Joachim Heierli's Anti-Crack Theory

[the_big_lepowski]

Anyone very familiar with this concept?

You can gain a pretty good understanding by reading "the discussion" portion of his paper, here:

http://www.fsavalanche.org/NAC/techP...ierli_etal.pdf

I read the whole article and a lot of the technical stuff seemed jibberish to my apparently stupid brain. Just wondering how many people have heard much about this.

Basically what I got from the end of the discussion portion is that an ECT pit on a flat plane will collapse and fracture easier than a pit on an angled slope due to the increased downward stress on a slab on a flat plane.
So how I interpreted this is that you can dig pits on flat and low angle terrain and if the fracture doesn't propagate on an ECT in these pits than it should not propagate on a higher angle slope due to the decreased downward stress on the slab.

Of course one major flaw with this theory would be that it could be very hard to find an area with representative snow conditions if you are not even on the slope you plan on shredding...

Please discuss, or maybe learn something new by reading the article

[jono]

I wrote a tome that I would find ridiculous were it not so tiny in comparison. So I'll give my two take-aways and leave the further paragraphs to your discretion/boredom. First, ECT numbers are suspicious when it comes to correlating them with avalanche risk, specifically since numbers appear to rise with increasing slope, while risk most likely also rises with slope. (I didn't see an attempt to address the correlation, but please correct me if I missed it). Second, the need to reinvent the wheel in the name of "snow science" seems to have created quite a gulf between snow science and science. Complex but probably still incomplete geek-speak to follow.

The potential for major flaws in the paper seems very high because the assumptions, stated and unstated, don't receive enough attention to validate them. There's also the problem that the perceptible signal to noise ratio could be improved substantially if the paper were simply organized linearly so that the reader is not jumping around in an attempt to locate variable definitions and apparently missing steps in the math. (Force is listed in N/m. I still haven't found the explanation that says which way the meters are being measured: vertically? Perpendicular to slope? Ski length?)

One potentially major foundational problem is that the data is taken in support of the hypothesis using a technique which purposely favors expediency over precision despite needing precise results in order to prove or disprove the hypothesis. Another is the assumption that elastic modulus is essentially the same for snow whose structure is noted to be different; it's possible that this is true but I didn't see evidence presented (did I miss that?) and it's easy to imagine a lot of reasons why this would not be the case, starting with simple density differences.

There is also the fact that the ECT measures a force that is increasingly non-parallel to the column as the slope angle increases. In other words, the likelihood of compressive buckling ("anti-cracking" per Heierli) as the primary failure mode should be expected to go down as slope angle increases simply because the load applied is not direct compression. I don't think this is necessarily a criticism of Heierli's math, since math allows the same fact to be proven a lot of different ways, but I'm not yet willing to accept that his math is right just because the data seems to lean the same way. (That might seem like a small quibble, but considering the substantially more complex mathematical basis he's arguing for than, say, borrowing well established math from stress flow/fracture mechanics, I think it's worth noting for those wanting to build on the subject.) There are other potential explanations. He should consider the impact of the slope on the absolute strength of the snow in the weak layer since increasing strength, residual shear or tensile stress (or any combination thereof) should be expected to impact his measurements. (It would be particularly interesting to note what the initial angle of repose was and whether it impacts consolidation--that seems to be another subject, but it can't be separated from his experimental results.)

While I don't see much point in providing 31 pages of re-inventing stress flow theory, the experimental data is very interesting in that it calls into question the relative or proportional strength of snow, as indicated by ECT, on steeper slopes. So if a pit digger gets used to one number meaning something on a 26 degree slope he should be wary of thinking a higher number on a 32 degree slope necessarily indicates better stability, since Heierle's data indicates that the 'same snow' will give higher numbers on steeper slopes. Be very, very wary of assuming this means higher stability with increasing slopes. Since virtually all empirical data indicates more risk with steeper slopes it would be better to conclude that ECT numbers are not well correlated with stability at different slope angles.

[powhoy]

The anti-crack theory definitely takes some time to get your head around. In terms of interpreting ECT results, there are a couple paper's that will help.

http://fsavalanche.org/NAC/techPages...sBirkeland.pdf

In addition to other things, this TAR article briefly lays out the effects of changing slope angles.

http://fsavalanche.org/NAC/techPages...T_SlopeAng.pdf
http://fsavalanche.org/NAC/techPages..._Simenhois.pdf
These two articles show the methods and data used to show that ECTs results are not slope angle dependent, and perhaps show a slight increase in force with increased slope angle. This is shown with both persistent and non-persistent weak layers. Obviously this only holds if the snowpack structure is staying consistent with the changing with slope angles.

The force it takes for fracture initiation has been shown to be very spatially variable, so even if an ECT takes an additional 5 taps on a steeper slope it shouldn't effect your decision making very much. Instead you should be paying more attention to the ability of the snowpack to propagation a fracture.

Lastly, getting a representative pit location is one of the most difficult things to accomplish in the backcountry. Obviously digging in the starting zone of the path you want to ski would be the best from an information perspective, but isn't practical from a safety perspective. These papers show that moving to a lower slope angle can still give you good results. But you still want to make sure the snowpack is as similar as possible in every other way. You are never going to have a perfectly representative pit, even within one path there is too much variability for that to be the case.

Bottom line is that being able to do stability test on slopes that are not 37 degree allows for safer information gathering in the backcountry.

[cyborg]

I was doing some reading on this last night... definitely have more to do before I can fully discuss it.

But you asked for a roll call, so here I am. I'm familiar with it.

Stupid bit: Can't help but think, what's the difference between a collapse and an "anti-crack"?

[powhoy]

Stupid bit: Can't help but think, what's the difference between a collapse and an "anti-crack"?

No difference. My understanding is that "anti-crack" is simply engineering talk for collapse. Unless you're an engi-nerd there is no reason not to just talk about it as the "collapse model" of fracture vs. the older shear model.

[cyborg]

I assumed that but I didn't wanna jump to any ignorant conclusions haha.

I actually am an enginerd, but a mere apprentice. Got a few classes, fe exams, hundreds of sleepless nights, etc to go. Oh well...

However, the application of engineering principles, specifically mechanics and materials science, to snowpack analysis is something that is starting to fascinate me. There are a few guys out there doing it (just saw something on the civil e department at mt state running some studies), and it's pretty interesting. I'd definitely like to know more.

[jono]

A little scanning around online makes it appear that this is a geology term, and it would be awfully ironic given the Simonhois-Birkeland article, but it looks like maybe it's being used incorrectly. Consistent with what "anti-" implies mathematically, the couple of geology articles I ran across use it to describe a singularity scenario whose sign is opposite that of a crack. Unless we're all reading him wrong, Heierle is using it to describe failure/fracture perpendicular to the primary crack direction.

Powhoy, if ECT results are not slope dependant, how can they also "perhaps show a slight increase in force with increased slope angle?" Do you just mean that propagation isn't slope dependant? Because if the ECT strengths increase with slope (as Heierle's data implies) and avalanche risk also increases with slope that seems like a serious obstacle to quantitative understanding through ECT values. But if you're just looking to spot a qualitative observation it'd just be a footnote.

[powhoy]

Jono, sorry for the confusion. I should have written "OR perhaps show a slight increase in force with increased slope angle?" instead of "and."

Looking at the Birkeland et al. (2010) data, two of the four datasets show an increase in initation force with slope angle. None of the data on non-persistent weak layers in Simenhois et al. (2012) show a relationship between slope angle and initiation force. The datasets in Heierle et al. (2011) show a mixed bag, some show initiation force increase with slope, while several do not.

Basically, the data seems to show that in some circumstances the initiation force increases with slope angles, but in some circumstances it does not.

Do you just mean that propagation isn't slope dependant?

Yeah, I meant that too. Regardless of whether force for fracture initiation increases with slope angle, that is separate from the propagation potential itself.

Because if the ECT strengths increase with slope (as Heierle's data implies) and avalanche risk also increases with slope that seems like a serious obstacle to quantitative understanding through ECT values.

The reason that avalanche risk increases with slope is because the balance between the force of gravity and friction changes. Basically this ECT slope angle research is showing that just as you can get whoomphing on a flat slope, showing the ability to initiate and propagate a fracture, you can get that same info from an ECT on a moderate slope. Your ECT column won't jump out of the wall on flat slope, just as you won't get an avalanche, but it'll still show you that the snowpack is capable of propagating a fracture.

[cyborg]

So, if I'm understanding correctly, an "anti-crack", geologically, would be a sort of closing of a fissure, perhaps two masses pushing together? Whereas he's labeling what is basically a collapse in the snowpack an "anti-crack"?

How much of a strength increase are we talking? Marginal, I would assume. I still need to read over the papers referenced, but was it enough of a change in ECT strength to be statistically significant?

[jono]

Cyborg, I am only scanning the few geology papers/abstracts that come up in a Google search, so really cursory, but mathematically speaking I got a similar idea. I'm not sure of the mechanism and I think that oversimplifies the whole thing, but in terms of identifying the direction of the vector (tensor?) your description seems in line with what I saw.

Powhoy, I agree with your statement of why risk rises with slope, and that's obviously an easier thing to look at mechanically; angle of repose being a fairly straightforward analogy. But the point I keep coming back to is the practical implication that if ECT numbers are found to rise slightly with slope and risk also rises (albeit much more and for different reasons) then ECT numbers should be understood to be of (even) less value if not taken on very similar slopes. Does that seem true? (I may not be convinced that Heierle has a strong handle on the why, but his and other data do seem to show it, regardless of why.) If so, and perhaps more to the point, how relevant is it? Does anyone find that a go/don't go ever hangs in the balance on the number of manual taps? I'd hope not, but if people report them it seems possible. Further, it seems at odds with TBL's tentative takeaway in the first post, so maybe that's the relevance.

[homemadesalsa]

As an avalanche educator who uses the ECT extensively, as do the folks I teach with, we have been de-emphasizing the importance of the strength score in all tests since we have been able to quantify shear quality and proceed from there to the ECT and the importance of propagation.

I find that is actually an easy concept even for beginners to grasp, and present as a matter of course the "stability wheel," that discusses strength, energy (which we are now calling propagation propensity) and structure. In level 1 we talk primarily about heavy over weak, but go into the lemons in level 2.

This doesn't help with the anti-crack, but I think you all are head and shoulders above me on that topic. It was taught/ explained to me as "it should have been called the collapse theory." and I then explain it in level 2 mechanics classes as a wave that is propagated through the bottom of the slab, and the uniformity of the slab, as well as the slab stiffness, help to continue that wave. I can't remember which of Bruce Jamieson's grad students talked about the race between the propagation and tensile failure in the slab above, but that makes it clear to students. I use the analogy of shaking a rug from one end to show the anti-crack, and the type of rug will determine how well the failure is telegraphed. With a jute rug, the shake turns to a dud, while a rug with a backing (a stiffer more consistent slab) will send the shake all the way to the end.

Not sure if this helps your conversation, but it is how I translate theory to practice in my teaching practice.

[jono]

Thanks! Good analogies, and I'm glad we're not reading this completely differently than people in your shoes. Ironic that a quantitative model should become so abstract, but such is math, I guess.

[stalefish3169]

I can't remember which of Bruce Jamieson's grad students talked about the race between the propagation and tensile failure in the slab above, but that makes it clear to students.

IIRC, you're referring to Alec van Herwijnen.

Here is a link to his 2009 ISSW paper MEASUREMENTS OF WEAK SNOWPACK LAYER FRICTION that Heierli co-authored which provides more details on some of the topics in this thread. It is on page 112.

http://www.wsl.ch/dienstleistungen/p...n/pdf/9854.pdf