Blog Archive - April 2008

Moulins, Calving Fronts and Greenland Outlet Glacier Acceleration

The net loss in volume and hence sea level contribution of the Greenland Ice Sheet (GIS) has doubled in recent years from 90 to 220 cubic kilometers/year has been noted recently (Rignot and Kanagaratnam, 2007). The main cause of this increase is the acceleration of several large outlet glaciers. There has also been an alarming increase in the number of photographs of meltwater draining into a moulin somewhere on the GIS, often near Swiss Camp (35 km inland from the calving front). The story goes—warmer temperatures, more surface melting, more meltwater draining through moulins to glacier base, lubricating glacier bed, reducing friction, increasing velocity, and finally raising sea level. A number of recent results suggest that we need to take another look at this story.

The Acceleration:

Jakobshavn Glacier, West Greenland, retreated 30 km from 1850-1964, followed by a stationary front for 35 years. Jakobshavn has the highest mass flux of any glacier draining an estimated 6% of the GIS. The glacier terminus region also had a consistent velocity of 19 meters/day (maximum of 26 m in glacier center), from season to season and year to year, the glacier seemed to be in balance, as I noted in a 1989paper. This is the fastest glacier in the world, no steroids needed. After 1997 it began to accelerate and thin rapidly, reaching an average velocity of 34 m/day in the terminus region. The glacier thinned at a rate of up to 15 m/year and retreated 5 km in six years. Jakobshavn has since slowed to near its pre-1997 speed, the terminus retreat is still occurring, but likewise is.

Helheim Glacier, East Greenland had a stable terminus from the 1970’s-2000. In 2001-2005 the glacier retreated 7 km and accelerated from 20 m/day to 33 m/day, while thinning up to 130 meters in the terminus region. Kangerdlugssuaq Glacier, East Greenland had a stable terminus history from 1960-2002. The glacier velocity was 13 m/day in the 1990’s. In 2004-2005 it accelerated to 36 m/day and thinned by up to 100 m in the lower reach of the glacier. Helheim and Kangerdlugssuaq combined drain 8 % of GIS. Hence, they are more than canaries in the coal mine. In 2006, the velocity of Helheim and Kangerdlugssuaq decreased to near the 2000 level, the terminus of Helheim advanced a bit (Howat et al., 2007).

The first mechanism for explaining the change in velocity is the "Zwally effect", which relies on meltwater reaching the glacier base and reducing the friction through a higher basal water pressure. A moulin is the conduit for the additional meltwater to reach the glacier base. This idea, proposed by Jay Zwally, was observed to be the cause of a brief seasonal acceleration of up to 20 % on the Jakobshavns Glacier in 1998 and 1999 at Swiss Camp (Zwally et al., 2002). The acceleration lasted two-three months and was less than 10% in 1996 and 1997 for example. They offered a conclusion that the “coupling between surface melting and ice-sheet flow provides a mechanism for rapid, large-scale, dynamic responses of ice sheets to climate warming”. The acceleration of the three glaciers had not occurred at the time of this study and they were not concluding or implying that the meltwater increase was the cause of the aforementioned acceleration. However, many others have made this assertion and are investigating (Stearns and Hamilton, 2007). Examination of recent rapid supra-glacial lake drainage documented short term velocity changes due to such events, but they had little significance to the annual flow of the large glaciers outlet glaciers (Das et.al, 2008).

The second mechanism is a "Jakobshavn effect", coined by Terry Hughes, (1986), where a force small imbalance of forces caused by some perturbation can cause a substantial non-linear response. In this case an imbalance of forces at the calving front propagates up-glacier. Thinning causes the glacier to be more buoyant, even becoming afloat at the calving front, and is responsive to tidal changes. The reduced friction due to greater buoyancy allows for an increase in velocity. This is akin to letting off the emergency brake a bit. The reduced resistive force at the calving front is then propagated up glacier via longitudinal extension in what R. Thomas calls a backforce reduction (Thomas, 2003 and 2004). For ice streaming sections of large outlet glaciers (in Antarctica as well) there is always water at the base of the glacier that helps lubricate the flow. This water is, however, generally from basal processes, not surface melting.

If the Zwally effect is the key than since meltwater is a seasonal input, velocity would have a seasonal signal. If the Jakobshavn effect is the key the velocity will propagate up-glacier, the terminus velocity will be impacted by tides, and there will be no seasonal cycle.

On Jakobshavn the acceleration began at the calving front and spread up-glacier 20 km in 1997 and up to 55 km inland by 2003 (Joughin et al., 2004). On Helheim the thinning and velocity propagated up-glacier from the calving front. Each of the glaciers fronts did respond to tidal variations indicating they had become afloat, detached from their bed (Hamilton et al, 2006). This had been the case at Jakobshavn for the last 50 years, but not for Helheim or Kangerdlussuaq. In each case the major outlet glaciers accelerated by at least 50%, much larger than the impact noted due to summer meltwater increase. On Jakobshavn the acceleration was not restricted to the summer, persisting through the winter when surface meltwater is absent.

As a result of the above Luckman et al. ( 2006) concluded:

“The most plausible sequence of events is that the thinning eventually reached a threshold, ungrounded the glacier tongues and subsequently allowed acceleration, retreat and further thinning. It is reasonable to believe that the 1998 Jakobshavn speed-up, also following a long period of stability, was triggered by the same processes of thinning but occurred earlier and after a shorter period of thinning because the tongue was already afloat.”

Examination of the acceleration of other glaciers such as the Petermann Glacier indicate a much smaller acceleration than that observed on three glaciers we have focused, and indeed it is in the summer and of a magnitude that the Zwally effect could explain (Rignot, 2005). Other large outlet glaciers such as the Rinks and Daugaard-Jensen have been stable since 1960 (Stearns et al, 2005). Many other lesser outlet glaciers have accelerated substantially.

That each of the three glaciers has a reduced velocity in 2006 and 2007 despite some exceptional melt conditions in 2007 further suggests that meltwater is not the dominant driver of the acceleration of the main outlet glaciers. Temporarily, there appears to be a force imbalance at the glacier fronts. This will reduce the annual contribution to rising sea level from glacier dynamic changes. The bad news is that the degree of acceleration that can occur via the Jakobshavn effect is greater in these cases than that from the Zwally effect. The Zwally effect is nonetheless real and also implies a direct sea level impact of greater melt.

The Jakobshavn is of particular importance as it has a bed below sea level for at least 80 km inland from the terminus. In this reach there are no significant pinning points, or abrupt changes in slope or width (Clarke and Echelmeyer, 1996) that would help stabilize the glacier during retreat. It is the only outlet glacier of GIS to lack these, and can then (via backforce reductions) tap into the heart of GIS. We know that surface melting is a slow process for raising sea level. but as Greenland’s major outlet glaciers have recently shown, rapid acceleration can quickly deliver large volume of ice to the ocean. The pace of change is not glacial.

Model-data-comparison, Lesson

In January, we presented Lesson 1 in model-data comparison: if you are comparing noisy data to a model trend, make sure you have enough data for them to show a statistically significant trend. This was in response to a graph by Roger Pielke Jr. presented in the New York Times Tierney Lab Blog that compared observations to IPCC projections over an 8-year period. We showed that this period is too short for a meaningful trend comparison.

This week, the story has taken a curious new twist. In a letter published in Nature Geoscience, Pielke presents such a comparison for a longer period, 1990-2007 (see Figure). Lesson 1 learned - 17 years is sufficient. In fact, the very first figure of last year's IPCC report presents almost the same comparison (see second Figure).

Pielke's comparison of temperature scenarios of the four IPCC reports with data

There is a crucial difference, though, and this brings us to Lesson 2. The IPCC has always published ranges of future scenarios, rather than a single one, to cover uncertainties both in future climate forcing and in climate response.

Any meaningful validation of a model with data must account for this stated uncertainty. If a theoretical model predicts that the acceleration of gravity in a given location should be 9.84 +- 0.05 m/s2, then the observed value of g = 9.81 m/s2 would support this model. However, a model predicting g = 9.84+-0.01 would be falsified by the observation. The difference is all in the stated uncertainty. A model predicting g = 9.84, without any stated uncertainty, could neither be supported nor falsified by the observation, and the comparison would not be meaningful.

Pielke compares single scenarios of IPCC, without mentioning the uncertainty range. He describes the scenarios he selected as IPCC's "best estimate for the realised emissions scenario". However, even given a particular emission scenario, IPCC has always allowed for a wide uncertainty range. Likewise for sea level (not shown here), Pielke just shows a single line for each scenario, as if there wasn't a large uncertainty in sea level projections. Over the short time scales considered, the model uncertainty is larger than the uncertainty coming from the choice of emission scenario; for sea level it completely dominates the uncertainty (see e.g. the graphs in our Science paper). A comparison just with the "best estimate" without uncertainty range is not useful for "forecast verification", the stated goal of Pielke's letter. This is Lesson 2.

In addition, it is unclear what Pielke means by "realised emissions scenario" for the first IPCC report, which included only greenhouse gases and not aerosols in the forcing. Is such a "greenhouse gas only" scenario one that has been "realised" in the real world, and thus can be compared to data? A scenario only illustrates the climatic effect of the specified forcing - this is why it is called a scenario, not a forecast. To be sure, the first IPCC report did talk about "prediction" - in many respects the first report was not nearly as sophisticated as the more recent ones, including in its terminology. But this is no excuse for Pielke, almost twenty years down the track, to talk about "forecast" and "prediction" when he is referring to scenarios. A scenario tells us something like: "emitting this much CO2 would cause that much warming by 2050″. If in the 2040s the Earth gets hit by a meteorite shower and dramatically cools, or if humanity has installed mirrors in space to prevent the warming, then the above scenario was not wrong (the calculations may have been perfectly accurate). It has merely become obsolete, and it cannot be verified or falsified by observed data, because the observed data have become dominated by other effects not included in the scenario. In the same way, a "greenhouse gas only" scenario cannot be verified by observed data, because the real climate system has evolved under both greenhouse gas and aerosol forcing.

Pielke concludes: "Once published, projections should not be forgotten but should be rigorously compared with evolving observations." We fully agree with that, and IPCC last year presented a more convincing (though not perfect) comparison than Pielke.

To sum up the three main points of this post:

1. IPCC already showed a very similar comparison as Pielke does, but including uncertainty ranges.

2. If a model-data comparison is done, it has to account for the uncertainty ranges - both in the data (that was Lesson 1 re noisy data) and in the model (that's Lesson 2).

3. One should not mix up a scenario with a forecast - I cannot easily compare a scenario for the effects of greenhouse gases alone with observed data, because I cannot easily isolate the effect of the greenhouse gases in these data, given that other forcings are also at play in the real world.

Target CO2

What is the long term sensitivity to increasing CO2? What, indeed, does long term sensitivity even mean? Jim Hansen and some colleagues (not including me) have a preprint available that claims that it is around 6ºC based on paleo-climate evidence. Since that is significantly larger than the 'standard' climate sensitivity we've often talked about, it's worth looking at in more detail.

We need to start with some definitions. Sensitivity is defined as the global mean surface temperature anomaly response to a doubling of CO2 with other boundary conditions staying the same. However, depending on what the boundary conditions include, you can get very different numbers. The standard definition (sometimes called the Charney sensitivity), assumes the land surface, ice sheets and atmospheric composition (chemistry and aerosols) stay the same. Hansen's long term sensitivity (which might be better described as the Earth System sensitivity) allows all of these to vary and feed back on the temperature response. Indeed, one can imagine a whole range of different sensitivities that could be clearly defined by successively including additional feedbacks. The reason why the Earth System sensitivity might be more appropriate is because that determines the eventual consequences of any particular CO2 stabilization scenario.

Traditionally, the decision to include or exclude a feedback from consideration has been based on the relevant timescales and complexity. The faster a feedback is, the more usual it is to include. Thus, changes in clouds (~hours) or in water vapour (~10 days) are undoubtedly fast and get included as feedbacks in all definitions of the sensitivity. But changes in vegetation (decades to centuries) or in ice sheets (decades(?) to centuries to millennia) are slower and are usually left out. But there are other fast feedbacks that don't get included in the standard definition for complexity reasons - such as the change in ozone or aerosols (dust and sulphates for instance) which are also affected by patterns of rainfall, water vapour, temperature, soli moisture, transport and clouds (etc.).

Not coincidentally, the Charney sensitivity corresponds exactly to the sensitivity one gets with a standard atmospheric GCM with a simple mixed-layer ocean, while the Earth System sensitivity would correspond to the response in a (as yet non-existent) model that included interactive components for the cryosphere, biosphere, ocean, atmospheric chemistry and aerosols. Intermediate sensitivities could however be assessed using the Earth System models that we do have.

In principal, many of these sensitivities can be deduced from paleo-climate records. What is required is a good enough estimate of the global temperature change and measures of the various forcings. However, there are a few twists in the tale. Firstly, getting 'good enough' estimates for global temperatures changes is hard - this has been done well for the last century or so, reasonably for a few centuries earlier, and potentially well enough for the really big changes associated with the glacial-interglacial cycle. While sufficient accuracy in the last few centuries is a couple of tenths of a degree, this is unobtainable for the last glacial maximum or the Pliocene (3 million years ago). However, since the signal is much larger in the earlier periods (many degrees), the signal to noise ratio is similar.

Secondly, although many forcings can be derived from paleo-records (long-lived greenhouse gases from bubbles in the ice cores most notably), many cannot. The distribution of sulphate aerosols even today is somewhat uncertain, and at the last glacial maximum, almost completely unconstrained. This is due in large part to the heterogenity of their distribution and there are similar problems for dust and vegetation. In some sense, it is the availability of suitable forcing records that suggests what kind of sensitivity one can define from the record. A more subtle point is that the 'efficacy' of different forcings might vary, especially ones that have very different regional signatures, making it more difficult to add up different terms that might be important at any one time.

Lastly, and by no means leastly, Earth System sensitivity is not stable over geologic time. How much it might vary is very difficult to tell, but for instance, it is clear that from the Pliocene to the Quaternary (the last ~2,5 million years of ice age cycles), the climate has become more sensitive to orbital forcing. It is therefore conceivable (but not proven) that any sensitivity derived from paleo-climate will not (in the end) apply to the future.

We've often gone over the Charney sensitivity constraint for the Last Glacial Maximum. There is information about the greenhouse gases (CO2, CH4 and N2O), reconstructions of the ice sheets and vegetation change, and estimates of the dust forcing. A recent estimate of the magnitude of these forcings is around 8 +/- 2 W/m2 (Schneider von Deimling et al, 2006). This implicitly includes other aerosol changes or atmospheric chemistry changes in with the sensitivity (or equivalently, assumes that their changes are negligible). So given a temperature change of about 5 to 6ºC, this gives a Charney sensitivity of around 3ºC (ranging from 1.5 to 6 if you do the uncertainty sums).

Hansen suggests that the dust changes should be considered a fast feedback as well (as could the CH4 changes?) and that certainly makes sense if vegetation changes are included on the feedback side of the equation. Since all of these LGM forcings are the same sign (i.e. they are all positive feedbacks for the long term temperature change), that implies that the Earth System sensitivity must be larger than the Charney sensitivity on these timescales (and for this current geologic period). So far so good.

Hansen's first estimate of the Earth System sensitivity is based on an assumption that GHG changes over the long term control the amount of ice. That gives a scaling of 6ºC for a doubling of CO2. This is however problematic for two reasons; first most of the power of this relationship is derived from when there were large N. American and European ice sheets. It is quite conceivable that, now that we are left with only Greenland and Antarctica, the sensitivity of the temperature to the ice sheets is less. Secondly, it subsumes the very special nature of orbital forcing - extreme regional and seasonal impacts but very little impact on the global mean radiation. Hansen's estimate assumes that an overall cooling of the same magnitude of the LGM would produce the same extent of ice sheets that was seen then. It may be the case, but it is not a priori obvious that it must be. Hansen rightly acknowledges these issues, and suggests a second constraint based on longer term changes.

Unfortunately, prior to the ice core record, our knowledge of CO2 changes is much poorer. Thus while it seems likely that CO2 decreased from the Eocene (~50 million years ago) to the Quaternary through variations related to tectonics, the exact magnitude is uncertain. For reasonable values based on the various estimates, Hansen estimates a ~10 W/m2 forcing change over the Cenozoic from this alone (including a temperature-related CH4 change). The calculation in the paper is however a little more subtle. Hansen posits that the long term trend in the deep ocean temperature in the early Cenozoic period (before there was substantial ice) was purely due to CO2 (using the Charney sensitivity). He then plays around with the value of the CO2 concentration at the initiation of the Antarctic ice sheets (around 34 million years ago) to get the best fit with the CO2 reconstructions over the whole period. What he ends up with is a critical value of ~425 ppm for initiation of glaciation. To be sure, this is fraught with uncertainties - in the temperature records, the CO2 reconstructions and the reasonable (but unproven) assumption concerning the dominance of CO2. However, bottom line is that you really don't need a big change in CO2 to end up with a big change in ice sheet extent, and that hence the Earth System sensitivity is high.

So what does this mean for the future? In the short term, not much. Even if this is all correct, these effects are for eventual changes - that might take centuries or millennia to realise. However, even with the (substantial) uncertainties in the calculations and underlying assumptions, the conclusion that the Earth System sensitivity is greater than the Charney sensitivity is probably robust. And that is a concern for any policy based on a stabilization scenario significantly above where we are now.

Blogs and peer-review

Nature Geoscience has two commentaries this month on science blogging - one from me and another from Myles Allen (see also these blog posts on the subject). My piece tries to make the point that most of what scientists know is "tacit" (i.e. not explicitly or often written down in the technical literature) and it is that knowledge that allows them to quickly distinguish (with reasonable accuracy) what new papers are worth looking at in detail and which are not. This context is what provides RC (and other science sites) with the confidence to comment both on new scientific papers and on the media coverage they receive.

Myles' piece stresses that criticism of papers in the peer-reviewed literature needs to be in the peer-reviewed literature and suggests that informal criticism (such as on a blog) might undermine that.

We actually agree that there is a real tension between a quick and dirty pointing out of obvious problems in a published paper (such as the Douglass et al paper last December) and doing the much more substantial work and extra analysis that would merit a peer-reviewed response. The approaches are not however necessarily opposed (for instance, our response to the Schwartz paper last year, which has also lead to a submitted comment). But given everyone's limited time (and the journals' limited space), there are fewer official rebuttals submitted and published than there are actual complaints. Furthermore, it is exceedingly rare to write a formal comment on an particularly exceptional paper, with the results that complaints are more common in the peer reviewed literature than applause. In fact, there is much to applaud in modern science, and we like to think that RC plays a positive role in highlighting some of the more important and exciting results that appear.

Myles' piece, while ending up on a worthwhile point of discussion, illustrates it (in my opinion) with a rather misplaced example that involves RC - a post and follow-up on the Stainforth et al (2005) paper and the media coverage it got. The original post dealt in part with how the new climateprediction.net model runs affected our existing expectation for what climate sensitivity is and whether they justified a revision of any projections into the future. The second post came in the aftermath of a rather poor piece of journalism on BBC Radio 4 that implied (completely unjustifiably) that the CPDN team were deliberately misleading the public about the importance of their work.

We discussed then (as we have in many other cases) whether some of the responsibility for overheated or inaccurate press actually belongs to the press release itself and whether we (as a community) could do better at providing more context in such cases. The reason why this isn't really germane to Myles' point is that we didn't criticise the paper itself at all. We thought then (and think now) that the CPDN effort is extremely worthwhile and that lessons from it will be informing model simulations some time into the future. Our criticisms (such as they were) were mainly associated instead with the perception of the paper in parts of the media and wider community - something that is not at all appropriate for a peer-reviewed comment.

This isn't the place to rehash the climate sensitivity issue (I promise a new post on that shortly), so that will be deemed off-topic. However, we'd be very interested in any comments on the fundamental issue raised - how do (or should) science blogs and traditional peer-review intersect and whether Myles' perception that they are in conflict is widely shared.