Best Radio Frequency for Backcountry Emergencies: VHF, UHF, or HF?

#

Ask six radio operators which band is best for backcountry emergencies, and you will get six confident answers, each of them true in a different sense.

The HF operator is the one with a wire still strung through the pines behind his truck, and he speaks with the unhurried certainty of a man who has answered a stranger eleven hundred miles away while standing somewhere with no road, no tower, and no satellite in view. High frequency, he will tell you, and only high frequency, reaches help over a mountain range. He is right. When the ridge, the valley, and the entire drainage between you and the county road lie across the path, the only signal that arrives is the one that leaves the atmosphere entirely, bends through the ionosphere, and comes back down where it is wanted. Press him about the two-thousand-foot walls of the canyon he is camped in and he will shrug, because his signal does not know those walls exist. Terrain is not part of his propagation path. That is not bravado; it is physics, and it is a capability no other band on this list can offer.

The VHF operator is the one with the repeater map laminated in his glovebox, and he finds the HF question almost impolite. HF, he will say, is useless when you need to talk to the person two ridges over, right now, before the light fails and the temperature drops. He is right too, and more right than his dismissiveness suggests. VHF transmits on a radio the size of a candy bar that runs for days on a battery you can charge from a power bank, from an antenna short enough to hang on a shoulder strap, and it is the band on which nearly every search-and-rescue organization in the country keeps its ears open. Reach, in his world, is measured in reliability at three miles, not in distance to the horizon.

The GMRS pragmatist is the third voice, and he has stopped listening to both of them. He cannot understand why the group is arguing about propagation at all when one inexpensive license covers his entire household for a decade, when he can hand a radio to a twelve-year-old and to a grandparent and they will both be legal and both be understood, and when his family has never once, in six seasons of camping, actually failed to stay in contact over the distances that matter. He is right as well, and his correctness is unfashionable: his problem is coordination, not rescue, and for coordination the compromise band is not a compromise at all.

The satellite messenger owner is the fourth, and he is the least interested in spectrum. His device is slow, its display is primitive, and it is good for roughly a hundred characters of text, and it has never once failed to deliver a message to his wife from a place with no cell service. He does not care about band plans because he buys bandwidth from satellites, and he will point out, correctly, that a short message from anywhere beats a long conversation from nowhere.

The CB traditionalist is the fifth, and he is the only one who is smiling. Twenty-seven megahertz, forty channels, no license, no exam, no fees, and he has talked across three states at dusk with a radio older than some of the people in the argument. He is right under specific conditions, and he knows the specific conditions precisely, which is why his certainty is not naïveté but a very narrow kind of expertise: he is the operator who understood that the question includes the word when.

And then there is the beacon owner, who says nothing, because he has nothing to add to a conversation about conversations. His equipment does not carry a voice, does not hold a dialogue, and does not care what the weather is doing to the ionosphere. It is also the only equipment in the group with a documented and auditable rescue record — tens of thousands of people recovered over four decades of satellite alerting^[1], alerts landing in a rescue coordination center within minutes, positions accurate to a hundred meters, identity and next-of-kin information already on file before the first radio call is placed. He is the only one of the six whose device is optimized for the actual worst day rather than for a good afternoon.

Every one of those six statements is defensible, and the reason is not that radios are mysterious or that operators are stubborn. It is that the six of them are answering six different questions and then mistaking the disagreement for a technical one. The HF operator is answering: how far can a signal travel with no infrastructure and no satellite? The VHF operator is answering: how do I speak to someone who is nearby but invisible? The GMRS pragmatist is answering: how do I legally equip an ordinary family for a weekend? The messenger owner is answering: how do I send a small, high-value message from anywhere to anywhere? The CB veteran is answering: what can I do with no license and infinite patience? The beacon owner is answering: how do I guarantee that somebody with a helicopter knows exactly where I am? Those are not competing answers to one question. They are answers to six questions, and the phrase “emergency communications” blurs them all into one.

That blur is the category error, and it is not a semantic quibble. It is the single most expensive mistake a backcountry party can make, because it is the mistake of believing that one well-chosen device has made a plan. In reality, “emergency communications” is three jobs wearing the same coat. The first is summoning rescue — getting a message to strangers with helicopters, and doing it once, reliably, with no second attempt. The second is local coordination — keeping a party together, reporting a route, calling off a search that has already been solved on foot. The third is situational awareness — weather warnings, closures, the general picture. To these a fourth is often added implicitly, and it is the one that saves more people than the other three combined: not needing to communicate at all — the itinerary left with a neighbor, the turnaround time honored, the two hikers who never separated because they agreed not to. The best emergency radio transmission, considered in the aggregate across all the rescues that never happened, is the one that was never sent.

Each of those jobs rewards a different part of the spectrum, and a band that is excellent at one is frequently poor at another. That is not a flaw in the design of the spectrum; it is the whole reason the spectrum has so many occupants.

“Best,” in turn, is not a property of frequency alone. It is the product of four constraints that pull against each other, and the resolution of those constraints is what an actual decision looks like.

The first is what propagation physics permits — whether the signal bends around a ridge, refracts off the ionosphere, hugs the ground, or requires an unobstructed straight line to the other antenna. The second is what antenna you can physically carry, and this is the most brutal of the four, because antenna efficiency is a function of size measured against wavelength, and a person on foot is working with a fixed budget of a few feet and perhaps a kilogram. At two meters, a quarter-wave antenna is nineteen inches of thin flexible wire. At eighty meters, the same fraction of a wavelength is a hundred and thirty feet of copper that must be supported by something, and the something is generally a tree that may or may not exist where you are. The third constraint is what your power budget allows — the weight of a battery, the duration you must sustain a watch, the temperature you must operate in — and it interacts viciously with the first two, because every watt of transmit power is paid for in grams and hours. The fourth, and the one most often ignored, is whether anybody is listening on that band, at that hour, in that region. A theoretically infinite range means nothing if the far end is silent. A theoretical 1.5-mile range means everything if the far end is your partner’s helmet and 1.5 miles is the whole problem.

Beyond those four there are two more constraints that are not physics but behave exactly like physics in practice: what the law in your jurisdiction permits you to transmit, and how much skill you have actually built. Both impose hard limits that no equipment purchase can remove. A fifty-watt radio in the hands of an operator who has never deployed its antenna is a smaller system than a five-watt radio in the hands of someone who knows what the top of a ridge is worth. Similarly, a beautifully capable band you are not licensed to use has a range of exactly zero for your purposes, however tempting it looks on a specifications sheet.

There is one final wrinkle in the word “best,” and it is statistical rather than physical. Communications for an ordinary trip and communications for an emergency are not the same optimization problem, because an emergency is by definition an event in the tail of the distribution. The day you need the system is not the average day; it is the day the weather turned, the trail was longer than the map said, the ankle rolled, the radio went into the creek. Gear chosen to perform best on average days and acceptable on bad ones has been optimized for the wrong days. What an emergency plan really buys is a guarantee of performance under the conditions you cannot choose, which is why the boring device with the documented record so often beats the elegant one with the impressive specifications, and why the constraints above deserve to be argued about in the cold and in the rain rather than in a forum.

This essay works through the physics first, then each band on its own terms, then a verdict. The short version, stated up front: VHF is the best single band for two-way backcountry voice, UHF carries the one signal most likely to save your life, HF is the only band that can reach beyond the horizon without a satellite, and LF and VLF are a beautiful lesson in electromagnetics that nobody can backpack.

The Question Behind the Question

#

The three jobs are easy to state and hard to keep separate, because they fail in different ways, on different timescales, and according to different metrics, so comparing them directly is a little like ranking a seatbelt against a spare tire. A concrete scene for each will make the distinction sharper than any definition.

The analogy is not merely decorative, because it predicts the shape of the correct answer. A spare tire and a seatbelt cannot be ranked against each other in the abstract, yet each is unbeatable on its own metric, and nobody argues that a car should carry only one of them. The same is true of the spectrum, with one important difference: a car and a backpack both have a fixed budget of weight, space, and money, so the jobs below are not simply different — they are competing for the same small allowance of grams and watt-hours. That competition, and not any property of electromagnetic waves, is what makes the question hard, and it is also what makes it answerable, because a budget forces a decision that a preference never will.

Picture a party of three in October, one of them on the ground at 3:40 in the afternoon with a broken ankle, six miles from the trailhead, a mile and a quarter down a steep timbered drainage, and no cellular coverage for thirty miles. This is the summoning job in its purest form, and note what it does not need. It does not need a conversation, a protocol, a clear channel, or a courteous reply. It does not even need much bandwidth. It needs exactly one thing: a message that leaves the drainage, crosses whatever distance is required, and arrives in the hands of people who own a helicopter and have already been told where to look. The correct metric is probability of a successful alert under the worst plausible conditions — the sky socked in, the battery cold, the operator frightened and inexperienced, the first two attempts apparently ignored. Everything about that list argues for a technology that does not require anyone at all to be listening at the moment of transmission, that repeats itself indefinitely without human intervention, and that carries its own location. In other words, it argues for a beacon rather than for a radio, which is precisely the intuition the HF and VHF operators in the first scene are missing: they are optimizing a conversational tool for a non-conversational job.

It is worth being precise about what a voice radio call actually requires in that drainage, because the limitations are structural rather than a matter of a better radio. A successful call requires, in sequence: that someone happens to be monitoring the frequency you chose, within range of a path that reaches you through eight hundred feet of rock and eighty feet of wet conifer canopy; that the same path works in the reverse direction, because a distress call you cannot hear being acknowledged is only half a transmission; that this stranger is competent and willing to act; that they correctly transcribe a location described verbally, by someone in pain, in poor light, under stress; and that they then relay it to an organization that can actually dispatch a helicopter. Every link in that chain is a place for the message to die, and the first link — a listener — is the weakest, because the frequency is not unattended in the abstract; it is unattended in the particular. Nobody owes the channel their attention. That is what makes the summoning job a lottery rather than a system, and a lottery is an odd thing to stake a rescue on.

There is also the uncomfortable question of who is operating the radio. The person with the swollen ankle, or the person watching them, is the least capable operator the party will field all day. Protocol recall is among the first faculties to degrade under stress, taking with it the correct channel, the correct call format, and the patience to hold the radio at the right height rather than against a chest. Speech itself degrades, and a voice message is a terribly inefficient container for the two facts that matter, which are an identity and a position. A voice carries fear, accent, unfamiliar place names, and the plain fact that the operator has never thought about any of this before, which is precisely why voice contact so often becomes a frustrating repetition of the phrase “say again” while the light drains out of the sky.

Contrast that with the architecture of a beacon, which is engineered to be detected rather than answered. Nothing about it depends on goodwill, attention, or timing. It does not need a channel, an operator who can speak clearly, or anybody at all to be listening at the moment of transmission, because the thing that hears it is not a person but a machine in orbit, and the thing that acts on it is not a helpful stranger but an institution whose entire reason for existing is to answer the alert. It repeats itself indefinitely without human intervention, which converts the summoning job from a one-shot attempt into a process that continues while the party is busy being cold. It carries a registered identity and a position, which dissolves the ambiguity that makes strangers hesitate to act, and it transmits the same message whether the operator is calm, weeping, or unconscious. The one thing it cannot do is hold a conversation, and a conversation is exactly what the summoning job never required.

Which leads to the failure point that sits upstream of every radio and every beacon, and that equipment cannot solve at all: the decision to activate. Injuries and weather rarely kill suddenly; they kill on a clock, and the clock is usually one of exposure, dehydration, or shock rather than of the injury itself. The transmission that summons a helicopter at four o’clock is by far the most valuable message in the entire scenario, and yet the most common reason it is not sent is not that the signal failed but that nobody sent it — because the party still believed they could walk out, because an earlier attempt had felt humiliating, because a rescue has a social weight and possibly a financial one, and because saying out loud that a situation has exceeded the group’s ability is a genuinely difficult thing to do. Every hour of delay is spent in the most expensive currency available. This is why the right summoning device is not merely the one with the longest reach; it is the one whose activation requires no negotiation, no explanation, no channel discipline, and no courage beyond pressing a single button. Friction in that one decision is the only specification on this list that is measured in body temperature.

There is one more design property worth noticing, because it explains some of the bureaucracy that surrounds beacons. A system built for a one-shot, worst-case, non-conversational job must assume that some of its messages are wrong, and it must therefore resolve the ambiguity of any given alert without compromising the response to a real one. Registration exists for exactly this reason: an alert that carries a vessel or person’s details can be cross-checked, next of kin can be telephoned, and a false alarm can be dismissed in minutes rather than triggering a search. The economics are deliberately asymmetric, because a false alarm costs a nuisance, an apology, and sometimes a fine, while a missed alert costs a life, and no rational design optimizes for the cheaper failure.

Now picture something much less dramatic and far more common. Two hikers separate at an unsigned junction at dusk, each assuming the other took the obvious fork. They are eight hundred yards apart, on opposite sides of a low forested shoulder, each within sight of the same lake and each invisible to the other. This is the coordination job, and its metric is nothing like the first one. Nobody is dying, the stakes at any single moment are trivial, and yet the transmission must succeed dozens of times over the course of a day, in the rain, while both people are walking, using equipment that costs less than a tank of fuel. The failure here is almost never a failure of range, because eight hundred yards is nothing. It is a failure of protocol — nobody agreed on a channel, nobody agreed on a check-in interval, and so both parties transmit into silence, conclude the other is out of range, and start making independent decisions that will turn a trivial separation into a genuine search. The band that solves this problem is the one that is small, cheap, boring, and always switched on, which is why the unglamorous compromise bands dominate the real world of backcountry voice.

Note what that scene implies about where the leverage actually is. The technology that solves it is not a radio at all; it is a protocol. Four decisions made in a parking lot before anyone puts boots on — we will all use the same channel, we will check in at every trail junction and at the top of every hour, nobody walks out of sight of the last person without saying so, and if we cannot raise each other within fifteen minutes we both stop where we are and wait, because the one thing we will never do is wander separately — cost nothing, require no license, and work identically on a five-dollar blister-pack radio and on a five-hundred-dollar one. A band cannot compensate for the absence of those four decisions, and the presence of them can make even the least capable band adequate. This is the least glamorous and most reliable truth in the entire field, and it is why the coordination job is dominated by human factors that no specification sheet lists.

The coordination job is also, uniquely among the four, an economics problem, because the device must be replicated across every member of the party rather than issued to one responsible adult. Cost per head becomes a real constraint, and with it a set of design requirements that sound trivial and are not: one battery chemistry and one charger type for the whole group, so that five radios can be topped up from a single power bank; buttons that work through gloves; water resistance sufficient for a hand that has just crossed a creek; a form factor that survives being dropped, sat on, and handed to an eight-year-old. The social dimension is equally real. The radio nobody uses because using it feels awkward or childish is not a communications device at all. Reducing the embarrassment of the radio check down to zero is a legitimate engineering requirement, because a group’s willingness to press the button is the actual specification being tested.

There is a commons to consider here as well. Coordination channels are shared with every other party in the region, which is why etiquette — keeping transmissions short, leaving a beat between exchanges, saying again rather than guessing — is not decoration but load-bearing infrastructure, and why crowded areas produce the characteristic coordination failure of talking over one another in the same narrow set of channels. This is also the only argument in favor of a less popular band that holds up on its own merits, since an uncrowded channel with a mediocre antenna often beats a famous channel at the trailhead register.

And there is a darker implication hiding in this job, which is where the category error acquires a body count. Radio packaging claims ranges in ideal conditions, and a party that has bought a coordination device believing it to be a summoning device has not merely failed to solve the second problem; it has changed its behavior in response to a capability it does not have. The comfortable feeling of carrying a radio that advertises thirty-five miles is exactly the sort of reassurance that permits a decision to continue past a turnaround, to split up at a junction, or to attempt a route at the edge of the group’s ability, which means the equipment has quietly increased exposure while appearing to reduce it. A coordination tool is worth having and is not a rescue plan, and any kit that blurs the two is worse for the blurring. Nor is the damage confined to the party itself: the searches generated by separated parties consume helicopters, volunteer hours, and alerting capacity from a pool that is finite on any given weekend, and they do so at the expense of whomever that weekend’s weather or luck has put in trouble, which is why a disciplined group is, in a modest but real sense, part of the region’s rescue infrastructure.

The third job looks passive and is therefore chronically underfunded. A thunderstorm line builds across the far ridge at two in the afternoon; a wildfire closure is announced at four; the pass that was open when the party left the pavement is shut when they reach it. This is the situational awareness job, and its metric is availability rather than performance — not “can this reach someone?” but “is it listening, right now, all the time, at no cost?” Its characteristic failure is not interference, distance, or terrain but the far more mundane fact that the receiver was left in the truck, switched off to save a battery, or tuned to a channel nobody broadcasts on. A great deal of wilderness safety is produced by a device that only receives, and the cheapest device that only receives may be the highest-value item in the entire kit.

What makes this job unusual is the character of its asymmetry, which is sharper than in either of the previous two. Information has a half-life. A closure posted at four in the afternoon is worth nothing at nine in the morning and everything at four, and the same is true of the cold front building across the far ridge, the river gauge upstream in a different drainage, and the fire that has been burning since the previous evening in a canyon nobody in the party has heard of. The fact does not degrade gently; it decays on the timescale of the hazard it describes, and weather decays in minutes. Late information does not merely reduce the comfort of a plan — it converts a plan into a mistake, and it does so without the party noticing, which is why this job’s failures are so often discovered only in the retelling.

The economics of the third job are also the best in the entire discipline, and they are the reason it is faintly embarrassing that it goes unfunded. There is no such thing as a false positive while listening: nothing you hear can hurt you, no listener need be licensed, no transmission can be misconstrued, and no etiquette applies. The equipment is the cheapest category in the field, frequently costs less than a replacement battery for the radios it accompanies, and is very often already built into the handhelds the party is carrying. Listening also protects the battery that the other two jobs depend on, so it pays for itself in the currency of uptime. A tool that costs almost nothing, carries no penalty for overuse, and reduces the probability that the other two jobs will ever be needed is, by any reasonable measure, the first thing anyone should buy — and it is routinely the last, because its benefits are invisible on the days it works.

It is worth adding that the receiver need not be electronic at all. The thunderstorm line building over the far ridge, the sudden shift in wind direction, the smell of smoke in still air, the fact that it has been raining hard two drainages upstream, the color of the water and the sound of the river, the grass bent flat along the current — a person who notices these things has situational awareness without a receiver, using instruments already in the kit. What an electronic receiver buys is range and authority: the same information, hours earlier, from an institution that has already done the analysis. Which is another way of stating the oldest rule of radio, that the ear is more useful than the mouth, and of explaining why listening deserves the same deliberate place in a plan as a check-in schedule does.

One more scene belongs here, and it is the quietest of the four because nothing happens in it. Two hikers turn back at three in the afternoon, not because of weather or injury but because three o’clock was the turnaround they agreed on in the parking lot, and they are eating sandwiches beside the truck by a quarter past five. No radio was switched on, no beacon was activated, no message was sent to anyone, and the rescue that might have consumed a helicopter, four volunteers, and a night of somebody’s sleep never had to be organized because it never had to happen. This is the fourth job, and it is the only one in the list that reduces the need for communication rather than improving the means of it.

Its instruments are the least technological items in the kit, which is precisely why they survive conditions that defeat everything else: an itinerary left with a named person who has been told not only where the party is going and when they expect to return, but what they should do when the time passes, whom they should telephone, and what the vehicle looks like. Nothing about that plan involves a battery, a frequency, an antenna, or a signal path, which is what makes it the last system standing after every electronic device in the party has failed, been soaked, or been left in the truck. Its weakness deserves to be stated plainly: it works on the timescale of hours, it requires somebody at the other end to notice and to act, and it can convey the fact of an emergency but never its urgency or its position. Its strength is that it costs nothing and it is entirely within the party’s control.

Its real difficulty is not design but compliance, because a turnaround that is negotiated in the moment is not a turnaround at all, and the person who most needs the discipline is the one having the best day. Plans of this kind fail in a characteristic way: they are made vaguely, shared with nobody, revised upward on the mountain, and then reported to a spouse who has no idea which forest district is involved. Writing the plan down, naming a person, and treating the turnaround as a commitment rather than a preference is unglamorous work that has almost certainly saved more people than any radio ever manufactured.

Place the four scenes side by side and the reason the argument never converges becomes obvious. The first job is one-shot and must be optimized for the worst case; the second is continuous and must be optimized for the everyday; the third has no transmission at all and must be optimized for availability; and the fourth is not an optimization problem at all but a discipline. They are measured in different units — probability of a single success, tolerance of repeated failure, simple uptime, and avoided need — and each is dominated by a different physical or human constraint. The one-shot job is dominated by propagation reach and by the question of who is listening without being asked. The everyday job is dominated by antenna practicality, battery life, and cost per unit, because you must equip everyone. The listening job is dominated by nothing at all except the discipline to keep the receiver on. The avoidance job is dominated by honesty about one’s own ability and about the time of day.

The economics of error diverge just as sharply, and this is the part that most often goes unexamined. In the summoning job, a false alarm is an embarrassment and possibly a fine while a missed alert is a fatality, so the correct design deliberately tolerates false positives, repeats itself without permission, and errs toward acting. In the coordination job, a false alarm costs a few seconds of awkwardness in a channel shared with strangers while a missed transmission costs a search that never should have been launched. In the listening job there is no error at all, because nothing you hear can harm you and nothing you fail to transmit can be misunderstood. In the avoidance job the accounting runs the other way entirely: the cost of caution is a shortened trip, and the cost of recklessness is borne by people who have never met you. A plan that optimizes the wrong job’s failure mode is how a party ends up with excellent equipment for a problem it does not have.

The consequence is that the four jobs want devices of different kinds, not merely different bands. The summoning job wants a purpose-built, single-button, one-way device that asks nothing of its operator and forgets nothing. The coordination job wants cheap, rugged, replicable handhelds with one battery chemistry and no learning curve. The awareness job wants a receiver simple enough that it is never left behind and a habit strong enough that it is never switched off. The avoidance job wants paper, a pencil, and a person at the other end who has been given specific instructions. Choosing one device to satisfy all four is how well-equipped parties end up, in the honest retelling afterward, unable to help themselves, and it is why the verdict at the end of this essay is not a single band but a sequence.

With the jobs separated, it is finally possible to ask the physics question honestly — which part of the spectrum serves which job — without the answer being contaminated by the question “which radio should I buy?” The next section takes the physics on its own terms, and the shopping follows much later, as a consequence rather than a premise.

The Physics That Decides Everything

#

Four physical facts, more than any specification sheet, determine which bands work where you actually are.

Stated compactly, and in the order the sections below take them, those facts are these. Above roughly 30 MHz a signal travels in a straight line, so its range is set by the radio horizon rather than by the power behind it. Every obstacle along that path bends the wave rather than blocking it outright, and the longer the wavelength the more graciously it bends, while free-space loss grows with the square of frequency — which means the lower bands reach around terrain and through wet canopy more cheaply, and the higher bands pay for their small antennas. Between roughly 3 and 30 MHz, a signal can be thrown at the sky and returned to earth by the ionosphere, carrying a modest transmitter across thousands of miles with no infrastructure at all, and behaving differently at noon than at midnight. And below HF, two slower and more earthbound mechanisms take over: ground wave, which hugs the surface of the planet and follows its curvature, and, in the lowest bands of all, a spherical duct formed between the ground and the ionosphere that can carry a signal much of the way around the world.

Those four mechanisms are not equally available at any given frequency, and their availability is what the rest of this section is about. The first two govern the entire line-of-sight half of the spectrum, which is to say VHF, UHF, and everything above. The third governs HF and nothing else. The fourth governs MF, LF, and VLF, where the planet itself becomes part of the antenna system. Every band this essay will eventually grade is descended from one of these four, and no band escapes its parent.

It is worth being clear about why these facts outrank the numbers printed on a radio’s box. A specification sheet describes what a device does into a dummy load on a bench under ideal conditions; it says nothing about what the environment will do to the signal a hundred yards from the antenna. Transmit power, receiver sensitivity, and modulation format all matter at the margin, and none of them chooses the path. A mediocre radio on the right band will succeed where an excellent radio on the wrong band fails, and a great deal of money spent on watts is spent attempting to buy a path that the physics was never going to provide. In the backcountry, the specification sheet describes the messenger and the physics describes the road.

There is a further property that makes these four facts more than a list of quirks, and it is the reason the spectrum behaves like a set of trades rather than a ladder from bad to good. All four are consequences of a single variable — frequency, and therefore wavelength — that cannot be optimized in both directions at once. Long wavelengths bend around ridgelines, lose less in foliage, and follow the earth’s curvature, but they demand antennas measured in tens or hundreds of feet and surrender bandwidth. Short wavelengths let a nineteen-inch whip do real work and carry enormous amounts of information, but they insist on a clear path and pay heavily for leaves, rain, and distance. Nothing in this section is a defect that a better budget can engineer away, because each is the same coin viewed from the other side. Choosing a band, therefore, is not choosing between a good and a poor option; it is choosing which penalties you would rather carry.

There is one more distinction inside these four facts that shapes the entire essay, and it is not about range at all but about character. Three of the mechanisms are comparatively stable and therefore comparatively predictable: the radio horizon can be reasoned about from a map, diffraction from the shape of a ridge, ground wave from the conductivity of the soil. Only the ionosphere is a living medium, changing with the hour, the season, and the eleven-year solar cycle, which is why HF’s reach is the greatest in the spectrum and the least guaranteeable. The practical consequence is a pairing that runs through every recommendation that follows: the bands that reach furthest are the ones whose reach you cannot promise, and the bands that never reach far are the ones that always arrive, which is why serious plans carry both rather than choosing between them.

Each of the four is developed below on its own terms, with its own consequences for a person on foot.

Line of Sight, and Why Elevation Beats Watts

#

Above roughly 30 MHz, radio behaves much like visible light. It travels in a straight line, does not follow the curvature of the earth, and is stopped by anything solid with a large gradient in its path. This is why VHF and UHF are described as line-of-sight bands, and why the radio horizon, not the transmit power, sets your range. The practical rule of thumb for a ground-level path is d_mi ≈ 1.23√h_ft: two people standing six feet tall with handhelds theoretically see about six miles between them, less in practice. Put one of them on a thousand-foot ridge, however, and that horizon stretches to nearly forty miles. That horizon is not even the geometric one; because the atmosphere refracts the wave gently downward, radio sees a little further than the eye does, roughly fifteen percent beyond the optical horizon, which is the small mercy that lets a signal occasionally arrive from beyond the visible crest of a hill.

The background to that formula is worth spelling out, because the whole geometry of a line-of-sight radio path falls out of it. The refractive index of air at sea level is about 1.0003 and decreases upward at roughly 4×10⁻⁸ per meter, or about forty N-units per kilometer, which means a ray launched horizontally in a standard atmosphere is not straight at all; it bends gently toward the earth. The conventional way to handle that curvature is a fiction: pretend the atmosphere has no gradient and inflate the planet instead, replacing the true earth radius R_E ≈ 6371 km with an effective radius kR_E where the standard factor is k = 4/3, giving about 8495 km. The distance to the horizon for an antenna at height h then follows from the Pythagorean theorem, d = √(2kR_Eh), which in convenient units becomes d_km ≈ 4.12√h_m and, in the units we actually think in, d_mi ≈ 1.23√h_ft. For two antennas the two horizons add, so the working form is d_mi ≈ 1.23(√h₁ + √h₂), which is why the number you want is not your elevation above sea level but the clearance of your antenna above the tangent line to the obstruction.

Nearly every practical number in this subsection falls out of that expression. Two hikers of six feet each share a horizon of about six miles. Raise one of them to thirty feet, on a rock outcrop or a truck roof, and the pair can see roughly nine and three-quarter miles. Put one of them on a thousand-foot ridge, with the other still six feet tall, and the path extends to nearly forty-two miles — an increase of sevenfold for the cost of a long walk uphill. Note also that k is not a constant but a weather variable. Temperature inversions and humidity gradients can push k above 4/3 and even bend the path into a tropospheric duct that carries VHF and UHF signals hundreds of miles over water or flat country, while sharp lapse rates push k below 4/3 and contract the horizon. This is why marine VHF occasionally astonishes its users, and why a path that worked on Tuesday may not work on Thursday, which is not a reason to distrust the formula but to distrust any single afternoon’s field test.

That single formula explains the most important operational lesson in this entire article, and it has nothing to do with frequency: a 5-watt handheld on a ridge routinely out-communicates a 50-watt mobile in a canyon. Radio is a geometry problem before it is a power problem. Any doctrine that starts with “buy more watts” and not “walk uphill” has skipped the physics.

To see how completely the geometry dominates, put the two descriptions of the same path side by side. The radio half is the link budget, and its governing equation is the Friis transmission formula, P_r = P_tG_tG_r(λ/4πd)², which in decibel form becomes P_r = P_t + G_t + G_r − FSPL, with the free-space path loss given by FSPL = 32.45 + 20 log₁₀ f_MHz + 20 log₁₀ d_km, or equivalently 36.6 + 20 log₁₀ f_MHz + 20 log₁₀ d_mi when distance is measured in miles. Every practical quantity in radio is logarithmic because the range of magnitudes involved is so vast: a five-watt handheld is 37 dBm, a receiver that hears 0.2 microvolts across 50 ohms is hearing roughly 8×10⁻¹⁶ watts, a little under a thousandth of a picowatt, which is −121 dBm, and the thermal noise floor against which those fractions of a femtowatt are detected is kTB, which for a 12.5 kHz channel at room temperature is only about −133 dBm.

Assemble those numbers and an uncomfortable result appears. A handheld transmitting 37 dBm, with a couple of decibels of gain at each antenna and a receiver threshold near −120 dBm, has an allowable path loss of roughly 160 dB. On the two-meter band, free-space loss reaches 160 dB at a distance of roughly sixteen thousand kilometers, better than a third of the way around the planet. The link budget of a modest handheld is therefore not remotely the binding constraint on where its signal can go. The horizon is. A five-watt radio is not weak; it is simply pointed at a wall of planetary curvature, and everything else in this section is a discussion of the shape of that wall.

This gives the quantitative form of the claim that elevation beats watts. Transmit power enters the budget logarithmically, so doubling it buys 3 dB, which in free space extends range by √2, or about 1.41 times. Antenna height enters the geometry as a square root, so multiplying the height by ten — six feet to sixty feet — multiplies the horizon by √10, or 3.16, which is exactly equivalent to 10 dB of link improvement: a tenfold increase in transmit power, purchased for the weight of a branch and a length of coax. Meanwhile the losses a ridge actually imposes are not gentle slopes but cliffs: a single obstructing ridgeline commonly costs 20 to 40 dB of knife-edge diffraction loss, which is to say a hundredfold to ten-thousandfold increase in power would be required to buy the same result, and no battery, antenna tuner, or license fee available to a person on foot can supply that. The corollary is one of the few genuinely free pieces of advice in this essay: before spending money on watts, spend twenty minutes with a topographic map identifying the ridge, bald, or fire tower the party could actually reach.

There is a second geometric ceiling that governs the flat case, where no ridge intervenes at all. When both antennas sit close to the surface, the direct ray and the ground-reflected ray arrive nearly out of phase and largely cancel one another, and the received power stops falling as 1/d² and begins falling as 1/d⁴, following P_r ∝ (h_th_r)²/d⁴, which is 40 dB per decade of distance instead of 20. The breakpoint where the steeper regime takes over sits near d ≈ 4πh_th_r/λ, and for two handhelds held 1.5 meters above the ground on the two-meter band that works out to about fourteen meters — walk past the end of a parking area and you are already in it. Real terrain is rougher and more absorptive than a perfect reflecting plane, and the measured exponent usually lands between 2.5 and 4, but the shape of the curve is the same, and it is the honest explanation for why a pair of handhelds is a one-to-three-mile tool on level ground while the same radios on packaging achieve twenty miles over water.

There is also a subtlety buried in the phrase line of sight, because a radio path does not need a thin string of visibility; it needs a corridor of a specific width. The first Fresnel zone — the region around the straight line where direct and indirect paths differ by less than half a wavelength — has a radius r₁ = √(λd₁d₂/(d₁+d₂)) at a point dividing the path, and the usual engineering rule is to keep sixty percent of that radius unobstructed. For a ten-kilometer link on two meters, that radius is about seventy meters at midspan, so roughly forty-two meters of clearance above the treetops is required. The same ten kilometers at 900 MHz needs a corridor only about twenty-nine meters wide, needing seventeen meters of clearance. The lower band is simultaneously more tolerant of a rough path and less tolerant of a tight one, which is one more way of saying that VHF wins the long shots and UHF wins the cramped ones.

One final frequency-dependent quantity belongs here, because it is the honest reason VHF is the band of choice for weak-signal work. Wavelength follows λ_m ≈ 300/f_MHz, so two meters and seventy centimeters differ by a factor of nearly three by definition. An antenna captures power in proportion to its effective aperture, which for an idealized half-wave dipole is A_e ≈ 0.131λ², with gain following G = 4πA_e/λ². Concretely, a two-meter dipole intercepts about half a square meter of wavefront while a seventy-centimeter dipole intercepts about a sixteenth of one — a difference of roughly eightfold, or nine decibels, between two antennas of the same physical style on the same path. Nine decibels is the difference between a signal and an interesting noise floor, and it is why a nineteen-inch whip on the low band will outperform a six-inch whip on the high band at the same distance whenever the path is marginal.

The consequence of all this arithmetic is that the formula d_mi ≈ 1.23√h_ft is not a curiosity to be memorized for a licensing examination but a map-reading tool. It tells you, before you leave the pavement, which ridges, balds, and fire towers are worth the detour, and it tells you that the party’s radios will be worth more on top of them than any conceivable upgrade in power. A serious backcountry plan should therefore specify not only what is carried but where it is intended to be used from, because in this band family the location of the antenna is a larger variable than every specification printed on the box.

Diffraction, Foliage, and the Cost of Frequency

#

Obstacles rarely block a signal completely. Radio waves bend around ridgelines and the edges of obstructions in a process called knife-edge diffraction, and the longer the wavelength, the more efficiently they do it. Free-space path loss, meanwhile, grows with the square of frequency: every tenfold increase in frequency costs 20 dB, meaning roughly a hundredfold reduction in power density at the same distance. For every additional decade of frequency, you must make that up in antenna gain, transmit power, or proximity.

Take the first claim seriously and it becomes a formula. When a ridge intrudes into the path, the obstruction behaves approximately as a knife edge, and the entire geometry of the problem collapses into a single dimensionless number, the Fresnel–Kirchhoff diffraction parameter ν = h√(2(d₁+d₂)/(λd₁d₂)), where h is the height of the obstruction above the straight line joining the antennas and d₁, d₂ are the two segments of the path. Everything that can be said about one edge follows from ν, and the frequency argument is visible in the square root of the denominator: because ν varies as 1/√λ, the identical ridge presents a numerically smaller edge to a longer wavelength. The loss attached to that parameter is given to good approximation by L = 6.9 + 20 log₁₀(√((ν − 0.1)² + 1) + ν − 0.1) decibels^[19], which rises monotonically with ν and therefore with frequency. Put numbers to it and the effect is real but sober. A ridge standing thirty meters above the line on a five-kilometer path, with the obstruction at the midpoint, yields ν ≈ 0.83 on the two-meter band for a diffraction loss of about 12.8 dB, while the same ridge at 900 MHz yields ν ≈ 2.08 for about 19.4 dB. The difference between the bands is thus roughly 6.6 dB per obstructing edge, which is worth having and is also not the orders of magnitude that the phrase “VHF bends around terrain” can suggest.

Where the effect becomes consequential is in the plural, because real mountain terrain is never a single edge. Engineering practice reduces a multi-obstruction path either to one equivalent edge by Bullington’s construction or to a sequence of edges by Deygout’s method, in which each obstruction is evaluated in turn and the individual losses are added with a small correction. Since diffraction losses accumulate rather than cancel, a path that must clear three successive ridges pays the penalty three times over, and a marginal path that costs 6.6 dB more per edge at UHF becomes 20 dB more expensive across the group, which is the difference between a conversation and a hiss. This is precisely why the two-meter band has the reputation it does in mountainous country, where five watts and a nineteen-inch whip routinely hold contacts of fifty to a hundred miles between a summit and a valley floor, and why seventy centimeters, for all its convenience, is the weaker companion in exactly the terrain where radios matter most.

One clarification about the cost of frequency belongs here, because it is routinely misunderstood. Free-space path loss, FSPL = (4πd/λ)², is not a loss of energy to any medium; nothing is absorbed and nothing is dissipated. The wave simply spreads over a sphere of area 4πd², and the λ² enters through the receiving antenna’s aperture, A_e = Gλ²/4π. Written in its honest form the received power in free space is P_r = P_tG_tA_r/4πd², an expression containing no frequency at all. The frequency penalty appears only when the aperture is held constant, which is the practical case, because a person carries an antenna of a given physical size and not a given gain. This is why a tenfold increase in frequency costs 20 dB, or 6 dB per octave, and it is also why the penalty is repayable: the same 20 dB can be recovered by antenna capture area, by tenfold proximity, or by a hundredfold increase in power. The entire lower-versus-higher band argument is a negotiation over which of those three currencies is available in a rucksack.

That is the honest scorecard on lower-versus-higher bands in the backcountry. VHF at 2 meters and MURS at 1.5 meters bend and creep over rolling terrain better than 70-centimeter or 900-MHz signals, and they punch through dense canopy with less loss. This is worth stating plainly, because you will read — including in some of my own earlier writing on this subject — that UHF penetrates foliage and obstacles better than VHF. That is a half-truth borrowed from urban radio. In cities, higher frequencies get into and out of buildings more effectively, scatter around concrete canyons, and let you build an efficient antenna the size of a pencil. In a forest, the physics reverses: leaves and branches are lossy absorbers, and attenuation increases with frequency. The correct formulation is not “UHF gets through foliage” but “UHF gets through cities, and does so with a radio you can put in a shirt pocket.”

The foliage half of the story rests on two mechanisms rather than one, and both of them point the same way. The first is absorption, whose root is water: a path through vegetation is largely a path through the water content of leaves and needles, and the dielectric loss of water in this part of the spectrum increases with frequency, which is the same property that makes 2.45 GHz the frequency of a microwave oven. The second is scattering, and it depends on the relationship between wavelength and the physical size of the scatterers. Twigs, needles, and leaves run from millimeters to tens of centimeters, which makes them electrically small at two meters, where they perturb the wave only mildly, and resonant — comparable to a half wavelength — in the UHF region and above, where they reradiate, depolarize, and scatter energy out of the wanted path. Both mechanisms therefore grow with frequency, both are strongest in leaf, and wet foliage adds further loss because the water film on every surface is thicker than the water inside it.

Empirical models capture this as a power law of the form L = Af^bd^c, with the frequency exponent typically between 0.2 and 0.5 and the depth exponent between 0.3 and 0.6, depending on canopy and campaign. Weissberger’s widely used fit for medium-density foliage, valid for depths of roughly fourteen to four hundred meters, is L ≈ 1.33 f_GHz^0.284 d_m^0.588 decibels^[20], and running that expression through a thirty-meter depth of trees gives about 5.7 dB at 145 MHz, 9.5 dB at 900 MHz, and 12.6 dB at 2.4 GHz. The honest conclusion is not quite the one the folklore predicts. Frequency does matter, and it matters in the direction claimed, but the difference between the two-meter band and the 900 MHz band in the same forest is a few decibels, while the depth of the forest is worth a dozen. Anyone choosing between bands on the strength of canopy penetration is optimizing the smaller term; anyone genuinely worried about the forest should be worried about how far the trees extend, since it is the exponent on depth that dominates. Dense conifer canopy in the 100–400 MHz region is commonly measured at a few tenths of a decibel per meter of depth, so forty meters of timber costs the better part of ten decibels at either band, and horizontal polarization, which couples more strongly to vertical trunks and branches, adds several more and is the reason a handheld should be held upright in woods.

So where does the familiar claim come from? From cities, where the physics genuinely does invert, and it is worth being specific about why. Building penetration loss — the deficit between a signal outside a structure and the same signal inside it — is broadly smaller in the UHF region than in the VHF region for the ordinary case of a signal arriving from outdoors: representative figures run around 20 to 30 dB for VHF entry into commercial structures and roughly 10 to 20 dB at 900 MHz. The reason is partly scale and partly geometry. At 100 MHz the wavelength is three meters, so a building is not large compared with the wave; walls, floors, and roof framing sit in the near field of the structure as a whole, and the wave is reflected and largely excluded rather than admitted through it. At 900 MHz the wavelength is a third of a meter, and windows, doorways, and interior corridors become apertures and waveguides, so the signal enters, scatters through the interior, and exploits multipath that VHF cannot use. The same logic is why the cellular industry built its networks at 800 MHz and upward, and why public safety agencies migrated from VHF to the 700 and 800 MHz bands when in-building coverage became the priority, and neither of those decisions says anything whatever about what happens beneath a canopy of wet fir. Cities, unlike forests, are full of doors.

That trade — terrain and canopy courtesy of the long wavelength, versus a small, efficient, battery-friendly antenna courtesy of the short one — runs through every band decision below.

The other side of that trade deserves arithmetic as well, since it is the side that actually decides what can be carried. The radiation resistance of an electrically short dipole falls as the square of its length measured in wavelengths, R_rad ≈ 197(l/λ)² ohms, while the loss resistance of the loading coil and the surrounding structure does not fall at all. A six-inch whip at 900 MHz is therefore close to a half-wave dipole with a radiation resistance in the vicinity of seventy ohms, radiating efficiently into whatever it is connected to, whereas the same six-inch whip at 145 MHz is about a thirteenth of a wavelength with a radiation resistance on the order of one ohm — the same order as the resistance of the matching coil that feeds it, which means the antenna is guaranteed to spend a substantial fraction of its energy as heat in its own matching network. A short whip does not become efficient through cleverness; it becomes efficient by being attached to a short wavelength.

Behind that remark sits a bound that no design can escape. The Chu–Harrington limit states that the radiation quality factor of any electrically small antenna satisfies Q ≳ 1/(ka)³ + 1/(ka), where k = 2π/λ and a is the radius of the smallest sphere that encloses the antenna. Since bandwidth falls as 1/Q, an antenna small relative to its wavelength is necessarily narrowband and narrowband antennas are necessarily lossy in any real implementation; the loaded, narrow, inefficient VHF whip on a handheld is not a design failure but a direct consequence of ka being tiny. Almost everything in this subsection follows from that single inequality, together with the absorption of water.

So here is the honest ledger between the two sides of the spectrum, in numbers rather than in allegiances. In terrain, the longer wavelength is worth roughly seven decibels per obstructing edge, and that figure compounds across the successive edges that actual mountains provide, so that a three-ridge path is worth some twenty decibels of advantage to VHF. In canopy, the frequency effect is a few decibels while the depth effect, which applies equally to both bands, is nearer ten. In the antenna, the short wavelength is worth tens of decibels if the comparison is between two pocket-sized whips, and nothing at all if the party is willing to carry the nineteen inches that two meters actually requires, which a handheld carries without complaint. Put those numbers side by side and the recommendation writes itself: where the goal is a signal that survives the shape of the land, take the lower band and the longer antenna, because the antenna is the one part of the system that a person on foot can repair with a length of wire and a tree branch, and the land is the part that cannot be moved at all.

The Ionosphere: HF’s Superpower and Its Mood Swings

#

Nothing else in this essay behaves like HF. Between roughly 3 and 30 MHz, signals launched at a shallow angle toward the sky can be refracted back to earth by the ionized layers of the upper atmosphere, skip off the ground, and be refracted again — a pinball sequence that can carry a modest transmitter thousands of miles with no infrastructure whatsoever. Terrain is irrelevant to a skywave signal, because the signal does not travel along the ground. A deep canyon that defeats every VHF radio in the party is, to a 40-meter signal, simply not in the conversation.

Take the mechanism apart and it turns out to be a story about electrons. The ionosphere is a weakly ionized plasma — a rarefied gas in which solar ultraviolet and X-ray photons have knocked electrons free of their atoms — and a plasma has a natural resonance frequency set by how many free electrons it contains and by almost nothing else. That plasma frequency is f_p ≈ 9√N_e hertz when the electron density N_e is expressed in electrons per cubic meter, and the refractive index of the medium, neglecting the earth’s magnetic field and electron collisions, is n = √(1 − (f_p/f)²). Two consequences fall straight out of that expression. First, whenever the wave frequency exceeds the local plasma frequency the index is real but less than one, which means the wave travels faster than it would in vacuum and, in a medium whose density varies with height, bends — this is refraction rather than reflection, and it is why a skywave ray curves smoothly back toward the ground instead of rebounding off a surface. Second, at the level where the plasma frequency rises to meet the wave frequency the index falls to zero and the wave turns around; anything above that altitude is never reached. HF propagation is therefore a contest between the frequency of the wave and the electron density profile of the atmosphere, and everything an operator knows about the behavior of the bands is a consequence of it.

The density profile is neither uniform nor stable, and the layers are named for their shape. The D region, from roughly 60 to 90 kilometers, is weakly ionized but still dense enough in gas that electron collisions dominate; it cannot refract HF usefully and it absorbs it instead. The E region, from about 90 to 120 kilometers, can refract the upper HF bands by day and hosts the patchy, intensely ionized sporadic-E clouds that open six and ten meters at distances ordinary propagation cannot explain. The F region, spanning roughly 150 to 500 kilometers and splitting into F1 and F2 in daylight, carries the electron densities that actually matter, and the F2 peak is where the maximum plasma frequency of the entire column occurs. That peak value defines the critical frequency, f_oF2 = 9√N_max, which is the highest frequency a vertically launched wave will return to the ground, and it is the single most useful number in HF propagation. Typical values run from 2 to 4 MHz in the small hours to 8 to 12 MHz on a summer afternoon in a moderately active solar year.

For oblique paths the usable frequency is higher, and in the flat-earth approximation the relation is pleasingly simple: the maximum usable frequency is the critical frequency divided by the sine of the take-off elevation angle, MUF = f_oF2/sin Δ^[23], which is equivalently f_oF2 sec φ with φ the angle of incidence at the layer measured from the vertical. At vertical incidence the two expressions coincide, which is the definition of the critical frequency; at the shallow angles used for long-distance work the magnification factor is typically three or four — the secant law alone would predict more, but the curvature of the earth limits how oblique a returning ray can be. The practical corollary is the entire reason the bands behave as they do. When the critical frequency stands at 10 MHz on an afternoon, a shallow-angle path can support 28 MHz and ten meters opens to the world; when it falls to 3 MHz after dark, the same geometry supports only about 9 MHz, twenty meters goes silent, and only forty and eighty meters still do any work. The bands do not open and close by magic, and they do not open and close on a schedule anyone chose; they open and close because the electron density of the upper atmosphere, and with it the critical frequency, follows the sun.

Geometry then sets how far a single reflection reaches. One hop to the F2 layer tops out somewhere in the neighborhood of 4,000 kilometers for a ray that grazes the layer and returns, because the curvature of the earth limits how far the chord can stretch before the ray escapes the atmosphere altogether; a longer path requires a second hop, using the earth or the sea as a reflecting surface sandwiched between two ionospheric reflections, which is why the pinball analogy is exact and why every additional hop adds loss and irregularity. Anything shorter than the minimum distance one hop can cover falls into the skip zone, and the width of that dead ring is not a constant. It depends on where the operating frequency sits relative to the critical frequency, because the steepest ray the layer can turn around is the vertically incident one. As the operating frequency drops toward f_oF2 the returning rays grow steeper, the hole contracts, and at the critical frequency itself the signal comes back down essentially on top of the transmitter. NVIS is not a different phenomenon from skywave; it is skywave walked to its limit.

This is HF’s superpower and its curse: the ionosphere is a living medium. Its behavior depends on the sun. The dense D layer absorbs lower frequencies during daylight, which is why 80 and 160 meters are night bands; the F layer’s maximum usable frequency rises and falls with solar activity, opening the upper bands at solar maximum and closing them as the cycle declines — Cycle 25 peaked around 2024–2025 and is now heading toward its minimum. Long-range skywave also creates a skip zone: from a given location, a high band may reach a station 1,500 miles away while the town 60 miles distant hears nothing at all.

Absorption is the other half of the diurnal story, and it too has an equation. The energy a wave loses in the lower ionosphere is governed by how often its electrons collide with neutral molecules and how many of them there are, and the non-deviative absorption accumulated along a ray takes the form A ∝ ∫ (N_eν/f²) dl, where ν is the electron collision frequency and the integral runs along the path. Two features of that expression deserve attention. The 1/f² dependence means absorption falls with the square of frequency, so the lowest bands pay a penalty that grows steeply as the wavelength lengthens; and the dependence on electron density, which in the D region is manufactured by solar radiation and disappears overnight as the electrons recombine, means the penalty is essentially a tax levied only in daylight. This is why 160 meters is not merely better after dark but effectively absent during the day, and why the same is true, less catastrophically, of 80 meters at midday.

Because the D region is manufactured by sunlight, absorption also scales with the solar zenith angle, roughly as the three-quarters power of the cosine of that angle in the standard empirical treatments, and that single dependence explains most of the seasonal behavior of the low bands. Summer brings a high sun, long days, and a persistent absorbing layer, so the bottom of the spectrum is noisy and the top is comparatively better; winter brings a low sun, long nights, and little absorption, so 160 meters comes alive and low-band long-distance work becomes the standard cold-season pursuit. There is even a documented winter anomaly in which mid-latitude absorption on some winter days exceeds what the zenith angle alone would predict, traced to changes in the neutral composition and electron-loss chemistry of the mesosphere — a useful reminder that the arithmetic in this subsection, unlike the arithmetic of free-space loss, is empirical and approximate rather than exact.

Solar activity enters through the same door, and it enters as a number. The standard indices are the sunspot number and the ten-point-seven-centimeter solar flux, the latter expressed in solar flux units of 10⁻²² W m⁻² Hz⁻¹ and running from about 70 at solar minimum to values above 200 at maximum. Both track the ionizing radiation that builds the F region, so both drive the critical frequency and therefore the maximum usable frequency, which is why the eleven-year cycle is not a curiosity to the operator but an operating parameter. Cycle 25 reached its smoothed peak in 2024 and 2025 with monthly sunspot numbers approaching 200^[18], and the decline now underway will take ten meters from worldwide openings to long silences, fifteen meters from dependable to intermittent, and twenty meters from the workhorse it is at the peak to a night band at the trough. Prediction is nevertheless a solved engineering problem rather than a black art, precisely because the physics described above is quantifiable: programs such as VOACAP^[25] compute the median maximum usable frequency, the expected signal-to-noise ratio, and the reliability of a specific path for a specific month, hour, and solar index, so the operator’s alternative to guessing is arithmetic.

There is one further frequency-dependent term, and for a party in trouble it is often the one that decides the matter. Unlike VHF, where the noise floor is essentially the thermal noise of the receiver itself, HF at low frequencies sits inside an enormous natural noise field generated by lightning worldwide and carried to the receiver by the same ionosphere that carries the signal, with man-made noise added on top. Median atmospheric noise at 3 MHz in a quiet rural location is typically 30 to 40 dB above the thermal limit^[21], falling to roughly 10 dB by 30 MHz and effectively vanishing at VHF. The consequence is not that HF is worse but that it is more expensive in a particular currency: the signal must be stronger, the antenna must be larger, and the station is built with a hundred watts and a wire rather than five watts and a rubber duck. What purchases that expense is the absence of the terrain losses that the previous subsection put at seven decibels per obstructing edge, so the comparison only becomes meaningful over the distances where VHF has already lost. A three-hundred-kilometer NVIS path at 5 MHz pays no diffraction loss at all, whereas the same three hundred kilometers on two meters crosses the horizon several times over and, with no repeater and no summit to relay from, does not exist. HF buys its reach with power, antenna, and the willingness to work a schedule — and it is the only band whose reach can be bought that way at all.

The mood swings themselves are a taxonomy of named phenomena, and for an emergency planner the taxonomy matters more than the monthly averages. A solar flare of sufficient class ionizes the D region within minutes and produces a shortwave fadeout, or sudden ionospheric disturbance, that blacks out HF across the sunlit face of the planet for anything from a few minutes to an hour. An energetic proton event produces polar cap absorption, which can hold high-latitude paths down for days and is the reason transpolar and auroral-latitude circuits are fragile. A geomagnetic storm, driven by a coronal mass ejection and measured by the planetary K index, depresses the critical frequency, raises absorption at middle and high latitudes, and can degrade or destroy HF circuits for days at a stretch; the same storms occasionally open 160 meters by way of compensation, because the absorption that harms the higher bands arrives from altitudes and geometries that do not trouble the lowest one. And above all of this, sporadic-E clouds open bands that theory says should be closed, frequently enough to be enjoyed by everyone and predictably enough to be relied upon by nobody. The practical consequence for someone whose plan depends on reaching a county sheriff three hundred miles away is uncomfortable but clear: HF can die at precisely the moment when the electromagnetic weather is worst, and the competent answer is a schedule of bands and hours together with a fallback that does not depend on the ionosphere in any form.

For backcountry use, the skip zone is the problem worth solving, and there is an elegant solution. NVIS — near-vertical incidence skywave — deliberately fires the signal nearly straight up, so it falls back to earth within a few hundred miles, filling in the region that ordinary skywave skips. On 40 meters during the day and 80 meters at night, a modest station can hold regional contact out to 300 or 400 miles with no repeaters, no infrastructure, and no line of sight. For someone deep in a mountain range who needs to reach a county sheriff’s office or an amateur emergency net on the far side of the divide, this is the capability that no other band offers.

NVIS is best understood as the limiting case of the geometry already described, and it happens to be unusually courteous to improvised stations. The requirement is a take-off angle at or above roughly sixty degrees from the horizon, corresponding to the near-vertical rays that a layer can only turn around when the operating frequency sits at or below the critical frequency; the resulting footprint is typically a disc three hundred to four hundred kilometers across, with the caveat that the coverage can become a ring rather than a filled circle if the antenna’s high-angle pattern has a null in the middle. The antenna then follows directly from the requirement. A horizontal dipole’s pattern depends on its height above ground because the direct ray and the ground-reflected ray interfere, and at heights below a quarter wavelength that interference cancels the low-angle lobes and reinforces the high-angle ones — the exact opposite of what a designer wants for long-distance work, and the exact property that makes such an antenna good here. A forty-meter dipole mounted three to five meters up is therefore close to ideal, while a vertical, which radiates most of its energy at low angles and along the ground, is close to useless for the purpose. This is why the canonical emergency HF station in the field is a length of wire strung between two trees at head height, which is also, conveniently, the cheapest and lightest antenna in the spectrum.

The rule that forty meters serves by day and eighty meters by night is a computation rather than a convention, and it follows from the two boundaries that bracket the usable frequency. The ceiling is the critical frequency, because a near-vertical ray above it penetrates the layer instead of returning; the floor is set by D-layer absorption, because the 1/f² term in that expression makes low frequencies unaffordable in daylight. On a moderately active afternoon the critical frequency lies somewhere between 7 and 10 MHz while the absorption floor sits above 3.5 MHz, so forty meters is comfortably beneath the ceiling and above the floor, while eighty meters lies beneath that floor or uncomfortably close to it and is reduced to ground wave and a few noisy short hops — a judgment that is firmer in summer than in winter, which is a distinction the next paragraph explains. After dark the critical frequency falls to 2 or 4 MHz and the D region has gone, so eighty meters is now below the ceiling and free of the daylight tax, while forty meters has become transparent to vertical rays and is traveling long instead of staying regional. The two bands exchange roles because the two boundaries move in opposite directions with the sun, and an operator who has internalized that sentence knows which band to try at any hour without consulting a table.

One further window is worth knowing about, because it is the kind of detail that converts a marginal plan into a working one. Along the dawn and dusk terminator the D region has already recombined on the dark side of the line while the F region remains substantially ionized, producing a moving strip of low absorption and good reflection; signals can be launched along that strip and travel with the twilight, so low-band contacts that are impossible in full daylight and noisy in full dark become easy for half an hour on either side of sunrise and sunset. The terminator sweeps westward at about 1,600 kilometers per hour at the equator and about 1,200 at the middle latitudes, so the window is narrow and it moves; military networks, maritime operators, and contesters have all built schedules around it, and a backcountry station that knows its own sunrise can use it to buy real capability for nothing.

HF is the only band where the terrain between you and help is irrelevant. It is also the only band whose performance changes with the hour, the season, and the eleven-year solar cycle — which is why HF operators live by schedules, and why an HF radio in a pack is worth less than an HF radio attached to a wire antenna and a habit.

Ground Wave and the Earth–Ionosphere Waveguide

#

Below HF, two other propagation mechanisms take over, and both are remarkable — and both are effectively unavailable to a backpacker.

At MF and LF, ground wave hugs the earth’s surface, following its curvature with relatively low loss, which is how AM broadcast stations cover a region on a single transmitter and why longwave time signals such as WWVB at 60 kHz^[15] reach across a continent.

Ground wave is a genuinely different animal from everything discussed above, and the difference is visible in the geometry. A signal traveling along the ground is not a ray that merely happens to be low; it is a wave that has become attached to the interface between air and earth, supported by currents induced in the ground beneath it and continuously replenished by those currents. Because it is attached it follows the curvature of the planet instead of leaving it, and because it is attached it is absorbed. The bookkeeping is done with the surface impedance of the ground, Z_s = √(jωμ/(σ + jωε₀ε_r)), which for any ground of practical interest reduces to Z_s ≈ (1 + j)√(πfμ/σ) with μ = 4π×10⁻⁷ H/m. Two consequences follow at once. The electric field is tilted forward from the vertical by an angle on the order of Z_s/η₀, where η₀ = 377 Ω is the impedance of free space, and that tilt is simultaneously the reason for the wave’s attachment and the reason for its loss, because the same forward-leaning field drives the ground currents that dissipate energy as heat. The depth to which those currents reach is the skin depth δ = √(1/(πfμσ)), which at 1 MHz is about 25 centimeters in seawater but a full seven meters in soil of average conductivity — two grounds differing in conductivity by three orders of magnitude, which the ground wave notices immediately.

This also produces a rule that operators of these bands treat as axiomatic: ground-wave service requires a vertically polarized antenna. A horizontal antenna near the ground is short-circuited by its own image in the surface, which is why every AM broadcast mast on earth is a vertical radiator and why the MF and LF bands pair naturally with tall monopoles and inverted-L structures that look nothing like anything a backpacker will ever carry. When the ground underneath is favorable, the mechanism is remarkably generous. At 60 kHz, where the wavelength is five kilometers, ground wave carries the WWVB time code across a continent; at AM broadcast frequencies the same mechanism delivers a dependable daytime service over something like a hundred kilometers of average soil^[22], and the range grows several-fold over sea water, which is why the maritime distress frequencies of the two-megahertz region — 2182 kHz most famously, with the Morse distress channel at 500 kHz before it — could be heard by ships hundreds of miles from a coast station. At HF the mechanism survives only for a few tens of kilometers, and at VHF it has vanished altogether.

The distance law of the ground wave is where the frequency dependence really lives. Over a perfectly conducting earth the field falls as 1/d, which in ordinary decibel language is a mere six decibels per doubling of distance. Over real earth the attenuation function departs from that ideal beyond a threshold set by Sommerfeld’s numerical distance, p = (πd/λ)(1/|N|)cos b, in which N² = ε_r − j60σλ and b is a phase angle determined by the complex permittivity of the ground. The essential feature is that p grows in proportion to d/λ, so the threshold is met at a distance that scales with wavelength and the same physical path is a difficult one at HF and an easy one at LF. Beyond that threshold the field decays as 1/d² — twelve decibels per doubling of distance — and at extreme range the decay becomes exponential, because the tilt of the field grows with distance and the induced currents rise with it. Read in frequency terms, the attenuation coefficient scales roughly as 1/|N| while |N| contains the term 60σλ, so a longer wavelength and a more conductive ground both reduce the loss, and the ground-wave range of any station is largely a property of everything underneath its path. This is also why the ground-wave region of the MF band is stable and largely fadeless in daylight, and why the AM band degenerates into a jumble of distant stations after dark: the D-layer absorption that attenuates eighty meters by day equally suppresses the MF skywave, so daytime reception is ground wave alone and clean, and when the sun sets and the ionosphere returns, skywave arriving from hundreds of kilometers away overlays the stable ground-wave service — which is exactly why American AM stations must reduce power or switch to directional arrays at nightfall.

Below about 30 kHz, in VLF, the planet becomes a waveguide: the conducting earth and the ionosphere form a spherical duct that carries signals around the globe with very little attenuation. This is why navies use VLF to reach submerged submarines, and why some of the largest radio transmitters ever built are arrays of masts spread over square miles.

That duct is the strangest of the three mechanisms in this essay, and the description of it as a waveguide is exact rather than metaphorical. The conducting earth and the ionized D region together constitute the two walls of a spherical waveguide, with the planet as the lower plate and the lower ionosphere — at a daytime altitude near 70 kilometers and a nighttime altitude of 85 to 95 — as the upper one. The duct behaves like any waveguide, and its cutoff frequencies are fixed by its height: in a parallel-plate approximation the higher-order modes cut off near f_c = nc/2h, which for a 70-kilometer height places the first of them in the neighborhood of two kilohertz, while the lowest mode, a quasi-transverse-electromagnetic mode consisting of a vertical electric field standing between two conducting walls, has no cutoff at all and propagates in principle down to direct current. Everything above a few kilohertz is therefore multimode propagation in which the modes travel at different phase velocities and interfere with one another along the way, which is the origin of the beating and the phase instability that make VLF a difficult medium for any purpose other than the ones it was designed for.

What makes the duct valuable is its attenuation. For the lowest mode at VLF frequencies the loss is on the order of two to three decibels per thousand kilometers in daylight and appreciably less at night, when the D region lifts, the walls of the guide separate, and the upper wall becomes less lossy. Those figures turn the entire planet into a single coverage area for a sufficiently large transmitter: a station of some hundreds of kilowatts at 20 kHz is audible everywhere, with a phase stable enough to serve as a worldwide frequency reference and to discipline clocks on the far side of an ocean, which is how the Omega navigation system worked and how VLF standard-frequency services continue to operate. The price of the modal structure is paid in bandwidth and in predictability, because the field at great distance is dominated by the first mode alone, the phase of that mode depends on the effective height of the ionospheric wall, and a transmitter wanting both range and reliability must accept whatever modulation constraints the medium imposes on it.

The most famous application of all follows from the conductivity of the sea. Seawater runs near four siemens per meter, so the skin depth δ = √(1/(πfμσ)) is about 1.8 meters at 20 kHz and roughly thirty meters at 76 hertz. That single comparison explains an entire era of engineering. VLF reaches a submarine only if the boat exposes an antenna or rises close to the surface, which is why the operational submarines of the VLF era trailed long buoyant wires and why their message traffic had to be terse; the navy therefore invested in an extremely-low-frequency system at 76 hertz, where the wavelength is four thousand kilometers and the only workable antennas were ground dipoles with arms tens of miles long laid across two states, radiating megawatts at a data rate of a few characters per minute. The system worked, reached submerged boats at useful depth for decades, and was retired in 2004 — which is the whole problem stated in a paragraph, that the price of penetrating deeper into the medium is paid in wavelength, and wavelength is paid in structures that no individual and very few nations can afford.

The catch is always the same. Antenna efficiency is a function of size relative to wavelength. At 20 kHz the wavelength is 15 kilometers; a quarter-wavelength antenna is roughly two and a half miles of wire. In practice, LF and VLF stations load enormous structures, accept a few percent efficiency, run in the hundreds of kilowatts to megawatts, and still transmit data at speeds measured in tens of bits per second.

The numbers behind that efficiency claim are worth setting out, because they show the limit to be arithmetic rather than conventional. The radiation resistance of an electrically short antenna falls as the square of its height measured in wavelengths, R_rad ≈ 197(l/λ)² ohms for a short dipole and roughly twice that for a monopole over a ground plane, which is the configuration of a tower, and it takes no interest in how large the transmitter is. At 60 kHz, where the wavelength is five kilometers, a two-hundred-meter tower — a structure of serious expense, taller than most buildings — is only four hundredths of a wavelength, and its radiation resistance works out in the neighborhood of six tenths of an ohm. Against that figure the resistance of the loading coil, the ground system, and the structure itself is measured in whole ohms. Efficiency is the fraction R_rad/(R_rad + R_loss), so the arithmetic lands these stations in the single digits to low tens of percent at best, and the megawatt installations that beat that figure do so by covering square kilometers with copper radials and by erecting the largest self-supporting structures ever built. A vertical radiator does not become efficient by being tuned well; it becomes efficient by being a substantial fraction of a wavelength, and at 20 kHz a quarter wavelength is two and a half miles of height.

Two further penalties follow from the same smallness, and both are equally unavoidable. The first is bandwidth. A small antenna behaves as a small capacitance with a large reactance to be tuned away, and the ratio of stored energy to radiated energy sets a quality factor whose approximate form for an electrically short radiator is Q ≈ 1/(2πfCR_rad); since fractional bandwidth runs roughly as f/Q, the bandwidth of a physically small antenna collapses as its size falls. That is why these transmitters possess bandwidths of hundreds of hertz at most, why they cannot be modulated quickly, and why the information rate is so miserable. The second penalty is stated by the Shannon capacity of the resulting channel, C = B log₂(1 + S/N): with a hundred hertz of bandwidth and a signal-to-noise ratio of twenty decibels the ceiling is a few hundred bits per second, and with the poorer ratios actually experienced at VLF the real rate is the tens of bits per second quoted above. The medium is not withholding anything that better engineering will eventually supply; it is delivering a verdict that a person can calculate in advance, in the same way that the sky delivers a verdict against a five-hundred-meter tower.

One caveat belongs at the end of this subsection, because it is the reason a backpacker is not entirely outside it. Everything above concerns transmitting, and transmitting is the difficult half. Reception is a different problem, since a receiver need not radiate and therefore does not care about radiation resistance at all; it needs only to place enough signal above the noise, and a ferrite rod of a few cubic centimeters manages that perfectly well at 60 kHz, which is why WWVB timekeeping fits inside a wristwatch and why the lightning sferics and whistlers that saturate these bands are audible on a handheld receiver with a coil of wire for an antenna. That asymmetry has a concrete consequence for this essay’s own accounting. The MF broadcast band delivers weather warnings, road closures, and emergency alerts out to a hundred kilometers or more on ground wave from stations running tens of kilowatts, and any pocket receiver with a loopstick can hear them, which makes it the longest-reaching receive-only information channel available to someone on foot without a satellite and without a license. It is not a way to call for help, and it cannot be, because the antenna needed to answer is the same antenna this subsection has spent two thousand words building and dismissing. The honest place to file LF and MF is therefore as the one region of the spectrum where a person on foot can listen but not speak — the same verdict the frequency has been delivering all along, heard this time from the quiet side of the wall.

That is not a limitation of engineering that better equipment will fix. It is the frequency telling you what it is for.

Band by Band in the Backcountry

#

VHF (30–300 MHz) — The Working Band

#

VHF is where most backcountry communication actually happens, and the reasons are practical rather than theoretical. The antenna is the whole argument: a quarter-wave whip for two meters is about nineteen inches of thin, flexible metal, mounted on a radio the size of a chocolate bar that will run for days on a small battery. Add the fact that VHF propagates a little better than UHF over rolling terrain and through canopy, and you have the best portability-to-performance ratio in the spectrum.

Look at the arithmetic behind the antenna argument and it becomes concrete. At 146 MHz the free-space wavelength is c/f = 3×10⁸/1.46×10⁸ ≈ 2.05 meters, so a quarter wave is fifty-one centimeters, or about twenty inches, and the nineteen-inch whip on a typical handheld is within a few percent of that figure — trimmed manually and matched by the network in its base. This is the entire miracle of the band: a resonant, efficient radiator that is also a thin flexible rod the length of a forearm, which is to say an antenna that survives being shoved into a pack and sat upon. The same handheld does even better for size at 900 MHz, where its helical stub is a good fraction of a wavelength; at 7 MHz it would have no antenna at all, which is the trade established two sections ago. Gain figures on this band are modest but honest. A quarter-wave whip over a proper counterpoise is worth about 5 dBi in theory against a perfect ground plane, while a real handheld with a real hand wrapped around it delivers something between three decibels below and two decibels above unity — which means the difference between a good antenna and a poor one on the same radio can exceed the difference between five watts and ten. That is why every experienced operator’s first recommendation is the antenna and the second is the battery, and why a handheld with a proper quarter wave will hold contacts that an identical handheld with a rubber duck cannot hear at all.

The battery arithmetic is equally unromantic and equally decisive. A two-meter handheld draws something on the order of twenty to sixty milliamperes while squelched and receiving, and between one and a half and two and a half amperes while transmitting at full power. Against a two-thousand-milliamp-hour pack that works out to roughly a hundred hours of listening or a single hour of continuous transmission — but the realistic number, at a one-percent transmit duty cycle, is a little over two days of genuine operation, which is longer than most backcountry trips last. Anyone who has watched a handheld die on the second morning has been operating at a higher duty cycle than they believed, in conversation rather than deliberately, and the cure is discipline rather than a bigger battery: listen more, transmit less, keep the radio warm, and treat the transmit button as a resource with a price on it. The power ceiling matters less than it appears, for a related reason. Five watts is the standard handheld maximum, while a vehicle-mounted rig runs twenty-five to seventy-five watts, and the vehicle’s advantage once its higher antenna and higher gain are included comes to roughly thirteen decibels. That figure is real and worth having, but on a path where received power falls as the fourth power of distance it buys about a factor of two in range, and a factor of two in range does not cross a ridge. It is the same lesson the physics section delivered, restated in the units of this band: watts buy distance along a path that already exists, and elevation buys a path.

VHF also owns the ecosystem. The amateur 2-meter band (144–148 MHz) is the most popular band in the world, saturated with mountaintop repeaters that turn a 5-watt handheld into a regional system; 146.52 MHz is the long-standing national simplex calling frequency^[13], and the informal Wilderness Protocol — monitoring that frequency at the top of each hour, and more formally every three hours from 07:00 to 22:00 — exists precisely so that a hiker with a handheld has somebody listening. The Automatic Packet Reporting System at 144.390 MHz periodically beacons a position and can be digipeated by satellites; the International Space Station carries FM voice and packet capability on 2 meters, meaning a handheld and a small directional antenna can reach an orbiting asset. Add 6 meters at 50 MHz, which in summer occasionally surprises everyone with thousand-mile sporadic-E openings on a VHF radio, and the automotive and marine worlds — marine VHF Channel 16 at 156.8 MHz, aviation’s 121.5 MHz emergency frequency — and the pattern is unmistakable: nearly every organization that rescues people listens on VHF.

The repeater ecosystem deserves more than a mention, because it is the single largest reason the band behaves the way it does. A mountaintop machine is not merely a louder station; it is a station with an antenna at two or three thousand feet and a clear view of a hundred miles of country, running fifty watts into a gain antenna so that its effective radiated power is several times that, listening with a receiver that will hear a whisper. In the United States the standard arrangement is a six-hundred-kilohertz offset, the repeater transmitting on the higher frequency and listening on the lower, with access controlled by a continuous subaudible tone — 88.5, 100.0, and 103.5 hertz most commonly — which the transmitting station superimposes on its audio and the repeater insists upon before it will open. That tone is not a courtesy; it is the key. A hiker with the correct frequency programmed and the wrong tone, or no tone at all, will hear the machine perfectly and be heard by nobody, which is the most common silent failure in amateur field work and the reason that pre-programming and testing a radio before a trip is worth more than any accessory. The arrangement exists because a repeater breaks the reciprocity of the path in terms of system margin rather than in terms of geometry. Radio paths are reciprocal, but radio systems are not: the mountainside station has roughly sixteen to twenty decibels more radiated power, a far better antenna, and a clear horizon, so the downlink — the repeater speaking to the hiker — closes at distances from which the uplink cannot be heard at all. The symptom is familiar and maddening, complete quieting from the machine and nothing but silence in return, and terrain deepens the asymmetry, because a signal descending from a summit bends over the ridgelines into the valley in a way that a handheld’s signal cannot manage going up.

What the repeater cannot do, simplex procedure attempts to do by agreement, and the Wilderness Protocol is the most successful piece of agreed procedure in the hobby. Its logic is worth unpacking because it solves a scheduling problem rather than a radio problem. Two parties with handhelds and no repeater cannot know when they will be within range of one another, and neither can afford to leave a radio monitoring all day, since a receiver that is squelched all day is a receiver and a set of ears spent for nothing. The protocol therefore manufactures a rendezvous in time: monitor the national simplex calling frequency, 146.52 MHz in North America, at the top of each hour and, where that is impractical, at least on the hours of 07:00, 10:00, 13:00, 16:00, 19:00, and 22:00. A caller following the same schedule therefore has a guaranteed listener, twice an hour or six times a day, for the cost of a few minutes of battery. The scheme’s weaknesses are structural rather than accidental: it assumes the other party knows the protocol, practices it, and can actually reach the person listening, and near cities 146.52 is congested enough that the calling channel is often occupied by stations following no protocol at all. It remains the best general answer available to the party that is unlicensed, unrepeatered, and equipped with nothing but handhelds, which describes most hiking parties who carry radios at all.

Two further VHF services deserve understanding precisely because they are so often oversold. The Automatic Packet Reporting System occupies 144.390 MHz in North America and transmits short bursts of data at 1,200 baud — position, telemetry, brief messages — which are digipeated onward by other stations and, where a gateway is within range, injected into the internet. Its virtue is that it requires no human listener: a party carrying a tracker appears on a map at home whether or not anyone is awake at the far end, and the infrastructure is volunteer and remarkably durable. Its limitations are equally clear. It broadcasts information one way rather than calling for help, 1,200 baud is slow enough that congestion is real on a busy weekend, and the data must reach a gateway or digipeater before it reaches anyone at all, which is the same terrain problem as voice with less margin to spare. The orbiting asset the band offers is more impressive than it sounds. The International Space Station carries FM voice and packet facilities in the two-meter band, with voice downlinks near 145.800 MHz and packet digipeating near 145.825^[14], and a five-watt handheld with a small directional antenna can work it during a pass. The engineering detail that ambushes newcomers is Doppler: the station moves at 7.66 kilometers per second, so the received frequency shifts by up to f v/c = 145.8×10⁶ × 7.4×10³/3×10⁸, on the order of plus or minus 3.6 kilohertz across the ten minutes of a pass. That is comparable to the entire width of a narrowband FM signal, so the operator must retune continuously or lose the contact, and a party expecting a satellite contact to be a rescue technique is expecting the wrong thing from the wrong duration.

The rest of the VHF neighborhood matters for the same reason the two-meter band does, which is listeners. Six meters at 50 MHz is the band that occasionally forgets it is VHF, since summer sporadic-E clouds and, at the peak of the solar cycle, even the F layer will support contacts over a thousand kilometers using the modest power and small antennas of the VHF world; the openings are unpredictable, brief, and capable of carrying a stranded party’s signal far beyond any repeater’s reach. The one-and-a-quarter-meter band at 222 to 225 megahertz in the United States is the overlooked middle child, with propagation nearly as good as two meters, antennas a third shorter, and far fewer users and radios on either end. Marine VHF channel 16 at 156.8 MHz is the maritime distress and calling frequency, monitored around the clock by coast stations nearly everywhere and, in American waters, by a Coast Guard shore network fitted with direction-finding and reachable twenty miles or more offshore; its modern companion, channel 70, carries digital selective calling, which automates the alerting that voice calling once did by ear. Aviation’s 121.5 MHz is the civil emergency frequency, monitored by air traffic facilities and by many aircraft in flight, and still the frequency to which the homing transmitter of a 406 MHz beacon leads searchers. Each of those frequencies has a population of professionals whose actual job is to listen on it, which is precisely what the two-meter band lacks in deep backcountry: the amateurs are there, but they are volunteers with day jobs, limited coverage, and no obligation to be awake.

The legal texture is worth stating precisely, because it is where good intentions meet the regulations. A two-meter handheld may be used in the amateur service only by a licensed operator, on amateur frequencies, under amateur rules. The marine band may be used from shore only in connection with a vessel, or to reach a vessel or coast station. The aviation emergency frequency is for aircraft and for ground stations in emergencies. The universal escape clause, recognized in the amateur service’s own rules and in international treaty, is that a station in distress may use any means at its disposal to attract attention and obtain assistance^[11] — and exercising that right is neither a crime nor an operating plan. It is a statement about the legal consequences of an act, not a claim that the act will work. A two-meter handheld cannot transmit on 121.5 MHz at all, and if it could it would be radiating frequency modulation into an amplitude-modulated band, which is a hardware mismatch before it is a legal question, so the airliner it hoped to reach would in most cases hear nothing at all. Willingness and equipment are different problems, and only one of them is a virtue.

MURS, the license-free service at 151–154 MHz, is the civilian entry point to the same physics: five channels, 2 watts, handhelds only, no repeaters, and — because it is underused — comparatively quiet channels. A MURS radio is not a rescue tool, but it is an honest one, and its band choice is the correct one for hiking parties.

MURS deserves more than its single paragraph, because it is the band choice most unlicensed hiking parties should be making and the reasons are structural rather than technical. The service provides five channels at 151.820, 151.880, 151.940, 154.570, and 154.600 MHz — VHF, and therefore the correct side of the spectrum for terrain — at a legal maximum of two watts, for handhelds, with repeaters and store-and-forward expressly forbidden^[9]. Two watts against the five watts of an amateur handheld is four decibels, or two thirds of a single S-unit, and in the fourth-power regime of a two-ray path it costs somewhere between a fifth and a third of the range depending on which distance law is doing the work: measurable, real, and not remotely decisive. What is decisive runs the other way, in that MURS is the only service in this discussion on which a party can legally hand a radio to a friend who holds no license at all. That single fact converts it from a personal tool into a group tool, and it is the reason MURS belongs in a discussion about emergency communications at all. The corollary is uncomfortable and should be stated plainly rather than softened: MURS is not a rescue band, because there is no repeater, no organized net, no professional listener, and no path to an aircraft or a coast station. It performs the coordination job rather than the summoning job, which makes it insurance against getting separated and no insurance whatever against getting hurt. The comparison with the UHF license-free services sharpens the point. FRS and GMRS offer more channels, more power, and in the GMRS case repeater access, a licensed population, and handhelds that will reach twenty miles from a good elevation — all at the cost of the shorter wavelength examined earlier, and at the cost of a band genuinely crowded at every trailhead and campsite in the country. MURS’s quiet channels are not a design feature but a consequence of low adoption, and quiet is exactly what a small party wants. One practical caveat belongs here: the two 154-megahertz channels were formerly itinerant business frequencies, the notorious dot channels of drive-through windows and construction crews, and they still carry that traffic wherever a trailhead abuts a highway. An honest MURS radio is honest in the specific sense that it does everything the physics says it can do and makes no claim to be a call for rescue.

The catches are real. Line of sight dominates, so valleys and ridgelines can reduce a nominal 20-mile path to a fraction of a mile. Repeater coverage is not uniform, and in genuine wilderness there may be no repeater at all. And legal use of the marine and aviation bands from the ground, without a vessel or aircraft, is restricted; in a true life-threatening emergency the general principle that a station in distress may use any means available applies, but that is a fallback, not a plan.

The catches deserve the same precision as the virtues. Line of sight dominates, and the sharpest way to see how completely is to compute the clearance a two-meter path actually requires. The first Fresnel zone radius at a path’s midpoint is r₁ = √(λd₁d₂/(d₁+d₂)), and at 146 MHz on a twenty-mile path — thirty-two kilometers, with the obstruction at the halfway point — that expression yields √(2.05 × 8000) ≈ 128 meters. A path is not considered clear unless sixty percent of that radius is unobstructed, which means the party needs a tube of some eighty meters radius and a hundred and sixty meters across, free of terrain, along the entire line between the antennas. A handheld carried at a meter and a half above ground, in rolling country, has nothing of the kind, and this one number explains why a nominal twenty-mile path becomes a fraction of a mile the moment the operator walks into a fold in the land. It also explains why elevation is worth so much more than power in precisely the situation where a party is in trouble: transmit power cannot clear a Fresnel zone, and there is no quantity of watts at which a ridge ceases to be in the way. Repeater coverage is likewise not the uniform blanket that coverage maps imply. Field strength in mountain terrain varies by thirty decibels or more over a few hundred meters, so a machine forty miles away may be usable from a summit and unreachable a kilometer into the next valley, and a repeater’s published footprint is a statement about probability rather than about certainty. There is also an organizational catch that no amount of physics will fix: many mountaintop systems are effectively private, built and maintained by clubs or emergency organizations that expect the courtesy of membership or support in exchange for access, and some are closed to general traffic altogether.

Which brings the discussion back to the conclusion this section opened with, and clarifies what carrying exactly one voice radio actually means. The reason to carry VHF is not that it wins at every job but that it is the only band where the three jobs coincide: it has an antenna a person will genuinely carry, a propagation advantage over the terrain a person will genuinely encounter, and a population of listeners — amateurs, coast stations, aircraft, and increasingly the search and rescue teams themselves — who are not there by accident. The third of those advantages is the one that summons help, and it is worth being honest that it is also the one bought with preparation rather than with money. The listeners exist on specific frequencies, in specific modes, with specific tone requirements, and a party that carries a two-meter handheld without the right channels programmed and the right tones set has purchased the antenna and the propagation and forfeited the part that matters most. The band is the working band because it is the band where all three advantages are available at once; whether a particular party has all three is a question about the party.

If you carry exactly one voice radio into the backcountry, carry something that transmits on VHF. It has the best antenna, the best terrain behavior, and — decisively — the most listeners.

UHF (300 MHz–3 GHz) — Small Antennas, Small Cells, and the Signal That Actually Saves You

#

UHF is the compromise band, and in the backcountry it is a good one. A quarter-wave antenna for 70 centimeters is about six inches long; a 5-watt handheld with an internal antenna is entirely practical, which is why every consumer service lives here. GMRS and FRS occupy 462 and 467 MHz, sharing channels and differing mainly in licensing and power — GMRS exempt from testing but licensed by application, covering the licensee’s immediate family, allowing 5-watt handhelds and 50-watt mobiles with repeaters^[6]; FRS limited to 2 watts and no repeaters^[7]. Europe’s equivalent, PMR446, sits at 446 MHz. The 70-centimeter amateur band (420–450 MHz) is here as well, with 446.000 MHz as the national simplex calling frequency, and it is a favorite in canyons, dense rock clutter, and urban-edge environments where reflection and scattering help rather than hurt.

The antenna arithmetic explains why every consumer service lives here. At 462 MHz the wavelength is sixty-five centimeters, so a quarter wave is sixteen centimeters — a shade over six inches — while at 446 MHz it is seventeen and at 902 MHz it is only eight. A six-inch whip therefore arrives at the UHF bands as a near-resonant radiator and arrives at two meters as a stub roughly a thirteenth of a wavelength with a radiation resistance near an ohm, which is the comparison drawn earlier in this essay and the reason an antenna that is efficient here is decorative there. The gain and pattern penalties that trouble VHF handhelds shrink as well, because a radiating element that is a large fraction of a wavelength depends less on the size of whatever counterpoise it can find and is less disturbed by the hand wrapped around the radio. The practical consequence is that five watts into a UHF handheld is closer to five watts actually radiated than the same power on two meters would be, and the preference of an entire industry for this band follows from that one geometric fact rather than from any marketing judgment.

There is a second and less often stated advantage, and it is a real one in canyons and rock clutter. The first Fresnel zone radius scales as the square root of wavelength, so the obstacle-free tube a path requires is smaller at higher frequencies: on a ten-kilometer path at seventy centimeters the first-zone radius at the midpoint is about forty meters, against roughly seventy-two meters at two meters. Thirty meters of extra clearance is thirty meters of extra clearance, and in a rock-walled canyon or a street of buildings it can be the difference between a path and no path at all. That geometric advantage is precisely what the claim about reflection and scattering helping rather than hurting is pointing at. When the wavelength is a close match for the size of the objects in the environment, walls and boulders and vehicle bodies stop being obstacles and begin behaving as mirrors and diffusers, and the total energy arriving at the receiver can exceed what a single diffracted ray would deliver. The price is fading. Multipath at UHF in a cluttered environment produces fluctuations of twenty decibels or more over movements of a fraction of a wavelength, which is why UHF voice in a city or a slot canyon alternates between perfectly clear and entirely absent, and why systems that intend to be reliable in such places either move the operator, move the antenna, or spread the signal across a range of frequencies so that not every component fades at once.

The service structure is worth laying out concretely, since it decides who can talk to whom. FRS and GMRS interleave on the same twenty-two channels in the 462 and 467 megahertz neighborhoods, where the 467 channels serve as repeater inputs and low-power interstitials, and the difference between the two services is licensing and power rather than physics. FRS permits up to two watts on most channels and half a watt on the interstitial ones, forbids repeaters, and requires no license. GMRS requires a license involving an application and a fee but no examination, covers the licensee’s immediate family, allows five-watt handhelds and fifty-watt mobiles, and — the decisive difference from a hiking party’s point of view — permits repeaters, which is what converts the band from a one-mile conversation into a ten- or twenty-mile one wherever somebody has put a machine on a hill. Europe’s equivalent, PMR446, sits at 446 megahertz with half a watt and no license, and the Japanese and Australian services cluster nearby under their own rules. The overlap carries one consequence worth remembering: because FRS and GMRS share channels, an FRS user in a busy area competes for airtime with licensed GMRS operators running ten times the power, which is a reliable recipe for being stepped on rather than heard.

UHF is also where modern off-grid data lives. LoRa radios, including Meshtastic mesh nodes, operate near 902–928 MHz in the United States, turning encrypted text into a resilient, license-free, low-power web of short hops with no infrastructure at all^[26]; RECCO avalanche reflectors use harmonic radar near 917 MHz^[27]. Both are examples of the same trade: superb portability, modest range, terrain-sensitive.

What makes LoRa interesting, and what distinguishes it from every voice system discussed so far, is that it buys sensitivity with time instead of with antenna size. The modulation is chirp spread spectrum: each symbol is a frequency sweep across the channel, and the receiver correlates over the whole sweep, so the effective processing gain is the ratio of the chirp bandwidth to the symbol rate and can exceed thirty decibels. A node configured at the highest spreading factor in a 125-kilohertz channel reaches a sensitivity near minus one hundred and thirty-seven decibels below a milliwatt, which lies below the thermal noise floor of a channel that wide and well below anything a voice receiver could use — at the cost of a data rate in the hundreds of bits per second and an on-air time for a short message approaching two seconds. That trade, of enormous spreading gain for a handful of bytes and of airtime measured in seconds rather than milliseconds, is why the LoRa mesh behaves like a shared party line rather than a network. A dozen nodes in the same valley spend most of their time waiting for a clear channel, a text that would be instantaneous over a cellular link takes a minute to cross five hops, and the network’s capacity is genuinely small. It is a coordination tool with a long tail of usefulness rather than a fast one: capable of moving a position or a short message between people a few kilometers apart with no infrastructure whatsoever, and incapable of doing anything quickly when twenty people all need to speak at once. Meshtastic, the most common open-source incarnation of the idea, runs in the 902 to 928 megahertz band in the United States with default settings near 906 or 907 megahertz, at power levels from tens of milliwatts to about a watt depending on hardware, and it adds what the older license-free services lack: text, position, and encrypted group messaging carried by a self-forming mesh that needs no repeater, no network, and no license. Its limitations are the same as the physics. The radios are still UHF, so terrain still rules and a ridge still blocks; the hop count is limited, so the mesh cannot relay across a valley with no node on the rim; and the airtime cost caps the capacity. What it genuinely improves is the coordination job — knowing where the members of a party are, confirming a rendezvous, getting a short message through where voice would not carry — which files it alongside MURS and GMRS rather than alongside a beacon.

RECCO is the most elegant piece of physics in this subsection and also the most misunderstood. It is not a transmitter and it contains no battery. It is a passive transponder, a small diode-loaded antenna resonating near 917 megahertz, and it works by harmonic radar: the searcher’s detector illuminates the slope, the tag absorbs energy at the fundamental, and the diode’s nonlinear junction re-radiates a small fraction of it at twice the frequency, around 1,834 megahertz. The trick is that virtually nothing in nature generates a coherent second harmonic, so the detector can pick a tag’s faint reply out of clutter that is thousands of times stronger at the fundamental. The physics also sets the range honestly. Because the interrogation must make the trip out and the response must make it back, the returned power falls as 1/R⁴ — twelve decibels of loss for every doubling of range — so the useful distance is on the order of tens of meters through air and considerably less through snow, degrading steeply with burial depth. This is why RECCO is a find-me-and-not-call-me technology: it never calls for help, never runs out of batteries, and only matters once a professional with a detector is already standing near the debris. In that respect it belongs in the same category as the avalanche transceiver, and in a completely different category from the beacon discussed next, which is the device that brings the professional in the first place.

Then there is the reason UHF deserves a higher rank than the pecuniary comparison suggests: the most reliable emergency signal available to a person on foot is a UHF signal, and it is not speech. A personal locator beacon transmits a 5-watt digital burst on 406 MHz, repeating with a slight random offset every fifty seconds, that is received by satellites in low, medium, and geostationary orbits. With GPS encoding, the position accuracy is roughly 100 meters^[4], and the newest MEOSAR constellation — hosted on GPS, GLONASS, and Galileo spacecraft^[3] — typically resolves and forwards an alert in under five minutes. The satellite part ended for 121.5 MHz alerting in 2009^[2]; that analog frequency survives only as a low-power homing tone for aircraft once they are already in the neighborhood. 406 MHz carries your identity too, which is why registration matters^[5]: the authorities know who you are and whom to telephone before anyone is airborne.

The engineering inside that sentence deserves unpacking, because 406 MHz is the rare case in this essay where the emergency system was designed from first principles around the physics of the problem rather than assembled from available parts. The band from 406.0 to 406.1 megahertz is reserved internationally for distress alerting alone, so there is no traffic to compete with and no reason for a receiver to listen anywhere else; the satellites that watch it are listening for one thing. A beacon’s transmission is a burst of about four tenths of a second, phase-modulated at four hundred bits per second, carrying the beacon’s identity, the type of distress, and, if the beacon has a fix, an encoded position — which means the whole message amounts to a few dozen bits that must be correct rather than a conversation that must be intelligible. The repetition interval is around fifty seconds and is deliberately randomized across beacons, because several beacons in the same area can appear on the same satellite at the same instant and identical timings would garble one another. The burst begins with a hundred and sixty milliseconds of unmodulated carrier, and that is not waste: the satellite measures the Doppler shift of that pure tone as it moves along its orbit, and from the shape of the resulting curve it can compute the beacon’s position to within a few kilometers even when the beacon carries no navigation receiver of its own. To make that measurement possible, the carrier has to hold its frequency to a couple of parts in a billion across the burst, which is why beacons use temperature-compensated oscillators rather than ordinary crystals. Registration is the other half of the design: the identity in the burst maps to a database record containing a name, a telephone number, and often an itinerary, so a rescue coordination center can place a call to a spouse or a hiking partner before it ever launches an aircraft. Since the overwhelming majority of alerts turn out to be false, that telephone call is not a bureaucratic afterthought but the cheapest filter in the whole system.

The second-generation system changed the geometry rather than the signal. The original Cospas-Sarsat payloads flew in low orbits at around eight hundred and fifty kilometers, which meant that a beacon waited for a satellite to pass over it and that a position required the whole pass to be measured. The medium-Earth-orbit payloads now hosted on GPS, GLONASS, Galileo, and BeiDou spacecraft orbit some twenty thousand kilometers up, and there are enough of them that any point on the earth has several in view at any instant; the alert is therefore detected and localized within a handful of minutes rather than tens of minutes, and the fix no longer depends on a single Doppler pass. With a navigation fix encoded in the message the accuracy is around a hundred meters, which turns a search area measured in square kilometers into one measured in hectares. All of that is delivered by a five-watt burst a little under half a second long. The energy arithmetic is the part that deserves to be famous. Each burst spends about two joules, so at one burst every fifty seconds a beacon’s average power draw is roughly forty milliwatts, and twenty-four hours of continuous alerting costs under four kilojoules — against the four hundred kilojoules a five-watt voice transmitter would consume merely holding a carrier for the same day. That is a hundredfold saving in battery for a message carrying more useful information than the voice transmission would, and it is why a personal locator beacon can be specified to transmit for at least twenty-four hours at minus twenty degrees Celsius on a battery that then sits in a pack for five years. A small packet, sent rarely, encoded against error, and aimed at a satellite is simply a better use of the same watt than a spoken sentence aimed at a person.

Which leaves the 121.5 megahertz homing transmitter, the part of a beacon’s design most often mistaken for a legacy leftover. It is not an alerting channel any longer, since the satellite service stopped listening to it in February 2009, and the frequency is now shared with a great many things that are not beacons; what remains is a continuous, unmodulated, very low power tone of tens of milliwatts, emitted alongside the 406 megahertz bursts so that an aircraft arriving in the area can direction-find the beacon with a standard airband receiver. Its range is short because its power is small, and that is the point of it: it is a terminal guidance aid for the last few kilometers of a search rather than a distress signal in its own right. A beacon whose 406 section has failed and whose 121.5 transmitter is still running is a beacon that will very likely be found by somebody who is already looking for it, and a beacon nobody is looking for. The same architecture appears in three packages — the personal locator beacon carried in a pack, the maritime EPIRB that floats free and activates itself, and the aviation ELT that fires on impact — differing in mounting, activation, and the assumptions about who is around to notice, but not in the signal.

Satellite messengers follow the same logic from the other direction. Iridium’s network, which carries the Garmin inReach system, operates near 1616–1626.5 MHz — still nominally UHF. The reliable SOS of the modern backcountry is not a voice transmission at all; it is a short, high-value, machine-managed packet delivered by satellite.

The commercial systems work on the same principle with a different business model. Iridium operates sixty-six satellites in low orbits at about seven hundred and eighty kilometers, in six planes inclined so that coverage is genuinely global including the poles, and its user links occupy 1616 to 1626.5 megahertz — nominally UHF even at the top of the band’s definition. Globalstar, whose simplex channels carried the early satellite messengers, transmits in the same neighborhood of the spectrum. The reason a pocket device with one or two watts can reach either constellation is the aperture arithmetic from the physics section, read in the other direction. The path loss to a satellite seven hundred and eighty kilometers away at 1.6 gigahertz is on the order of a hundred and fifty-five decibels, which is ruinous for a handheld; what rescues the link is that the far end has a large, well-designed antenna on a stabilized platform with an unobstructed view of an entire hemisphere. The satellite supplies the aperture that the handheld cannot carry, and it is the one station in this whole essay that a party does not have to bring with them.

That dependence has a practical consequence worth stating without embellishment, because it is where the useful advice lives. The link margin on these devices is small — single-digit decibels in the ordinary case — which is why the manufacturers insist on a clear view of the sky, why a messenger can work flawlessly on an exposed granite slab and fail completely under a thick canopy, and why a party’s position should be logged before it descends into a canyon. Recall that a thirty-meter depth of foliage costs several decibels at these frequencies; on a terrestrial link with thirty decibels of margin that penalty is an inconvenience, and on a satellite link with three or four it is the entire outcome. Nothing about the radio has changed in either case. Everything about the geometry has.

The two approaches also differ in who is on the other end, and the difference is not small. A 406 megahertz alert travels through an internationally standardized, government-funded architecture — beacon to satellite, satellite to mission control center, mission control center to rescue coordination center — with no commercial intermediary anywhere in the chain and no requirement that anyone find the alert profitable. A satellite messenger’s request for help arrives first at a commercial coordination center staffed by people whose job is to triage messages and then telephone the relevant authority, which in practice is usually fast and occasionally very fast, and which depends on a company continuing to operate and on a human deciding to escalate. The systems are complementary rather than competing, and the honest pattern for a serious backcountry party is a beacon for the alert and a messenger for the conversation: the beacon to summon the aircraft, and the messenger to tell the family, correct the destination, and explain that the aircraft is no longer needed. That is the shape of the modern answer, and it explains the recommendation with which this subsection closes. A beacon is not a better radio; it is not a radio at all. It is the only device in this entire essay whose whole purpose is the single job that matters most, which is why it repays its cost more reliably than anything else in backcountry communications, and why it succeeds by refusing to attempt the easy half of the problem.

In UHF you get the trade the backcountry rewards most: a small, efficient antenna, a radio that runs all weekend on a pocket battery, and the one channel in this entire essay with a documented, measurable rescue record — 406 MHz. Funding a beacon before upgrading a radio is the single highest-return purchase in backcountry communications.

HF (3–30 MHz) — Range That Ignores the Horizon

#

HF is the only band that can summon help from beyond the horizon without satellites. That sentence is the whole case for it, and it is a strong case. The 80-meter and 40-meter bands, both HF, subdivide naturally into a night pair and a day-and-night regional workhorse; 20 meters and above carry long-haul traffic when the sun cooperates; 60 meters offers a small number of channels near 5.3 MHz with modest power limits that behave much like 40 meters. Below all of them, 160 meters at 1.8–2.0 MHz is technically in the medium-frequency range — a superb nighttime regional band whose half-wave antenna, however, is well over two hundred feet long.

Each rung of that ladder sits where it does for a reason, and the reasons were derived earlier as arithmetic. A wavelength of nearly eighty meters puts eighty meters at the low end, where vertical rays return to earth once the critical frequency has fallen with the sun, and where the daylight absorption penalty is at its most brutal because the one-over-f-squared term in the D region is largest for the longest waves; it is a night band by construction rather than by habit. Forty meters sits just beneath the daytime critical frequency, which is the whole trick of it: low enough for most of the day to bring near-vertical signals back down regionally, and high enough at night that it stops paying the daylight tax, which makes it the only band that works both shifts. Twenty meters sits above the daytime critical frequency and below the daytime maximum usable frequency, which is the definition of a long-haul band, and it therefore lives and dies with the solar flux. Ten meters needs a critical frequency that can be magnified up to its own ceiling by a sufficiently oblique path, which is why it is legendary near solar maximum and silent near minimum, and why the years immediately ahead will be mostly silent on it.

The antenna arithmetic is what makes the lower rungs uncomfortable, and it is worth doing because it explains the shape of every backcountry HF decision. A half-wave dipole comes out at roughly four hundred and sixty-eight divided by the frequency in megahertz, in feet, so eighty meters wants about a hundred and twenty feet of wire, sixty meters about ninety, forty meters sixty-five, twenty meters thirty-three, ten meters sixteen, and the two-meter band, for reference, three. That is the real reason 160 meters belongs to the medium frequencies and to a particular kind of operator: its half wave is around two hundred and fifty feet, a length given above and equal to the length of a short city block. On forty meters, though, a sixty-five-foot dipole is a wire that rolls into a ball the size of a fist, and the practical art of backcountry HF consists entirely of getting such a wire high enough to matter — which is to say twenty or thirty feet up a tree and not a hundred, because what governs a high-angle NVIS pattern is height measured as a fraction of a wavelength, and a fraction of forty meters is a small thing. Sixty meters, being just below forty, behaves like forty, which is why the channelized arrangement near 5.3 megahertz — five discrete channels, upper sideband only, and a hundred watts of effective radiated power in the United States^[10] — has become the band of choice for the kind of emergency traffic that would otherwise clutter its larger neighbor.

For the backcountry, three properties of HF matter most. The first is that it is terrain-proof: the ridge, the canyon, and the forest canopy between you and the nearest town are simply not on the propagation path. The second is that it is infrastructure-free in the way that counts; a wire thrown over a tree branch and 10 or 20 watts is a complete station, and amateur emergency traffic nets exist on a schedule in most populated regions. The third, and the least pleasant, is that it demands knowledge: band choice by time of day, antenna deployment, propagation forecasting, and a discipline of listening. An HF radio in the hands of someone who has not operated it is worth roughly a kilogram of dead weight.

Terrain-proof is the correct description with one qualification that operators learn the hard way, and it is worth stating plainly because it separates a working station from a frustrated one. The ionosphere does not care about the ridge, but the launch does. A skywave signal departs the antenna at an elevation angle, the angles useful for long-distance work are shallow — five to twenty degrees above the horizon — and the antenna must therefore be able to see the horizon at those angles. A ridge five hundred meters away and two hundred meters tall sits about twenty-two degrees above the antenna’s horizon, and it permanently caps the take-off angle: the steep paths still work, the shallow ones are gone, and the station is reduced to whatever the local geometry permits. In a genuine canyon this is not hypothetical. The practical consequence is the mirror image of the VHF advice: where a VHF operator climbs to see over the terrain, an HF operator in a hole needs a path steep enough that the rim cannot obstruct it, or a launch angle shallow enough to clear the horizon in another direction entirely. Full NVIS performance needs take-off above sixty degrees, which almost any canyon allows, and that is why the NVIS compromise is the backcountry HF mode of necessity rather than of convenience.

The forest canopy, meanwhile, is a much smaller problem here than at VHF, and the earlier arithmetic says exactly why. The same power law that made a thirty-meter depth of foliage cost about six decibels at two meters makes the same thirty meters cost roughly two decibels at eighty meters, because the frequency term drives the loss and eighty meters is nearly forty times lower in frequency. A party under a canopy that has already taken twenty decibels from a two-meter signal has lost almost nothing at all at forty meters. Set beside that the path-loss arithmetic and the reason a ten-watt HF station is not a toy becomes clear. Free-space loss on a five-megahertz path of six hundred kilometers, up and back down again, comes to about a hundred and two decibels, which is less than the hundred and sixteen decibels the same formula assigns to a mere hundred-kilometer line-of-sight path on two meters. A wave that leaves your wire, strikes the ionosphere, and lands three hundred kilometers away has paid a smaller geometric toll than a VHF signal traveling a sixth of the distance, which is the entire reason HF closes links at ten watts that no quantity of VHF power will close at all. What HF does not escape is the receiver’s companion, the noise floor thirty-odd decibels above thermal at five megahertz, and against that adversary the answer is the antenna rather than the amplifier.

Infrastructure-free deserves its own honesty, because the phrase can mislead. The nets exist and they exist on schedules: regional emergency and traffic nets meet daily or weekly on fixed frequencies at fixed hours, the maritime and weather nets that serve blue-water traffic hold watch on twenty meters around the clock, and the whole apparatus is staffed by volunteers who have agreed to listen at a particular time rather than continuously. That is the difference between hearing a repeater’s carrier and hearing a person. On VHF the listener is present whenever the machine is on; on HF the listener is present when the net is, which makes the operating plan for a backcountry station a schedule rather than a calling channel. Two arrangements genuinely work. The first is a pre-arranged sked with a home station — a licensed friend or a club that has agreed to listen on a specific frequency at a specific hour — which is a private appointment against a public resource and, by a wide margin, the most reliable HF plan available to a party on foot. The second is checking into a scheduled net at its published time, which requires knowing its frequency, hour, and reporting procedure in advance. Both share a property that no plan on any other band requires: somebody at the far end is expecting to hear from you. Both also share the chain this essay keeps tracing, since the amateur on HF rarely launches anything. They telephone, and the telephone call becomes the rescue. HF’s summoning is mediated, and its reliability is a property of the mediator.

What remains is the knowledge, and there is no way to buy it short of operating. The controls are few but they interact: which band, chosen by the hour and the solar flux; how high and how oriented the antenna, chosen by whether the goal is regional NVIS or long-haul; whether the ground or counterpoise at the feed point is adequate, since a random wire with a poor return path radiates much of its energy in the wrong directions or down the shield of the feedline; and whether the noise the receiver reports is the band’s or the station’s own. An operator with ten watts, the right band, and a well-deployed wire will work stations that a hundred watts on the wrong band never hears, which is a twenty-decibel swing and the reason the remark about dead weight is not rhetorical. The competence is worth acquiring, and the only way to acquire it is to take the radio out, then take it out again, until the band plan has stopped being a table and become a habit.

There is a civilian bridge here too. CB at 27 MHz is HF, and despite its reputation it is a genuine off-grid voice option at 4 watts AM or 12 watts single-sideband in the United States, license-free, with forty channels and a long, awkward antenna. SSB CB at dusk can reach out thirty to fifty miles over good ground, and it is more capable than most people who dismiss it realize — though its popularity has faded to the point that the far end is often silent, which is the fatal weakness of any low-population band.

CB is worth the same precision. The band runs from 26.965 to 27.405 megahertz in forty channels at ten-kilohertz spacing, with channel 9 at 27.065 as the emergency convention and channel 19 at 27.185 as the highway channel, and it is license-free in the United States, where the ceiling is four watts on amplitude modulation and twelve watts peak envelope on single sideband^[8]. The gap between those two modes is wider than the numbers imply, because an amplitude-modulated signal spends most of its power on a carrier that conveys nothing at all while a single-sideband signal puts everything into one sideband and compresses it into roughly half the bandwidth. In practical talk power the sideband channel is worth something like ten decibels more than the amplitude-modulated one, which is the difference between a voice that is intelligible and a voice that is merely present, and it is why the older users who cared about range used sideband while the ones who cared about compatibility did not.

The range arithmetic splits into two regimes, and an honest account keeps them apart. The first is horizon-limited, and the calculation from the line-of-sight section applies unchanged: two stations with antennas ten meters up each see about seven miles to the horizon and fourteen miles between them, thirty meters of mast gives twenty-four miles, and a hundred meters of effective elevation at both ends gives about forty-five. The thirty-to-fifty-mile figure the band is credited with is therefore arithmetically sound if and only if both ends are elevated, which is exactly the situation of two ridge-top vehicles and exactly not the situation of two hikers in a valley with whips on their packs; a handheld CB, whose antenna is a compromise of a compromise, will not manage a mile. The second regime is the one that produced the band’s mythology. Twenty-seven megahertz is HF, so it can be carried by the ionosphere: sporadic-E clouds in late spring and summer, and the F layer near the top of the solar cycle, will lift a four-watt signal a thousand miles on a good evening, and the celebrated skip contacts of the CB era were entirely real. What they were not was dependable. Skip follows the mood of the ionosphere, arrives and departs within minutes, and is as likely to deliver somebody else’s conversation on top of yours as it is to deliver yours anywhere useful.

The antenna is also why CB is a vehicle technology rather than a pack technology. A quarter wave at 27 megahertz is a whisker under three meters, a nine-foot whip, and a half wave is eighteen feet; neither is something a person carries on a hike, though both are unremarkable on a truck. The other half of the problem is the thin population, and it deserves emphasis because it applies to every low-density band in this essay. CB’s popularity collapsed after the 1970s, and what remains is a modest population of rural users, hunters, ranchers, and over-the-road drivers — enough that a conversation on a logging road is plausible, and not enough that a distress call is likely to be heard by anyone able to act on it. There is no CB equivalent of a coast station, no published schedule, no net, and no organization whose purpose is to listen; the old requirement that licensees monitor channel 9 at all times disappeared decades ago and left the emergency channel a convention with no custodian. A band whose physics is sound and whose population is thin fails for reasons that have nothing to do with physics, which is the lesson the low bands have been teaching throughout, and which is why the verdict of this essay does not change when the wavelength becomes interesting.

MF and LF (300 kHz–3 MHz, 30–300 kHz) — The Skipped Middle and the Deep End

#

Notice that the question as posed jumps from HF straight to LF, omitting the medium frequencies in between. That gap deserves a moment, because two useful things live in it. The first is the AM broadcast band and 160 meters, discussed above: real, functional, and hopeless for a person on foot, because the antenna must be a meaningful fraction of a wavelength measured in hundreds of feet.

Filling in the gap is worthwhile because the medium frequencies are not empty; they are where broadcasting, maritime distress, and the bottom of the amateur spectrum all pile up. In the Americas the AM band runs from 530 to 1700 kilohertz, and in most of the rest of the world from 531 to 1602 at nine-kilohertz spacing, and its engineering is identical to what the physics section described for ground wave: a vertically polarized field attached to the earth, stable and fadeless by day, delivered a hundred kilometers or more by stations radiating from quarter-wave antennas. The height arithmetic is the whole problem in miniature. A quarter wavelength at one megahertz is seventy-five meters, about two hundred and fifty feet, and while a broadcast station either owns such a tower outright or loads a shorter one with a matching network, a person on foot owns neither the structure nor the network. Everything else that lives in this region of the spectrum carries the same sentence with it: the maritime MF channels where voice distress calling was born, the 160-meter amateur band at 1.8 to 2.0 megahertz with its two-hundred-and-fifty-foot half wave, and the 500-kilohertz Morse distress channel of the wireless era, retired from service but not from history. The medium frequencies’ remarkable ground-wave reach is real, and it is shelved, because on these bands reach is purchased in wavelength and wavelength is purchased in steel.

There is also an observation buried in the way the question itself was posed, which is that its list skipped the medium frequencies entirely, jumping from HF straight to LF. The omission is itself a form of answer. Even the person asking assumed that a band whose antennas are measured in hundreds of feet could not figure in the decision, and they were right; the medium frequencies have always been the bands of the built environment, of cities with transmitters and of ships with half-wavelength antennas hung from masts, and the backcountry has never been their address. The gap in the question was not an oversight but a competent instinct, and this section exists mainly to vindicate it while also rescuing from it the one device that matters.

The second is easy to miss and genuinely important: avalanche transceivers operate at 457 kHz^[28], which is MF. They are the one life-critical radio device a backcountry traveler routinely carries on the lower bands, and their design teaches the lesson this whole essay is circling. They are simple to the point of crudity, use a single international frequency, transmit a continuous tone rather than a message, and achieve a range of only tens of meters — deliberately. The device is optimized for one job, in one environment, with one failure mode. It works because the engineering question was never “what is the best band?” but “what is the most reliable band for this specific problem?”

The avalanche transceiver makes that sentence concrete, because every one of its design decisions is visible in its physics. Its frequency, 457 kilohertz, was not chosen by market habit but by the properties of the medium it must see through. Snow and ice are lossy dielectrics whose attenuation climbs steeply with frequency — the same water-dielectric mechanism that makes foliage cost more at UHF than at VHF, with the stakes raised to the point where a 900-megahertz signal would be swallowed within the first meter of wet snow and a 457-kilohertz one escapes a two-meter burial with most of its margin intact, because its attenuation in consolidated snow is only a small fraction of a decibel per meter. The wavelength at that frequency is six hundred and fifty-six meters, which sounds as though it should forbid any hand-held antenna, and it does forbid an electric one; the transceiver therefore uses a magnetic antenna, a ferrite rod wrapped in a coil, which stays compact precisely because the dimensions that matter for a loop are those of the rod, not of the wave.

That ferrite rod is also why the search has a geometry to it. A small loop radiates and receives in a figure-eight pattern with a null along the rod’s axis, so a buried transceiver is heard differently depending on how it lies and from which direction it is approached — and this is not a defect to be engineered away but the basis of the search itself. Rescuers do not walk the debris on a grid; they walk the flux line, the magnetic field path that emerges from the buried unit’s loop and curves back toward its two nulls, because walking along that line leads directly to the transmitter. The procedure, refined over decades, is a coarse pass until the signal is strong and the direction stable, a fine search that follows the flux line and watches the reading tighten, and then the probe and the shovel; the whole sequence is a guided walk down a magnetic field curve rather than a radio conversation. The range is therefore deliberately short, on the order of tens of meters, because the people who will find you are standing next to the debris and the debris field has the dimensions of a slope, not of a valley; a longer range would multiply the ambiguities of the search rather than help it, since in a multiple burial the strongest signal normally wins the searcher’s attention and the weakest victim needs it most. And the time budget is minutes, since survival after a complete burial falls steeply after the first quarter of an hour, which is another way of saying that the device is designed for the person who was already present, not for anyone who must be summoned.

The single worldwide frequency is the most deliberate decision of all. The alpine rescue authorities and the radio regulators settled on 457 kilohertz as the one frequency every transceiver transmits on and every transceiver listens to, with no channels, no programming, and no user choice, because the only failure mode that cannot be tolerated is that a transceiver of one maker, from one country, bought in one decade, should fail to converse with a transceiver of any other maker from anywhere else. A second channel would be a second feature, and a second feature would be a second failure mode. Modern sets modulate the carrier with a digital frame that carries identity and lets the display point the way, but the core specification has not moved. Now consider what the device refuses to do, because the exclusions are as instructive as the inclusions. It does not transmit a voice. It does not call for outside help, since there is no repeater, no satellite, and no rescue authority at the far end. It does not even report its own depth with any precision. Each of these capabilities was considered and refused, and the refusals are what make the instrument reliable for the single thing it does: letting the person standing on the debris find the person under it. The avalanche transceiver is therefore the control case for this entire essay. It carries no antenna argument, no band trade, no license question, and no learning curve, because the engineering question was never which band is best but what is the most reliable possible device for the moment when everything has already gone wrong within a few dozen meters of the one person who can help. The answer is a continuous tone at a frequency that sees through snow, a short range, and a figure-eight pattern. Everything else is elaboration.

Above 457 kHz, the amateur 630-meter band (472–479 kHz) and the 2200-meter band at 135.7–137.8 kHz exist as experimental allocations with watt-level effective radiated power and enormous, inefficient antennas. They are fascinating, they are mathematically humbling, and they belong to the laboratory, not the trailhead.

The humbling part is best appreciated as arithmetic, because these two allocations are where the antenna physics earlier in this essay reaches its purest form. The 630-meter band, 472 to 479 kilohertz, sits just above the old 500-kilohertz distress channel, and the 2200-meter band at 135.7 to 137.8 kilohertz is the lowest allocation an amateur may use anywhere. In both, the license limits are set in single-digit watts of effective radiated power, which means the transmitter itself may run a hundred watts or more while the legal ceiling applies to what remains after the antenna has consumed the difference. A twenty-meter vertical — already a structure no one carries — is thirty-one thousandths of a wavelength at 472 kilohertz, with a radiation resistance of a few tenths of an ohm and an efficiency against any realistic ground system in the low single digits of percent, so a hundred watts becomes three radiated and still crosses a continent on a good night because the propagation follows the long-wave rulebook. On 2200 meters the same twenty-meter vertical is nine thousandths of a wavelength, its radiation resistance is a few hundredths of an ohm, its efficiency is below half a percent, and a hundred watts becomes a half watt that is nevertheless heard thousands of kilometers away. The bandwidth is the companion humiliation: at these quality factors the usable channel is a few hundred hertz at most, which constrains the message forms themselves to hand-keyed telegraphy and very slow digital modes. There is a further structural irony, which is that the ground system carries much of the current and therefore much of the responsibility: on these bands the ground plane is part of the antenna in a literal sense, and a fully worked installation wants radials scattered over an area measured in the tens of meters, which is to say a campsite full of copper wire belonging to a person who is not going anywhere.

None of this is a criticism of the allocations, which are among the liveliest experimental grounds in the hobby, the place where loading-coil lore, ground-system science, and the patience of the human ear all meet. It is a statement of their address. Everything since the ionosphere section has been built on trades, and these are the bands where the trade is all one-sided: nothing to carry, nothing to hear that a satellite cannot deliver in a tenth of a second, and nothing to say that a UHF packet cannot say better in a tenth of a second. The medium and low frequencies are not where the backcountry decision happens; they are where the physics of wavelength shows itself in its most humbling form, which is worth a section of an essay whose subject lies elsewhere — and which brings it, next, to the one region even below these, where the wavelength stops being a technical constraint and becomes a physical condition of the planet itself.

VLF (3–30 kHz) — The Physics Nobody Can Backpack

#

VLF is where the planet itself becomes part of the antenna system. Ground-hugging waves and the earth–ionosphere waveguide carry signals across oceans with remarkable stability, which is why VLF is used for time dissemination, navigation, geophysical surveying, and submarine communication, and why it penetrates soil and seawater better than anything else in the spectrum. It is also, molecule for molecule, the worst practical choice for a person in trouble. Wavelengths of ten to one hundred kilometers mean antennas measured in kilometers, efficiencies measured in single-digit percent, transmitted bandwidths of a few hundred hertz at best, and data rates that make a telegram look impulsive. There is no version of this that fits in a rucksack, and there is nobody on a mountaintop listening to 20 kHz.

The sentence about the planet is literal before it is rhetorical. A wave whose wavelength is measured in tens of kilometers cannot treat the ground as a surface it travels over; the ground becomes the return conductor it travels with, and the current that completes the antenna circuit returns through the earth over an area that spans kilometers of soil and rock. The transmitting station’s counterpoise, in other words, is the continent. The sky is no less involved, because the same plasma physics that refracts HF explains why the ionized lower atmosphere is opaque to these waves: at twenty kilohertz the wave sits far below the D region’s plasma frequency, so the refractive index in that region is imaginary and the wave cannot enter it — the sky presents a wall rather than a window, and the wall is the second conductor of a duct. Two conductors a few tens of kilometers apart, a spherical guide around the whole planet, and attenuation rates of one to three decibels per thousand kilometers: that is the system the big stations live inside, and it is why a transmitter in Maine is heard in Japan with a stability good enough to discipline clocks.

Naming the services attaches the physics to real machinery. The time services are the most familiar: WWVB at 60 kilohertz in the United States and DCF77 at 77.5 in Germany^[16] are strictly the low-frequency extension of the same duct, the reason a radio-controlled wristwatch can set itself to an atomic clock on another continent, and the reason a clock is the one thing a person actually gets from these bands. The navigation systems are historically the maritime and military uses, from the VLF Omega network of the navigation era to the Russian Alpha system still transmitting position signals in the ten-to-fifteen-kilohertz range; the geophysical surveyors borrow the navy’s transmitters and measure the distortion of their fields over a target area, which is how the same station that talks to a submarine maps a mineral body. The submarine service itself deserves its scale: the stations that reach the submerged fleet are among the largest radio installations ever built, with towers in the hundreds of meters, antenna systems a kilometer or more in span carrying a megawatt or more of transmitter power, and one of them, the Alexanderson alternator station at SAQ in Sweden^[17], still radiating 17.2 kilohertz on its anniversary days from a hundred-year-old mechanical generator that is a UNESCO world heritage site. None of that equipment has been shrunk in a hundred years, and none of it can be.

The penetration claims quantify themselves the same way the earlier sections have quantified everything else. Skin depth in a standard medium goes as the inverse square root of frequency, so twenty kilohertz in rock of average conductivity reaches about a hundred meters where one hundred megahertz in the same rock reaches about a meter and a half: seventy-fold in depth, and the same story in seawater, where the earlier arithmetic gave two meters at twenty kilohertz against twenty-five centimeters at a megahertz. That is why these wavelengths are how institutions look into the ground and under the sea, and also why none of that capability survives contact with the trailhead: the transponder end of a rock-penetrating system is a transmitter site, and the receive end is a geophysical instrument, and there is no middle-sized version because the physics is the size. And the data reality completes the humbling: with a hundred hertz of bandwidth even a perfect channel carries under seven hundred bits per second, and real VLF circuits deliver tens, so a short telegram occupies the better part of a minute of continuous transmission and a voice conversation is not a meaningful thing to ask for at all. A system can be enormous and still be slow, because the two properties are purchased with the same currency, and VLF spends everything it has on being enormous.

Which makes the final clause of the paragraph the truest sentence in it. There is nobody listening, because the listeners at twenty kilohertz are machines — clocks, receivers on submarines, instruments on a rock — backed by institutions, and a person with a radio in the mountains confronts a band that answers none of the three jobs: it cannot summon, it cannot coordinate, it conveys only the time to someone who already owns a watch. It is not a failure of the band that it serves no one on foot. It is a statement that the band’s customers are not individuals, and the essay’s question was about individuals.

The correct verdict on VLF is not that it is useless. It is that its advantages — enormous range, rock penetration, immunity to terrain — were purchased with size, power, and bandwidth, and they accrue only to institutions that already own those things.

That verdict generalizes, and it is worth naming what it generalizes to, because it is in fact the thesis of the entire spectrum as this essay has walked it. Every advantage in the spectrum is purchased in some currency, and the currencies trade against one another along a straight line: at the bottom of the spectrum the price of range is size, paid by institutions who can afford transmitter sites and megawatts; at the top the price of portability is range, paid by individuals who accept that their radios speak only to the near horizon; and in the middle of the spectrum, at HF, the price is knowledge, paid by individuals who will never own a transmitter site but can own a skill. The bands do not differ in kind. They differ in who pays, and in what. VLF is simply the end of that line where the payment is so large that the customer is a navy, a national laboratory, or a broadcast authority — the allocation of enormous range demands the purchase price of enormous structure, and the two sides of the transaction have never once been divided into a backpack.

There is one more honest observation to file beside the verdict, and it concerns the receiving side, which this essay has occasionally distinguished from transmitting. At VLF the distinction nearly collapses. A receiver on these frequencies can be small — the loop antenna in a radio-controlled clock is the size of a thumbnail — because receiving does not need radiation resistance, which is why a person genuinely can extract the time signal on a wristwatch. But that is all the band offers to a person: the time, a datum, burned one-way from an institution. There is no conversation to be had, no position to be exchanged, no help to be asked, and no answer to be given; the information that a VLF clock delivers is precisely the information whose absence this essay has been treating as the fourth and least acknowledged job of emergency communications. Time is the one datum that makes every other plan hold — the schedule of a net, the hour of a sked, the patience to wait for a pass — and a person on foot can carry it in a watch without carrying a radio at all. That, and not an antenna farm, is the most useful thing VLF has ever handed to an individual.

Which is where the middle of the spectrum meets the top, and where the essay’s question finally resolves itself. The institutional bands below HF are where the physics of wavelength is purchased wholesale; the highest band in the essay, which follows directly, is where the same physics is purchased in a different currency altogether.

Microwave and Above (1 GHz upward) — Where Data Lives

#

Conventionally, microwaves begin around 1 GHz and run up to the edge of infrared. The physical signature is short wavelength and correspondingly small antennas, which is exactly what satellites need, and the baggage is strict line of sight, sensitivity to rain, snow, and wet foliage, and the need to point something accurately. This is the band of GPS downlinks, of Iridium and Globalstar, of Starlink’s Ku- and Ka-band links, of Wi-Fi and Bluetooth.

The signature is a scale, and the scale is the story. At one gigahertz the wavelength is thirty centimeters; at ten it is three; at twenty, where the Ku band carries the downlinks of the modern commercial constellations, it is a centimeter and a half; and at thirty, the Ka band, it is a single centimeter. Small antennas follow as a matter of definition, since an efficient radiator wants to be an appreciable fraction of a wavelength, and a resonant dipole at one gigahertz is fifteen centimeters long while the same dipole at ten is a centimeter and a half — which is why a satellite can carry an antenna at all, and why a person can carry a terminal. The same shrinkage carries the section’s baggage with it, because the loss mechanisms that were background noise at three hundred megahertz become the foreground story here. The water-dielectric mechanism that cost twelve and a half decibels for thirty meters of wet foliage at 2.4 gigahertz costs nineteen at ten and twenty-four at twenty-four, and at these frequencies the mechanism no longer waits for a canopy: rain is the atmospheric expression of it, a slant path to a satellite crosses tens of kilometers of weather, and the specific attenuation that a VHF net would never notice runs to fractions of a decibel per kilometer in drizzle at fourteen gigahertz and to a couple of decibels per kilometer at twenty in a hard shower. That is why the commercial constellations budget ten to twenty decibels of rain margin and degrade their coding rates rather than their links when a storm cell moves across the beam — why a satellite phone call degrades into a scrambling voice or a dropped line precisely during the squall that the caller most wants to report. Snow is the surface version of the same story, a wet radome or a snow-loaded reflector costing exactly those decibels at exactly the moment an alpine emergency is being described through it. And the deepest expression of the physical signature lies above the bands anyone uses: by sixty gigahertz the atmosphere’s own oxygen absorbs a signal at ten or more decibels per kilometer, and the frequency span between thirty gigahertz and the formal boundary with the far infrared is where the air itself refuses to cooperate — which is why the millimeter bands are reserved for short, deliberately contained links and why no emergency has ever ridden on them.

It is worth disposing of one classification fact before the services are named, because a reader coming straight from the UHF section will notice it immediately. The GPS downlinks at 1575.4 and 1227.6 megahertz, Iridium’s user band at 1616 to 1626.5, and Globalstar’s at 1610 to 1626.5 all sit below three gigahertz, which makes them UHF under the strict band plan and microwave under the convention this section has adopted, and the overlap is not pedantry. It is the reason the satellite services have occupied two subsections of this essay: they are the only things in either band that an individual actually lifts, and they work by renting another institution’s altitude rather than by owning any frequency of their own. The classification dispute is the economics in miniature.

The signature service of the region deserves its own paragraph, because it is the quietest signal this essay has discussed and one of the most consequential. The civil GPS signal arrives at the earth’s surface at roughly a hundred and sixty decibels below a watt — a figure so small that it lies about twenty decibels below the thermal noise of the receiver’s own front end, which is to say that the navigation signal a hundred million devices depend on is, at the moment it arrives, a whisper quieter than the noise of the listener. It is recoverable only because its designers knew in advance what it would say: the receiver multiplies the noise by a copy of the 1.023-megachip-per-second spreading sequence, and one millisecond of coherent integration buys about thirty decibels of processing gain, enough to lift the whisper out of the roar. The price of that recovery is the same price this essay watched LoRa pay in the UHF section — sensitivity bought with time and with foreknowledge rather than with power — and the decoded prize is a data stream of fifty bits per second carrying a clock and an ephemeris. The defining technology of the modern outdoors is therefore, molecule for molecule, a data radio working twenty decibels below its own noise floor; and the symmetry with the avalanche transceiver of the MF section is exact. Both are instruments that refuse every task but their own; the difference is that the transceiver’s perfection is a loud continuous tone at 457 kilohertz, and the constellation’s is a whisper below the noise of everything that listens.

The need to point something accurately completes the baggage, and it has the most concise arithmetic in the section. The beamwidth of an aperture runs at roughly seventy times the wavelength divided by the diameter, in degrees, so a one-meter dish at twenty gigahertz — a meter being larger than any antenna this essay has allowed a person to carry — produces a beam about a degree wide, against twenty-one degrees for the same meter at one gigahertz. Direction at these frequencies is no longer a category of operation but a mechanical act, and the modern terminals have answered the act with electronics rather than with skill: the phased array in a satellite terminal, and in the patch arrays inside a telephone, steers its beam electrically across a wide hemisphere with no moving part, turning pointing from a practiced gesture into a property of the hardware. But a phased array is a small aperture pretending to be a large one, and small apertures at high frequencies are where the aperture arithmetic comes due in the other direction: the terminal that needs no manual alignment accepts less gain per radiated watt and a sharper demand for open sky. The instrument has not abolished the physics; it has moved the physics into the structure and the coding.

Microwave is not a voice band for the backcountry. It is the band that carries the modern “phone call”: texts, positions, photographs, weather data, even voice over a satellite link that is really data pretending to be telephony. It matters enormously in an emergency plan, but as bandwidth and data, not as a radio conversation. It is also the band most dependent on a clear view of the sky, which is precisely the condition a person in trouble cannot always provide.

The verdict of the section deserves to be folded into the essay’s framework before the comparison table begins, because it is the strongest single confirmation of what the spectrum has been saying since the ionosphere section. At this end of the spectrum the three jobs have become data transactions: a summoning is a short burst to a constellation, a coordination is a position pin, an awareness is a weather file, and the “phone call” itself is data pretending to be telephony, compressed to a few thousand bits per second and sounding like a voice arriving from the bottom of a well. The modern emergency call is not a radio conversation; it is a network operation conducted through a radio, and the distinction matters because it changes what fails. A radio conversation fails for radio reasons — propagation, interference, power — and those reasons are the ones a second radio, a higher antenna, or a different band can sometimes answer. A network operation fails for network reasons that no amount of transmitting will cure: the constellation’s schedule, the terminal’s registration, the message queue, the account balance. The greatest change the microwave region has worked on the practice of backcountry radio is therefore not that it is faster or clearer; it is that the operator’s relationship to the equipment has shifted from owning a channel to renting a transaction. The walker does not own the satellite, the gateway, the network, or the spectrum; the walker rents the sky one burst at a time, which is why the emergency-planning literature treats a messenger not as a radio to be mastered but as a subscription to be maintained.

The dependence on a clear view of the sky is the honest price of all of it, and the price is worse than it looks, because the correlation is perverse. The conditions that close a microwave link — canyon walls, storm decks, wet canopy, rain on the slant path — are precisely the conditions under which an emergency is most likely to be reported from the backcountry, and the narrow margins the UHF section described for satellite terminals do not improve with frequency: the decibels a messenger fights for through a wet canopy at 1.6 gigahertz become tens of decibels on a Ka uplink in a storm^[24]. The band that carries the modern rescue is the band the weather can shut down, and the weather that produces the incident and the weather that loads the link are the same weather. None of that makes microwave useless, and the honest restatement is exactly as important: the rescue system of the twenty-first century terminates in these bands, with the messenger’s position pin, the weather file, the voice from the well, and the direct-to-phone text that the newest constellations have begun delivering to unmodified telephones through trials and early service in 2025 and 2026 — even when the last meter of the path, the 406-megahertz burst that actually summons the aircraft, travels on UHF. The point of the section is placement, not dismissal. Microwave is where the modern emergency is coordinated, monitored, and documented; it is the planner’s band and the network’s band, and the grade is not a demotion but a taxonomy: it is not the band for a radio conversation, because the band itself no longer holds conversations. It holds transactions.

And with that the band-by-band walk is complete; the summary that follows compresses it into a table and a verdict.

Summary: The Bands at a Glance

#

Which Band for Which Emergency

#

The Verdict

#

If the question is taken literally — which single band should a person on foot depend on — the answer is VHF, and it is not a close race among the practical candidates. It wins on the three criteria that actually determine success in the backcountry: the largest number of potential listeners per watt, the best terrain and canopy behavior of any band with a carryable antenna, and the best antenna-to-weight ratio in the spectrum. A 5-watt handheld and a ridge will outperform equipment costing twenty times as much, and the people who perform mountain rescues are listening on that band.

The verdict’s first clause deserves its own arithmetic, because it is the only clause in the essay that stacks three independent wins on the same band. Consider the listeners first. The largest number of potential listeners per watt is not a technical property; it is a social geometry. VHF is the band on which the infrastructure of casual attention lives: the national simplex calling frequency that every amateur in the country is expected to guard, the repeater machines that clubs maintain on summits and water towers, the marine channels that every vessel with any radio watches, the aviation band, and the weather transmitters that a whole population leaves on in kitchens. A watt at two meters addresses an ecosystem of monitoring that no other band has, because no other band carries the professional services and the volunteer culture on the same spectrum. The terrain behavior is the physics this essay has already walked, now stated as a competition: thirty meters of wet foliage costs 5.7 decibels at two meters against 9.5 at nine hundred, a knife-edge ridge costs roughly six and a half decibels per edge at two meters and measurably more on the bands above it, because the longer wave bends around the corner with less to give up, and the lower band then surrenders those decibels to a far better horizon — the line-of-sight horizon that answers two pocket radios at three and a half miles stretches past a hundred when both ends stand on a two-thousand-meter ridge. No band with a carryable antenna does better on all three, and the antenna itself closes the argument: a half wave at two meters is a meter of wire that rolls into a pocket, where the same half wave at forty meters is twenty meters of wire that must be thrown over a branch, and at 160 meters is eighty meters that cannot be thrown over anything a person will find. The contrast with the twenty-times-more-expensive station is therefore not a slogan but a ratio: the expensive station’s advantage in reach is real, and it is real only where the operator has already supplied the knowledge the cheap band gives away. And the final clause of the paragraph is the one to underline, because rescue services are not hobbyists: the people whose working duty it is to find you — park rangers, mountain rescue teams, sheriff’s deputies, and the aircrews on the aircraft band — staff their radios on VHF, and every listener on that band is a potential first link in the chain that ends with a helicopter.

If the question is taken seriously rather than literally — which band is most likely to save your life — the answer changes to UHF, for reasons that have nothing to do with voice. The 406 MHz alert, and the satellite links that carry modern messengers, are the only backcountry emergency technologies with a demonstrated, auditable rescue record. Every one of them is UHF or above. In a genuine emergency, the feeling of being in contact by voice is not the same as having an alert in a rescue coordination center; the beacon wins, and it is not close.

The change of question is the whole argument, and it deserves a paragraph of its own before the rescue record is invoked. VHF wins the question of which band a person on foot should depend on, because depend means talk, coordinate, and maintain contact. UHF wins the question of which band is most likely to save a life, because save means a rescue system that acts whether or not anyone hears the caller. The two answers are different currencies, and the confusion of them is the most expensive mistake in backcountry radio: the hour spent raising a friend on a repeater while a beacon sits unregistered in a pocket is the hour in which the rescue machinery is doing nothing. The demonstrated record is the second half of the argument, and it is genuinely auditable. The 406-megahertz alert path, and the satellite links that carry the modern messengers, are the only backcountry emergency technologies whose outcomes are counted: the constellation that listens for beacons has assisted in the rescue of 66,956 people in 21,054 distress situations since the system opened in 1982, including 3,211 people in 2024 alone^[1], in a system whose every detection is recorded, dated, and attributed by the rescue coordination centers that act on it. No voice band in existence can produce an audit trail like that, because voice carries no automatic record of its own effect; a QSO that saves a life is indistinguishable, at the moment it happens, from one that does not. And the precision matters as much as the volume: the modern alert path is minutes from keypress to coordination center, the beacon embeds its position from the same satellite navigation constellation this essay has already described, its registration tells the coordinator whom to contact on shore, and the only part of the chain below UHF is the final homing signal at 121.5 megahertz that the rescuers’ own radios chase over the last hundred meters. The band of the summons is UHF and the band of the search is VHF, which makes the two verdicts of this section a division of labor rather than a contradiction. The last sentence of the paragraph is the summary of the whole point: the feeling of being in contact is a sensation produced inside a radio, while an alert in a coordination center is a sensation produced inside a system that has already begun moving. One of them describes the emergency to a friend; the other starts the machinery. The beacon wins, and it is not close, because the question was about saving life, and the machinery is what saves it.

HF is the third pillar and the most underrated in genuinely remote country, because it is the only band that does not require you to be visible to the person you are calling. It is also the most demanding, and its demands are paid in knowledge rather than money.

The third pillar earns its paragraph less for what the band does than for what it refuses to rent. Every other answer in this section requires that the far end be visible to the caller — a repeater on a summit, a satellite in a sky, a station behind a ridge must all stand somewhere the wave can be drawn to. HF alone works without a line and without a lens: the wave goes up a few hundred kilometers, comes down a few hundred kilometers away, and the person at the far end need not exist within any horizon that geometry can draw, which is the property the paragraph names and the one that matters in country where the catalog of visible listeners is empty. The demands are the counterweight, and they are exactly as stated, paid in knowledge rather than money. Money buys a one-thousand-five-hundred-dollar transceiver and a quarter of that again in accessories, and the equipment sits dead in a rucksack unless the operator already knows which band will close at which hour — when the absorption floor of the ionosphere lifts at dusk and the low bands come alive, when forty meters goes regional at night, when twenty meters opens another continent and is useless to anyone within four hundred miles of the operator — knows how to throw a dipole into a tree and load it against a tuner, knows the traffic procedure of a net and the discipline of a sked, and holds a license that was earned in an examination room rather than purchased at a counter. Those pieces of knowledge are the antenna. They weigh nothing, and they cannot be improvised: the first HF contact in an emergency is a contact with an unfamiliar radio under a deadline, and that is the one recipe the band will not forgive. VHF is a dial and a push-to-talk; HF is a repertoire. That is the entire difference in one sentence, and it is also why the paragraph calls the band underrated: the repertoire is learnable, free, and permanent, and it is the only purchase in the spectrum that improves with the years instead of aging with the firmware.

MF, LF, and VLF lose for one reason that no amount of enthusiasm overcomes: you cannot carry the antenna. Between them they contain wonderful physics, one indispensable safety device at 457 kHz with an intentionally tiny range, and no viable voice channel for a person on a trail. Microwave is not a voice band at all, but it is where data, and therefore increasingly where help, actually travels.

The last paragraph of the section is a pair of verdicts that share a single logic, and the logic is worth stating once, because it is the whole spectrum summarized in a sentence. The lower bands lose not on physics but on carriage: the 160-meter half wave is eighty meters of wire, the 630-meter quarter wave is one hundred and sixty, the 2200-meter allocation hangs from structures worked against a wavelength of more than two kilometers, with radiation resistances in the low tens of milliohms and efficiencies below half a percent, and every one of those measurements is a complete sentence — the band does not have an antenna problem, the band is an antenna problem. The 457-kilohertz beacon is the exception that proves the sentence, because it abandoned everything the other lower-band services would have to carry: no voice, no range, no generality, only a continuous tone from a ferrite rod that fits in a palm and a figure-eight pattern that guides a partner walking the debris — and it is precisely by refusing every job but one that it became the one indispensable device on the lower bands. The microwave verdict closes the pair from the other end of the scale. It is not a voice band, which the earlier sections have established down to the arithmetic of rain, canopy, oxygen, and pointing, and it is where help actually travels, because help has become data before it becomes a helicopter: the position pin, the weather file, the messenger text, the direct-to-phone message crossing the constellations in trials and early service, the burst that arrives in a coordination center while the caller still has charge to spare. The same sky-dependence that disqualifies microwave as a conversation band leaves it supreme as a cargo band, and the two clauses of the paragraph are a single judgment. The person on foot carries VHF to talk and UHF to summon; HF covers the country where neither answer reaches; everything below HF is physics for its own sake; and everything above UHF is data that increasingly arrives with a rescue inside it. That distributes the verdicts cleanly across the spectrum, which is exactly what makes the essay’s final question answerable: whether the person carrying the kit matters more than the kit itself.

What Actually Matters More Than the Band

#

Three habits matter more than the band you choose, and they cost nothing.

The three habits are worth their place at the end of a long essay, because the physics has been building toward one unglamorous conclusion: the spectrum decides what participation is possible, and habits decide whether participation happens. A band is a property of the planet; a habit is a property of the person, and every difference this essay has measured between the bands is smaller than the difference between a person who acts on a band and one who merely carries it. Hardware is bought once; habits are renewed on every trip, which is the literal sense in which they cost nothing — they tax the account that draws on experience rather than the account that spends money. They also align with the four jobs the essay opened with. Elevation serves every band at once; the schedule converts coordination and summoning from intentions into machinery; and redundancy guards the job the essay treated as the fourth and least acknowledged — the one no signal performs, which is not needing one.

Elevation, geometry, and antenna height. Move uphill, get off the canyon floor, hold the radio away from your body, and mount the antenna as high and as clear as possible. This single practice outperforms every upgrade in transmit power available to a backpacker.

The practice outperforms every power upgrade because height and power are not currencies of equal value. Power is bought in decibels at the transmitting end only; height is earned at both ends of the path at once, and its arithmetic comes straight from the line-of-sight section. Two radios at pocket height answer each other across a horizon of about three miles; if one operator climbs a hundred meters above the canyon floor, that number stretches to nearly fourteen — a fivefold change for the price of walking uphill, with the decibel account banked before the first transmission. The same rule scales the antenna side of the habit: the essay’s own six-to-sixty comparison, in which a tenfold gain in height multiplies the horizon by the square root of ten and is worth ten decibels of path budget — the same ten decibels that would require ten times the transmitter power to buy. There is a second, less famous reason height matters, and it is the Fresnel zone: the first Fresnel radius of a ten-kilometer two-meter path bulges nearly seventy meters above the straight line between the ends, and any obstacle intruding into that invisible ellipsoid costs real decibels, which is why an antenna at the top of the pack pays even on paths where the horizon is not the limiting factor. The body is the nearest obstacle of all: a handheld pressed against the torso wastes several decibels, sometimes six or more, into the operator’s own tissue — the water-dielectric mechanism this essay has been quantifying since the foliage section, with the attenuator now strapped to the radio — which is to say that the way a radio is held can eat an entire upgrade from five watts to twenty. And the habit is the only one in the essay that multiplies every band at once: horizon for VHF, sky view for the satellites, takeoff angle for HF — and even the bands that like low antennas want the operator high, since NVIS asks the dipole to hug the ground but asks that the ground beneath it be a summit rather than a canyon floor.

A schedule and a listening habit. A radio that is switched off is a brick. Amateur and GMRS nets meet at known times; the Wilderness Protocol gives the 2-meter calling frequency a monitoring convention; NOAA weather broadcasts continuously on VHF. A written check-in plan, in which someone expects a call at a specific hour on a specific frequency, converts a modest handheld into an actual safety system.

The schedule is the most misunderstood of the three habits because it looks like a bureaucratic act when it is actually a physics act. A channel is defined by frequency, and a radio is defined by time: the two coordinates in which a signal exists are where it is transmitted and when it is sent, and the written plan fixes both, turning a transceiver that competes with silence into a transceiver that arranges to be heard. The Wilderness Protocol is the canonical example: a monitoring convention on the two-meter calling frequency, 146.52, in which stations guard the first five minutes of every hour and treat the windows at seven in the morning, ten, one, four, seven in the evening, and ten at night as the hours when transmissions are expected. Amateur traffic nets institutionalize the same idea on the lower bands, with a check-in roster that creates the listener before the emergency exists; NOAA weather broadcasts run continuously so that the receiving side never needs scheduling at all; and the family or party plan does the same work at the smallest scale, with a fixed hour, a fixed frequency, and a fixed expectation. The deep property of a written plan is what a missed check-in becomes: silence is converted into signal. A radio that was merely carried contains no information when it is quiet; a radio on a schedule contains a message even when it transmits nothing, because its silence at the appointed hour is itself a report that triggers the watcher’s telephone call — and the summoning chain suddenly has two independent starts, the keypress and the missed sked. The schedule also budgets the battery, since a radio whose transmissions are planned spends its chemistry almost entirely listening, and it prices the habit in a currency this essay has already met: a one-minute check-in inside a four-hour window is a duty cycle of less than half a percent, which is the same parsimony the avalanche transceiver and the position-reporting beacon practice by design.

Redundancy with different physics. Do not stack three VHF devices and call it redundancy. The point of a plan is that the tools fail for different reasons: a satellite beacon that works when the terrain defeats everything line-of-sight, a VHF handheld that works when the sky is socked in, an HF station that works when there is no infrastructure and no satellite pass, and a written itinerary that works when every battery in the group is dead.

The redundancy paragraph is the place where the engineering habit of mind pays for itself, because its subject is not the gear but the failure modes. Three VHF handhelds share one physics, and therefore share the fate of that physics: the same ridge that defeats one defeats all three, and the stack is a budget line rather than a plan. Real redundancy requires that the layers fail for different reasons, and each of the four layers chosen here occupies a different corner of the failure space. The beacon works when terrain defeats line-of-sight because its far end hangs thousands of kilometers up rather than on the next ridge, and it fails in the corners a user can see: a canyon that walls off the low sky, and water over the antenna, since conductive water absorbs UHF within centimeters. The VHF handheld fails when the ridge stands between, and is untouched by the sky entirely, because a two-meter wave passes through cloud as if the cloud were not there. The HF station fails when the operator’s skill fails, or when the band itself is dead at that hour and season, and is untouched by both terrain and weather. And the itinerary fails only at the receiving end, if nobody reads it — the single layer with no electronics and no propagation, which is exactly why it is the layer that still works when every battery is flat. The arithmetic of diversity is that independent failures compound by multiplication rather than by addition: if each layer has a fair chance of failing on its own, the four fail together only when the failures coincide, and coincidences of independent mechanisms are the rarest events in engineering. That is the entire justification for carrying physically different tools, and it is also why the essay’s verdict could not have been a single band: a band is a single layer, and a plan is a stack.

Since a genuinely life-threatening emergency generally permits a station in distress to use any means available, the most valuable preparation is not a license or a band plan but familiarity: knowing what the far end sounds like, what the antenna does when it is nine feet up rather than thirty, and how your own hands behave when the stakes are real. See also my earlier treatments of [[Satellite messengers vs. Radios]], [[Ham, GMRS, MURS]], and [[Overlanding-Radio]], and the more aggressive position I took in [[Why You Should Transmit on ANY Vialble Frequency in a SHTF Situation]] — the physics in this essay is what makes that argument safe to make only as a last resort.

Familiarity is the habit the other three rest on, and it is worth unpacking the paragraph’s two examples, because each points at a different kind of knowledge. Knowing what the far end sounds like is the acoustic side of the schedule: the difference between a repeater’s tail and a simplex contact, the rhythm of a net, the sound of a station that is hearing you and one that is not, the way a distant signal acquires a hollow, watery edge as it fades — all pattern recognition that only repetition installs. Knowing what the antenna does at nine feet rather than thirty is the geometric side of elevation made tactile: the arithmetic of that particular change is about five decibels, the same budget as tripling the transmitter power, and a person who has felt it once on a real ridge will climb the extra twenty feet in weather without needing to be convinced. And knowing how the hands behave when the stakes are real is the part no specification sheet can hold: motor skill that has been rehearsed survives the adrenaline, while motor skill that has only been intended does not, which is why the amateur service’s institutional answer to this exact problem was always the drill — the net, the exercise, the annual Field Day — a rehearsal of the ordinary so that the extraordinary has something to fall back on. The paragraph’s closing reference to the more aggressive doctrine is the essay’s own cautionary coda: transmitting on any frequency when everything else has failed will work precisely as well as the operator’s familiarity with the receiving end, because an unrecognized voice on an unknown band is only noise to the listener who could help. The physics in this essay is what gives that last-resort doctrine its one honest justification, and the practice this paragraph demands is what keeps it from being a dangerous joke. All of which compresses into the honest answer that closes this section.

The honest answer to “which band is best” is a sequence, not a frequency. Carry a beacon that transmits on 406 MHz, because that is the signal that produces a helicopter. Carry a VHF handheld because that is the band with the most ears and the best antenna. Carry HF, or a satellite messenger, if your definition of “remote” includes the word “beyond.” Get high, get on a schedule, and practice in daylight on a clear day what you intend to do at night in a storm — because the band you chose was never the hard part. Being the person who knows how to use it is.

References

#

The factual claims in this essay rest on primary documents rather than secondary summaries: the regulations as written, the operators of the distress systems, and the ITU-R recommendations behind the propagation arithmetic. Numbers derived in the text, and judgments that are mine, carry no citation.

1. Cospas-Sarsat. “Cospas-Sarsat Distress Alerts.” System totals: 66,956 persons assisted in 21,054 distress situations since 1982, including 3,211 persons in 1,076 situations during 2024. https://www.cospas-sarsat.int/en/system-overview/cospas-sarsat-distress-alerts
2. Cospas-Sarsat. “Beacon FAQ.” Confirms that legacy analog beacons transmitting only on 121.5 or 243 MHz are no longer monitored by the satellite system, and that satellite alerting requires a 406 MHz signal. https://www.cospas-sarsat.int/en/beacon-faq
3. Cospas-Sarsat. “Transition to MEOSAR.” Development began in 2004; payloads fly on Galileo, GLONASS, GPS, and BeiDou; initial operational capability was declared in April 2023, with full operational capability anticipated as early as 2026. https://www.cospas-sarsat.int/en/system-overview/transition-to-meosar
4. NOAA SARSAT. Search-and-rescue brochure on GPS-enabled 406 MHz beacons (PDF). Detection within three minutes and a location estimate within five minutes of activation for GPS-encoded beacons. https://www.sarsat.noaa.gov/wp-content/uploads/SAR-GPS_Brochure_lowres.pdf
5. NOAA SARSAT. Program home page and beacon registration guidance, including why registration shortens a false-alarm investigation and speeds a real response. https://www.sarsat.noaa.gov/
6. 47 CFR § 95.1767, “GMRS transmitting power limits”: 50 W for mobile, repeater, and base stations on the 462/467 MHz main channels; 15 W for fixed stations; 5 W ERP and 0.5 W ERP on the two sets of interstitial channels. https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-95/subpart-E/section-95.1767
7. 47 CFR § 95.567, “FRS transmit power”: 0.5 W ERP on channels 8–14 and 2.0 W ERP on channels 1–7 and 15–22. https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-95/subpart-B/section-95.567
8. 47 CFR § 95.967, “CBRS transmitter power limits”: 4 W mean carrier power for AM and FM voice, 12 W peak envelope power for single sideband. https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-95/subpart-D/section-95.967
9. 47 CFR §§ 95.2763 and 95.2767, MURS: five channels at 151.820, 151.880, 151.940, 154.570, and 154.600 MHz, with a 2 W transmitter power limit. https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-95/subpart-J/section-95.2763 and https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-95/subpart-J/section-95.2767
10. 47 CFR §§ 97.303(h) and 97.313(i), the 60 m band: five discrete channels between 5330.5 and 5406.4 kHz, and a 100 W PEP effective radiated power limit. https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-97/subpart-D/section-97.303 and https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-97/subpart-D/section-97.313
11. 47 CFR § 97.403, “Safety of life and protection of property”: no provision of the amateur rules prevents a station from using any means of radiocommunication at its disposal for the immediate safety of human life. https://www.ecfr.gov/current/title-47/chapter-I/subchapter-D/part-97/subpart-E/section-97.403
12. NOAA National Weather Service. “NOAA Weather Radio.” Seven VHF channels from 162.400 to 162.550 MHz, continuous broadcast, no license required to receive. https://www.weather.gov/nwr/
13. ARRL. “Band Plan.” Lists 146.52 MHz as the National Simplex Calling Frequency in the two-meter band. https://www.arrl.org/band-plan
14. ARISS. International Space Station amateur radio operations, including FM voice downlink near 145.800 MHz. https://www.ariss.org/
15. NIST. “Radio Station WWVB.” The 60 kHz time and frequency service. https://www.nist.gov/pml/time-and-frequency-division/time-distribution/radio-station-wwvb
16. Physikalisch-Technische Bundesanstalt. “DCF77.” The 77.5 kHz legal-time dissemination service. https://www.ptb.de/cms/en/ptb/fachabteilungen/abt4/fb-44/ag-442/dissemination-of-legal-time/dcf77.html
17. Alexander Association. SAQ, the Alexanderson alternator station at Grimeton, Sweden, still transmitting on 17.2 kHz for anniversary events. https://alexander.n.se/en/
18. NOAA Space Weather Prediction Center. “Solar Cycle Progression.” Observed and predicted sunspot number and radio flux for Cycle 25. https://www.swpc.noaa.gov/products/solar-cycle-progression
19. ITU-R P.525, “Calculation of free-space attenuation,” and ITU-R P.526, “Propagation by diffraction,” the source of the knife-edge loss curve used in the diffraction section. https://www.itu.int/rec/R-REC-P.525/en and https://www.itu.int/rec/R-REC-P.526/en
20. ITU-R P.833, “Attenuation in vegetation,” which collects the empirical foliage fits, including Weissberger’s and the COST 235 model. https://www.itu.int/rec/R-REC-P.833/en
21. ITU-R P.372, “Radio noise,” the basis for the atmospheric noise figures quoted for the MF and HF bands. https://www.itu.int/rec/R-REC-P.372/en
22. ITU-R P.368, “Ground-wave propagation curves for frequencies between 10 kHz and 30 MHz,” and ITU-R P.684, “Prediction of field strength at frequencies below 150 kHz.” https://www.itu.int/rec/R-REC-P.368/en and https://www.itu.int/rec/R-REC-P.684/en
23. ITU-R P.533, “Method for the prediction of the performance of HF circuits,” and ITU-R P.531, “Ionospheric propagation data and prediction methods required for the design of satellite services and systems.” https://www.itu.int/rec/R-REC-P.533/en and https://www.itu.int/rec/R-REC-P.531/en
24. ITU-R P.676, “Attenuation by atmospheric gases and related effects,” and ITU-R P.838, “Specific attenuation model for rain for use in prediction methods.” https://www.itu.int/rec/R-REC-P.676/en and https://www.itu.int/rec/R-REC-P.838/en
25. VOACAP. The ITU-R-based HF propagation prediction engine referenced in the ionosphere section. https://www.voacap.com/
26. Meshtastic. Project documentation for the license-free LoRa mesh, including the 902–928 MHz North American band and its airtime and range trade-offs. https://meshtastic.org/docs/
27. RECCO. “Technology.” The harmonic-radar principle behind the battery-free reflector. https://recco.com/technology/
28. Avalanche.org (American Avalanche Association and National Avalanche Center). “Transceiver / Beacon.” The single-frequency, send-and-search architecture described in the MF section. https://avalanche.org/avalanche-resources/transceiver/