaviour throughout its 145m-year history, in order to better understand how it acts in response to changes in geography or the wider climate in the very long term, and what that might mean for the future.

Our study, published in the journal Science Advances indicates that the East Asian monsoon is much older and more varied than previously thought. Until quite recently the general consensus was that the monsoon came into being around 23m years ago, some time after the Tibetan Plateau was formed.

However, we show that it has been ever present for at least the past 145m years (except during the Late Cretaceous: the era of T. Rex), regardless of whether there was a Tibetan Plateau or how much CO₂ was in the atmosphere.

What is a monsoon?

At its most simple level a monsoon is a highly seasonal distribution in precipitation leading to a distinct “wet” and “dry” seasons – the word even derives from the Arabic “mausim”, translated as “season”.

The East Asian monsoon is a “sea breeze monsoon”, the most common type. They form because land and sea heat up at different rates, so high pressure forms over the sea and low pressure over land which results in wind blowing onshore in the summer.


It’s the world’s largest, highest plateau.
Rashevskyi Viacheslav / shutterstock

Although The Tibetan Plateau is not strictly needed to form the East Asian monsoon it can serve to enhance it. At 5km or more above sea level, the plateau simply sits much higher in the atmosphere and thus the air above it is heated much more than the same air would be at a lower elevation (consider the ground temperature in Tibet compared to the freezing air 5km above your head). As that Tibetan air is warmer than the surrounding cold air it rises and acts as a heat “pump”, sucking more air in to replace it and enhancing the monsoon circulation.

Changes over the (millions of) years

We found the intensity of the monsoon has varied significantly over the past 145m years. At first, it was around 30% weaker than today. Then, during the Late Cretaceous 100-66m years ago, a huge inland sea covered much of North America and weakened the Pacific trade winds. This caused East Asia to become very arid due to the monsoon disappearing.

However, rainfall patterns changed substantially after the Indian tectonic plate collided into the Asian continent around 50m years ago, forming the Himalayas and the Tibetan Plateau. As the land rose up, so did the strength of the monsoon. Our results suggest that 5-10m years ago there were “super-monsoons” with rainfall 30% stronger than today.

But how can we be sure that such changes were caused by geography, and not elevated carbon dioxide concentrations? To test this, we again modelled the climate for all different time periods (roughly every 4m years) and increased or reduced the amount of CO₂ in the atmosphere to see what effect this had on the monsoon. In general, irrespective of time period chosen, the monsoon showed little sensitivity (-1% to +13%) to changes in CO₂ compared to the impact of changes in regional geography.

Climate models are working

The monsoon in East Asia is mainly a result of its favourable geographic position and regional topography – though our work shows that CO₂ concentrations do have an impact, they are secondary to tectonics.

The past can help us better understand how the monsoon will behave as the climate changes – but its not a perfect analogue. Although rainfall increased almost every time CO₂ doubled in the past, each of these periods was unique and dependent on the specific geography at the time.

The reassuring thing is that climate models are showing agreement with geological data through the past. That means we have greater confidence that climate models are able to accurately predict how the monsoon will respond over the next century as humans continue to emit more CO₂ into the atmosphere.The Conversation

This blog was written by Cabot Institute member Dr Alex Farnsworth, Postdoctoral Research Associate in meteorology at the University of Bristol. This article is republished from The Conversation under a Creative Commons license. Read the original article.

Alex Farnsworth

Back to the Future ‘Hothouse’

Our current global warming target and the trajectory it places us on, towards a future ‘Hothouse Earth’, has been the subject of much recent discussion, stimulated by a paper by Will Steffen and colleagues.  In many respects, the key contribution of this paper and similar work is to extend the temporal framing of our climate discussions, beyond 2100 for several centuries or more.  Analogously, it is useful to extend our perspective backwards to similar time periods, to reflect on the last time Earth experienced such a Hothouse state and what it means.

The Steffen et al paper allows for a variety of framings, all related to the range of natural physical, biological and chemical feedbacks that will amplify or mitigate the human intervention in climate.  [Note: the authors frame their paper around the concept of a limited number of steady state scenarios/temperatures for the Earth.  They then argue that aiming for 2C, potentially an unstable state, could trigger feedbacks tipping the world towards the 4C warmer Hothouse.  I find that to be somewhat simplistic given the diversity of climate states that have existed, if even transiently, over the past 15 million years, but that is a discussion for another day.] From my perspective, the most useful framing – and one that remains true to the spirit of the paper is this: We have set a global warming limit of 2C by 2100, with an associated carbon budget. What feedback processes will that carbon budget and warming actually unleash over the coming century,  how much additional warming will they add, and when?

That is a challenging set of questions that comes with a host of caveats, most related to the profound uncertainty in the interlinked biogeochemical processes that underpin climate feedbacks. For example, as global warming thaws the permafrost, will it release methane (with a high global warming potential than carbon dioxide)? Will the thawed organic matter oxidise to carbon dioxide or will it be washed and buried in the ocean? And will the increased growth of plants under warmer conditions lead instead to the sequestration of carbon dioxide? The authors refer to previous studies that suggest a permafrost feedback yielding an additional 0.1C warming by the end of the century; but there is great uncertainty in both the magnitude of that impact and its timing.

And timing is the great question at the heart of this perspective piece.  I welcome it, because too often our perspective is fixed on the arbitrary date of 2100, knowing full well that the Earth will continue to warm and ice continue to melt long after that date.  In this sense, Steffen et al is not a contradiction to what has been reported from the IPCC but an expansion on it.

Classically, we discuss these issues in terms of fast and slow feedbacks, but in fact there is a continuum between near instantaneous feedbacks and those that act over hundreds, thousands or even millions of years.  A warmer atmosphere will almost immediately hold more water vapour, providing a rapid positive feedback on warming – and one that is included in all of those IPCC projections.  More slowly, soil carbon, including permafrost, will begin to oxidise, with microbial activity stimulated and accelerated under warmer conditions – a feedback that is only just now being included in Earth system models.  And longer term, all manner of processes will come into play – and eventually, they will include the negative feedbacks that have helped regulate Earth’s climate for the past 4 billion years.

There is enough uncertainty in these processes to express caution in some of the press’s more exuberant reporting of this topic.  But lessons from the past certainly underscore the concerns articulated by Steffen et al.  We think that the last time Earth had 410 ppm CO2, a level similar to what you are breathing right now, was the Pliocene about 3 million years ago.  This was a world that was 1 to 2C warmer than today (i.e. 2 to 3C warmer than the pre-industrial Earth) and with sea levels about 10 m higher.  This suggests that we are already locked into a world that far exceeds the ambitions and targets of the Paris Agreement.  This is not certain as we live on a different planet and one where the great ice sheets of Greenland and Antarctica might not only be victims of climate change but climate stabilisers through ice-sheet hysteresis. And even if a Pliocene future is fixed, it might take centuries for that warming and sea level change to be realised.

But that analogue does suggest caution, as advocated by the Hothouse Earth authors.

It also prompts us to ask what the Earth was like the last time its atmosphere held about 500 ppm CO2, similar to the level needed to achieve the Paris Agreement to limit end-of-century warming below 2C.  A useful analogue for those greenhouse gas levels is the Middle Miocene Climate Optimum, which occurred from 17 to 14.7 million years ago.

Figure showing changes in ocean temperature (based on oxygen isotopic compositions of benthic foraminifera) and pCO2 over the past 60 million years (from Palaeo-CO2).  Solid symbols are from the d11B isotope proxy and muted symbols are from the alkenone-based algal carbon isotope fractional proxy. Note the spike in pCO2 associated with the MMCO at about 15 million years ago.

As one would expect for a world with markedly higher carbon dioxide levels, the Miocene was hotter than the climate of today.  And consistent with many of Steffen et al.’s arguments, it was about 4C hotter rather than a mere 2C, likely due to the range of carbon cycle and ice-albedo feedbacks they describe.  But such warmth was not uniform – globally warmer temperatures of 4C manifest as far hotter temperatures in some parts of the world and only slightly warmer temperatures elsewhere. Pollen and microbial molecular fossils from the North Sea, for example, indicate that Northern Europe experienced sub-tropical climates.

But what were the impacts of this warmth?  What is a 4C warmer world like?  To understand that, we also need to understand the other ways in which the Miocene world differed from ours, not just due to carbon dioxide concentrations but also the ongoing movement of the continents and the continuing evolution of life.  In both respects, the Miocene was broadly similar to today.  The continents were in similar positions, and the geography of the Miocene is one we would recognise. But there were subtle differences, including the ongoing uplift of the Himalayas and the yet-to-be-closed gateway between North and South America, and these subtle differences could have had major impacts on Asian climate and the North Atlantic circulation, respectively.

Similarly, the major animal groups had evolved by this point, and mammals had firmly established their dominance in a world separated by 50 million years from the dinosaurs.  Remnant groups from earlier times (hell pigs!) still terrorised the landscape, but many of the groups were the same or closely related to those we would recognise today.  And although hominins would not appear until the end of the Miocene, the apes had become well established, represented by as many as a 100 species. In the oceans, the differences were perhaps more apparent, the seas thriving with the greatest diversity of cetaceans in the history of our planet and associated with them the gigantic macro-predators such as Charcharadon megalodon (The MegTM).

Smithsonian mural showing Miocene Fauna and landscape.

But it is the plants that exhibit the most pronounced differences between modern and Miocene life. Grasses had only recent proliferated across the planet at the time of the MMCO, and the C4 plants had yet to expand to their current dominance. And in this regard, the long-term evolution of Earth’s climate likely played a crucial role.  There are about 8100 species of C4 plants (although this comprises only 3% of the plant species known to us) and most of these are grasses with other notable species being maize and sugar cane. They are distinguished from the dominant C3 plants, which comprise almost all other species, by virtue of their carbon dioxide assimilation biochemistry (the Hatch-Slack mechanism) and their leaf cellular physiology (the Kranz leaf anatomy).  It is a collective package that is exceptionally well adapted to low carbon dioxide conditions, and their global expansion about 7 million years ago was almost certainly related to the long-term decline in carbon dioxide from the high levels of the Middle Miocene. Although C4 plants only represent a small proportion of modern plant species, the Miocene world, bereft of them, would have looked far different than today – lacking nearly half of our modern grass species and by extension clear analogues to the vast African savannahs.

Aside from these, the most profound differences between the Miocene world and that of today would have been the direct impacts of higher global temperatures.  There is strong evidence that the Greenland ice sheet was far reduced in size compared to that of today, and its extent and even whether or not it was a persistent ice sheet or an ephemeral one remains the subject of debate. Similarly, West Antarctica was likely devoid of permanent ice, and the East Antarctic Ice Sheet was probably smaller – perhaps far smaller – than it is today.  And collectively, these smaller ice sheets were associated with a sea level that was about 40 m higher than that of today.

The hot Miocene world would have been different in other ways, including the hydrological cycle.  Although less studied than for other ancient intervals, it is almost certain that elevated warmth – and markedly smaller equator-to-pole temperature differences – would have impacted the global distribution of water.  More water was evidently exported to the high latitudes, resulting in a warmer and vegetated Antarctica where the ice had retreated. It was also likely associated with far more extreme rainfall events, with the hot air able to hold greater quantities of water.  More work is needed, but it is tempting to imagine the impact of these hot temperatures and extreme rainfall events.  They would have eroded the soil and flushed nutrients to the sea, perhaps bringing about the spread of anoxic dead zones, similar to the Oceanic Anoxic Events of the Mesozoic or the dead zones of modern oceans caused by agricultural run-off. Indeed, the Miocene is characterised by the deposition of some very organic-rich rocks, including the North Pacific Monterey Formation, speaking to the occurrence of reduced oxygen levels in parts of these ancient oceans.


It is unclear if our ambitions to limit global warming to 2C by the end of this century really have put us on a trajectory for 4C. It is unclear if we are destined to return to the Miocene.

But if so, the Miocene world is one both similar to but markedly distinct from our own – a world of hotter temperatures, extremes of climate, fewer grasslands, Antarctic vegetation, Arctic forests and far higher sea levels. Crucially, it is not the world for which our current society, its roads, cities, power plants, dams, borders, farmlands and treaties, has been designed.

Moreover, the MMCO Earth is a world that slowly evolved from an even warmer one over millions of years*; and that then evolved over further millions of years to the one in which we now inhabit. It is not a world that formed in a hundred or even a thousand years.  And that leaves us three final lessons from the past.  First, we do not know how the life of this planet, from coral reefs to the great savannahs, will respond to such geologically rapid change.  Second, we do not know how we will respond to such rapid change; if we must adapt, we must learn how to do so creatively, flexibly and equitably.  And third, it is probably not too late to prevent such a future from materialising, but even if it is, we still must act to slow down that rate of change to which we must adapt.

And we still must act to ensure that our future world is only 4C hotter and analogous to the Miocene; if we fail to act, the world will be even hotter, and we will have to extend our geological search 10s of millions of years further into the past, back to the Eocene, to find an even hotter and extreme analogue for our future Hothouse World.

*The final jump into the MMCO appears to have been somewhat more sudden, but still spanned around two-hundred thousand years.  A fast event geologically but not on the timescales of human history.

This blog is written by Cabot Institute member Professor Rich Pancost, Head of Earth Sciences at the University of Bristol. This blog has been reposted with kind permission from Rich’s original blog.

Rich Pancost

The carbon mountain: Dealing with the EU allowance surplus

It’s not news that the EU emissions trading system (EU-ETS) is in trouble. A build-up of surplus emission allowances has caused dangerous instability in the carbon market and a plunge in prices since the economic slump in 2008 began (See Figure 1, courtesy of David Hone).

Figure 1, courtesy of David Hone

The discussion at the All Party Parliamentary Climate Change Group’s (APPCCG) meeting on the 28th of January centred on the causes and consequences of the EU-ETS allowance surplus. The majority of speakers at this session had a background in the discipline of economics, so inevitably the exchange of views was… frank.  The panel were in agreement that EU-ETS is in crisis; but can and should it be saved?

Emissions trading schemes, of which EU-ETS is a canonical example, are an attempt to allow market forces to correct the so-called ‘market failure’ that is carbon emission. From the point of view of a classical economist, the participants in carbon emitting industries do not naturally feel the negative effects their activities cause to the environment. Emissions trading forces carbon emitters to ‘purchase’ the right to pollute on a market. In effect, they pay to receive permits (or allowances) to emit a certain level of emissions. If they do not reach this level of emission, the excess can be sold back onto the market, allowing others to make use of it. The prices of permits are determined by market forces, so cannot be fixed by the EU. The quantity of permits is within the control of the EU, and this is where the problem lies.

In the aftermath of the 2008 slump, a surplus of allowances began to build up, leading to a crash in the price of allowances. Many commentators blame EU economic forecasting for this problem, as the recession and consequent reduction in economic activity was not factored in to the EU-ETS control mechanism. Criticism has been forthcoming for the economic models used, and some go as far as to liken the mismanagement of EU-ETS to the ‘wine-lake and butter-mountain’ days of the 1980s, where the Common Agricultural policy was allowed to consume over 70% of the EU’s budget. Perhaps the models are too simple – James Cameron, a speaker at the APPCCG event, spoke of the ‘premium on simplicity’ that exists in creating policy. Maybe that approach has extended itself into the mathematical models used to predict the performance of EU-ETS, rendering them over-simplistic?

Personally, I see things a little differently. It’s clear that economic models are often far from perfect; however, I’m not sure that’s where the problem lies. In the implementation of policy, decision makers have to draw on the implications of many separate models; for instance, they must consider the GDP growth of EU member states, their adoption rate of new energy efficiency standards and the relative industrialisation of their economies. To my mind, the greatest source of error is in the gaps and interfaces between these economic models. Policy makers must make decisions on how to interpret the way economic predictions will interact with one another, and these interpretations are always subject to value judgements. What we need is a more joined-up approach.

Climate science has long used ‘macro-models’ to incorporate a variety of physical processes into their predictions, an approach that could be adopted by economists as well. While the first economic macro-models may not achieve even a fraction of the accuracy of climate models, that is not to say they cannot be improved through collaboration and quantitative criticism. Perhaps now is the time to make a start?

This blog is written by Neeraj Oak, Cabot Institute.



Neeraj Oak