Chapter

The Large Scale Dynamics Are Reasonably Well Understood, Un-Certainty Lies in the Parametrization of Small-Scale Processes

Andrew J. Weaver is professor and Canada Research Chair in the School of Earth and Ocean Sciences at the University of Victoria/Canada. His main research interests are climate dynamics, the role of the oceans in the climate system, and Earth system modeling. He has served as a lead author for the 2nd, 3rd, 4th and 5th IPCC Working Group I reports. He has published numerous articles and several books on climate change. Andrew Weaver is a Guggenheim Fellow, a Fellow of the Canadian Meteorological and Oceanographic Society, a Fellow of the American Meteorological Society and a Fellow of the Royal Society of Canada.

Prof. Weaver, it is widely known that ocean currents have an effect on the climate. The Gulf Stream, for example, influences the climate in western Europe. What about the opposite effect? How do changes in the climate influence the oceans?

In fact, the ocean and the atmosphere are really a coupled system. It is not so much one influencing the other, as them influencing each other simultaneously. On the timescale of days to weeks, the atmosphere interacts mainly with the ocean's sea surface temperatures. But on longer timescales, you begin to involve the upper layer of the ocean; on very long timescales you also affect the deeper

Andrew J. Weaver is professor and Canada Research Chair in the School of Earth and Ocean Sciences at the University of Victoria/Canada. His main research interests are climate dynamics, the role of the oceans in the climate system, and Earth system modeling. He has served as a lead author for the 2nd, 3rd, 4th and 5th IPCC Working Group I reports. He has published numerous articles and several books on climate change. Andrew Weaver is a Guggenheim Fellow, a Fellow of the Canadian Meteorological and Oceanographic Society, a Fellow of the American Meteorological Society and a Fellow of the Royal Society of Canada.

ocean. You probably know about El Niño—this is a fully coupled phenomenon that involves an interaction between the upper ocean and the atmosphere. On the decadal to century timescales the thermohaline circulation (or overturning circulation) interacts with the atmosphere. But it is hard to actually say one would be driving the other as they are fully coupled.

The oceans form a very complex system, with their temperature distribution and all their currents and circulations etc. How stable is the current state of this system? Are there instabilities, where small changes in the boundary conditions can lead to dramatic change in the system?

Over the last decades, the community has come to understand quite well the actual mechanism behind how a large scale instability of the sinking in the North Atlantic can occur—not the precise amount of fresh water, required to cause it, but the mechanism itself. The mechanism is the following: You have surface water that is relatively salty in the North Atlantic. But should that surface water become very fresh then it can become lighter than the deeper water and therefore it can stop sinking. And that has a feedback on the transport of heat from lower to higher latitudes. So this instability is reasonably well understood.

The question is: To what extent do we understand the thresholds beyond which this effect is triggered? And of course there is a lot of variation between model projections, with different parametrizations of small scale mixing processes showing different results. But the main statement is that the large models as a collective really do not show this as something that is in the cards over the next century or two. This is simply because the perturbation on the fresh water is too small.

But as I said, there is much evidence that this has happened many times in the paleo-record over the last glacial cycle. The difference between then and now is that there were vast quantities of ice on land which provided a much greater potential for fresh water sources than we have today. We also know that the threshold depends on the mean state itself. Colder climates are inherently more unstable than warmer climates.

Is there any characteristic phenomenon you would expect to happen with the ocean currents as the global temperatures raise?

Historically, when we look at the paleoclimate record, there is an awful lot of evidence that the sinking or the thermohaline circulation in the North Atlantic has been variable. And this has led to somewhat dramatic shifts and reorganizations of ocean circulations in the Atlantic. What is thought is that much of this was driven by changes in the amount of fresh water entering the North Atlantic from the melting of various ice sheets.

There is clear evidence from climate models that one can expect an increase of precipitation at higher latitudes as we warm. At the same time there is also evidence in the observational record that this is occurring. Many have asked how this may affect the overturning. It turns out that as we have learned more and more about the thermohaline circulation we have become confident in the finding that you would expect a slight weakening of this so-called overturning and a weakening of the transport of heat from lower to higher latitudes to take place over the next century. But we do not expect a rapid transition or shutdown of the overturning that has been seen in the paleo-record as a consequence of warming. The reason why: there is not enough fresh water from the atmosphere, so it would require a rather dramatic melting from the Greenland ice sheet.

Can you give an indication over which timescales ocean currents typically change?

There are several key timescales in the ocean. There are very long timescales, associated with slow diffusive properties. There are centennial timescales that are associated with the overturning circulation. There are decadal timescales associated with gyre circulation. And there are very short timescales associated with convection. If you are asking how long it took abrupt changes in the past to occur: There is a lot of evidence that the transitioning from on to off states or on to weaker states of the north Atlantic overturning took place over a few decades— quite rapid. In the paleorecord, this was probably very common.

You already mentioned climate models. You have developed and used numerical models to investigate the oceans for quite some time. What is the state of the art with these models? What can you model well and what is still uncertain or unknown?

I think the situation is very similar to atmospheric models. The large-scale dynamics are reasonably well understood. These involve the equations governing the motion of the fluid itself—the Navier-Stokes equations. The processes involving the large-scale transport of heat and tracers (so-called conservation equations) are also well known.

Where we have uncertainty is in the parametrization of the small-scale processes that are potentially playing an important role in the kind of instability thresholds we just talked about. For example: convective processes happen at very small scales in the real ocean. In global ocean models, we have to somehow represent this on the scale of grid-cells that are approaching 50 x 50 km. In an ideal world we would be able to integrate high resolution, fully three-dimensional, non hydrostatic models, for many centuries. This would allow us to capture these convective processes. But we are not there yet. I am also thinking of parametri-zations of other mixing processes: there are internal waves that propagate and break and when they break they actually mix water properties. These are again very small-scale processes. We have internal waves in our models, but their properties are much more sluggish. These are really the biggest uncertainties in the ocean models: the very small scale mixing processes that are parametrized.

But one remark: While you can have great uncertainty in a particular small-scale process it doesn't mean that reducing this uncertainty by improving our understanding of its physics will lead to different large-scale results. That is, the existing parametrizations may be doing a very fine job. As our knowledge increases we may find that improvements to the parametrizations have little effect. This is something that is not always clear to everyone.

Was this article helpful?

0 0

Post a comment