Deco Theory 101, 201, 301, and 401

Please register or login

Welcome to ScubaBoard, the world's largest scuba diving community. Registration is not required to read the forums, but we encourage you to join. Joining has its benefits and enables you to participate in the discussions.

Benefits of registering include

  • Ability to post and comment on topics and discussions.
  • A Free photo gallery to share your dive photos with the world.
  • You can make this box go away

Joining is quick and easy. Log in or Register now!

stuartv

Seeking the Light
ScubaBoard Supporter
Scuba Instructor
Messages
11,573
Reaction score
8,124
Location
Lexington, SC
# of dives
500 - 999


A ScubaBoard Staff Message...

This thread was moved from the Advanced forum leaving a "Permanent Redirect" link. Don't be confused it you suddenly find yourself in this forum.

This presentation, by @Dr Simon Mitchell, titled "What is optimal decompression" is, quite simply, awesome! It will take you from the very basics right through to the very advanced. It includes analysis of the current status (as of May, 2020) of the known science comparing deep stops and bubble models to shallower stops and dissolved gas models.

It is 2 hours long, including the Q&A session at the end, but there is great info that is covered in the Q&A as well, so if you have time to watch it, and you are not already an expert on decompression science, I think it is worth the time.


After watching it, I now have a couple of questions for Dr Mitchell. Dr Mitchell, if you are reading this, I hope you won't mind me asking these questions here in this venue.

1) You presented analysis of the Integral Supersaturation (IS) of various dive profiles. The analysis showed that the IS predicts outcomes that match the experimental data - which showed that (putting it in simple terms for the sake of discussion) shallower stops have less risk of DCS than using deeper stops.

If I understand the IS calculation process correctly, the IS is calculated by taking the IS of each individual compartment (of the 16 compartments defined by Buhlmann ZHL-16C) and summing them to result in 1 number, which is then compared against the IS from other dive profiles.

My question is: Why do you sum the 16 numbers? Why not take the Maximum, from among all the 16 compartments and compare that?

2) I corresponded with you 2 or 3 years ago, when you had said that you had settled on use of GF50/80 for your personal diving. I asked you then if there was any reason to think that GF80/80 would be more or less likely to result in DCS - because it seemed to me that using the same number for GFLo as you use for GFHi would be safer than using a number for GFLo that was lower. At the time, my takeaway from your response was that using 80/80 might be better than 50/80, but that that was a big departure from Best Practice up to that point, so you felt it was best to approach that slowly, rather than jumping straight to 80/80.

Through most of the presentation, it seemed like the data showed support for the notion that using the same number for GFLo and GFHi would be "less risky". The IS for any two ascent plans of the same total length would always (I think) be lower, when GFLo was higher.

But, a presentation I saw recently, by Alessandro Marroni, seemed to suggest that using a GFLo that is actually lower than your GFHi might be advantageous. E.g. GF50/80 might actually be safer than GF80/80.

Also, towards the end of your presentation, you said that your current settings are still GF50/70 or 50/80, depending on the dive.

My question here is: What are your current thoughts on using 70 or 80 for GFLo (i.e. the same as whatever GFHi value you are using)? Any change in your thinking since the time we spoke about this in the past?
 
@stuartv Mark Powell is running another deco theory online course/workshop thing starting on the 27th, I believe he is covering all sorts of ungodly time zones. Maybe you should join one,

As I understand it, from reading all the entertaining threads that kept people off the streets and lead to various colourful visualisations for IS, the point is to integrate risk over time. Is it riskier to be over saturated in just the blood or the blood and the brain? If the blood is extremely oversaturated and the brain a bit less is that the same as the blood extremely oversaturated and the brain not saturated at all?

This is different to the simplistic “no compartment may cross this (per compartment) line”. You could imagine a planner that worked by trying out all the possible plans and finding the one with the minimum integral supersaturation for a given bottom time and run time. Such a plan might have really high instant supersaturation if it made the other compartments lower on average.
 
After watching it, I now have a couple of questions for Dr Mitchell. Dr Mitchell, if you are reading this, I hope you won't mind me asking these questions here in this venue.

Hello Stuart, thanks for the positive comments. It does move a little slowly because I was trying to speak slowly for the Polish translation. No problem at all with questions!

1) You presented analysis of the Integral Supersaturation (IS) of various dive profiles. The analysis showed that the IS predicts outcomes that match the experimental data - which showed that (putting it in simple terms for the sake of discussion) shallower stops have less risk of DCS than using deeper stops.

That is fundamentally correct, though the conclusion ("shallower stops have less risk of DCS than using deeper stops") is only correct for the profiles I compared. There is a danger of implying that endlessly shallower stops must be better. If one omits too much 'deeper' decompression, then supersaturation will be increased and outcomes will likely be worse.

If I understand the IS calculation process correctly, the IS is calculated by taking the IS of each individual compartment (of the 16 compartments defined by Buhlmann ZHL-16C) and summing them to result in 1 number, which is then compared against the IS from other dive profiles.

My question is: Why do you sum the 16 numbers? Why not take the Maximum, from among all the 16 compartments and compare that?

Its because, among other things, bubble formation is influenced by by both the magnitude of gas supersaturation, and the duration over which supersaturation persists. So, putting it simply, a higher magnitude will drive bubble formation more strongly (bubbles will form more quickly), and a greater duration will allow more bubbles to be formed. If you consider the maximum only, then you are only considering one of these two important influences. We sum all 16 compartments because all potentially contribute to bubble formation and risk and ignoring all the others just because one compartment has a higher supersaturation at a particular time makes no pathophysiologic sense.

2) I corresponded with you 2 or 3 years ago, when you had said that you had settled on use of GF50/80 for your personal diving. I asked you then if there was any reason to think that GF80/80 would be more or less likely to result in DCS - because it seemed to me that using the same number for GFLo as you use for GFHi would be safer than using a number for GFLo that was lower. At the time, my takeaway from your response was that using 80/80 might be better than 50/80, but that that was a big departure from Best Practice up to that point, so you felt it was best to approach that slowly, rather than jumping straight to 80/80.

Through most of the presentation, it seemed like the data showed support for the notion that using the same number for GFLo and GFHi would be "less risky". The IS for any two ascent plans of the same total length would always (I think) be lower, when GFLo was higher.

But, a presentation I saw recently, by Alessandro Marroni, seemed to suggest that using a GFLo that is actually lower than your GFHi might be advantageous. E.g. GF50/80 might actually be safer than GF80/80.

Also, towards the end of your presentation, you said that your current settings are still GF50/70 or 50/80, depending on the dive.

My question here is: What are your current thoughts on using 70 or 80 for GFLo (i.e. the same as whatever GFHi value you are using)? Any change in your thinking since the time we spoke about this in the past?

The choice of GF Lo is effectively the choice of how much emphasis you wish to place on deeper stops. As you know, the lower the number, the more emphasis will be placed on deep stops and vice versa. At this point in time we have reasonable evidence to suggest that the degree of emphasis placed on deep stops by bubble models was inefficient, that is, you could achieve less risk of DCS in a decompression of the same duration as prescribed by the bubble model by not placing so much emphasis on deep stops. On the other hand there is some evidence to support the hypothesis that endlessly shallower stops are not the right approach either. In that regard I am speaking of Sandro Marroni's recent (as yet unpublished) experiments using 40m decompression dives that a decompression that approximated GF100:100 was not as good as one that approximated GF20:100. Not a particularly surprising result. Anyway, the point is that in choosing how much emphasis to place on deep stops, I have looked for a compromise between bubble models (which in GF terms would often have given a GF Lo of 10 or 20) and raw Buhlmann (GF Lo of 100). I don't have any hard data to inform this choice, but intuitively I have landed on somewhere around 50. I would not consider 60 a stupid choice, but I suspect that 80 may be bordering on too high. The sweet spot for GF Lo and how to choose it for a particular dive is still something we don't have a definitive answer on.

Simon M
 
That is fundamentally correct, though the conclusion ("shallower stops have less risk of DCS than using deeper stops") is only correct for the profiles I compared. There is a danger of implying that endlessly shallower stops must be better. If one omits too much 'deeper' decompression, then supersaturation will be increased and outcomes will likely be worse.

Understood. I liked the way you put it in the presentation. If (unqualified) shallower stops were better, then we would make our first stop at the surface. Which is obviously not the right approach... LOL :)

Its because, among other things, bubble formation is influenced by by both the magnitude of gas supersaturation, and the duration over which supersaturation persists. So, putting it simply, a higher magnitude will drive bubble formation more strongly (bubbles will form more quickly), and a greater duration will allow more bubbles to be formed. If you consider the maximum only, then you are only considering one of these two important influences. We sum all 16 compartments because all potentially contribute to bubble formation and risk and ignoring all the others just because one compartment has a higher supersaturation at a particular time makes no pathophysiologic sense.

I think you may have misunderstood what I was asking. Or I have misunderstood your response.

I am interpreting your response to mean that you interpreted my question of "why don't we look at the Maximum, instead of the sum" as looking at the maximum instantaneous supersaturation. That is not what I meant. What I meant is, why don't we look at the IS of each of the 16 compartments and, instead of summing them, look at the maximum IS of any one compartment?

Example:

One dive profile results in one compartment having an IS of 10 and all fifteen of the other compartments had an IS of 1 each. The sum - the IS, the way your calculating it - is 25.

A different profile (for the same depth, bottom time, and run time) results in having nine compartments with an IS of 2, and the remaining seven compartments having an IS of 1. That also yields an IS of 25.

The implication is that these two dive profiles are equally at risk for resulting in DCS.

My question is why don't we score the first profile as a 10 and the second one as a 2 (instead of scoring them both a 25)? Intuition (which can obviously be VERY wrong) suggests that the profile that had an IS of 10 in one compartment is more likely to result in bubble formation that could lead to DCS - as compared to the profile that had a maximum IS in any one compartment of 2. This would seem to be in line with how the Buhlmann algorithm works, where it always determines which compartment is the controlling compartment, but looking at the maximum over pressurization in each one individually. Why not look at IS the same way - identifying a controlling compartment for IS, instead of summing them all?


Also, thank you so much for being so willing to share your time and knowledge!
 
Understood. I liked the way you put it in the presentation. If (unqualified) shallower stops were better, then we would make our first stop at the surface. Which is obviously not the right approach... LOL :)



I think you may have misunderstood what I was asking. Or I have misunderstood your response.

I am interpreting your response to mean that you interpreted my question of "why don't we look at the Maximum, instead of the sum" as looking at the maximum instantaneous supersaturation. That is not what I meant. What I meant is, why don't we look at the IS of each of the 16 compartments and, instead of summing them, look at the maximum IS of any one compartment?

Example:

One dive profile results in one compartment having an IS of 10 and all fifteen of the other compartments had an IS of 1 each. The sum - the IS, the way your calculating it - is 25.

A different profile (for the same depth, bottom time, and run time) results in having nine compartments with an IS of 2, and the remaining seven compartments having an IS of 1. That also yields an IS of 25.

The implication is that these two dive profiles are equally at risk for resulting in DCS.

My question is why don't we score the first profile as a 10 and the second one as a 2 (instead of scoring them both a 25)? Intuition (which can obviously be VERY wrong) suggests that the profile that had an IS of 10 in one compartment is more likely to result in bubble formation that could lead to DCS - as compared to the profile that had a maximum IS in any one compartment of 2. But, this would seem to be in line with how the Buhlmann algorithm works, where it always determines which compartment is the controlling compartment, but looking at the maximum over pressurization in each one individually. Why not look at IS the same way - identifying a controlling compartment for IS, instead of summing them all?


Also, thank you so much for being so willing to share your time and knowledge!
Good question, Stuart. Looking forward to the answer!
 
My question is why don't we score the first profile as a 10 and the second one as a 2 (instead of scoring them both a 25)?

I suspect if you want to go down that rabbit hole, you'd want to also see which compartment was the 10: fast ones are supposed to tolerate greater overpressure and although it's already factored in (via M-value), you may still want to give them different "weight" for the scoring. While at it, you might want to give less weight to TCs associated with skin bends and more: to those linked to neurological DCS, and so on. Sticking to KISS principle may be the better option.
 
Or maybe, instead of the simple "integral supersaturation", it might be interesting to look at the integral of the percentage of the way to the M-value. I.e. if you spend a long time at a GF99 of 100, is that better or worse than a short time at GF99 of 150?

The thing I don't like about this idea is that IS has no relation to the Buhlmann M-values. The only thing it takes from Buhlmann is the number of compartments and their half-lives. And those are, to be fair, somewhat arbitrary, I think.

Meaning, you have 16 compartments that range in half-life from 4 minutes to 635 minutes, or something like that. But, there is nothing that inherently requires exactly that. You could construct a very similar model with 18 compartments and half-life values from 3.5 minutes to 650 minutes if you wanted and end up with results that are virtually identical (in practice).

If you start analyzing integrals of calculations that depend on M-values, then I think you are adopting a platform of assumptions that you really might not want.

What I think would be REALLY interesting is an analysis of IS that decouples on-gassing and off-gassing rates. E.g. Instead of a 4-minute compartment, which is defined to be one that goes halfway from current pressure to (a higher) ambient in 4 minutes and then halfway from it's current down to a lower ambient in 4 minutes, maybe the compartment has an on-gassing half-life of 4 minutes, but an off-gassing half-life of 5 minutes.

If you think about a physical model, an analogy might be a magnet attracting particles of iron. They move towards the magnet quickly and easily. To move away from the magnet is slower and harder. What if on-gassing and off-gassing is similar? I.e. some effect of binding energy between the molecules that makes a certain compartment on-gas with a shorter half-life than it off-gasses?
 
  • Like
Reactions: OTF
The thing I don't like about this idea is that IS has no relation to the Buhlmann M-values. The only thing it takes from Buhlmann is the number of compartments and their half-lives. And those are, to be fair, somewhat arbitrary, I think.

Meaning, you have 16 compartments that range in half-life from 4 minutes to 635 minutes, or something like that. But, there is nothing that inherently requires exactly that. You could construct a very similar model with 18 compartments and half-life values from 3.5 minutes to 650 minutes if you wanted and end up with results that are virtually identical (in practice).

I expect it's implicitly between the ambient and M-value, unless it's a dive with "blown deco", but that's where the larger M-value for fast TC comes in and lets you have the "IS of 10 in one compartment".

Arbitrary compartments are indeed the point of ZH-L16: he figured out that `a = 2 * t ^ -1/3` and `b = 1.005 - t ^ -1/2` works for any half-time t. (At least within the range of empirical data available.)
 
https://www.shearwater.com/products/teric/

Back
Top Bottom