Gradient Factors & Dive Computers

Taking Control of Your Decompression

Gradient factors affect every diver using a modern dive computer — including recreational divers doing 18–30 m dives on air. Understanding what they are, and how your computer uses them, helps you make better ascent decisions and customise your computer to match your physiological risk profile.

How Dive Computers Track Decompression

Most technical dive computers use the Bühlmann ZH-L16C model — a dissolved-gas model that divides the body into 16 tissue compartments with nitrogen half-times ranging from 4 to 635 minutes, calculates how much nitrogen each compartment is holding at any moment, and uses M-values (maximum tolerable inert gas pressure) to define the ceiling — the shallowest depth you can safely ascend to at any point in the dive.

M-values were formalised by Robert Workman (USN NEDU, 1965). Each compartment has its own M-value that increases linearly with depth — at higher ambient pressure, tissues tolerate more supersaturation before bubble formation becomes likely.

What Are Gradient Factors?

Gradient factors (GF) are a user-configurable conservatism dial applied on top of ZHL-16C. They treat M-values as 100% of the theoretical limit and allow you to set your own working margin below that limit. They are written as two numbers: GF Low / GF High.

GF Low controls how deep your first decompression stop is. GF High controls how close to the M-value you are permitted to reach at the surface. A GF of 0.85 means the computer only allows 85% of the theoretical M-value to be used — the ceiling appears deeper than unmodified Bühlmann would require. Lower numbers are more conservative.

The GF Interpolation Formula

Between the deepest stop and the surface, the effective GF changes linearly. Baker's original formulation:

GF_at_current_depth =
  GF_Lo + (GF_Hi − GF_Lo) × (depth_first_stop − current_depth) / depth_first_stop

At the deepest stop, the effective GF equals GF_Lo. At the surface, it equals GF_Hi. Every intermediate stop uses a proportionally interpolated value, creating a smooth ramp of progressively relaxing conservatism as the diver ascends.

Example: GF 30/85, first stop at 21 m, diver currently at 9 m.

GF = 0.30 + (0.85 − 0.30) × (21 − 9) / 21
   = 0.30 + 0.55 × 0.571
   ≈ 0.61

At 9 m on this profile, the computer is enforcing 61% of the M-value — more permissive than the initial stop, but still well below the 85% surface limit. This interpolation is what produces the characteristic stepped decompression schedule that gets progressively more efficient as you ascend.

The Deep Stops Controversy

For years, bubble models like VPM and RGBM pushed the idea of deep stops — beginning decompression much deeper, at pressures high enough to keep bubble nuclei compressed. This seemed mechanistically sound.

The NEDU deep stop study (2011) tested this directly and found the opposite: redistributing stop time from shallow to deep increased DCS incidence. Slow tissue compartments — those with half-times of 100–600 minutes — continue loading gas throughout any stop spent at depth. A deep stop helps fast compartments off-gas while simultaneously loading slow compartments further. Since slow compartments have low surface M-values and require long shallow stops to clear, this trades a manageable shallow stop obligation for a harder-to-clear deep one.

The 2008 Decompression and Deep Stop Workshop concluded that available evidence does not support deep-stop schedules over conventional shallow-stop decompression.

Current practice among experienced technical divers has shifted accordingly: GF_Lo of 30–40 is now considered more evidence-based than very deep settings like GF_Lo 20. The goal of GF_Lo is not to add a maximally deep first stop — it is to initiate decompression somewhat before the strict M-value ceiling while keeping the bulk of off-gassing time in the shallow zone where it is most efficient.

What the Belgian Military Study Found

When the Belgian Defence replaced their Cochran EMC-20H computers with Shearwater Perdix units in 2018, they used the default GF settings of 30/70. The result was decompression times substantially longer than those produced by DCIEM and USN air tables — tables with established safety records.

De Ridder et al. (2023) analysed why. Lower GF_Lo values place the first stop deeper, which loads slower tissue compartments more. For air dives in the 23–60 m range, this can actually increase DCS risk by building slow-tissue obligation without proportional benefit to fast compartments.

Their analysis found the optimal match to established safe tables was achieved with GF_Lo 100, GF_Hi 75–100 — no deep stops, conservatism applied only at the surface. The most conservative practically achievable compromise within Perdix software constraints was 75/75 to 95/95. The Belgian Navy was advised against using the default 30/70 settings for air dives.

This does not mean GF 30/85 is wrong for all diving. It means the assumption that lower GF_Lo is always safer is incorrect. For trimix dives where slow tissue loading is less of a concern (helium off-gasses faster), deeper initial stops may behave differently. No single GF setting has been validated as universally optimal across all gas mixes and depth profiles.

Ceiling-Controlled vs Staged Decompression

Angelini, Tonetto, and Lang (2022) compared two approaches to executing ZH-L16C decompression: staged decompression, where the diver waits at fixed 3 m depth increments as nitrogen washes out; and ceiling-controlled decompression, where the diver follows the ceiling continuously upward, maintaining the maximum available pressure gradient throughout.

Ceiling-controlled decompression was 4–12% shorter in total decompression time. The tradeoff is that fast tissue compartments experience slightly higher supersaturation during ascent. The practical implication: modern dive computers with real-time ceiling tracking implement something between these extremes. The 3 m stop increment that most computers use today traces back to Haldane's 1908 practice — chosen partly because of the precision of available pressure gauges, not because 3 m is physically special.

Evidence-Based GF Settings

No large-scale randomised trial has established a single optimal GF setting. What the evidence supports:

For recreational divers whose computers use preset profiles (Conservative / Normal / Aggressive), these typically correspond to approximately GF 20/70 through 40/85 without exposing the numbers. If your computer lets you set GF manually, start with 40/85 and do not go lower on GF_Lo without a specific reason based on your gas mix and depth profile.

Oxygen During Decompression

Risberg and Midtgaard (2024) reported on a Norwegian commercial diving project where 11 divers completed 91 dives (23.5–36.2 m) with planned surface oxygen decompression — but breathed air instead of oxygen due to a technical error. Only 2 divers developed clinical DCS (2.2%), though 5 of the remaining 9 later reported retrospective mild symptoms.

The study reinforces two points relevant to GF selection. First, moderate omission of decompression has a smaller DCS impact than probabilistic models might suggest — the relationship is not binary. Second, oxygen dramatically accelerates nitrogen washout during decompression: replacing in-water nitrogen-breathing stops with surface oxygen decompression substantially reduces total nitrogen elimination time. For divers using nitrox or oxygen as a deco gas, the effective decompression efficiency is considerably better than the ZHL-16C nitrogen-breathing model alone would predict.

GF99 and Surface GF

Modern Shearwater computers display two real-time values that make GF practical to use operationally.

GF99 is the current leading-tissue supersaturation as a percentage of its M-value at your present depth. If GF99 = 92, the leading tissue is at 92% of its M-value right now. If GF99 exceeds 100%, the M-value is exceeded — hold depth or descend slightly.

Surface GF answers the question: if I surfaced right now, skipping all remaining stops, what would GF99 be at the surface? Hold the stop until Surface GF drops to your GF_Hi setting.

These two numbers transform gradient factors from an abstract planning parameter into a live operational tool.

Practical Guidance for Recreational Divers

Even on no-deco dives, how you ascend matters. Regardless of your computer's GF settings: ascend at 9 m/min or slower in the shallows — many divers ascend too fast in the last 10 m. Do your full safety stop — 3–5 minutes at 5 m is not optional. After dives deeper than 18–20 m, a staged ascent (1 minute at 9 m, 2 minutes at 6 m, 3 minutes at 3 m) adds meaningful off-gassing time. Avoid yo-yo profiles; each depth oscillation reloads fast compartments. Be more conservative after long dives, cold dives, exertion, or multiple dive days — individual physiology (dehydration, age, PFO, fatigue) is not captured by the algorithm.

References

• Angelini SA, Tonetto L, Lang MA. Ceiling-controlled versus staged decompression: comparison between decompression duration and tissue tensions. Diving Hyperb Med. 2022;52(1):7–15. doi:10.28920/dhm52.1.7-15
• Baker EC. Clearing Up the Confusion about "Deep Stops". 2010. (Gradient Factor papers, published online)
• De Ridder S, Pattyn N, Neyt X, Germonpré P. Selecting optimal air diving gradient factors for Belgian military divers: more conservative settings are not necessarily safer. Diving Hyperb Med. 2023;53(3):251–258. doi:10.28920/dhm53.3.251-258
• Doolette DJ, Mitchell SJ. Recreational technical diving part 2: decompression from deep air dives. Diving Hyperb Med. 2013;43(4):215–220
• Fock AW. Analysis of recreational closed-circuit rebreather deaths 1998–2010. Diving Hyperb Med. 2013.
• Bühlmann AA. Decompression–Decompression Sickness. Springer, 1984 (updated 1992).
• Risberg J, Midtgaard H. Decompression sickness in surface decompression breathing air instead of oxygen. Diving Hyperb Med. 2024;54(3):242–248. doi:10.28920/dhm54.3.242-248
• NEDU TR 11-06. An Evaluation of Decompression Algorithms for Use in a Fleet Dive Computer. US Navy Experimental Diving Unit, 2011.

Train With Me

Gradient factors and decompression strategy are a core part of every technical course I teach. Enquire about training →