Birthing a litter of drumlins Quite appropriately, Glacial Till won the new the latest edition of “Where on (Google) Earth?”, hosted here yesterday. The location I picked is the subject of a new paper by Mark Johnson and colleagues appears in the current issue of Geology (October 2010). It shows a place in Iceland where a piedmont-style outlet glacier called Múlajökull is pooching out to the southeast from the Hofsjökull ice cap. Here’s a more zoomed-out view of the glacier’s terminus:


Here, I’ve jacked the contrast up a bit, so you can see what’s so cool about this location — note the radial array of elliptical meltwater lakes…


The other outlet glacier, seen just to the west, is Nauthagajökull. With this context established, we can take a look at Figure 1 from the Johnson, et al. (2010) paper:


The red ellipses are between the lakes I pointed out earlier. They are drumlins, elliptical hills of glacial till. Drumlins are examples of the sub-set of glacial geomorphology which includes features made by deposition of glacial sediment (till). They are taller at the upstream end, and taper out downstream, a shape something like an “upside-down spoon.” Long-term readers will recall the time that I shared the experience of visiting some drumlins in New York, where I learned that “spoon” analogy from Paul Tomascak.

There are a lot of drumlins left over from the Pleistocene glaciation, but we don’t totally understand how they form. That’s what’s so exciting about the recession of Múlajökull: it’s exposing the world’s only known active drumlin field for geologic scrutiny. Johnson, et al., have documented 50 separate drumlins emerging from beneath the ice. Their field works has yielded some new observations that may shed light on how these distinctive landforms develop.

First off, they note that Múlajökull is a “surge-type” outlet glacier, which means that it pulses forward rapidly (4 times in the past 60 years), which isn’t the case for other glaciers, like neighbor Nauthagajökull. See the comparison in Figure 1d — where Nauthagajökull is relatively smoothly retreating, but Múlajökull has fits and starts. This may be important: Nauthagajökull hasn’t produced any drumlins.

Second, they documented various aspects of the drumlins at Múlajökull. They have an aerial aspect ratio of about 3.0, which is similar to what we see in the drumlin zones of New York and other Pleistocene drumlin fields. So that makes uniformitarians happy — maybe the dynamics of Múlajökull are analogous to the Laurentide ice sheet! Another, more detailed study, was made of the internal structure and stratigraphy of the drumlins, as exposed in channels carved into the drumlin laterally by flowing meltwater. The guts of the drumlin show multiple till units, the most recent of which truncates the ones below it in a subtle but discernible angular unconformity.The uppermost till can be traced to the end-moraine produced by the most recent (1992) surge of the glacier, but not beyond it.

They also note the presence of orange-colored water-escape structures, cutting across the till units and filled with fine sediment, and a pebble fabric which is parallel to the drumlin’s long axis (and ice-flow direction).

A final class of data is gained by taking a look at what the glacier’s snout looked like before it revealed its internal drumlins. Here’s Figure 5 from the new paper, which overlays the traced drumlin boundaries from Figure 1 on an air photo from 1995, a time after the glacier surged forward in 1992, but before the most recent recession of the terminus that revealed the drumlins:


The authors note that the crevasse pattern on the 1995 glacier is clearly related to the location of the drumlins that have recently emerged. A V-shaped pattern of crevasses may be seen immediately upstream from many of the drumlins’ positions.

After the 1992 surge, the glacial ice at the terminus of Múlajökull has been essentially stagnant: there are no recessional moraines between the 1992 surge end-moraine and the current ice front. Without moving ice, the authors find it difficult to imagine how drumlins could be formed. They infer that the drumlins formed during the surging stage of the glacier’s movement. The erosional basal contact of the upper till unit seen inside the drumlins suggests that erosion (as well as deposition) is an important part of the processes which form drumlins. Stress differences under and between crevasses cause slight differences in the rates of erosion vs. deposition the glacier bed. More till builds up beneath crevasses, less till accumulates between them. Time goes by, the glacier surges, and a big batch of new till gets added to the top of the drumlins. Amplifying feedback enlarges the drumlins with each successive surge, mainly on the upstream end and the sides of the drumlin. The authors interpret the drumlin’s internal stratigraphy of multiple till units as the record of multiple surges.

The authors of the new paper conclude by examining the two principal models for drumlin formation: a subglacial bed-deformation model from Boulton (1987), and a meltwater model proposed by Shaw (2002). They point out the truncated stratigraphy they observed inside the Múlajökull drumlins as evidence for the Boulton model, and a lack of sufficient meltwater to support the Shaw hypothesis.

Right now, Múlajökull is our only functional modern analogue for drumlin formation in the Pleistocene, but others may soon emerge. The authors also predict that as glacial recession continues to play out all over the world, we may someday observe other active drumlin fields, and gain further insights into what’s happening beneath continental glaciers.


Boulton, G.S. (1987). A theory of drumlin formation by subglacial sediment deformation, in Menzies, J., and Rose, J., eds. Drumlin symposium: Rotterdam, Balkema, p. 25-80.

Johnson, M., Schomacker, A., Benediktsson, I., Geiger, A., Ferguson, A., & Ingolfsson, O. (2010). Active drumlin field revealed at the margin of Mulajokull, Iceland: A surge-type glacier Geology, 38 (10), 943-946 DOI: 10.1130/G31371.1

Shaw, J. (2002). The meltwater hypothesis for subglacial bedforms. Quaternatary Interational, v. 90, p. 5-22. DOI: 10.1016/S1040-6182(01)00089-1.


The Purgatory Conglomerate

After my thesis defense at the University of Maryland, my mentor and friend E-an Zen asked me if I had ever heard of the Purgatory Conglomerate. I had not. Over the years, E-an has been a great source of new ideas and information to me, and so when he raises a notion, I pay attention.

In my thesis, I had done some strain analysis on volcanic clasts in a meta-ignimbrite that had developed foliation and lineation in Mesozoic shear zone in California’s high Sierra, and that reminded E-an of a rock he had once seen which was screaming for similar treatment: the Purgatory Conglomerate.

On my summer travels this year, I finally had the opportunity to swing through Newport, Rhode Island, and check it out in person. To me as a structural geologist and Zen devotee, this was like nirvana. Check it out:



I was very excited to be there. Here’s me enthusiastically embracing a watermelon-sized clast:


Looking along the trend of the stretching lineation (which is pretty much non-plunging):


Most of the clasts are quartzites of various flavors… Depositionally, it’s a relatively mature conglomerate.


Here’s looking “down the barrel” of the stretched clasts in a big boulder sitting atop the outcrop:


Here’s a really big clast, with local beer cans for scale (not mine, I swear):

Here’s a really long clast:
Recall that my Swiss Army knife is 11 cm long, but even without the specific unit, you can see that this clast has an axial ratio (on the plane of the outcrop) of roughly 7:1.

Here’s another long one with an axial ratio of 7:1, with a bonus feature. It displays internal bedding (of the sandstone it was originally derived from):


This is totally awesome. These cobbles, boulders, and pebbles have flowed into elongated shapes! We can use the geometric term “prolate” to describe their cigar-like or hot-dog-like forms.

It’s not all conglomerate there. There are some meta-sandstone bodies too. Here’s an example of sandstone meeting conglomerate:

Even cooler: here’s bedding: lenses of conglomerate within sandstone:

Annotated copy of that same photo:purgcong_37

(I once showed you something similar from the Sierra Crest Shear Zone: check photo C of this archived post.)

So how did the Purgatory Conglomerate get so distinctively deformed? Close examination of the rock suggests the main mechanism was pressure solution:

purgcong_10In the photo above, look below the Swiss Army knife for a triangular clast, and trace out its boundaries. You will see that it impinges on the hot-dog-shaped clasts immediately next to it. This triangular grain is encroaching on its neighbors’ territory! Now, one way to interpret this is that the original clasts had shapes which, jigsaw-puzzle-like, were perfectly formed to accommodate their neighbors’ shapes. But that seems rather unlikely, especially when you consider the ten gazillion clasts in this outcrop, all perfectly locked together.

Instead, the idea is that high pressure points (the edge of one round cobble touching an adjacent round cobble, for instance) are sites where certain minerals will go into solution. Quartz is both a common mineral and a mineral which will dissolve under high pressure and re-precipitate under lower pressure. Calcite pulls the same trick — that’s where stylolites come from. [Many nice examples of stylolites and other pressure-solution features here.]

Here’s another nice example showing how the individual clasts lock together with one another, suggesting part of their outer edge has dissolved away:purgcong_13

Here, too. See if you can pick out a few examples of where one clasts impinges on its neighbor. A refresher course may be found here.purgcong_15

Time for a different perspective. Unlike most of the previous pictures, this one is taken looking along the long axis of the clasts.


Zooming in for a closer look at that same photograph, the yellow areas highlight areas where one grain impinges on a neighboring grain:purgcong_zoom

If there are particularly large clasts, they may shelter smaller neighbors in their “pressure shadow,” immediately adjacent to them. Think of a building collapsing during an earthquake, with a strong central pillar. If you stand next to that pillar, you’re less likely to have the ceiling collapse on your head, since the pillar is protecting you. With that in mind, examine this part of the outcrop:

Since the long axes of the clasts runs left-to-right, that suggests that they were squeezed top-to-bottom. Therefore, the area immediately to the left of the giant clast would be “protected” from the highest pressures by the bulk of its large neighbor. If we zoom in there…

purgcong_28…we see clasts whose long axes are not aligned with the rest of the outcrop. They are pointed in several other directions, and/or are not as prolate as the rest of the clast population.

Here’s an annotation of the zoomed-in view, showing the orientation of the long axis of the grains in this outcrop plane:purgcong_28

The implication is that these “protected grains” were less subject to pressure solution than the grains which weren’t lucky enough to have a giant neighbor immediately “next door” (along strike).

In addition, it seems that the strain (deformation/stretching) of the clasts was more severe in some locations, and less severe in other locations. Here, my hands bracket a zone of less deformed (more spheroidal, less prolate) clasts within the overall outcrop of strongly deformed clasts:purgcong_29

The Purgatory Conglomerate is preserved at a spot called Purgatory Chasm. Here’s a shot of the chasm itself, cutting through the conglomerate outcrop down to the Atlantic Ocean. I’d guesstimate that it’s 10 m deep or so:

Here’s the Chasm, further out where the rock is clean of vegetation:purgcong_12

For a nice perspective on the whole area, check out this Quicktime 360° view.

So why is the Chasm there? We may get some insights by taking a look at another feature seen at the outcrop: a joint set which runs ~perpendicular to the long axis of the clasts:purgcong_25

A wider view, showing the same orientation of joints cutting across the conglomerate, ~perpendicular to the stretching lineation. Bikini babes at upper left for scale:

Now let’s go back to the Chasm, and take a look into it. Windsurfing board (washed out in daylight beyond the cleft) for scale:purgcong_34

There are a whole lot of joint faces there, all (a) perpendicular to lineation, and (b) parallel to the Chasm. You can see them all as parallel lines to the left of the Chasm. The large concentration of fractures in the area of the Chasm suggests that the Chasm was eroded out along a zone of more pervasively fractured rock. As you stand there and peer in, waves will come in and slosh towards the back end of the Chasm. But why is it so fractured here? I’m not sure.

In other places, you can see fractures that have “healed” into quartz veins:purgcong_23

This is where some of that dissolved quartz ends up, sealing shut these cracks. But not all the quartz veins I saw were perpendicular to lineation; there were some that were ~parallel to it, as in this photograph:


As a totally gratuitous bonus, the Purgatory outcrop also features glacial striations. In both of the following two photographs, the striae run from upper right to lower left:

Another spot, showing the same thing:purgcong_16

Annotation showing which lines are bedding traces internal to the quartzite cobbles, and which are glacial striations:purgcong_16

I’ve  got a few more photos of my visit to Purgatory in this Flickr photo set.

So: Thanks, E-an, for another great idea! What a cool place; I can’t wait to bring students back here…

Mineral phantasms?

Ice… serpentine… halite… What do they all have in common?

I’ve discussed mineral “ghosts” here before — really, those are only pseudomorphs, where one mineral’s chemistry becomes unstable due to a change in conditions, and then a new mineral forms in the same space. I’ve also brought up the issue of clasts of minerals which are unstable over the long term (ice).

Last night, at the final meeting of the Geological Society of Washington for the spring season, Bob Hazen of the Carnegie Institution of Washington gave the Bradley Lecture. Bob discussed his ideas about mineral evolution, and gave a compelling talk.

One of the key implications about thinking about minerals evolving over time is that new mineral species can evolve when conditions change and permit their growth, but so too can old mineral species go ‘extinct’ when conditions change and no longer promote their growth.

This got me thinking about that ice-clast breccia again (link above),  and how that would be interpreted by future geologists, assuming the ice itself has melted away. Consider the geologic record of a superwarm planet, where temperatures never dip low enough to form ice. Would we be imaginative enough to invoke ice as the cause of glacial landforms, of striations and deposits of till? How would we explain dropstones and ice wedges if ice were an “extinct” mineral on Earth?

And so after the talk was over, I went up to Bob and introduced myself and asked him if he could think of (or imagine) other minerals which could profoundly affect the geologic record, yet disappear after they have done their work. As we were talking, it occurred to me that halite in the form of salt domes could perturb the local stratigraphy, then the salt diapirs could rise up to the surface and be eroded (or re-dissolve into the ocean), leaving a piercing trail of destruction in their wake.

Bob came up with another one: serpentine at a subduction zone: hydrothermal alteration of oceanic crust produces serpentine, but then the serpentine is unstable when it gets subducted. It dehydrates (gives off water), and (poof!) there’s no more serpentine minerals. However, this dehydration is super duper important geologically: the addition of that water to the hot rocks of the subduction zone lowers the melting temperature of the rocks, and helps generate magma: the magma that rises to feed volcanic arcs. If we didn’t have oceanic crust to look at, would we have imagined serpentine beneath our convergent boundaries, a humble transformer of the world above?

Readers, I put the same question to you: Which minerals cause big effects, but then disappear? Who are the prime movers who flee the scene of the crime? These are minerals that aren’t just ghostly; they’re downright phantasmic!  I’ll be eager to read your suggestions, or hear your thoughts on the three I’ve noted here.



Glacial striations, southern Central Park

New York City has some cool geology: Paleozoic metamorphics scraped by Pleistocene glaciers.

Rockies course applications open

For those of you who are potential NOVA students (really, that’s pretty much anyone on the planet), I wanted to let you know that applications are now open for the July 2010 Regional Field Geology of the Northern Rockies course that I co-teach with Pete Berquist of Thomas Nelson Community College. A more detailed description is available on my website.

Contact me via e-mail if you want more information or download an application here.

To whet your appetite, here’s Rockies 2009 student Jason Von-Kundra mapping Mississippian-aged carbonates in the Bridger Range of Montana:


Faulted moraine

Continuing with the recounting of geological sights in the Owens Valley, California, area… This one is in the Pine Creek area. Take a look at this photo:


No, that’s not just a portrait of Jeff Lee and his awesome handlebar mustache. Look behind Jeff, on the hillside above.

See the little step down that the hill takes? Let’s zoom in:


Still can’t see it? Here, allow me to annotate it for you:


That’s a fault! A normal fault, with the Jeff side of the landscape dropping down relative to the mountain side (in the distance). Great, you might think. A subtle fault scarp. Big deal.

Oh, but you should not be so quick to dismiss it! After all, the material that the fault cuts across turns out to be a significant clue to the timing of when this fault happened.

This Google Map shows these two very well-developed lateral moraines extending out of Pine Creek Canyon:

In the Pleistocene, a valley glacier glided down out of the Sierran highlands into the Owens Valley to the east. As it flowed, it brought ground-up Sierran rocks down with it, depositing the sedimentary debris as glacial till. The fault above cuts through the northern lateral moraine. The moraines (made of till) are therefore Pleistocene in age, and since the fault cuts across the moraines, it must be more recent than the Pleistocene.

This is not a shocker: the boundary between the Sierra Nevada and the Owens Valley is well known to be a normal fault, as are most of the recent faults in the Basin and Range province. But being able to say “ten feet of offset have occurred on this fault since the Pleistocene” is a significant piece of data.

Cool, huh?