Preface

The theoretical groundwork for skydiving altitude indication described in this document does not arise from a purely analytical perspective, where the theory was first developed and then devices were built around it. Rather, it is the formalized result of over a decade of experience building a diverse array of skydiving altimeters. The goal is to aid a more complete understanding of a subject that has seen almost no formal research: to establish a general taxonomy and initial understanding with which the subject can be more precisely analyzed and discussed. Consideration of these ideas can promote good practices related to altitude awareness and can lead to better altitude indication design.

Document structure

This document proceeds in five parts as follows. First, we describe the problem — what altitude awareness is and how it fails (Part I). We then build a theoretical framework for altitude indication consisting of a taxonomy of the indications themselves (Part II), a description of the sensory modalities through which those indications are conveyed and the jumper who receives them (Part III), and an imported body of theory — cockpit alerting (Part IV), the closest existing formal vocabulary. Finally, we derive design principles from the preceding Parts and make recommendations concerning altitude indication design (Part V). An afterword closes with the author's own advice as a skydiver. Five appendices follow: barometric altimeter error (A), the reasoning behind the tolerance envelope in Figure 5 (B), quantifying the indication space (C), studies and measurements (D), and the revision history (E). In the web edition, several figures are interactive (marked ▶︎ interactive; e.g., Figure 13 and Figure 14), allowing the reader to hear and see the concepts described in the text.

Part I — The Nature of Altitude Awareness

1.1 Scope and motivation

This document covers altitude awareness and how altitude information reaches the jumper. How altimeters measure altitude only appears in this theory where it affects awareness: their errors set a floor on how accurate the jumper's estimate can be (§1.6, with the barometric errors in Appendix A); for a full introduction, see "A skydiver's guide to how barometric altimeters work" and "Altitude types: pressure, indicated, true, density."

Maintaining altitude awareness in skydiving is paramount to a jumper's safety. The USPA's Skydiver's Information Manual (SIM) tells students on their first jump "altitude awareness is your most important task" (SIM, ISP Category A). But the importance of altitude awareness does not end after initial training: without it — regardless of what the intentions are on a skydive — jumpers can easily find themselves in dangerous and potentially fatal situations; all other considerations on a jump are secondary to a jumper's safety.

We begin with a set of primitive definitions; further terms are defined as they arise, building on these, and all defined terms are collected in the Glossary at the end.

1.2 Primitive definitions

• Skydive — according to the glossary in the 2026 SIM (significantly rewritten in 2025): "the descent of a person to the surface from an aircraft in flight when he or she uses or intends to use a parachute during all or part of that descent" (SIM, Glossary). This document extends the definition to include ascent, since, for the purposes of altitude indication, it matters just as much as it does for freefall and canopy.
• Altimeter — a device designed to measure altitude and indicate it to a jumper; one altimeter can emit indications across multiple sensory modalities.
• Indication — any stimulus delivered to a jumper by design that carries altitude information. Ambient cues (visual ground reference, light gradient, temperature) and the internal clock (§3.4) are not indications: an indication is sent — by an altimeter, or by another jumper signaling on purpose (a wave-off, an instructor's pull signal); an ambient cue is found by the jumper.
• Modality — the sense carrying the indication (auditory, visual, tactile).
• Parameter — a perceptual dimension within a modality along which an indication can vary to carry information (auditory: pitch, loudness, tempo, spatial position; visual: brightness, hue, position). An indication maps altitude onto one or more parameters (cf. parameter-mapping sonification, §2.4).
• Audible altimeter — an altimeter whose indications use the auditory modality.
• Stage — one of the three parts of the skydive: ascent, freefall, and canopy.
• Threshold and phase — two altitude-related elements of a skydive; the fundamental altitude components that any skydive can be broken down into. A threshold is an altitude-pinned point at which an action must begin (breakoff, deployment, decision altitude); a phase is an altitude-extended activity between thresholds (working time, tracking, the downwind leg, the base leg).
• Cutaway — releasing a malfunctioning main canopy, usually by pulling a handle on the harness, before deploying the reserve; also called a breakaway.
• Decision altitude — the altitude at which a decision must have been made whether to cut away or keep a main canopy.
• Hard deck — the lowest altitude at which one can safely cut away a main canopy with adequate altitude for inflation of the reserve. Often confused with decision altitude.
• Reserve-only altitude — in freefall or in an emergency exit, the altitude below which it is only safe to pull the reserve, as one would otherwise risk an automatic activation device (AAD) fire and potential main-reserve entanglement; may be different on ascent than descent.
• Altitude awareness — a jumper's internal estimate of current altitude, whether numeric or not, together with its uncertainty, error, and salience; graded, not binary (§1.5). A lapse is the growth of that error, allowing for its uncertainty, beyond current tolerance: the jumper can no longer be sure they are within it.
• Altitude misjudgment — in contrast to a lapse, an error in directly perceived altitude at a more primitive cognitive level where no represented value is in play at all. Non-human animals also demonstrate such misjudgments due to various factors like sensory decline, poor lighting, etc. The two are not exclusive: in the landing pattern a jumper has a represented estimate all the way to the ground as direct perception sharpens — both errors can be present at once.
• Altitude picture — what a given altitude looks like. Swoopers use the related and to-an-extent overlapping term sight picture to refer to how a landing visually appears — theirs is a more complex definition and includes, for example, the way a landing turn appears.

1.3 The altitude structure of a skydive

Every skydive has an implicit altitude structure — its own profile of thresholds and phases across the three stages. Altitude structure shapes the way a jumper conceives of a skydive; planned in advance, it also serves as a conceptual map of what happens where — or at least what is supposed to happen. Jumpers "plan the dive and dive the plan," but emergencies can force changes to a skydive's intended altitude structure. Figure 1 shows the altitude structure of a hypothetical skydive by a typical fun jumper. As one might expect, thresholds become more frequent as one approaches the ground.

1.4 Lapses in altitude awareness

Altitude awareness lapses lie on a spectrum from mild to severe: from low pulls to AAD activations to fatalities. We can establish a working scale ordered by severity: how far a lapse goes before recovery, and then what it costs. We define six grades of severity in altitude awareness lapses, investigating their characteristics and relations. What specifically the scale refers to differs at different depths. Grade 3 is qualitatively different from the rest in that it involves a voluntary decision of whether or not to include a life-saving device — a jumper can take this grade out of the equation, escalating any lapse that passes grade 2 to grades 4–5. Those last two grades — the most severe — are different from the first four in that they are physical outcomes of a lapse, reachable from grade 2 on.

• Grade 0 — benign uncertainty. Some uncertainty about altitude coupled with inaccuracy of estimation is inevitable — the default state of a skydive, with no negative consequences; arguably not a violation at all, hence grade 0.
• Grade 1 — tolerance exceeded, no consequence. The jumper's error, allowing for its uncertainty, exceeds tolerance, but recovery occurs without consequence; this is most visible in missed thresholds (e.g., a student does not lock on at 6,000 ft as planned but still pulls at the planned altitude) but can also simply occur when a jumper's estimate is significantly off.
• Grade 2 — late action. Awareness recovered past tolerance; the jumper acts, but late and hence lower (e.g., a late breakoff or low pull).
  • Note — injuries and fatalities begin here. Every penetration past tolerance has a cost. A late breakoff on a big-way can result in being closer to other jumpers than desired, risking a collision. A low pull can put a jumper in a bad location and takes time away from other safety-critical tasks such as a canopy control check, landing area selection, pattern setup, and obstacle avoidance. In such a way, a jumper who acts late may open safely and still wind up being injured or killed by an unintended landing in a dangerous area.
• Grade 3 — automation rescue. AAD save: for whatever reason the jumper never recovered awareness; the AAD fired and actually saved the jumper's life. Presupposes a functional AAD (SIM 4-3.G; Airtec 2025).
  • Grade 3, special case — concurrent activation (the race zone). The jumper pulls late and the AAD fires while the main is opening: the lapse reached the automation envelope (hence is escalated to grade 3), but both the human and the automation layers acted. The AAD's inference is correct on its information — a main pulled late, still in its snivel, has not yet adequately slowed the jumper, so the device cannot distinguish late recovery from no recovery. The outcome in this case can be worse than a standard AAD fire, carrying entanglement risks that can escalate the event to grade 4–5. Consequence for altitude error tolerance: deployment tolerance is bounded by AAD firing altitude plus opening distance plus an additional safety margin — i.e., the main must have slowed the skydiver's descent enough not to trigger the AAD at its activation altitude.
• Grade 4 — injury to which a lapse was a significant contributing cause.
• Grade 5 — fatality to which a lapse was a significant contributing cause.
  • Note — contributing, not sole, causes. Injuries and fatalities rarely have a single cause: a lapse is usually one link in a chain that can also include equipment, conditions, other jumpers and decisions made long before the jump; risk stacking increases the likelihood of an escalation of a lapse to grade 4 or 5. Just as in accident investigation, which names contributing factors alongside a probable cause (ICAO Annex 13; e.g., NTSB 2010), grades 4 and 5 ask only whether the lapse was among them. That question is harder to answer when it is no longer possible to get the jumper's perspective on what went wrong.

What the grades refer to changes with the scale as noted above: grades 0–2 refer to the jumper's state (an uncertainty, an exceeded tolerance, a late action); grade 3 refers to a system response (an AAD save); and grades 4–5 name a physical outcome (injury, then fatality). The scale thus evaluates three different things in sequence, in order of their severity: the lapse itself, what intervenes (assuming an AAD is present and functional), and the consequences.

Loss of altitude awareness remains the dominant factor in AAD saves across the board, according to activation data presented by Karl Meyer of Advanced Aerospace Designs, which makes the Vigil AAD. Low or delayed cutaways appear frequently and, among students, overload and task fixation are cited as major contributors (Meyer 2026).

From 1983 through 1987 — a time when almost no one used an AAD and pull altitudes were lower (a double whammy) — no- or low-pull scenarios accounted for 40% of US skydiving deaths; this would be considered an astonishing figure by 2026 standards. From 2017 through 2024, in contrast, USPA recorded no accidental no-pull deaths (suicides are counted separately), and only two jumpers died as a result of a low deployment. The streak ended in 2025, when three jumpers died in separate accidents without a parachute deploying. USPA attributes the change to AADs and higher deployment altitudes, and notes that many no- or low-pull cases that once would have been fatal "now have survivors who can tell us what happened. Common causes are loss of altitude awareness because of distraction or lack of ground references" (USPA Fatality Summaries 2017, 2024, 2025). Lapses now end mostly at grade 3 rather than at grades 4–5. Where loss of altitude awareness still appears in fatal accidents, it is mostly under a malfunctioning canopy, below decision altitude (§1.9) — a fact that deserves wider reporting than it gets.

As one would expect, the observability of a grade in practice runs more-or-less opposite to its severity, so the most benign lapses most often go unnoticed or unrecorded: epidemiology often records grade 4 and is obligated to record grade 5; AAD manufacturers and the USPA (for students and tandems) log grade 3; grades 0–2 — the overwhelming majority of all lapses — are for the most part unrecorded.

Group dynamics and human psychology can lead to multiple jumpers on the same jump simultaneously experiencing an altitude awareness lapse. A group lapse is a lapse shared by two or more jumpers in which their lapses are linked — either by a common cause, or by one jumper's lapse causing or sustaining another's. Individual lapses that merely coincide are not group lapses; group lapses are correlated. At least three mechanisms underlie them. Diffusion of responsibility: each jumper assumes someone else is watching the altitude — the bystander effect applied to monitoring (Darley & Latané 1968). Shared capture: the jumpers are each other's distraction, so a single stimulus captures everyone at once, and the group's momentum — the shared eagerness to finish one more point — reinforces it like a feedback loop. Cue substitution: the group breaks off on a signal — as sometimes happens in an angle track — rather than on each jumper's own indication, so when whoever gives the signal lapses, everyone does. All three of the following documented cases are two-ways, perhaps in part because the more jumpers are on a jump, the more likely it is that at least one of them is going to be altitude aware. Often all it takes is one jumper becoming altitude aware and beginning to track away to re-trigger the sampling cycle (§1.7) in other jumpers: Skydive Mag reports two jumpers from Argentina whose AADs both fired because of "the joy of having finally found each other in freefall" (Hübner 2023); a confidential incident report describes two freeflyers who lost altitude awareness together and deployed their mains, only for both AADs to fire at the same moment, around 700 ft (Hewitt n.d.) — a shared case of concurrent activation (grade 3, above); and two Belgian freeflyers in 2015 carved head-down together through their audibles' beeps until both CYPRES units fired about a second apart — "We were just too comfortable with our audibles, carving around and waiting for the beeps to come," one of them said afterward, and the other blamed a summer of big-group jumping, in which breakoff is cued by the group (De Clercq and Hoorens, in CYPRES Team 2020).

1.5 A model of altitude awareness

Altitude awareness is not a binary state of being either aware or unaware. Rather, it is an internal estimate with a continuously varying degree of uncertainty, error, and salience: one is uncertain about altitude to a degree; one's actual estimate also has a separate degree of error; and it is salient in one's mind to varying degrees throughout a jump.

A lapse in altitude awareness is the growth of the error, allowing for its uncertainty, beyond current tolerance — determined chiefly by the next threshold, but also by those after it, and by the demands of the current phase (§1.2). Error tolerance is coupled to time, as the faster you descend, the less time altitude affords you; this is critical because thresholds require actions that take some minimum amount of time to complete regardless of descent rate. Indications reset uncertainty and narrow the error, but must first be registered: an altimeter that is not seen or heard is of no use. Between such resets the estimate decays and salience wanes.

Altitude error tolerance is the altitude estimate error the next altitude-pinned action, and those after it, can absorb while the outcome remains acceptable: safe, also within what the other jumpers on the load will accept (jumper etiquette), and what the jumper's own altitude discipline (below) demands (e.g., error at breakoff is compensated for by adequate separation and a pull that is not considered unusually low). At a single threshold it can be written as ε_max = b − L − Δ. Here b is the threshold — 3,500 ft at deployment — the altitude at which the action is planned to begin. L is the completion limit, the lowest altitude the action may finish by: the container must be open by 2,500; it does not move. Δ is the completion distance assumed for the action: what the reaction and the action itself are expected to consume between the two. What remains of the altitude margin b − L after Δ is the tolerance (Δ is explained below; the three terms are drawn in Figure 4). No one pulls at exactly 3,500, and that is specifically what tolerance is about: the estimate is never exact and the timing never perfect, and the tolerance is what makes that survivable. If the jumper acts when the estimate reads b but the estimate is high by ε, the action actually begins at b − ε and must still finish by L, so ε can be at most b − L − Δ (at deployment: believing 3,500 while actually at 3,400, the pull begins at 3,400, consumes 500, and the container is open at 2,900 — 400 ft above the 2,500-ft limit; an error of 500 ft would use that margin up entirely). A reaction slower than the one Δ allows spends the same margin, even on a correct estimate. Altitude error tolerance can be visualized as a tolerance envelope around a jumper's actual altitude — as in Figure 5 — and has the following characteristics:

• A function of altitude. The tolerance profile T(A) is unique to each skydiver on each skydive and represents permissible error as a function of altitude at a given descent rate; drawn around the jumper's actual altitude, it forms the tolerance envelope. Generally speaking, T(A) narrows with a faster descent rate, because the act of responding (opening, or any timed action in freefall or under canopy) consumes more altitude the faster you are descending. A tolerance profile is therefore a two-dimensional slice of something with multiple parameters, altitude being the primary one: permissible error as a function of altitude alone, drawn as the envelope around actual altitude. The skydive sets the other parameters: its descent rate, seen as the slope in Figure 5, and the completion distance of each threshold, the altitude its action consumes, which sets how wide the envelope is at that point. Although T(A) depends on the type of jump — a wingsuit or high pull tolerance profile would certainly look quite different than a speed skydiver's — there is some convergence between most jumps around breakoff or pull time: for the most part, people track away or at least return to belly-to-earth position and pull — from there canopy flight is often discipline-independent and so T(A) is as well.
• What shapes it. Among the most significant factors that shape T(A) are: the altitude margins (b − L), the gap between each threshold and the limit its action must finish by, both altitudes coming from USPA minimums, drop-zone policy and the plan itself; descent rate, which turns every second of reaction into altitude and depends on discipline, stage, canopy and wing loading, or a malfunction; the task and the jumper's response, meaning what each action costs (a track, an opening, a turn) and how quickly the jumper reacts, which depends on experience and state; the available opportunities to check one's altitude before the next action, taking into account how much time there is and whether the eyes are free to look (almost none between breakoff and deployment, much more in the pattern); who else is in the air, meaning the size of the group, other groups on the load, the exit order and the traffic in the pattern; acceptability (community norms, drop-zone culture and the jumper's own discipline); and conditions such as light, visibility, wind and the landing area. The altitude margins, descent rate and task set the tolerance at the thresholds, with primary focus on the late side; the opportunities to check altitude shape the envelope between them; the early side is governed primarily by who else is in the air, and by the need to avoid people being where others don't expect them or where they are not allowed to be. Tactical operations change the profile's shape drastically (§5.9).
• Dynamic. T(A) changes in real time during the jump as its parameters change: for example, a high-speed malfunction changes the expected descent rate after deployment. This makes it impossible to draw in advance what T(A) will actually look like on a real jump; what can be drawn is a hypothetical jump that goes exactly as planned, with nothing unexpected coming up, which in practice never happens (e.g., pull altitudes are always approximate). The hypothetical profile in Figure 5 represents one such jump, following the plan of Figure 1: at each moment, its envelope takes into account what could still go wrong from that point on, not what has already been ruled out (once the canopy is open, for example, a bag lock can no longer occur and is no longer taken into consideration for decision altitude).
• Partly subjective. Acceptability is not safety alone: a deviation can be survivable and still unacceptable, because it puts a jumper where the others don't expect them, or deviates from the plan. What counts as an acceptable outcome is therefore in part a matter of opinion — especially for the higher thresholds, where more altitude remains to correct an error — and jumpers may disagree (example 1 below). T(A) therefore has an assessor: a jumper's own envelope for themselves and the envelope other jumpers would draw around them do not coincide; Figure 5 represents one perspective, the author's own.
• Asymmetric. It is always best to have zero error in an altitude estimate, but one direction is far more dangerous than the other. Believing oneself to be lower than one actually is leads to actions (e.g., pulling) coming early, which can even work in the jumper's favor — an early pull leaves more altitude to deal with a malfunction; believing oneself higher leads to late actions — clearly the more dangerous direction. As a consequence, the tolerance envelope generally extends further below a jumper's actual altitude than above it; an exception is the landing pattern, which is shared with other jumpers, making accuracy more important in order to avoid collisions — there both sides of the envelope are much closer to equal.
• Narrowing toward the ground. Although the overall trend is that T(A) narrows as the ground approaches, the envelope does not shrink uniformly; rather, it tightens into each threshold and changes once that threshold has passed, usually expanding again. This follows from the definition: tolerance is measured primarily against the next altitude-pinned action (later thresholds are taken into account as well, however, as breakoff's padding for the pull shows). As such, each completed action works almost like a reset of the envelope, moving the reference to the next action, focusing the mind further down, with more altitude left to check and correct before the next threshold approaches. Whether it relaxes depends in great part on those opportunities to check altitude in between thresholds: T(A) assumes that the jumper's normal sampling continues (§1.7), so it opens up where there is time to check before the next action and not at all where there is none, as between breakoff and the pull (though some licensed jumpers have the habit of checking their altimeter at the end of the track, just before pulling, adding to the reaction time, and so to the completion distance Δ); a lapse is what happens when the sampling fails and the error grows past the envelope. What happens to T(A) after each threshold is partly subjective, a matter of each jumper's judgment, but something always changes once the action is complete (for deployment, once the canopy is open), because the reference moves to the next action and the jumper's mind turns to it: usually the envelope opens up; between breakoff and the pull it stops narrowing and holds; after decision altitude it even tightens slightly as the reference shifts to entering the pattern more accurately to avoid traffic issues. Between thresholds the error must still leave room for every potential action ahead, and whichever of them leaves the least room for altitude error, given the descent rate and the time its action takes, sets the limit; Figure 5 draws each stretch as a straight line from the value just after one threshold to the value at the next, and treats decision altitude as an exception, for reasons given in Appendix B.
• Largely determined by margin. The tolerance differences between phases (larger values in the working portion of freefall, smaller ones in the landing pattern under canopy) are primarily attributable to altitude margin, not descent rate. The margin b − L can sometimes be more than a thousand feet in freefall but only a couple hundred at most in the landing pattern, so the same ε_max = b − L − Δ narrows under canopy because its first term is small. At deployment the descent rate drops by an order of magnitude, yet tolerance tightens rather than loosening. Descent rate varies proportionally more under canopy than in freefall, and turns increase it in the landing pattern exactly where tolerance is tightest. From phase to phase, the effect of descent rate on the narrowing of T(A) is secondary to the effect of the thresholds' altitude margins.
• Completion distance Δ is determined by task and descent rate. The altitude consumed by an action is understood to be: Δ = (the action's own altitude cost) + (current descent rate × reaction time). The first term indicates how much altitude the action itself consumes and is set by the task's own dynamics — a track's fall rate, a reserve's inflation characteristics, a canopy's turn — rather than by the descent rate at the moment of the threshold in question: a standard wave-off-and-pull has a small Δ, whereas proper implementation of emergency procedures has a larger one — so a pending emergency carries a tighter tolerance than a routine pull at the same altitude, regardless of descent rate. Tolerance therefore tightens both when you descend faster and when the pending action costs more altitude; the descent-rate term is only half of Δ.
• AAD not taken into account in determination. The automation layer widens survivability, not tolerance: skydivers are trained to act as if the AAD were not there (though training includes the AAD as a factor in two-out scenarios), remaining altitude aware and pulling at the correct altitude. No one is taught to purposely use the AAD as a backstop.
• Distinct from the demand for awareness. Tolerance is permissible error; it is not the same as the vigilance a phase demands. Every jumper is held to the same standard — a belly jumper falling at 120 mph and a freeflyer at 180 must both break off on time — but the freeflyer has less room for error, because the response consumes more altitude at higher speed (the descent-rate term of Δ, above). Speed also changes the required sampling rate: the estimate goes stale faster the faster you descend, so a quicker descent demands more frequent samples (§1.9).
• Threshold criticality. Missing certain thresholds costs more than others, but tolerance does not scale — at least not directly — with the stakes, for two reasons. The stakes are already considered when determining the threshold altitudes themselves, which are set with more margin where a lapse costs more; and best practice is to treat every threshold as vital, because it trains good altitude awareness and altitude discipline: the characteristic of maintaining planned thresholds tightly in practice (e.g., the tandem instructor who always pulls at 6,000 ft demonstrates this).

As the ground approaches, vision generally takes over from indications completely, for several reasons: the jumper is much more familiar with judging altitude at such low elevations; this is where accuracy matters most, and the eye resolves altitude best this close to the ground; and it is best not to depend on barometric altimeters for such a fine resolution of altitude, as ambient air pressure can drift between takeoff and landing enough that even the most accurate altimeters will display a different number on the ground in the same location at the start and at the end of the jump.

Altitude awareness tolerance is something that really exists, yet it remains almost entirely implicit and only rarely discussed: jumpers carry it as an instinctual sense developed through experience, not a computed set of numbers. One reason is its complexity, made apparent from the points above. Because it is learned from experience and rarely examined, a jumper's own T(A) can be poorly tuned: every jump that works out with a thin margin carries a subtle implication that the margin was enough, the "it always worked before" by which small deviations become normal — an individual analogue of what Vaughan (1996) called the normalization of deviance. It is only generally determined using an exact number where something external requires it — a student pull altitude, the AAD firing altitude, or final approach. Elsewhere it stays inexact, and for better or for worse: for most skydives there are no exact numbers given for T(A) at all. A companion interactive version of the profile in Figure 5 — in which a jumper's estimate can be followed through the jump, and one jump that does not go as planned, where the jumper has a bag lock, drawn in its place — is available on the FDS website: Altitude tolerance visualizer. Both the figure and the visualizer are deliberate simplifications: they are not intended to represent any real jump, but rather to make the structure of T(A) visible and, in the case of the visualizer, how it changes when something unexpected happens.

Two illustrative examples follow. The first demonstrates the subjective nature of T(A); the second applies the formula ε_max = b − L − Δ to a freefall deployment and then to what is generally the final threshold of a skydive, the turn to final approach.

1) Planned breakoff is 5,500. Breakoff occurs at 4,500 — a thousand feet lower — and everyone calls it a lapse; if the breakoff had occurred at 5,000 — five hundred lower than planned — it would be arguable, especially if the final planned point was turned. This would not be a huge deal on a four-way, but could present a real problem on a sixteen-way where separation was already tight, and different jumpers will have different opinions regardless. This shows how the tolerance boundary is a fuzzy zone, and not a line — defined by the specific jump and viewed differently by different jumpers. In this case, the threshold's tolerance is determined in part by group size: the SIM requires at least 1,500 ft between breakoff and the highest planned deployment for groups of five or fewer and 2,000 ft for six or more (SIM, ISP Category H) — and group size also increases the risks associated with any lapse, since every foot lost is shared with more jumpers who need clear airspace to deploy.

2) The formula applied to a freefall deployment and then to the landing pattern at a moderate wingloading. Take a main canopy deployment first: ε_max = b − L − Δ has b ≈ 3,500, L ≈ 2,500 (the minimum container opening altitude for B-, C- and D-license holders, SIM 2-1.I) and Δ ≈ 500 (about three seconds from the moment the estimate reads 3,500 to an open container, at tracking speed), so ε_max ≈ 3,500 − 2,500 − 500 = 500 ft. If you believe you are at 3,500 when you are actually at 2,800 (ε = 700, 200 ft past the budget), you open at 2,300 — 200 ft below the 2,500-ft limit L: the excess of ε over ε_max is exactly the distance by which the limit is missed. Now put the same formula on the turn to final (standard pattern: downwind ~1,000, base ~600, final ~300, flare ~10). On that leg, b ≈ 300 ft (planned final turn), L ≈ 100 ft (the completion limit — the lowest the turn can finish and still level the wing to flare), and assume Δ ≈ 75 ft (an estimate of the altitude a normal 90° toggle turn consumes at a moderate wingloading, erring on the high side — a diving or riser turn consumes more). So ε_max ≈ 300 − 100 − 75 = 125 ft: the entire error budget at the final turn is about a hundred and twenty-five feet — about eight seconds of level flight on the base leg (~900 fpm). If you believe you are at 300 when you are actually at 150 (ε = 150, 25 ft past the budget), the turn finishes at 75 — 25 ft into the low-turn zone. (Figure 5 nonetheless draws the final turn at about 100 ft: there the community's norms are tighter than this safety limit; see Appendix B.) In both cases we apply the same three-term formula, but the margin term b − L fell from 1,000 ft to 200 ft and the tolerance shrank fourfold. This shows how the tightest tolerance is at the lowest end of the skydive. Both are drawn in Figure 4.

The values drawn in Figure 5 are the author's own judgments for one hypothetical jump, a 4-way from 13,000 ft flown to the plan of Figure 1, each sized for the worst plausible case rather than the typical one: about 500 ft late at breakoff and at deployment, about 350 ft at decision altitude, and about 200, 150 and 100 ft at the three turns of the landing pattern, with more room on the early side throughout. Other jumpers would draw their envelopes differently; the reasoning behind each value is given in Appendix B, so that they can compare it with their own.

Every lapse in altitude awareness has its own geometry in time — its lapse geometry — but all fit two general shapes, the second of which occurs only in exceptional situations (Figure 6):

• Decay — attentional suppression, jumper conscious. The sampling loop is suppressed, and the salience of altitude in the jumper's mind wanes while the estimate goes stale. The decline need not be smooth: salience can go back up without an altitude check — a passing thought of altitude, an ambient cue, a teammate glancing at their altimeter — but if it carries no altitude information, it refreshes salience, not the estimate. Causes are diverse but share this signature: internal fixation (absorbed in the primary task — a formation, a docking sequence, camera work) and external capture (a sudden or salient stimulus pulls attention away from altitude — another jumper's malfunction, a shoe coming off on a dock, a near-collision) (Yantis & Jonides 1984; Theeuwes 2010). Because the jumper is conscious, this class is recoverable by an exogenous re-trigger of the sampling loop — a well-timed, salient indication breaks the spell, so-to-speak. This is the failure mode indication is designed to remedy.
• Step (an instantaneous drop to zero) — incapacitation, awareness vanishes. A freefall collision, a hard-enough riser strike, a medical event: salience drops off with no decay, a flat blackout (the internal clock is out too — no dead reckoning), then, if consciousness returns, awareness resumes lower, often in a dazed state with no sense of the time lost. This class is not re-triggerable — you cannot alert the unconscious — and it is here that an AAD is indispensable.

Severe hypoxia is an exceptional case and falls between the two: its onset is gradual, but then it eventually leads to total incapacitation, then finally death. Long before that, it impairs judgment, so that indications may not be registered or processed in an effective manner (SIM 5-6; FAA 2020). Hypoxia occurs on a spectrum, so most skydives from typical exit altitudes begin in a mildly hypoxic state — the degree depends on exit altitude, DZ elevation, time at altitude, and level of physical activity (e.g., scooting along a bench to the door from the front of a Cessna 208 and then performing a poised exit with a heavy tandem passenger at 18,000 ft): a jumper making normal exits from 13,500–14,000 ft pressure altitude was below 90% blood-oxygen saturation (SpO₂; 95–100% is normal at sea level) at exit and did not recover until under canopy (Bradke & Everman 2020), and on big-way loads to 15,000 ft MSL one jumper fell to 73% (Galdamez 2024). Controlled studies confirm the effect is mild, but measurable (Figure 7).

1.6 The character of altitude awareness

A jumper thinks about altitude largely in terms of the altitude structure outlined in §1.3, with particular attention paid to the thresholds. This has similarities to the way a delivery driver might think of arrival points — the stops along the route, not the road between them. The driver's mind may drift while driving away from the arrival points toward other distractions and then, occasionally, attention is somehow drawn back to the arrival points. However, in contrast to the driver, who in theory could stop at any point, a skydive has gravity as the engine and nothing equivalent to brakes.

Salience: inevitability and fear. Two interrelated psychological factors drive a threshold's salience: inevitability and fear. Inevitability: every threshold beneath the jumper will be visited and the descent cannot be stopped — only when, and in what state, are unknown. The hazard is never the altitude (its arrival was always certain) but the jumper's state at that point (cf. §4.6). This sense of inevitability weighs on a jumper, and it can either sharpen timing and reactions or cloud one's judgment in the case it leads to fear; the thresholds are fear-weighted according to the perception — accurate or not — of what failure at that point could potentially cost, or what kind of a situation such failure could lead to. The weighting is not abstract but visceral, as failure can sometimes mean not only a shot to the ego, but injury or death: students often feel some fear even at the mere mention of critical altitudes — the numbers themselves acquire conditioned affective charge even before a jump is made.

Occurrent vs. dispositional. Altitude awareness is occurrent when it is actually in consciousness at the moment. Between such moments a dispositional estimate persists below awareness: reportable if asked, and going stale as descent continues. The vast majority of the skydive occurs in the dispositional state of altitude awareness due to the sheer intensity of skydiving and the fact that there is nearly always some other task to fulfill on the jump unrelated to altitude awareness. There are many times on a jump when altitude does not really seem to be on a jumper's mind at all (recall the driver analogy). This distinction separates two kinds of decay: attentional decay is the fading of the occurrent state — the process behind the decay-shaped lapse of §1.5; estimate decay is the drift of the dispositional estimate. A lapse can be categorized by which failed. In a sampling failure, the jumper stopped checking: the occurrent state was not renewed. In a calibration failure, the stored estimate was wrong beyond what T(A) allows — it either drifted or was wrong from the start.

Coexistent, not exclusive. The occurrent state of altitude awareness rides alongside the primary task, resembling the monitoring of a running clock with specific points in time at which activities are scheduled to take place (some student training devices are actually timers at their core) — the mind holds altitude as a second thought concurrent with the activity.

Asymmetric dynamics — sawtooth-shaped. Occurrent awareness wanes but does not gradually wax (Figure 8), though it is not entirely unidirectional — as noted for the decay-shaped lapse (§1.5), attention can momentarily be brought back to altitude without a true refresh. Its general shape, however, is that it comes on suddenly, then wanes as the task reabsorbs attention, meaning it has a sharp onset and gradual decay. Each tooth's peak is a re-acquisition: the jumper is reading the altitude, so salience is at its highest and the estimate at its freshest. The peaks are not all the same height — the weight and surprise of each check (below) set how saliently it registers — and even the highest peak is not zero error: a reading is only as accurate as the indication and the jumper's reading of it, and altitude is constantly changing.

The decay is slow-then-fast, because altitude awareness is fundamentally a tracking problem. Altitude is a moving target: the task is not retaining a fixed value but keeping a moving one current, so the estimate must be propagated, not merely remembered. A fresh reading is held and advanced: the jumper dead-reckons it forward — "6,000 two seconds ago and falling at fast belly rate, so ~5,600 now." While the propagation runs, the estimate stays current — salience holds and uncertainty stays tight (a freshness plateau in altitude awareness); when the mental thread is dropped, currency is lost fast and salience drops away with it. Hence slow-then-fast: the flat part is currency actively maintained, the steep part is currency abandoned. The propagator has two inputs — the internal clock (§3.4), which supplies elapsed time, and a perceived fall rate, which turns it into altitude lost (altitude falls as time × rate): the plateau is propagation working properly, the fall-off is propagation abandoned. Corollary: an unusual or changing fall rate (belly→sit, an unfamiliar suit, a new DZ) complicates the propagation. The entire sampling cycle (§1.7) exists in the first place due to imperfect propagation; a theoretically perfect dead-reckoning jumper would only need a single glance at altitude at the top.

Psychological intensity of re-acquisition. Each renewal of awareness is characterized by both a level of surprise and a weight (gravity). Surprise: the intensity of the onset tracks how far the stored estimate has drifted from the truth — faint when the check confirms the estimate, startling when it was far off and a signal to recalibrate. Weight: independent of surprise, a check carries the stakes of the altitude it lands on as felt gravity — the anticipation of a critical altitude is significant even when the estimate is accurate; a re-acquisition in freefall at 3,500 ft with the same level of error correction is not experienced the same as one at 12,000 ft, even when both confirm the estimate. Weight is the visceral psychological representation of the tolerance profile T(A) (§1.5): how the jumper experiences T(A) psychologically. The two axes are separable: a check can be surprising but light in weight (wrong, but high) or heavy but not surprising (accurate, but low), or both at once — the dangerous case, as illustrated in Figure 9.

Fear: a double-edged sword. Fear can promote altitude awareness and is in fact the primary motivating force behind the sampling loop (§1.7), but past a certain intensity it erodes the very awareness its purpose is to encourage, in three ways. First, it degrades sampling and perception directly: danger raises arousal, and arousal narrows attention onto the most central cues (Easterbrook 1959; see also §3.2) — an effect documented in parachutists and scuba divers, though much weaker in experienced ones (Baddeley 1972). Second, it weakens the control that schedules a check: anxiety impairs goal-directed attention — above all the shifting between tasks that every check requires — and hands attention to whatever is most salient (Eysenck et al. 2007), so a check may be postponed, or made without registering (§1.7). Third, it corrupts the propagator: fear distorts the internal clock (§3.4), so the dead-reckoned estimate degrades while feeling maintained — the altitude estimate thread still runs, but on a bad clock. The nearest experimental evidence comes not from skydiving but from SCAD diving, about 2.5 seconds of true freefall from an amusement-park tower into a net (Stetson, Fiesta & Eagleman 2007): participants judged their own fall to have lasted about a third longer than one they had watched, yet a wrist display flashing numbers just too fast to read stayed unreadable — fear lengthened remembered time without speeding up perception. Real skydiving shows the same retrospective effect: among 76 novice tandem skydivers, those who reported more fear judged their jump to have lasted longer, and those more excited judged it shorter (Campbell & Bryant 2007). If the running clock bends the same way, the direction of a dropped estimate's drift is set by the jumper's state, not fixed: a fearful jumper, for whom time drags, tends to believe they are lower than they are — the early side — while an excited or absorbed one, for whom time flies, tends to believe they are higher, the dangerous direction. The dangerous drift is ironically on the enjoyable jump, not the frightening one. Laboratory work agrees that arousal lengthens judged durations (Droit-Volet & Meck 2007). Strictly, then, the direct evidence is that fear distorts duration judgment, and retrospectively; whether a skydiver's internal clock is distorted mid-freefall — in a trained jumper, with a task — is untested, though in the author's experience jumpers often describe the feeling of "time slowing down" in the event of malfunctions. Fear is therefore a self-limiting engine of altitude awareness: too little and the loop is not driven, too much and the readings it produces are corrupted.

Three quantities of altitude awareness: salience, uncertainty and error. Salience alone disregards both how sure the jumper is and how wrong they are. An estimation band — not to be confused with the tolerance envelope T(A) shown in Figure 5 — can be used to represent both: its width is the jumper's uncertainty (the spread of answers you might get if you could ask the jumper their altitude repeatedly) and its offset is the error (how far the middle of that spread lies from the true altitude). Figure 10 shows all three over a jump, with salience shown by the band's opacity, which decreases as altitude slips from mind.

The lapse, geometrically. The jumper's altitude estimation band and the jump's tolerance envelope T(A) (§1.5) can be drawn on the same time axis. The whole model appears at a glance in Figure 10 (salience, uncertainty and error can all be seen): the band can be seen drifting and widening between checks and snapping tight at each re-acquisition inside the envelope. A lapse is the band escaping the envelope: possible error exceeding permitted error, at the altitude where it happens. Once outside, a lapse deepens the longer it lasts: the band keeps drifting while the envelope narrows toward the ground.

• Dynamics — it fades and widens: between refreshes the band fades (salience wanes — attentional decay) and widens (uncertainty grows — estimate decay); re-acquisition restores both. Width generally narrows toward the ground (ambient cues and natural altitude perception sharpen, §3.4; sample rate often rises — §1.7).
• Felt uncertainty vs. true error: the band's width is the jumper's felt uncertainty; true error (deviation from actual altitude) is unknown until a refresh reveals it. Calibration is indicated by the band containing the actual altitude; overconfidence is represented by a narrow band in the wrong place — and when that band is also fully salient (opaque in Figure 10), it is the confident-but-wrong case of Figure 11. Complacency and overconfidence thereby have distinct visual shapes: a fading, widening band vs. a tight band off-center.
• The band never closes to a line. Altitude never holds still: a check samples a value that keeps changing as it is read — about 90 ft in half a second of belly-flying freefall — and both the display and the jumper round what they see. The instrument adds its own error on top: a refresh collapses cognitive error (mental estimate vs. instrument reading) but not measurement error (instrument reading vs. real AGL). Writing A for the actual altitude, Â for the jumper's estimate and I for the indicated altitude, the total error splits into the two: e = (Â − I) + (I − A), cognitive plus measurement. Both terms are present at once and both are signed, so they can compound — an absorbed jumper drifting late on a cold day, when the altimeter also reads high — or partly cancel. A refresh zeroes only the former; no amount of checking can correct the latter. Altitude measurement, which makes only a limited appearance in this theory (§1.1), enters the awareness model here. A barometric altimeter's error comes in two forms. Temperature shifts the band (humidity slightly, and only in the protective direction): on a cold day the altimeter reads high, so the jumper is lower than indicated — confident but wrong, and shared by every barometric altimeter and AAD. The temperature shift grows with altitude above the DZ: tens to a couple of hundred feet at pull altitude, but up to a thousand feet or more at exit on a very hot or cold day (Appendix A). Lag adds a signed bias of its own — in a fast descent the altimeter trails the true altitude and reads high, the dangerous direction — while noise widens the band.

Figure 11 shows salience against signed error on two axes, revealing four extremes — every altitude estimate occurs somewhere in this space.

1.7 The sampling cycle

The jumper acts as a sampler with an unfixed sample rate, not a continuous monitor: conscious thought is directed to altitude; the estimate refreshes; the working task reabsorbs consciousness; altitude fades from the foreground; something re-triggers the check; the cycle repeats. The sample rate often rises as critical thresholds approach.

Re-triggers. What fires a check is either endogenous — trained habit, the internal clock, prospective memory (remembering to remember; Einstein & McDaniel 1990), task transitions such as finishing a point — or exogenous: indications, the appearance of the ground, other jumpers' behavior. Expertise automatizes endogenous re-triggering; stress can disrupt it (§1.6). Seen from the indication's side, the same split separates indications read on demand — a display that waits to be read — from those that come on their own, of which there are two sorts: background, present without a sampling act (a peripheral LED altimeter), and interrupting, an indication that announces or shows itself and is built to distract the jumper from what they are doing. A background indication keeps the estimate up while the loop runs; only an interrupting one can re-trigger a loop that has stopped; Part V decides which to use where. Seen from the loop, a sampling failure (§1.6) is the re-trigger not firing before the estimation band escapes the tolerance envelope T(A); task fixation is the cycle suppressed by a competing task that keeps winning attention. Re-acquisitions keep the estimate calibrated, since each check resets it to the indication, but nothing can remedy a check that itself is off: a misread altimeter, an indication not registering in the mind (a registration failure, below), or an indication that is itself off — zeroed wrong, lagging, or biased by temperature and humidity (§1.6, Appendix A).

The failed sample. A sample can fire without registering — a registration failure, the third kind alongside the sampling and calibration failures of §1.6: the jumper glances at the altimeter but does not really read and think about what it says, so the value never reaches awareness. Surprisingly, even a loud beeping audible altimeter can go unregistered (SIM 2023–2024, 5-3.J.4.c.(2); Dehais et al. 2014). This is a distinct failure from not looking or listening (attentional decay, §1.6) and from an unreadable signal (masking, §3.3): here the jumper looks and the information is right there, or the alarm sounds, yet no reset occurs, because the reset is on registration, not emission (§1.5). Human-factors research calls this a looked-but-failed-to-see error (Mack & Rock 1998; Wolfe, Kosovicheva & Wolfe 2022); its auditory counterpart might be called listened-but-failed-to-hear. It is especially deleterious because it feels like a completed sample — the jumper believes they checked — so the estimate is neither refreshed nor flagged as stale: a possible pathway to a confident-but-wrong estimate (§1.6); overconfidence in altitude estimation can have deadly consequences. A close sibling is misreading: here the sample does register, but delivers the wrong value (a needle read against the wrong graduation, a transposed thousands digit). It is in one sense worse than the failed sample, which merely leaves the estimate stale — misreading actively installs a wrong estimate held with full confidence — a calibration failure created at the moment of the check. This is the most direct pathway to a confident-but-wrong estimate — it does not even require estimate decay. Fittingly, altimeter misreading is itself a founding human-factors problem: in a 1947 analysis of 270 pilot instrument-reading errors, the altimeter was the most frequently misread instrument, its hundred-, thousand-, and ten-thousand-foot pointers confused for one another (Fitts & Jones 1947); later comparisons ranked the three-pointer design lowest, and designs moved to displays that pair a pointer with a numeric counter (Grether 1949; Hill & Chernikoff 1965; ICAO 2003). The NTSB could not determine why United Air Lines Flight 389 descended into Lake Michigan in 1965, but noted that the crew "could have misread 6,000 feet to be 16,000 feet" (NTSB 1967).

Fear drives the sampling loop, and is depletable. Fear is the primary motivating force behind endogenous re-triggering (§1.6). We can distinguish two types of relevant fear: (1) a diffuse, sub-clinical fear of heights — not vertigo or acrophobia, which would be disqualifying, but the ordinary background unease of exposure — which diverts the mind to altitude more often, raising the baseline sampling rate simply by keeping the ground salient; and (2) the fear weighting carried by the critical thresholds themselves (§1.6). A safety-conscious jumper has enough fear of heights to keep sampling elevated, but not so much as to be incapacitated. Students are taught to check often and fear re-triggers the check constantly, sometimes even disruptively. But fear habituates with exposure — arousal before a jump follows a markedly different course in experienced parachutists than in novices (Fenz & Epstein 1967) — and its slow-but-sure dwindling removes the primary impetus of the checks.

Sampling among experienced jumpers. The shift with experience is not simply audible-for-visual substitution but a near-abandonment of visual instrument sampling. High-jump-number skydivers hardly look at their visual altimeters at all — mostly around breakoff and to set up the landing pattern (both threshold checks) — while dispositional awareness through freefall is carried by ambient cues and the internal clock (§3.4). Whether that near-abandonment is maturity or complacency depends on what replaced the checks; one answer is the higher-order awareness of §1.8.

Fixed intervals in training. A fixed check interval is prescribed in student training. The current SIM states it outright, with the conversion built in: "Check your altimeter every 5 seconds (approximately 1,000 feet of freefall), every time you complete a maneuver, anytime you encounter difficulty, and if you simply do not know your altitude" (SIM 4-1); the student program gives the arithmetic behind it: "The first 1,000 feet takes 10 seconds, and the next 1,000 feet takes 5 seconds" (SIM, ISP Category D). The student chapter asks for more — "every three seconds" (SIM, ISP Category A) — and drop-zone first jump courses teach their own versions, such as "every 5-6 seconds … until you have reached 6000'" (Skydive City 2017). The interval is an implied tolerance claim: at belly-to-earth fall rates, 5–6 s ≈ 1,000 ft of descent, so the rule is really more like "check your altitude every 1,000 ft." If the taught interval is adequate, a student's tolerance must be larger than the ~1,000 ft that passes between checks: the prescription is a conservative margin, not the boundary. The ISP also states a particular allowance outright: the Category A advancement criterion is to "initiate main deployment within 1,000 feet of the assigned altitude" (SIM, ISP Category A). That is not T(A) itself but a graded standard set well inside it: from a 5,500-ft assigned altitude and the 3,000-ft student minimum (SIM 2-1.I), the formula gives a student a deployment tolerance of about 2,000 ft. Note the student is taught to perform periodic checks down to the 6,000-ft threshold, then to lock on (keep looking at the altimeter) for continuous monitoring — a clear implication of a tightening tolerance envelope — until pull (SIM, ISP Category A).

1.8 Higher-order altitude awareness

Everything to this point describes first-order awareness. At the pinnacle of altitude awareness lies a higher order meta-awareness (Schooler 2002). Reaching this level requires a certain degree of mastery of skydiving; however, simply pondering the concepts included within this document can prime a jumper and plant the seeds for the development of such higher-order altitude awareness. Jumpers who practice this have an awareness of and can monitor the state of their own altitude awareness while skydiving: its fuzziness, its freshness, the characteristics of the sampling loop. In the terms of metacognition research, this is the meta-level of Nelson and Narens's (1990) framework: it monitors the object level — here, first-order altitude awareness — and controls it, chiefly by re-triggering the sampling loop.

• It does not monitor altitude itself. Rather, in relation to what is described in §1.6, it monitors salience ("Is altitude on my mind?"), staleness ("When did I last check?"), and the width of the estimation band ("How sure am I right now?").
• Experience and currency play a major role. Experienced and current jumpers have a head start: they can apply these concepts with less effort than those still learning the basics, which take up enormous attentional bandwidth, or those out of practice. Its seeds may be seen in the rare AFF jumper (a wind tunnel instructor, for example) who may actually have the capacity in freefall to wonder, "Am I checking my altimeter enough like my instructor said?" Its mature development sees its implementation in a measured, matter-of-fact manner.
• It tends to produce better first-order altitude awareness. Those who are actively monitoring their own patterns and habits of monitoring altitude will notice quickly if the sampling loop has stopped or the altitude estimate has gone stale. Often this causes a re-triggering of the sampling loop. Meta-awareness thus tends to lift first-order awareness above baseline — but only as far as its monitoring is accurate. Self-monitoring rests on cues rather than direct access to one's own state, and it can be miscalibrated (Koriat 1997): a jumper can feel current while the estimate is stale, as in the failed sample (§1.7).

1.9 Stage-specific characteristics

In addition to a host of commonalities established in §§1.5–1.8, each stage of a skydive has unique demands and characteristics when it comes to altitude awareness (Figure 1 depicts the stages and their thresholds).

Ascent. This is the only stage in which altitude is generally increasing, so a stale estimate from a jumper's mind that has entirely drifted elsewhere often errs on the safe (low) side: the aircraft keeps climbing past it. An estimate dead-reckoned forward, though, can err either way — the climb may have slowed — and in an aircraft emergency, overestimating altitude is the dangerous direction. Although the SIM's definition of a skydive begins at exit, for the purposes of indication, ascent is part of the skydive (§1.2): maintaining altitude awareness in an aircraft is important for safety reasons; above a certain altitude it can be safe to make an emergency exit, while getting out low can have fatal results. The SIM makes this concrete: below the drop zone's seat-belt-off altitude, jumpers land with the aircraft; above it but below their decision altitude, they exit on the reserve; higher still, on the main (SIM, ISP Category E). Every one of those choices depends on knowing the altitude at the moment the emergency happens. AADs, too, arm themselves partway up the climb (Airtec 2025; Advanced Aerospace Designs 2021). It is almost always the longest stage of the skydive and the stage where jumpers most often "tune out." It has the lowest task load of any stage. In addition to the expected visualization and anticipation of the jump they are to carry out, many people — especially those with a greater experience level or those working in the sport — simply think about other things on the way to altitude. It may even seem rational to be altitude unaware during this stage: a jumper needs to mentally prepare for the skydive; the pilot in command is not the jumper; there is usually a group of other jumpers; and it isn't officially part of the skydive. But it only appears rational until one experiences an emergency, at which point the value of altitude awareness becomes apparent. Ascent is a good stage for jumpers to calibrate their altitude awareness by looking out, estimating altitude, and comparing their estimate against a numeric altimeter.

Freefall. For those who have never experienced freefall, it would be hard to imagine just how high-intensity the experience really is and how different it is from ordinary life. At least with activities such as motorcycle racing, extreme skiing, or big-wave surfing, you are still in a more familiar environment. In short, the experience can be distracting, which can negatively affect altitude awareness. Freefall demands the highest sampling rate of any stage: altitude is lost fastest, so the estimate goes stale an order of magnitude faster than under canopy. Even belly-to-earth, the slowest standard freefall position, a jumper falling at about 120 mph loses 1,000 ft in under six seconds (about 5.7 s).

Canopy. Canopy flight has a significant number of important tasks that have to be completed in order to descend safely and compliantly to the ground: establishing clear airspace, a canopy control check, housekeeping tasks (e.g., loosening the chest strap, adjusting the slider or removing the slider, deployment bag and pilot chute entirely), a flight to a holding area, a well-executed pattern with defined altitude thresholds, a well-timed landing flare. All of these tasks should ideally integrate good altitude awareness, not only to stay within margins but because it makes a jumper predictable to the other canopies in the pattern. The closer we get to the ground the more usable our natural, non-instrument-aided visual sense of altitude becomes. Canopy has the tightest tolerances of the jump, as outlined in §1.5. It is the stage in which lapses in altitude awareness and altitude misjudgments most frequently have fatal consequences. The fatality record bears this out: the recent deaths in which USPA names a loss of altitude awareness are mostly malfunctions under canopy, where a jumper working on a problem passed decision altitude without acting — "Focusing on the problem can lead to a loss of altitude awareness, and before you know it you are well below your decision altitude" (USPA Fatality Summary 2022). In one tandem accident, the pair had a spinning malfunction and reached the 3,500-ft decision altitude just 18 seconds after deployment (USPA Fatality Summary 2020). There is a point at the end of canopy flight below which the jumper no longer acts on the estimate and flies on directly perceived altitude (§1.5); the numeric altitude estimate itself persists, but is no longer taken into consideration in the landing.

Part II — A Taxonomy of Altitude Indications

Skydivers today have a broad range of devices to choose from when it comes to altimeters. Whereas the earliest skydivers used no altimeters at all, it isn't unusual today for a jumper to be wearing two or even three altimeters at once. These include numeric altimeters with dials and displays of various sorts, peripheral-vision LED altimeters, speaking audibles, beeping audibles, GPS-based altimeters, and more. The list continues to grow as manufacturers continue to explore new ways to convey altitude.

2.1 The five dimensions of an indication

The taxonomy of altitude indication uses a faceted system of classification, meaning it is built from several independent dimensions (facets) that combine to locate a specific indication; these dimensions form a coordinate space for comparison, not a set of bins. In contrast, hierarchical classification sorts each item into one branch of a tree, such that each item has exactly one location. Here, a location in that coordinate space — a single value in each dimension — is called a cell. Real-world indications exhibit characteristics unique to themselves that no dimension can capture (the specific timbre of a siren, an LED altimeter displaying several colors at once). And many are multiply coded, occupying several cells at once (the needle of an analog dial read in a single glance as a needle position, in the yellow zone, and a value of 2,800 ft all at the same time). We restrict multiple coding by definition to a single signal: a tone and a light, even when synchronized, are considered two indications in two modalities — redundancy across modalities (§5.4). As a rule in the following, we establish a dimension of indication only if it earns its place — that is, if it usefully tells designs apart, reveals a trade-off, or exposes a structurally empty cell (a combination that cannot exist).

Five dimensions emerge as having such characteristics. An indication is located by its span (ranged or non-ranged, §2.2), its encoding (how the signal represents altitude, §2.3), its motivation (whether that representation resembles altitude or is merely conventional, §2.4), its modality (the sense it addresses), and its role (phase or threshold, §2.5). Four are covered in this Part, while modality is substantial enough to have its own dedicated Part (Part III). §2.6 lays all five out in an indication space. Multiple coding can occur in encoding (the analog dial's needle, numerals and colored arc, §2.3), in motivation (the breakoff light bar, iconic in its flashing and conventional as a police light, §2.4), and in span and sometimes role (where a signal marks a threshold and then persists, §2.5) — but never in modality: a signal in a second modality is a second indication.

One scope note governing all of the dimensions that follow: they classify indications by the characteristics of their emission — what the altimeter delivers by design — rather than their perception. How an indication is registered by the jumper, and what it does to their awareness, is the subject of Parts I and III; where this Part does touch on perception — as in the sawtooth discussion of §2.2 — it says so explicitly. Emission and perception do not always line up.

2.2 Ranged vs. non-ranged

We can classify indications by whether they are conceived of as occupying an altitude range (i.e., a span) or a point; this is not to be confused with our discussion of the phases and thresholds of a skydive — here we are concerned purely with the properties of the indications themselves. A point indication takes up an incidental altitude range — it ends when the signal ends, not at another altitude — so we call it non-ranged. Ranged indications are what the name suggests: they begin at one altitude and end at another.

The standard human-factors terms for these are continuous and discrete; we mention them so the taxonomy can be compared to that literature. In medical alarms, for example, discrete alarms — “the sounds that signal when some clinical parameter in the patient has been exceeded” — are contrasted with continuous auditory feedback (audification or sonification) (Edworthy 2013). The correspondence is close but not exact, because continuous is used in two ways: it can mean that a signal persists across an interval, or that it varies smoothly rather than in steps. We use ranged in the first sense: that the indication occupies an altitude span. Outside this paragraph, this document uses ranged / non-ranged throughout; continuous is kept only for a signal that is literally unbroken in time (a steady tone). We will also introduce the terms static and dynamic as properties of a range when discussing encoding in §2.3.

• Ranged indication — persists over a specific altitude span.
• Non-ranged (point / "one-shot") indication — fires once at a defined altitude crossing (e.g., a single four-second beep at 12,000 ft on ascent). Necessarily takes up an incidental altitude range due to its temporal nature but conceived and intended to be perceived as an event, not a state; maximally precise at the instant of delivery. (Role note: typically serves a threshold — §2.5. Non-ranged is its temporal nature; threshold indication is its role; the two are different axes.)
• A train of non-ranged pulses (beeps at 5,500 / 5,000 / 4,500) approximates a ranged indication: identical pulses approximate a static one, and pulses that vary — in pitch or tempo, or spoken values — a dynamic one (the sampling-rate idea below). A simple test: if a signal's start and end are set by time, it is non-ranged; if they are set by altitude, it is ranged. For a train of pulses, apply it to whatever carries the meaning: a beep every 1,000 ft is a train of non-ranged indications; beeps accelerating through the last 1,000 ft before breakoff form one ranged indication.

(A note on implementation, in an era with few analog systems left: at a fine enough grain every digital indication is stepped, and only the analog instruments of the sport's early days, such as the aneroid needle, emitted truly smooth signals (see indices, §2.4). Here ranged is defined by span occupancy, not smoothness, and steps below perceptual resolution fuse psychophysically; the classification is made at the scale of the task and the limits of the jumper's perception, not the device's underlying electronic characteristics.)

These two forms, ranged and non-ranged, work in a complementary manner on the sawtooth of occurrent altitude awareness:

• Non-ranged indication can effectively be used to add a tooth, so to speak — it can re-trigger awareness at a point. More non-ranged indications = finer teeth = higher effective sample rate; the occurrent-awareness sag between them remains.
• Ranged indication can work to remedy the decay — it can keep the estimate up between refreshes, and flatten the sag toward a plateau. We will see this is especially true when dynamic indications, which change according to altitude, are used.

Real-world constraints:

• Flattening requires a monitorable modality and a low, standing allocation of attention. A ranged display consulted by occasional glancing is still sampled glance by glance (it adds teeth, not sag-fill): a ranged emission does not by itself make for sustained perception (§2.1). Only a modality that arrives without a sampling act — audition, peripheral vision, touch — genuinely flattens the sawtooth (a background indication, §1.7), and even then not for free: noticing without actively looking still requires some level of attention, and the channel must be held open by a low, standing allocation rather than by focused checks. This is why peripheral-vision LED altimeters (ColorAlti technology) can remedy the decay where an analog altimeter's needle cannot.
• Raising the floor of altitude awareness. Fully flat awareness would keep altitude permanently in conscious attention, competing with the primary task; the sag is partly a feature — it frees attention between refreshes, enabling the jumper to concentrate on the tasks of the skydive. A design paradigm for ranged indication: raise the floor (lowest point) but keep the sawtooth — keep the dispositional estimate within tolerance T(A) without consuming occurrent conscious attention. (The design principle this yields is stated in §5.2.)

2.3 Encoding: how altitude is represented

Encoding options depend on whether an indication is ranged or non-ranged (§2.2). We have two encoding groups separated by the manner in which the signal represents altitude:

• Direct — the signal's physical form carries the jumper's relation to a set altitude: whether they are within a set span, where within it (e.g., the hue of an LED), or that a set altitude has just been reached (e.g., a beep at 300 ft). The altitudes themselves are assigned, by the device's settings or its design, and the jumper must know them: no direct indication carries its own value. It divides into three subgroups by range (§2.2):
  • Static indication (ranged only) — form uniform across the span; can be an unvarying stimulus or a repeating pattern; conveys simply that the jumper is within an altitude range and nothing else.
  • Dynamic indication (ranged only) — varies across the span (rising pitch, accelerating beeps, brightness, LED hue change); position within the span is carried as a magnitude. Helps a jumper to know not only that they are within an altitude range, but also where within that range.
  • Event-marker (non-ranged) — a bare point signal (e.g., a two-second beep at 5,500 ft in freefall): the signal itself carries only "the set altitude arrived" (nominal/occurrence). The non-ranged counterpart to static — presence, but of an event or threshold rather than a window or phase. Cannot be static or dynamic (conceptually it has no span to be uniform or vary across, as the altitude span it takes up is incidental).
• Symbolic — the signal itself carries the value or a named meaning, by assignment: it is a sign (e.g., a spoken value, displayed digits, or a beep pattern that tells breakoff from pull) whose meaning is set by learned convention rather than by its form. Its meaning is almost always a point value, in which case it is non-ranged — "5,400" names an altitude, not a span — so range applies only to its carrier and not to altitude: a callout ends when its words end, while a display persists for an arbitrary amount of altitude as it steps from value to value. An interesting exception: an LCD displaying "BREAKOFF" from 5,500 to 5,000 is clearly a ranged symbolic indication, but a spoken "Breakoff!" at 5,500 is non-ranged — its altitude span ends when the speech ends. Resolution — how finely the value is given: to the nearest 10 ft, the nearest 100 ft ("5.4"), or once every 1,000 ft — sets symbolic indications apart from one another, as does the decoding process they demand. An event marker and a symbolic value are both point-like, but conceived entirely differently: the latter carries its value in the sign itself, while the former carries only its arrival; because it works by convention, symbolic also sits outside the realm of static/dynamic indications.

The analog dial altimeter is a hybrid indication, multiply coded (§2.1) across both groups. The needle tip's position varies smoothly with altitude (direct/dynamic), while the numbers on the scale assign that position an exact value (symbolic) — and the colored arc adds a third (direct/static) encoding: presence in a region. The position provides a rough magnitude at a glance, while the scale enables one to ascertain an actual value after decoding.

A new experimental altimeter developed by Freefall Data Systems is another hybrid design — a digital example — arranging the same two groups, direct and symbolic, differently. WatchAlti (in development) shows altitude on two concentric rings of ten LEDs — the outer ring the thousands, the inner the hundreds, with a single LED just above the center of the face for ten-thousands (21 total LEDs). Its structure is therefore place-value: the quantity is decomposed by an assigned rule — which ring means which decimal place — into values that are each rendered directly: the outer ring as a filled arc (an extent), the inner ring as a single lit position (an angle) or, at the user's option, a filled arc. The difference from the analog dial is the manner in which the assignment does its work: the dial's numerals label a single direct parameter, whereas WatchAlti's place-value rule combines three direct parameters — two rings and the ten-thousands light — into one value. The inner ring's parameter — the hundreds — is therefore periodic in altitude rather than monotonic: it sweeps around and resets every thousand feet, resembling a magnitude only within its period, and each reset doubles as an event marker — a built-in tick every thousand feet. Unlike numerals, which must be looked at to be read, its LEDs can often be noticed in peripheral vision when they change, and the fill gives a coarse reading at a glance; the place-value decoding is needed only for the exact value. The instrument thus sits between an on-demand display and a background one (§1.7). In its single-position mode the instrument is, structurally, the three-pointer altimeter of §1.7. In its all-lit mode — in which all 21 LEDs are lit at the same time — the same device essentially becomes a ColorAlti altimeter on the wrist (§2.4).

The classic "one-shot" alarm (e.g., a single beep at 300 ft) is therefore an event marker, and thus non-ranged (§2.2). A spoken "three hundred" at the same altitude is also non-ranged but differs in encoding: it actually names the value, so it is symbolic. The one-shot beep is not a further encoding type but simply an event marker.

Key relations:

• A numeric display with a high refresh rate changes with altitude, yet it is symbolic, not dynamic: the digits are signs whose values are assigned, not a parameter of the signal that varies with altitude. What separates direct from symbolic encoding is relation vs. value — whether the signal's form carries the jumper's relation to a set altitude (direct: static, dynamic, event-marker) or the sign carries the value itself, by convention (symbolic) — and not whether the indication changes.
• Only symbolic can provide an exact, unambiguous value. A beep, for example, cannot say "300," even if you know that is where it is supposed to occur. So while the direct group can indicate the arrival of a threshold (both ranged and non-ranged), it doesn't provide an exact value as unambiguously as symbolic does. By contrast, the ranged subgroup of direct can convey approximate altitude (within a span) at any point in time.
• The trade-off, for indications of position: symbolic buys exactness at the price of a decoding process (learned, working-memory-dependent, slower, worse under arousal); dynamic buys immediacy at the price of precision. Indications of action, such as "Pull!", fall outside it: they carry no value to be exact about. (§3.2 covers the cognitive-load side.)
• One cell is structurally empty: non-ranged + dynamic — non-ranged indications are conceived as a point, which has no span to vary across.

2.4 Motivated vs. arbitrary: the iconic–conventional spectrum

Whether the encoding of an altitude indication — direct or symbolic — is motivated or arbitrary is a property of its mapping. Arbitrary is not used in a negative sense here; motivation and arbitrariness simply form the two ends of a spectrum — Saussure's "arbitrariness of the sign" and its opposite, the motivated sign (Saussure 1916). Direct encodings tend to be motivated, i.e., their mapping is constructed in some way to resemble altitude (e.g., sweeping through the hues from blue to red would typically be seen as representing going from high to low). Symbolic encodings tend to be arbitrary: the number "300" itself doesn't resemble any altitude region the way a color can — it is only incidental, for example, that it contains fewer digits than "3,000," not a consequence of indication design. Neither tendency is a rule, however: a hue sweep on an unintuitive palette is direct and arbitrary; a tally of beeps, or a filled arc read by counting, is symbolic and motivated, because a count resembles its number. The tally is not hypothetical: a 1997 patent has an audible altimeter confirm its stored breakoff altitude as one beep per 1,000 ft, with one trailing beep adding 500 ft (5,500 ft = five beeps, then one) — another way for a number to be conveyed with no display and no speech (Pilkington 1997). This is why the arbitrary pole is called conventional here rather than symbolic (a term we use in §2.3 for encoding type): the two dimensions do not coincide.

• Motivated (iconic) — the sign→meaning mapping is grounded in resemblance or natural association (e.g., warmer = lower altitude). Often guessable, cross-cultural, or reconstructible to a degree from the form itself.
• Metaphor — the middle of the spectrum. Between resemblance and convention lies a mapping that corresponds to a metaphor. A sustained flat tone set at 300 ft — a flatline — resembles no altitude but does indeed resemble the sound of a flat EKG, suggesting the most critical phase of the skydive is being entered. The flatline has been over-dramatized by Hollywood over the years, as many EKG monitors do not produce such alarms for a flatline, instead often favoring harsher intermittent sequences; nonetheless, the association remains strong. The sign is borrowed from a code everyone already knows, the meaning is metaphorical, and it carries all of its urgency semantically while paradoxically being, acoustically speaking, the least urgent signal in the parameter space — no pulse, no onset, no roughness (in timbre), often not in the most sensitive range of hearing (cf. §5.5–5.6).
• Arbitrary (conventional) — the mapping is conventional rather than resembling. May not be cross-cultural (e.g., a callout in a language the jumper does not speak). The word conventional does not mean common here, but agreed upon or assigned.

One further distinction within the arbitrary pole matters for much of what follows: whether the convention is local — set by whoever configures the device, its manufacturer or the jumper, reassigned at no cost, and memorized per device (two beeps for breakoff could as well have meant deployment) — or inherited from a shared code the jumper already has (a spoken "three hundred"; the numerals on a scale; an analog dial with 12 numbers on its face, read like a clock face). Inherited conventions can be overridden only by fighting the prior convention, and they need no learning here: they were learned when the jumper learned to speak and read.

• The reversal test (diagnostic for motivation): If the reversal of a mapping seems illogical, the mapping is motivated; if the reversed mapping would serve just as well, it is arbitrary. (Reversing a color sweep to red-high / blue-low will not be seen as logical to many — motivated; reassigning which beep pattern means what violates no intuition — arbitrary.)

  • A case that leans arbitrary: beep tempo. In freefall, beeps that accelerate toward the ground are motivated — reversing them would be extremely unusual; the tempo resembles closure, and faster is more urgent. Under canopy it is not uncommon for this to be reversed: fewer, longer beeps as the ground approaches — against the urgency mapping, since lower is more urgent. The likely reason is concentration: an alarm beeping faster and faster as the jumper enters final approach would distract at the point where the eyes have taken over most of the work. So here the urgency mapping yields to concentration during the most dangerous phase of the skydive. The lesson is general: the same direct parameter can lean iconic in one stage and conventional in the next, because the polarity of a mapping is a design choice.
  • A contrasting case: WatchAlti's fill. More lit LEDs for more altitude is motivated in two senses — by quantity (a bar grows with what it measures; more marks = more) and by the inherited convention that gauges drain as a resource (altitude) is spent. Reversal in such a case — a bar that grows as altitude is lost — would have to fulfill a particular purpose, such as leading up to a threshold, to be seen as logical. The contrast with beep tempo shows the reversal test grades motivation: illogical reversal points to a uniquely motivated polarity (the color sweep, the fill); plausible reversal suggests a motivated mapping with a competing motivation for the other polarity (canopy beep tempo, above); indifferent reversal points to an arbitrary mapping. In cases where both polarities are motivated, the designer's choice between them is a convention between motivations that must be learned.

• Failure layers: every indication can fail at the registration layer (Level 1 in Endsley 1995; §3.3): it can go unheard under masking or with blocked sinuses, be drowned out by task load, or be lost in a failed sample (§1.7). No choice of encoding protects against that; anything can fail a jumper who does not register the signal. This is likely the novice's most common failure. Overload and task fixation feature prominently in AAD saves (§1.4), and in aviation about three-quarters of situation-awareness errors (Endsley's term; "situational" in everyday use) arise at the level of perception (Jones & Endsley 1996). What the encoding can mitigate is a second failure point: a misinterpretation. Motivated mappings and inherited codes are understood once they register, which is why speaking altimeters do not fail in the same way beeping audibles do (a spoken "Pull altitude" at 3,500 ft is unambiguous if it registers, whereas the meaning of a beep at that altitude is not intrinsically comprehensible). This supports the use of motivated or inherited encodings for the most critical thresholds and phases of a skydive. Two further layers, response failure and incapacitation, belong to the jumper rather than to the signal (§3.3).

• Motivated ≠ convention-free. Every indication has a mapping (how variation in the signal encodes value) and a referent (the quantity the value refers to). Resemblance works only on the mapping; the referent is conventional for virtually every indication — a blue→red sweep would be exactly as motivated for airspeed or fuel remaining, since the gradient says "a lot → a little," never specifically "of altitude."

• We distinguish three targets of iconic encoding — magnitude, urgency, and action — magnitude being the only one to require a ranged indication. Each is something a signal can resemble:

  • Iconic-of-magnitude ("how high") — requires a span to vary across, so it is ranged only (= dynamic; a train of beeps at 1,000-ft intervals falling in pitch counts: its meaning lies in the sequence, so by §2.2's test it is one ranged indication): a lone pitch, loudness or duration cannot resemble a value, since it has nothing to compare against. Self-scaling signals are the exception: a count supplies its own unit, and a part of a visible whole — a gauge lit once at a fraction of its length — supplies its own reference, so both resemble a number without needing a span. This is how devices without a display are able to use counted beeps to confirm a stored setting (e.g., five beeps for preset 5).
  • Iconic-of-urgency ("how much this matters") — carried by the signal's character (harshness, loudness, tempo, roughness, onset), so it is range-independent — urgency can apply to a one-shot beep or a continuous tone equally, though a span lets it escalate: beeps speeding up toward a threshold say the moment is growing more urgent. Urgency is strongly motivated. It is grounded in cross-species signaling, where harsh, low sounds signal threat (Morton 1977), and in the psychoacoustics of urgency (Patterson 1982; Edworthy, Loxley & Dennis 1991; Hellier, Edworthy & Dennis 1993). It is also partly learned: some urgent patterns, like a siren's wail, carry their meaning by convention as much as by nature. Reversal test passes: a gentle chime for "PULL ALTITUDE" is illogical.
  • Iconic-of-action ("what to do") — the signal resembles the response required rather than the altitude or its urgency. This is the rarest of the three: most skydiving actions — pull, track, turn — have no form a signal can imitate. One example might be a jumper's siren set to sound only in freefall at 2,000 ft, meaning "pull the reserve" — an emergency response sound initiating an emergency response. Another is the flatline at 300 ft — an unbroken tone with no events in it, iconic of a jumper who on final approach best makes minimal inputs: stillness in the signal iconic of stillness in the flying. The siren resembles the response by metaphor; the flatline, by its form. Iconic-of-action is range-independent, like urgency.
  • Consequences: a one-shot alarm can sound urgent but cannot intrinsically tell the jumper how high they are; and of the three targets, urgency is the one with a mature design science behind it — the cockpit urgency-mapping literature of §4.2 is, in these terms, the science of iconic-of-urgency.

Indices. A third kind of sign sits off the motivated–arbitrary axis: the index (Peirce's term; Peirce 1932). An index is tied to its object by cause or contiguity rather than by resemblance or convention — smoke for fire is the classic example. Being an index is a separate property rather than a third point on the axis, so one sign can be both indexical and iconic. Most of the ambient cues of §3.4 are indices: ear pressure, ground rush, warming air and wind noise are all caused by altitude or speed. Every instrument indication is an index in this sense — a digital readout no less than an analog needle, its chain of cause merely longer (sensor, converter, firmware, display) — but the needle is the clearest case: the air physically moves it, so its position is at once indexical and iconic (§2.3).

Roots in the field of auditory display. In the auditory modality, the motivated↔arbitrary distinction corresponds to the established auditory icon vs. earcon distinction from auditory-display / sonification research (see The Sonification Handbook, Hermann, Hunt & Neuhoff eds., 2011, open-access). The field's sound types each land somewhere in the indication space. An auditory icon (an everyday sound representing by resemblance; Gaver 1986; Brazil & Fernström 2011) is motivated (iconic). An earcon (an abstract, structured motif whose meaning must be learned; Blattner, Sumikawa & Greenberg 1989; McGookin & Brewster 2011) is arbitrary in what it means — the field calls earcons "symbolic" in the sense this document calls conventional — though its character can still carry iconic urgency. Parameter-mapping sonification (a data value driving a sound parameter, e.g., altitude→pitch; Grond & Berger 2011) is dynamic and iconic-of-magnitude. Speech ("five thousand") is a symbolic encoding carrying an inherited convention, and auditory display rather than sonification (sonification is defined as the use of non-speech audio to represent information; Walker & Nees 2011). A single alarm beep or short beep pattern is what the field calls a one-element earcon, the simplest kind, which "may be only a single pitch or have rhythmic qualities" (McGookin & Brewster 2011). In §2.3's terms, a lone beep at a set altitude is an event marker: direct in when it sounds, arbitrary in what it means. When the sound also tells one meaning from another — a two-beep breakoff alarm set apart from a three-beep pull alarm — it is symbolic as well. A beep rapidly increasing in tempo according to altitude over a range is parameter-mapping sonification; a "falling" sound (like one hears in cartoons) would be an auditory icon. The field also offers a mature body of design results to import, especially Ch. 19, Intelligent Auditory Alarms (Guillaume 2011).

Worked example — ColorAlti technology (visual; dynamic/iconic; magnitude), the light on MetaAlti, WatchAlti and the DigiAlti 3 LightBar. On MetaAlti the default sweep runs blue→red through the spectrum from 13,000 to 5,500 ft. Triply motivated — making the reverse unthinkable even to a non-jumper: (1) aerial perspective — the air scatters blue light, so distant and high things look blue (the sky, far mountains); (2) warm-advancing / cool-receding — warm hues feel closer/imminent, so warming toward the ground reads as "closing"; (3) red = danger — an association proposed to have a dual origin (Meier et al. 2012): biologically predisposed (primate threat displays) and reinforced by convention (stop signs, warning lights). The association is implicit (Pravossoudovitch et al. 2014) and pushes toward avoidance in evaluative situations (Meier et al. 2012). The sweep carries magnitude ("which band"), not exact value — trading exactness for immediacy and peripheral monitorability (§2.2: it remedies the decay). At breakoff (5,500) the sweep stops and switches to a flashing red/white/blue "police light-bar" — a multiply coded indication: pre-attentive flicker (involuntary attention capture — iconic urgency) + the learned police-emergency convention (arbitrary, inherited) + the abrupt steady→flash mode change marking the phase transition (iconic).

2.5 Form vs. role

An indication has a form — everything the dimensions so far describe: its span, encoding and motivation (§2.2–2.4) and its modality (Part III) — and a role, the task-structure element it serves (§1.2's phase/threshold distinction). Of the form's dimensions, span is the one that constrains role:

• Phase indication (role) — serves a phase: supports awareness through an altitude-extended activity.
• Threshold indication (role) — serves a threshold: marks an altitude-pinned action point.

A non-ranged indication can serve either role: the classic breakoff alarm serves a threshold, while periodic beeps — events every N feet, signaling no specific action — serve a phase, as exogenous re-triggers of the sampling cycle (§1.7). A ranged indication usually serves a phase: it occurs during some activity, even when that activity is setting up for a threshold (the final seconds of working time before breakoff, the base leg before landing); the urgency ramp of siren-type devices is such a case (Figure 17). The inverse is also possible: a threshold followed by a phase can be represented by a single indication. The breakoff light bar (§2.4) marks breakoff with its abrupt onset at 5,500 ft and then flashes on through the track to 4,000 ft: its onset is a point and its flashing a span, so the one signal takes both values of span, serving first a threshold and then a phase. Nor does a span always mean a phase: role follows what the span is intended to cover. The flatline at 300 ft (§2.4), sounded as a point, marks the turn to final; held down to about 200 ft, where the turn's tolerance runs out (Figure 5), it becomes ranged but still serves the threshold — a point with a span attached (§5.2); held to the ground, it would cover the final approach and serve a phase. Indications can also be jumper-sent: the wave-off ("a distinct wave-off to signal jumpers who may be above them," SIM 5-1.D) and an instructor's hand signals are non-ranged, symbolic and visual, and sent with intention, which makes them indications — a special and imprecise case: each carries only the sender's own estimate, so its error is the sender's error, which is one way one jumper's lapse can spread to others (§1.4). By contrast, what another jumper does without meaning to signal anything — tracking off, glancing at an altimeter — is not an indication but an ambient cue (§3.4).

2.6 The indication space

Indications are located by five largely independent dimensions (§2.1), laid out in Figure 18, with exceptions that all hinge on span: it constrains encoding (§2.3), iconic magnitude (§2.4) and role (§2.5). Only designed indications are located this way; the ambient cues of §3.4 fall outside the space — they are considered distinct from indication (§1.2).

1. Span (ranged / non-ranged, §2.2)
2. Encoding (direct — static / dynamic / event-marker — or symbolic, §2.3)
3. Motivation (motivated ↔ arbitrary, §2.4) — with its sub-axis, iconic target (magnitude / urgency / action)
4. Modality (visual / auditory / tactile) — with its parameters, Part III
5. Role (phase / threshold, §2.5)

The five differ in kind — span and role are binary, encoding and modality are unordered categories, and only motivation forms a true spectrum. Counted coarsely, with motivation reduced to its two poles, the space contains 5 × 2 × 3 × 2 = 60 cells. Span and encoding enter as a single factor, since only five of their eight pairings can exist (3 ranged, 2 non-ranged; §2.3); motivation (2 poles), modality (3) and role (2) combine freely with each. Divided as §2.4 divides it — the motivated pole by its three iconic targets, the arbitrary pole into inherited and local conventions — motivation brings the count to 138, as one further constraint appears; Appendix C lays out both counts.

The product landscape and the indication space. If we restrict the modality dimension to visual and auditory (Part III) and encoding to its two base groups direct vs. symbolic (§2.3), the space projects onto a smaller 2×2 matrix whose cells accurately reflect the categories of products the industry produces and sells: visual-symbolic — the numeric readout; visual-direct — the peripheral-vision light; auditory-symbolic — the speaking altimeter; auditory-direct — the beeping audible. Hybrids occupy multiple cells: the analog dial and WatchAlti each combine visual-direct and visual-symbolic (Figure 15). There are not yet any successful tactile products, and the Part III preamble discusses the primary reason why (the freefall masking floor, §3.1). A terminological consequence follows for this document and beyond it: these four cells are indication types, not modalities. An indication type is any location in the space, coarse or fine; these four are the coarsest the taxonomy affords (the full space gives the finest). "Modality" is reserved for the sense (Part III). So an altimeter like MetaAlti does not offer three modalities but three indication types — peripheral-vision light, speaking, and beeping audible — across two modalities (auditory and visual). Freefall Data Systems' own product categories — Numeric, Speaking, Audible, Peripheral-Vision LED — predate this framework and name exactly these four cells.

Part III — Sensory Modalities and the Jumper

This Part covers the sensory modalities through which indications reach the jumper and, in §3.4, how the same senses can ascertain altitude on their own, independent of any hardware. The sense an indication addresses forms the fourth dimension of altitude indication outlined in Figure 18 (§2.6). The sense of altitude has to be trained, as we do not have any sort of dependable biological altimeter. This is not to say there are no differences among people in their natural ability to estimate altitude; in that respect it resembles absolute pitch (naming or producing a specific pitch without any reference): at some point one has to be taught a reference pitch, for example, what 440 Hz sounds like. Technological and non-technological means often complement one another in practice — the former trains the latter through association, as when a speaking altimeter announces altitudes while a tandem instructor, jump after jump, imprints a series of altitude pictures in their mind. In some cases an altimeter can override or stand in for the senses, or vice versa: an altimeter can correct a misread altitude picture at a new DZ; a jumper may be made aware of the breakoff threshold not by their altimeter, but simply by others beginning to break off. Technological and non-technological means of altitude awareness share sensory channels and can therefore interfere with one another as well (e.g., the ground altitude picture competes for attention with a visual altitude display).

Altitude indication is the display of altitude through a sensory modality. The two primary modalities are:

• Visual modality → visualization of altitude (data-visualization / visual-display research).
• Auditory modality → sonification of altitude (auditory-display / sonification research — ICAD, The Sonification Handbook). Precisely: sonification is non-speech audio conveying information; auditory display is the umbrella that also covers speech (Kramer et al. 1999; Walker & Nees 2011). So an audible altimeter's tones are sonification; a speaking altimeter is auditory display but not sonification.

A third modality, tactile, is included in our taxonomy — in principle a vibrotactile display can occupy every cell in our space (Part II) — though its usefulness depends on the stage of the skydive and other factors discussed below:

• Tactile modality → haptification of altitude (haptic display / haptics research).

In practice, in freefall the jumper's entire body experiences intense vibration — turbulent airflow and clothing flutter, whose strength scales with dynamic pressure — on top of a steady pressure load on the skin. This masks most tactile input and drastically limits the modality's utility. One tactile signal known to cut through the noise is an AFF instructor's shake, as it clears the floor in amplitude and is of lower frequency, though even it can fail.

During ascent or canopy flight, the vibration floor is much lower and vibrations can be felt, though sensitivity is limited by airspeed under canopy or by the aircraft's own vibration on ascent (a Cessna 182, for example). Vibrotactile indication is plausibly viable in these stages, though unlikely to match the salience of auditory or visual indications unless very high in amplitude (to a startling or annoying extent).

In theory, it may be possible to create electrotactile indications — electrical stimulation of the nerves in the skin — that could be felt in freefall (Kaczmarek et al. 1991; Zhou et al. 2022). How salient even this would be in freefall is an open question with significant obstacles. The working consensus in the sport seems to be that tactile indication is more trouble than it is worth.

The remaining senses — taste, smell, and the vestibular sense, which registers acceleration but not altitude — cannot seriously be considered for the purposes of altitude indication.

3.1 The sensory environment by stage

Each stage of a skydive differs markedly in character and brings unique challenges to the sensory environment. In every stage, visibility depends on available light — it is harder to ascertain one's altitude at night.

Ascent. A skydiver has the best chance to attend to the senses during the ascent stage; it is the only stage in which someone else is flying, and the jumper is generally only tasked with being attentive and responsive to safety-related issues. The sensory environment in this stage varies dramatically with the aircraft. With few exceptions, ascent is a relatively loud stage of the skydive — in some cases even nearing the volume of freefall. An exception would be a hot air balloon, which during parts of ascent is nearly silent. The sounds of ascent can mask the sounds produced by audible altimeters (§3.3), depending on helmet and earplugs. Visibility of the ground is affected by seating location, the airframe, window quality, whether the door is open, and aircraft attitude; visual indications themselves, however, are generally easy to see on ascent. Depending on the aircraft used, vibrations while climbing can be significant as pilots often use full power to get to altitude as fast as possible; this vibration can mask tactile indications. Cabin events can re-trigger one's attention to altitude (§1.7).

Freefall. The greatest sensory load is experienced during freefall: the broadband noise of wind above 100 mph, the sight of open earth all around, the intense vibrations produced by turbulence. All of these things and more can contribute to sensory overload and cause a failed sample (§1.7): looking at an altimeter without registering what it says. Wind noise depends on equipment (helmet and earplugs) as well as discipline and head position, and can be useful as a cue to fall rate (or airspeed in wingsuit flight), but it is notorious in the sport for masking audible indications. The brightness at altitude — sometimes above white, reflective clouds — or the sun itself can likewise mask visual indications or reduce their salience. In some cases, body position can additionally compromise the visibility of an altimeter. As noted above, tactile input is of relatively little use in this stage. Any sort of activity during this stage can be distracting enough to draw attention away from altitude.

Canopy. The canopy flown and its wing loading have great influence on the sensory environment — at the extreme, the high-performance canopies of XRW (extreme relative work — canopy pilots flying with wingsuiters), where canopy flight and freefall overlap. In general, however, canopy flight is the quietest stage. Audible alarms can often be set to a lower volume with equivalent effect, though airspeed will affect their audibility. In addition to the same issues of brightness and the sun, canopy control checks, the landing area, other canopies, and the traffic pattern all compete with visual indications for attention. Tactile input is similarly limited by vibration in this stage, though less so than in freefall.

3.2 Cognitive load

A jumper's cognitive load can have a significant influence on their sensory perception. Attention is not a single pool but divides into separate resources by modality (visual / auditory), code (verbal / spatial), and processing stage (perception / response); two tasks interfere more the more resources they share (Wickens 2002, 2008). Under canopy, a student taking radio instructions from the ground is already listening to speech, so a spoken altitude competes with the instructor's voice for both the auditory and the verbal resource; a beep shares only the modality, and a visual display shares neither. High load and stress can cause attention to narrow and peripheral input to drop out (Easterbrook 1959), which can lead to task saturation and a failure of the sampling cycle (§1.7). Some of this load is intrinsic to the skydive and inevitable (with exposure and experience, its intensity wanes); part of it is extraneous, created by the altitude indication. For example, the decoding step of symbolic encoding can add what cognitive load theory calls extraneous load — load imposed by the form of the presentation rather than by its content (Sweller 1988, 1994; Sweller, van Merriënboer & Paas 1998). Good indication design keeps extraneous load low and takes into consideration experience level (practice makes decoding nearly automatic), discipline, and natural ability to ascertain altitude.

3.3 Failure modes across modalities

The mechanisms by which an indication fails to register are shared across modalities; what differs by modality is the source — masking, for instance, comes from glare or ambient brightness level for the eyes, wind noise for the ears, and buffeting for the skin. These are all failures of the registration layer of §2.4, before any question of interpretation arises: in situation-awareness terms, Level 1 errors, failures to perceive (Endsley 1995). The one exception is startle, the last row of Figure 20: there the onset itself disrupts the response, not the registration.

Sensory masking vs. attentional failure. While sensory masking and attentional failure may have similar outcomes — a jumper not reacting — they have different causes and remedies. Masking is driven by the environment: the signal fails to clear the noise floor (§3.1). Sinus pressure can affect hearing as well, unevenly across frequencies; this is not masking in the strict sense — it filters signal and noise alike — but it can push an already faint signal below audibility. Attentional failure is driven by the jumper's cognitive load (§3.2): the signal physically reaches the jumper but the brain never brings it into awareness (Lavie 2005). The 2023–2024 SIM mentioned both: "Extreme background noise of freefall and a jumper's attention to another event can render audible altimeters ineffective" (SIM 2023–2024, 5-3.J.4.c.(2)) — and elsewhere warned against "focusing on an unimportant goal and losing track of the more important" (SIM 2023–2024, 6-2.E.2.e). The 2025 rewrite (carried into the 2026 edition) drops the general warning, narrowing it to pertain to the fastest disciplines: in freeflying, "audible altimeters may also be hard to hear with the increased wind speed of the vertical position" (SIM 5-2.E); in speed skydiving, "the wind noise … may drown out the audible beeps," so flashing visuals inside the helmet are recommended (SIM 5-11.C), and "issues with sinuses can severely hinder … the ability to hear audible altimeters" (SIM 5-11.E).

Masking can be remedied to a point by amplifying the signal, while attentional failure is harder to remedy. In the author's experience, jumpers often report "not hearing" an audible that in all likelihood cleared the noise floor. The same failure has been measured in pilots: in a flight simulator, 11 of 28 general-aviation pilots (39.3%) missed an alarm during a windshear landing that every pilot heard on a normal landing (Dehais et al. 2014). Acute stress can produce what the police literature calls auditory exclusion: officers in shootings commonly report their own gunfire as muffled or inaudible — 84% of the 157 officers Artwohl (2002) surveyed reported diminished sound, and in Klinger's (2001) interviews auditory blunting was reported in 70% of 113 shootings at the moment of firing — based on self-report, but consistent with load-induced inattentional deafness (Raveh & Lavie 2015).

For attentional failure, amplifying the signal does not always help; what helps is whatever wins the competition for attention: a distinctive signal with a crisp attack or a harsh timbre (promising recent evidence: Riedinger et al. 2026), personal relevance (one's own name can break through even unattended speech: the own-name effect; Moray 1959; Wood & Cowan 1995), repetition that offers several chances, timing away from peak-load moments, and a backup in another modality (a partial remedy; see below). A third possibility is that the signal was perceived but never reached memory — inattentional amnesia (Wolfe 1999) — so the jumper cannot report it, and still did not act.

Two layers exist beyond registration. Misinterpretation of what an indication represents occurs in a jumper's mind, but is significantly affected by encoding methods (design decisions), not modality (§2.4): a motivated or inherited code is understood once it registers, while a local convention depends on training and on recall under load (§3.2). Response failure — the signal registered and was understood, but the action did not follow (freeze, task fixation; §1.7) — belongs to the jumper's state rather than to the modality. One failure level precedes them all: incapacitation (§1.5 — collision, a hard enough riser strike, a medical emergency) removes the perceiver, so no indication in any modality can help; this is a good reason to use an AAD.

Redundancy and common-mode failure. Two indications in the same modality share every failure in that modality's column of Figure 20: two audibles, one per ear, are masked by the same wind and habituated by the same repetition; two FlexiLights (peripheral-vision flexible gooseneck lights), one per eye, are masked by the same ambient light and habituate alike. That is a common-mode failure: one mechanism takes out both. Pairing across modalities (a light and a tone) means that no single mechanism defeats both. This makes a case for cross-modal redundancy (§5.4), though with a limitation: attentional failure is partly common-mode across modalities, since a jumper saturated by the task can miss the light and the tone together — visual load alone measurably dulls hearing (Raveh & Lavie 2015). Cross-modal redundancy remedies the environment, but only lower cognitive load, or a signal that is able to gain the jumper's attention, defeats task saturation.

3.4 Non-instrument altitude information: ambient cues and the internal clock

Long before skydivers had instruments, they had to rely on ambient cues (§1.2) and internal processes. Unlike today, these could not be refined by repeated comparison to an indication, but had to be developed on their own, over time and through experience. Many of the following — those that depend on vision — are in part or entirely dependent on available light. Three pieces of information are noted along with each cue: its source (ambient, found in the environment, or endogenous, from within the jumper), the sense that picks it up, and whether it is always on or conditional (on light, weather or circumstance).

• The internal clock (endogenous, temporal; always on). Perhaps the most significant of all for experienced jumpers is a natural sense of timing that develops over repeated jumps. It is not a particularly safe method, because the altitude lost in a given time varies with fall rate; but it nonetheless becomes an important and undeniable tool. Predictable jumps like tandems in particular acquire a rhythm of what happens when — one even becomes accustomed to what a jump from 18,000 ft "feels like" temporally compared with one from 13,000. The internal clock also supplies the elapsed time for the dead-reckoning propagator (§1.6) that holds a recent altimeter reading alive between checks — a jumper extrapolates the last known altitude forward at a perceived fall rate. This is what produces the freshness plateau in the sampling cycle and what keeps a fresh reading usable. The internal clock can be compromised in its accuracy — even causing confident inaccurate estimations — by unfamiliar exit altitudes, fall rates, and disciplines. Skill at time estimation varies between individuals and improves with practice. Fear also distorts time judgments, most clearly in recollection (Grondin 2010; §1.6), so the internal clock is least trustworthy for the jumpers who most need to develop and rely on it.

• Visual ground detail and appearance (ambient, exteroceptive; conditional). Another survival skill, and one that ranks very high, is the ability to ascertain altitude simply from the way the ground looks — through the altitude picture (§1.2) alone. The ability is refined through repeated comparison to altimeter indications. The 2023–2024 SIM taught it as a drill: "a. Looking at the ground. b. Looking at the altimeter and note the altitude. c. Repeat this procedure several times per jump to develop the ability to eyeball the altitude," and it ranked the two: ground and time were "a primary reference," while "altimeters provide excellent secondary references for developing and verifying primary altitude-recognition skills" (SIM 2023–2024, 5-3.J.3.a–b, J.5). The indication thus trains the ambient cue that later replaces it to an extent. The faculty is calibrated to a training set — a home DZ, for example — and detunes when jumping in circumstances that contrast with the training set. Orientation is one such circumstance: "Head-down and sit-fly positions present a different visual picture of the earth; freeflyers may not be visually aware of their altitude" (SIM 2023–2024, 6-2.E.2.b). This is a skill in itself, like absolute pitch (Part III preamble), but with one added challenge: the ground is continuously moving — the picture is learned in motion, and read in motion. Both the rate at which the ground approaches (closing rate, with an assumed fall rate) and the speed over the ground (while spotting before exit or in wingsuit flight, for example) affect the use of terrain appearance as an indicator of altitude. A jumper exiting a jet (Skydive Perris' DC-9, for example), who sees the ground streaming past considerably faster than usual, carries a silently assumed ground speed from previous experience and may read the altitude picture as slightly lower than it is. As altitude decreases, the ground resolves into finer features (fields → roads → vehicles → people), and their angular size and its rate of change grow. Below roughly 3,000 ft comes ground rush — the accelerating optical expansion of the ground, negligible higher up and overwhelming below ~1,000 ft (Figure 21). Ground rush works differently from altitude picture: it is tau (τ), time-to-contact specified directly by the rate of optical expansion, which is familiar from the horizontal domain in everyday life (braking a car, for example). Tau needs no assumed speed — but for the same reason it carries no altitude information, only time remaining; converting it to altitude would require the same fall-rate assumption the altitude picture in part relies on (Gibson 1979; Lee 1976; Lee & Reddish 1981). The visual cue is therefore nonlinear and weaker higher up, where deployment decisions are made, and strongest where the time to act is dwindling.

• Light gradient (ambient, exteroceptive; conditional). The view below darkens with descent as the haze between jumper and ground thins or once clouds are passed — less scattered light veiling the terrain (aerial perspective, cf. §2.4) — most noticeably on hazy or cloudy days or in mountainous regions. (Geometry contributes less than one might expect: the horizon sits only ≈2° below eye level at 13,000 ft and ≈0° on the ground, so in a flat area, the ground's share of the visual field barely changes during descent.) An interesting case is twilight: at dusk or dawn the ground can be completely shadowed, the sun already set or not yet risen there, while at altitude the sun is still up.

• Temperature (ambient, somatic; always on). Lapse rate ≈ 2°C per 1,000 ft; descent is felt as warming, particularly on winter jumps from high exit altitudes — slow and coarse, but pervasive. On ascent, dropping temperature can be a clear cue to increasing altitude — generally more noticeable than during freefall or canopy flight. Temperature inversions — warm air above cold air — can draw a jumper's attention to specific altitude regions, as can the sudden drop-off at a marine layer.

• Barometric physiology (ambient, interoceptive; always on). Ears equalize at a certain rate and the sinuses produce characteristic sensations as ambient pressure rises in descent or falls on ascent — it is a grossly inaccurate way to determine altitude, but genuinely altitude-coupled. High-altitude flatus expulsion (HAFE) — the expulsion of intestinal gas that expands as pressure falls, from around 11,000 ft MSL (Auerbach & Miller 1981) — is a phenomenon familiar to every experienced tandem instructor.

• Cloud and smoke layers (ambient, exteroceptive; conditional). Cloud bases and layers at known altitudes serve as fixed markers when present and are often noted by skydivers during ascent — the 2023–2024 SIM taught: "Observe the cloud bases on the ride to altitude to use later as an altitude reference" (SIM 2023–2024, ISP Category C). These tend to shift throughout the day; the cloud base, for example, can rise with temperature. Smoke layers in fire-prone areas work the same way — often trapped under a temperature inversion at a fixed altitude — and are unique in that they can also be sensed by smell: on a smoky day, the smell of smoke can mark the top of the layer (e.g., beginning around 6,000 ft).

• Social and situational cues (ambient, exteroceptive; conditional). Certain activities or situations have a tendency to call people's attention to altitude and re-trigger the sampling cycle: other jumpers taking off their seat belts at 1,000–1,500 ft, the door opening for a low pass, another jumper checking their altimeter, a group breaking off. The hazard of the shared reference is mentioned in the SIM: the camera flyer "needs to maintain independent altitude awareness and never rely on the student or instructor(s)" (SIM 2025 and 2026, 5-7.F).

In general the following characteristics are true for altitude estimation based on ambient cues: low precision, condition-dependent availability, uncalibrated and drifting, mostly unavailable or distorted under stress — yet free, unpowered, and in some cases always on.

Notably, the 2025 rewrite (and the 2026 edition after it) drops nearly all of the 2023–2024 edition's guidance on training the eye — the ground-and-altimeter cross-checks on the climb and after deployment, cloud bases as altitude references, and the ranking of ground and time as primary references with the altimeter secondary (SIM 2023–2024, ISP Category C; 5-3.J). Not everything is discarded, however; the internal clock is utilized to a limited extent: static-line and IAD students still "count out loud by thousands to 5,000, then check the parachute" (SIM 2025 and 2026, ISP Category A). In freefall, the SIM now emphasizes instruments for altitude awareness.

What kind of sign each cue is. Here we mean sign in the technical sense of semiotics (§2.4): anything that stands for something beyond itself — in this case, altitude — to someone able to read it. Most of these cues are indices — tied to altitude by cause, not by resemblance or convention — but not all in the same way. Ground detail, the light gradient, temperature, and ear pressure are pure indices: each is directly caused by altitude, and reading it takes practice but no knowledge particular to the jump. Cloud layers and smoke layers must be read with knowledge of the day's conditions: a cloud base marks an altitude only if one knows where the base is. Social cues rest on knowing what jumpers do at what altitude: another jumper's opening canopy means "pull altitude" only because one knows the altitude they intended to pull at. The internal clock is not a sign at all: nothing outside the jumper points to the altitude; it is dead reckoning, an inference from elapsed time and fall rate.

Part IV — Imported Theory: Cockpit Alerting

Skydivers and aircraft pilots face an overlapping set of problems, and cockpit alerting systems are the closest relatives of what we design for skydiving altitude indication. It is therefore useful to import concepts from the theory and taxonomy of such systems, which have developed over more than a half century. The literature's focus lies primarily on the airliner cockpit, which has the most highly developed systems of all aircraft for such purposes; however, rotorcraft have systems of their own (14 CFR 27/29.1322; helicopter terrain awareness and warning systems, HTAWS) and problems that overlap with those of skydiving (noise, single-pilot operation). Military aircraft add a further layer: systems for threats — radar tracking, missile lock, detected launches — and for offensive operations (the AIM-9 Sidewinder's seeker tone, sent to the pilot's headset, sonifies the lock-on sequence). Where the mapping to skydiving breaks is of as much interest as where it is seamless. We can also import some concepts from medical alarms (alarm fatigue) and industrial alarm management (alarm rationalization) to supplement what aviation has not addressed.

4.1 The alert hierarchy

The tiers. Airline pilots deal with a large number of independent systems — far more than skydivers do — each with its own characteristics and failure points. To help pilots prioritize in such an environment, cockpit alerts are tiered. Although the exact appearance of the alerts, sounds of the alarms, and signatures of the tactile alerts are not strictly defined, FAA regulations require that warnings and cautions be given specific colors (red for warnings, amber or yellow for cautions) and that both, unlike the lowest tier (advisories), target at least two senses, in any combination of visual, auditory, or tactile (14 CFR 25.1322; FAA AC 25.1322-1; EASA CS 25.1322).

• Standardization through regulation, certification, and training. A pilot does not assign meaning to any indication; rather, the meanings are inherited conventions, standardized across aircraft and internalized through training.
• Mapping to skydiving. There is no direct mapping of this system to skydiving. This is partly because we have no comparable warnings about our gear, which account for a large share of what pilots see; imagine, for example, a warning such as "Pilot chute coming loose!" as a jumper is sitflying — this would certainly justify the equivalent of a master warning. A rare exception — an indication of the gear's state rather than of a problem — is Airtec's Wing Suit CYPRES, an AAD with two sets of activation criteria, one for wingsuit and one for canopy flight: once it has registered more than ten seconds of descent at canopy speed, an audio unit in the helmet plays a melodic sound for ten seconds to announce the switch to canopy mode, and only then are aggressive canopy maneuvers safe (Airtec 2025, §2.6; now produced only in small quantities). Overlap does exist where an indication tells a jumper they are somewhere they should not be — as a pilot can be — such as having descended through the intended deployment altitude without pulling (a warning-tier event). A parallel caution-tier event in skydiving could be an indication to break off from a group.
• Two differences. In skydiving, there is no standardization: a jumper does indeed assign meaning to a chosen indication. And in skydiving the response to any warning- or caution-tier event is ideally immediate, whereas in cockpit alerting only a warning demands immediate action (often from memory); a caution demands immediate awareness but only a subsequent response — typically the relevant non-normal checklist (FAA AC 25.1322-1).

4.2 Urgency mapping

Urgency is the one iconic target of motivation (§2.4) with a mature design science behind it, in cockpit alerting.

• Patterson's guidelines. In 1982, Roy Patterson produced a landmark study and design guide for auditory warning systems in aviation, drawing on — and extending — psychoacoustics. In it, he stresses the ergonomics of the warnings themselves. Warning sounds are built from pulses grouped into bursts, and he describes their spectral content and how they map to perceived urgency. Pulse onsets (attacks) are ramped over 20–30 ms to reduce the startle effect — existing flight-deck warnings went from silence to over 100 dB in under 10 ms — and for the same reason, bursts begin on a quieter pulse and build. A spectrum of several harmonically related components (numbers below, under Spectral placement) adds redundancy, so that no single spurious noise source can mask the whole sound. Finally, each warning takes three forms: an initial burst at moderate urgency; a quieter and slower background form repeated every few seconds so the crew can still talk; and an urgent form (higher, faster, louder) that returns only if the condition goes unattended, with a voice message included to carry the meaning. The pulse is specified in detail; the burst is defined only vaguely — "four, clustered pulses followed by two, irregularly-spaced pulses" in his prototype, the precise gaps to be "established empirically for each new warning." The Sea King report, which applied the guidelines to a helicopter (§4.5), is much more detailed: it specifies each of its warnings pulse by pulse — spectrum, envelope, timing, pitch and level (Lower et al. 1986, Figs. 7–20). Patterson's work was foundational in standardizing the systems used today; at the time, systems on aircraft were disparate and suboptimal (Patterson 1982; Patterson 1990; Lower et al. 1986).
• Two quantitative rules from Patterson's guide recur throughout this document. First, level is set relative to masking: warning indications should sit roughly 15 dB above their masked threshold in the ambient noise, but not more than about 25 dB above it — loud enough to be heard, but not so loud as to startle, distract, or disrupt. Second, the number of distinct warnings is limited to four to six immediate-action sounds, because in a learning study naive listeners acquired the first four to six sounds of a set quickly and the rate of acquisition slowed markedly beyond six; Patterson explicitly declines to transfer the figure to pilots directly, but recommends that the number of immediate-action warnings "should not exceed about six." The application of these principles to auditory indication design is discussed further in Part V (Patterson 1982; Patterson 1990; Lower et al. 1986).
• Spectral placement. Patterson recommends placing tones where the masked threshold is lowest: spectral components in the range 0.5–5 kHz, fundamentals between 150 Hz and 1 kHz, with at least four main components between 1 and 4 kHz, harmonically related so that they fuse into a single pitch. For helicopters the window was narrowed to 0.5–3.2 kHz because the helmet earphones rolled off above that. Skydiving has three stages with differing noise floors and masking characteristics, so different spectral placement and content are necessary for different stages (the design consequences are drawn in §5.5; Patterson 1982, via Lower et al. 1986).
• Perceived urgency is a designable psychoacoustic quality. Edworthy, Hellier, and colleagues quantified how perceived urgency scales with pulse rate (speed), fundamental frequency, harmonic structure (inharmonic spectra come across as more urgent), amplitude envelope and onset, rhythm, and pitch contour — with speed being the dominant parameter — and fit the relations as power functions, so that a target urgency can be specified (Edworthy, Loxley & Dennis 1991; Hellier, Edworthy & Dennis 1993).
• The mapping is cross-modal. Two later literatures supply the visual and tactile counterparts to the work in the auditory domain. Warning-label research rated the perceived hazard of colors and signal words (red highest, with orange and yellow below it and not reliably told apart; DANGER > WARNING > CAUTION) — empirical backing for the tier colors of §4.1 (Chapanis 1994; Braun & Silver 1995; Wogalter 2006). Automotive display research then mapped urgency across auditory, visual, and tactile warnings and found that rate — pulse, flash, vibration — carries urgency in every modality (Baldwin & Lewis 2014).
• Perceived urgency should correspond to situational urgency (Edworthy & Adams 1996). A signal without appropriate urgency for a critical event can be missed, or the reaction to it can be too slow; an over-urgent signal for a routine event can startle, distract, or annoy — and annoyance is what gets an alarm muted or disabled (§5.6). The principle is stated as a design rule in §5.6.
• Urgency and annoyance are correlated but separable. The parameters that raise perceived urgency also raise annoyance, but the relationship is not one-to-one: some parameters buy urgency more cheaply than others, and indication design should take that into account (§5.6; Marshall, Lee & Austria 2007).
• Uniqueness and discriminability. Warnings that are acoustically similar can be confused, and confusion rises with the number of warnings in the set. Patterson's learning study found that warnings sharing the same pulse-repetition rate were confused even when their spectra differed grossly — the shared rhythm outweighed the difference in timbre — and later warning sets that gave each warning a distinctive temporal pattern were confused far less (Patterson 1982; Patterson 1990; Lower et al. 1986). For audible alarms in skydiving, it becomes an important design principle that the jumper can discern which alarm has sounded.
• Arbitrary versus motivated approaches (§2.4): earcons and auditory icons. The international medical-alarm standard IEC 60601-1-8 (from the International Electrotechnical Commission; first edition 2003) offered, in an optional annex, one abstract melody per alarm category: earcons designed by a committee. Studies found the melodies hard to learn and easy to confuse, and a 2020 amendment added auditory-icon alarms built on Edworthy's later work (a heartbeat-like sound for cardiac, and so on) as an alternative set, the melodies remaining in the standard. It is a documented case of a motivated approach being brought in where an arbitrary one had failed (IEC 60601-1-8:2006+AMD2:2020; Sanderson, Wee & Lacherez 2006; Edworthy et al. 2017).

4.3 The cry-wolf effect

• The effect. Frequent exposure to false alarms or alarms that require no action leads to such alarms increasingly being ignored. This has been documented both in laboratory conditions and in practice. In the laboratory, people answer an alarm about as often as it is true: subjects told that only half of the alarms interrupting their task were genuine responded to about half of them, and the match held at 25% and 75% as well (Bliss, Gilson & Deaton 1995, "probability matching"). Sorkin's 1988 note "Why are people turning off our alarms?" collected cases in which warning systems had been disabled before an accident — cab warnings taped over before the 1987 Conrail–Amtrak collision, the electronic-warfare audible switched off aboard the USS Stark, the takeoff-warning system in the Northwest 255 crash, nuclear-plant operators silencing alarms that "had become 'old friends'" — and gave two reasons: alarms that are aversive or interfere with the task, and false-alarm rates perceived as excessive. For the second he quoted the nuclear-plant study's diagnosis: "The net results of the many false alarms is a 'cry-wolf' syndrome which leads to a lack of faith in the system and a casual attitude toward the constant presence of certain alarms" (Seminara et al. 1977, quoted in Sorkin 1988). Breznitz's Cry Wolf (1984) is a book-length treatment of the false-alarm effect, theory and experiments, and Parasuraman and Riley (1997) called the behavior disuse — neglect of an automated aid because of its false-alarm rate — in their taxonomy of automation use and misuse.
• The canonical case: GPWS → EGPWS. First-generation ground-proximity warning systems (mandatory on large aircraft from 1974) looked down with a radio altimeter, not at the terrain ahead: over a sharp rise in terrain they could not detect the closure until it was too late for evasive action (IATA 2016), and older units were "plagued with nuisance and false alarms" (FAA 2000). The regulator drew the cry-wolf conclusion explicitly — "if an aural cockpit alarm sounds too often as a false alarm, the flight crew will either begin to ignore it or will be tempted to disable the system" — and controlled-flight-into-terrain accidents (CFIT: an airworthy aircraft under full control flown into the ground because the crew did not know where the terrain was, aviation's altitude-awareness lapse) continued to occur in GPWS-equipped aircraft (FAA 2000). The enhanced system, EGPWS, introduced in the mid-1990s, compared the aircraft's GPS position with a terrain database — a forward-looking view — giving earlier warning, and the FAA's 2000 rule made such terrain awareness and warning systems (TAWS) mandatory, requiring alerting times long enough to act on but otherwise "as short as possible," to minimize nuisance alarms (IATA 2016; FAA 2000). Two lessons can be learned from this: reliability restores response, and warning time matters independently — a true warning that arrives too late is as useless as a false one.
• Medical alarms: alarm fatigue. By the Joint Commission's estimate, 85–99% of hospital alarms do not require clinical intervention, and alarm fatigue is a recognized patient-safety hazard (Joint Commission 2013; Edworthy 2013); in one study of 461 intensive-care patients, 88.8% of annotated arrhythmia alarms were false (Drew et al. 2014).
• Industrial alarm management: rationalization. The process-control standards used in many tightly controlled factories and chemical plants define an alarm as a signal that requires an operator response. In a review called alarm rationalization, every alarm must be justified against that definition, and alarms that fail are removed or given a lower priority. The standards also set quantifiable load benchmarks — on the order of one or two alarms per operator per ten minutes in steady operation (ANSI/ISA-18.2-2016; EEMUA Publication 191).
• The cancel reflex. A different path to not responding was documented by Patterson: the alarm registers, is believed, but the crew's first act is nonetheless to silence it. He had reported it as early as 1982: when warnings are expected, "flight crew may be detailed to sit with a finger poised over the cancel button so that the warning can be terminated the instant it sounds" (Patterson 1982, quoted in Sorkin 1988). Patterson opens his 1990 paper with a confidential incident report of a night flight in which a stall warning, a stick shaker, and several warning lights were all present at once: "I was frightened numb for several seconds and drawn off instruments trying to work out how to cancel the audio/visual assault rather than taking what should be instinctive actions… action is invariably taken to cancel the cacophony before getting on with the actual problem." The 1986 report that carried Patterson's guidelines into the Sea King helicopter (Lower et al. 1986; §4.5) says the same in a sentence: "Often the crew's attention will be directed, not to the real problem, but to the problem of finding the warning-cancel button." An alarm loud or persistent enough to be intolerable converts the response it was meant to trigger into a response to the alarm itself. The literature's own countermeasure is the fall-back structure of §4.2 — sound once with urgency, recede to a background form that permits work, return only if unattended — so that the alarm never has to be canceled to get on with the task. The equivalent of silencing or canceling an active indication is not generally possible while skydiving; the closest equivalent is the removal of alarms or indications between jumps, or the removal of an altimeter entirely (Patterson 1990, pp. 485, 489; Lower et al. 1986).
• Urgency inflation. In listing the failings of the airliner warnings its new designs were meant to replace, the Sea King report includes: "some of the less important warnings even sound more urgent than the more important warnings." When indications for routine events are artificially amplified in their urgency, the cry-wolf effect is transferred from reliability to urgency itself, making the hierarchy discussed in §4.1 more or less meaningless (see §5.6). A jumper who has every indication set to maximum urgency has no hierarchy at all (Lower et al. 1986).
• Limiting the sounds. Patterson's "about six" (§4.2) became a budget in practice. The Sea King project set out to give every priority-one warning its own sound and found it "impracticable" across helicopter types — "far too many different warning sounds would be needed" — so it grouped most priority-one warnings under a single sound, all priority-two under a master caution, all priority-three under one more, and let the voice message that followed convey the specific meaning (§4.4): ten to twelve sounds for the whole military fleet, six to eight in any one aircraft, "of which three would be the general priority one, two and three sounds." Patterson's hospital work reached the same compromise — three category sounds plus a handful of specific urgent ones (Lower et al. 1986; Patterson 1990, p. 491).
• The skydiving analog (applied in §5.7). Skydiving alarms are practically never false in the sense they are in a cockpit — they fire at an altitude set in advance, not on a condition the device has to detect (a descent-rate alarm, §4.6, is the exception), and the altitude always arrives (§1.6) — so they can only be early or late, set incorrectly, or accurate but non-actionable: sounding at a point of the skydive that demands no action. It is through this last kind that the cry-wolf mechanism enters.
• Three ways not to respond. Cry-wolf is learned mistrust: the alarm registered, was judged unreliable or empty, and was ignored. Inattentional deafness (§3.3) is non-registration: it never reached awareness. The cancel reflex (above) is a third: registered and believed, but answered by silencing. All three end in not acting, and they have different remedies — reliability and pruning for the first, load and salience for the second, and level, persistence, and the fall-back structure for the third.

4.4 Voice versus synthesized tones

• Cockpit alerting's design. A cockpit warning commonly pairs an attention-getting sound composed of synthesized tones with a voice message that names the problem and the action. In the Sea King design, a voice message follows every burst of sound (Lower et al. 1986, Fig. 2; §4.3). Time-critical warnings receive distinctive voice alerts (e.g., "PULL UP," "TERRAIN," "WINDSHEAR," "TRAFFIC"; FAA AC 25.1322-1). The tone is intended to capture attention and the voice then directs action — though an early simulator study found that an alerting tone placed before a synthesized voice warning slowed pilots' responses and added little (Simpson & Williams 1980).
• The pros of voiced indications. Voice has the shortest learning curve and the least ambiguity, as it is a pre-learned, inherited convention (§2.4): a spoken "pull" cannot be mistaken for anything else.
• The cons. In aviation — and especially in skydiving — synthesized alarms can cut through the noise far more effectively than speech. A word or short phrase also often takes longer to deliver than an equivalent alarm composed of synthesized tones. In freefall, we have only seconds to react before a significant amount of altitude is lost. Speech also occupies the verbal channel and so competes with verbal task content (§3.2). Less important but also relevant: voiced indications are language-dependent.
• Urgency in the delivery. Both words and the way they are spoken carry urgency: "danger" outranks "caution," and a speaking style intended to sound urgent is — not surprisingly — rated more urgent than a flat one. A spoken indication's urgency is therefore designable in its delivery, not only its content (see §5.6; Hellier et al. 2002).
• Transfer to speaking audibles. In skydiving, even a caution-tier event must be handled within an altitude budget, which at freefall speeds amounts to only seconds — far less than aviation allows. Thus the two-stage structure (attention, then message) does not apply as cleanly as one might hope. The reverse division might be argued for skydiving: the most urgent events get the most urgent-sounding indications — synthesized, not voiced — and voice is reserved for general altitude information or where there is time to act. On the other hand, voiced indications are the least ambiguous (above), so an argument can be made for them as well (§5.6).

4.5 Rotary-wing aircraft

• Similarities and contrasts. Although there are some similarities for rotary-wing aircraft — air-ambulance helicopters are often flown single-pilot, the pilot wears a helmet in an environment with a high noise floor, and in low-level work terrain is seconds away — larger contrasts remain: a seated pilot with a visual panel, warnings through earphones at a controlled level, and alerts that are indications of potential issues rather than scheduled altitudes (§1.6).
• The Sea King re-derivation. The 1986 Sea King project is the one documented instance of Patterson's guidelines being re-derived for a new noise environment. The rules held and the numbers did not — the level floor became "15 dB above masked threshold or 70 dB SPL, whichever is higher," the frequency window shrank to the earphones' bandwidth, and a priority-one burst had to complete in 1.4 s — a figure derived from a temporal response budget. The pilot should begin responding within about three seconds. Part of that time goes to confirming the warning on the panel, and the burst should be able to play twice if necessary before the three seconds are up. So the initial burst must be recognizable within about a second and finished within 1.4 s, which, at the pulse durations and spacings the Sea King sounds used, meant at most nine pulses per burst (Lower et al. 1986).
• A ranged altitude tone. The same report describes a ranged auditory altitude indication, the military radio-altimeter tone that continuously tells the pilot whether the aircraft is at, above, or below a set height, because at low level "the need to glance away continually at the altimeter would be dangerous"; and the 1990 paper that reported the work also recounts an altitude accident — a helicopter flown into the sea in fog while its crew never saw the flashing low-altitude warning light, a registration failure (§3.3) that an auditory indication "would almost undoubtedly have averted" (Lower et al. 1986; Patterson 1990, citing Cooper 1984).

4.6 What does not transfer

• The environment. The noise floor in freefall differs significantly from that experienced in a cockpit (wind noise at the ear makes freefall far louder, §3.1).
• The operator. In skydiving, there is no second crew member to cross-check or challenge decisions; under canopy, one's eyes and hands must be committed to the task at all times. Skydivers' reaction budgets are usually single-digit seconds. Jumpers also configure the alarms themselves (§4.1).
• Expected indications. Many skydiving indications are of altitude itself rather than of an action, and are therefore expected: the altitudes will inevitably be passed through. The cockpit's closest counterparts are the automatic radio-altimeter callouts on approach ("five hundred," "one hundred," "minimums"): scheduled altitude indications that, like an audible altimeter's, are certain to arrive.
• The framework transfers, not the findings. Aviation supplies some relevant questions and variables — level relative to masking, urgency, alarm count, voice versus tone, reliability — but the solutions must be re-derived for skydiving, as the Sea King project re-derived them for helicopters (§4.5). Some aspects of the framework do not map well or have no equivalent relevance in skydiving: the response-time hierarchy (§4.1), fixed and trained semantics, the two-stage voice alert (§4.4), and the alert conceived primarily as a last line of defense (below). The response-time hierarchy does, however, apply to emergency procedures: a high-speed malfunction demands immediate action, whereas a low-speed one allows time to assess and attempt a fix, with decision altitude (§1.2) as the floor of that window (the SIM's categories are total and partial; SIM 2026, 4-1.E). An altimeter that discriminates descent rate can even indicate which it is: a freefall-rate alarm that sounds below deployment altitude is, in effect, a high-speed-malfunction warning, since there only a no-pull or a high-speed malfunction keeps a jumper near freefall speed. Its threshold must sit above what a canopy can reach in a dive: a steep landing turn can exceed the 78 mph at which a standard AAD fires, for which reason the Speed CYPRES fires at 102 mph instead (Airtec 2025; Parachutist 2022).
• Contrasting roles. In a cockpit, indications mainly alert the pilot to problems that may be occurring; in skydiving, they mainly keep problems from occurring in the first place. For instance, in skydiving the AAD as a line of defense sits beneath the human; the alert's job in such a case is therefore not to be the last defense, but to help keep the jumper out of the race zone (§1.4; §5.2). One fixed-wing parallel is the F-16's Auto-GCAS, which flies the aircraft away from terrain if the pilot does not — for instance after losing consciousness under G — integrated into the existing F-16 fleet starting in 2014, with its first operational save on a late-2014 combat mission (Lockheed Martin 2015).

Part V — Design Principles

The characteristics of altitude awareness, the taxonomy, the nature of the sensory modalities, and the theory imported from cockpit alerting and related literature together suggest certain design principles. The purpose of this Part is not to mandate or standardize altitude indication systems — standardization makes sense for pilots who fly many aircraft, but for skydivers the advantages of customizing one's own indications outweigh it: a far greater number of possible combinations within and across modalities (Appendix C), systems tailored to specific types of jumps or to a jumper's own points of vulnerability, and simple personal preference. This Part is instead a guide to designing more effective customized systems of altitude indication on a solid theoretical and practical foundation. It can also help avoid outright errors in indication design, as opposed to matters of taste: a signal that cannot be perceived (a low-pitched alarm in freefall, where spectral masking makes it inaudible, §5.5); an indication placed without taking into account reaction or execution time (§5.2); or two signals for different actions that are not sufficiently dissimilar (§5.7).

After §5.1 on the choice of altimeters, this Part runs through the logic of indication design. First, we take a demand from Part I — a task element (threshold or phase), with its tolerance T(A) and its criticality. In §5.2–5.4 we set the coordinates of each indication in the taxonomy — form (§5.2), encoding (§5.3), modality (§5.4) — to meet the demand. Next, in §5.5 we realize the signal against its masking floor, and in §5.6–5.7 we apply the system-wide constraints (urgency across the hierarchy and economy across the set of indications). Finally, in §5.8 and §5.9 we tune the design for the jumper and the jump itself. In practice the order is not strict: the system-wide constraints, the jumper, and the jump shape every step from the start. Figure 25 shows an example of an indication system based on the skydive from Figure 1.

5.1 Altimeter selection

One of the most important decisions a skydiver makes for training and maintaining altitude awareness is which types of altimeter to use. It bounds the entire indication design space, dictating which cells of the taxonomy the jumper's indications can occupy. As the previous Parts have shown, this is not simply a case of more is better; many factors have to be taken into account: jumper experience level, level of cognitive load while skydiving, natural disposition to pay attention to altitude as well as natural ability to ascertain it from ambient cues (§3.4), the disciplines of skydiving a jumper practices, limitations of certain sensory modalities (e.g., far-sightedness or deafness), use or non-use of an AAD, and personal preference. The range of possibilities is wider than it has ever been: digital and analog numeric altimeters in various forms, GPS altimeters, beeping audibles, speaking audibles, dedicated peripheral-vision LED altimeters, and hybrids like WatchAlti (§2.3). There is no single correct answer; different combinations offer different options and achieve different results.

The consensus among instructors seems to be that beginners should not be overloaded with altimeters: one wrist- or hand-mounted altimeter is sufficient and most appropriate, freeing cognitive bandwidth that multiple sources would consume. The altimeter may use several encodings — DigiAlti, with its LightBar above the LCD, or an analog dial with yellow and red zones — but what matters is that it is a single device that targets the visual modality alone. This is sound, and it falls in line with the cognitive-load findings of §3.2: a second source of altitude information is extraneous load for a jumper whose decoding is not yet automatic, and it competes for the same attentional resources. On the cockpit model (§4.1), however, a strong argument can be made for one warning-tier alarm on an audible at an altitude where a student should no longer be in freefall — for example, a freefall-rate alarm at 4,000 ft, by which point a student should be under canopy (§4.6). Provided it is designed to balance startle against salience over the noise floor (§5.5), it adds a second modality as a fallback where the visual indication has failed (§5.4). A useful question for study: does an audible in addition to a visual altimeter reduce low pulls among students? In the author's experience, instructors tend to dismiss audibles for students without providing any real-world evidence. The 2023–2024 SIM said "Students should use audible altimeters only after demonstrating a satisfactory level of altitude awareness" (SIM 2023–2024, 5-3.J.4.c.(3)); the 2025 rewrite no longer contains the sentence.

As a jumper's cognitive load lessens, more attention can be paid to altitude, and other factors influence selection. For experienced jumpers the primary consideration is coverage: the task elements of their disciplines, stage by stage (§5.2). The USPA recommends redundancy across modalities — an audible in addition to a visual altimeter — for advanced disciplines such as freeflying, camera flying, and wingsuiting (SIM 5-2.C, 5-7.D, 5-8.C). The SIM recommends redundancy within a modality for three disciplines: two audible altimeters for freeflying (SIM 5-2.C), movement jumps (SIM 5-10.C), and speed skydiving. For speed skydiving it acknowledges that masking can render both audibles ineffective, recommending "two audible altimeters, preferably including flashing visuals inside the helmet, since the wind noise on a speed skydive may drown out the audible beeps," and it states "jumpers must increase the volume of all audible altimeters to maximum" (SIM 5-11.C). The SIM's wingsuit guidance — "a visual altimeter mounted on the chest strap or mud flap" (SIM 5-8.C) — results in less accurate altimeter readings than may be acceptable to most (Mongoven, in preparation a): the SIM itself warns that an altimeter in a burble "may read inaccurately" (SIM 4-3.J). A jumper without an AAD — a configuration that should be rare — should weigh the elevated, potentially fatal consequences of a lapse particularly carefully and choose accordingly. The SIM's own line on the AAD applies to everyone: "An AAD is only a backup device, and no jumper should rely on it to deploy a parachute" (SIM 4-3.G).

With experience of a particular setup, a hierarchy of use emerges, different for each stage of the skydive, with one altimeter dominant. Historically the primary altimeter (the one relied on most, which should essentially always be available) has been a numeric one in every stage — the type skydivers learn on — but in the author's experience, many jumpers who use a speaking audible now rely on it more than on their wrist altimeter; both are symbolic encodings, with the advantage of providing actual numbers. There is a strong case for combining the two for redundancy across modalities; a jumper need not decide in advance which is primary. A jumper cannot glance at a speaking audible, so it is not accessible the way a numeric altimeter is; but a numeric altimeter has to be looked at to register anything, so in another sense the speaking audible is more available, especially with frequent announcements: it arrives without a sampling act (§1.7, §2.2). Speaking audibles can also announce other altitude-related information and commands (e.g., "Pull!" or "decision altitude").

Each of the four indication types in §2.6 has its own advantages and disadvantages:

• Numeric readout (visual, symbolic, on demand): provides an exact altitude value, short learning curve; must be intentionally sampled by looking at it (§1.7), its decoding suffers under cognitive load (§3.2).
• Peripheral-vision light (visual, direct, background or interrupting): present without sampling (§2.2), unaffected by noise; more subtle than a beeping audible, masked by bright sunlight (§5.5).
• Speaking audible (auditory, symbolic, background or interrupting): provides an exact value without looking at anything, unambiguous; slower to deliver, harder to perceive over wind noise, and less dramatic than a beeping audible (§4.4).
• Beeping audible (auditory, direct, background or interrupting): the most jarring of all, cuts through noise; more ambiguous than a symbolic indication — a jumper has to know what altitudes a device is set to (§2.3).

The weaknesses of one indication type can often be remedied by the strengths of another, making a strong case for combining them (§5.4). Additionally, jumpers should think about balancing out their on-demand, interrupting, and background indications (§1.7).

The physical placement of altimeters on the body is an underappreciated aspect of indication design. Freefall Data Systems is carrying out a detailed study of the effect of mount location on the accuracy of barometric altimeters across skydiving disciplines; its results are pending (Mongoven, in preparation a). Placement affects not only accuracy but accessibility: an audible's level falls by about 6 dB with every doubling of its distance from the ear (the inverse-square law), so even a small move, such as from 4 to 2 cm, can significantly increase its audibility; a peripheral LED altimeter placed too far from central vision can be impossible to see when looking in certain directions.

5.2 Matching form to task

The task itself is paramount in considering the form an indication is to take. Starting with the fundamentals of our taxonomy — in which an indication can be either ranged or non-ranged, conceptually occupying either a span or a point, respectively — ranged indications map cleanly to the phases of a skydive and non-ranged indications equally well to its thresholds. This is true for ascent, freefall, and canopy. Nonetheless, a ranged indication (static or dynamic) can be useful at a threshold too, as an indication that the tolerance T(A) is being exhausted — essentially a point with a span attached (e.g., an audible alarm starting at a pull threshold of 4,000 ft and stopping precisely at 3,500 ft, the stop marking the moment the tolerance window has passed).

For phases, a ranged indication helps keep the estimate up between samples — raising the floor of altitude awareness, suppressing the decay of dispositional awareness — while periodic non-ranged re-triggers force a sample by one of two means: by indicating the altitude directly (the jumper knows what altitude has just been indicated), or by drawing a jumper's attention to another altimeter (the jumper glances at a wrist-mounted altimeter to get a fresh reading after an audible alarm). The re-triggers add teeth to the sawtooth of occurrent altitude awareness (§1.6). The goal, as previously stated, is to keep the dispositional altitude estimate inside tolerance without consuming the attention the phase needs for its own task. Our descriptive model from §1.5 also suggests scheduling indications in consideration of the tolerance profile T(A), lending preference to non-ranged indications where T tightens and ranged indications — especially dynamic ones — where the estimate must track over a range; no indication at all is affordable where T is loosest.

The opposite of the above-mentioned point with a span attached is what we can appropriately call a set-up phase: a phase that ends in an action, such as the last seconds of working time before breakoff, the base leg before final, or the set-up for a swoop (§2.5). This subgroup of phases lends itself particularly well to a ranged dynamic indication followed by a non-ranged event marker at the threshold: the ramp-up dials in the approach and creates anticipation, and the mark indicates the threshold itself with maximum precision.

One important task altimeters have is keeping the jumper out of the AAD race zone (§1.4). It is thus highly recommended to include indications specifically aimed at such prevention — for example, unique indications unlike any others or spoken words marking a jumper's decision altitude; a decision altitude indication should ideally be configured to occur whether in freefall or under canopy, so that it is perceived either way. Even if it is not needed in 99.9% of cases, having it in helps train a jumper's eyes to what this critical altitude looks like (altitude picture training). The fatality record supports the recommendation: in recent years, jumpers who died while trying to fix a malfunctioning main passed decision altitude without noticing (§1.9).

For each phase and threshold — working time, breakoff, deployment, decision altitude, hard deck, canopy pattern — latency and error tolerance should be taken into account; stage comes into play primarily because the masking floor and the altitude available differ between stages. The skydiving counterpart to the Sea King temporal response budget (§4.5) is an altitude budget, where the designer's focus lies on placement in an altitude range rather than the temporal length of the indication; an indication for an action must occur at the altitude by which the action must be complete (L, §1.5) plus the altitude consumed by registering it, reacting, and completing the action at an assumed fall rate. A non-ranged signal's own duration costs altitude too — conceived as a point to mark a threshold, it is nonetheless delivered over a span: a two-second voice message is about 350 ft belly-to-earth and about 530 ft head-down — so an indication for an action must be recognizable within a fraction of the altitude it is placed above that action. Some jumpers set an indication slightly above its threshold for exactly this reason, though for most thresholds, the planned altitude already allows for the reaction time and response. Time enters the equation only through the fall rate; thus a placement that is adequate belly-to-earth may be inadequate for an angle track (§5.9) — again, indication design is discipline-specific. How much altitude each task element consumes between indication and completed action, by stage and discipline, has not been measured.

5.3 Choosing the encoding

Another fundamental design decision is which encoding an indication should use. Between the two main encoding groups (§2.3), symbolic has the advantage of exactness (e.g., it can give an exact number) but involves a decoding process — a process that is learned, working-memory-dependent, and liable to be impeded by cognitive load. This is the type of encoding that is used most often in primary altimeters. Direct encoding has much more immediacy — there is nothing to decode — and, built as it usually is from beeps, tones and lights, much more viscerality (a beeping audible would even get the attention of a dog); but it is not as precise. These two types of encoding are ideally mixed and balanced principally according to the task structure of the skydive.

Within the direct encoding group we have three subgroups. Static direct indication (e.g., a green light while a speed skydiver is within the measurement zone) is useful for informing a jumper that they are somewhere within a specified altitude range. Shifts between such static indications (e.g., a sudden shift to red when the same speed skydiver exits the measurement zone) can be perceived to be more abrupt — and therefore more salient — than the gradual shift of a single direct dynamic indication. But dynamic indications are the most useful for additionally informing a jumper where they are within the specified range. Event-marker indications (a four-second red light at 1,000 ft) inherently contain less information, but are still useful in situations like ascent in an aircraft, where the altitude threshold itself is important while the altitude passed after the onset of the indication is not.

As noted in §2.3, the subgroups of direct encoding do not apply to symbolic encoding: a sign carries its value by convention, not through its form. Symbolic encoding is best for cases that require the least ambiguity. When it leans on inherited conventions such as language and numbers, little learning is necessary beyond the decoding itself, which makes it well suited to student use and primary altimeters; a local code, such as a manufacturer's beep pattern, must be memorized (§2.4).

On the motivated–arbitrary spectrum of §2.4, the most critical thresholds and phases should lean toward motivated mappings, or inherited conventions such as language or numbers, as opposed to local conventions such as those of a manufacturer. This is because well-chosen encodings mitigate misinterpretation (the failure points of §2.4).

5.4 Modality and redundancy

As skydives increase in complexity and a larger task load is taken on, the importance of redundancy of indications across modalities increases. We have seen how this is supported by the USPA's recommendations (§5.1) and by cockpit alerting (§4.1). Duplication within modalities as well as different forms of indication within the same modality — even combined into a single instrument such as DigiAlti 3's numeric readout plus LED LightBar — have their place. Altimeters are always placed in slightly differing locations even when an attempt is made to place them symmetrically, so one will inevitably be louder or more prominent than the other, even if they are providing the same indications. Altimeters are also mechanical or electronic devices, and it is always possible that one fails or runs out of battery. The two kinds of redundancy therefore guard against different things: duplication within a modality guards against device failure and placement asymmetry, while only redundancy across modalities guards against the perceptual failures that all indications in one modality share (§3.3).

5.5 Avoiding masking

Every sensory channel has sources of masking — secondary stimuli that impede or entirely block the perception of an indication (e.g., frequency masking due to wind noise). In order to be clearly and reliably perceived, an indication must be built with characteristics that allow it to clear the masking floor. The three senses differ in their characteristics, each one presenting its own challenges in terms of sensory masking that are also dependent on stage of the skydive and the individual. The general method that can be used for each channel is equivalent to Patterson's method for audio (§4.2): find the threshold the masker imposes and set the signal's strength a fixed margin above it; cap that signal's maximum strength at a certain margin above that in order to avoid startle or — in the most dramatic case for vision — blinding.

Auditory.

Freefall presents the greatest challenges when it comes to auditory masking while skydiving. Patterson's recommended harmonically related components are of little use in such a harsh noise environment. Earphones aside — they both reduce wind noise and pipe the sound directly into the ear — anything below 1 kHz is drowned out, and even 2 kHz does not reliably clear the masking floor; so instead we concentrate on the upper end of the most sensitive region of hearing. Freefall Data Systems generally targets 4 kHz as a center frequency for synthesized alarms intended for freefall; spoken indications use higher female voices and are processed such that the higher frequencies — which are most important for the comprehension of human speech in any case — are artificially boosted to clear the noise floor. We therefore do not recommend musical alarms for freefall: tones lose their musical quality and melodic intervals become unreliable above about 5 kHz (Attneave & Olson 1971), only about a major third above the 4 kHz around which freefall alarms are centered, so a melody has little room to move between the masking floor and that limit. Canopy flight, the quietest stage (§3.1), is where musical alarms are of most use. SirenAlti's altitude-reactive tunes are an example: a familiar melody, such as Reveille or the William Tell Overture, plays faster as the jumper descends — a ranged, dynamic indication in which the music is the carrier and the tempo represents the altitude. Because the tunes are already known, they also avoid the problem of learning the invented melodies of IEC 60601-1-8 (§4.2). In a quiet helmet they may also be suitable for wingsuit flight.

The wind noise experienced in freefall has a broad frequency spectrum that can mask indications. This spectrum changes dynamically with airspeed as well as the orientation, position and movement of the jumper's head, and is additionally dependent on the type of helmet and the use of earplugs — which not only attenuate but change the overall shape of the spectrum. So the designer should aim to sufficiently clear the floor (Patterson's recommendation is by at least 15 dB) at around 4 kHz for the worst-case scenario, which sets an absolute floor for the loudness of the indication; the startle effect is of less concern in an already incredibly loud environment. For spoken indications, besides boosting the higher frequencies, another option is to shift the entire frequency profile of a recorded spoken indication up — a pitch shift via STFT (short-time Fourier transform) or even simply speeding the sound up (faster delivery also has its own advantages to a point) — and produce something like the so-called chipmunk voice, which may not sound pleasant but may in fact be more perceptible.

Figure 23 shows the noise at the ear inside two helmets, one open-face and one full-face, through the three stages of a skydive (Appendix D). The ear hears in overlapping bands known as critical bands, roughly a whole tone wide above about 500 Hz and wider below (Fletcher 1940), and a frequency is drowned only by the noise within its own band; so the top row shows the noise critical band by band critical band, as it is perceived psychoacoustically, and the bottom row as a fine spectrum. The line along the bottom of the top row is the threshold of hearing in quiet (ISO 226:2023, the standard behind the modern equal-loudness contours that succeeded the Fletcher–Munson curves; Fletcher & Munson 1933): the quietest level at which a tone can be heard at each frequency, lowest between about 2 and 5 kHz, where the ear is most sensitive. Above it run the masked thresholds, highest at low frequencies and falling as frequency rises: the noise summed over one critical band (Glasberg & Moore 1990), less about 4 dB (Scharf 1970), which is the level a tone must reach to be heard at all. The shaded band above each is Patterson's range, 15 to 25 dB above the masked threshold: the room the designer has to work with.

The dots mark the same 4 kHz beep from one audible altimeter, measured at the ear on its side, its margin above the masked threshold drawn as a stem. It cleared the masked threshold in every stage and both helmets. In freefall in the open-face helmet it landed inside Patterson's band, near its top; in the full-face helmet it arrived about 10 dB louder and sat above the band, and in the climb and under canopy, where the noise is about 20 and 35 dB lower, it sat far above it in both. At the ear opposite the audible, it fell just below the masked threshold in freefall, though a fine FFT of that ear still shows it clearly, 6 to 17 dB above the noise around it: the FFT listens through slots under 2 Hz wide, the ear through a whole critical band.

Visual.

While auditory masking is — at least for our purposes — primarily spectral in nature (restricted to frequency overlaps), visual masking is inherently more complex. It has a spectral component — a hue close to its background, such as blue against sky — but more relevant factors come into play, ones that highlight the limitations of this most relied-upon of all senses.

Depending on the characteristics of the visual altimeter — whether, and how strongly, it is self-luminous — daylight can either improve or worsen readability. For self-luminous altimeters such as peripheral-vision LED altimeters, ambient light masks the signal; for reflective displays, which require ambient light to be read, darkness masks the signal (most numeric altimeters integrate some form of lighting for night jumps in order to mitigate such masking). Relevant in both cases is the contrast threshold: a signal is seen only if it differs from its background — in luminance, in color (chromatic masking), or both — by a contrast the eye can resolve at its current adaptation level. Two cases follow from it: sunlight reflected off the top glass of an altimeter (reflected glare) adds the same veiling light to the digits and their background alike, flattening the contrast between them; and a display in shade, seen by an eye adapted to bright sky or cloud, can be too dark to be readable. So the salience of indications from a visual altimeter changes throughout a skydive and throughout the day depending on available light. A background's pattern matters as well as its brightness: a textured background — a complex urban landscape below, an intricately patterned jumpsuit behind an altimeter mounted on the forearm — raises the contrast an indication needs in order to be picked out (contrast masking: Legge & Foley 1980). Figure 24 shows each case on a generic altimeter.

Every experienced skydiver has had instances where glare has impeded their vision. The sun has the power to mask all indications and a daylight-adapted eye generally needs very bright LEDs for them to be visible. Not only direct sun can produce this effect: a bright source anywhere — sunlit clouds, sun glint off another jumper's visor, desert sands or snow-covered terrain — scatters light inside the eye and lowers contrast everywhere, not only near the source (disability glare).

Just as the auditory sense has different levels of sensitivity to different frequencies, the field of vision has its own related psychophysical properties: peripheral vision has lower acuity and worse color discrimination than central vision but retains high sensitivity to flicker and changes in brightness. So for peripheral-vision LED indications, temporal (e.g., flashing) and brightness parameters play a more important role.

Tactile.

Experience has shown that the tactile sense is of limited use for altitude indication in skydiving; during the stage that matters most for altitude indication, the one in which altitude is lost fastest, its primary masker — ambient vibration, which also carries its own spectral signature — makes a tactile indication nearly impossible to perceive reliably. In instances where it can be used reliably, however, the rule remains the same in essence: sufficiently clear the vibrotactile floor, either through signal strength or through electrotactile stimulation.

5.6 Urgency and hierarchy

The urgency hierarchy inherited from cockpit alert theory should be taken into account in indication design; perceived urgency of indications should approximate situational urgency. In terms of raw appearance, a digital altimeter simply counting down altitude is a display devoid of any inherent urgency — the digits never change with criticality. Any sense of urgency is supplied by the jumper, through their own psychological state and the learned meaning of the displayed number at a particular stage of a skydive. This is, however, the type of urgency that fails under high cognitive load (§3.2). Indications should therefore carry urgency of their own.

The character of a signal itself can create a sense of — or be iconic of — urgency in various ways (§2.4): through temporal means (flashing or beeping, pulse rate and spacing); through spectral means (hue and hue combinations, or pitch, timbre, and inharmonicity); through signal strength and onset sharpness (brightness or loudness); and, for spoken indications, through the delivery itself (§4.4). Whether an indication should be spoken or synthesized can be argued both ways, coming down primarily to personal preference: synthesized signals are more immediate in their delivery and cut through noise better, while a spoken word is the least ambiguous signal there is (§4.4), an advantage that grows with the number of indications in the set, since only about six distinct sounds can be reliably told apart (§4.2). A signal can also lean on conventions and have connotations that increase the sense of urgency: a siren is generally associated with urgency in everyday life; flashing blue, red, and white lights mimic emergency vehicles. The signal should be structured, however, to maximize effect while minimizing annoyance (§4.2) — a difficult balance. Across the set, urgency should be graded to criticality, since a jumper who sets every indication to maximum urgency has no hierarchy at all (§4.3). The grading also leaves room for levity at its low end: on indications that call for no urgent action, a familiar tune or a humorous sound can lighten a jump — in the author's experience, jumpers often ask for indications in well-known voices. At a threshold that demands action, levity fails the reversal test of §2.4, in the same way a gentle chime for "PULL ALTITUDE" does.

5.7 Indication economy

Although altitude indication in skydiving addresses a different and more specific problem than cockpit alerting systems, Patterson's research makes clear that one should be mindful of the number of indications for a given stage or phase of a skydive. The number of indications a jumper can feasibly work with in a particular stage is the indication budget, and it is to be spent frugally. The budget works differently than in a cockpit, however: only a pilot's final approach resembles what a skydiver experiences — a descent in which a set of threshold altitudes will be crossed (§4.6). A spoken countdown of altitude is closer to a pilot's radio-altimeter callouts than to their warnings or cautions — for many jumpers a speaking audible functions as the primary altimeter (§5.1). It therefore draws on the budget less dramatically, especially when included alongside synthesized alarms, which then gain more prominence against the backdrop of the spoken announcements, which may to an extent become habituated to. The same happens in the cockpit: "After many hours of flying with an altitude alerter the chime and light could blend into the background noise of other cockpit alerts" (McClellan 1987). An exact cap on the number of indications per stage is hard to specify, since requirements and ability differ, but non-actionable indications beyond the primary altimeter and any countdown should be kept to a manageable number (§4.3) in order to limit the cry-wolf effect. Indications for different actions should also be sufficiently distinguishable from one another, whether auditory or visual, so that they cannot be confused with one another; the more indications in a stage, the more difficult this becomes (§4.2).

Skydivers also develop a familiarity with their indication systems that pilots and crew members do not have the luxury of developing. This again has to do with the different nature of the two systems: apart from routine callouts such as the radio altimeter's, a pilot will not routinely hear the same set of alarms or see the same indications, whereas a jumper hears and sees theirs on practically every jump. Patterson's limit of about six sounds came from a learning study with naive listeners (§4.2); a jumper who has heard the same set over hundreds of jumps is anything but naive. Over time, as a jumper becomes accustomed to their indications and their cognitive bandwidth increases, more indications can be added (§5.1). The indication system also has a level of personal relevance to the jumper — they designed it themselves — that a cockpit alerting system does not have for a pilot, and this works to the jumper's advantage.

5.8 The jumper

Personal characteristics and limitations of the jumper belong in the design process. Experience level matters here as it did in choosing altimeters (§5.1); how it changes will vary from jumper to jumper and situation to situation: Is the jumper engaging in new disciplines of skydiving, fighting complacency issues as fear is waning, or have they gained cognitive bandwidth they can dedicate to additional indications? Generally speaking, beginners should not be overloaded with learning all sorts of new indications when skydiving itself is entirely new to them. Experienced jumpers should periodically re-evaluate, refining and improving their indication design.

About 8% of men and 0.4% of women of European descent have an inherited red–green color deficiency — the highest rate of any population surveyed; men of Chinese and Japanese ethnicity run at 4–6.5% (Birch 2012) — restricting the available color space for such jumpers; color palettes for indications should be adapted to accommodate the common deficiencies. Presbyopia (stiffening of the lens) makes it more difficult for older jumpers to focus on digits close up, requiring more reliance on other means for maintaining altitude awareness and possible selection of an alternate primary altimeter.

Hearing loss through skydiving and aviation should be avoided actively through the consistent use of hearing protection. It is recommended to have one's hearing assessed periodically by medical professionals to track any loss. Any type of hearing loss or deafness should be addressed and can be compensated for through good altimeter choice and indication design. Noise-induced hearing loss (such as from firearms exposure) is known to commonly produce a notch at around 4 kHz — the very center frequency recommended for synthesized freefall alarms (§5.5; McBride & Williams 2001). Deaf jumpers should use a peripheral-vision LED altimeter as a backup to their primary visual altimeter, as their entire indication structure runs through a single modality.

5.9 Jump configuration

The configuration of a jump involves an extensive list of parameters, including but not limited to: critical planned altitudes and activities, light conditions, tactical constraints, discipline, gear and accessories, and the AAD and its activation altitude. In a sense, jump configuration is everything else jump-related that the designer needs to take into account — everything other than the jumper (§5.8) and the indications themselves, which are the design.

The planned thresholds and phases supply the main demands the indications are designed to meet. Per stage they are not numerous, so the number of indications beyond the primary altimeter can roughly match them. Some examples for ascent: 1,500 ft for seat belts; 2,000 ft for safe to exit on main; 5,000 ft for a tandem passenger interview; 12,000 ft for a gear check. Freefall examples might include: a 5,500-ft breakoff followed by tracking to a 4,000-ft pull; a decision altitude of 2,500 ft (high-speed malfunctions); a reserve-only altitude of 2,000 ft. For canopy one might consider: a decision altitude of 2,000 ft; a hard deck of 1,000 ft; the legs of the landing pattern (1,000, 600, and 300 ft).

The level of available light drastically affects the presentation of indications targeting the visual modality. In the dark, the brightness of self-luminous indications needs to be managed and numeric displays need to be bright enough to see; the SIM lists a lighted altimeter among the special equipment for night jumps (SIM 5-3.D).

Jumpers on tactical operations have a host of factors unique to their situation, and the task structure of such jumps can be considerably more complex than a civilian skydive's. Civilian operations typically take place within a limited geographical area, where the reference air pressure at the takeoff airport and the landing area is nearly the same; this is not true of tactical operations, which have no such geographical limits, and therefore may require the jumpers to set altimeters while in flight (zeroing them to landing zone elevation). Equipment or circumstances may preclude the use of certain technologies (e.g., a military jumper may be wary of using an audible altimeter during a stealth operation). Such jumpers should consider what sets their jump apart from a civilian one and whether indication design can improve safety in light of it. Altitude error tolerance tightens in many tactical operations, as a jumper can find themselves in a more vulnerable position than they normally would be, further from help (e.g., in the middle of a forest with fire approaching, in a combat zone far from the intended landing area).

The discipline of skydiving a jumper practices plays a fundamental role in shaping indication design. The SIM's minimum altimeter setups for particular disciplines are covered in §5.1. In any discipline in which body movement is significantly restricted, it is wise to use a speaking audible or a peripheral-vision altimeter (a numeric altimeter is not visible in every body position). Extremely loud disciplines like angle tracking and speed skydiving, whose wind noise carries stronger high-frequency components, also suggest the use of peripheral-vision altimeters, as the masking floor can in some cases be a challenge to overcome. A jumper who practices more than one discipline therefore need not have a single indication system: wingsuiting and speed skydiving, for example, clearly benefit from different designs, and the system can be reconfigured from one jump to the next. The cost is familiarity, which grows with every jump made on the same system (§5.7); signals shared between a jumper's systems should keep the same meaning in each, so that a sound learned as breakoff in one never means pull in another.

Gear, clothing, and accessories also affect how indications are perceived. Those wearing earplugs can effectively use louder indications that clear the noise floor by a greater margin than those without. Those wearing sunglasses will likely require the maximum brightness level from a peripheral-vision LED altimeter. Thick clothing can make tactile input difficult if not impossible to feel.

We conclude with a hypothetical design for a system of indications for the jump described in Figure 1. In addition to the indications themselves and their corresponding altitudes, we note for didactic purposes where they lie in our coordinate space. This example is not a recommendation but can serve as a reference and impetus for ideas for jumpers who are interested in more carefully considering their indication design.

Afterword

The main goal of altitude awareness is to prevent injury and death. In addition to developing safe practices and good habits in how a jumper deals with aircraft, gear, and all those involved in all stages of the skydive, a major focus of a newer jumper should be on the development of this critical skill. Experienced jumpers should focus on maintaining and sharpening it, ideally keeping it in check by attaining some level of meta-awareness. In general, jumpers should be aware of the psychological and circumstantial conditions that can lead to a lapse in altitude awareness, and recognize that no one is immune to them: any jumper can have a grade 4 or grade 5 lapse.

Appendices

Appendix A — Barometric altimeter error

Skydivers and pilots typically use barometric altimeters, which always have some error, to tell their altitude. This error stems from various sources, each of which shifts the reported altitude in a particular direction (a signed shift, i.e., up or down), spreads it out (a widening, i.e., a wider variation in reported values at the same altitude), or both:

• Temperature and humidity bias — signed shift only, not a widening. A barometric altimeter converts air pressure to altitude. By convention — in order to keep barometric altimeters in sync across aviation — this conversion assumes a standard atmosphere: air void of any humidity at all, a sea-level pressure of 29.92 inHg, and a temperature of 15 °C (59 °F) at sea level, with the air assumed to cool at a fixed rate of 6.5 °C per km (about 2 °C, or 3.6 °F, per 1,000 ft) as altitude increases (ICAO Doc 7488). Although digital altimeters typically automatically zero to ground elevation to set their own reference point (and do this regularly throughout the day to account for the natural ambient pressure drift that occurs), zeroing cannot correct for the differences between what the altimeter assumes and the day's actual conditions, for example:

  • the temperature might be 20 °C instead of 15 °C
  • the rate of cooling is never actually a fixed rate as assumed
  • the air is also never completely dry

  What zeroing does is remove only the pressure difference: it substitutes in the day's current pressure at the current elevation for the assumed 29.92 inHg. The rest of the model described above remains, however: the altimeter still uses the assumed theoretical temperature profile (void of any humidity) of the air column above and below its reference point.

  Of the three differences noted above, temperature is by far the largest source of bias. If the air on a particular day cools at the standard rate, the true altitude above the DZ is the indicated altitude multiplied by the ratio of the ground temperature to the standard one, both in kelvin (absolute temperature; at a sea-level DZ the standard is 15 °C, or 288 K). This works out to about 0.35% of altitude above the DZ for every degree C (about 0.19% for every degree F) the ground temperature differs from standard. Pilots round it up to 4% per 10 °C (ICAO PANS-OPS, quoted in SKYbrary n.d.), which overstates the error by about 15%; the values below use the exact ratio. Worked values at a deployment altitude of 3,500 ft (as in Figure 1), at a sea-level DZ (still assuming dry air and the fixed cooling rate), show how widely the readings spread at temperature extremes:

  • 0 °F (−18 °C): ~400 ft lower than indicated — the altimeter reads high. Pilots are taught about this case so they do not impact terrain due to inaccurate altimeter readings.
  • 32 °F (0 °C — freezing): ~180 ft lower.
  • 59 °F (15 °C — standard): zero error — this is the atmosphere the instrument assumes; indicated = true.
  • 90 °F (32 °C): ~210 ft higher than indicated — the altimeter reads low (the safe direction).

  Realistic bound: luckily, a jumper does not normally see this 0–90 °F spread; a home DZ's practical range might look more like 50–95 °F; this results in only ~60 ft lower than indicated at the cold (50 °F) end and ~240 ft higher at the warm (95 °F) end. So within normal jumping conditions, the dangerous-direction error is tens of feet and most of the range actually puts a jumper higher than indicated. Because the standard is an average and most jumping happens in warmer weather, the bias usually just happens to work in the jumper's favor.

  Humidity works the same as temperature, only more weakly: moist air is lighter than dry air, so a humid day acts like a slightly warmer one, and its bias runs only in the protective direction (as the standard atmosphere assumes dry air); this puts the jumper higher than indicated — at most about 50 ft at 3,500 ft, on the hottest, most humid days.

  These biases shift the indicated altitude band (which represents the reported values at a particular altitude) off-center. They also scale with altitude above the DZ, so the error is ~0 at the ground and largest at altitude. AADs are barometric too. On a cold day they fire at a true altitude below their setting, costing the reserve some potential inflation time; on a warm day they fire early, at a true altitude above their setting — protective for the save function, but also slightly increasing the likelihood of a two-out scenario. For a more complete discussion, see Altitude types: pressure, indicated, true, density; to compute the bias for a particular drop zone, temperature and dewpoint, see the Barometric altimeter error calculator. (See Figure 26 and Figure 27.)

• Sensor lag, noise, and processing — a shift up when descending, and a widening. Digital pressure sensors and microcontrollers alike are electronic devices and as such have their own inherent processing lag, either due to data and signal processing techniques (which vary by manufacturer) or the speed of their processing. The pressure sensors also contain some inherent unavoidable inaccuracy in the values they report (noise). Response lag in rapid descent shifts the band: during descent the instrument trails the true altitude and reads high — the more so the faster the descent: every tenth of a second of delay is worth about 18 ft in belly-to-earth freefall, but under 2 ft under canopy ("Altimeters may lag during both ascent and descent; expect inaccuracies of up to 500 feet, plus or minus," SIM 4-3.J). The noise and precision limits of the pressure sensor widen the band. All three of these sources of error vary from device to device and none can be eliminated entirely.

• Turbulence and dynamic pressure — a signed shift and a widening, determined by the altimeter's location. An altimeter reads the pressure at exactly one location and not in the undisturbed air around the jumper. Air flowing past the body raises the pressure on windward surfaces and lowers it in the wake (the burble), so the reading is offset by the local pressure coefficient C_p times the dynamic pressure. Dynamic pressure grows with the square of airspeed, so the effect in freefall is normally much more significant than under canopy. A mount facing into the relative wind reads low, putting the jumper higher than indicated; a mount located in a jumper's burble (wake) reads high, which is the dangerous direction: a jumper thinks they are higher than they are. As the body moves, the flow around each mount changes as well, so the reading fluctuates. The SIM itself warns that an altimeter in a burble "may read inaccurately" (SIM 4-3.J); in fact nearly every mounting location sits in disturbed air, and most read high: air pressing on a windward mount can raise the pressure only up to a ceiling (the stagnation pressure, where the air stops dead), while the suction on the sides and behind a falling body theoretically has no such limit and covers most of its surface. So the question is not whether a mount reads off but by how much and in which direction — and unfortunately, the direction is more often the dangerous one. AADs read the pressure at one location too, and the same effect moves their firing altitude: the Vigil in Pro mode is reported to fire between 840 and 1,100 ft depending on body position (Austin 2023). The size of the effect at each mounting location is the subject of a companion study (Mongoven, in preparation a; Appendix D).
• The zero reference — a signed shift. A barometric altimeter reports altitude above the point where it was zeroed, not above the ground beneath the jumper: over terrain or a landing area higher than the DZ the jumper is lower than indicated, and over lower ground the reverse (operations without a fixed DZ may set altimeters in flight, §5.9). Ambient pressure also drifts after zeroing: between takeoff and landing on a stable day only by a few feet (§1.5), but a passing front can shift surface pressure by 5–10 mb over a day, well over 100 ft. Falling pressure makes the altimeter read high, the dangerous direction — pilots learn it as "from high to low, look out below." Digital altimeters re-zero on the ground throughout the day; an analog one keeps whatever zero it was last given.

The signed shifts in altitude add up. In freefall on a warm day, the temperature bias offsets some of the lag (and of a burble-side mount's error); on a cold day the errors stack up — all in the dangerous direction. Under canopy all three shrink: lag and dynamic-pressure error with the slower descent, and the temperature bias as the jumper approaches the ground.

Appendix B — The reasoning behind Figure 5's tolerance envelope

The tolerance envelope drawn in Figure 5 (§1.5) is hypothetical and based on the author's own subjective judgments; other jumpers' envelopes would all look different. The reasoning behind each judgment is explained below so that other jumpers might compare and contrast their own opinions and personal T(A):

• Assumptions made. T(A) depends on more parameters than any figure could list — theoretically an infinite number of continuously changing parameters — so the profile assumes only the most salient and that the jump goes exactly as planned: one B-licensed jumper wearing a single wrist-mounted visual altimeter and a rig equipped with a reserve static line (RSL) and an AAD, on a 4-way belly jump from 13,000 ft, the four jumpers making up the entire load, falling at about 120 mph, with the plan of Figure 1 (breakoff at 5,000, pull at 3,500, decision altitude 2,000, pattern at 1,000, 600 and 300 ft), under a canopy descending at about 15 ft/s. Every other condition is taken as ordinary (daylight, good visibility, light winds, an uneventful climb); significant changes in any of them would change the envelope's shape.
• Within reason, plan for the worst-case scenario. In instances where a quantity varies (an opening distance, a malfunction, a reaction time), the profile takes the upper limit a jumper must be ready for, not the average. More altitude and more time are always better, and a margin sized for the typical case is exceeded by the first instance of an atypical case — the classic "It always worked before!" creeping into the structure of one's tolerance envelope.
• Every threshold is treated as vital (see Threshold criticality in §1.5), so no threshold is drawn looser simply because missing it would cost less; missing one threshold can mean missing the next, and so on.
• At the top: about 2,000 ft late, 5,000 ft early. Immediately following exit on our assumed jump, there is no threshold until breakoff. Believing oneself to be about 2,000 ft higher than one actually is (believing 15,000 at 13,000), approximately eleven seconds of freefall, would mean the jumper's internal clock — which in part times the sampling cycle — had been compromised too significantly (§3.4); an error on the other side (believing one is lower than one is) corrects itself, since a jumper who believes breakoff is near simply looks at the altimeter with the check refreshing the estimate. From there the envelope narrows in a straight line to breakoff.
• Breakoff: about 500 ft. This one comes more from a sense of community consensus and etiquette than a rule: in the author's experience, a breakoff up to about 500 ft low is commonly accepted in the skydiving community for a planned breakoff at 5,000, and it still leaves enough clearance for tracking away and deploying at a safe altitude. The USPA provides numbers for that clearance: for groups of five or fewer, breakoff should be at least 1,500 ft above the planned deployment altitude, which allows about 8 seconds of tracking, and the jumper should "wave off and deploy at the planned altitude whether or not you have turned and tracked" (SIM, ISP Categories G and H). The plan in our figure is based exactly on that minimum. A late breakoff therefore costs separation, not opening altitude: 500 ft late still leaves about 1,000 ft, close to 6 seconds of tracking, roughly two-thirds of the recommended time; much later than that and the space between jumpers might grow uncomfortably narrow.
• The track: no tolerance relaxation. From breakoff to deployment the envelope stays at 500 ft. Breakoff, track and deployment are one sequence of about eight seconds, and the jumper should be tracking with awareness and not focused on a visual altimeter, so there is no time to check and no reason for the envelope to relax.
• Deployment: about 500 ft. The container must be open by 2,500 ft, the USPA's minimum container opening altitude for B-, C- and D-license holders (SIM 2-1.I). After breakoff everyone tracks belly-to-earth at about 120 mph, where pulling and opening the container take about 500 ft (about three seconds from the decision to pull to an open container), so a pull planned at 3,500 can come up to 500 ft late and still open by 2,500. The opening holds the same tolerance until the canopy is safely open.
• Decision altitude matters on every jump. Even after a good canopy control check, a problem can appear just above it, so its tolerance is computed as a contingency for the worst malfunction still possible under an open canopy: a fast spin descending at about 75 mph, following a stall gone awry; a tear that develops in the canopy; or a collision with another canopy, a real possibility on a 4-way, with four canopies opening near each other — all rare, but possible. Among the highest-speed malfunctions is a bag lock, in which the main is deployed but the canopy stays in the deployment bag, such that the jumper keeps falling, essentially at freefall speed. A jumper who deploys at 3,500 ft and hesitates with a bag lock is at 2,000 ft in about eight seconds. On the jump drawn in Figure 5, though, a bag lock is possible only between deployment and the canopy opening, and in that window deployment's own 500-ft tolerance is tighter than a bag lock's would be: a bag lock looks like a slow opening at first and is recognized only about four seconds after the pull (an estimate), at about 2,800 ft, which still leaves about 1,450 ft before the emergency procedures must begin (below). Deployment is not complete until a canopy is open, so its 500 ft holds until then. Once the canopy is open, a bag lock can no longer occur (see Dynamic in §1.5).
• Decision altitude: about 350 ft. The decision itself to cut away is made above decision altitude. At decision altitude, only the execution of emergency procedures remains — cutting away and deploying the reserve — and the first of them, the release of the main, must happen by the 1,000-ft hard deck. Both altitudes are based on USPA rules: decision altitude is no lower than 2,000 ft for B- and C-license holders, and below the hard deck "it becomes too low to safely cut away" (SIM 4-1). A malfunction changes the descent rate, not the procedure, and the tolerance follows from three terms: the descent rate, the time to release the main and the hard deck. The descent rate is taken to be that of the fast spin, about 75 mph (110 ft/s). The time to release the main is taken as about six seconds, the low end of a published average: jumpers take six to eight seconds on average just to release the main once they have decided to cut away (Thomas & Booth 2005, reported in Crouch 2014, in the context of spinning canopies). Experts can assess, decide, execute and complete the whole procedure in about five seconds (Wallace 2019), but a B-licensed jumper is not assumed to be one. In theory, with the RSL, the reserve starts to deploy the moment the main is released, but just like the AAD, it should not be counted on. The hard deck itself also has room built in: a reserve must be functionally open within 300 ft of pack opening (PIA TS-135 §4.3.8), so a cutaway completed at 1,000 ft still leaves several hundred feet to spare. Six seconds at 110 ft/s costs about 660 ft, leaving about 350 ft of tolerance: a jumper who believes they are at 2,000 while actually at 1,650 still releases the main at about the hard deck and has the reserve open by about 700 ft.
• A bag lock costs altitude. In a bag lock the jumper keeps falling at about 176 ft/s (or potentially even faster due to body position changes): after recognition, which can easily take four seconds, releasing the main, the first of the emergency procedures, takes about two seconds, and the RSL deploys the reserve at that moment (both estimates). The release time is circumstantial: it is assumed here to be shorter than the spin's six seconds above, a published average, but on a given jump it can just as well be the other way around. To release the main by the hard deck, the procedures must begin by about 1,350 ft, and by about 1,700 ft to leave the fast spin's 350-ft safety margin from above: lower, not higher, than decision altitude. What a bag lock costs is altitude, about 175 ft every second: a jumper who recognizes it at about 2,800 ft and acts at once has the reserve open by about 1,900 ft; one who waits for decision altitude still releases the main above the hard deck, at about 1,650 ft, but has the reserve open only by about 1,100 ft. The Altitude tolerance visualizer depicts this case: a bag lock recognized about four seconds after the pull and acted on at once.
• Decision altitude is an emergency threshold. What it guards against outranks every other concern under canopy, so until it has passed, its own tolerance alone sets the envelope. Once it has passed and no emergency has occurred, the reference moves to the pattern, where constraints are tighter: it is the one threshold after which the envelope tightens, if only slightly, instead of relaxing. The rule applies to the late side only; the early side is discussed below. Just after the opening the envelope relaxes only modestly, to about 750 ft, since the first seconds are dedicated to a canopy control check; just after decision altitude it tightens to about 300 ft, narrowing to pattern entry.
• The landing pattern: about 200, 150 and 100 ft. Here the USPA does not give exact numbers for deviations, so the values are based on perceived norms. The USPA's own patterns differ, 1,000, 600 and 300 ft in the 2023–2024 SIM and 900, 600 and 300 ft since 2025 (ISP Category A), while drop zones that work in meters commonly fly 300, 200 and 100 m (about 985, 655 and 330 ft); in the author's experience 800, 400 and 200 ft is flown as well, though not taught to students, while discussing a turn to final below 200 ft is practically unheard of; and one drop zone's guide to the usual 1,000, 600 and 300 ft adds that "these altitudes can sometimes fluctuate a hundred or so feet" (Skydive Monroe 2023), a spread the metric pattern above falls within. Pattern entry is given the most leeway, about 200 ft: entering late is fairly common for jumpers who opened further out, and done carefully it does not add significant danger, since jumpers are expected to be joining the traffic flow at the start of the pattern. For safety, the window tightens as the ground approaches: about 150 ft at base, and about 100 ft at the turn to final. At the final turn the norms leave room for error: the safety calculation of example 2 in §1.5 allows about 125 ft before the turn ends in the low-turn zone, and low turns "are one of the biggest causes of serious injury and death in skydiving" (SIM 4-1). The tolerance is drawn a little tighter, at about 100 ft, so that the turn never comes later than 200 ft, below which no commonly given pattern puts it. Between the turns the envelope relaxes a little, to about 250 ft after entry and 175 ft after base, since the eyes should be focused on the landing area and traffic; after the final turn it does not relax at all, and it ends where direct perception takes over, at about 50 ft.
• How much wider the early side is. Following the asymmetry described in §1.5, acting early can cost other people as well, not just the jumper: an early breakoff essentially ends the jump for the rest of the group earlier than planned; an early opening can put a canopy in the path of jumpers still in freefall; a new jumper on final approach at 600 ft might irritate a tandem instructor entering the pattern at just above that altitude. The early side therefore depends greatly on how many people are in the air: the more jumpers and groups, the more the whole envelope tightens, the early side most significantly. On this jump, the four comprise the entire load, so in freefall the early side is drawn wide, at two and a half times the late side, all the way from exit. Under canopy above decision altitude it is drawn at one and a half times, since believing oneself far lower right after opening might rush the control check or send one looking for a landing area too soon. Below decision altitude, with no emergency, the jumper is heading for pattern entry, where all the canopies converge, so the early side is only slightly wider, at one and a quarter times the late side, to keep the pattern predictable: in traffic an early turn is partly a matter of etiquette as well as safety, while a late one, nearer the ground, weighs safety far more heavily. So the early side narrows relative to the late side as the jump goes on, because acting early costs more the lower one is and the more crowded the airspace gets. These factors too are subjective judgments.
• Between thresholds. The shape is based on the definition: the envelope narrows into each action and opens up again where there is time to check before the next action — little or none in the track, more in the pattern, most on the long stretches under canopy. Whether and how far it opens (or contracts) after each threshold is a matter of each jumper's judgment; the values drawn in this case are the T(A) the author would draw for this jumper on this hypothetical jump.

Appendix C — Quantifying the indication space

In §2.6, the cells in the indication space are counted coarsely, with motivation reduced to its two poles, motivated and arbitrary: 5 × 2 × 3 × 2 = 60. Figure 28 lays them all out. Of the 96 combinations of the five dimensions, three whole rows are empty by nature: static and dynamic encodings are properties of a span, so neither can be non-ranged, and an event marker is a point, and as such cannot be ranged (§2.3). Every constraint falls on the pairing of span and encoding; motivation, modality and role combine freely with each of the five pairings that remain. The figure's second panel places the four indication types of the product landscape (Figure 19) in the same space: each is a block of 8 or 12 cells rather than a single cell, and together they reach 40 of the 60, leaving the 20 tactile cells, where no successful products have yet been developed (§2.6).

The 60 cells, numbered as in the figure, each with the indication type of the product landscape it falls under:

We can view the space at a finer resolution by dividing motivation as the theory itself does: the motivated pole into its three iconic targets — magnitude, urgency and action — and the arbitrary pole into inherited and local conventions (§2.4). One constraint then appears that the coarse count cannot see: resembling a magnitude takes a signal that varies across a span or carries its own scale (§2.4), so a static signal, which does not vary, and an event marker, which is a point, cannot; only dynamic signals and self-scaling symbols can. Of the 40 combinations of span, encoding and motivation, 17 are empty by nature and 23 remain; modality and role still combine freely with each, so the space holds 23 × 3 × 2 = 138 cells, not the 150 that five motivation values would give unconstrained (the figure's third panel). The 23 combinations are listed below, each with an example from this document where one exists; a dash only indicates that this document gives none, not that such an indication cannot exist.

Appendix D — Studies and measurements

Aspects of the theory are based on data that are still being collected in a number of overlapping studies. As results come in, they are added to the main text.

• Altitude tolerance survey (§1.5). A short web survey in which skydivers judge individual deviations in altitude from one planned jump (a B-licensed 4-way from a Cessna 182) as acceptable, borderline, or not acceptable. Each threshold is probed on the late and the early side; this provides an actual measured tolerance to set beside the hypothetical profile of Figure 5. Hypotheses and analysis sealed before the first response; collecting since October 2026; analyzed first at 100 responses, with collection then continuing.
• Altimeter placement survey (§5.1). Where jumpers wear their altimeters, of any kind, on the jumps they normally do. Its ranking of locations by popularity is the measure the location study's first hypothesis uses. Collecting since October 2026.
• Altimeter location study (§5.1, Appendix A). The pressure error each mounting location produces, measured with synchronized loggers worn at 24 locations on the body, flow with different body positions and disciplines, from canopy flight to speed skydiving (Mongoven, in preparation a). Results pending.
• In-helmet noise (§5.5). Noise recorded at both ears inside the helmet on two jumps during every stage, one in an open-face and one in a full-face helmet, with the jumper's own audible beeping at 4 kHz (Mongoven, in preparation b). Its first results are reported in Figure 23. The recordings are not yet calibrated, so absolute levels rest on the microphones' rated sensitivity (about ±3 dB); comparisons between helmets, between stages, and of an alarm against the noise around it are unaffected, since each is made with the same microphones and recorder. Calibration is planned, and the study stays open to more helmets, body positions, disciplines, and earplugs. Recorded November 2026.

Appendix E — Revision history

• v1.0 (10 November 2026). The first release: Figure 23 measured, with sound (Appendix D).
• v0.1–v0.9 (July–November 2026). The first draft and seven full revision passes.

Glossary

Terms as defined by the text, with the section where each is introduced. In the main text bold marks both a definition at first use and paragraph lead-ins and emphasis; here definitions are collected by themselves.

• Alarm rationalization — in process-control alarm management (ISA-18.2), the review in which every alarm must be justified against the definition of an alarm, a signal that requires an operator response; alarms that fail are removed or given a lower priority (§4.3).
• Alert tiers — the cockpit hierarchy: a warning demands immediate awareness and immediate response, a caution immediate awareness and a subsequent response, an advisory awareness only (§4.1).
• Altimeter — a device designed to measure altitude and indicate it to a jumper; one altimeter can emit indications across multiple sensory modalities (§1.2).
• Altitude awareness — a jumper's internal estimate of current altitude, whether numeric or not, together with its uncertainty, error, and salience; graded, not binary. It is occurrent when actually in consciousness and dispositional between such moments, when the estimate persists below awareness, reportable if asked and going stale as descent continues (§1.2, §1.5, §1.6).
• Altitude budget — the skydiving counterpart to a cockpit's temporal response budget: an indication for an action must occur at the altitude by which the action must be complete plus the altitude consumed by registering it, reacting, and completing the action at a particular fall rate (§5.2).
• Altitude discipline — the characteristic of maintaining planned thresholds tightly in practice (§1.5).
• Altitude error tolerance — the altitude estimate error the next altitude-pinned action, and those after it, can absorb while the outcome remains acceptable: safe, within what the other jumpers on the load will accept, and what the jumper's own altitude discipline demands. At a single threshold, ε_max = b − L − Δ (§1.5). See tolerance profile.
• Altitude margin — the altitude between a threshold b and its completion limit L; tolerance is what remains of it after the completion distance Δ (§1.5).
• Altitude misjudgment — in contrast to a lapse, an error in directly perceived altitude at a more primitive cognitive level, where no represented value is in play at all; non-human animals make such misjudgments too, through sensory decline or poor light, for example (§1.2).
• Altitude picture — what a given altitude looks like (§1.2). Cf. sight picture.
• Altitude structure — a skydive's own profile of thresholds and phases across the three stages (§1.3).
• Ambient cues — non-instrument altitude information that is found by the jumper rather than sent: visual ground detail, the light gradient, temperature, barometric physiology, cloud and smoke layers, social and situational cues. Considered distinct from indications (§1.2, §3.4).
• Announcements — a speaking altimeter's regular spoken altitude readouts (§5.1).
• Arbitrary — of a mapping: conventional rather than resembling, where conventional means agreed upon or assigned, not common. The convention may be local (set by whoever configures the device, reassigned at no cost, memorized per device) or inherited from a shared code the jumper already has, such as language, numerals, or the clock face (§2.4).
• Attentional decay / estimate decay — the two kinds of decay between checks: attentional decay is the fading of the occurrent state; estimate decay is the drift of the dispositional estimate (§1.6).
• Audible altimeter — an altimeter whose indications use the auditory modality (§1.2).
• Auditory icon / earcon — the auditory-display terms matching the motivated and arbitrary poles: an everyday sound representing by resemblance versus an abstract, structured sound learned by convention (§2.4).
• Background indication — see on demand.
• Calibration failure — a lapse in which checking continued, but the stored estimate was wrong beyond tolerance (§1.6). Cf. sampling failure, registration failure.
• Cancel reflex — an alarm registers and is believed, but the first act is nonetheless to silence it (§4.3).
• Cell — see indication space.
• Check / sample / refresh / re-acquisition — four words describing aspects of one moment of the sampling cycle (see that entry), kept distinct. A check is the jumper's act of looking at the altimeter (the SIM's word); a sample is the model's name for that same act, the jumper being a sampler with an unfixed sample rate; a refresh is what a check does to the estimate, resetting it to the indication; re-acquisition is the outcome of a successful check: salience and estimate restored, the peak of a tooth in the sawtooth. A failed sample is a check with no re-acquisition (§1.6, §1.7).
• Common-mode failure — one mechanism taking out two redundant indications at once, as when two indications in the same modality share one or more failures of that modality resulting in the indications not being perceived (§3.3).
• Completion distance (Δ) — the altitude assumed to be consumed by an action from its beginning to its completion, reaction included; determined by task and descent rate (§1.5).
• Completion limit (L) — the altitude by which an action must be complete; this does not move (§1.5).
• Cry-wolf effect — frequent exposure to false alarms, or to alarms that require no action, leads to such alarms increasingly being ignored. In medicine the same hazard is called alarm fatigue (§4.3).
• Cutaway — releasing a malfunctioning main canopy, usually by pulling a handle on the harness, before deploying the reserve; also called a breakaway.
• Dead reckoning — carrying the last altimeter reading forward at a perceived fall rate. The dead-reckoning propagator is the internal clock, supplying elapsed time, with a perceived fall rate turning it into altitude lost (§1.6, §3.4).
• Decision altitude — the altitude at which a decision must have been made whether to cut away or keep a main canopy (§1.2).
• Direct encoding — the signal's physical form carries the jumper's relation to a set altitude: within a set span, where within it, or its arrival. The altitudes themselves are assigned; no direct indication carries its own value. Three subgroups: static, dynamic, event-marker (§2.3).
• Dynamic indication — a ranged direct indication that varies across its span (rising pitch, accelerating beeps, LED hue change), carrying position within the span as a magnitude (§2.3).
• Encoding — the manner in which the signal represents altitude, direct or symbolic; the second dimension of the taxonomy (§2.3).
• Estimation band — a representation of the jumper's altitude estimate: its width is the uncertainty, its offset the error, and, in the figures, its opacity the salience (§1.6). Not to be confused with the tolerance envelope (§1.5) or the indicated altitude band (Appendix A).
• Event-marker — a non-ranged direct indication: a bare point signal carrying only the information "the set altitude arrived" (§2.3).
• Failed sample — see registration failure.
• Failure layers — the levels at which an indication can fail to produce action: registration (sensory masking, attentional failure), misinterpretation, and response failure; incapacitation precedes them all by removing the perceiver (§2.4, §3.3).
• Form / role — an indication's form is everything the taxonomy's dimensions describe (span, encoding, motivation, modality); its role is the task-structure element it serves: a phase or a threshold (§2.5).
• Freshness plateau — the flat part of the slow-then-fast decline of awareness between checks, while dead reckoning actively keeps the estimate current (§1.6).
• Group lapse — a lapse shared by two or more jumpers in which their lapses are linked, by a common cause or by one jumper's lapse causing or sustaining another's. Three mechanisms: diffusion of responsibility (each jumper assumes someone else is watching the altitude), shared capture (the jumpers are each other's distraction), and cue substitution (breaking off on a group signal rather than one's own indication) (§1.4).
• Hard deck — the lowest altitude at which one can safely cut away a main canopy with adequate altitude for inflation of the reserve. Often confused with decision altitude (§1.2).
• Higher-order altitude awareness — awareness of the state of one's own altitude awareness: how fuzzy and how fresh it is, and how the sampling loop is running; the level that monitors and regulates ordinary altitude awareness (§1.8).
• Hybrid indication — an indication multiply coded across both encoding groups, such as the analog dial: needle position direct and dynamic, the scale symbolic, the colored arc direct and static (§2.3).
• Iconic-of-magnitude / iconic-of-urgency / iconic-of-action — the three things a motivated (iconic) signal can resemble. Magnitude ("how high") needs a span to vary across, so it is ranged only, self-scaling signals are the exception. Urgency ("how much this matters") is carried by the signal's character (harshness, loudness, tempo, roughness, onset), so it is range-independent. Action ("what to do") is the rarest: the signal resembles the response required, by its form (the flatline on final resembles the stillness required for straight flight) or by metaphor (a siren, the sound of emergency response, for the reserve) (§2.4).
• Index — Peirce's third kind of sign, tied to its object by cause rather than by resemblance or convention; a separate property from the motivated–arbitrary axis, so a sign can be both (the analog needle). Most ambient cues are indices of altitude (§2.4, §3.4).
• Indicated altitude band — the spread of altitudes an altimeter reports at one true altitude: a signed shift moves it off-center, a widening spreads it out. The instrument's counterpart to the jumper's estimation band, whose error it compounds (Appendix A).
• Indication — any stimulus delivered to a jumper by design that carries altitude information. An indication is sent, by an altimeter or by another jumper signaling on purpose; an ambient cue is found (§1.2).
• Indication budget — the implicit limit on the number of indications for a given stage or phase that a jumper can feasibly respond to most efficiently, to be spent frugally (§5.7).
• Indication space — the coordinate space formed by the five dimensions, in which every designed indication has a location; each location, one value per dimension, is a cell (§2.1, §2.6).
• Indication type — any location in the indication space, coarse or fine. The four product cells, numeric readout, peripheral-vision light, speaking audible, and beeping audible, are the coarsest, in the 2×2 matrix (§2.6).
• Internal clock — the natural sense of elapsed time that develops over repeated jumps; endogenous, so neither an ambient cue nor an index. It also supplies the elapsed time for dead reckoning (§1.2, §3.4).
• Interrupting indication — see on demand.
• Jump configuration — everything jump-related that the designer of an indication system must take into account other than the jumper and the indications themselves: planned altitudes and activities, light conditions, tactical constraints, discipline, gear and accessories, and the AAD and its activation altitude (§5.9).
• Lapse — the growth of altitude estimate error, allowing for its uncertainty, beyond current tolerance: the jumper can no longer be sure they are within it. Geometrically, the estimation band escaping the tolerance envelope, as in Figure 10 (§1.2, §1.6).
• Lapse geometry — the two shapes a lapse can take: a decay (attentional suppression, jumper conscious, recoverable by an exogenous re-trigger), caused by internal fixation (absorbed in the primary task) or external capture (a sudden or salient stimulus pulls attention away), or a step (an instantaneous drop to zero: incapacitation) (§1.5).
• Lapse grades — a working scale of severity: 0 benign uncertainty; 1 tolerance exceeded, no consequence; 2 late action; 3 automation rescue; 4 injury; 5 fatality. Grades 0–2 describe the jumper's state, 3 a system response, 4–5 a physical outcome (§1.4).
• Local / inherited convention — the two kinds of arbitrary (conventional) mapping. A local convention is set by whoever configures the device, its manufacturer or the jumper; it can be reassigned at no cost and is memorized per device (two beeps for breakoff). An inherited convention comes from a shared code the jumper already has (a spoken number, the numerals on a scale, a clock-face dial); it needs no learning and can be overridden only by fighting the prior learned convention (§2.4).
• Masked threshold — the level a tone must reach to be heard at all against a given noise; Patterson's rule places warning sounds 15 to 25 dB above it (§4.2, §5.5).
• Masking — secondary stimuli that impede or entirely block the perception of an indication, in any modality: wind noise for the ears, glare and ambient light for the eyes, vibration for the skin (§3.3, §5.5).
• Metaphor — the middle of the motivated–arbitrary spectrum: a mapping that resembles not altitude but something that stands for its meaning, such as a sustained flat tone at 300 ft that resembles a flat EKG (the flatline) (§2.4).
• Misreading — a check that registers but delivers the wrong value (a needle read against the wrong graduation, a transposed thousands digit), installing a wrong estimate held with full confidence: a calibration failure created at the moment of the check (§1.7). Cf. registration failure.
• Modality — the sense carrying the indication: auditory, visual, or tactile; the fourth dimension of the taxonomy (§1.2, Part III).
• Motivated (iconic) — referring to a mapping: grounded in resemblance or natural association, such as warmer meaning lower. The theory distinguishes three targets of iconic encoding: magnitude, urgency, and action, of which only magnitude requires a ranged indication (§2.4).
• Motivation — the motivated–arbitrary (iconic–conventional) spectrum, a property of an indication's mapping; the third dimension of the taxonomy and the only true spectrum among the five (§2.4).
• Multiply coded — of an indication: occupying several cells of the indication space at once, within a single signal; a tone and a light set off together are two indications, not one (§2.1, §2.4).
• Non-ranged (point) indication — fires once at a defined altitude crossing; conceived as an event, not a state; maximally precise at the instant of delivery (§2.2).
• Occurrent / dispositional awareness — altitude awareness actually in consciousness at the moment (occurrent), and the estimate that persists below awareness between such moments, reportable if asked and going stale as descent continues (dispositional) (§1.6). Cf. attentional decay / estimate decay.
• On demand / background / interrupting — three kinds of indication by how they reach the jumper. On demand: a display that waits to be read. Background: present without a sampling act, such as a peripheral light or a ramp of accelerating beeps, holding the estimate up while the loop runs. Interrupting: an alarm that announces itself; the only kind that can re-trigger a stopped loop (§1.7).
• Parameter — a perceptual dimension within a modality along which an indication can vary to carry information (auditory: pitch, loudness, tempo, spatial position; visual: brightness, hue, position) (§1.2).
• Parameter-mapping sonification — a data value driving a sound parameter, such as altitude driving pitch; in this taxonomy, a dynamic encoding, iconic of magnitude (§2.4).
• Phase — an altitude-extended activity between thresholds, such as working time, tracking, or the downwind leg. A phase indication serves a phase (§1.2, §2.5).
• Primary altimeter — the altimeter relied on most, which should essentially always be available (§5.1).
• Probability matching — people have been found to answer an alarm about as often as it is true (§4.3).
• Race zone — the altitude window in which a late pull can lead to concurrent activation: the AAD fires while the main is still opening (§1.4).
• Ranged indication — persists over a specific altitude span (§2.2).
• Re-trigger — whatever fires the next check: endogenous (trained habit, the internal clock, prospective memory, task transitions) or exogenous (indications, the ground appearing, other jumpers' behavior) (§1.7).
• Registration failure (failed sample) — a sample that fires without registering: the jumper looks at the altimeter, or hears the audible, but the value never reaches awareness; "looked-but-failed-to-see", or for an audible, "listened-but-failed-to-hear" (§1.7).
• Reserve-only altitude — in freefall or in an emergency exit, the altitude below which it is only safe to pull the reserve, as one would otherwise risk an AAD fire and potential main-reserve entanglement; may be different on ascent than descent (§1.2).
• Reversal test — the diagnostic for motivation: if reversing a mapping seems illogical, the mapping is motivated; if the reversed mapping would serve just as well, it is arbitrary (§2.4).
• Role — see form / role.
• Salience — the degree to which altitude is present in the jumper's mind at a given moment, driven in part by inevitability and fear; with uncertainty and error, one of the three quantities of altitude awareness (§1.2, §1.5, §1.6).
• Sampler — the jumper's mode of monitoring altitude: samples at an unfixed rate, not continuous monitoring (§1.7).
• Sampling cycle (also sampling loop) — check, refresh, reabsorption by the task, fading, re-trigger, check (§1.7).
• Sampling failure — a lapse in which the jumper stopped checking: the occurrent state was not renewed (§1.6).
• Sawtooth — the shape of altitude awareness over a descent: a sharp rise at each check and a slow-then-fast decline between checks; each check adds a tooth. The decline is not necessarily one-way: salience can rise between checks — from a passing thought of altitude, an ambient cue, a teammate glancing at their altimeter — and if what raised it carries no altitude information, the estimate is not refreshed (§1.5, §1.6, §2.2).
• Self-scaling signal — a signal that resembles a number without needing a span, the exception to magnitude requiring a ranged indication: a count supplies its own unit, and a part of a visible whole (a gauge lit once at a fraction of its length) supplies its own reference. This is the same mode through which devices without a display use counted beeps to confirm a stored setting (§2.4).
• Set-up phase — a phase that ends in an action, such as the last seconds of working time before breakoff or the base leg before final (§5.2).
• Sight picture — the swoopers' term for how a landing visually appears; related to, and partly overlapping with, the definition of altitude picture (§1.2).
• Skydive — "the descent of a person to the surface from an aircraft in flight when he or she uses or intends to use a parachute during all or part of that descent" (SIM, Glossary), extended here to include ascent (§1.2).
• Span — whether an indication occupies an altitude range (ranged) or a point (non-ranged); the first dimension of the taxonomy (§2.2).
• Spectral placement — placement of an auditory signal's frequency components. Best placed where the masked threshold is lowest, so it is heard over the noise; Patterson's guideline puts components in the range 0.5–5 kHz. Skydiving's three stages have different noise floors, so placement differs by stage (§4.2, §5.5).
• Stage — one of the three parts of the skydive: ascent, freefall, and canopy (§1.2).
• Standard atmosphere — the conventional atmosphere a barometric altimeter assumes in converting pressure to altitude: dry air, 29.92 inHg and 15 °C at sea level, cooling at 6.5 °C per km; real air departs from it, which shifts the reading (Appendix A).
• Static indication — a ranged direct indication uniform across its span, conveying only that the jumper is within an altitude range (§2.3). Can be a single uniform signal or a repeating pattern.
• Symbolic encoding — the signal itself carries the value or a named meaning, by assignment: it is a sign, such as a spoken value, displayed digits or a coded beep pattern, whose meaning is set by learned convention rather than by its form. Its meaning is almost always a point value (§2.3).
• Tau (τ) — time-to-contact specified directly by the rate of optical expansion; the mechanism behind ground rush, carrying time remaining but no altitude (§3.4).
• Threshold — an altitude-pinned point at which an action must begin, such as breakoff, deployment, or decision altitude; b in the tolerance formula of §1.5 (Figure 4). A threshold indication serves a threshold (§1.2, §2.5).
• Threshold criticality — the cost associated with missing certain thresholds, tolerance not scaling directly with the stakes (§1.5).
• Tolerance envelope — the tolerance profile T(A) drawn around the jumper's actual altitude (§1.5).
• Tolerance profile T(A) — permissible error as a function of altitude at a given descent rate; unique to each skydiver on each skydive (§1.5).
• Uncertainty / error — the two measures of an altitude estimate: uncertainty is the spread of answers you might get if you could ask the jumper their altitude repeatedly (the band's width); error is how far the middle of that spread lies from the true altitude (its offset) (§1.6).
• Urgency inflation — amplifying the urgency of indications for routine events, which transfers the cry-wolf effect from reliability to urgency itself (§4.3).

References

Works are cited in the text by author and year; standards and regulations by their designation. The Skydiver's Information Manual is cited as SIM, by section; without a year, the 2026 edition is meant. Secondary citations ("quoted in", "reported in") list only the source that was read.

Works cited

• Advanced Aerospace Designs. 2021. Vigil Cuatro User's Manual. US version 2021.05.
• Airtec. 2025. CYPRES 2 User Guide. Revision 01/2025, art. no. 991002. https://www.cypres.aero/wp-content/uploads/2025/09/991002_cypres_2_user_guide-en.pdf.
• Artwohl, A. 2002. "Perceptual and Memory Distortion during Officer-Involved Shootings." FBI Law Enforcement Bulletin 71 (10): 18–24.
• Attneave, F., and R. K. Olson. 1971. "Pitch as a Medium: A New Approach to Psychophysical Scaling." American Journal of Psychology 84 (2): 147–166.
• Auerbach, P. S., and Y. E. Miller. 1981. "High Altitude Flatus Expulsion (HAFE)." Western Journal of Medicine 134 (2): 173–174.
• Austin, A. 2023. "AAD Activation Altitudes of Leaders." Skydive Mag, July 14.
• Baddeley, A. D. 1972. "Selective Attention and Performance in Dangerous Environments." British Journal of Psychology 63 (4): 537–546.
• Baldwin, C. L., and B. A. Lewis. 2014. "Perceived Urgency Mapping across Modalities within a Driving Context." Applied Ergonomics 45 (5): 1270–1277. https://doi.org/10.1016/j.apergo.2013.05.002.
• Birch, J. 2012. "Worldwide Prevalence of Red-Green Color Deficiency." Journal of the Optical Society of America A 29 (3): 313–320. https://doi.org/10.1364/JOSAA.29.000313.
• Blattner, M. M., D. A. Sumikawa, and R. M. Greenberg. 1989. "Earcons and Icons: Their Structure and Common Design Principles." Human–Computer Interaction 4 (1): 11–44. https://doi.org/10.1207/s15327051hci0401_1.
• Bliss, J. P., R. D. Gilson, and J. E. Deaton. 1995. "Human Probability Matching Behaviour in Response to Alarms of Varying Reliability." Ergonomics 38 (11): 2300–2312.
• Bouak, F., O. Vartanian, and K. Hofer. 2019. "Performance and Health Effects of Mild Hypoxic Hypoxia in Simulated 6-Hour Exposures between 2,438 and 3,048 Metres." Journal of Military, Veteran and Family Health 5 (2). https://doi.org/10.3138/jmvfh.2018-0055.
• Bouak, F., O. Vartanian, K. Hofer, and B. Cheung. 2018. "Acute Mild Hypoxic Hypoxia Effects on Cognitive and Simulated Aircraft Pilot Performance." Aerospace Medicine and Human Performance 89 (6): 526–535. https://doi.org/10.3357/AMHP.5022.2018.
• Bradke, B. S., and B. R. Everman. 2020. "Mild Hypoxia of a Skydiver Making Repeated, Medium-Altitude Aircraft Exits." Aerospace Medicine and Human Performance 91 (2): 110–115. https://doi.org/10.3357/AMHP.5497.2020.
• Braun, C. C., and N. C. Silver. 1995. "Interaction of Signal Word and Colour on Warning Labels: Differences in Perceived Hazard and Behavioural Compliance." Ergonomics 38 (11): 2207–2220. https://doi.org/10.1080/00140139508925263.
• Brazil, E., and M. Fernström. 2011. "Auditory Icons." In Hermann, Hunt, and Neuhoff 2011, 325–338.
• Breznitz, S. 1984. Cry Wolf: The Psychology of False Alarms. Hillsdale, NJ: Lawrence Erlbaum Associates.
• Campbell, L. A., and R. A. Bryant. 2007. "How Time Flies: A Study of Novice Skydivers." Behaviour Research and Therapy 45 (6): 1389–1392.
• Chapanis, A. 1994. "Hazards Associated with Three Signal Words and Four Colours on Warning Signs." Ergonomics 37 (2): 265–275. https://doi.org/10.1080/00140139408963644.
• Crouch, J. 2014. "The RSL: Separating Fact from Fiction." Parachutist, January.
• CYPRES Team. 2020. "Cypres Save: Reinout and Simon's Story." Skydive Mag, June 30. https://skydivemag.com/new/cypres-saves-reinout-and-simons-story/.
• Darley, J. M., and B. Latané. 1968. "Bystander Intervention in Emergencies: Diffusion of Responsibility." Journal of Personality and Social Psychology 8 (4, pt. 1): 377–383. https://doi.org/10.1037/h0025589.
• Dehais, F., M. Causse, F. Vachon, N. Régis, E. Menant, and S. Tremblay. 2014. "Failure to Detect Critical Auditory Alerts in the Cockpit: Evidence for Inattentional Deafness." Human Factors 56 (4): 631–644. https://doi.org/10.1177/0018720813510735.
• Drew, B. J., P. Harris, J. K. Zègre-Hemsey, T. Mammone, D. Schindler, R. Salas-Boni, Y. Bai, A. Tinoco, Q. Ding, and X. Hu. 2014. "Insights into the Problem of Alarm Fatigue with Physiologic Monitor Devices: A Comprehensive Observational Study of Consecutive Intensive Care Unit Patients." PLoS ONE 9 (10): e110274. https://doi.org/10.1371/journal.pone.0110274.
• Droit-Volet, S., and W. H. Meck. 2007. "How Emotions Colour Our Perception of Time." Trends in Cognitive Sciences 11 (12): 504–513.
• Easterbrook, J. A. 1959. "The Effect of Emotion on Cue Utilization and the Organization of Behavior." Psychological Review 66 (3): 183–201.
• Edworthy, J. 2013. "Medical Audible Alarms: A Review." Journal of the American Medical Informatics Association 20 (3): 584–589. https://doi.org/10.1136/amiajnl-2012-001061.
• Edworthy, J., and A. Adams. 1996. Warning Design: A Research Prospective. London: Taylor & Francis.
• Edworthy, J., S. Loxley, and I. Dennis. 1991. "Improving Auditory Warning Design: Relationship between Warning Sound Parameters and Perceived Urgency." Human Factors 33 (2): 205–231.
• Edworthy, J., S. Reid, S. McDougall, J. Edworthy, S. Hall, D. Bennett, J. Khan, and E. Pye. 2017. "The Recognizability and Localizability of Auditory Alarms: Setting Global Medical Device Standards." Human Factors 59 (7): 1108–1127. https://doi.org/10.1177/0018720817712004.
• Einstein, G. O., and M. A. McDaniel. 1990. "Normal Aging and Prospective Memory." Journal of Experimental Psychology: Learning, Memory, and Cognition 16 (4): 717–726. https://doi.org/10.1037/0278-7393.16.4.717.
• Endsley, M. R. 1995. "Toward a Theory of Situation Awareness in Dynamic Systems." Human Factors 37 (1): 32–64. https://doi.org/10.1518/001872095779049543.
• Eysenck, M. W., N. Derakshan, R. Santos, and M. G. Calvo. 2007. "Anxiety and Cognitive Performance: Attentional Control Theory." Emotion 7 (2): 336–353.
• FAA (Federal Aviation Administration). 2000. "Terrain Awareness and Warning System; Final Rule." Federal Register 65 (61): 16736–16756, March 29.
• FAA (Federal Aviation Administration), Civil Aerospace Medical Institute. 2020. Hypoxia. Pilot safety brochure OK-20-0926. https://www.faa.gov/pilots/safety/pilotsafetybrochures.
• Fenz, W. D., and S. Epstein. 1967. "Gradients of Physiological Arousal in Parachutists as a Function of an Approaching Jump." Psychosomatic Medicine 29 (1): 33–51.
• Fitts, P. M., and R. E. Jones. 1947. Psychological Aspects of Instrument Display. I: Analysis of 270 "Pilot-Error" Experiences in Reading and Interpreting Aircraft Instruments. Memorandum Report TSEAA-694-12A. Wright Field, OH: Aero Medical Laboratory, Air Materiel Command. DTIC ADA800143.
• Fletcher, H. 1940. "Auditory Patterns." Reviews of Modern Physics 12 (1): 47–65. https://doi.org/10.1103/RevModPhys.12.47.
• Fletcher, H., and W. A. Munson. 1933. "Loudness, Its Definition, Measurement and Calculation." Journal of the Acoustical Society of America 5 (2): 82–108.
• Freefall Data Systems. n.d. Research articles and tools at freefalldatasystems.com: "A Skydiver's Guide to How Barometric Altimeters Work" (research_how_altimeters_work); "Altitude Types: Pressure, Indicated, True, Density" (research_types_of_altitude); "Barometric Altimeter Error Calculator" (research_altimeter_error_calculator); "Altitude Tolerance Visualizer" (research_tolerance_visualizer).
• Galdamez, L. 2024. "Studying Hypoxia in Skydivers." Parachutist, December 12.
• Gaver, W. W. 1986. "Auditory Icons: Using Sound in Computer Interfaces." Human–Computer Interaction 2 (2): 167–177. https://doi.org/10.1207/s15327051hci0202_3.
• Gescheider, G. A., and J. H. Wright. 1968. "Effects of Sensory Adaptation on the Form of the Psychophysical Magnitude Function for Cutaneous Vibration." Journal of Experimental Psychology 77 (2): 308–313.
• Gibson, J. J. 1979. The Ecological Approach to Visual Perception. Boston: Houghton Mifflin.
• Glasberg, B. R., and B. C. J. Moore. 1990. "Derivation of Auditory Filter Shapes from Notched-Noise Data." Hearing Research 47 (1–2): 103–138. https://doi.org/10.1016/0378-5955(90)90170-T.
• Grether, W. F. 1949. "Instrument Reading. I. The Design of Long-Scale Indicators for Speed and Accuracy of Quantitative Readings." Journal of Applied Psychology 33 (4): 363–372. https://doi.org/10.1037/h0058374.
• Grond, F., and J. Berger. 2011. "Parameter Mapping Sonification." In Hermann, Hunt, and Neuhoff 2011, 363–397.
• Grondin, S. 2010. "Timing and Time Perception: A Review of Recent Behavioral and Neuroscience Findings and Theoretical Directions." Attention, Perception, & Psychophysics 72 (3): 561–582. https://doi.org/10.3758/APP.72.3.561.
• Guillaume, A. 2011. "Intelligent Auditory Alarms." In Hermann, Hunt, and Neuhoff 2011, 493–508.
• Hellier, E. J., J. Edworthy, and I. Dennis. 1993. "Improving Auditory Warning Design: Quantifying and Predicting the Effects of Different Warning Parameters on Perceived Urgency." Human Factors 35 (4): 693–706.
• Hellier, E., J. Edworthy, B. Weedon, K. Walters, and A. Adams. 2002. "The Perceived Urgency of Speech Warnings: Semantics versus Acoustics." Human Factors 44 (1): 1–17. https://doi.org/10.1518/0018720024494810.
• Hermann, T., A. Hunt, and J. G. Neuhoff, eds. 2011. The Sonification Handbook. Berlin: Logos. Open access at https://sonification.de/handbook.
• Hewitt, A. n.d. "Confidential Reports – Training." skydive-safety.com. http://www.skydive-safety.com/Confidential-Reports-Training.htm.
• Hill, J. H., and R. Chernikoff. 1965. Altimeter Display Evaluation. NRL Report 6242. Washington, DC: Naval Research Laboratory. DTIC AD0610664.
• Hollins, M., A. K. Goble, B. L. Whitsel, and M. Tommerdahl. 1990. "Time Course and Action Spectrum of Vibrotactile Adaptation." Somatosensory & Motor Research 7 (2): 205–221.
• Hübner, C. 2023. "If I Ever Need My AAD, I'll Stop Skydiving." Skydive Mag, February 10. https://skydivemag.com/new/if-i-ever-need-my-aad-ill-stop-skydiving/.
• IATA (International Air Transport Association). 2016. Controlled Flight into Terrain: A Study of Terrain Awareness Warning System Capability and Human Factors in CFIT Accidents 2005–2014. 1st ed. Montreal: IATA.
• ICAO (International Civil Aviation Organization). 2003. Appendix A: Amendments to Annex 6 Concerning Altimeters. Information paper AN-Conf/11-IP/8, Eleventh Air Navigation Conference. Montreal: ICAO.
• Joint Commission. 2013. "Medical Device Alarm Safety in Hospitals." Sentinel Event Alert, no. 50, April 8.
• Jones, D. G., and M. R. Endsley. 1996. "Sources of Situation Awareness Errors in Aviation." Aviation, Space, and Environmental Medicine 67 (6): 507–512.
• Kaczmarek, K. A., J. G. Webster, P. Bach-y-Rita, and W. J. Tompkins. 1991. "Electrotactile and Vibrotactile Displays for Sensory Substitution Systems." IEEE Transactions on Biomedical Engineering 38 (1): 1–16. https://doi.org/10.1109/10.68204.
• Klinger, D. 2001. Police Responses to Officer-Involved Shootings. Final report to the National Institute of Justice, NCJ 192286. Washington, DC: US Department of Justice.
• Koriat, A. 1997. "Monitoring One's Own Knowledge during Study: A Cue-Utilization Approach to Judgments of Learning." Journal of Experimental Psychology: General 126 (4): 349–370.
• Kramer, G., B. N. Walker, T. Bonebright, P. Cook, J. Flowers, N. Miner, J. Neuhoff, et al. 1999. The Sonification Report: Status of the Field and Research Agenda. Prepared for the National Science Foundation by members of the International Community for Auditory Display. Santa Fe, NM: ICAD.
• Lavie, N. 2005. "Distracted and Confused? Selective Attention under Load." Trends in Cognitive Sciences 9 (2): 75–82. https://doi.org/10.1016/j.tics.2004.12.004.
• Lee, D. N. 1976. "A Theory of Visual Control of Braking Based on Information about Time-to-Collision." Perception 5 (4): 437–459. https://doi.org/10.1068/p050437.
• Lee, D. N., and P. E. Reddish. 1981. "Plummeting Gannets: A Paradigm of Ecological Optics." Nature 293 (5830): 293–294. https://doi.org/10.1038/293293a0.
• Legge, G. E., and J. M. Foley. 1980. "Contrast Masking in Human Vision." Journal of the Optical Society of America 70 (12): 1458–1471. https://doi.org/10.1364/JOSA.70.001458.
• Lockheed Martin. 2015. "Lockheed Martin F-16 Auto GCAS Team Wins Aviation Week Award." Press release, November 11.
• Lower, M. C., P. D. Wheeler, R. D. Patterson, J. Edworthy, M. J. Shailer, R. Milroy, G. M. Rood, and J. A. Chillery. 1986. The Design and Production of Auditory Warnings for Helicopters 1: The Sea King. ISVR Report AC527A. Southampton: Institute of Sound and Vibration Research.
• Mack, A., and I. Rock. 1998. Inattentional Blindness. Cambridge, MA: MIT Press.
• Marshall, D. C., J. D. Lee, and P. A. Austria. 2007. "Alerts for In-Vehicle Information Systems: Annoyance, Urgency, and Appropriateness." Human Factors 49 (1): 145–157. https://doi.org/10.1518/001872007779598145.
• McBride, D. I., and S. Williams. 2001. "Audiometric Notch as a Sign of Noise Induced Hearing Loss." Occupational and Environmental Medicine 58 (1): 46–51.
• McClellan, J. M. 1987. "Altitude Alert." Flying, February, 44–48.
• McGookin, D., and S. Brewster. 2011. "Earcons." In Hermann, Hunt, and Neuhoff 2011, 339–361.
• Meier, B. P., P. R. D'Agostino, A. J. Elliot, M. A. Maier, and B. M. Wilkowski. 2012. "Color in Context: Psychological Context Moderates the Influence of Red on Approach- and Avoidance-Motivated Behavior." PLoS ONE 7 (7): e40333. https://doi.org/10.1371/journal.pone.0040333.
• Meyer, K. 2026. "What's in a AAD Life Save." Seminar recording, British Skydiving, February 22. YouTube. https://www.youtube.com/watch?v=jrS2Enn-75U.
• Mongoven, C. In preparation a. The Effect of Barometric Skydiving Altimeter Location on Instrument Accuracy. Freefall Data Systems.
• Mongoven, C. In preparation b. Spectral Profiles of In-Helmet Wind Noise While Skydiving. Freefall Data Systems.
• Moray, N. 1959. "Attention in Dichotic Listening: Affective Cues and the Influence of Instructions." Quarterly Journal of Experimental Psychology 11 (1): 56–60. https://doi.org/10.1080/17470215908416289.
• Morton, E. S. 1977. "On the Occurrence and Significance of Motivation-Structural Rules in Some Bird and Mammal Sounds." American Naturalist 111 (981): 855–869.
• Nelson, T. O., and L. Narens. 1990. "Metamemory: A Theoretical Framework and New Findings." In The Psychology of Learning and Motivation, vol. 26, edited by G. H. Bower, 125–173. San Diego: Academic Press.
• Nesthus, T. E., L. L. Rush, and S. S. Wreggit. 1997. Effects of Mild Hypoxia on Pilot Performance at General Aviation Altitudes. Report DOT/FAA/AM-97/9. Washington, DC: FAA Office of Aviation Medicine.
• NTSB (National Transportation Safety Board). 1967. Aircraft Accident Report: United Air Lines, Inc., Boeing 727-22, N7036U, in Lake Michigan, August 16, 1965. Washington, DC: NTSB.
• NTSB (National Transportation Safety Board). 2010. Loss of Control on Approach, Colgan Air, Inc., Operating as Continental Connection Flight 3407, Bombardier DHC-8-400, N200WQ, Clarence Center, New York, February 12, 2009. Aircraft Accident Report NTSB/AAR-10/01. Washington, DC: NTSB.
• Parachutist. 2022. "Keep an Eye Out: AAD Activation During Swoop." By USPA staff. Parachutist, November 30.
• Parasuraman, R., and V. Riley. 1997. "Humans and Automation: Use, Misuse, Disuse, Abuse." Human Factors 39 (2): 230–253. https://doi.org/10.1518/001872097778543886.
• Patterson, R. D. 1982. Guidelines for Auditory Warning Systems on Civil Aircraft. CAA Paper 82017. London: Civil Aviation Authority.
• Patterson, R. D. 1990. "Auditory Warning Sounds in the Work Environment." Philosophical Transactions of the Royal Society of London B 327: 485–492.
• Peirce, C. S. 1932. Collected Papers of Charles Sanders Peirce. Vol. 2, Elements of Logic, edited by C. Hartshorne and P. Weiss. Cambridge, MA: Harvard University Press.
• Physio-Control. 2015. LIFEPAK 12 Defibrillator/Monitor Operating Instructions. Publication date 08/2015, MIN 3207254-033. https://www.stryker.com/content/dam/stryker/ems/resources/operating-instructions/international/3207254-033_int-eng_lifepak_12_operating_instructions.pdf.
• Pilkington, G. 1997. Audible Altimeter for Parachutists. US Patent 5,638,059, filed January 29, 1996, and issued June 10, 1997.
• Pravossoudovitch, K., F. Cury, S. G. Young, and A. J. Elliot. 2014. "Is Red the Colour of Danger? Testing an Implicit Red–Danger Association." Ergonomics 57 (4): 503–510. https://doi.org/10.1080/00140139.2014.889220.
• Ramírez-delaCruz, M., et al. 2024. "Effects of Acute Hypoxic Exposure in Simulated Altitude in Healthy Adults on Cognitive Performance: A Systematic Review and Meta-Analysis." Biology 13 (10): 835. https://doi.org/10.3390/biology13100835.
• Rankin, C. H., et al. 2009. "Habituation Revisited: An Updated and Revised Description of the Behavioral Characteristics of Habituation." Neurobiology of Learning and Memory 92 (2): 135–138.
• Raveh, D., and N. Lavie. 2015. "Load-Induced Inattentional Deafness." Attention, Perception, & Psychophysics 77 (2): 483–492. https://doi.org/10.3758/s13414-014-0776-2.
• Riedinger, F., C. Suied, R. Keane, J.-F. Sciabica, and F. Dehais. 2026. "Rough Auditory Alarms Mitigate Inattentional Deafness during Piloting-Like Task." Human Factors 68 (8): 1078–1097. https://doi.org/10.1177/00187208261431654.
• Sanderson, P. M., A. Wee, and P. Lacherez. 2006. "Learnability and Discriminability of Melodic Medical Equipment Alarms." Anaesthesia 61 (2): 142–147. https://doi.org/10.1111/j.1365-2044.2005.04502.x.
• Saussure, F. de. 1916. Cours de linguistique générale. Paris: Payot. English translation: Course in General Linguistics, translated by W. Baskin. New York: Philosophical Library, 1959.
• Scharf, B. 1970. "Critical Bands." In Foundations of Modern Auditory Theory, edited by J. V. Tobias, 1:159–202. New York: Academic Press.
• Schooler, J. W. 2002. "Re-representing Consciousness: Dissociations between Experience and Meta-consciousness." Trends in Cognitive Sciences 6 (8): 339–344. https://doi.org/10.1016/S1364-6613(02)01949-6.
• Seashore, C. E. 1938. Psychology of Music. New York: McGraw-Hill.
• SIM 2023–2024. United States Parachute Association. 2022. Skydiver's Information Manual, 2023–2024 edition. Fredericksburg, VA: USPA.
• SIM 2025. United States Parachute Association. 2025. Skydiver's Information Manual, 2025 edition (the narrative rewrite). Fredericksburg, VA: USPA.
• SIM 2026. United States Parachute Association. 2026. Skydiver's Information Manual, 2026 edition. Fredericksburg, VA: USPA. Cited as "SIM" without a year.
• Simpson, C. A., and D. H. Williams. 1980. "Response Time Effects of Alerting Tone and Semantic Context for Synthesized Voice Cockpit Warnings." Human Factors 22 (3): 319–330. https://doi.org/10.1177/001872088002200306.
• SKYbrary. n.d. "Altimeter Temperature Error Correction." SKYbrary Aviation Safety. https://skybrary.aero/articles/altimeter-temperature-error-correction.
• Skydive City. 2017. AFF/ISP Guide. Student ground-school manual. Zephyrhills, FL: Skydive City. https://skydivecity.com/wp-content/uploads/AFF-ISP-Guide-2017.pdf.
• Skydive Monroe. 2023. "Skydiving Landing Patterns Explained." March 23. https://skydivemonroe.com/blog/skydiving-landing-patterns/.
• Sorkin, R. D. 1988. "Why Are People Turning Off Our Alarms?" Journal of the Acoustical Society of America 84 (3): 1107–1108.
• Stetson, C., M. P. Fiesta, and D. M. Eagleman. 2007. "Does Time Really Slow Down during a Frightening Event?" PLoS ONE 2 (12): e1295. https://doi.org/10.1371/journal.pone.0001295.
• Sweller, J. 1988. "Cognitive Load during Problem Solving: Effects on Learning." Cognitive Science 12 (2): 257–285. https://doi.org/10.1207/s15516709cog1202_4.
• Sweller, J. 1994. "Cognitive Load Theory, Learning Difficulty, and Instructional Design." Learning and Instruction 4 (4): 295–312.
• Sweller, J., J. J. G. van Merriënboer, and F. G. W. C. Paas. 1998. "Cognitive Architecture and Instructional Design." Educational Psychology Review 10 (3): 251–296.
• Theeuwes, J. 2010. "Top-Down and Bottom-Up Control of Visual Selection." Acta Psychologica 135 (2): 77–99. https://doi.org/10.1016/j.actpsy.2010.02.006.
• Thompson, R. F., and W. A. Spencer. 1966. "Habituation: A Model Phenomenon for the Study of Neuronal Substrates of Behavior." Psychological Review 73 (1): 16–43.
• Troxler, I. P. V. 1804. "Über das Verschwinden gegebener Gegenstände innerhalb unseres Gesichtskreises." Ophthalmologische Bibliothek 2 (2): 1–53.
• USPA (United States Parachute Association). 2018–2026. Fatality summaries for 2017–2025, published annually in Parachutist and at uspa.org; cited as "USPA Fatality Summary" with the year covered. 2017: "Malfunction, Malfunction, Malfunction"; 2018: "A Record Low"; 2019: "Striving for Zero"; 2020: "Another Record Low"; 2021: "A Positive Direction"; 2022: "A Step Backward"; 2023: "A Widespread Improvement"; 2024: "A Milestone in Safety"; 2025: "A Need for Vigilance".
• Vaughan, D. 1996. The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA. Chicago: University of Chicago Press.
• Walker, B. N., and M. A. Nees. 2011. "Theory of Sonification." In Hermann, Hunt, and Neuhoff 2011, 9–39.
• Wallace, B. P. 2019. "Don't Delay, Cut Away!—The Five-Second Rule: A Time-Based Approach to Emergencies." Parachutist, July 1. https://parachutist.com/Article/dont-delay-cut-awaythe-five-second-rule-a-time-based-approach-to-emergencies.
• Wickens, C. D. 2002. "Multiple Resources and Performance Prediction." Theoretical Issues in Ergonomics Science 3 (2): 159–177. https://doi.org/10.1080/14639220210123806.
• Wogalter, M. S., ed. 2006. Handbook of Warnings. Mahwah, NJ: Lawrence Erlbaum Associates.
• Wolfe, J. M. 1999. "Inattentional Amnesia." In Fleeting Memories: Cognition of Brief Visual Stimuli, edited by V. Coltheart, 71–94. Cambridge, MA: MIT Press.
• Wolfe, J. M., A. Kosovicheva, and B. Wolfe. 2022. "Normal Blindness: When We Look but Fail to See." Trends in Cognitive Sciences 26 (9): 809–819. https://doi.org/10.1016/j.tics.2022.06.006.
• Wood, N., and N. Cowan. 1995. "The Cocktail Party Phenomenon Revisited: How Frequent Are Attention Shifts to One's Name in an Irrelevant Auditory Channel?" Journal of Experimental Psychology: Learning, Memory, and Cognition 21 (1): 255–260. https://doi.org/10.1037/0278-7393.21.1.255.
• Yantis, S., and J. Jonides. 1984. "Abrupt Visual Onsets and Selective Attention: Evidence from Visual Search." Journal of Experimental Psychology: Human Perception and Performance 10 (5): 601–621. https://doi.org/10.1037/0096-1523.10.5.601.
• Zhou, Z., Y. Yang, J. Liu, J. Zeng, X. Wang, and H. Liu. 2022. "Electrotactile Perception Properties and Its Applications: A Review." IEEE Transactions on Haptics 15 (3): 464–478. https://doi.org/10.1109/TOH.2022.3170723.

Standards and regulations

• 14 CFR 25.1322. Flightcrew Alerting. Code of Federal Regulations, Title 14, Part 25 (transport category airplanes).
• 14 CFR 27.1322 and 29.1322. Warning, Caution, and Advisory Lights. Code of Federal Regulations, Title 14, Parts 27 and 29 (normal and transport category rotorcraft).
• ANSI/ISA-18.2-2016. Management of Alarm Systems for the Process Industries. Research Triangle Park, NC: International Society of Automation. Supersedes ANSI/ISA-18.2-2009.
• EASA CS 25.1322. Flight Crew Alerting. In Certification Specifications and Acceptable Means of Compliance for Large Aeroplanes (CS-25), Amendment 28, 2023. Cologne: European Union Aviation Safety Agency. In its present form since Amendment 11, 2011.
• EEMUA Publication 191. 2024. Alarm Systems: A Guide to Design, Management and Procurement. 4th ed. London: Engineering Equipment and Materials Users Association.
• FAA AC 25.1322-1. 2010. Flightcrew Alerting. Advisory Circular. Washington, DC: Federal Aviation Administration.
• ICAO Annex 13. 2024. Annex 13 to the Convention on International Civil Aviation: Aircraft Accident and Incident Investigation. 13th ed. Montreal: ICAO.
• ICAO Doc 7488. 1993. Manual of the ICAO Standard Atmosphere (Extended to 80 Kilometres (262 500 Feet)). Doc 7488/3, 3rd ed. Montreal: ICAO.
• IEC 60601-1-8. Medical Electrical Equipment — Part 1-8: General Requirements for Basic Safety and Essential Performance — Collateral Standard: General Requirements, Tests and Guidance for Alarm Systems in Medical Electrical Equipment and Medical Electrical Systems. Geneva: International Electrotechnical Commission. 1st ed. 2003; 2nd ed. 2006; Amendment 2, 2020.
• ISO 226:2023. Acoustics — Normal Equal-Loudness-Level Contours. Geneva: International Organization for Standardization.
• PIA TS-135. 2010. Technical Standard 135, Revision 1.4: Performance Standards for Personnel Parachute Assemblies and Components. Parachute Industry Association. https://www.pia.com/wp-content/uploads/TS-135v1.4.pdf.