SCIENCE AT THE SHINE DOME canberra 7 - 9 may 2008

Symposium: Dangerous Climate Change: Is it inevitable?

Friday, 9 May 2008

Dr John Church
CSIRO Marine and Atmospheric Research

John Church is an oceanographer specialising in the role of the ocean in climate, particularly anthropogenic climate change. He was co-convening lead author for the chapter Changes in sea level in the Intergovernmental Panel on Climate Change Third Assessment Report and co-chaired the World Climate Research Programme workshop, Understanding sea-level rise and variability. John presented the 2006 Roger Revelle Lecture at the Intergovernmental Oceanographic Commission (IOC). He was awarded the 2006 Roger Revelle Medal by the IOC, a CSIRO Medal for Research Achievement in 2006, the 2007 Eureka Prize for Scientific Research, and presented the 2008 University of New South Wales AMOS RH Clarke Lecture. He was co-chair of the international Scientific Steering Group for the World Ocean Circulation Experiment from 1994 to 1998, and chair of the Joint Scientific Committee of the World Climate Research Programme from 2006 to 2008.

Global sea level: How well can we predict its future?

(Click on image for a larger version)

Thanks very much, Amanda, and thank you to the convenors for inviting me to give this talk. My apologies for being late. Thank you to those who actually made it possible for me to get here in time at least to give a late talk.

The title that was given to me was 'Global sea level: How well can we predict its future?' so in the next 20 minutes or so I am going to try to bring you up to date on what our current understanding is.

(Click on image for a larger version)

If you look back through time, using the palaeo evidence, then you see that about 120,000 years ago, at the onset of the last ice age, sea level was roughly 4 to 6 metres greater than it is in today's time, at temperatures that we would expect late this century.

Very recent evidence is suggesting that at that time rates of sea-level rise were over a metre per year, 1½ or perhaps even 2 metres per year in the extreme case, at temperatures and sea levels similar to today's, or that we at least expect this century.

Then, through the last ice age, sea level fell by about 120 metres to its minimum about 20,000 years ago, then rose rapidly from 20,000 years ago until about 7000 years ago, at peak rates approaching 4 metres per century.

Then, over the last 6000 years, sea level has been much more steady. Indeed, over the last couple of thousand years, sea level has varied little from about 2000 years ago until the start of the 18th century.

(Click on image for a larger version)

One example of that is a nice piece of work that our President here, Kurt Lambeck, has done, looking at sea levels around Italy and using ancient Roman fish tanks and looking at the levels implied by those. The indications are that there has been no net change in sea level from about the time of these fish tanks, roughly 2000 years ago, until about 1800.

(Click on image for a larger version)

At the time of the last IPCC [Intergovernmental Panel on Climate Change] report, we didn't have a time series of global average sea level. By combining satellite altimetry, which gives us near-global coverage, with the historical sea-level measurements around the world, we have constructed an estimate of global average sea levels (shown here by the red curve) from 1870 up to, essentially, the present.

We see that during that period sea level has risen. The rate of rise averaged over that period is about 1.4 millimetres per year. It is somewhat faster during the 20th century, at 1.7 millimetres per year, so there has been an acceleration both during the period from 1870 and also during the 20th century.

Shown in the little box on the top right, in green, is the satellite altimeter data, which is the only truly or near global coverage of sea level we have. During that period, both the in situ data and the satellite data show that sea level has been rising at over 3 millimetres per year.

(Click on image for a larger version)

So what contributes to this rise?

Firstly, thermal expansion is perhaps the best understood of those terms. We have recently re-estimated the amount of ocean heat content change, which is essentially linearly proportional to thermal expansion – this is for the upper 700 metres of the ocean. The previous estimates are shown here in red and blue, and the climate models were unable to simulate the decadal variability, particularly the large rise in ocean heat content in the 1970s.

Our revised estimate is shown in black. The grey shading around that is our estimate of the error bars. As you go back in time those error bars grow substantially, but since about 1970 they are much smaller.

There is also decadal variability in our estimate, and the trend from the 1970s up to the present is roughly 50 per cent larger than the previous estimates for both ocean heat content change and thermal expansion.

(Click on image for a larger version)

One of the causes of the variability in the ocean heat content and steric sea level is violent volcanic eruptions, such as shown in this picture from Mt Pinatubo in 1991. I noticed the very recent eruption in South America, and it will be very interesting to see the ocean's response to that over the coming years.

These have to be major, violent volcanic eruptions that get aerosols into the stratosphere, and they cause a cooling of the climate system and of the ocean.

(Click on image for a larger version)

We can now compare the models from the most recent WCRP [World Climate Research Programme] model intercomparison project, used in the IPCC report – they are shown in the coloured lines here. These are only the models that include all the natural forcings, including the volcanic forcing, compared with our estimate of sea level and heat content, shown in black.

There is good agreement, particularly in the variability. There is now significant correlation between the observed and the modelled variability, and significant agreement in the overall trend. On average, the models tend to underestimate the trend in ocean heat content change since the 1960s, but there are a number of different potential reasons for that, including differences in forcings in the models.

(Click on image for a larger version)

The second major contribution to sea-level rise is melting of non-polar glaciers. These are glaciers in places like Switzerland, Patagonia, Alaska et cetera. The data show that in the 1960s glaciers were melting, contributing to sea-level rise at about 0.4 millimetres per year, and this has increased in the latter half of the 20th century and particularly during the 1990s to rates of about 1 millimetre per year. Recent estimates are even larger than demonstrated in this diagram.

So again we now have a time series of changes in contributions from glaciers which we did not have before.

(Click on image for a larger version)

The two ice sheets are the two potentially largest contributions to sea level in the longer term. The Greenland ice sheet has been subject to increasing surface melting over the last several decades, in fact, and we now have a whole series of estimates of changes in the mass of the Greenland ice sheet from different techniques, as shown in the right-hand panels: from satellite altimetry, from laser-borne aircraft altimetry, from synthetic aperture radar and also from gravity measurements.

The estimate is that the contribution to sea level was about 0.2 millimetres per year during the 1990s. This is thought to be a recent increase in the contribution from the Greenland ice sheet to sea level.

(Click on image for a larger version)

The content of the Antarctic ice sheet in total approaches 60 metres of equivalent sea level, and is potentially the largest contribution. The west Antarctic ice sheet gives reason for particular concern, because it is grounded below sea level.

Estimates of the mass balance of the Antarctic ice sheet are uncertain. Again there are satellite altimetry measurements and, particularly in recent years, gravity measurements. These indicate probably a loss of mass, particularly from the west Antarctic ice sheet, with a small gain of mass from increased precipitation in the east Antarctic ice sheet, and a net mass loss from the total Antarctic equivalent to about 0.2 millimetres per year, but with significant uncertainty on this number.

(Click on image for a larger version)

We have combined those estimates in a recent paper that is currently in press. Shown in the top panel are the various contributions – the glacier contribution, shown here by the dark blue; an estimate for the ice sheets based on the most recent IPCC numbers, then grading down to a smaller contribution or zero contribution from Greenland in the 1960s but a steady contribution from the Antarctic ice sheet; our estimates of ocean thermal expansion, shown here in red; and estimates of terrestrial water storage. (This is water stored in aquifers and snow and soil moisture.)

When you add all those together, you get the blue curve shown in the bottom panel, compared with our estimates of sea-level rise re-computed by using a longer satellite altimeter record than was used in the IPCC reconstruction and a larger number of in situ measurements. That is shown in black. And the yellow dashed curve is another estimate of global average sea level by a completely different technique, which agrees very well with our own estimate. And the satellite altimeter record over the last few years is shown by the red dashed line.

So the trend for the first time between the sum of components is approximately equal to the observed rate of rise, but there is significant decadal variability which we are still unable to understand.

(Click on image for a larger version)

So what do the projections for the future say? I have combined here the projections from the Third Assessment Report of 2001 – shown by the curves and the shading, starting from 1990 and going to 2100 – and the projections from the more recent IPCC report, shown by the two bars. The mauve bar is from the model projections, and that agrees reasonably well with the light shading, which is essentially the equivalent from the IPCC Third Assessment Report, with the exception that the lower bound is slightly higher.

So that is for roughly a 20 centimetre to 60 centimetre rise over this period.

The Third Assessment Report and also the Fourth Assessment Report made an extra allowance for uncertainties associated with ice sheets, and that is shown by the top solid line for the Third Assessment Report and the red shading for the Fourth Assessment Report. Given the different time-scales, again there is good agreement between these. However, the Fourth Assessment Report also made the comment that, given the recent and rapid changes in ice sheets, 'Larger values cannot be excluded.'

(Click on image for a larger version)

So how well are these projections holding up to the test of time?

(Click on image for a larger version)

Shown here is a comparison of the Third Assessment Report projections. The upper black dashed line and the lower dashed black line are the two limits on the IPCC projections. That upper limit is a line that leads to an 88 centimetre rise by 2100. Shown in red are our in situ sea-level reconstructions available when this report was done; and, in blue, the satellite altimeter observations. Both of those lines are tracking near the upper limit of the IPCC projections. That is the line that leads towards an 88 centimetre rise by 2100.

That is not to say that they will continue to track that line, but that is where they are at the moment.

(Click on image for a larger version)

Our uncertainty of how to represent ice-sheet models, particularly the rapid response that is being seen in both Greenland and Antarctica to changes, has caused some people to question whether the IPCC sea-level estimates are underestimates, and has caused people to pursue alternative approaches. One of these is the Stefan Rahmstorf model, which is a simple statistical relationship, relating our sea-level records to observed surface temperature and using that to project into the future. That leads to projections of roughly 70 centimetres to almost 140 centimetres by 2100, compared with the upper range of the IPCC projections of closer to 80 centimetres – so, somewhat larger than the IPCC projections.

(Click on image for a larger version)

What about the longer term? Sea-level rise will continue long after greenhouse gas concentrations are stabilised in the atmosphere, firstly from ocean thermal expansion, shown here as a series of model experiments where greenhouse gases are stabilised in the top panel at various levels, and in the bottom panel is the thermal expansion of the ocean response to those. You hardly see a change in slope at the time of stabilisation of the greenhouse gases. So sea-level rise from thermal expansion continues for centuries, and stabilisation of emissions means even larger rates of change.

(Click on image for a larger version)

There is significant concern about the ice sheets. First I will talk about the Greenland ice sheet.

The palaeo evidence, as I said previously, is that at the last interglacial sea level was perhaps 4–6 metres higher than present-day, with a significant component of that coming from the Greenland ice sheet. Recent palaeo evidence is suggesting sea-level rise rates of a metre per century or larger at that time, at sea levels close to present-day sea levels.

Surface melting has been observed to increase over recent decades. Projections are that for a sustained warming of over 4.5° over Greenland, with an uncertainty of about 1°C –that is equivalent to global average surface warmings of about 3° – then melting exceeds precipitation, leading to an ongoing melting of the ice sheet, and if those temperatures are maintained, essentially an elimination of the Greenland ice sheet.

From surface melting alone, that takes millennia, but it does contribute of the order of 7 metres to sea level.

(Click on images for larger versions)

However, there is also concern about the melt water which has been observed to be flowing off the Greenland ice sheet and then flowing down through cracks and crevasses in the ice sheet, as demonstrated schematically in the top panel and the picture on the bottom panel, to the base of the ice sheet. This has actually been observed now with water flowing roughly a kilometre through the ice sheet down to its base, rapid response of the ice sheet to this water, and the potential for – and, in fact, observations of – a faster flow of the ice sheet in response to this bottom lubrication.

However, the observations most recently reported are that this increases the rate of flow of the ice sheets by perhaps 50 up to 100 per cent for the ice sheets. This is not thought to be the contribution or the main factor leading to rapid responses of the outlet glaciers. This only results in a much smaller change in flow of the outlet glaciers.

(Click on image for a larger version)

However, there are other processes which are important in both the Greenland ice sheet and the west Antarctic ice sheet, associated firstly with the loss of ice shelves. As shown in the middle panel here, if you lose the ice shelves you lose some of the buttressing of the ice sheets and a more rapid flow of the glaciers occurs. This rapid flow in the outlet region is rapidly propagated upstream. There are recent models and observations of this process starting to be developed, but it is not yet incorporated in the global models.

Secondly, as shown in the bottom panel, there is a penetration of water underneath the west Antarctic ice sheet and some of the outlet glaciers in Greenland, potentially leading to a more rapid movement, again, of the ice sheets.

(Click on image for a larger version)

In the interests of time I will not go into detail about these slides.

(Click on image for a larger version)

Sea level will be felt not only through changes in the mean level, but also through changes in the frequency and intensity of extreme events. Shown here is an example: we have taken the Fremantle record, which is one of Australia's longest sea-level records, divided it into pre-1950 data and post-1950 data, and asked: what is the frequency of return of an extreme event of a given magnitude?

What we see is that, in the case illustrated here, a 1.6 metre sea-level event occurred once every five years in the pre-1950 data and about once every two years in the post-1950 data. So we have already seen, this century, a more than doubling in the frequency of extreme events of a given magnitude. And the Sydney record shows almost the same result.

(Click on image for a larger version)

You can take that and look at gauges around Australia to ask: what will that mean for, say, a 0.5 metre rise in sea level? (As we have seen from the projections, 0.5 metres is by no means an extreme sea-level rise by 2100.) What that implies is that the present 1 in 100 year event could be happening several times a year by 2100.

(Click on image for a larger version)

So, is it dangerous? We need to think about what are the key impacts in the coastal zone, from both sea-level rise and variability – it is the interconnection of these two phenomena that is critically important.

It causes inundation, whether it is in Pacific Island nations or in developed Western nations, as in New Orleans, shown here on the left.

(Click on image for a larger version)

It causes storm surges and waves, with their impact.

(Click on image for a larger version)

It causes coastal erosion.

(Click on image for a larger version)

It has impacts on emergency and escape routes. We need to consider questions of environmental refugees, and it is not 'if' but 'when and where and how will we respond?'

(Click on images for larger versions)

And we need to consider what are the associated issues of food, water and health security, et cetera.

(Click on image for a larger version)

To look at one of the areas of major impact, the Bay of Bengal: there have been 23 surge events since 1737, with over 10,000 people killed. The most recent is in 1991, with 140,000 people killed and 10 million people made homeless.

(Click on image for a larger version)

Also, Europe has been subjected to these effects. I have just returned from talking with people advising the Netherlands government on building their dykes higher.

(Click on image for a larger version)

In summary, the science shows that ongoing sea-level rise is inevitable. It is an issue that is affecting us now, it will affect us during the 21st century, and it will affect us in the longer term. There are some remaining gaps in our ability to predict how much the sea-level rise will be and, particularly, how rapidly it will rise, and also changes in extreme events.

Is it dangerous? Well, the question is at what level and for whom. This is a judgment which depends on far more than the science, but it would seem to me that some of the numbers speak for themselves.

We are going to need to adapt. The impacts will be felt through extreme events which will become more frequent and more severe, even without changes in the intensity or the frequency of meteorological events. The least developed countries and the poor are most at risk. This requires local and regional planning, and planning to avoid the impacts of severe events. And we need to narrow uncertainties.

We will need to mitigate if we are to avoid the most extreme scenarios. Without significant, urgent and sustained action we could pass a threshold during the 21st century that would commit the world to metres of sea-level rise over the subsequent centuries.

Environmental refugees are an issue. We already have them; we will have more in the future.

It is essential and, I believe, urgent that science, government, business and community partnerships are strengthened. We have seen two very recent examples where this partnership has failed to deliver what it could have delivered. We knew the New Orleans levees were inadequate for storm surge, but no action was taken. We knew that surge was approaching – Katrina was well predicted – yet people died in their hospital beds. There was a six-day forecast of the most recent Myanmar disaster, which was quite accurate. Yet we did not respond adequately.

(Click on image for a larger version)

For those interested in more information on the scientific issues, there is a report on the WCRP web site and we are in the process of completing a book.

Thank you.

Discussion

Question: This is just a question on the rate. You said that in the last deglaciation, 6000–10,000 years ago, the rate was 3–4 metres per century. What are the forcings now as compared with that period, do we know? I want to get some understanding of the implications of that period for the next hundred years.

John Church: Those rates of up to 4 metres per century were the maximum rates from roughly 20,000 years ago to 7000 years ago, over that period, with average rates somewhat slower than that. At that time there were much larger ice sheets, particularly in the Northern Hemisphere, which could contribute rapidly to sea-level rise. Now the major player is the Greenland ice sheet, which we should think of as a relic from the last ice age, given its latitude. It only contains about 7 metres of sea-level on land, as compared with the 120 metres or so that we have lost since the Last Glacial Maximum. So that would indicate to me, at least – and I am not a glaciologist – a lower rate of contribution.

But there is recent palaeo evidence indicating that when sea level was at values close to today's – and above today's, in fact – rates of sea-level rise were a metre per century or perhaps higher.

Question: You are probably aware that James Hansen has done what may be a rather simple calculation. He said that if the recent rate of sea-level rise (or maybe it is recent rate of ice loss from Greenland, I forget which) has approximately doubled in a decade, and if that represented a non-linear phenomenon kicking in and if that continued to double every 10 years, you could have 5 metres or so of sea-level rise by the end of this century. That may be conjectural and rather uncertain, but I think the question that we scientists need to address is: can that be ruled out with confidence?

John Church: Again, I am not a glaciologist. My glaciological colleagues, I think, would say that that is rather an extreme number. We actually don't have good estimates prior to the 1990s of how much the ice sheets contributed, so the doubling is rather arbitrary. The time-scale of doubling and assumption of continued doubling are also rather arbitrary.

I personally think those numbers are rather too high, and I haven't heard any of my glaciological colleagues, really, talking about numbers that large. But we are talking about things we don't understand. James Hansen makes the valid point that ice sheet decay is a wet process, where water can rapidly lubricate movement of the ice sheets. So I think we have to be very cautious. We have to continue to monitor sea levels to try and understand better the contributions, and I think the ice sheets in particular are an absolutely urgent and critical problem for us to address.

Question: When will Bangladesh cease to exist, and where will its 90 million people go?

John Church: I can't answer either of those questions. One of the things I missed out in my talk is that we need to think more than just about global average sea-level change; we need to think about the local responses. There are regional distributions in sea-level rise, there are regional distributions in land motion. And so in places like Bangladesh, not only are they low but also those sediments are compacting. The rate of relative sea-level rise is larger in most of these deltaic regions around the world than the global average.

I believe environmental refugees will be a major issue for our children to deal with.

Question: I was just wondering: if we decided we didn't like the idea that the Arctic ice was to disappear completely in summer times, and if we put in place – this is now the hypothetical part of my question – a regime that restored the Arctic ice to the extent that we might have expected 30 years ago, what would be the prognosis for sea-level rise in terms of what you have been talking about? You said sea-level rise was inevitable and you were talking about thermal inertia. But a number of these things are kick-on effects assuming that we don't make major changes in policy.

John Church: I will make a couple of comments. I think we do need to make significant changes in policy, both on the mitigation side and on the adaptation side. We can significantly lessen the impacts through both of those actions.

With respect to your specific question about maintaining the sea ice over the Arctic Ocean, I am not quite sure how you are suggesting we do that. The surface melting of Greenland is not primarily related to the Arctic cover, I believe, so I am not sure that that will be a major contributor.

Question (cont.): To get the Arctic ice back you would probably have to have about a third of a degree reduction in temperature compared with where we are now. If you had a third of a degree reduction in global average temperature across the world, how would that affect your prognosis in relation to the other things you were talking about?

John Church: Certainly if you had a third of a degree reduction in global average temperature you would lessen all of these issues – the ice-sheet issues, the ocean thermal expansion and also the glacier contributions. I am afraid I don't see that happening in the near term.

Question (cont.): That wasn't my question!

Amanda Lynch: Thanks very much, John. That was great.

I would like to thank all of the speakers we had in this session. I think they gave very interesting talks, if depressing for some of us, reminding us of these issues and their immediacy.