Connect with us

Education

Part Two: “Tech Divers, Deep Stops, and the Coming Apocalypse”

In part two of this four-part series on the history and development of GUE’s decompression protocols, GUE founder and president Jarrod Jablonski discusses the lack of appropriate mixed gas decompression tables in the early days of technical diving. He goes on to discuss the initial development of “ratio decompression” and the early thinking and rapid adoption of “deep stops,” which have come under pressure as a result of new research that questions their efficacy and safety. He also discusses the emergence of the now nearly ubiquitous gradient factors (GF) used in most dive computing. Feel free to DIVE IN and share your thoughts.

Published

on

If you have not read Part One of the “Decompression, Deep Stops, and the Pursuit of Precision in a Complex World” series please find it here.

by Jarrod Jablonski

Header Photo by JP Bresser

Technical diving means different things to different people, opening legitimate arguments about different time periods and personalities involved in shaping this activity. Our purpose is more confined as I wish to focus mainly upon the development of decompression procedures. Leading figures like Jacques Cousteau (1910–1997) performed reasonably significant “technical” dives including both deep and overhead exposures. Yet, his dives were relatively short, often on air, and absent a community of fellow tech divers with whom to evolve varying strategies. Meanwhile, military or commercial activities reached significant depth with notable exposures but used different technologies and procedures. Hopefully, the following overview will be independently interesting—although I also hope to illustrate what I consider an intriguing wrinkle currently lacking in most debates over the best ascent schedule for a decompression dive.

For our purposes, I mark the 1980s as the period in which technical diving truly started to become a globally recognized activity. Of course, numerous significant projects occurred before this period, but the 1980s established a growing global awareness that amazing diving feats were possible by enthusiasts. The 1990s and the developing internet age brought these ideas progressively into the mainstream. Most importantly, these global communities could now easily communicate information and, of course, disagree about the best way to do one thing or another. 

Jarrod Jablonski, George Irvine, and Brent Scarabin preparing for dives in Wakulla Springs (late 1990s). Photo from the GUE archives.

The 1980s and 1990s were also an interesting period of development because people were doing more aggressive dives but still lacked important support tools that would take another decade to materialize. For example, most early tech dives were using some form of military, commercial, or scientific tables, such as the U.S. Navy, Oceaneering International, Inc. or the National Oceanic and Atmospheric Administration (NOAA), which were often not well-suited to the dive at hand. This is because there were no decompression programs available. Divers were unable to calculate their own profiles and would trade whatever tables they could gather. In many cases, they were forced to choose from tables that were calculated at a different depth, time, and/or breathing mix to the planned dive. 

George Irvine after a dive with the WKPP. Photo from the GUE archives.

Over time, the physiologist Dr. R.W. “Bill” Hamilton (1930–2011) began creating custom tables, but the lack of flexibility, as well as the cost, discouraged frequent use for most divers, especially when considering a range of depth profiles and their various time adjustments. This problem was especially prominent in deep cave diving where variable profiles and long bottom times created numerous complications . This landscape encouraged, if not demanded, that exploration divers be creative with their decompression practices. For example, Woodville Karst Plain Project (WKPP) explorers George Irvine and I began to explore ways to extrapolate from existing tables. The birth of what is now called ratio decompression originated with this practice. 

By exploring the way decompression schedules develop, we began outlining simple ratios that allowed divers to adjust their decompression based upon a ratio of time spent at depth. Then, in 1998, I went on to form Global Underwater Explorers (GUE), and such practices became part of the process for helping divers appreciate the structure of decompression while supporting adjustments to profiles when depth or time varied from that expected. Regrettably, some individuals took this too far and began promoting complex adjustments and marketed them as superior to the underlying algorithms from which they had been derived.

We should leave a more in-depth review of ratio deco for another time. For now, I intend to illustrate that “rules” such as ratio deco had their roots in limited availability of decompression tables and evolved as a useful tool for understanding and estimating decompression obligation. The fact that early tech divers had limited ability to calculate profiles encouraged a “test-and-see” philosophy, further fueling the early popularity of ad hoc adjustments to decompression profiles. This was most notable in the DIR and GUE communities through ratio-style adjustments and also prominent with non-DIR advocates who used a modification known as Pyle-stops, originating from ichthyologist Richard Pyle’s early deep dives and his attempts to refine efficient ascent protocols. One common adjustment in these approaches involved a notable reduction in ascent speed, adding additional stops known as “ deep stops

Driven by self-discovery, the migration toward deep stops resulted in rare agreement among technical divers, and the practice developed a life of its own in conference symposiums, being advocated far and wide by a healthy share of technical divers. The earliest phases of these modifications were driven in part by limited access to decompression tables, though by the end of the 1990s there were a variety of decompression programs available. This was a big improvement, although it was still true that divers had a limited baseline of successful dives while planning mixed-gas dives in the 50 to 100 meter range and beyond. Today, most divers take for granted that technical dives planned with available resources and within today’s common use (a few hours around 90m/300 ft) are relatively safe in terms of decompression risk. When these programs first came to market, there remained many question marks about their efficacy. 

Jarrod Jablonski and Todd Kincaid use Doppler while evaluating decompression profiles in 1995. Photo from the GUE archives.

The availability of decompression programs was a big advantage for technical divers, as they now had more sophisticated tools at their disposal. Yet, the output from many of these programs produced tables that were sometimes different to profiles thought to be successful by some groups. Typically, the difference related to an increase in the total decompression time and/or a distribution of stops that varied from developing consensus. These divers were keenly interested in the developing tools but reluctant to change from what appeared successful, especially when that change required additional hours in the water. A debate developed around the reason for these differences, bringing the already brewing interest in bubbles well into the mainstream. 

At the time, all tables were based upon dissolved gas models like Buhlmann, and thus not directly modeling bubble development during an ascent. Dissolved gas models preference ascents that maintain a reasonably high gradient between the gas in tissues and the gas being breathed, which should be supportive of efficient elimination from tissues. Dissolved gas models manage—but do not explicitly control for—bubbles and were thus labeled (probably unfairly) as “bend and mend” tables. In fact, dissolved gas models are designed to limit supersaturation (explicitly), because this is supposed to limit bubble formation. Haldane references this control of supersaturation in his pioneering publication .

Those advocating for different ascent protocols imagine that a lack of deep stops creates more bubbles, which then need to be managed during a longer series of shallow decompression stops. In this scenario, a slower ascent from depth, including “deep stops,” would reduce the formation and growth of bubbles while reducing time that might otherwise be needed to manage previously formed bubbles. Support for these ideas gained momentum in diving and professional communities, although some of these individuals were arguably conflicted by a vested interest in promoting deep stop models. Eventually, the practice received more critical consideration but during the intervening years deep stops were largely considered as common knowledge.

Jarrod Jablonski and George Irvine at the last stop of a 15-hour decompression in 1998. Photo from the GUE archives.

The idea of controlling bubbles became extremely popular through the 1990s, encouraging deeper stops designed to limit bubble formation and growth. The “test-and-see” approach developed by early tech divers appears to have fueled the promulgation of deep stops though the determining characteristics remained poorly defined. Early tech divers embraced the uncertainty of their activity, realized that nobody had the answers, and decided to “experiment” in search of their own answers. In fact, divers were experimenting with a lot more than deep stops, altering gases breathed, the placement of various decompression stops, and the total amount of decompression time utilized. This meant that divers were sometimes aggressively adjusting multiple factors simultaneously. 

For example, divers using a decompression program might input significantly less helium than was present in their breathing mixture. This was done because the algorithm increased decompression time with elevated helium percentage, sometimes known as the helium penalty . In other cases, divers would completely change the structure of stop times. For example, divers would invert the way a profile should be conducted by doing more time near a gas switch and less time prior to the next gas switch, believing the higher-oxygen gas only 3m/10ft away was time better utilized. 

I do not intend an exhaustive or detailed review but only to assert the many, sometimes radical approaches being taken by tech divers who often perceived success with these various strategies. In some cases, it appears these practices may have been leading the way toward improvements (eliminating the Helium penalty) while other adjustments might prove disadvantageous. In both cases, it is important to note that divers understood—or should have understood—these actions to be potentially dangerous, accepting risk as a natural part of pushing into uncertain territory where definitive guidance and clear borders are rarely available. 

None of this is to argue that divers should engage in aggressive decompression or challenge convention or place themselves at risk to unknown complications. I merely wish to clarify the atmosphere under which these adjustments were conducted, while highlighting that conventional ideas of risk mitigation are inherently complicated against a backdrop of novel exploration. A sense of relative risk dominated most of these trials since there are also risks associated with lengthy, in-water exposure. Eliminating what might be an avoidable decompression obligation could reduce risk from other obvious factors, including changing weather, dangerous marine life, being lost at sea, hypothermia, and oxygen exposure. Decompression experimentation was but one of many attempts to establish new protocols during extensive exploration projects. 

During this time, deep stops and similar adjustments were part of the “norm” for aggressive technical explorers who were sometimes notably reducing in-water time as compared to available tables. It is interesting to note that individual differences in susceptibility seemed among the most prominent variables across the range of tested adjustments, and we will return to this in a later discussion. For now, we should acknowledge that the “success” being achieved (or imagined) is greatly complicated by a small sample size of self-selected individuals who were simultaneously experimenting with a range of variables. On the other side, I do not want to entirely discount the results being seen by these divers. We should remember that development of safe decompression ascents for the general diving community was not the goal of most tech divers. These divers were interested in maximizing their personal and team efficiency during decompression. These strategies may or may not have been objectively successful or broadly applicable, but many teams imagined them so, at least within the narrow scope being considered. Later, we will return to these important distinctions with careful consideration for the potentially different interests being pursued by divers evaluating different decompression profiles.

Fortunately, a desire to create broadly useful tools was high on the list of priorities for some individuals in the technical diving community, leading to a relevant and important contribution. This inventive approach would introduce a new way to think about decompression, remaining to this day at the center of the debate about deep stops. 

Creating a New Baseline

Despite the popularity of deep stops and other modifications presented previously, technical divers lacked a common language for comparing the results, especially across different profiles. Compounding this problem was the variable way decompression programs considered a dive to be more or less “safe.” For example, some programs developed “safety factors” which increased total decompression time by an arbitrary factor, i.e., made them 10% longer. Other programs used different strategies, though it was not clear whether any of the various safety factors actually made the decompression safer. Whether safer or not, these factors were typically inconsistent and added complications when comparing various profiles. Just as these debates were reaching a fervor, help arrived from an unlikely place. An engineer by trade and decompression enthusiast by choice, Erik Baker had developed a novel solution. 

George Irvine passing time during a long decompression. Photo from the GUE archives.

Baker was seeking a way to establish consistency in considering the safety or lack of safety between various profiles. The term Baker applied was Gradient Factor (GF). I will not invest considerable time exploring the science behind GF, as many useful resources are widely available. For our purposes, it is sufficient to say that GFs allows a user to establish a lower threshold than the maximum recommended by a dissolved gas algorithm. This maximum pressure or M-value is assigned to “compartments” having an assumed amount of tolerance relative to the flow of blood they receive. Decompression theory is complicated by many factors but when the pressure of a gas in a compartment nears the M-value, it is thought that the risk of decompression sickness becomes higher. By adjusting a profile through the use of GF, one presumably reduces the risk. However, this also means there is less gradient between the gas in tissues and the gas in blood, reducing the driving force for the removal of gas and probably requiring additional decompression time. 

Gradient factors took an important step toward using a consistent language when talking about a variety of adjustments divers might make to their decompression. As part of his work on GF, Baker had developed a keen interest in the decompression adjustments by leading technical divers. Baker and I began working together while evaluating some of our most extreme diving profiles. This collaboration led to a number of productive developments including GUE’s DecoPlanner, released in 1997, and among the first to utilize GF methodology. These collaborations further highlighted what appeared to be a discrepancy between the decompression expected by dissolved gas algorithms and the decompressions being conducted by many technical divers. 

Decompression experimentation tended to cluster in small groups whose size was likely affected by self-selection with members who would stop or reduce aggressive dives when experiencing decompression sickness. Yet, it seemed possible that there was more to the story. These divers were doing several things that should have exposed them to notable risk and yet were repeatedly completing decompressions of more than 15 hours. Deep stops were only part of the story as these divers were ignoring conventional wisdom in several areas while notably reducing decompression time. What was behind this discrepancy? Were deep stops right or wrong? Was the conservative approach to helium right or wrong? Were these individuals lucky? Were they unusually resistant to decompression sickness, or were there other factors lurking in the background?

Note: I will outline many of these developments in the upcoming Part Three, where we more directly consider the modern challenges to deep stops and most especially the assertion they are dangerous. In the interim, I hope to hear from our readers. Do you have different experiences from this period? Do you think such experimentation is reckless or inadvisable? Please let us know your thoughts.


Jarrod is an avid explorer, researcher, author, and instructor who teaches and dives in oceans and caves around the world. Trained as a geologist, Jarrod is the founder and president of GUE and CEO of Halcyon and Extreme Exposure while remaining active in conservation, exploration, and filming projects worldwide. His explorations regularly place him in the most remote locations in the world, including numerous world record cave dives with total immersions near 30 hours. Jarrod is also an author with dozens of publications, including three books.

1 Comment

1 Comment

  1. Robert

    October 9, 2019 at 9:25 am

    I think the main obstacle to meaningful experimentation is that the observable is so unclear and noisy. “I felt better after the dive” can depend on so many factors including in particular confirmation bias. Even more objective criteria like “presence of clinical DCS symptoms” or Doppler counts are not much better due to their intrinsic variability. To be sure that a given profile leads to less than one accident per thousand dives (a risk that is still considered too high by many) you need to do significantly more than 1000 dives with that profile (while controlling all other relevant factors. Ideally, you want to know as well that the profile is optimal in the sense that shortening it leads to too high rate of accidents. Good luck with confirming that empirically! Apart from that everything is just anecdotal evidence of questionable use. Don’t get me wrong, research here is extremely useful. It’s just that progress is slow and we will see many detours of procedures though of as useful by a group of people for some time that later turn out to be wrong.

Leave a Reply

Your email address will not be published. Required fields are marked *

Education

Decompression Series Part Four: Finding Shelter in an Uncertain World

In the final of this four-part series on the history and development of tech decompression protocols, GUE founder and president, Jarrod Jablonski weaves together various forays into decompression science, including Brian Hill’s pioneering pearl diver study, the NEDU’s work on deep stops, evidence of individual susceptibility, and probabilistic decompression models in an attempt to define the state of our understanding. It may give you pause to stop. Feel free to add your comments.

Published

on

By

By Jarrod Jablonski

Header photo courtesy of the GUE archives

Did you miss Part III? Read it here.

The human quest to explore below the water’s surface began some 5,000 years ago . Since that time, our species has pursued deeper and longer immersions, charting a course through hundreds of years of diving activity and associated research. Many of the advances in procedure, technique, and equipment are a direct result of the compelling and valuable data and experience documented during underwater explorations. As with many novel activities, this process of advancement required pushing physical and intellectual barriers. 

During the 1980s and 90s, advances in technology supported an activity that became known as technical diving. This diving led to the development of ascent practices which were somewhat different from those of scientific, military, and commercial divers. A unique set of needs and limited relevant examples encouraged a great deal of experimentation among these early explorers, including adjustments to breathing gases and the distribution of decompression stops used during their ascent. Some technical divers began using a slower ascent from depth, in the hope this would control the formation of bubbles. These slow ascents became known as “deep stops” and were practiced in the hope they could reduce decompression stress and/or shorten decompression time. 

In fact, the idea of bubble control was not new. During the 1960s, physiologist Brian Hills sought to characterize the profiles of pearl divers who had been operating since the late 1800s. These divers were interesting because they were ascending in two-thirds of the time required by Navy tables, a time that would cut even more decompression from most modern-day ascent schedules. Hills believed the reduced decompression times were the result of a unique ascent profile, including stops deeper than those called for by the Navy tables. Years later, technical diving explorers started adopting similar techniques while reporting reductions in total decompression time. It is difficult to qualify if this perceived success was actually occurring since the groups were relatively small, not carefully monitored, and simultaneously adjusting numerous other factors during their ascent. Even absent these complications, the generally low risk of decompression sickness can greatly complicate evaluations between different strategies. 

The enthusiasm for deep stops likely reached its peak in the late 1990s and was dealt a serious blow by the previously discussed Navy Experimental Diving Unit (NEDU) study that was released in 2011. This study, and others, propose that deep stops are less efficient and may actually increase the risk of decompression sickness. The reader should refer to part three of this series for discussion and references. This series contends that opposition to deep stops is supported by prevailing research, but that a range of other variables need to be considered in order to effectively develop best practices. These aspects are particularly relevant to experienced divers, who report decompression sickness problems when eliminating slower ascents from depth.

Jarrod Jablonski with the Halcyon PVR-BASC semi-closed rebreather aka “The Fridge.” Photo courtesy of GUE archives

It is not my intent to re-litigate the previous three sections of this article, but an interesting, and I believe, underappreciated aspect of Brian Hills’ pearl diver study provides a nice segue. What I find most interesting are the roughly 3,000 deaths and injuries of an unknown quantity that helped shape those unique ascent profiles. In other words, how was this conclusion affected by the elimination of those who are more susceptible to injury, and how much was due to a lack of rigor in the study ? Hills concluded that the success of the profiles was “due to the much deeper initial decompression stops used” by the pearl divers. In a similar way, technical divers took note of the history, the encouragement from experts, and the perceived success by those in their community. 

Given new and mounting evidence against deep stops, can we now definitively conclude that Hills, the pearl divers, and the tech divers were wrong? Are we sure the perceived success was imagined? If some success occurred, was it more about the generally low levels of risk in decompression sickness? Or could something else worth considering be at play? Asked another way, we might inquire how the conclusions reached by Hills and those technical divers are different from the way modern-day decompression tables have come into being.

The history of pearl diving and deep stops is very different from that of most decompression research in at least two substantial ways. The first difference has to do with methodologies, and the second with objectives. In terms of methodology, most decompression research is conducted using the scientific method: developing a testable hypothesis and, hopefully, crafting well-devised experiments in order to interrogate the hypothesis. Open publication of methods and results, internal and external debate, and reproducibility of results are among the many ways in which a hypothesis will be tested over time, narrowing the results toward either a more or less trusted conclusion. The history of deep stops, and possibly to a lesser extent, that of pearl divers, share few, if any, of the rigors commonly associated with the scientific method.

Individual Susceptibility

Looking to the history of decompression research, the objective of a particular study is implicit, if not explicit, in the development and testing of a hypothesis. With decompression profiles, we seek to balance the safety of the majority while not unduly affecting the whole. For example, we seek ascent profiles that keep a high percentage of individuals from being injured while not greatly extending the decompression time of the group as a whole. What would the results look like if we instead sought the most efficient decompression for a select minority of individuals?

 Some researchers joke among themselves that they already know who will get bent among a group of test individuals. This is because research trials require a lot of volunteers among a relatively small population of willing participants, meaning that some of the same individuals are often involved in multiple experiments. This is not to say that a few individuals have skewed all research, but rather to say that a minority of subjects in all research projects can affect the outcome by being particularly susceptible to decompression stress. 

This individual susceptibility is likely no surprise to anyone and is relatively well established among researchers, as is the variability in one individual from one day to the next. We see such variability in almost every conceivable area of our lives, affecting the way we respond to everything from drugs and alcohol to food and criticism. How could it be otherwise? We are all a kind of genetic experiment, refined through time with an endless series of personal and species-wide successes and failures. If we are variably sensitive to decompression stress, as seems almost certain, then in what myriad of ways might that be playing out? 

It appears that some individuals bubble more and some less on the same profile. Might they also be more or less sensitive to whatever collection of bubbles are generated? Is it possible that we develop different collections of symptoms to various types of decompression stress? That we are individually more or less sensitive to similar symptoms? Some of these factors we believe to be true and some we might suspect to be true. Many others lurk in the background, and all impact our sense of what we might call decompression stress. 

Casey Mckinley, Jarrod Jablonski, and George Irvine before a dive with the Woodville Karst Plain Project (WKPP). Photo courtesy of the GUE archives.

Given a world filled with individuals, we must do our best to bridge the divide. The good news is that we do this relatively well seeing that some differences are important but most are not usually extreme. The tail of the distribution represented by a small number of resistant individuals may well be quite small. This means that building profiles for resistant individuals might not have much impact and/or might be unreasonably dangerous. Either way, this individual variability is highly relevant and holds promise for the future. The next big advancement in health care will likely involve personalized medicine. Most of us may not live to see the usefulness of these developments in medicine, much less in decompression research, but the process is nonetheless hopeful. For example, research on heart rate variability ( HRV ) might be one such development, allowing a theoretical computer to monitor your individual stress and adjust the ascent accordingly.

Managing individual susceptibility to a fluctuating range of variables is complicated, especially when many of these variables remain undiscovered, or at least poorly understood. Clearly, all is not lost, as we do a very good job managing the problem of decompression sickness. Depending upon our measure of success, we could say this problem is effectively solved. The fact that we are arguing about the nuances of decompression-stop arrangement and obsessing about relatively small adjustments to our total decompression time speaks to this success. We are likely refining along the margins beyond the point of diminishing returns. However, we should not fool ourselves into thinking that we have all the answers. 

It’s The Data, Stupid

Another way to look at the science of decompression is to say it has mostly been a data-gathering exercise around which we fit slowly evolving boundary conditions. The boundary conditions are prescribed by algorithms and work quite well as long as we stay roughly within their range. It is quite possible we are not capturing any kind of truth about the way things work but rather refining our boundaries as we gather more data. It is true that we briefly foray into the field for some bubble-dynamics or that we strive to define the boundaries with process-markers like immune response, but none of these aspects has yet to produce a credible change in current practices. 

By far the most useful part of decompression research has been the accumulation of data and the refinement of algorithms that capture these outcomes. Ideally, these algorithms would extend well beyond the data they describe, supporting “safe” diving profiles where sparse or even no data exists. Yet, evidence suggests that our models are especially bad in these outlier territories including very deep and/or very long dives. Most divers with meaningful experience in the 100+ meter range will admit they have little assurance of a clean ascent absent any symptoms of decompression sickness. These aspects further suggest that we are working in the proverbial dark, or at least just barely within the distant illumination of modern knowledge.  This appears true at least with respect to specific determinations of cause (mistakes made) and effect (DCS incidence). Attempts to manage this uncertainty are in process among researchers spanning the globe. 

Most experts are convinced that bubbles play a role in developing symptoms of decompression sickness, and most of these believe the effect is significant. In this regard, we have perhaps not come so far from Haldane or Buhlmann, who were both well aware of bubbles but lacked the tools to manage their development throughout a diver’s ascent. Likewise, the most recent deep stop studies do not propose that bubbles are irrelevant, only that deep stops appear inefficient and, in at least some cases, can increase risk. On the other side, we have evidence that slower ascents and/or deeper stops can reduce bubbling, but we remain unclear about the degree of importance the bubbling itself represents, especially over the long ascents conducted by technical divers. 

Even a perfect model of bubbles might fail to predict or appreciably reduce decompression sickness, given the many complications in asserting the specific effect of bubbles in a given individual or within a particular injury. We are probably far from a perfect bubble model and perhaps even farther from determining how the wide array of variables might impact different individuals over time. 

Perhaps we can find a way to manage our uncertainty while still progressing our understanding of the likelihood of a given outcome. For good reason, this process is reminiscent of mysteries coming to light in other fields. We seem to be discovering that more knowledge in a given area does not always result in a clearer understanding. Less than 50 years ago, most people were convinced we had “solved” the mystery of elementary particles, bundling the atom in nice packages of three constituents with simple-sounding names. Now the more we learn, the better we measure, the deeper we look, the more unsettling is the complexity. 

Probabilistic Models and Uncertainty

Despite the confusing world around us, we have managed to achieve a high degree of success, and this continues despite our uncertainty. Management of this uncertainty can be mitigated by the use of probabilistic models and is currently common in other disciplines. This is an interesting and promising field, though it seems unlikely probabilistic decompression models will greatly change our current decompression profiles. This assumption may be wrong but seems appropriate, partly because we already have very low levels of decompression sickness, and partly because we have many supporting dives validating current time/depth profiles. 

Jarrod Jablonski towing decompression bottles at the surface during a GUE project dive. Photo courtesy of the GUE archives.

Adjustments like deep stops temporarily promised to reduce decompression time, perhaps by as much as one-third, but failed to materialize when tested more rigorously. This seems likely to remain true, at least as long as we assert a primary objective in maintaining very low DCS risk for the overwhelming majority. There may be a variety of small improvements to be found, but our current approach seems broadly “correct,” at least within the bounds of most active diving profiles. 

In some ways, we already manage uncertainty but do so indirectly by assigning a very low level of acceptable risk to the profiles that we test. This ultimately impacts the resulting decompression schedule. Using probabilistic models might allow us to permit a high level of risk, which could conceivably shorten decompression time. However, it remains to be seen if these models will be released in a way that allows users to accept high levels of risk. Even if such options become available, I wonder how many divers would use them in an aggressive way. Regardless of these factors, probabilistic models might allow a rational selection of risk, especially for those with the requisite understanding. 

Current and foreseeable models may not be describing any sort of truth, but they do appear good at determining useful boundaries (time and depth limitations) around which a desired outcome (limited DCS risk) appears most likely. I do not mean to belittle that success in the least. We maintain a high degree of confidence we will not suffer decompression sickness on most dives, and that is no small achievement. Yet, it also brings us full circle and back to the idea that modern-day decompression tables are largely determined by those most susceptible to decompression sickness. 

The NEDU study was stopped when it reached a threshold relating to DCS outcome. In this case, 10 of 198 dives resulted in DCS symptoms. Most were mild, late onset, Type I, but with two cases of rapidly progressing CNS manifestations. Two of the DCS cases were experienced by one individual. Ethical considerations require that a manned diving trial with DCS as an end point be designed to limit unnecessary injury to divers by maintaining a low level of DCS risk. This is a sensible and inevitable outcome of human trials. 

I am not advocating for a change to this strategy, but I am curious how this process affects our understanding of DCS, since we know little about the reactions occurring in more than 90 percent of test subjects. Would these individuals begin experiencing low-level symptoms after longer exposures? How much longer? Would we suddenly start seeing dangerous Type 2 symptoms in a rapidly escalating percentage of individuals? This rapidly escalating risk seems likely based upon experience with provocative profiles, but the details remain poorly defined. 

Team of divers descending into the cave. Photo courtesy of the GUE archives.

Maybe some individuals are more resistant to bubble formation while they or others are less sensitive to the bubbles that form. We can find many cases of prolific bubbling absent DCS symptoms. Meanwhile, DCS symptoms can be present with no detectable bubbles. This is to be expected, as symptoms are at least partly related to where bubbles are located. But these results might also hint at other differences in our response to bubbling. What if some divers form bubbles easily and/or experience high susceptibility to any formed bubbles? How would that knowledge affect any decompression recommendations? Is it conceivable that what works well for one diver or even the majority of divers is not optimal for all divers? 

All of this ambiguity should lead a thinking person to question the certainty of their pronouncements. We might be inclined to reduce our deep gradient and ascend more quickly from depth,  as the developing evidence indicates. But we should also respect the dive buddy that says they get bent when moving quicker in deep water. We can’t definitively say what works best, but we can say what seems to work well in the majority of cases for the majority of people. For most divers, these debates are largely academic, since the differences in profiles amount to minutes in one direction or another. 

Technical divers are progressively more affected by changes in recommended ascent profiles in relation to the length of their dives. Yet, even tech dives of relatively modest lengths show impacts of less than 10 minutes and are usually not worth nearly as much anxiety as one can find in the community. Having said this, it is easy to appreciate the desire to maximize efficiency. I am merely trying to suggest one should not be in a big hurry to change what seemed successful in the past. Those wishing to balance experience with evolving science might begin to raise their deeper gradients in a progressive fashion over time while paying attention to how they and their dive buddies respond. Or a person that perceives success with their current approach might choose to hold tight and make few, if any changes. I am arguing that we should recognize both opinions have merit and that we should take each perspective into account when working within our team to establish a given ascent schedule. 

The one definitive thing we can say about decompression is that it works well in the vast majority of cases, and when it doesn’t work, we probably will not know the exact reason. That reality is unlikely to change in the foreseeable future, although we certainly need to keep trying. A knowledgeable friend of mine once said that if we get bent, it is because we did not do enough decompression. Truer words have never been spoken. 

Personal Note:

I am very curious to hear about your experiences and opinions regarding evolving decompression science. Are most of you convinced that deep stops bring no value? How many think they are dangerous? Do you think I make too much of individual susceptibility, or do you see that in your own experiences? I welcome all points of view, critical and otherwise. Let the games begin :-).


Jarrod is an avid explorer, researcher, author, and instructor who teaches and dives in oceans and caves around the world. Trained as a geologist, Jarrod is the founder and president of GUE and CEO of Halcyon and Extreme Exposure while remaining active in conservation, exploration, and filming projects worldwide. His explorations regularly place him in the most remote locations in the world, including numerous world record cave dives with total immersions near 30 hours. Jarrod is also an author with dozens of publications, including three books.

Continue Reading

In this video, GUE President Jarrod Jablonski and Technical Administrator Richard Lundgren discuss some of the early hurdles associated with decompressing from deep tech and cave dives. This discussion explores the lack of readily available deep-diving tables and most especially those capable of managing multi-level profiles. They also explore the early development of ratio deco and similar “experiments” conducted by early technical divers.

What is Ratio Deco?

The basic idea of ratio deco involves establishing a ratio between the time spent at a given depth and the associated decompression. This is possible because the curve describing the relationship between bottom time and total stop time (TST) at a single depth can be approximated by a straight line. The straight-line approximation breaks down beyond a certain range but can be useful toward estimating decompression time. For example, the ratio in a typical tech dive at 45 m/ 150 ft using the appropriate gases is 1:1, meaning that a dive at 45 m/150 ft for 30 min will result in a decompression of 30 min while assuming appropriate bottom and decompression gasses.

We can also follow with adjustments for deeper dives so that each 3 m/10 ft deeper than planned would result in a decompression extension of five minutes. At some point the increase or decrease in the baseline parameter will break down, leading to profiles that are too conservative or too liberal and we need a new ratio such as the GUE standard for 75 m/250 ft at 2:1 where 30 minutes of bottom time results in 60 minutes of decompression, assuming appropriate decompression gasses arranged through a properly staged ascent.

What exactly are deep stops?

Deep stops are technically any stops added below what would otherwise be established by a typical dissolved gas model. Dissolved gas models establish faster ascents when compared to models that intend to reduce development of bubbles. Bubble-oriented models seek to reduce bubble development by ascending more slowly, usually by adding pauses or "deep stops".

“The formation of gas bubbles in the living body during or shortly
after decompression evidently depends on the fact that the partial
pressure of the gas or gases dissolved in the blood and tissues is in
excess of the external pressure. But it is a well-known fact that
liquids, and especially albuminous liquids such as blood, will hold gas
for long periods in a state of supersaturation, provided the super
saturation does not exceed a certain limit. In order to decompress
safely it is evidently necessary to prevent this limit being exceeded
before the end of decompression.”

Boycott AE, Damant GCC, Haldane JS. The prevention of compressed-air illness. J Hygiene (Lond ) 1908;8:342-443.

The extra decompression time calculated by various algorithms when breathing a helium mix is a consequence of the long held belief that helium, which is lighter than air, enjoys faster uptake by the body than nitrogen (in the case of the Buhlmann algorithm 2.65 times faster).

Compartments are hypothetical tissues which are intended to model how gas moves in and out of tissues where blood flow varies. Fast compartments are intended to model movement in a tissue with a lot of blood flow and slow compartments strive to model behavior in tissues with very little blood flow.

Historical records reveal the Greek philosopher Aristotle describing the use of a snorkel, relating the occurrence of ruptured eardrums, and outlining the use of the first diving bell by Alexander the Great.

Summary by Mitchell SJ, Doolette DJ. Extreme scuba diving medicine.

“The few studies available at the time of adoption of deep stops by technical divers [53,55] have been interpreted to support this notion. The earliest of these papers, an observational study of the practices of pearl divers in the Torres Strait of Australia [53], often cited as unqualified support for deep stops, is difficult to obtain and worth summarizing here. These pearl divers performed air dives to depths up to 80 msw followed by empirically-derived decompression schedules that had deeper stops and were somewhat shorter than accepted navy decompression schedules. Thirteen depth/time recordings were made of such dives, and these dives resulted in 6 cases of DCS (46% incidence). The remaining data was a count of dives performed from four fishing vessels over a two month period and these 468 man-dives resulted in 31 reported cases of DCS (7% incidence). It takes a certain cognitive dissonance to interpret these high incidences of DCS as supporting a deep stops approach.”

In: Feletti F, editor. Extreme sports medicine. Basel: Springer International Publishing; 2016. p. 313-33.

Note: Data regarding the 468 man dives was collected by interviewing Japanese surface tenders and looking at their (Japanese) logs, relying mainly on their memories for what the decompression profiles were, and how many DCS cases occurred.

LeMessurier DH, Hills BA.
Decompression sickness: a thermodynamic approach arising from a study of Torres Strait diving techniques.
Hvalradet Skrifter 1965;48:84.

Heart rate variability is the physiological phenomenon of variation in the time interval between heartbeats. It is measured by the variation in the beat-to-beat interval. Other terms used include: "cycle length variability," "RR variability," and "heart period variability".

https://www.health.harvard.edu/blog/heart-rate-variability-new-way-track-well-2017112212789

Statistics includes the process of finding out about patterns in the real-world using data such as the incidence of injury for a given time at a given depth.

When solving statistical problems, it is often helpful to make models of real world situations based on observations of data, on assumptions about the context, and on theoretical probability. The model can then be used to make predictions, test assumptions, and solve problems.

A deterministic model does not include elements of randomness. Every time you run the model with the same initial conditions you will get the same results. Most simple mathematical models of everyday situations are deterministic. For example, calculating the return on a loan with a given interest rate over x number of years. Simple statistical statements, which do not mention or consider variation, could be viewed as deterministic models.

A probabilistic model includes elements of randomness. Every time you run the model, you are likely to get different results, even with the same initial conditions. A probabilistic model is one which incorporates some aspect of random variation. Deterministic models and probabilistic models for the same situation can give very different results.

Probabilistic decompression models are designed to calculate the risk (or probability) of decompression sickness (DCS) occurring on a given decompression profile. These models can vary the decompression stop depths and times to arrive at a final decompression schedule that assumes a specified probability of DCS occurring. The model does this while minimizing the total decompression time.

Probabilistic models allow selection of risk in ways that support rational choices. As with most tools, this power can also be used irrationally though the tool should not to be blamed for such abuse. For example, one might do three dives a day for five days, each with an established one percent risk. The risk of at least one DCS in that series is 1-(1-0.01)^15=0.14.

Alternatively, a diver might select one big dive with all the risk captured in that one dive, i.e. 14%, benefiting from the associated faster decompression. The extent to which modelers might allow users to make those choices remains an open question. Meanwhile, the consequences of accepting higher or even very high risk remain largely unknown. For example, to what extent would a 20% risk of DCS make me vulnerable to serious forms of DCS sickness? Should I be allowed to take those chances if I choose. If so, what sort of disclaimer is needed and/or what should be required of me to ensure I understand the risk I am taking?

More on Probabilistic Models:
https://en.wikipedia.org/wiki/Decompression_theory#Probabilistic_models