Looking at the ocean floor, it’s difficult to say exactly how rough the bottom is. There are individual rocks and coral heads, then large and relatively flat features. There are small coral organisms growing, each with their own size and shape.

Instead of trying to measure each feature from the centimeter scale on up, oceanographers roll all of that into one number called the friction factor, which is really just a measure of how much a particular bottom grabs at the water sliding over it. It is a bit like the grit of sandpaper, telling us whether the bottom is generally coarse or fine.

Over a smooth floor only a thin layer right at the bed is really dragging. If you put obstacles in the way, the water has to go around them, peeling off the back of each one and leaving a swirl behind it. Those swirls break down into smaller and smaller ones until the motion ends up as heat. Taller obstacles, spaced so the water has to thread between them, mean more swirls and a thicker band of churned up water above the bed for the wave to drag along with it.

A sandy bottom hardly takes any energy from a wave. The ripples are small and rounded and the water mostly slides over the top of them, which is why models of the surf zone often leave friction out altogether and get an answer close enough to the real deal. A rock bottom does more, since ledges and boulders give the water a bit more to catch on. A healthy coral reef sits so far past both that the bottom can take more out of a wave than breaking does over the whole crossing.

The difference between sand and coral is not only the size of the features, it’s how far up into the water the resistance reaches. Over sand the pushing and pulling happens on ripples a couple of centimeters tall and stays pinned near the bed. Over coral it acts on every branch, plate and lobe on the reef. A single head presents several times more surface to the flow than the patch of seafloor it sits on, and the heads are packed closely enough that water threading between them moves noticeably slower than water passing over the top. The dragging layer now fills the canopy from the substrate to the tips of the coral, all of it working on the wave at once.

How much any of this costs depends steeply on how fast the water at the bed is moving. The energy pulled out goes as the cube of that speed, so water moving twice as fast over the same bottom takes eight times as much out of the wave overhead.

Depth is also an important factor in this. Orbital motion weakens with distance below the surface, so in deep water the roughness barely sees the wave at all, and as the water shallows the same wave drives progressively stronger flow across the reef. Long-period swell carries its motion deeper than short-period wind swell, so it starts paying the friction tax further offshore. A wide, shallow reef in front of the break subtracts from the swell across the entire crossing.

All in all, friction takes energy out of the wave and never gives it back, so the wave arrives at the break smaller than it otherwise would have. It does that continuously across the whole shallow crossing rather than all at once, which is why the width and depth of the rough ground matter as much as how rough it is.

Further Reading:

Paper on reef friction

Share

What makes the horseshoe different from other shapes is that the peel angle changes along the ride. It starts at zero at the peak and opens up as the break walks outward along each arm.

In the figure above, the blue lines are the crest at three moments as it comes in. They are drawn straight, since by the time a wave is this close to the beach refraction has already done most of its work. The red line is the edge of the bank where the water is shallow enough to break the wave. At the peak the crest and the red line are parallel, so the angle between them is zero and that part of the wave breaks all at once. That is where these waves start to throw their lip.

As the crest keeps coming in, it meets the bank further out on each arm, where the edge has curved back toward the beach, and the angle between the crest and the edge opens up. In the drawing it goes from zero at the peak to around 30 degrees and then past 40.

What this looks like from a surfer’s perspective is a steep drop where the angle is near zero, a fast section right after it, a wall you can work as the angle passes through the range from last week, and a shoulder to kick out on as the arm drops into the deeper water on either side. A beautiful ride, and in this case, symmetrical on the left and right.

Let’s say the wavefront hasn’t fully refracted or the horseshoe isn’t directly perpendicular to shore. When the swell arrives at an angle, the crest meets the bank off center. The peak shifts toward the side the swell is coming from (the left in the figure above), and the two arms get different angles, because on one side the crest and the bank edge are lined up with each other and on the other they are working against each other. In the second figure the left runs at around 25 to 35 degrees and the right at 30 to 45, so the left is the faster wave. This is why one direction is better than the other on a given swell, and why the same bank can favor the other direction a week later when the swell has swung around.

The curve does one more thing, and it is why the peak is bigger than the shoulder. Shallow water slows a wave down, so the part of the crest over the middle of the bank falls behind the parts over the deeper flanks, and the crest bends in toward the apex. The direction a wave carries its energy is the direction the crest is facing, so a crest bent inward sends energy inward, and more of the wave arrives at the peak than would have if the bottom were flat. The peak is taller as well as shallower, and that forces a bigger lip.

So two factors make a difference here: one, the peel angle evolves during the ride which creates makeable sections, and two, the shape focuses wave energy at the takeoff, leading to a lip you can duck under on a good day.

Further Reading:

Peel angle

Share

In studying waves, we tend to quantify this as an angle rather than a speed. Imagine hovering above the break and watching it roll in towards the beach. The angle that the white water makes with the crest of the wave is the peel angle. In this frame, zero degrees is a closeout, with all of the wave breaking at once and there being a line of white water parallel to the beach.

A higher peel angle, such as 60 degrees, means the wave is running out of depth progressively along its length rather than all at once. The break has to walk sideways down the crest to get to the end of the wave, and you get to ride ahead of it.

Let’s get a bit geometrical. If the wave is moving at speed c and the peel angle is α, the breaking point travels along the crest at c divided by the tangent of α. You have a harder job than the breaking point, because you have to move toward the beach as well as sideways, so your required speed is c divided by the sine of α. Both get larger as the angle shrinks. This means that as the angle gets smaller, you have to go faster. For a perfectly aligned closeout, you’d literally have to be moving infinitely fast to surf it all the way.

The practical floor for surfing was put at 30 degrees in 1974, from work out of the Look Laboratory at the University of Hawaii. Most of what we ride sits between 45 and 66 degrees. Above about 70 degrees, the wave stops peeling in any useful sense and crumbles toward the beach instead of down the line, which is fine for some white water surfing but not much more than that. Those numbers shift with wave size, since a bigger wave gives you a steeper face to drop down and more speed to work with, so an angle that shuts you out at two feet can be fine at six.

Peel angle also isn’t fixed along a single wave. The bottom changes underneath it as it comes in, and the angle changes with it. Where the angle drops, the break gets out ahead of you and you have a section to make. Where it rises, the break slows down and hands you a shoulder to work with. A wave that keeps collapsing and recovering that way is what we’re describing when we call a wave sectiony.

Swell direction has less to do with the angle at the break than you would think. Refraction turns the crest toward the beach for the entire time a wave is in shallow water, and by the time it breaks it has mostly finished the job. A long-period swell aimed 30 degrees off the coast arrives at around six, and no offshore direction gets it much past twelve. Both are effectively closeouts.

A wave breaks where it runs out of depth, so the line the breaking point traces is the depth contour where that happens, and the peel angle is the angle between the crest and that contour. Since refraction has already pulled the crest nearly parallel to the beach, what you are really measuring is how crooked the shallow ground sits relative to the shoreline. A point, a reef, a ledge, or a bank of sand angled across the beach all do the same job, holding a line that the beach itself does not. The shape of the bottom is the shape of your ride. The best shape for it is a curve, shallow in the middle and deeper on both flanks, but that will have to wait until the next one.

Further Reading:

Paper on peeling waves

Share

The definition of breaking is more specific than a wave falling apart. A water surface wave is a shape moving through the water while the water itself mostly stays put, circling in place and ending up about where it started. Breaking is when that stops being true. The water up in the crest gets moving nearly as fast as the shape is traveling, runs out of wave to circle back down into, and just keeps going forward on its own.

All it requires is that the wave's own size changes how fast it travels. Big parts have to travel at one speed and small parts at another.

A few hundred meters down, warm water sits on top of cold water, and waves run along that boundary the same way surface waves run along the boundary between water and air. The two layers differ in density by a fraction of a percent instead of the factor of eight hundred between water and air, so gravity makes the waves come out enormous and slow. The biggest we’ve measured stand over two hundred meters tall. They shoal up the continental slope, they steepen, and when the water inside one catches the wave, they break. The overturning lip can be a hundred meters of it, with nothing on the surface to show for it but a slick.

In a sound wave the air moves back and forth along the same line the wave travels, not in circles across it. Even though the waves are fundamentally different forms, again, what matters is only whether a wave’s size changes its speed.

The squeezed parts of a sound wave are denser and warmer than everything around them, and the air in them is itself moving forward. Both make that part of the wave travel faster, the motion more than the warmth. So the loud part is always gaining on the quiet part in front of it, barely, for as long as the sound is traveling. Every loud sound leans forward as it goes, and given enough distance, they all end up in the same shape, a sawtooth, no matter what they started as.

You’d think the front would eventually fold over itself, but that would mean the air pressure at one spot had three different values at the same time. Air doesn’t do that. So instead of a fold you get a cliff, with the pressure jumping from one value to another across almost no distance. This is called a shock, and thunder, sonic booms, and the crack of a whip coming off a tip that went supersonic are all examples of one. It’s the sound wave’s version of breaking.

Brightness doesn’t change the speed of light in empty space. A bright patch of a light wave travels exactly as fast as a dim one, so the bright part can never gain on the dim part, and nothing ever leans. Quantum effects break that rule at some field strengths, but for any practical purpose, light in a vacuum is the one wave that isn’t allowed to break.

Pushing enough light through glass, though, causes the light to slow down where the beam is brightest. Intensity starts setting the speed, and the leaning starts up again, backward this time. The bright peak lags its own dim edges, so the back of the pulse steepens toward a cliff of its own, an optical shock.

In air, a pressure can’t hold two values at once, so a barrel gets replaced by a sonic boom. In water, a wave breaks as it does because the surface of the ocean is a shape hanging in space instead of a number at every point, and a shape has somewhere to fold into. Breaking is something waves generally do. A surface is just the only place with room to throw a lip. And light in a vacuum is just crazy.

Further Reading:

Waves

Transverse waves

Longitudinal waves

Share

This piece is about measuring how that canyon affects the swell that gets to Black’s. How much of Black’s is actually the canyon, which swells it works on, where the extra energy comes from, and what breaks have their energy sapped by it.

In this experiment, I use an open-source wave model applied to the local bathymetry. Celeris is a GPU solver fast enough to simulate every individual wave in a sea state on my laptop. I ran a swell over the real bathymetry, then ran the identical swell again over a seafloor with the canyon smoothed away. Any difference in how the swell arrives at the break between those two oceans is therefore the canyon’s doing. I ran twenty swells this way, four periods from 14 to 20 seconds across five directions from due south to west-northwest, spanning the window this stretch of coast receives.

This is the third piece in the First Order series, one question surfers have, answered in layers you can drag. The interactive map runs all twenty comparisons over the real seafloor, with breaks from Torrey Pines to Windansea marked: open the explorer →

Zeroth order: a plain shelf

On a featureless shelf the seafloor treats every beach the same. We have covered refraction before: waves feel bottom, slow down, and turn toward shore, and on a plain shelf that turning is even, so every stretch of sand receives the same wave the same way. In this picture two beaches three kilometers apart should be near-identical. Along most of this county’s sand they are, which is exactly what makes the exception worth an article.

First order: a canyon is a lens for waves

A kilometer off the sand, this shelf holds a remarkably steady twenty meters of water. Follow that line seven kilometers from La Jolla Shores to Torrey Pines and the bottom barely moves, except twice. Directly off the Shores, and again just north of Scripps Pier, the floor falls out to more than 120 meters: the two heads of the La Jolla and Scripps submarine canyons, each just a few hundred meters across.

A wave only feels the seafloor once the water is shallower than about half its wavelength. For a 6 second windswell that is 28 meters, so short-period chop skims over this entire shelf, canyon and all, without feeling much of anything. An 18 second groundswell feels bottom at 250 meters, and thus the whole gash is in play on its way in. The canyon does not exist for windswell, but it exists for groundswell. The longer the period, the more of the canyon exists.

Shallow water slows waves down. The stretch of a crest passing over the deep slot keeps its speed while the parts on either side drag on the shelf, so the crest bows and the wave’s energy paths, the rays, bend away from the deep. Behind the canyon, the rays spread apart, and just outside its rims, they crowd together. Crowded rays mean the same energy squeezed into less coastline, so bigger surf. Spread rays mean the opposite, weaker surf. A trench in the shelf is a lens, and like any lens, it cannot brighten one spot without darkening another.

In 1947, Walter Munk and M. A. Traylor published wave refraction diagrams for the La Jolla canyons, tracing swell paths over the measured bathymetry by hand to predict where the energy lands, methods that grew out of the wartime surf-forecasting problem. The grey rays in the explorer are the same construction, computed live in your browser over the same seafloor, almost eighty years later.

The experiment

Ray diagrams predict where the energy should go. The model actually moves it, wave by wave, and the two oceans (real and canyon-removed) let us see how the canyon changes things: divide the wave height over the real seafloor by the wave height with the canyon erased. Red is where the canyon makes that swell bigger, blue is where it makes it smaller. Working in ratios also cancels the model’s main weakness, a slow numerical bleed of wave energy with distance, because both oceans bleed identically.

On southerly groundswell the focus lands on South Peak at Black’s: about 80 percent bigger at 16 to 20 seconds, the canyon close to doubling that swell at the peak. A second hot band sits on the canyon’s north rim, halfway between the Shores and the pier, and it holds the largest number in the whole matrix, peaking above plus 100 percent on an 18 second swell from due south. The energy for both of these comes out of La Jolla Shores. In those same souths, it gives up 40 to 65 percent, growing worse as the period stretches.

From due south the hot bands sit at the rim by the pier and at South Peak. By west-southwest the brightest band is South Peak proper. From the west and west-northwest the focus strays from Black’s and lands on Torrey Pines. Stretching the period at a fixed direction causes the band to migrate the same way, north.

The shadows scale simply: the longer the period, the deeper the shadow, the Shores sliding from minus 40 percent on a 14 second south to minus 65 on a 20 second one, exactly what the depth-reach argument predicts. The focus is messier. Amplification at a fixed spot can drop as the period stretches, not because the canyon’s effect weakened but because the focal band moved. From the west, South Peak reads plus 72 percent at 14 seconds and minus 6 at 20, while Torrey Pines, the next stop north, holds between plus 30 and plus 54 across the same stretch. Longer period does not simply mean a stronger canyon effect at your break. It can also mean the focal band is now somewhere else.

In the real-seafloor runs, the Black’s stretch comes out bigger than the pier and the Shores in nineteen out of nineteen readable swells. But on many of those swells, the canyon is barely boosting Black’s at all. Rather, it is gutting the neighbors, taking 64 percent off the pier and 89 off the Shores during a long west-southwest.

The reputation for Black’s Beach is half amplification and half everyone else’s shadow, and luckily, the shadow half has been measured directly. When Magne and colleagues ringed this canyon with wave buoys for a 2007 study, they measured wave height dropping by a factor of five from one side to the other, over a distance shorter than a single wavelength.

Second order: which part of the canyon does the work

The canyon is really two things: a gorge hundreds of meters deep running offshore, and a set of narrow fingers cutting the last kilometer of shallow shelf. Because we can erase it in two layers, we can ask which portion does the most to the swell. Filling only the shallow fingers and re-running a 16 second west swell, the pattern along the beach comes back nearly unchanged. Leaving only the deep gorge still moves plenty of energy, about 60 percent as much, but in broad strokes, offshore. The detail that determines which peak breaks where comes from the fingers. This is one swell, so hold it loosely, but the part of the canyon that matters most to your session appears to be the last shallow kilometer, not the deep gorge behind it.

And now, a word on what this can’t do. The maps show the shape of the canyon’s effect, not a wave-size forecast, and the model’s absolute heights are not trustworthy by design, which is why every number here is a ratio between the two oceans. Between the five modeled directions the explorer blends neighboring runs, a smooth guide rather than model output. Swells at 12 seconds broke the model, the grid is too coarse to resolve them, and parts of the 14 second runs arrive too weak to read, which the tool shows as grey. The last few hundred meters into the sand, sandbars and shorebreak included, are beyond what we resolved.

What does hold up, though, is the geometry. Like the swell window from the first piece, the canyon’s map is a property of the place. For a given period and direction the pattern repeats exactly (minus effects such as tidal elevation, background currents, wave-wave interaction, etc.), and all twenty patterns are in the explorer →

This is the third post of our First Order series. Tools accompany each piece, so they arrive when the tool is done rather than on a schedule. All tools are free and open-source, built for the curious surfer. Check this one against your favorite forecast next groundswell and let me know how it does.

Further Reading:

Munk & Traylor (1947)

Magne et al. (2007)

USGS CoNED Southern California topobathymetric DEM

Celeris

Share

Seawater is close to transparent on its own. It’s the stuff in it that affects how clear it looks. A sand grain that gets split into a thousand pieces adds no mass to the water relative to its whole form. Each piece is a tenth of the original diameter, so the total surface has multiplied by ten, and all of that new surface is available to bounce light around. The particles are what you cannot see through, and it is their combined area in your line of sight that does it, not their mass.

Wavelength sets how far a wave’s orbital motion reaches down, and wavelength comes from period, so period is what decides whether a swell touches the bottom at all. A one meter wave in 20 meters of water moves the bed back and forth at about 11 cm/s when its period is 6 seconds, and 28 cm/s at 12 seconds.

Medium sand, around two tenths of a millimeter, needs roughly 25 cm/s at the bed under a 12-second wave before it lifts. In 20 meters of water, that’s about where a 1-meter swell of 10 seconds or longer sits. At 6 seconds, the same wave height only puts 11 cm/s on the bed, which moves nothing.

How long the sediment stays up depends mostly on grain size. A 200 micron sand grain falls through 10 meters of water in about 7 minutes, and 20 micron silt takes closer to 8 hours. That’s still-water math, though. Any turbulence in the water column keeps the fine material up far longer, which is why the murk outlasts the swell that made it by days.

After the sediment is suspended, it gets advected around. Speiser and colleagues put Sentinel-2 imagery over 31 sites in Northern California in 2025 and fit an exponential decay to turbidity heading offshore, and they found the turbid zone grew with wave energy and reached kilometers out during large events. Sandy beaches threw a lot of sediment into the water close to shore, but it faded quickly with distance. Rocky shores put in less and carried it much farther out, which is about what you’d expect from better organized rips running over uneven bottom.

Some of the murk shows up from below and has nothing to do with the swell. Internal bores run up the shelf on the tidal beat, scraping sediment off the slope as they go. Masunaga and colleagues tracked bores that displaced isotherms 20 meters vertically in 40 meters of water, with the strongest resuspension right at the head of the bore and the suspended material peeling off into a layer that spread offshore above the thermocline. That’s the structure you’re swimming through when the top few meters are warm and clear and everything below is cold and brown.

Biology also varies how much stuff is in the water. Upwelling brings nutrients toward the surface, a bloom follows within days to weeks, and it can cut visibility the same way sediment does. Runoff adds its own pulse of fine terrestrial material after a storm, and that’s the one case where bad visibility and bad water quality reliably show up together. The rest of the time they’re unrelated. Water you can see a long way through isn’t necessarily water you want in your sinuses, and water you can’t find your own hands in can be perfectly clean.

Further Reading:

Komar paper

Speiser paper

Masunaga paper

Share

This is the second piece in the First Order series, a tool paired with a simple analysis and reporting to answer a surfing question. This one builds upon the first one. That piece held the seafloor roughly fixed and asked which directions can reach a break, and there refraction, the seafloor’s hand on the swell, was the last correction it added. Here the seafloor is the subject, so refraction moves to the front. There are two explorers now, one for the Southern California Bight and one for the New York Bight, both built over the same real bathymetry, GMRT, so we can start looking at shelf-scale effects on surf.

Zeroth order: the swell goes straight

A swell in deep water does not know the bottom is there. Under each crest the water moves in circles that shrink with depth, and about half a wavelength down the motion is gone. As long as the seafloor stays below that depth the wave rolls on untouched and dead straight. At this order the two coasts are identical, and the seafloor might as well not exist. Everything that separates them happens once the wave starts to feel it.

First order: the bottom steers

Once the seafloor rises into those orbits, the bottom drags and the wave slows. Where one end of a crest slows before the other, the whole crest pivots toward the shallow side, the way a shopping cart pulls to the side with a bad wheel. That pivot is refraction, and it is working on the wave the entire time it is in shallow water, right up until it breaks. Because a wave breaks in a couple meters of water no matter the coast, it arrives nearly straight in no matter the coast. The angle a swell makes with the sand is set by how shallow the water is when it breaks, not by the shelf it crossed to get there, barring very extreme bathymetry.

How far out the swell feels the bottom is set by the period, because deep-water wavelength grows with the square of the period. A 6 second swell feels bottom in about 28 meters of water, nearly at the sand. An 18 second swell feels it at 250 meters. So the shelf a long swell has to cross, and the distance over which the bending is spread, is whatever sits above 250 meters. Here is where the two coasts are very different.

The first correction: how much bottom there is to bend over

Off New Jersey the shelf barely dips. The 30 meter contour sits about 16 kilometers out at Rockaway and nearly 40 at Long Beach, and 200 meters is more than 90 kilometers offshore, past the edge of the map. An 18 second swell feels bottom the moment it reaches the shelf and keeps feeling it across every remaining kilometer.

Southern California does the opposite. The bottom falls away almost at once, 30 meters within a few kilometers of most breaks and 200 meters by about six. A swell stays in deep water until the last mile or two, so all of its bending is packed into that thin strip. This shelf is less of a ramp than the lip of a table.

The second correction: early and drawn out, or late and abrupt

The tools trace each swell from deep water into about a mile offshore and stop there, short of the final approach. By then the East swell is most of the way done turning. Every East break has already bent its 18 second swell 13 to 39 degrees, climbing with period, Manasquan running 1, 16, 31, then 38 across the four periods. By this mark, the West swell has barely started turning. Oceanside, Lower Trestles, and Salt Creek read 6 degrees or less at the same point, not because they bend less in the end but because their bending is all still ahead of them, in the last mile the steep shelf saves it for. The two exceptions, Long Beach and Ventura at 42 and 33, are the West breaks that happen to sit over a wider patch of shelf.

A swell bent gradually across tens of kilometers is also moved as it is turned. The rays show the East swell walked a median of 58 kilometers down the coast between deep water and the beach, coming ashore at a break that a straight line from its deep-water heading would have missed by fifty-some kilometers. The West swell is displaced a median of under two kilometers. It lands close to where it was aimed. In some sense, it is easier to understand the path a swell will take to hit SoCal than the Northeast.

And again with the caveats on this model. A ray is the path energy would take if a swell moved like a beam of light, bending over the seafloor but never spreading sideways. A direction that reaches a break in the tool has a clean, full-strength path in. A direction that misses is not getting nothing: real waves diffract, leaking into the shadows the rays leave empty, weaker but not gone, which is the soft-shadow story from the first piece. Where they crowd together, the seafloor is focusing energy and the surf is bigger; where they fan apart, it is smaller. The rays map which directions arrive at full strength by a direct route, not every bit of energy that reaches the shore.

The takeaway here is that an identical 18 second swell breaks facing the sand on both coasts, but it got there two different ways. Off Oceanside it held its line across deep water and turned in the last mile, landing a stone’s throw from its aim. Off Rockaway it was bent the whole way across a shallow shelf and set down fifty kilometers from where it started pointing.

This is the second piece in the First Order series. The tools are free and open-source; please mess around with them.

Further reading:

GMRT bathymetry

The Southern California explorer

The New York explorer

Share

Every break receives swell through a certain window, and no two breaks have exactly the same window. That is the whole answer, and the rest of this piece is about learning to see it.

To get there, we will borrow the standard physicist’s trick for problems too messy to solve outright: solve a cartoon of the problem first, then correct the cartoon in layers. The first pass is deliberately too simple. Each pass after that adds back one piece of ignored reality, and you stop when the answer is good enough for the decision in front of you. The jargon for this is orders of approximation: zeroth order for the cartoon, first order for the first serious correction, second order for the fine print, and higher orders beyond that. When a physicist calls a number “right to first order,” they mean it is the napkin answer: not exact, but close enough to act on, and honest about what it ignored.

That is where this series gets its name. First Order is a set of interactive tools, each built around one question surfers argue about, each answered in layers you can see and drag. The articles walk up the ladder; the tools let you play with it. This first one takes on the question in the title, because swell direction is the corner of forecasting where the napkin answer does the most work. To make it visible I built an interactive map of the Southern California Bight, the long bend of coast between Point Conception and San Diego, over its real seafloor, for ten of its well-known breaks. Every number below comes from it, and you can check each one yourself: open the explorer →

Zeroth order: swell goes straight

A storm is a wave generator parked over deep water. For a day or two it drags on the sea surface, piling chop into bigger and bigger waves, and when those waves finally outrun the wind that made them they become swell: organized, long-crested, and pointed wherever they were heading when they escaped. That direction is fixed at the moment of escape, like a bullet out of a barrel. Deep water barely touches a swell afterward. A 14 second swell covers about 940 kilometers of open ocean a day, and left to its own devices, does not turn. Whatever heading it left the storm with is the heading it arrives with, days and thousands of kilometers later.

So the simplest model of surf, the zeroth order cartoon, is a ruled sheet: parallel crest lines sweeping across the map from wherever the storm sat. Most of a forecast lives inside this model. When the chart says 285 degrees at 15 seconds, it is describing the simplest picture. The first number is where the lines come from. The second is their period, the time between one crest and the next.

First order: draw the sight lines

From the lineup, let’s imagine drawing a line straight out to the open ocean and swinging it across the horizon. The arc where that line reaches deep water without running into anything is your swell window.

Everything you ever surf (I know, I know: rivers, cruise ship waves, weird quirks aside) comes through it. Whatever sits inside that arc, an island, a headland, the bend of the coast itself, blocks the swell that would have arrived from behind it. The window is a property of the place, not the weather. It doesn’t change with what the buoy is reporting, and on any timescale that matters to a surfer, it never changes. A swell either comes from inside your arc or, to first order, it does not exist for you.

To a wave, an island is a wall. Swell cannot pass through rock, so a body of land drops a shadow on the water behind it, stretching downswell the way a building blocks the sun, and the Channel Islands are walls that run tens of kilometers end to end. A break parked in one of their shadows misses out on swells from a huge patch of ocean offshore. Not weakened, gone, at least in the straight-line version of the story.

The explorer draws this for ten breaks, using the islands’ real outlines traced from the bathymetry, with every sight line starting at the sand, and the spread is hard to ignore. Oceanside and Lower Trestles tie for the widest window of the ten, 88 degrees each. Long Beach sees 42, and most of what it lost sits in plain sight: Palos Verdes, the breakwater, and Catalina stacked across its horizon. If you have ever wondered why Long Beach is the flattest town on this coast, its window is the answer.

The blocking wall that matters is often the one down the beach. Point Dume covers a wedge of Malibu’s west. Dana Point takes Salt Creek’s south. La Jolla eats swell south of Blacks. None of these headlands is large by island standards, but blocking is about the angle a wall covers, not its size, and a point a few kilometers away covers more horizon than an island forty kilometers out, the same way your thumb at arm’s length covers the moon.

Baja California runs about 1,200 kilometers past the southeast corner of the frame in our tool, but it still closes up to 11 degrees of window here, Malibu most of all. Putting every wall together, no single direction reaches all ten breaks at once. At most, a particular swell hits seven out of ten. At 177 degrees, the best the south can do, the three missing breaks are El Porto, Salt Creek, and Blacks, every one of them shut out by its neighboring headland rather than an island. The wave model W. C. O’Reilly and R. T. Guza built for this bight, the one CDIP still runs, treats the islands as near-solid blocks, because for the directions that matter here that is what they are.

The ten breaks are just samples. Computing the window every two kilometers of shoreline helps us tell the story for the whole coast at once. Reading the figure below left to right is like you are walking the coast from northwest to southeast. Above Point Conception the coast stares at about 180 degrees of open Pacific, the half of the compass not blocked by the state of California. Around the corner into the bight, the window collapses, bottoming out at 41 degrees along the Santa Barbara waterfront, the most sheltered stretch of open coast in Southern California, pinched between Conception wrapping off one shoulder and the four northern islands fencing the other. The water inside the Long Beach breakwater rounds all the way to zero, though that one is by design for ships. From there the window rebuilds slowly, reaching about 107 degrees at Point Loma. Santa Barbara’s summer flat spells are pretty easy to conceptualize with this one figure.

The straight-line model also makes a claim that should bother you. From 300 degrees, the classic winter west-northwest, it reaches zero of the ten breaks. Every one of them hides behind Conception, a headland, or an island from that angle. Yet west-northwest swell is the backbone of a Southern California winter. A model that declares the season’s busiest direction impossible is missing three things. The first is about what actually arrives: a storm is not a point on the map but a weather system hundreds of kilometers wide blowing for days, so the swell it ships is a fan of directions 20 to 40 degrees across, not a single line. The 300 on the chart is the middle of the fan, and the fan’s southern flank reaches down toward whatever slivers the windows leave open. The other two are physics the straight lines ignore: shadows are not hard-edged, and the seafloor steers. Those are the higher-order corrections, and they are the rest of this piece.

The first correction: shadows are soft

Waves do not travel in perfectly straight lines. They bend into the space behind an obstacle, a process called diffraction, and you already know it intuitively from sound: you can hear someone around a corner you cannot see them through. Sound waves are meters long and bend easily around obstacles their own size; light waves are a millionth of that and barely bend at all, which is why you can’t see around those same walls. Ocean swells, with wavelengths in the hundreds of meters, live on the sound side of that divide.

Penney and Price worked out the math for water waves in 1952, borrowing the optics of light slipping past a straight edge, and their results are very useful to us. Right at the edge of the ruler-drawn shadow, a wave stands at about half the height it carried in the open, and it keeps dropping the deeper into the shadow you go. Real seas do a little better than the math, because a real swell arrives as that fan of directions rather than a single line, and parts of the fan slip past the island’s edge at friendlier angles; accounting for the fan and the shadow’s edge pushes it closer to seventy percent of open height. Either way, a shadow is not a clean wedge of dead water. It is darkest straight behind the island and feathers out toward the edges. I did some work on a project looking at the diffraction of internal waves around an island, and I can tell you what was told to me when I got stuck: diffraction is difficult. We’ll leave the math here for this one because it’s gnarly.

How far the swell bleeds into the shadow depends on its wavelength. Deep-water wavelength grows with the square of the period, from about 156 meters at 10 seconds to 625 at 20, and a wave bends more when its wavelength is large compared to whatever is blocking it. An island tens of kilometers across always throws a real shadow, but a 20 second groundswell wraps into a lee that a 10 second windswell runs past. Long-period south fills corners that short-period south never reaches.

The second correction: the seafloor steers

A wave in deep water does not know the bottom exists, because a wave is not just the moving hump you see. Under every crest the water is circling, and the circles shrink with depth until, about half a wavelength down, the motion dies out. If the seafloor sits below that depth, the swell passes over it untouched. If the seafloor pokes up into the circles, the wave feels the drag and slows. This is the basis of refraction.

A 6 second swell, wavelength 56 meters, wakes up to the bottom in about 28 meters of water, 90-odd feet. In the grand scheme of the ocean, this is practically at the beach. An 18 second swell feels bottom in 250 meters, which in this bight means it is already steering as it threads between the islands. And where one end of a crest slows before the other, the whole crest pivots toward the shallow side, the way a shopping cart with a janky wheel pulls towards the jammed side. Long swells begin turning tens of kilometers out. Short swells run straight until the final moments.

To show this in the explorer, I’ve actually coded this process backwards. Wave paths are reversible, the route from deep ocean to the sand is the same route walked in either direction, so the explorer traces rays outward from each break across the real seafloor and keeps every path that reaches open water. This allows you to pick a break and see the whole family at once, every route a swell of that period can take to get in.

Try it: drag the direction and the swell’s route goes bold. Adjust the period and the fan reshapes a bit: some routes detach at long period, bending off grazing approaches before they arrive, and new ones appear where the seafloor wraps swell around a corner. At 18 seconds, Long Beach picks up west swell bent around Palos Verdes from directions no 6 second swell can deliver.

Below, I show the same physics run forward, one swell at two periods: every ray pair starts together offshore, and by the time they cross the island shelves the 18 second rays have peeled away from their 6 second twins, hooking into the Santa Barbara Channel while the short-period rays run straight past it.

Putting the whole coastline on one line, the first order picture and the second order corrections snap together. Let’s trace a realistic 18 second swell from the SSW, a fan spread across 190 to 200 degrees rather than a single bearing, and log the energy the rays deliver along the shore, scored against an open-coast baseline: 1.0 is what the same swell would hand a straight beach with nothing in the way. Where the windows are open, from Trestles through Oceanside, the fan lands at essentially full strength. Where the islands stand in the way, the model delivers almost nothing: the Santa Barbara and Rincon stretch reads near zero, and Catalina drops a second hole over Santa Monica Bay. Those zeros are the point. The real ocean is never that dark, and the reason is the leak from the previous section: diffraction backfills the shadows the ray model leaves empty. The gap between this curve and what the buoys record on a real south swell is diffraction.

The winter mystery resolves the same way, piece by piece. The fan does the first part of the work: a west-northwest storm’s southern flank reaches the slivers the windows leave open near 263 to 285 degrees. Diffraction picks up the second part, feathering energy past the island edges into the lees. Refraction does the last, turning what survives toward the beach as it crosses the shelf. None of these paths exist in the straight-line picture but are important to consider for realistic modeling.

As fun as this project is, I need to share a word on what the explorer is not. It is geometry, not a forecast: it does not tell you what will be working today, tomorrow, or any specific time. The ten breaks are measured from the sand while the coast-wide curve is measured from a standard three kilometers out, the traced ray paths start just offshore so the last stretch to the beach is not modeled, diffraction is not drawn (the windows are not as stark as they seem here), and the energy numbers are relative to an open-coast baseline, not wave heights. We are also completely ignoring currents, which do affect waves. When it disagrees with the buoy, trust the buoy: that is the current wave state. Note: I am not saying the models derived from the buoys, I am saying the buoys themselves. The windows, though, are just geography, and geography does not update: explore your break’s window →

To end this on a concrete note, the sharpest version of all of this is two south swells in the same stretch of summer. One born off New Zealand and one born off South America both showing up out of roughly the south, but the few degrees between their approach angles decide your surf. One threads the gaps and lights up the western channel while the other lands almost wholly in the islands’ shadow, shut out from the west end before it reaches the coast. What gets through is set by where the swell was born.

This is the first post of our First Order series. Since tools will accompany each of these, they will be sporadic and only written when complete. All tools are free and open-source with the goal of providing resources for the curious surfer. I hope you have explored and enjoyed the tools. Check it against your favorite forecasting platform next time and let me know how it does.

Further Reading:

Penney and Price paper

O’Reilly and Guza paper

Buoy observations

Bathymetry data

Share

The arsenal itself has warmed the ocean by nothing you could measure, which is the good news. The heat came from the sun. In the instrumental record, 2024 was the warmest the global sea surface has ever been.

In 2025, La Niña spread across the Pacific and pulled the surface back down, about a tenth of a degree below 2024, taking it to the third-warmest year instead of first. Underneath, the heat content of the upper two kilometers kept climbing to a new record.

In February 2026, NOAA changed how it defines El Niño. The old index compared a patch of the tropical Pacific to a fixed thirty-year average, but the whole tropical ocean has warmed so far that the fixed average kept flagging ordinary heat as an event. The new one subtracts the background warming first and measures only what stands above it. When you have to redefine the baseline to keep a word meaningful, that tells you something about the baseline.

Warm water is bigger water, and a warmer ocean is a taller one. In 2024 the global sea surface rose 0.59 centimeters against an expected 0.43, and for once the reason was mostly expansion rather than melt. In a normal year about two-thirds of the rise is meltwater off the ice sheets; in 2024 about two-thirds was the ocean swelling as it warmed and became less dense.

Interestingly, 2025 sea level barely budged, a mere 0.08 centimeters. The same La Niña that cooled the surface dumped its rain over the Amazon and a few other basins, and for a few months the land held onto water that would otherwise have been in the sea.

Cumulatively over the two years, the ocean grew by about 2,400 cubic kilometers. Poured over California, that is a layer almost six meters deep, coast to state line.

Down at the bottom of the planet, the ocean did the opposite of what was generally forecasted. The projections all had the Southern Ocean surface freshening as ice melted into it. Instead, since about 2015, the surface south of 50 degrees has been getting saltier. Alessandro Silvano and colleagues laid this out in PNAS this year, pairing satellite salinity with float measurements, and over that same stretch the Southern Ocean has shed sea ice on the order of the area of Greenland.

A hole in the ice near Maud Rise reopened around the same time, the first time it has done that since the 1970s, which is tied to the same saltier, less layered surface. This is the part that got mangled in the press as an ocean current reversing, essentially the plot of the movie “The Day After Tomorrow.” No current reversed. A decades-long freshening trend reversed, which is a different and much less impactful sentence. The saltier surface mixes more easily, letting heat from below reach the ice from underneath.

The Antarctic Circumpolar Current runs about 173 Sverdrups through Drake Passage, a Sverdrup being a million cubic meters a second, which works out to something like 10 million cubic kilometers of water shoved past a single point every two years. Putting the two numbers side by side: the entire two-year rise in global sea level, all 2,400 cubic kilometers of it, would slide through Drake Passage in under four hours.

The coolest thing the ocean did in these two years, in my opinion, it did on December 21, 2024. A storm in the northwest Pacific threw up a sea state so large that the SWOT satellite, passing overhead, measured a significant wave height of 19.7 meters. That is the largest wave height ever measured from orbit in 34 years of satellite altimetry, past the previous record set in 2011. Fabrice Ardhuin and colleagues reported it in PNAS, and from those numbers the biggest individual waves in the storm were likely over 35 meters, taller than a ten-story building.

The storm didn’t have a name, because storms in that corner of the Pacific don’t get them, so the scientists named it themselves. They called it Eddie, because the swell it made ran the Eddie Aikau at Waimea Bay the next day. The day after that, the same swell lit up Mavericks for what people are still calling one of the best days that wave has ever produced.

SWOT and the modelers tracked it radiating out for 24,000 kilometers, running south down the eastern Pacific, slipping through the gap at Drake Passage, and curling up into the tropical Atlantic more than two weeks after the storm that made it had already died. This wave jumped oceans.

On June 11, 2026, NOAA named a new El Niño and gave it better than even odds of turning strong by fall. El Niño winters tend to drag the North Pacific storm track south toward California, which usually means more swell, and more sand pulled off the beaches by spring. This one might do all of that. With all of the surprises these past years, who knows what’s in store for us in the next two.

Thank you for two years.

Further Reading:

Sea ice paper

Large wave paper

Share

Sunlight arrives at Earth in parallel lines. Near the equator it lands close to head-on, and a beam of a given width warms a patch of ocean about that same width. Toward the poles, an identical beam comes in at a slant, spread thin across a much larger patch. Same energy and more area, so less heat per square meter of water. Averaged over a full year, the equator takes in roughly two and a half times the sunlight the poles do at the top of the atmosphere, and the gap at the sea surface is wider still, because slanted light also has to punch through more atmosphere to get there and tends to glance off the water rather than soak in.

To complicate the explanation, the atmosphere and the ocean spend every hour of every day hauling heat out of the tropics and dumping it toward the poles. Warm air climbs, storms carry moisture and heat north and south, currents drag tropical water up the western edges of ocean basins, and eddies peel off those currents and carry warm water further still. Near 35 degrees latitude, where the transport peaks, the combined system moves something like five and a half petawatts poleward, a frankly unfathomable amount. Every power plant, engine, and furnace humans operate, added together, runs a couple hundred times short of that. If this were the only fact you knew about the ocean, you’d expect the whole thing to have evened out a long time ago.

All of that circulation makes the poles warmer than they would otherwise be, not colder. Shutting it off would cook the tropics while the poles froze harder than they already do. The temperature difference between the water off Baja and the water off Alaska is a push and pull of the heat imbalance and the restoring currents. The system never finishes, because the sun refills the imbalance every morning. Weather and currents moderate the temperature difference between poles and equator.

This is the general trend of the system, but there are exceptions around the world. Northern Europe is mild for how far north it sits because warm water is delivered to its coastal waters. California is cold for how far south it sits because northerly winds push surface water offshore and cold water rises from below to take its place, a process called upwelling. There is a tongue of genuinely cold water sitting directly on the equator in the eastern Pacific, which sounds like a mistake. None of it changes the trend. Average around the globe at a fixed latitude and the temperature lands where it always has (global warming waters aside). The same holds over time. Weather moves the water temperature day to day, eddies week to week, and El Niño year to year, but averaging enough of them together and the leaves the pattern unchanged.

Looking at the ocean as 3D helps close the story. A thousand meters below the surface anywhere in the open ocean, the water sits within a few degrees of freezing, including directly beneath the equator, underneath the warmest surface water there is. That deep water was made at high latitudes. It got cold, got heavy, sank, and crept out along the bottom of every basin until it filled them. The warm water we surf in is a thin skin stretched across the top of a fundamentally cold ocean, and the poles are what put the cold there.

So the water gets colder toward the poles because the planet is round and the sun is far away, and no amount of churning changes the angle at which the light shows up. The currents, the storms, the eddies, the heatwaves, all of it is the ocean mitigating this imbalance.

Further reading:

Solar radiation

Share

Broken-wave water is full of air. Void fraction, the share of a volume that’s air rather than water, can reach roughly 65% in the collapsing cavity just after the lip lands. It holds above 20% through the mixed region below, and it lingers, because bubbles take time to rise up and out. For a stretch of seconds, you are immersed in a seawater-foam mixture.

This mixed fluid has a different density than straight seawater, significantly lower. In still water, a wetsuited surfer with full lungs runs a little less dense than seawater, which is the reason we float. But if the density of the surrounding water drops by a fifth, the margin can shrink or even flip. Water/foam beats surfer.

But less float doesn’t really explain the feeling of being stuck near the bottom, waiting for permission to leave. A hard plunging wave does drive some air down with it, enough to dent your buoyancy at depth, but the foam thins the deeper it goes and can’t account for all of it. So what is actually holding you down there? In shallow water, where the column is thin enough for the wave to reach the bottom, it comes down to where the wave dumps its energy.

When a wave pitches out and lands, it throws a jet of water down into the column ahead of it. In a strong plunging wave, the vortices that the jet spins up penetrate nearly the whole depth as soon as the wave breaks. The vertical velocities inside these bursts run larger than the wave’s own orbital motion, an order of magnitude above the turbulence ordinary bottom friction makes, somewhere around 1 m/s in surf-zone measurements. This speed is in the same range as the speed your buoyancy lifts you, so it is possible that it actively pushes you down.

This is likely why two waves of the same height treat you so differently. A wave that stands up and throws its lip forward, the way it does over a shallow reef or a slab, drives that jet deep and stirs the whole water column at once. A wave that slumps and slides down its own face on a gentle sandbar spreads its turbulence slowly and weakly, and leaves the water near the bottom mostly alone. Face height is the factor that many would assume to be the biggest driver, but whether you get parked on the bottom also depends on how the wave breaks, which is mostly a question of what the seafloor looks like where it does. 3ft shore breaks have hit me way harder than 8ft crumblers.

In summary, the answer is two-fold. Waves plunge water and air down into the water column, both actively pushing you down and taking away your buoyancy at the same time. Just remember to count in your head, it’s never as long as you think down there.

Further Reading:

Void fraction

Velocity paper

Share

A fin moves through water at a Reynolds number of a few hundred thousand to half a million. In most basic terms, the Reynolds number is a number that allows us to predict how a fluid will flow at certain speeds, viscosities, densities, etc. That number fixes the thickness of the boundary layer, the thin film of water that clings to the fin and gets hauled along with it. For fins on a surfboard, that layer is about a millimeter thick.

Anything sharper than that millimeter is buried inside the boundary layer. Water flowing past cannot tell a honed edge from one rounded off with a few passes of sandpaper, because both sit beneath that film, down in the slow viscous layer the free stream never directly touches. So you can polish the edge to a razor and the water keeps seeing the same blunt millimeter of sluggish fluid wrapped around it.

There is one real catch, and it sits at the trailing edge. A thick, blunt back end drags, so thinning that edge genuinely helps. But the gain is about thickness, not sharpness. Once the trailing edge is down to roughly a millimeter, it is as fast as it will ever be. Grinding it past that into a blade does nothing the water can feel, and only leaves you an edge that chips on the first rock.

On the other side of the fin, a sharp nose makes the water whip around a tight corner, and the flow peels off the fin too soon. The fin loses its grip and the tail slides out. Rounding the nose off causes the water to bend around it smoothly, so the fin holds longer and lets go gently when it finally does, instead of all at once.

The rails are no different from the trailing edge. They sit under the same millimeter of clinging water, so sharp or rounded, the flow cannot tell. The tip is the only place a point even looks useful, because that is where water spills around the end and spins off into a little vortex. But how strong that vortex gets depends on the size of the tip and the angle it sweeps back at, not on whether the last few millimeters come to a needle.

So, let’s stop making knife blades on the bottom of boards, shall we? There is little benefit and boy can it hurt you or others.

Further Reading:

Reynolds number

Share

A big set rolls through, and for about a minute after it passes, the surface where it went is glassy. The wind chop is just gone. Then the wind rebuilds it and we’re back to the chop. The wind never let up the entire time, so it isn’t that causing it. It really does look like the big wave annihilates it.

My first thought is surface tension. Think of pulling a wrinkle out of a sheet by stretching both ends. But, it just can’t be the answer. Surface tension only governs the smallest ripples, the capillary waves shorter than about 1.7 centimeters, where the water really does behave like it’s straining against itself. The chop getting wiped out is bigger than that.

Next thought, Doppler shifts. A big set shoves a mass of water toward the beach. A current running with the chop will stretch the wind waves out longer and lower, which is a real effect. The problem is two-fold. The surge, and the backwash that follows it, relax in seconds, ten or twenty at most, while the glass hangs around for a minute or two. And stretching the chop doesn’t erase it. You’d be left with longer, mellower waves, not the flat water we actually see.

Now, we aren’t totally in the dark, because we know how to make glassy water in other situations. Convergence zones pile up the thin organic film on the surface and damp out the ripples, the same effect that draws those slick lines off internal waves. But it doesn’t give an obvious reason for a one-minute patch that shows up the instant a set breaks. There’s nothing about a passing set that should concentrate and hold a film against a wind that’s actively working the surface.

As of now, my thinking is focused on bubbles. When the set breaks, it drives a plume of air down under the surface and leaves a raft of foam behind. Bubbles near the surface can damp short waves, and this is the guess whose timescale actually fits, since a bubble raft lingers for tens of seconds. These trapped air bubbles attenuate small surface waves by dissipating their energy, knocking down the wind chop.

There’s even precedent for this idea, just in a different context. One proposed explanation for how rain calms the sea is exactly this, the drops forcing air bubbles under the surface to soak up the chop. But, nobody has measured it in the surf, and if bubbles were the whole story the glass should sit only on the bubbly water, which isn’t always where you see it. As far as I can tell at least.

Each of these potential ideas has a hole in it, and it seems to be an open question. So keep an eye out for this smooth lull and let me know if you have any ideas as to what is causing it.

Thank you all for reading 100 of these. I’m excited to continue on with more surf science and welcome any questions you’d like answered.

Further Reading:

Rain paper

Share

Emerita analoga, aka sand crabs or sand fleas, live buried in the swash zone, dug in tail-first and facing the water, with only its eyes and a pair of antennae poking above the surface. Each time a wave drains back over it, the crab uncoils a second set of feathery antennae and combs the backwash for plankton. They migrate up and down the beach with the tide to stay in that band, and on the right fine-sand beach they pack in by the thousands per square meter, so one bed can stipple a whole patch of sand with V’s.

Backwash thins and speeds up as it runs down the slope, and by the time it’s a sheet a few millimeters deep it’s moving fast enough that anything sticking up throws a standing wake, the same way a rock in a quick creek kicks up a fixed little peak. It holds in place because the water is draining downhill faster than the wake can climb back up against it, so the disturbance gets pinned right where the obstacle is. The flow scours a shallow horseshoe around the crab and trails that wake behind it, which is why the point of the V sits at the animal and the arms open seaward. Finding bait is pretty easy when they’re around.

But most of the V’s on the sand aren’t crabs. Walk ten feet up the face to where the sand is tiled edge to edge with the same diamond pattern and start checking the points and you’ll find nothing there. No crab, no shell, no pebble, no bump at the apex of any of them. These are rhomboid marks, and they don’t need an animal, or an obstacle of any kind. The water makes them on its own.

As the thin sheet of backwash drains down the slope, it organizes itself into a repeating lattice, the way fast shallow water patterns up into a regular grid instead of running smooth. The draining flow sets up standing waves, with the angle of the diamonds tied to how fast the sheet moves against the slope. The whole field freezes the direction of the last sheet to leave the beach.

In order to figure out what caused the shape, check the apex. A V with a little lump at its landward point, especially a few of them clustered together, is something solid sitting in the flow, usually a crab, sometimes a shell or the tube of a buried worm. A clean grid of V’s across open sand with nothing at any of the points is just the backwash itself.

Further reading:

Sand crabs

Backwash

Share

Turbines pull momentum out of the air, and downwind of the array sits a wake of slower wind that can run for tens of kilometers. Less wind over the water means less energy feeding the waves growing under it. The modeling puts the result at a few percent off the wave height near the farm, fading to almost nothing within twenty kilometers or so. It lands mostly on the local chop being generated right there, not on the long-period groundswell that was generated far away.

A bottom-mounted turbine sits on a single pile maybe ten meters across. Setting that against a swell whose crests run two or three hundred meters apart, the pile is a pencil standing in a doorway to the shelf. It scatters a sliver of energy, nudges the direction a touch, and the effect on wave height comes in under a few percent.

A few hundred piles in tidy rows looks like it should reflect waves back out to sea, the way evenly spaced bars on the seabed bounce a swell (Bragg reflection). The usual dismissal is that the spacing is wrong, that piles a kilometer apart would need a two-kilometer wave to resonate and surf is far shorter than that. The dismissal itself is wrong, however. The reflection also happens at higher multiples of the spacing, so a coarse grid does not fail to resonate, it just resonates with the swell at a higher order.

The piles block maybe a percent of each row, the rows are never perfectly spaced, and real swell shows up smeared across a band of periods and angles. Sharp reflection needs a dense, precise lattice and one clean wavelength, and a wind farm is none of those. A little energy comes back at a few specific frequencies, but most slips through.

So for the farms standing on the seabed today, the swell crosses the array and arrives about the way it left. If anything the water is slightly cleaner, since the wind the farm skims off would otherwise be stacking up chop.

Floating turbines are a bit of a different story. They do not stand on a pole, they ride on a platform of fat, half-submerged columns, and a column that wide compared to the wave does not diffract neatly. Idealized modeling has a single floating platform cutting the wave height behind it by more than half under the right conditions, with a shadow trailing about a kilometer downwave.

Whether a shadow a kilometer long still matters several kilometers down the line is genuinely unsettled, tangled up in how fast the wave field heals around the gap. The waves we ride should mostly survive the steel bolted to the seabed. The ones riding on top are worth watching.

This is an article based on a comment from Dr. Lonneke Goddijn-Murphy of SeaLetters on a previous article. If anyone has a question they want answered, I’d love to hear it.

Further reading:

Wind farms

Wind farms affecting surface waves

Floating wind farms

Share

At Scripps, my research focused on internal waves on the inner shelf, the region of the coastal ocean between roughly 10-100m depth and home to lots of physical and biological processes. I studied two specific types of internal waves: superinertial coastal trapped waves (CTWs) and shoaling non-linear internal waves (NLIWs).

The first chapter is a rewrite of my first peer-reviewed paper, where I describe how these CTWs drive stronger currents than the tides do. The paper answers a question going back nearly 60 years in this region: why is the sea surface elevation at the semidiurnal frequency not correlated with the currents on the inner shelf?

To follow up this paper, we ran a nonlinear model that showed these waves are coherent over long distances, as in 100s of kilometers. We also reanalyzed the data to show that the inner shelf currents can actually be predictable, finding that more than half of the current 90 days out is explainable by this CTW signal. This paper is currently in review and one that I feel very proud of.

And to close out the dissertation, the final paper pivoted to detection of the NLIWs on a fiber optic instrument. Using machine learning algorithms from self-driving cars, I built an algorithm that automatically detects these waves on the instrument in real time and used it to show how these shoaling waves lead to a significant portion of the mixing on the inner shelf. This one will be submitted soon as I finish up a postdoc at Scripps.

All together, these three chapters add a bit of understanding to the inner shelf dynamics of the Southern California Bight, with applications to other regions where similar signatures have been observed: Tasmania, Hawaii, etc. for the CTWs and the Oregon coast, offshore of Japan, etc. for the NLIWs.

It has been a process of ups and downs, wave jokes a necessary evil in the field. This is one of the ups and it is nice to share some of the work that I have done with you; I hope you enjoyed this one.

Further Reading:

When my dissertation is published, I will send it to anyone interested.

Share

The gray was never really about cold water by itself. It requires cold water at the surface and a band of warmer air parked on top of it like a lid. The bigger the temperature jump between the two, the stronger the lid, and the better it traps the damp air underneath that cools and condenses into our morning cloud. The cold water matters mostly because it widens the gap, which is why our coast, fed by chilly northern water, gets so much more of this than the warm-watered East Coast ever does.

Warming the surface closes the temperature gap some. A smaller jump is a weaker lid, and a weaker lid holds less gray. So yes, warmer water should mean less marine layer. But the honest version of this argument is that the water is not dimming the clouds directly. It is loosening the lid that lets them pool near the coast. Warming waters look like the cloud base creeping higher, until some days it never really commits at all. The sun normally has to chew through a thick gray ceiling but meets a thinner one, and clears it earlier in the day.

You can watch this dial move across a single summer without any heatwave involved. June is the grayest month along the Southern California coast, then the gloom starts surrendering through July. A good chunk of that is the local water warming as the season runs on, and warming faster here than it does up north. Up in Northern California, where the water stays cold longer, the grayest stretch doesn’t show up until July, a full month behind us. The effect happens every year on its own without a heatwave, so a heatwave just turns the same knob harder.

So should the warmest ocean on record hand us a clear June this spring? The 20.6°C headline is mostly a northern, open-water number, with the core of the heat sitting up toward the Gulf of Alaska across something like eight million square kilometers of sea. May Gray answers to a much smaller patch, a thin band of water right off our coast. The California Current still drags cold water down from the north, and upwelling keeps pulling more up from the deep, so the coastal strip can hold its own temperature while the rest of the basin cooks.

It is plausible we get a shorter June Gloom this year, if that warm anomaly works its way down the coast and settles into the water off the Bight. But it leans on a lot of moving parts. Where the warm water actually pools, how hard the coastal winds blow, whether upwelling keeps the nearshore cold, and whether the El Niño now building on top of the heatwave nudges things one way or the other. Whether our particular stretch of coast gets to tan this spring is a toss-up.

Further Reading:

Marine layer

Warmer waters

Share

A 17-second North Pacific groundswell can hold waves at the coast for a week, the dominant period stepping down a second or two each day. A 9-second windswell can show up and be flat within a day or two. The difference between the two has to do with the generation offshore rather than anything at the break itself.

Wind dragging on the ocean surface starts by making short, choppy waves at the top of the spectrum. That wind blowing over a big enough patch of ocean works the surface into longer, bigger waves. A 20-second peak only emerges after days of sustained generation. Shorter storms only output shorter periods.

Back to the big storm: while the long end of the spectrum fills in slowly, the wind is still making 14-second, 12-second, 10-second, and shorter waves at the same time. We talk about “an 18-second swell” because that’s the dominant period at the buoy, but the whole mess is leaving the storm together.

Walter Munk’s 1963 work tracking Southern Ocean swells to Alaska showed how dramatically period controls travel speed in deep water. A 20-second wave moves at about 35 mph, but a 10-second wave moves at just half that. The further the waves travel, the more they spread out in time. Because of this, a storm 4,000 miles offshore sends its 18-second leading edge to the buoy more than a day before the 12-second component arrives.

But when the buoy reads 8 to 10-second primary with no long-period leading edge, we can expect a shorter run of swell. The storm didn’t blow long enough or over enough fetch to develop the longer period swell, so there is a smaller window of time to surf it. Short-period waves also dissipate much faster as they travel. The longer end of the wave spectrum crosses a basin without much energy loss, but the short stuff burns itself out within a few hundred miles. Most short-period swells at the coast are generated locally in the past day or two, not days to weeks before like the long period.

The one exception is a productive storm that sits close to the coast. North Pacific lows in winter sometimes track right past California, throwing 15 to 17-second swell from a few hundred miles offshore. The storm still had to last long enough to make those periods, so the spectrum is still there. But without much ocean distance to do the dispersion sorting, the longer and shorter components arrive bunched together. The event fires for a day or two and disappears. There are plenty of waves, just no slow tapering tail. Gotta get ‘em while they’re hot.

Further Reading:

Dispersion

Spectrum

Share

Behind your face are four pairs of air-filled cavities called the paranasal sinuses: maxillary under each cheekbone, frontal above the eyebrows, ethmoid between the eyes, and sphenoid tucked deep behind the nose. Each is lined with mucus and tiny hairs called cilia that constantly sweep debris toward the throat. They connect to the nasal cavity through small openings called ostia, typically 1 to 4 mm wide, about the width of a wooden pencil tip.

The maxillary sinus is the biggest of the four, roughly walnut-sized, and its drainage ostium sits not at the bottom of the cavity but near the top of the medial wall. The official medical phrasing is “poorly positioned for free drainage.” The frontal sinus is the other troublemaker, with what the literature calls the most complex and variable drainage of any paranasal sinus, routing everything through a twisting frontonasal duct before it can reach the nose. Basically, the drains are on the ceiling for both of these.

Physics is a pesky part of this problem as well. Water in a tube only a few millimeters across forms a concave meniscus, the molecules at the surface clinging to the walls and to each other strongly enough to support the column’s weight. The effect is called capillary action, and it’s the same mechanism that lets paper towels wick up spilled coffee against gravity. In your sinus ostia, it creates an air lock, preventing gravity from clearing the last close out.

At roughly 3.5% salinity, seawater is nearly four times saltier than the fluid in your tissues, hypertonic enough to pull water out of mucosal cells and stress the lining on contact. Cold makes this even worse, triggering vasodilation and a rush of mucus in the sinuses. The lining swells, narrowing the ostia further, sometimes closing them completely. The already small exit gets clamped down, and the water sits awaiting an awkward moment to emerge.

Eventually something tips the balance, often the first or second inversion of the day. Maybe you bend over to towel off your feet and the head angle finally beats the meniscus. Maybe the swelling subsides after an hour or two. Or maybe a sneeze creates enough pressure to break the seal on an unsuspecting guest. Around 35% of swimmers report nasal congestion after a session, and the timing of release is mostly unpredictable. The phenomenon is recognized enough that ENTs have a name for the chronic version, “swimmer’s sinusitis.”

Our sinuses evolved for breathing air, warming and humidifying it on the way to the lungs, with the mucus-and-cilia system handling the occasional bit of dust. Repeated head-first dunks in cold salt water came along later, and the design hasn’t caught up.

Further Reading:

Sinuses

Meniscus

Share

California Department of Fish and Wildlife confirmed in late March that starvation, not avian flu, is the primary cause. Brown pelicans and common murres are dying alongside the cormorants, almost all of them young birds in their first year. 2025 was an unusually strong breeding year for these species, which means a larger-than-usual cohort of juveniles is now competing for whatever food is available.

On September 9, 2025, the northeast Pacific reached the highest average temperature ever recorded in the basin: 20.6°C, nearly half a degree above any previous record. The marine heatwave responsible, designated NEP25A by NOAA, first emerged in the Gulf of Alaska that May and expanded through the summer to roughly 8 million square kilometers (about half a Russia), larger by area than any prior northeast Pacific heatwave on record. It shrank through October and November, then swelled again along the coast in December. As of early 2026 it persists, with an El Niño developing on top of it.

Warm surface water pushes anchovy and sardine schools deeper into cooler layers, away from some seabirds’ reach. A brown pelican can only dive about six feet, so a small change is significant. Common murres can dive much further, hundreds of feet, but each dive is intensive and burns lots of calories. A juvenile that hasn’t dialed in its hunting yet runs its energy budget into the red trying.

Warm surface water also resists mixing with the colder water below it. That matters because the deep water carries the nutrients (nitrate, phosphate, silicate) that phytoplankton need to grow. Less mixing means less fertilizer reaching the sunlight, and phytoplankton, the base of the food web, runs short. The community shifts from large, fat-rich diatoms toward smaller dinoflagellates. Their predators, zooplankton, follow: warm-water copepods are smaller and carry less fat than the subarctic species they replace, so a fish eating the same number of copepods gets fewer calories. Krill hit an 18-year low in southern California during the 2013-2016 Blob. Pyrosomes, gelatinous tube-shaped colonies that are essentially low-calorie filler, proliferate and absorb energy that would otherwise have moved up the chain.

Each link in the chain loses calories. A fish eating smaller, leaner copepods gets less energy from a meal, and a cormorant eating that fish gets less still. The bird has to catch more fish to compensate, but the fish are deeper than they used to be.

The 2013-2016 Blob, the previous benchmark for this kind of event, killed roughly a million common murres and 400,000 Cassin’s auklets along the same coastline. NEP25A is bigger by area and now in its second year. The juveniles washing up this spring are the cohort that would have started breeding around 2029.

Further Reading:

Marine heatwave tracker

Effect of heatwave on phytoplankton

Share