The Blogcast

Forecaster’s Note · Friday, September 4, 2026

The crew page is showing 100% probability of precipitation this morning and that is the least useful number on it, because the drizzle is already on the windshield. A probability forecast stops being a forecast the moment the thing happens. The number that actually matters today is 0.06 inches.

Yesterday I told you six of seven models loaded the volume onto Friday. They did, and Friday turned out to be worth six hundredths of an inch. That is the whole event. Read that as good news, sequence around a wet morning, and put your attention on the two things this day is genuinely carrying: a fourth consecutive Lake Wind Advisory, and the fact that this is the coldest day of the sixteen-day run before the humidity starts falling out from under us on Monday.

Right Now, and the Station Finally Agrees With the Grid

KTRK at 5:35 this morning: 48.2°F, dewpoint 42.8°F, relative humidity 81%, wind 8 mph out of the southwest, light drizzle. Twenty minutes earlier it was 50.0°F with a 22 mph gust. The grid at the same hour has the basin at 47.6°F and 83%.

Six tenths of a degree and two points of humidity apart. After three mornings of writing about how far the station and the model were diverging, they have converged, and the reason is the same reason they were splitting before: the column is saturated and mixed. Rain falling through a well-stirred boundary layer forces the air toward its wet-bulb temperature, and a wet-bulb is a number the model and the thermometer are both good at. When the basin decouples — clear sky, dead calm, cold air pooling in the Martis flats — the two disagree by a lot. When it is raining and blowing, they agree. Model-versus-station agreement is not a measure of how good the model is. It is a measure of how boring the atmosphere is.

Surface pressure is 818 hPa and the sky went to overcast between four and seven. This is a shortwave dragging a shower band along and just north of the I-80 corridor, exactly where Reno put it at 2:42 this morning.

Today: GO, and the Wet Half Is the First Half

High 57°F, low 43°F, minimum humidity 61%, UV index 3.0, sunrise 6:32, sunset 7:26. All systems GO for outdoor work. Nothing here stops a crew. Bring rain gear and plan the sequence.

The shape of the day is the inverse of yesterday’s. Yesterday the wet half was the back half; today it is the front. Hourly probability runs 21% at six, 20% at nine, 23% at ten, then falls off a cliff — 10% at noon, 8% at one, 4% at two, 3% by three. Cloud cover holds at 100% until two, breaks to 57% at three, 41% at four, 25% by six, and the basin is effectively clear by eight tonight. Reno’s aviation section has MVFR at KTRK through the morning with shower activity tapering after 15Z — eight o’clock our time — and mountain obscuration on the crest most of the day.

The instruction is the mirror image of yesterday’s: back-load the day. Anything needing dry ground or a clean subgrade waits until after one. The morning is for the work that does not care — pipe in the trench, shoring, hauling, layout, anything under cover. Six hundredths of an inch on a Sierra jobsite is a nuisance and a slick ladder, not a shutdown. Fifty-seven degrees with an 81% humidity and a southwest wind on you is the kind of cold that gets into people by ten o’clock. Dry socks in the truck.

No flatwork pour today. No final grade on fine-grained material until the afternoon at the earliest, and probably not then — minimum humidity only gets to 61%, which means the surface never really lets go of what it took on overnight. Compaction will fight you on anything silty. That is the honest cost of today, and it is a scheduling cost, not a safety one.

The Wind: My Grid Is Low and I Am Not Defending It

This is the fourth straight day of a posted Lake Wind Advisory — CAZ071 through CAZ073, until 9 PM tonight, covering Donner, Stampede, Prosser and Boca. Same product, four days running. If you have a crew on or near water today, that is your governing document.

Here is where I disagree with my own page. My grid has today’s gusts topping out around 26 mph through the working afternoon with a 31 mph spike at daybreak. Reno has southwest to west gusts increasing to 30 to 40 mph this afternoon, with ridge gusts 45 to 65 mph all day. The eight-model ensemble behind this page splits the difference at a 34 mph seventy-fifth percentile.

Take the office number, not mine. This is a known and repeatable failure: a gridded model at this resolution cannot resolve what terrain does to a southwest flow across the Sierra crest and down into this basin. It smooths the ridgeline, so it smooths the momentum transport, so it under-forecasts the gust. Reno’s forecasters are looking at 30 to 35 knot winds at ten thousand feet and mountain-wave turbulence, and they are pricing in the mixing my grid is quietly ignoring. Plan for 40 mph in the open this afternoon and be pleasantly surprised.

What that means: second strap on tarps, silt fence and anything light and stacked, done this morning, not this afternoon. Sheet goods stay flat and weighted. A boom pick this afternoon needs a stand-down number written down in the yard before the first lift, and 40 mph is a defensible place to put it. The one mercy is that the wind dies hard tonight — gusts collapse from 21 mph at seven to 6 mph by nine.

Model agreement on the temperature is excellent by contrast: seven models put today’s high between 55 and 58°F. Three degrees. Trust the temperature. Argue about the gust, and when you argue, side with Reno.

Tonight: The First Clear Calm Night in a Week

Worth flagging on its own, because it is the setup this basin is famous for. Sky clears by eight, wind collapses by nine, and the Martis flats get their first real chance to decouple since last weekend. Published low is 43°F and the model spread runs 41 to 49°F.

On a night like this the low-lying ground — open trench bottoms, the flats, anywhere air can drain and sit — runs several degrees under whatever the basin average says. This is not a frost call. Forty-three is nowhere near it and I am not going to manufacture a scare out of a September night. But if you are staging anything temperature-sensitive, or you have a water line exposed in a trench, the bottom of that trench is the coldest place on the job at five in the morning and it always has been.

The Weekend: Saturday Is the Clean One, Sunday Is the Wildcard

Saturday the 5th is the day. High 66°F, low 41°F, minimum humidity 38%, gusts down to 21 mph, probability of anything falling out of the sky in the low teens. Reno’s own line is “drier and less windy weather prevails on Saturday.” No product posted, no wind worth planning around, dry ground by mid-morning. If you have a Saturday crew or a punch list, that is your window and it is the only clean day between here and Monday.

Sunday the 6th is where it gets interesting. High 64°F, minimum humidity 32% — and the wind comes back. My grid has gusts to 28 mph; the ensemble seventy-fifth percentile is 38 mph, the highest number anywhere in the ten-day run. Reno is explicit: gusty winds and shower and thunderstorm chances both return Sunday afternoon, renewing wind impacts on holiday travel.

On the rain, the office and I agree and the agreement is the useful part: 30 to 50% from Lassen County to the Oregon border, 20% or less down here. My ensemble has Truckee at 11%. Reno adds the caveat that matters — those numbers may go up or down depending on the depth and track of a compact upper cyclone that has not arrived yet. Circle Sunday afternoon for wind, not for water. And if you are working north of the basin, treat the thunderstorm mention as real; a September cell over dry fuel is a different conversation than a September cell over Truckee.

Reno also frames the whole holiday weekend as running 10 to 20 degrees below seasonal averages, highs in the 60s and 70s. That is a genuinely pleasant three days to be outside in and it is the last of them for a while.

Fire Weather: Nothing Until Monday. Then Circle Monday and Tuesday.

Today has 61% minimum humidity and measurable rain on the fuels. There is no fire weather conversation to have. Saturday at 38% and Sunday at 32% are both comfortably above my attention line.

Then it turns, and it turns on the holiday. Minimum relative humidity, Truckee, day by day: Monday 18%, Tuesday 16%, Wednesday 35%. Highs of 72, 72 and 68°F. Below 20% gets mentioned here. Below 15% is critical. Truckee does not cross into critical in this run — it gets close on Tuesday and backs off.

Reno on the same three days runs 19%, 8% and 25%. That 8% on Tuesday the 8th is the number in this entire forecast that would make me stop and read it twice, and it is on the Nevada side with an 82°F high behind it. The valleys east of us will very likely draw a fire product for Tuesday. We probably will not. Do not let the second half of that sentence erase the first — plenty of Ruppert work happens on the other side of the state line.

The saving grace on Monday and Tuesday is the same one I wrote about last week: the wind is not there. Gusts run 15 mph Monday and 23 mph Tuesday on my grid, with ensemble seventy-fifth percentiles of 23 and 29. That is a real distinction and it drives the whole assessment. Dry air without wind is an ignition problem, not a spread problem. The risk sits on the equipment side, which is the side you control completely: spark arrestors checked before the holiday, no hot exhaust parked against cured grass, grinding and cutting and chain drag moved to first light when the humidity is still in the fifties.

Six hundredths of an inch buys the fine fuels roughly a day and a half, not a week. By Monday afternoon that moisture is gone and the fuels are back where they were Wednesday. The rain does not save Monday. It barely reaches Saturday.

Further out there is a genuinely ugly pair — the 13th and 14th, with minimum humidity at 13% and 8% and highs near 82°F, and Reno at 7% and 91°F on the 14th. I am flagging those two days and doing nothing else with them, for a reason I will explain in the next section.

The Fight Over Next Week

There is a straight disagreement in the guidance right now and you should know about it, because it decides whether the back half of next week is a cool breezy shower pattern or the driest stretch since August.

Reno’s position: eastern Pacific troughing persists through the middle of next week — cooler than normal, breezy afternoons, low shower chances. They add that there is high uncertainty about remnant tropical moisture pushing north, and if it does, shower chances go up mid-week.

My deterministic grid’s position: a ridge builds. 77°F on the 10th, 78°F on the 11th, 81°F on the 13th, 86°F by the 15th, 88°F on the 17th. Warm, dry, done.

Those cannot both be right, and the model spread tells you which one to believe: on September 13th, the eight-model pack ranges from 59°F to 81°F on the high. Twenty-two degrees on one day. On the 10th it is 59 to 78. A twenty-two degree spread is not a forecast. It is a coin toss with a decimal point on it.

When the office is talking about a persistent trough and my single deterministic run is drawing a ridge, the office wins on principle — they are reading the ensemble means and the teleconnections, and I am reading one member that happens to be plotted on my page. Discount the mid-month warm-up until the spread closes. I will keep publishing those numbers because that is what the grid says, and I will keep telling you not to plan on them.

Confidence

How this page has actually been scoring. 82 days verified. Last 30 averaging 84 out of 100. High bias −0.1°F, wind bias +0.1 mph — both effectively unbiased. The GO/NO-GO call is 30 for 30 over the last thirty days. Precipitation is still the weak leg: 86% false alarm rate, 40% miss rate, with 23 correct dry calls, 6 false alarms and 1 miss in the last thirty.

Yesterday scored 52, and that is the worst day on this page in a month. I am putting it at the top of this section rather than burying it, because a verification record you only quote when it flatters you is worth nothing.

Two things went wrong and they went wrong for the same reason. The high verified at 59°F against a 60°F call — one degree, excellent. The low verified at 50°F against a 38°F call. Twelve degrees. And the archived overnight snapshot carried 0% probability of rain for a day that produced 0.02 inches, which the scorer correctly marked a miss.

The twelve-degree low is not a mystery, it is the thing I wrote about in this space yesterday morning and then failed to apply. Thursday night had 89% cloud cover and a wind that never dropped below 5 mph. A lid on top and a mixed column underneath means no radiational cooling and no cold pool, so the basin does not fall to its clear-night number. I described that mechanism correctly and the number underneath it was still 38°F. The verification record has now promoted the KTRK station to primary truth precisely because of this — the gridded analysis products run 2.9 to 7.1°F colder than the station on overnight lows. The correction is real, it is not a constant, and it is largest on exactly the nights the pool cannot form.

The precipitation miss is the more interesting one. The snapshot taken Wednesday evening had Thursday at 0%. By the time I wrote Thursday’s note at dawn, the ensemble had it at 94% and I published that. Thursday delivered two hundredths of an inch. So the machine was wrong at 5 PM, right by 6 AM, and the archive graded the 5 PM version. That is a thirteen-hour swing from zero to ninety-four on a shortwave, and it is a fair description of how much a September shortwave in the Sierra is worth at range.

Today, temperature: locate. Marked, exposed, verified. Seven models inside three degrees, and the rain is already falling, which pins the whole thermal profile.

Today, wind: locate on the day, pothole on the number. A posted advisory, a fourth consecutive day of the same pattern, and a clear physical reason my grid reads low. I am confident it blows. I am not confident whether that is 26 or 40, so plan the 40.

Today’s rain: locate, and it barely matters. It is falling now. Six hundredths of an inch, done by mid-morning.

Saturday: locate. Highest-confidence call on this page. Dry, warming, light wind, no product posted, office and models in full agreement.

Sunday wind: locate. Sunday rain: pothole. Everybody sees the wind. The rain depends on a compact upper cyclone whose depth and track are still open, and Reno says so in as many words.

Monday and Tuesday fire weather: locate on the humidity, locate on the absence of wind. Every model dries the air out on the holiday and none of them puts a wind signal behind it. That combination is unusual and it is the reason this reads as a mentionable rather than a warnable. If a gust signal shows up in this weekend’s runs, those two days change character entirely and I will say so here Sunday morning.

Next week’s pattern: old as-built. A forecast office arguing for a trough, my grid drawing a ridge, and a twenty-two degree model spread underneath the disagreement. Nobody knows. Anyone who tells you they know is quoting one member.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: it is raining now and it is over by mid-morning, so back-load the day and put the dry-ground work after one. Strap the light stuff this morning and plan the afternoon around 40 mph, not the 26 on my own gauge — fourth day of the Lake Wind Advisory, and Reno is right about this one. Dry socks and rain gear, because 57°F at 81% humidity gets into people. Take Saturday, it is the only clean day in the weekend. Watch Sunday afternoon for wind. And put a mark on Monday the 7th and Tuesday the 8th — 18% and 16% humidity on fuels that will have forgotten this rain by then, with Reno at 8% on Tuesday. Dry air, no wind, all of the risk on the equipment side and all of it inside your control.

Forecaster’s Note · Thursday, September 3, 2026

Yesterday I told you Thursday and Friday were both suspect and that the office and the guidance did not agree on which one carried the water. Twenty-four hours later the question has sharpened into something more useful: everybody now agrees it rains. Nobody agrees on which day gets the volume.

Reno’s 1:04 AM discussion is unambiguous: showers creeping into northeast California at daybreak, greatest chances this afternoon and evening, 60 to 90% across the Tahoe Basin, totals 0.01 to 0.25″ with locally over half an inch on the high peaks. The eight-model ensemble behind this page is higher than the office — it puts today at 94% and tomorrow at 90%. On the probability question there is no argument left to have.

The Lake Wind Advisory is still posted, CAZ071 through CAZ073, until 11 PM tonight. Donner, Stampede, Prosser and Boca. Same product, third day.

Right Now, and the Station Is Warmer Than the Grid Again

KTRK at 4:35 this morning: 50.0°F, dewpoint 28.4°F, relative humidity 43%, wind 10 mph out of the south-southwest, cloudy. The grid at the same hour had the basin at 46.3°F and 59%.

The station is 3.7°F warmer and sixteen points drier than the model. That is a bigger gap than the last two mornings, and it goes the opposite direction from the cold-pool story I have been building here all week. Which is exactly what it should do. Cloud cover is 89% and the wind never dropped below 5 mph all night. A lid on top and a mixed column underneath means no radiational cooling and no cold air pooling in the Martis flats — so the basin does not undercut the grid, it overshoots it. The pool needs a clear calm night. It did not get one.

The humidity half is the more useful number. A 28.4°F dewpoint at 43% RH says the surface air over this basin is genuinely dry right now, drier than the grid thinks. Moisture has to arrive before rain falls out of it, and at 4:30 this morning it had not arrived. That does not argue against the rain — the forcing is upstream and on its way. It argues about when, and it is the reason I am telling you the front half of today is workable rather than telling you to expect a wet morning.

Today: GO, and Take the Morning

High 62°F, low 46°F, minimum humidity 38%, UV index 7.3, sunrise 6:31, sunset 7:28. All systems GO for outdoor work. Nothing on today’s board stops a crew, and the first half of the day is clean.

The shape of the day is what matters more than the daily number. The onset curve is flat through the working morning and then climbs hard: near nothing at noon, still low at four, lifting at five, and the crew page has rain arriving after 7 PM. Humidity does the same thing from the other side — 38% at noon, 42% at two, 57% at four, 73% at five, 88% by six. Fifty points of moisture in four hours. That is a moisture return you can feel on your skin before you see it on a gauge.

The instruction is simple: front-load the day. Anything that needs dry ground, a clean subgrade, or a surface that will not slick over gets done before three. Anything that can be finished wet goes to the back half. Reno’s aviation section expects MVFR to IFR at KTRK by noon with mountain obscuration — that is ceilings dropping onto the crest by the middle of the day, and it is the first visible sign the moisture arrived.

Wind runs all day and is the thing to manage. Gusts: 27 by six this morning, 28 at eight, 26 at ten, 28 at noon, 30 through the afternoon, peaking at 31 around five, then easing. Reno has widespread 30 to 40 mph across the region with 50 to 60 mph in the wind-prone US-395 corridors, which is not us. Dust control matters until the rain does it for you. Tarps, silt fence and anything light and stacked wants a second strap this morning, not this afternoon. A boom pick at 30 mph gusts is workable with a written stand-down number decided in the yard.

Model spread on today’s peak gust: MeteoFrance 28, GFS 31, UKMO 33, GEM 34, ECMWF 42, ICON 61 mph. ICON is the high model on gusts for the fourth day running and it has not been right once. It mixes momentum down harder in terrain than anything else in the pack, and in this basin that is a known overshoot. Discount the magnitude. Do not discount the day — a model that is wrong about how hard can still be right about when.

The seven-model cluster on today’s high runs 58 to 62°F. Four degrees across seven models. Trust the temperature. Argue about the gust.

The Rain: High Confidence It Happens, Low Confidence on the Day

Probability and accumulation are two different forecasts, and this is a week where they are telling you two different things. The probability is settled — 94% today, 90% tomorrow on the ensemble, 60 to 90% from Reno. The accumulation is not settled at all.

Here is the split, in inches.

Today: GFS 0.15″, MeteoFrance 0.03″, UKMO 0.02″, ECMWF 0.02″, GEM 0.00″, ICON 0.00″, JMA 0.00″. Tomorrow: ECMWF 0.29″, MeteoFrance 0.22″, GEM 0.18″, UKMO 0.17″, JMA 0.15″, ICON 0.07″, GFS 0.03″.

One model — GFS — puts real water on today. Six of seven load Friday instead. Yesterday that vote was three of four. It got stronger, not weaker, which is usually how a guidance signal behaves when it is right.

Read those two paragraphs together and the picture is coherent rather than contradictory: today is very likely to get wet, and tomorrow is very likely to get wetter. Reno’s own numbers say the same thing — showers today, lingering into Friday morning. High probability of something, most of the volume on the back end.

The one place I lean harder than my own grid is toward more rain, not less, and there is a documented reason. This page’s verification log has rain days under-warned four times out of five. The standing note on that pattern says the safe band is gone — days forecast at single-digit probability with a zero accumulation have produced measurable rain repeatedly since mid-August. When the quantity forecast on this page is wrong about rain, it is usually wrong on the dry side. Treat GEM’s and ICON’s and JMA’s 0.00″ for today as a floor, not a forecast.

On a jobsite the distinction mostly collapses anyway. A wet Thursday evening and a wet Friday morning are the same problem, and both are already inside the window.

What to actually do: rain gear on the truck before you leave the yard, not staged at the shop. No flatwork pour, no hydro-seed, no final grade on fine-grained material today or tomorrow. Friday’s minimum humidity is 51% and showers linger into the morning, so compaction is going to fight you on anything silty even if the gauge stays dry. And the counterweight, stated plainly, because a forecast that only warns is not worth reading: this page’s precipitation false-alarm rate is 86%. When it says rain, it is usually wrong about the amount. A quarter inch is a nuisance and a schedule problem, not a washout — this is a reason to carry gear and sequence the day, not a reason to send anybody home.

The Weekend: Dry, Warmer, and a Sunday Wildcard

Friday the 4th is the transition — high 55°F, low 42°F, showers thinning through the morning, gusts to 31. It is the coolest day of the run and the last wet one for a while.

Then it opens up. Saturday the 5th: 69°F, minimum humidity 31%, gusts down to 22. Sunday the 6th: 68°F, humidity 27%, gusts 22. Reno has winds relaxing to the usual afternoon zephyr, 20 to 25. Those are two clean, workable, warm days with no product on them and no wind worth planning around. If you have a Saturday crew or a punch list, that is the window.

The wildcard is Sunday afternoon. A low currently spinning over the Pacific Northwest sags south and Reno carries light shower and thunderstorm chances for northeast California and northern Washoe. The ensemble barely sees it for Truckee — 19%, no accumulation from any member. Given the under-warn record I just cited, treat a Sunday afternoon storm cell as possible rather than absent, particularly if you are working north of the basin.

Labor Day itself looks dry. Hurricane Marie — the tropical fork I have been tracking since Monday — is now expected to track further out to sea, and the holiday-weekend soaking scenario has largely fallen out of the guidance. Reno flags that if the northern low digs far enough south it could still pull tropical moisture up regardless, but nothing arrives before Tuesday at the earliest. Two days ago that system was the story. Today it is a footnote, and that is worth remembering the next time a named storm shows up eight days out.

Fire Weather: Nothing Today, Circle Monday Through Wednesday

Today is not a fire day. Thirty-eight percent minimum humidity, rain arriving on top of it, and fuels that get their first wetting since the last front. Nothing on the board.

Then it dries out hard. Minimum relative humidity, Truckee, day by day: Saturday 31%, Sunday 27%, Monday 18%, Tuesday 17%, Wednesday 18%. Highs on those last three: 72, 72, 69°F. Reno on the same three days runs 13, 9 and 12% — the Nevada side will almost certainly draw a fire product.

Circle Monday the 7th through Wednesday the 9th. Three consecutive days under twenty percent is the first sustained run below my attention line since the Red Flag Warning expired Tuesday. Below 20% gets mentioned. Below 15% is critical, and Truckee does not go there in this stretch — Reno does.

What those three days do not have is wind. Gusts collapse to 11 mph Monday, 23 Tuesday, 16 Wednesday. That distinction is the whole assessment: dry air without wind is an ignition problem, not a spread problem. The risk sits entirely on the equipment side and entirely inside your control. Spark arrestors checked. No exhaust parked against cured grass. Grinding, cutting and chain drag moved to first light, when humidity is still in the fifties and sixties. A quarter inch of rain tonight buys you roughly seventy-two hours on fine fuels — and Monday is exactly seventy-two hours out. The rain does not save Monday. It just delays the clock to it.

Further out, the sixteen-day run has minimum humidity at 11 to 13% from the 12th onward with highs climbing toward 90°F. I am flagging that and nothing more, because the spread out there is not a forecast. On September 14th, GFS has 87°F and ECMWF has 64°F. Twenty-three degrees apart on the same day. Anyone handing you a mid-month heat forecast right now is reading one model and calling it consensus.

Confidence

How this page has actually been scoring. 81 days verified, last 30 averaging 86 out of 100. High bias +0.4°F, low bias +0.8°F, wind bias −0.1 mph — all three effectively unbiased across the record. The GO/NO-GO call is 30 for 30 over the last thirty days. Precipitation remains the weak leg at 86% false alarm, 25% miss.

Yesterday scored 86. The high came in 5°F cooler than published — 61°F against a 66°F call — and the low came in 6°F warmer, 45°F against 39°F. Wind within one, no-rain call correct. Worth naming: I wrote yesterday that the cold air had been arriving earlier and deeper than the models had it and told you to treat 60°F as a ceiling rather than a target. It verified at 61°F. The warm-side low miss is the same standing pattern the learning log has flagged for three weeks — forecast overnight lows on this page run cold, and the correction is to add several degrees before writing any frost or cold-start language. I have not written frost language this week for that reason.

Today, temperature and wind: locate. Marked, exposed, verified. Seven models inside four degrees on the high, a posted advisory with explicit numbers, and every decision identical anywhere in the gust range.

That it rains: locate. Which day carries it: pothole. I know this setup cold — shortwave energy dropping in from the north, moisture into the Basin, showers peaking late in the day, quarter-inch class totals. The ensemble at 94% and 90% on back-to-back days is about as confident as a precipitation call on this page ever gets. What nobody has pinned is where the volume lands: the office weights this afternoon, six of seven models weight tomorrow. Prepare both days. Commit to neither.

The weekend: locate. Dry, warming, light wind, no product. Highest-confidence call on the page.

Sunday afternoon storms: old as-built. One office mentions it, my grid does not see it, and the feature responsible is still four days from arriving. Low ceiling on how much anyone knows.

Monday through Wednesday fire weather: locate on the humidity, pothole on the wind. Every model agrees the ridge builds and the air dries. None shows a wind signal, and wind is the variable that turns a careful day into a dangerous one. If a gust signal appears in this weekend’s runs, those three days change character entirely and I will say so here.

Mid-month heat: old as-built, and barely that. A 23°F model spread on a single day is not a forecast, it is a placeholder.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: work the morning hard and get the dry-ground work done before three. Strap the light stuff at first light while the gusts are still in the twenties. Rain gear on the truck tonight for an evening and a Friday morning that are both live. Take Saturday and Sunday — they are the two clean days in this run. And put a mark on Monday the 7th, because three straight days under twenty percent humidity land on fuels that will have dried out again by the time you get there.

Forecaster’s Note · Wednesday, September 2, 2026

Yesterday I told you the second week of September was a genuine fork — a GFS ridge against an ECMWF tropical remnant — and to watch it Friday. It moved faster than that. Reno’s 1:19 AM discussion names the system: Tropical Storm Marie. The fork is still open, but it is no longer a shape in the model output. It is a storm with a track, and tracks get better fast.

The other thing that changed overnight is that the Red Flag Warning expired and nothing replaced it. There is no fire product on any zone in this forecast area this morning. What is still posted is the Lake Wind Advisory, and it got bigger: extended through 11 PM Thursday and expanded into Mono County, covering CAZ071 through CAZ073. Donner Lake, Stampede, Prosser and Boca are all named. Southwest 15 to 25, gusts to 30, one to three foot waves.

Right Now, and the Grid Nailed It

The airport station read 44.6°F at 4:15 this morning, dewpoint 32°F, relative humidity 61%, wind 7 mph, sky clear. The model grid had the basin at 44.6°F and 57% at the same minute.

Dead on the temperature. Four points on the humidity. That is the second morning running that the station and the grid agree, and it is the second morning running with wind in it — gusts never dropped below 18 mph overnight and touched 30 mph at two in the morning. This is the pattern I described yesterday holding up: the grid and the station converge on windy nights and diverge on still ones, because a mixed atmosphere never lets cold air pool in the Martis flats and the grid does not model that pool. Two data points is not proof. But it is the right two data points, and the mechanism is not complicated.

Note the dewpoint: 32°F. That is the number to carry into the weekend section below, because a dewpoint at freezing is the floor the overnight low can reach if the air ever stops moving.

Today: GO, Cold and Breezy, Take the Whole Day

High 60°F, low 42°F, zero precipitation, minimum humidity 38%, UV index 6.5, sunrise 6:30, sunset 7:30. Reno calls it “fall-like” and they are not wrong — meteorological fall started Monday and this is the coolest first week of September this basin has seen in a while. All systems GO for outdoor work.

Nothing on today’s board is a fire problem. The humidity ladder runs 62% at six, 59% at eight, 43% by ten, and bottoms at 39% around noon before recovering to 49% by six and 70% by ten tonight. Thirty-eight percent is the daily minimum and it is nowhere near a line I care about. Below 20% gets my attention, below 15% is critical, and we do not go there for five more days. Reno at 25% is drier than us and even Reno is not in trouble today.

The elevation gap is doing its usual work: Truckee 60°F, Reno 68°F — eight degrees and thirteen points of humidity bought by 3,000 vertical feet. That gap is why the fire products keep stopping at the crest.

Wind is the only thing to manage. Basin gusts: 23 mph by eight, 29 by ten, peaking at 31 around noon, holding near 30 through four, easing after six. Reno has the same picture — widespread 30 to 35, with 40 to 45 in the wind-prone US-395 corridors, which is not us. Dust control tightens after ten. Silt fence and tarps want a second strap. Anything light and stacked stays down. A boom pick is workable at 30 mph gusts but it deserves a written stand-down trigger, not a judgment call made at noon with a load in the air.

One item worth reading twice: the model spread on today’s peak gust is GFS 31, GEM 34, ECMWF 40, ICON 50 mph. ICON is the outlier again — same as yesterday, same reason, it mixes momentum down harder in terrain than the others. It has now been the high model on gusts three days running and it has not been right yet. Weight it accordingly, but do not zero it out; a model that is wrong about magnitude can still be right about which day is the windy one.

The models agree on the high to within two degrees today — ECMWF 58, ICON 59, GFS 60, GEM 60. That is as tight a same-day cluster as this page has produced in two weeks. Trust the temperature. Argue about the gust.

Thursday: The Wet One, and the Models Disagree on Which Day It Is

Thursday the 3rd: high 57°F, low 42°F, 63% chance of measurable rain, gusts to 40 mph mid-afternoon. That is the hardest gust in the entire sixteen-day run, and it lands on the same afternoon as the showers.

Reno is clear on the timing: shower chances start late Wednesday night, peak in coverage and probability Thursday afternoon and evening at 45 to 70%, confined to northeast California, the Tahoe Basin, and northern Washoe. Totals 0.02 to 0.25″, higher along the crest. Thunderstorm chances 10% or less, with a stray strike possible from forcing aloft. Aviation expects MVFR or IFR at KTRK with mountain obscuration as early as tonight.

Here is the honest uncertainty, and it is a timing problem rather than a yes-or-no problem. GFS puts 0.11″ on Thursday and 0.09″ on Friday. ECMWF puts 0.01″ on Thursday and 0.23″ on Friday. GEM and ICON both hold Thursday nearly dry and load Friday. Three of four models want Friday to be the wetter day; Reno wants Thursday. That is a real disagreement between the office and half the guidance, and I am not going to paper over it.

What it means for a schedule: plan both days wet and be pleasantly surprised by one of them. Rain gear on the truck tonight, not Thursday morning. Do not put a flatwork pour, a hydro-seed, or anything needing a clean dry subgrade on Thursday or Friday. Friday’s minimum humidity is 66% in the grid, which tells you the air stays saturated even if the rain gauge stays empty — compaction on fine-grained material is going to fight you either day.

And remember the receipts before you cancel anything: this page’s precipitation false-alarm rate is 86%. When it says rain, it is usually wrong. A 63% probability from a source with that record is a reason to bring gear, not a reason to stand down a crew.

The Frost Call: Reno Is Saying It Out Loud Now

Reno’s discussion: overnight lows start to flirt with near freezing tonight for Sierra communities, continuing through the weekend. My grid has the lows at 42, 42, 43, 39, 39 and 40°F tonight through Monday.

Two days ago I told you to add a five-degree warm correction to those numbers. Yesterday I walked that back, because the verification record says the standing overnight-low bias on this page is +1.1°F, not five. That correction still stands and I am not moving it again. But the shape of the risk has sharpened: Saturday and Sunday nights are the two to watch, at 39°F on the grid, post-frontal, clear air, and the wind finally quitting.

Add the dewpoint from this morning — 32°F — and you have the ingredients. Radiational cooling on a clear calm night takes the air down toward the dewpoint, and the Martis flats sit several degrees under the airport on any night that actually goes calm. 39°F on the grid with a 32°F dewpoint is a frost morning in the low spots. It is not a frost morning at the shop.

The instruction is unchanged and it is the right one: check the station at five and make the call each morning. Blankets staged, not deployed. Water lines and any exposed fresh concrete get covered Friday night regardless, because Saturday is the first morning with no wind in it.

Next Week: The Fire Numbers Everyone Should Be Looking At

Nobody is talking about next week because this week has rain in it. Look at it anyway.

Minimum relative humidity, Truckee, day by day: Saturday 34%, Sunday 25%, Monday 19%, Tuesday 13%. Highs climbing 67, 68, 71, 76°F across the same four days. That is a ridge building back over fuels that will have taken a quarter inch of rain and then dried out completely inside seventy-two hours — which is exactly how long fine fuels need.

Circle Tuesday the 8th. Thirteen percent humidity is below my critical line, and it is the driest number on this page in sixteen days. What Tuesday does not have is wind — gusts collapse to 15 mph, and low humidity without wind is a spark-ignition problem, not a fire-spread problem. That is a meaningful difference. It means the risk is entirely on the equipment side and entirely inside your control. Spark arrestors, exhaust contact with cured grass, grinding and cutting, chain drag. Hot work goes to first light, when humidity is still in the fifties.

Monday the 7th at 19% is the first day over the line and deserves the same treatment one notch softer. Reno at 11 to 13% both days will almost certainly draw a fire product on the Nevada side; whether it reaches our zone depends on whether any wind shows up with it, and right now none does.

The second-week fork has narrowed, and it narrowed toward hot and dry. Yesterday GFS built a ridge and ECMWF brought the tropics. Today ECMWF has switched sides: it now runs the 8th, 9th and 10th at 72, 80 and 85°F with humidity down to 14%, which is the ridge solution with the volume turned up. GFS has the same three days at 76, 73 and 77°F. Only GEM holds the wet story, and it holds it hard — 0.47″ on the 10th, the wettest single day anywhere in the run.

So the vote is now two for the ridge, one for the rain, and one model that does not reach that far. That is not a resolved forecast, and Reno says so plainly: “forecaster confidence remains very low.” Their number for late Sunday night is 15 to 20% for rain, and they flag that Marie’s moisture may show up as nothing but cloud cover. Labor Day weekend may or may not get soggy, and anyone telling you which is guessing.

Confidence

How this page has actually been scoring. 80 days verified, last 30 averaging 86 out of 100. High bias −0.1°F, low bias +1.1°F, wind bias +0.1 mph — all three effectively unbiased. The GO/NO-GO call is 30 for 30 over the last thirty days. Precipitation is still the weak leg at 86% false alarm, 25% miss. Yesterday scored 85: the high missed by 8°F on the warm side — I published 66°F and the front delivered 58°F — the low was within two, wind within three, and the no-rain call was correct. That miss is worth naming. The cold air arrived earlier and deeper than every model had it, which is the classic behavior of a dry frontal passage and a good reason to treat today’s 60°F as a ceiling rather than a target.

Today: locate. Marked, exposed, verified. Four models inside two degrees on the high, a posted advisory with explicit numbers, and a station reading that matches the grid to the tenth. The only open item is the peak gust, and every decision is the same at 31 or at 50.

Thursday and Friday: pothole. I know the setup cold — shortwave energy from a low to the north, moisture into the Basin, showers peaking in the afternoon. What nobody has pinned is which of the two days carries the water. Prepare both, commit to neither.

Saturday and Sunday nights, frost: pothole on the pattern, old as-built on any single morning. Cold is certain, clear is certain, the dewpoint supports it. Which morning actually reaches 32°F in the flats depends on whether the wind quits, and a four-day wind forecast does not answer that.

Monday and Tuesday, fire weather: locate on the humidity, old as-built on the wind. Four models agree the ridge builds and the air dries out. None of them has a credible wind signal yet, and wind is the variable that separates a careful day from a dangerous one. If a wind signal appears in Friday’s runs, Tuesday changes character entirely.

Marie and the holiday weekend: old as-built. Named system, real moisture, no agreement on track. Reno’s own confidence is “very low.” Two of four models have already jumped from the wet solution to the dry one in twenty-four hours, which tells you how little the guidance is anchored. Keep the weekend loose.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: today is a good cold dig day — work it, manage the dust, strap the light stuff, and write the boom trigger before you leave the yard. Rain gear on the truck tonight for a Thursday and Friday that are both suspect. Cover water lines and fresh concrete Friday night for Saturday’s frost. And put a mark on Tuesday the 8th — 13% humidity on cured September fuel, no wind to spread it, and every ignition source riding on your trailer.

Forecaster’s Note · Tuesday, September 1, 2026

Saturday I told you to circle today. Sunday I said Reno was telling us where a Red Flag was going to land. Monday the watch posted. At 12:12 this morning the Reno office converted it: Red Flag Warning, 11 AM to 11 PM today, zones CAZ270 and CAZ278 on the California side, NVZ420 and NVZ458 on the Nevada side. The call verified end to end, four days out.

Now read the zone list again, because the useful part of a verified call is knowing exactly what it covers. Truckee is still not in the warning. What we got instead is a Lake Wind Advisory, 11 AM to 8 PM, covering CAZ071 and CAZ072 — the Greater Lake Tahoe Area, and Reno names Donner Lake specifically. That is the first alert of any kind on our zone in this stretch, and it is a wind product, not a fire product.

Right Now, and the Gap Ran the Other Way

The airport station read 55°F at 4:15 this morning, dewpoint 34°F, relative humidity 44%, wind 6 mph out of the southwest, sky clear. The model grid had the basin at 50°F and 40% at the same minute.

For two mornings running I told you the grid was about five degrees warm and thirty-some points dry before dawn. This morning it is five degrees cold and the humidity gap has collapsed to four points. That is not the grid getting better overnight. That is the pre-frontal southwest flow already working: the wind never went calm, so the cold air never pooled in the Martis flats, so there was no pool for the grid to miss. The station and the grid agree on windy nights and diverge on still ones. Know which kind of night you had before you trust either number.

The verification record backs this up and corrects something I said yesterday. Across 79 scored days the overnight-low bias on this page is +1.1°F, not five. Yesterday morning it ran three degrees cold on the low. The five-degree warm gap was a two-day clear-and-calm signature, not a standing bias, and I overstated it. Apply that correction to still nights only.

Today: GO, and It Is a Wind Day

High 66°F, low 45°F, zero precipitation, UV index 7.1, sunrise 6:29, sunset 7:31. That is roughly ten degrees below normal for the first of September and it is the coolest daytime number since the middle of last week. All systems GO for outdoor work. Grading, pipe, compaction, flatwork, dust-generating passes — today takes all of it. The thing today asks you to manage is wind, and the thing it asks you to watch is elevation.

The humidity ladder: 64% at six, 30% by nine, 26% from noon through one, 30% by three, 39% by six. Twenty-six percent is dry but it is not my line. Below 20% gets attention, below 15% is critical, and the basin does not go there today. Compare that to Reno, which bottoms at 12% this afternoon. Our 3,000 feet of elevation is buying twelve degrees and fourteen points of humidity today, and that entire margin is the reason the warning stops at the crest.

Basin gusts build through the morning: 24 mph by nine, 23 at noon, 26 mid-afternoon, easing after six. Those are ordinary numbers. Dust control tightens after ten, tarps and silt fence want a second strap, and anything light and stacked should already be down from yesterday’s note. Nothing here shuts a job down on the valley floor.

The Ridgetop Number Is the One That Matters

Reno has ridgetop gusts increasing to 70 to 80 mph today, valley gusts of 30 to 40 mph from Topaz Lake north to the Oregon border, and 20 to 30 everywhere else. Aviation is carrying 60 to 70 knot winds at 10,000 feet with mountain wave turbulence and low-level wind shear at every Sierra and Sierra Front site.

My four model streams do not agree on how much of that reaches the deck here, and the spread is the story. For today’s peak basin gust: GFS 27, GEM 34, ECMWF 35, ICON 48 mph. Twenty-one miles an hour of disagreement on a same-day forecast. ICON is the outlier and ICON is also the model that mixes momentum down most aggressively in terrain — which is exactly the mechanism in play when you have 70 mph over the crest and a dry frontal passage. I would not bet the job on 48, but I would not plan a boom pick assuming 27 either.

The practical rule: if your work is above the basin floor, use the ridgetop forecast, not the valley one. Ridge access roads, hillside cuts, anything at Palisades elevation. Crane and boom work high on the hill should have a stand-down trigger written before the crew leaves the yard, not negotiated at two in the afternoon with a load in the air. On the water, the advisory is the whole message — Donner and Tahoe will be building chop from eleven onward.

Fire side, even though we are outside the warning: hot work goes to first light, when humidity is still in the sixties and the wind has not stood up. Spark arrestors verified, not assumed. Extinguisher and charged water riding with the crew. Check the exhaust side of every skid steer and mini-ex before you set it down in cured grass. September grass in the Sierra does not care which zone the warning polygon stopped at.

Wednesday and Thursday: The Cold, Wet Middle

Wednesday: high 60°F, low 40°F, humidity recovering to a minimum of 45%, gusts 35 mph — and ICON again louder at 51. Fire risk drops off a cliff with that humidity recovery; Reno says so explicitly. Wednesday is a cold, breezy, entirely workable day with a rigging problem attached. Note that the strongest basin gusts of this week land Wednesday and Thursday, not Friday, which reverses what I wrote yesterday — Friday has backed down from 30 mph to 22 in this morning’s run.

Thursday the 3rd is the wet one, and it is also where the models fight. High: ICON 49, GFS 55, ECMWF 61, GEM 64. Fifteen degrees of spread at two days out is unusual and it comes down to how much cloud and shower each model puts over the crest. Precipitation is light in all of them — 0.08″ on the raw run, a 19% probability — but the humidity profile is the part that matters for work: the grid holds Thursday above 90% relative humidity from evening through the following morning. Reno has shower chances expanding to 30–55% and dipping south into the Tahoe Basin Thursday into Thursday night, with a 5–10% embedded thunderstorm mention Wednesday night.

Plan Thursday as cold and damp. Rain gear on the truck Wednesday night. Do not schedule anything that needs a dry surface or a clean subgrade, and do not put a hydro-seed or a flatwork pour on it. Friday recovers to 62°F and Reno calls Friday and Saturday relatively dry; the weekend climbs back to 67 and 68°F.

The Frost Call: Walk It Back One Notch

Reno is still explicit that overnight lows drop near to below freezing for Sierra communities like Truckee through the back half of the week. My grid has the lows at 40, 40, 40 and 39°F Wednesday through Saturday. Yesterday I told you to add a five-degree warm correction to those and call it frost. The record says do not do that. The standing low bias is +1.1°F and it ran cold yesterday.

So the honest version: the pattern is right — post-frontal, clear, dry air, long nights — and the Martis flats will run several degrees under the airport on any night the wind actually goes calm. But 40°F on the grid is not 32°F on the ground unless the drainage sets up, and Wednesday night still has gusts in it. Check the station at 5 a.m. and make the call each morning. Blankets staged, not necessarily deployed. This is a real risk, not a scheduled event.

Next Week Is a Genuine Fork, and I Am Not Going to Pretend Otherwise

Two models are telling completely different stories about the second week of September, and the difference is not detail — it is the whole pattern.

GFS builds a ridge. The 8th, 9th and 10th at 74, 81 and 81°F with minimum humidity of 13, 15 and 16%, and gusts collapsing to single digits. That is a hot, bone-dry, dead-calm stretch on fully cured September fuels. Wind is the only missing ingredient for a real fire-weather setup, and on that solution there is no wind.

ECMWF brings the tropics. Same three days at 68, 63 and 56°F, with 0.11″ on the 9th and 0.53″ on the 10th — the wettest day anywhere in the sixteen-day run from any model. That is remnant tropical moisture riding north into the Sierra.

Twenty-five degrees and half an inch apart at nine days. Reno flags the same thing and puts it plainly: precipitation chances of 15–25% Sunday onward as models hint at tropical moisture, with plenty of uncertainty on the track. Both of those futures are physically coherent. Neither one is more likely because it is on the page you happened to open.

What to do with a fork: keep the second week loose and do not commit to either story. Do not schedule a pour that needs four dry days. Do not stand down a crew for rain that has a one-in-five chance. Watch it Friday, when the tropical system’s track will be inside the range where the models start converging. And remember the receipts — this page’s precipitation false-alarm rate is 86%. When it promises rain a week out, it is usually wrong.

Confidence

How this page has actually been scoring. 79 days verified, last 30 averaging 86 out of 100. High bias −0.6°F, low bias +1.1°F, wind bias +0.4 mph — all three essentially unbiased. The GO/NO-GO call is 30 for 30 over the last thirty days. Precipitation remains the weak leg: 86% false-alarm rate, 25% miss rate. Yesterday scored 95 — high exactly right at 77°F, low three degrees off, wind two off, no-rain call correct. Trust this page on temperature and wind. Do not trust it on rain.

Today: locate. Marked, exposed, verified. Posted Red Flag Warning, posted Lake Wind Advisory, a named frontal passage, and specific ridgetop numbers from the office. The only open item is the peak basin gust, where the models spread 27 to 48 mph — and every decision that matters is the same at either end.

Wednesday: locate. Cold, breezy, humidity recovered, fire risk down. Reno and all four models tell the same story.

Thursday: pothole. I know the setup and the shower chance is real, but a fifteen-degree spread on the high at two days out means nobody has the cloud cover pinned. Plan for the cold end and let the number sort itself out.

The frost, Wednesday through Saturday: pothole on the pattern, old as-built on any individual morning. Cold is certain. Which morning actually reaches 32°F in the flats depends on whether the wind quits, and four-day wind forecasts do not resolve that.

The 8th onward: old as-built, and worse than usual. This is not a case of the models disagreeing on digits. They disagree on whether a ridge or a tropical remnant owns the second week of September. Right general layout does not even apply — there are two layouts. Take nothing from that range except the instruction to stay flexible.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: work the day, it is a good one. Use the ridgetop forecast if you are working high, and write the boom stand-down trigger before you leave the yard. Hot work at first light. Rain gear on the truck Wednesday night for Thursday. Check the station at five each morning for frost instead of trusting the grid. And leave the second week of September wide open, because right now there are two of them.

Forecaster’s Note · Monday, August 31, 2026

Saturday I told you to circle Tuesday the 1st. Sunday I said that when the Reno office writes widespread and critical in the same sentence three days out, they are telling you where a Red Flag is going to land. At 1:28 this morning they posted it: a Fire Weather Watch, Tuesday morning through Tuesday evening, covering fire zones CAZ270, CAZ278, NVZ420 and NVZ458 — Surprise Valley, eastern Lassen, northern Washoe, and the Northern Sierra Front including Reno and Carson City.

Read the zone list before you get excited or complacent. Truckee is not in it. Our point still shows zero active alerts. The watch sits on the leeward side of the crest, where the front dumps downslope and the humidity bottoms out. What we get is the same front on its way over the top, and the ridgetop numbers are the ones that should hold your attention.

Right Now, and the Station-vs-Grid Gap Again

The airport station read 41°F at 4:15 this morning, dewpoint 36°F, relative humidity 81%, wind flat calm, sky clear. The model grid I pull for this page had the basin at 46°F and 44% at the same minute.

That is the second morning running with the same signature: the grid is roughly five degrees warm and thirty-seven points dry on the pre-dawn number. It is not a broken model. Cold air drains off the slopes into the Martis flats every clear calm night and pools there, and a smoothed grid box cannot see a pool. If you are calling a crew start time or deciding whether a pour needs blankets, use the station overnight and the grid for the afternoon. Your phone app is using the grid.

Today: The Last Easy One

High 76°F, low 45°F, zero precipitation, UV index 7.5, sunrise 6:28, sunset 7:33. Reno calls it “seasonably warm and dry today with typical afternoon zephyr breezes.” That is the right description. Gusts stay under 10 mph until mid-morning, reach 16 mph around noon and top at 20 mph near four o’clock. All systems GO for outdoor work. Grading, pipe, compaction, concrete, lifts, crane picks — today takes all of it.

Humidity is what today asks you to manage, not wind. The ladder: 81% at dawn, 37% by eight, 25% by ten, 19% from noon through two, back to 25% by six. Below 20% deserves your attention. Below 15% is where I call it critical. We touch 19 and hold for a couple of hours — dry, not alarming, and about two points drier than yesterday at the same hour.

On the ground that means the same thing it meant yesterday, only slightly more so. Trench bottom and fine grading want the morning while the overnight moisture is still doing half your compaction. By early afternoon the same pass throws a plume, and with light wind it hangs in the cut instead of clearing. Respirators on saw work. Water truck cycle tightens after eleven.

Calibration against the valley: Reno runs 87°F today against our 76°F. Our 3,000 feet buys eleven degrees. Reno’s fire discussion has humidity near 10% today in parts of northeast California and northwest Nevada with gusts around 25 and isolated near 30 — but again, that is Lassen and Washoe, not the basin. If you have crews split, the valley crew is having the harder day.

Tuesday: The Front. Plan It Today.

A dry cold front crosses tomorrow. Locally: high 65–68°F depending on which model you ask, minimum humidity 18% at midday, gusts 23 mph out of the southwest. Those are workable numbers on their own. The context around them is not ordinary.

Reno has North Sierra ridgetop gusts near 50 mph tonight, building to near 70 mph by Tuesday evening, with valley gusts of 30 to 40 mph north of US-50 in the afternoon and evening. Aviation is carrying mountain wave turbulence and low-level wind shear all day. Blowing dust on the playas. If your work sits high — ridge access roads, a hillside cut, anything above the basin floor, anything at Palisades elevation — you are much closer to the ridgetop number than to the 23 mph the valley grid gives you.

Write the fire plan today while nobody is in a hurry. Hot work — grinding, cutting, welding — goes to first light, when humidity is still near 40% and the wind has not stood up. Nothing outside a cleared pad after ten. Spark arrestors verified, not assumed. Extinguisher and a charged water source riding with the crew, not parked at the gate. No dragging chains on rock. Check the exhaust side of every skid steer and mini-ex before you set it down in cured grass. A hot muffler in late-August cheatgrass is how a real share of these actually start.

The rigging side of Tuesday matters as much as the fire side. Anything light and stacked — tarps, silt fence, rigid foam, form panels, empty conex doors, jobsite signage, the porta-john — gets strapped or laid down this afternoon, not tomorrow morning. Boom and crane work: read the ridgetop forecast, not the valley one, and be ready to stand down after midday.

Wednesday Is a Frost Problem, Not a Wind Problem

Behind the front, Wednesday is the week’s coldest day: high 62°F, and Reno is explicit that low temperatures in the colder Sierra valleys “may dip below freezing each morning from Wednesday through the start of the weekend.” My grid gives Wednesday morning a low of 36°F — and remember the gap I opened this note with. The grid ran five degrees warm on the overnight low both of the last two mornings. Apply that correction and Wednesday morning is at or below freezing in the Martis flats.

That is a concrete call, a water line call, and a hydro-seed call, and it needs to be made Tuesday afternoon at the latest. Blankets on anything green. Bleed exposed lines. Do not schedule an early flatwork pour for Wednesday and hope.

Humidity recovers to 34% Wednesday, which knocks the fire risk down even though gusts hold at 24 mph. Wednesday is a normal, cold, breezy work day.

Friday Is the Real Story Nobody Is Looking At

Thursday is a placeholder — 61–62°F, humidity back to 42%, gusts 24 mph, and a shower chance the two model streams cannot agree on: the raw run says 13% with a trace, the bias-corrected ensemble behind the crew page says 29%. Reno puts the actual shower odds at 20–50% but places them over western Lassen and far northern Washoe. North of us.

Friday the 4th is where the week turns — and it is also where my two model streams split the widest they have all week, so I am going to show you both rather than pick one.

The raw run has Friday at a high of 51°F with minimum humidity 81%, meaning the day never dries out. The bias-corrected ensemble that feeds the crew page has it at 59°F. Eight degrees apart at four days out is a wide split, and the reason is worth knowing: the ensemble has learned this grid runs cold on cloudy days and warms it back up, while the raw run takes the cloud deck at face value. Both put precipitation probability at 27% with roughly 0.08″, and both carry gusts to 30 mph — the strongest single-day gust anywhere in the sixteen-day run.

Here is the tell that makes me lean cold. Reno sits at 70°F on Friday. Against the raw 51°F that is a nineteen-degree spread across 3,000 feet of elevation, which is far more than a dry lapse rate produces on its own. A spread that large is cloud, moisture and orographic lift sitting on our side of the crest while the valley stays in the clear — a coherent physical story, not a model artifact. Against the ensemble’s 59°F the spread is eleven degrees, which is just an ordinary day. Reno’s own text gives the Tahoe Basin only a 10–20% shower chance Friday, so nobody is promising much rain either way.

Plan Friday as a raw day and let the exact number sort itself out. Somewhere between 51 and 59°F, wet-feeling, with 30 mph gusts is the combination that ends productivity and starts hypothermia complaints regardless of which end it lands on. Rain gear on the truck Thursday night. Do not schedule anything that needs a dry surface.

The Two-Inch Storm Is Gone. Told You.

Yesterday I flagged a single deterministic run that had put 2.10 inches on Monday the 7th with a 57°F high, when the same model’s wettest day twenty-four hours earlier was 0.012 inches. I said do not plan around it, do not close a trench for it, and do not repeat it to anybody as a forecast.

Today’s raw run has the 7th at 81°F and zero precipitation; the ensemble is cooler and damper at 72°F with a 24% shower chance. Neither one has anything resembling two inches. The storm never existed. That is what a single-run outlier looks like when it washes out, and it is exactly why the confidence section is at the bottom of every one of these.

The ridge behind it is back, too. Saturday I had the 8th through the 11th at 85–89°F with humidity minimums of 12–15%. Sunday’s run softened it to 76–82°F and 23–29%. Today it firms back up to 83, 84, 86 and 81°F with minimums of 20–26% — most of the way back to where Saturday had it. Three runs, three different answers, one week out. Take the shape, ignore the digits. Further out, the 13th and 14th show 86 and 90°F with humidity minimums of 13%, which would be a genuine fire-weather setup if it verifies. It is fifteen days away. It probably will not verify in that form.

One more thing worth carrying: Reno is watching the possibility that moisture from a tropical system gets drawn north into the Sierra Sunday into early next week. They say confidence is very low and it depends on how the system interacts with an offshore low. My grid shows nothing — 0.01″ on the 6th and dry after. Two sources, two answers, and the honest read is that nobody knows yet. Watch it Wednesday.

Midday Update — The 7th Just Flipped Back

I wrote that last paragraph this morning and told you to watch it Wednesday. It did not wait until Wednesday. The midday run reversed the 7th inside of four hours: this morning the raw model had Monday the 7th at 81°F and bone dry, and the run I am looking at now has it at 62°F with 0.40″ and a rain code on it — the wettest single day anywhere in the sixteen-day outlook. That is a nineteen-degree swing and a fifth of an inch appearing out of nothing, on the exact day Reno told us to watch the tropical moisture.

Do not plan around it. I am going to be blunt about why, because I have the receipts: across 75 verified days, this grid’s precipitation false-alarm rate is 85%. When it promises rain a week out, it is wrong far more often than it is right, and the probability field on the 7th is still only 10% even while the accumulation field says 0.40″ — the model is arguing with itself. What actually changed is that a wet solution is now on the table where this morning there was none. That is a reason to keep Sunday and Monday loose on the schedule. It is not a reason to move work.

The ridge behind it moved too, and it moved the direction that matters. This morning I had the 8th through the 11th at 83, 84, 86 and 81°F with humidity minimums of 20–26%. The midday run has 75, 85, 86 and 79°F with minimums of 18, 13, 11 and 14%. Same heat, five to twelve points drier. Thursday the 10th at 86°F and 11% humidity is the one to put on the calendar — that is a hot, dry, second-week-of-September afternoon on cured fuels, and it is the first day in this run that would stand up a Red Flag on its own merits if the wind shows up with it. Right now it does not; gusts stay in the single digits through that stretch. Wind is the missing ingredient, and wind is also the thing nobody can forecast ten days out.

Nothing in the first five days changed. I re-ran every number in this note against the midday data before publishing the update. Today still 76°F. Tuesday still 68°F with 23 mph gusts and the Fire Weather Watch still sitting on the Nevada side of the crest. Wednesday still 62°F with a 38°F grid low, which is still a frost call once you apply the warm bias. Thursday still low-60s. Friday still 50°F, still 0.098″, still 30 mph gusts — the strongest in the run. Reno’s Friday came down from 70 to 67°F, which narrows the cross-crest spread from nineteen degrees to seventeen. That tightens the story rather than breaking it. Every call in this note stands.

Confidence

How this page has actually been scoring. 75 days verified, and the last 30 average 88 out of 100. Daytime highs are running essentially unbiased — mean high bias 0.0°F. Wind is unbiased. The overnight low is the weak spot at 3.1°F mean error, biased warm, which is the same cold-air-pooling gap I opened this note with and the reason the Wednesday frost call gets made off the station and not off the grid. The GO/NO-GO call itself is 30 for 30. Precipitation is the one to distrust: 85% false-alarm rate, 25% miss rate. This page is good at temperature and wind and bad at rain, and you should weight it accordingly.

Today: locate. Marked, exposed, verified. Clear, calm morning, warm dry afternoon, light zephyr wind. The grid, the station and the Reno discussion all say the same thing and the physics is trivial.

Tuesday: locate. Upgraded from yesterday. There is now a posted Fire Weather Watch, a named frontal passage, and specific ridgetop and valley gust ranges from the office. The event is not in question. The only open item is the exact ridgetop peak, and 50 versus 70 mph does not change a single decision you should be making.

Wednesday’s frost: locate on the pattern, pothole on the number. That it gets cold is certain — post-frontal, clear, calm, dry air. Whether the Martis flats hit 30°F or 34°F is the difference between a real frost call and a nuisance, and the grid has proven it runs warm on exactly that number. Check the station at 5 a.m., not the app.

Friday: pothole. I know this area and I know this setup. A 26-degree drop with a 19-degree Reno spread is a coherent, physical story. But the two sources disagree on precipitation amount by a factor of two, and four-day peak-gust forecasts routinely move ten miles an hour either direction.

The 6th onward: old as-built. Right general layout, do not scale off it. The ridge has moved twice in three days — three times counting the midday run above — the tropical moisture is a coin flip by Reno’s own admission, and yesterday’s two-inch storm is a fresh reminder of what day-seven guidance is worth. The 7th flipping nineteen degrees between breakfast and lunch is the same lesson arriving twice in one day.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: spend today on the work that needs still air and a dry surface, because it is the last of both for a while. Strap the light stuff this afternoon. Write Tuesday’s fire plan before you leave the yard tonight. Make the frost call Tuesday for Wednesday morning. Put rain gear on the truck for Friday. And leave Sunday and Monday loose until Wednesday tells us whether the 7th is real.

Forecaster’s Note · Sunday, August 30, 2026

The station at the Truckee airport read 34°F at 3:55 this morning with relative humidity at 86% and the wind at 3 mph out of the south. Dewpoint 30°F. That is a four-degree spread between air and dewpoint on a clear calm night, which is the exact recipe for what Reno is already forecasting: patchy fog across the Martis Valley and the Truckee flats through about sunrise, gone shortly after.

Worth noting for anyone who checks a phone app before they leave the house: the model grid I pull for this page had the basin at 44°F and 49% at the same moment. The station says 34 and 86. Neither one is broken. The grid is a smoothed average over a box, and the box does not know that cold air drains downhill into the Martis Valley every clear night and pools there. Trust the station for the 5 a.m. number and the grid for the afternoon. If you are dispatching crews on the overnight low, you are ten degrees colder than the app thinks.

Today: The Calm One. Take It.

High 77°F, low 42°F, zero precipitation, UV index 7.5, sunrise 6:27, sunset 7:34. Gusts top out around 15 mph. Reno’s discussion says it plainly — “warm and dry conditions with lighter winds this afternoon.” All systems GO for outdoor work. Grading, pipe, compaction, concrete, lifts, anything you have been holding for a still day. This is that day, and it is the last one for a while.

The one thing today asks of you is dust and dry air, not wind. Humidity runs the ladder down hard: 86% now, 29% by 10 a.m., 17% from noon straight through 4 p.m. Seven straight hours at or under 20%. Below 20% deserves your attention on any day; below 15% is where I start calling it critical. We touch 17 and hold, which is dry but not the alarm.

What that means on the ground: the water truck cycle that worked on Friday is wrong for this afternoon. Fine grading and trench bottom work want the morning, while the overnight moisture is still doing half your compaction for you. By two o’clock the same pass throws a plume. And with light wind, that plume hangs in the cut instead of clearing — better for the neighbors, worse for the guy in the hole. Respirators on the saw work.

For calibration, Reno today runs 86°F with humidity bottoming at 11%. Our 3,000 feet of elevation buys us nine degrees and six points. If you have crews split between the basin and the valley floor, the Reno crew is having a materially worse day than yours.

The Week: Circle Tuesday the 1st

A low builds off the Pacific Northwest starting tonight and we spend the rest of the week on the dry, breezy gradient side of it. Monday is the warning shot: 76°F, minimum humidity 22%, gusts to 19 mph. Reno flags poor overnight humidity recovery Monday with elevated to near-critical conditions — but the areas they name are northern Washoe, Surprise Valley and eastern Lassen, not us. Monday is a normal work day here.

Tuesday the 1st is the one. High drops to 68°F, minimum humidity 16%, local gusts 25 mph out of the southwest — and Reno is carrying 30 to 40 mph for the Sierra Front with minimum RH in the 10–20% range. Their words, not mine: “Tuesday will see the best potential for widespread critical fire weather concerns.” When the office writes widespread and critical in the same sentence three days out, that is a forecaster telling you where a Red Flag Warning is going to land. There is no watch or warning posted for our point right now — zero active alerts as I write this — but Tuesday is the day one shows up if one shows up.

Put the plan on paper today, not Tuesday morning. Hot work — grinding, cutting, welding — moves to first light, when humidity is still in the fifties and the wind has not stood up. Nothing outside a cleared pad after ten. Spark arrestors verified on every machine, not assumed. Extinguisher and a charged water source riding with the crew, not sitting at the gate. No dragging chains on rock. Check the exhaust side of every skid steer and mini-ex before you park it in cured grass, because a hot muffler in August cheatgrass is how a real percentage of these things actually start.

The rest of the week reads like this: Wednesday cools hard to 61°F and humidity recovers to 28%, which knocks the fire risk down even though the wind stays up at 24 mph. Reno agrees — roughly ten degrees of cooling from Monday to Wednesday. Then Thursday the 3rd is the wind day: gusts to 34 mph, the strongest number anywhere in the sixteen-day run, on a 67°F afternoon. That is not a fire number, that is a rigging number. Crane and boom work, panel setting, tarps, silt fence, rigid foam, empty conex doors, anything stacked and light — Thursday is the day it goes for a walk. Strap it Wednesday.

Rain: Reno Says Maybe, My Grid Says No, and One Run Lost Its Mind

Locally there is nothing. Precipitation probability stays at or under 11% every day through the 6th and the accumulation column is a straight line of zeroes. Reno carries 15–30% shower chances Thursday into the start of the weekend as moisture filters back in, but read the geography carefully — they place it across far northeast California and northwest Nevada. That is north and east of us. We are on the edge of it at best.

Then there is the day-nine problem. One run this morning puts 2.10 inches on Monday the 7th with a high of 57°F. Twenty-four hours ago the same model’s wettest day in a sixteen-day window was 0.012 inches. A model does not go from a dirty windshield to a two-inch soaker in one cycle because the atmosphere changed — it does that because a single deterministic run caught a feature nobody else has. Do not plan around it, do not close a trench for it, and do not repeat it to anybody as a forecast. It goes in the notebook as something to watch, and if it is still there Wednesday it becomes real.

One correction to Saturday’s note while I am here. I flagged the 8th through the 11th as a hot dry ridge at 85–89°F with humidity minimums of 12–15%. This morning’s run has that same block at 76–82°F with minimums of 23–29%. Still warm, still dry, but meaningfully softer than what I wrote. That is what a nine-day forecast is worth, and it is why the confidence section exists.

Confidence

Today and Monday: locate. Marked, exposed, verified. Light wind, dry afternoon, the fog burning off by mid-morning. The grid and the Reno discussion say the same thing and the physics is obvious.

Tuesday: locate on the pattern, pothole on the peak. That the day is dry, windy and dangerous is not in question — two independent sources and a textbook setup. What is not pinned is whether the Sierra Front gust maximum lands at 30 or 40, and that is the whole difference between elevated and critical. Re-read Tuesday at first light before you commit any hot work.

Thursday’s wind: pothole. I know the area. A 34 mph gust nine days into a breezy trough pattern is credible, but single-day peak winds at four days out routinely move ten miles an hour in either direction.

The 7th onward: old as-built. Right general layout, do not scale off it. Two sources disagree about whether there is any moisture at all, one run has a two-inch storm that appeared out of nowhere, and Saturday’s ridge already moved nine degrees in a day.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: today is the still day — spend it on the work that needs still air. Strap the light stuff Wednesday for Thursday. And write Tuesday’s fire plan down today, while nobody is in a hurry.

Forecaster’s Note · Saturday, August 29, 2026

If you are standing outside right now you are standing in the wettest air you will feel all week. 49°F, relative humidity 71%, wind essentially nothing at 2 mph out of the northwest, pressure 821.8 hPa at station level. Reno is calling patchy fog for the Martis Valley and the Truckee flats this morning off the leftover moisture from midweek, and that is exactly what a clear calm night over damp ground does to a basin. It will burn off by mid-morning.

Then the floor drops out. Hour by hour today: 79% RH at 8 a.m., 40% by 10, 32% by noon, 24% by 3 p.m. That is fifty-five points gone in seven hours. Reno’s discussion puts the afternoon drying at 40–50 points below yesterday for everything south of I-80 that caught the wetting rain. Nothing about it is unusual for the last week of August. It is just a big number to move through in one shift, and it changes what the ground and the air will do to you between morning break and cleanup.

Today: Dig Day, With Two Notes

High 74°F, low 47°F, no precipitation, UV index 7.5, sunrise 6:26, sunset 7:36. All systems GO for outdoor work. Compaction, pipe, grading, concrete — nothing on the board today says stop.

Two things to plan around. First, the wind builds through the afternoon: gusts hit 23 mph by noon and peak near 27 mph between 2 and 3 p.m. out of the southwest. Reno has Sierra ridgetops gusting 50+ mph and moderate turbulence at 10,000 feet, so if you are working anything elevated or exposed, that is your number, not the valley one. Tarps, rigid insulation, empty conex doors, silt fence, anything on a stack that isn’t strapped — strap it before lunch, not after it lets go.

Second, the dust arrives with the drying. Wet ground at 7 a.m. and 24% humidity at 3 p.m. means your water truck cycle from yesterday is wrong for this afternoon. Front-load the trench work and the fine grading into the morning while the moisture is still doing the compaction work for you.

For calibration: Reno sits at 83°F today with RH bottoming at 17%. That 3,000 feet of elevation is worth about nine degrees and seven points of humidity to us right now. If you have crews split between the basin and the valley floor, they are not having the same day.

The Week: A Ridge, Then a Trough That Doesn’t Pay Rent

Sunday and Monday are the pleasant end of it — 76°F both days, winds backing off to 11–14 mph, RH minimums of 19% and 18%. Warm, dry, workable, unremarkable.

Then troughing sets up over the Pacific Northwest and we sit on the dry gradient side of it. Tuesday the 1st the high drops to 67°F and the low to 39°F — and the humidity goes with it, minimum 14%, with gusts back to 27 mph. Wednesday holds the pattern at 64°F and 22 mph. That is a trough that takes the temperature and gives back nothing: no moisture, no rain, just cold dry air moving fast.

Circle Tuesday the 1st. Fourteen percent humidity with a 27 mph gust on it is the combination that matters, and Reno is already using the words elevated to critical fire weather for Tuesday and Wednesday across the Tahoe Basin and northeast California. There is no Red Flag Warning posted for Truckee as of this writing — zero active alerts for our point — but that is the day a watch would land on if one lands at all.

What that means on the ground, well before anybody issues anything: no grinding or cutting outside a cleared pad Tuesday afternoon. Move the hot work to first light, when RH is still in the sixties and the wind hasn’t stood up. Spark arrestors verified, not assumed. Extinguisher and a charged water source with the crew, not back at the truck. Chains off the deck — a dragging chain on rock in cured cheatgrass at 14% is the whole story of how a lot of these start. Check the exhaust side of every skid steer and mini-ex before you park it in grass.

None of that is alarm. It is Tuesday’s tailgate topic, written down on Saturday so it is not improvised on Tuesday.

Rain: Not This Week, and I’d Hold the Line on Next

Sixteen days of model output and the largest single-day total anywhere in it is 0.012 inches on the 4th. That is not rain, that is a dirty windshield. Precipitation probability stays under 20% the whole run.

Reno is more generous, carrying 15–30% chances Thursday into next weekend as moisture starts working back in under the persistent breeze. That is a real disagreement between the deterministic run I am reading and the forecaster who has the whole ensemble in front of him, and when those two split I take the human. Call it a genuine but low chance late next week. Do not schedule around it and do not leave an open trench on the strength of it either way.

Behind that, the 8th through the 11th builds a proper late-summer ridge: 85°F, 88°F, 89°F, 88°F with RH minimums of 15%, 14%, 12%, 13%. Nine days out, so treat the numbers as shape rather than fact — but the shape is hot and very dry, and September heat on cured fuels is how this region historically gets its worst weeks. Worth knowing it is on the board.

Confidence

Today through Monday: locate. Marked, exposed, verified. The drying, the wind ramp and the weekend warmth are all agreed on and physically obvious.

Tuesday through Thursday: pothole. I know the pattern — dry gusty northwest flow under PacNW troughing — and Reno says the same. What is not pinned down is the timing and the peak strength of the wind, which is the difference between elevated and critical. Re-read Tuesday morning before you commit hot work.

The 5th onward: old as-built. Right general layout, do not scale off it. Two independent sources disagree about whether next week gets any moisture at all, and the 88°F block sits at the far end of a 16-day run.

Live conditions and the GO/NO-GO call update all day on the crew weather page. Full model detail on the weather engine. Both free, no ads, no sign-in.

Short version: work the morning, strap the light stuff before the afternoon, and put Tuesday’s fire plan on paper today.

Every forecast app on your phone is quoting a model. Almost none of them tell you which one, and none of them tell you whether that model has been right lately in the place you actually stand. In a basin like Truckee that gap matters, because the terrain does things to the air at night that a global grid cannot see.

So I built the scoreboard. Eight global models, five variables each, checked every day against ERA5 reanalysis at 39.328 N, 120.183 W. As of this week that is 120 scored days, running 2026-05-01 through 2026-08-28. Here is what fell out of it.

Daytime Highs Are Solved. Stop Worrying About Them.

All eight models land inside 1.2°F on the daytime high. The spread across the entire ensemble is under 2°F, and RMSE runs 1.28 to 2.06. UKMO is the tightest at 1.28, ICON right behind at 1.35. ECMWF carries a steady 1.19°F cold bias, which sounds bad until you notice the word steady — a consistent bias is a correctable bias, and the engine corrects it.

Practical version: if all you need is tomorrow’s high, grab whatever app is closest to your thumb. They agree.

Overnight Lows Are Where It Falls Apart

This is the chart at the top of the post, and it is not close:

• Trustworthy: ECMWF IFS (−1.23°F), ECMWF AIFS (+0.31), UKMO (+0.50)
• Drifting: CMA (+2.10), GFS (+2.27)
• Not usable: JMA (+5.15), ICON (+5.51), GEM (+5.88)

Positive means the model forecasts a warmer night than actually happened. That bottom group is running five to six degrees warm with RMSE near 6.5, which on a clear calm night is worse than just assuming normal conditions for the date. Three of the eight models are actively subtracting information.

If your app says 34°F tonight and it is quoting GEM, plan for 28°F.

UKMO Is the Trap

UKMO has the best temperature numbers in the whole set. It also has the worst precipitation detection: probability of detection 0.39, meaning it misses roughly six wet days in ten. ECMWF catches 90% of them.

That combination is genuinely dangerous, because skill on one variable reads like skill in general. It is not. A model can nail every temperature for a month and still not tell you it is going to rain on your open trench. Weight each model per variable or do not weight it at all.

Wind has the same shape. GFS runs about 3 mph hot, JMA 3.7 mph light, GEM 4 mph light. Four miles an hour is not a rounding error when it decides whether a lift or a crane works that day.

The Part That Actually Costs Money

I split the overnight-low error by season. Every model that runs warm at night runs warmer in spring:

• UKMO: +3.30°F in spring, −0.48°F in summer
• JMA: +6.81°F in spring
• ICON: +6.23°F in spring
• GFS: +3.72°F in spring

The physical story is a cold-air pool settling into the basin over leftover snowpack, and a global grid is too coarse to resolve it. The operational story is worse. Spring is exactly when you are making the frozen-ground call, deciding whether tomorrow is a dig day or a stand-around day, whether that pour survives the night. That is the season the models are least honest with you, and the season where being wrong is most expensive.

ERA5 Itself Runs Warm

One more layer, and it is the humbling one. I started archiving observations from the Truckee airport station and compared them to ERA5 over seven complete days. ERA5 came in 9.7°F warmer than the station on the overnight low.

Which means the table above flatters every model. Against what the basin actually feels at 4 a.m., UKMO is +5.8, ECMWF is +8.4, GEM is +17.1.

Seven days is a small sample and it is all summer, so I am treating the size of that number as a rough estimate and the ranking as solid. I am storing it as a measurement rather than folding it into a correction, because the engine already corrects against ERA5 and stacking the two would double-count. But the direction is unambiguous: reanalysis smooths terrain, and Truckee is terrain.

What I Do Not Know Yet

There is no winter in this dataset. The window opens May 1, so every cold-season conclusion above is extrapolated from spring and summer. Given that the spring split already shows the bias getting worse as you move toward the cold end of the year, I would expect winter to be worse still — but expecting is not measuring, and I am not going to pretend otherwise. Extending the window back through last winter is the next job.

Same caution on the whole exercise: 120 days is enough to see a five-degree systematic bias clearly. It is not enough to rank ECMWF against AIFS with confidence when they sit half a degree apart.

The Short Version

Highs, trust anybody. Lows, trust ECMWF or UKMO and nothing else. Rain, trust ECMWF and specifically do not trust UKMO. In spring, assume the night is colder than you were told.

All of this feeds the ensemble weighting on the Truckee weather engine, which is free and has no ads. The scoring re-runs nightly and the window rolls forward on its own, so these numbers move. I will post again when the winter sample lands.

So, a Bluetooth speaker is built for a backyard patio. Mine kept ending up in the back of a UTV, running a ten hour snow shift, or sharing a truck bed with a chainsaw and a jug of hydraulic fluid.

They died in that order. Four funerals was enough — I quit buying and started welding.

The Case Does the Heavy Lifting

The enclosure is a Pelican 1600. Crushproof, dustproof, watertight out of the box, the same case people use to fly camera bodies and rifles around the world. If it survives baggage handlers, it survives my crew. Mine gave up the watertight part on purpose. I'll get to why.

Inside is real car audio, not consumer parts. Alpine SPR-60C components — two 6.5 inch woofers, two silk soft dome tweeters, passive crossovers — mounted in the lid behind a welded bracket panel. A four channel Rockford Fosgate Prime amp drives them off a 12V sealed AGM battery sitting in a steel cradle I welded and bolted to the case floor. The entire interior is lined with Dynamat, and that part isn't optional. A bare Pelican case is a plastic drum. Line it or your bass turns into a rattle.

Three Things I Learned

Heat kills amps, water doesn't. Everybody building for a jobsite worries about moisture getting in. The thing that actually cooks an amplifier is ten hours of running with nowhere for the heat to go. So I cut three louvered slots into the case side, angled so rain and snow shed off while air still moves across the amp. That's the watertight rating gone, and it's the difference between a build that lasts a season and one that lasts years.

Vibration is the real enemy. Forest service roads at 6,000 feet will shake a build apart faster than any weather. The battery is AGM specifically because it's sealed and vibration rated, and it rides on foam inside that welded cradle. A flooded battery bouncing loose in a case thrown into a truck bed is a bad day waiting to happen.

Put a voltmeter on it. Cheapest part in the whole build and the one I use most. A digital DC voltmeter on the side panel tells me exactly how much runtime is left instead of making me guess. Dual USB right next to it, 1A and 2.1A, because halfway through a long shift the crew's dead phones matter more than the music does.

The Small Stuff

Controls are marine grade illuminated rocker switches, the kind you'd find on a boat console. They're recessed into the side panel so the lid still closes flush and latches. Every wire run is labeled, fused, and dressed, because the version of me troubleshooting this thing in a parking lot in February won't remember what any of it was for.

It pairs as EXILE-BT. First time I fired it up in the shop and Spotify came through clean, I stood there grinning like an idiot for a solid minute.

Was It Worth It

About $850 in parts. That's real money, more than a stack of consumer speakers. But this one has been through jobsites, off grid camps, and a trip to the Dominican Republic, and I haven't replaced a single component. The math worked out a long time ago.

If you build one, don't skip the Dynamat and don't skip the vents. Everything else you can adapt to whatever case and parts you can get your hands on.

Full build breakdown with all 17 photos, the complete parts list, and the wiring detail is on the Pelican Boom Box build page.

Third post in the Godzilla El Niño series. See the original post (June 27) and the July update.

We Said Pay Attention. The Ocean Just Screamed.

So, in June we read the SSTs and said a monster was loading up in the Pacific. In July NOAA confirmed it, and we flagged the August discussion as the next real checkpoint. That discussion landed Thursday. Not one number came back softer. Or better, depending on how you feel about powder.

This is no longer a forecast discussion. This is NOAA telling us the Pacific has rearranged itself in a way they've almost never seen, and there's a 69% chance this El Niño exceeds every event in their records going back to 1950. Not a small chance. Not an outlier scenario buried in the tail of a probability chart. The most likely outcome is now a historic event.

Everything Moved in the Wrong Direction

Here's how fast this escalated. In our July post the Niño 3.4 anomaly was +1.2°C. July closed at +1.4°C. Niño 3 came in at +1.7°C. Out east, the Niño 1+2 index near South America climbed from +2.7°C to +2.9°C, with eastern equatorial Pacific anomalies running past +2.0°C. This isn't one hot patch anymore. The whole eastern and central tropical Pacific is moving together.

The probability of a very strong event went from 81% to over 90% for the autumn and winter window. For October through December specifically, it's 95%. That's not a forecast that hedges. That's a forecast that's telling you to get ready.

Ten Degrees Under the Surface

But the number that stopped me cold is under the water. In July the equatorial upper-ocean heat content was running about +2.3°C — a big fuel tank already. NOAA's August data goes further: the equatorial subsurface temperature index climbed through July, and anomalies reached +10°C at depth. The thermocline has been shoved down hard, and warm water is now sitting where cold water belongs. The vertical structure of the Pacific has been rearranged.

In excavation, we have a saying: once the grade breaks, you're committed. The Pacific's grade broke months ago. The warm water is deep, it's massive, and there's no mechanism to turn it around before winter. This thing runs on its own fuel now.

The Atmosphere Locked In

Here's the part that turned this from serious to locked. A major El Niño doesn't run on warm water alone. It needs the atmosphere to couple with the ocean, to form a feedback loop where westerly winds push warm water east, the warm water drives more thunderstorms, and those thunderstorms weaken the trade winds even further. NOAA's August report confirms that coupling is now fully in place.

Low-level westerly wind anomalies now extend from the western Pacific into the east-central Pacific. Tropical convection has shifted east, away from Indonesia and into the central and eastern basin. The Southern Oscillation Index dropped to nearly two standard deviations below normal. The atmosphere isn't just responding to the ocean anymore. It's reinforcing it. It's like watching a loaded excavator on a slope: once it starts sliding, the weight feeds the motion.

What 69% Historic Means

NOAA uses RONI, the Relative Oceanic Niño Index — Niño 3.4 measured against the tropical ocean's own average instead of a fixed baseline. It exists because the whole tropical belt has warmed, which quietly inflates the traditional index. A three-month RONI of +2.5°C or higher in October through December would beat every El Niño NOAA has on record back to 1950 — past 1982-83, past 1997-98, past 2015-16. NOAA just put a 69% probability on exactly that.

That's not an edge case. That's the most likely bucket in the forecast. Weak and moderate outcomes have essentially disappeared from the late-year outlook.

What This Means for Tahoe and the Sierra

When Niño 3.4 pushes past +2.0°C, the subtropical jet gets shoved south and locks onto California like a fire hose. In 1982-83, Norden near Donner Pass recorded nearly 804 inches. In 1997-98, Squaw Valley got 68 inches in a single 24-hour period. Those two peaked at +2.2°C and +2.4°C. We're sitting at +1.4°C in mid-August and El Niño doesn't peak until late fall. The runway in front of this thing is the part that gets my attention.

If NOAA's 69% scenario plays out and this thing peaks above +2.5°C on the RONI scale, we're looking at conditions that exceed the template for every big snow year in living memory. Five hundred inches at elevation would be conservative. Six hundred or more is on the table. The ski season could start in October and run through July.

I grew up on the Squaw Valley Ski Team and carried the Jamaican flag in Vancouver. I've skied everywhere. And I have never seen a setup this aggressive, with this much subsurface fuel, this far ahead of winter.

Get Ready

The next NOAA diagnostic discussion is scheduled for September 10. Season pass secured. Avalanche gear serviced. Real chains for the truck. Stock the pantry deep. If you're a backcountry skier, take the AIARE course, because a season like this is an avalanche season first and a powder season second.

I run a crew at 6,000 feet. I know what a big winter does to a schedule. I also know what a 600-inch winter looks like from a chairlift, and it's the reason I live here. This one's coming, and the Pacific just told us exactly how big.

Tracking this since June. Next update when NOAA's September 10 data drops. The monster isn't slowing down.

Somebody on the crew asked me this week when I start training for ski season. I laughed, because the question assumes there's a version of my year where I stop. There isn't. The offseason is a lie. There's just the part of the year where the training doesn't look like training.

Back in the ski cross days it was structured — gym blocks, testing weeks, dryland camps, a coach with a clipboard telling me my vertical jump was down two centimeters and acting like somebody died. By the time I got to the start gate in Vancouver in 2010, I knew exactly what my body could do because we'd measured all of it. That's what a national team buys you: certainty.

Now my clipboard is a grade rod. And here's the thing nobody tells you about trench work at 6,000 feet: it's a better base program than most gym memberships. Ten hours of climbing in and out of a cab, checking grade in a hole, moving pipe, shoveling the corners the machine can't reach — none of it is cardio on paper, and all of it adds up.

My legs go into November stronger than they came out of April. The mountains don't care how you built the engine, just that it runs.

But base isn't everything, and forty-year-old knees keep a ledger. So the deliberate work starts now, about ninety days out.

Weighted hikes up whatever ridge is closest. Step-ups and single-leg squats, because skiing is a one-leg-at-a-time sport no matter what it looks like. Ankle and hip mobility, which is the least glamorous thing I do and the reason I can still get low in a turn.

At this age the game isn't power — it's showing up to the first icy groomer with tissue that's ready for it. Your knees don't forgive a cold start the way they did at twenty-five.

Ski cross taught me the rule that still runs my life: you don't rise to the occasion, you sink to your preparation. That was true in a start gate with three guys who all wanted the same line, and it's true on a jobsite, and it'll be true on the first chair in November.

And there's a reason I'm starting the clock this week. Everything I've been tracking says this El Niño is still loading up offshore, and if it delivers even close to what the pattern suggests, this winter is going to be one you want legs for. Deep days don't wait for you to get fit. Ninety days. Clock's running.

The plans said dirt. The plans always say dirt. Then partway into a trench this week the machine stopped mid-pull — not slowed, stopped — and everybody on the crew looked up at the same time, because we all know that sound. That's the rock lottery, and up here you play it every time you open the ground.

Here's the geology in one breath: Truckee sits on granite that's been rotting in place for a couple million years, and it doesn't rot evenly. Most of what we dig is decomposed granite — the flour-fine stuff that's currently eating my air filters — and it digs like heavy sand. But scattered through it are corestones: rounded chunks of intact granite the weathering never got to. There's no pattern to them. No plan sheet, no soils report, nothing tells you there's a boulder the size of a chest freezer sitting exactly at pipe grade. You find out when the bucket does.

The seat tells you first. The bucket's pulling a clean rhythm and then the stick loads up and the whole machine leans into it. New operators keep digging at it, and that's how sticks get bent and teeth end up somewhere in the spoil pile. The right move is the opposite: back off, work around it, expose the edges. You can't make a decision about a rock you can't see.

Once it's exposed, it's just math. If it weighs less than the machine wants to lift, you thumb it and walk it out — that's what's happening in the photo up top. Too big for that, and it's breaker time, or if spec allows, you over-excavate beside the trench and bury it below grade — dig the rock a grave and move on. What you never do is fight a boulder that outweighs your counterweight. The rock has all day. You don't.

The crew part is my favorite. The second a big one starts showing, the size calls start flying. Everybody's a geologist until the thing actually comes out of the hole, and it's always bigger than the low guess and smaller than the loud one. Closest call wins nothing except being right, which on a crew is worth more than money.

Rock time is also why honest schedules up here carry padding. We get five months to do a year's worth of digging, and the lottery doesn't care. A boulder handled right costs you an hour. A boulder handled wrong costs you a stick, a schedule, or worse. Play it patient, and the ground usually lets you win.

So, it hasn't rained in Truckee in weeks, and everything on my sites — the iron, my truck, my lungs — wears the same coat of fine gray powder. Everybody's watching the winter that's coming. I get it. But snow mostly parks a machine. This powder kills them while they work. Dust season.

Most of what we dig up here is decomposed granite. Run a track machine across it in late July and it grinds down into a grit that gets into everything, and unlike valley topsoil, this stuff is basically lapping compound. Sharp, hard, doesn't dissolve, doesn't wash out, doesn't quit. Here's where it does its damage and what we do about it.

Air filters. At 6,000 feet your diesel is already down on air before dust ever shows up — I covered what altitude takes from a machine earlier this summer. A choked filter on top of thin air and you're paying for a derated machine at full rate. Watch the restriction indicator daily, and here's the part that surprises people: a slightly dirty filter actually filters better than a fresh one.

Don't pull it to eyeball it — every time you drop that filter you dump grit straight down the intake. And don't hit it with compressed air. You're not cleaning it, you're blowing holes in the media you can't see. Replace it when the indicator says so, not before, not after.

Cooling. Those same fines pack into radiator and oil cooler cores a few thousandths at a time until one 95-degree afternoon the machine derates on the far side of the site. Blow the cooling package out weekly — opposite the direction of normal airflow, in the morning while the cores are cool. Thin air already moves less heat. Don't give the mountain help.

Undercarriage. The most expensive consumable on a track machine, and abrasive fines are what eat it. Two habits pay for themselves: run track tension correct for the ground — tight tracks in abrasive fines chew through bushings and sprockets at double speed — and clean the packed material out of the frames at the end of shift, not the start. It sets up overnight like weak concrete.

Grease. Dust plus grease equals grinding paste, right in your pins. Wipe every zerk before the coupler touches it, then grease daily so fresh grease purges the contaminated stuff out of the joint. Thirty seconds a fitting. Cheapest insurance on the site.

Why this matters right now: if the models are right about this winter, every machine on the mountain needs to roll into November healthy, because there won't be a quiet week to fix anything until April. The machine that dies in February almost always started dying in July. Take care of it now, while the dying is still optional.

So, it's the middle of July, the lake is full of paddleboards, and I'm writing about a heater. There's a reason. If you read last week's post, you know NOAA just confirmed the El Niño everybody's been watching. The winter that's coming doesn't care what the thermometer says today, and winter gear built in January is gear built in a panic. July is when you build it right.

Here's the problem this thing solves. Winter jobsites up here drop below zero, and crews need heat somewhere — a conex, a cure tent, an enclosure over fresh concrete. The default answer is propane, and propane in an enclosed space is asking for trouble. It's a combustion source pumping moisture and worse into the air you're breathing, the bottles are heavy, and they sulk when it gets seriously cold. I wanted heat I could trust, move with one hand, and forget about for a full shift.

The heart of the build is an HCALORY TB-MAX diesel heater. These Chinese diesel heaters have been keeping truckers and vanlifers alive for years, and the design is solid: sealed combustion, so the burn air and exhaust never touch the air you're heating. It sips fuel and draws about 40 watts running — one to four amps at 12 volts, with a brief 12-amp spike at ignition. That low draw is what makes the rest of the build possible.

The heater lives inside a weatherproof steel tongue box with sealed flexible ducting out the side, so you park the box outside and duct warm, dry air to wherever it's needed. Power comes from two 120Ah LiFePO4 batteries in their own box — 240 amp-hours total. Lithium iron phosphate matters here: you can actually use 80 to 90 percent of the capacity without hurting the batteries, where lead acid taps out around half. Do the math on a 40-watt draw and the 20-plus hours I'm claiming is the conservative end — at that rate the pack's got days in it. Overnight concrete cure? Covered, with margin.

Everything connects with Anderson quick-disconnects, and the charging system rides in its own orange Apache case. That's the part most people skip and regret. Loose chargers and bare cable ends don't survive a winter of truck beds and snowbanks. Give every piece of the system its own sealed home and it'll still work in February.

Total build cost: $1,047.35. For that you get a heat source rated to 20 below, no cords to a generator, no propane bottles to babysit, and clean dry air instead of combustion moisture fogging up the enclosure. It's been battle-tested on real winter sites and it earns its spot on the trailer every single cold morning.

Full parts list, specs, and every build phase in photos are over in the DIY section. If the models are right about this winter, build yours now. The mountain doesn't wait for you to be ready.

UPDATE to the original Godzilla El Niño post (June 27, 2026)

NOAA Put It in Writing

So, back in June the ocean was already telling on itself. SSTs climbing, warm water rolling east under the surface, all 24 models leaning the same way. We published the Godzilla El Niño post and told you to pay attention.

NOAA just made it official. They've issued a formal El Niño Advisory, meaning conditions aren't developing, they aren't emerging, they are present. Right now. The tropical Pacific has crossed the threshold and there's no ambiguity left. The monster is real.

The Numbers Have Gone Vertical

When we wrote the original post, the chance of a very strong El Niño was 63%. Here's where the probabilities stand now:

The Niño 3.4 SST anomaly is at +1.2°C above average and climbing. The easternmost Niño-1+2 index has already hit +2.7°C. All 24 IRI/Columbia forecasting models are in strong agreement: this event is strengthening, not fading. Weather.com is calling it one that could "rank among the largest events in historical record since 1950."

"Very strong" El Niño means the Niño 3.4 index exceeds +2.0°C above average. NOAA has only recorded a handful of events that strong since 1950. Models now project we cross that threshold in the October-December window, right when the winter storm track starts paying attention.

The Kelvin Wave: The Engine Under the Surface

Here's what most weather coverage misses. Beneath the surface, a downwelling Kelvin wave has spent months shoving warm water east across the equatorial Pacific, and the fuel gauge shows it: upper-ocean heat content is running about +2.3°C above average across the eastern half of the basin. That warm water is surfacing now, feeding the SST climb. The engine's running and nobody can shut it off.

Every Model Agrees: This Is Going Up

Look at the forecast plume below. Every spaghetti line lives in El Niño territory through winter. The CFSv2 median peaks just above +2.0°C in the October-December window, over the very-strong line, with the warm end of the ensemble pushing +2.5°C. The NMME and IRI multi-model runs tell the same story: this event strengthens through fall and holds deep into 2027.

The Big Three Might Get Company

Let's put this in context. The three biggest El Niños in the modern record peaked at +2.2°C (1982-83), +2.4°C (1997-98), and +2.6°C (2015-16). The July runs put this one's median right at the bottom of that club, with the warm end of the ensemble near +2.5°C. If the warm end verifies, this event pulls up a chair with those three.

And we're stacking this on top of the hottest ocean baseline in recorded history. The equatorial Pacific hit 26.99°C in April. Those model numbers are relative anomalies — the same reading means more actual heat than it did in 1997. The monster's swimming in a warmer ocean.

What This Means for Tahoe: 400-600+ Inches

We said it in the original post: regular El Niños are a gamble for Tahoe. Monster El Niños deliver. With an 81% chance of very strong by the October-December window, we're firmly in monster territory. Here's what history tells us about seasons like this:

1982-83: Norden near Donner Pass recorded nearly 804 inches. Snowpack hit 15 feet by February. Atmospheric rivers stacked for weeks. It snowed every single weekend from January 15 to May 15.

1997-98: Snowpack exceeded 200% of average. Squaw Valley reported 68 inches in 24 hours on New Year's Day. Lake Tahoe rose 6.5 feet from runoff.

If the very-strong projections verify, those two seasons are the blueprint. We're talking about above-average precipitation, potential for 400-600+ inch snowfall seasons at elevation, extended atmospheric river events, storm cycles that stack for weeks, rain-on-snow at lake level, and pure powder at the peaks. Early-season storms could hit by late October. The ski season could run November through July.

What We're Watching Next

The August NOAA update is the next big checkpoint. If SSTs keep climbing at this rate through August, the October-December projections get locked in. We'll also be watching for the first atmospheric river setup of the fall. Historically, monster El Niños announce themselves with an early-season bomb before Halloween.

I carried the Jamaican flag at the 2010 Olympics. I've skied on every continent. I run heavy equipment at 6,000 feet for a living and ski 60+ days a year. I've been tracking this since March, and I've never seen models this aggressive, this unified, this early in the season.

The original Godzilla post has the full baseline data, the SST trajectory chart, and the historical snowfall comparisons. This update confirms what the ocean already told us: the monster is real, and it's bigger than we thought.

Get your season pass. Service your avalanche gear. Buy real tire chains. Stock the pantry. If you're in the backcountry, take the AIARE course. A 600-inch season is an avalanche season.

We'll keep tracking this. Next update when the August NOAA data drops (scheduled for August 13). Bookmark the page. The monster isn't slowing down.

So, the skis went into storage months ago and the whole town's acting like winter never happened. Bikes out, paddleboards on every Subaru, dust on the trailheads. That's the civilian version of the off-season, and there's nothing wrong with it. But racing taught me a different calendar, and I still run on it.

The Off-Season Myth

Ski cross is four people banging elbows down a course at highway speed. What decides those races isn't talent in December — it's what your legs did in July. Dryland season was the real season. Squats until your vision went sideways, trampoline work for air awareness, balance drills that made you feel like a toddler learning to walk.

By the time snow flew, the racing was mostly already decided. You were just finding out who did their summer homework.

Some summers we chased snow anyway — up on the glacier at dawn, lapping a salted course that turned to mashed potatoes by 11. Ski boots and sunscreen at the same time messes with your head the first time. You get used to it.

Training at 6,000 Feet, Then and Now

These days my off-season program looks different. Nobody writes it on a whiteboard and there's no trampoline. It's ten hours a day on a jobsite at 6,000 feet — in and out of trenches, humping pipe, walking grade all day in the thin air.

I used to pay a trainer to make me do step-ups with a weighted vest. Now the vest is high-vis and the step-ups are free. The altitude base I spent years building as a racer just quietly maintains itself, one shift at a time.

The Photo Up Top

That shot of me floating over Lake Tahoe is a Keoki Flagg frame from 2009. People see a photo like that and think it's one perfect moment. It's not.

It's hundreds of reps, a photographer who knows the light, and legs that spent all summer getting ready to be that calm in the air. The moment is winter. The work is now.

Your July Homework

If you ski, here's the honest version: the season you'll have in January is being decided right now. Hike something steep. Ride the bike uphill, not just down. Do the boring single-leg stuff. And check your gear this month — boots, bindings, edges — while the shops are empty and the techs are bored.

Winter always comes back to Tahoe. The only question is whose legs show up ready.

So, the Fourth just came through Truckee and did what it always does — doubled the population, tripled the boat traffic, and turned the grocery store into a contact sport. Great weekend. But if you dig holes for a living up here, the holiday is also a calendar alarm you can't snooze: the season is half gone.

The Season Math

Down in the valley, an excavation crew works more or less year-round. At 6,000 feet, the ground doesn't give you that. It stays locked up or soaked well into May, and the first real storm can land in October like it owns the place — because it does.

Call it five workable months, maybe six if the fall is kind. Whatever a valley outfit spreads across a year, we compress into that window. Every clean, dry, 75-degree day in July isn't just a nice day. It's inventory.

Holiday Week on a Crew

Here's the part nobody puts in the brochure: half the crew wants the Fourth off, and the ground couldn't care less about the calendar. So you rotate. Somebody takes the lake day with their kids, somebody covers, and next holiday it flips.

A crew that never gets a summer day with their family burns out by August, and a burned-out crew makes mistakes around open trenches. Keeping people fresh isn't soft — it's the cheapest safety program there is.

Night Moves

Some of the work only happens when the town is quiet. Utility crossings and conduit runs under pavement get done at night, under light towers, when the road can actually be closed without turning the whole grid into a parking lot. There's something clean about night work — no traffic, cool air, just the sweep of the conduit dropping into the pull box exactly the way the drawing said it should.

The Screen That Keeps Us Honest

The other half of the sprint is not wasting motion. Machine control has changed what a short season can hold. The screen in the cab shows cut and fill in real time against the design surface — no waiting on stakes, no re-checking grade three times, no digging it twice. When you've only got five months, the hours that screen saves aren't a convenience. They're the season.

By October the trenches will be closed, the pavement patched, and the machines staged for snow. Winter always collects up here. The whole job in July is making sure that when it does, you don't owe it anything.

So, nobody warns you about this when you move your operation up into the Sierra — the mountain takes a cut off the top of everything. Your engine, your crew, your patience. Every machine I run sits around 6,000 feet, and the thin air doesn't care what the spec sheet promised down at sea level.

The Engine Math

A naturally aspirated diesel gives up roughly three percent of its power for every thousand feet of elevation. Run that math at 6,000 and you've handed back close to eighteen percent before you turn a bolt. That's the difference between a bucket that breaks rock and a bucket that polishes it.

The turbocharged machines claw most of it back — the turbo crams the thin air down to something the cylinders still recognize. The little naturally aspirated stuff just runs leaner and works harder for the same hole. You feel it in the hydraulics on a hot afternoon.

Heat Is the Other Half

Thin air cools worse than thick air. Same radiator, less air mass moving through it, so coolant and hydraulic temps climb faster than they ever did down in the valley. Up here I keep the coolers blown out clean and I watch the gauges in July like a hawk. A clogged cooler at sea level is an annoyance. A clogged cooler at 6,000 feet in the afternoon sun is a tow bill.

The Ground Fights Back

Then there's what you're digging into. The Sierra is granite, decomposed granite, and the occasional boulder the size of a chest freezer that the plans swore wasn't there. Teeth wear fast. You learn to read the ground by feel through the sticks — back off before you stall the machine, reposition, take a smaller bite. Patience saves more steel than horsepower does.

Cold Mornings, Short Windows

We froze at the tail end of June this year. That tells you everything about the season up here. Block heaters earn their keep, you let the hydraulics warm before you ask the machine for anything real, and you respect the cold start.

The dig season is short, so when the ground's open, you move. Alarm at 5:30, truck by 6:15, boots on the jobsite by 7, off by 3:30 if the day cooperates.

The Crew Runs on Thin Air Too

Don't forget the people. Altitude pulls the water right out of you, and the sun up here is brutal because there's less atmosphere standing between you and it. New guys gas out by lunch their first week and figure they're out of shape. They're not. They just aren't acclimated yet. Water, shade, sunscreen, and a hard hat that actually stays put.

None of this is in the operator's manual. That book was written for sea level. Out here you learn the mountain's edits the hard way, then you build them into the plan. Respect the altitude and it'll let you work. Ignore it, and it'll cook a machine just to prove the point.

The Setup

The Pacific Ocean is the hottest it's ever been measured. The equatorial Pacific hit 26.99°C in April 2026, blowing past the previous record. On June 11, NOAA issued an El Niño Advisory. Two weeks later the Niño 3.4 anomaly is at +0.9°C and still climbing. That's not a gradual warming. That's a rocket ship.

All 24 IRI/Columbia forecasting models agree El Niño will strengthen through the fall. NOAA puts the probability of El Niño during December 2026 through February 2027 at 96%. The chance of a very strong event: 63%. The CFSv2 forecast plume has the median peaking just above +2.0°C in the October-December window, with the warm end of the ensemble near +2.5°C. The previous record holders peaked at +2.6°C (2015) and +2.4°C (1997). If the warm end verifies, this event sits at the same table as both of them.

A downwelling Kelvin wave is moving east across the equatorial Pacific, dragging a deep pool of warm water with it — and NOAA's upper-ocean heat content charts are climbing right alongside the surface numbers. That's the engine no one can shut off. We're stacking a monster El Niño on top of the hottest ocean baseline in recorded history.

What It Means for Tahoe

I'll be honest: the correlation between regular El Niño and Tahoe snowfall is historically weak. Of 27 El Niño winters on record, only 15 produced above normal snow. But the two verified very strong events, the true monsters of 1982-83 and 1997-98, both delivered exceptional Sierra snow. When the SST anomaly pushes past +2.0°C, the jet stream overwhelms the geographic uncertainty and funnels moisture directly into California. Regular El Niños are a gamble for Tahoe. Monster El Niños deliver.

The Numbers

We just limped through 348 inches. Below average. Now look at what the monster years produced. In 1982-83, Norden near Donner Pass recorded nearly 804 inches. Snowpack hit 15 feet by the end of February. In 1997-98, snowpack exceeded 200% of average and Squaw Valley reported 68 inches in 24 hours on New Year's Day.

SST Trajectory

The Niño 3.4 anomaly was negative as recently as March. It crossed the +0.5°C El Niño line in June and hit +0.9°C by the last week of the month. That rate of change is aggressive.

The previous record El Niños peaked at +2.42°C (1997) and +2.57°C (2015). We're at +0.9°C in June with five months of strengthening ahead. Hold that curve through fall and we're knocking on the all-time record.

Ski Season Forecast

If this event reaches +2.0°C, the historical analogs point to a 600 to 700+ inch season at Palisades Tahoe. Nearly double the average. Early season storms could hit by late October. We could be skiing into July. In 1982-83, it snowed every single weekend from January 15 to May 15.

Expect stacked atmospheric rivers, rain on snow events, 100 mph ridge winds, and avalanche cycles that run for days. Lake Tahoe rose 6.5 feet in 2016-17. A stronger event pushes that higher. This isn't just a big winter. It's a force of nature.

Get Ready Now

Season pass, gear check, real tire chains, a stocked pantry, and a roof rake. If you ski the backcountry, get your avalanche gear serviced and take the AIARE course. A 700 inch season is an avalanche season.

I carried the Jamaican flag at the 2010 Olympics. I've skied every type of snow this planet produces. I run a CAT 299 for a living and ski 60 plus days a year. When they start talking Kelvin waves, I start checking my boot liners.

The mountains looked tired by March. Now the ocean is running a fever nobody alive has seen: the warmest equatorial Pacific ever measured, a Kelvin wave feeding it from below, and every model on the board pointing the same way.

Respect the mountain. Get ready. See you out there.

This is a developing story. I'll be tracking the models, the SSTs, and the Kelvin wave all through the fall. Updates coming as the data evolves. If this thing stays on track, we're looking at a genuine Godzilla El Niño. Bookmark this page.

So, it's June 24th and I'm writing about a freeze warning. Let that land for a second. Most of the country is sweating through triple digits right now, firing up the AC, complaining about humidity. Up here at 5,900 feet in Truckee we're about to watch the thermometer fall 47 degrees in three days — 77° midweek to 30° by Sunday morning. Thirty. Below freezing. In late June. If you've lived in the Sierra long enough, this doesn't surprise you. If you just moved here, welcome. The mountain doesn't read the calendar.

Here's what the models are showing. Friday brings the front door. Sustained winds out of the west-northwest climbing through the afternoon, gusts to 34 mph by evening. That's enough to blow tarps off a grade, send unsecured plywood sailing, and make any crane work a hard no. Saturday the upper-level trough digs in and drags cold Pacific air right over the crest. We're looking at up to a half inch of snow above 7,500 feet. Snow. On June 27th. The high stays right around 50, which means if you're working in a trench all day you're reaching for a hoodie by two o'clock. Then Sunday morning the bottom drops out. The snow shuts off, the sky clears, radiational cooling does its thing, and Truckee wakes up at 30°F.

The daily temperature swings this weekend are borderline absurd. We're talking 30 to 35 degree spreads between the overnight low and the afternoon high. You'll start the morning scraping frost off windshields and end the afternoon in a t-shirt. That kind of swing is hard on people, hard on equipment, and especially hard on fresh concrete.

For construction crews, this isn't a novelty forecast. This is a work-stopper if you're not paying attention. Concrete poured Thursday or Friday needs protection. If the surface temperature of a fresh pour drops below 40°F in the first 24 hours, you're looking at reduced strength, surface scaling, and potential microcracking that doesn't show up until months later. Blankets go on. Curing compounds go down. No exceptions.

Any exposed waterlines that haven't been buried and backfilled need to be drained or heat-traced by Friday afternoon, because a 30°F morning will freeze an exposed half-inch copper line before the sun gets to it. Fresh grading work is at risk too. Frost heave on an uncompacted subgrade can push your elevation a quarter inch overnight, and that quarter inch is the difference between passing and failing a grade check on Monday.

Our crew runs a GO/NO-GO system for exactly this kind of event. Every morning the 8-model weather engine pulls forecasts from GFS, NAM, HRRR, ECMWF, and four other sources, stacks them against each other, and flags any day where conditions cross a threshold. Wind above 25 sustained is a yellow flag. Freezing temps on active pour work is a red flag. The system doesn't make the call for us, but it puts the data in front of the foreman before the crew leaves the yard. This weekend popped a red flag on Wednesday morning, three full days before the trough digs in. That's three days to cover pours, drain lines, and reschedule any wind-sensitive work. Three days because the models agreed early and the engine caught the convergence.

For the community, the playbook is simple but it matters. Bring the pets inside Saturday night. Cover any tomato plants, pepper starts, or anything else you optimistically put in the ground in early June thinking summer was here. It wasn't. Disconnect and drain garden hoses. If you have exposed exterior plumbing on the north side of the house, a couple wraps of pipe insulation and twenty minutes of work saves you a burst pipe and a Sunday morning you didn't plan for. And if you're driving over Donner Pass this weekend, check conditions before you go. Snow at elevation combined with gusts pushing 36 mph means chain controls aren't impossible, and I've seen CHP close the summit for less.

If you want to track this front in real time as it moves through, the full dashboard is live at errolkerr.com/weather. Eight models, hourly resolution, wind and temperature overlays, all updated every six hours. Built it from the cab of a truck on a jobsite because the existing forecasts weren't cutting it for field decisions, and events like this weekend prove exactly why.

The Sierra giveth and the Sierra taketh away. Last week we had 15 hours of daylight and 80 degree afternoons. This weekend we might see frost on the windshield and snow on the ridgeline. Same mountain. Same town. Different week. If you're not ready for that range, you're not ready for Truckee. Stay warm out there.

So, back in late 2017 I bought a plain 6x12 enclosed cargo trailer. White aluminum box, diamond-plate floor, three clearance lights, and one dome bulb bright enough to find your own hand and nothing else. I didn't want another pile of tools to load and unload every morning. I wanted a shop that drives to the site. So I built one, and eight years later it's cut every two-by-four on every job I've run since.

Here's the lesson that matters most, and it lands before you ever touch a saw: wait. That trailer sat in my driveway for three months before the first cut. Three sets of sketches, two material lists, one full rebuild of the plan. A cargo trailer punishes hasty decisions. Drill the wrong stud, mount a saw at the wrong height, run your wires in the wrong order, and you pay for it for the next decade. The patience up front is the build.

Then the heaviest thing goes in first. I dropped a loaded Kobalt tool chest into place before I framed a single wall, because once you build cabinetry around 280 pounds of steel you can't pull it out again without tearing the whole cabinet down. Set your anchor, square your bench height off it, and build everything else to that line.

One thing nobody warns you about: the interior plywood on these trailers won't hold a lag bolt. It's stapled to thin metal channels behind the wall. Find those channels with a stud finder, every sixteen inches, and cleat your framing into the steel. Skip that step and your nice new cabinet rips off the wall the first time you hit a pothole on the way to the grade.

The phase I'm proudest of is the wiring, which is exactly where most trailer builds die. Somebody runs wires straight off a battery with no fuses, no labels, and no plan, then spends ten years wondering why nothing works. I did mine like a boat. Marine-grade tinned copper, every circuit on its own breaker and its own switch, and an eight-channel lit rocker panel so I can see at a glance what I left running. Tinned wire costs about double what regular automotive wire does and lasts about ten times as long. Trailers get damp. Spend the money.

Power runs two ways. Plug into shore power if the site has it, run off a deep-cycle battery if it doesn't. A shunt on the battery negative tells me exactly how much I have left instead of guessing, and guessing is what kills batteries. The saws sit on platforms aimed straight out the open ramp, so I can rip sixteen-foot stock without dragging anything outside. That single design choice drove the entire layout.

Three weeks of nights and weekends got me a finished shop on wheels. Eight years on, it's just part of the truck. I documented the whole thing end to end, every phase and every part and every mistake, over in the DIY section. Steal the whole build. That's what it's there for.

So, it's Prime Day week and I'm sitting in the cab of my truck on a construction site outside Truckee, hotspotted off my phone, pushing code through a cellular connection that would make most developers cry. And I'm thinking about feedback loops.

Here's the deal. Every product on my deals page is something I actually use or have vetted for field work. The DeWalt impact driver. The Klein multimeter. The boots. When you click through and buy something, a small commission comes back to me. Standard affiliate setup. But here's where it gets interesting.

That commission doesn't go into a vacation fund. It goes straight back into building this platform. Better hosting. More API calls for the weather engine. More time I can spend tuning the 8-model ensemble that tells my crew whether tomorrow is a GO or a NO-GO. The weather system on this site runs real forecasts for real crews making real decisions about whether to put people on a grade in the Sierra.

I built the entire weather engine from places like this. Tailgates, truck cabs, hotel rooms after 10-hour days in the dirt. Same mentality I had training for the Olympics. You don't wait for perfect conditions. You work with what you have and you make it count.

So when I say every click matters, I mean it literally. A drill bit purchase funds a weather API call that might keep a crew off a ridge during a lightning cell. That's the loop. That's why I'm out here building this thing from a construction site instead of waiting for a desk and a fiber connection.

The platform doesn't care where the code comes from. It cares that the code is right.

Click something. Build something. That's how it works.

So, we just came through the longest stretch of daylight of the year, and for once the sky played along. If you were outside in Truckee this past week you already know the drill: warm afternoons, cold mornings, and a couple of afternoons where the clouds loaded up like they meant business, then quit before dinner.

I check the forecast ensemble every morning before the truck rolls. Here's how the week actually went, and how the models did against it.

The Setup

A broad ridge of high pressure parked itself over the Great Basin for most of the week. That's the classic late-June pattern up here. High pressure means sinking air, sinking air means clear skies and warming temperatures, and clear skies at 6,000 feet means the heat bleeds right back out overnight. Daytime highs ran in the upper 70s to low 80s in town. Mornings dropped into the high 30s and low 40s. That 40-degree daily swing is normal for us, and it catches flatlanders off guard every single summer.

Models vs Reality

The 8-model ensemble nailed the big picture days out. Every member agreed on the ridge, agreed on the warming trend, and agreed the valley would stay dry. When the models all shake hands like that, you can plan a week of work with confidence. We did.

Where they argued was the middle of the week. A few members tried to break out afternoon thunderstorms over the crest, pumping up CAPE values as a little moisture sneaked in from the south. Two models went all-in on storms. Three kept it bone dry. The rest split the difference with isolated afternoon cells.

What actually happened was a classic Sierra tease. The cumulus built tall and dark over the Carson Range one afternoon, we heard a couple of rumbles up high, and then the whole thing fell apart by dinner without dropping much on the valley floor. The members that hedged toward isolated cells got it right. The two models screaming severe over-cooked it, which is exactly what high-resolution models tend to do with mountain convection. They see the instability and assume the worst.

Why It Matters on the Grade

This isn't just weather-nerd stuff. When the models disagree on afternoon storms, that's a GO with a watch, not a NO-GO. We worked, we kept an eye on the sky, and we had a plan to pull off the high ground if a cell organized. It never did. Reading the spread between the models is how you avoid both burning a workday for a storm that never comes and getting caught flat-footed by one that does.

Looking Ahead

The ridge holds into early next week. Expect more of the same: warm, dry, and breezy in the afternoons. The one thing I'm watching is a weak disturbance the long-range members hint at toward the end of the month that could drag monsoon moisture north and bump our thunderstorm odds back up. Long range, low confidence, but it's on the board.

Enjoy the long days while they're here. They don't last.

So, this video has been sitting on my YouTube channel since December 2013. Twelve years. Two hundred and change views. I never promoted it, never shared it, never really talked about it. But it's one of the most honest pieces of footage from that entire chapter of my life, and I think it's time people saw it.

The documentary is about twelve minutes long and it covers me and my crew during the ski cross racing years. This was the period where I was grinding through the circuit, chasing FIS points, trying to qualify for international competition, and doing it all with basically no budget and no federation behind me. Just a crew of people who believed in the mission and were willing to show up.

What You're Looking At

The footage opens on the National Brotherhood of Skiers website. The NBS has spent decades putting Black athletes on the U.S. team and getting them to the Olympics. They run the largest continuously operating ski convention in America. Night parties, competitions, fundraisers, youth camp, the whole program. Their Olympic Scholarship Fund helped make it possible for athletes like me to compete when nobody else was writing checks.

That NBS connection wasn't a footnote. It was a lifeline. When you're a Black ski racer trying to put together a campaign with no money and no infrastructure, the people and organizations who step up are the difference between racing and watching from the couch. The Brotherhood stepped up.

The Race Sheet

There's a shot in the documentary of a race results board. Handwritten. Old school. And if you look close you can see it: KERR, E. Highlighted. That's what it looked like. You'd show up to a race, check the start list taped to the timing shack, find your name, and that was your whole world for the next ninety seconds. Everything you'd been training for, every gate you drilled, every hill you hiked because you couldn't afford a lift ticket, it all came down to what happened between the start wand and the finish beam.

I loved that part. The simplicity of it. You're either fast enough or you're not. The clock doesn't lie. The clock doesn't care about your story or your background or how hard you worked. It just measures. And on the days when the number next to your name was lower than everyone else's, nothing in the world felt better.

The Crew

The title of this documentary is "Errol Kerr and Crew" and that word matters. Crew. None of this happened alone. You see people in this footage, on the mountain, at the races, in the parking lots, and every single one of them was part of making it work. Coaches who donated their time. Friends who drove twelve hours to hand me a pair of skis because mine broke. My mom, Catherine, who held the whole operation together from the administrative side while I was out there trying not to crash into a fence at 60 miles per hour.

You don't race ski cross solo. It's a team sport disguised as an individual one. Somebody waxes the skis. Somebody films the runs so you can study your line. Somebody makes the phone calls to get you registered. Somebody picks you up when you yard sale into the B-net and can't feel your left knee. That's the crew. And this documentary catches a piece of that.

The Soundtrack

We cut the whole thing to Alter Bridge and Michael Franti. Two Alter Bridge tracks off the Blackbird album and a Michael Franti and Spearhead track from Yell Fire. That was the music of that era for me. Alter Bridge is heavy enough to match the intensity of racing but melodic enough to carry a story. Myles Kennedy's voice over footage of the mountains just works. And Franti brought the soul. The positivity. The "keep going" energy that you need when everything around you says quit.

Music matters in a project like this. It tells you what the filmmaker was feeling. And what I was feeling was: this is hard, this is beautiful, and I'm not stopping.

Why Post This Now

I'm sitting in Truckee in 2026. I run excavation crews for a living. Lindsay and I have a baby on the way. I'm a completely different person than the guy in this documentary. But I'm also exactly the same person. The guy who shows up early, works until the job is done, surrounds himself with good people, and doesn't quit when it gets hard.

Ski cross taught me that. The crew taught me that. And this little twelve-minute documentary that's been collecting dust on YouTube for over a decade is proof of it.

So hit play. Watch the whole thing. See what it looked like before the Olympics, before the flag, before any of it made sense to anyone but us. It's raw, it's real, and it's twelve minutes of my life that I'm proud of.

Two hundred views in twelve years. Let's fix that.

So, let me take you back. Born in Flatbush, Brooklyn. April 12, 1986. My parents met in Jamaica — my mom Catherine was on vacation, met my father, fell in love, got married on the island. Jamaica was in my blood before I took a breath.

But I wasn’t in Brooklyn long. Before I turned one, my mother moved us to California. Lake Tahoe. The Sierra Nevada. And that’s where everything starts.

I grew up on the mountain. Not visited. Not vacationed. Grew up. Snow wasn’t some novelty I discovered later. Snow was my whole childhood. While kids in Brooklyn were playing basketball on concrete, I was standing at the top of a mountain watching the sun come up over the ridge. Somebody put me on skis when I was four, and that was it. Game over. The closest thing to flying without leaving the ground.

I joined the Squaw Valley Ski Team — Palisades Tahoe now — and that’s where I learned to race. Real racing. Gates, courses, the whole program. Same mountain as Daron Rahlves, Marco Sullivan, Travis Ganong. Squaw produced some of the best ski racers in the country, and I was right there in the mix. A Black kid on the ski team in the ’90s in Tahoe? I got the looks. But I wasn’t a tourist. I was a local. And the mountain doesn’t care what you look like. Gravity works the same on everybody.

I was also racing BMX and motocross, which sounds unrelated until you realize that reading terrain at speed and fighting for position in a pack is exactly what ski cross demands. I just didn’t know it yet.

The U.S. Ski Team

Then ski cross entered the picture. Four racers, one course, first to the bottom wins. No judges, no style points. Just racing. It combined everything — alpine technique from Squaw, terrain instincts from motocross, pack-racing mentality. The sport was built for me.

I made the U.S. Ski Team. My coach was Tyler Shepherd, who ran the U.S. skicross program. My rookie World Cup season, 2007-08: twelfth at Les Contamines. Eighth at Meiringen-Hasliberg. Fifth at X Games Buttermilk Mountain — in my first year. Ranked 24th in the world. Stars and stripes, World Cup starts, the whole dream.

And then I walked away from it.

For the Land of My Father

My father was Jamaican. He died when I was twelve. That changes a kid. I’d been wanting to ski for Jamaica for nearly a decade before it finally happened.

So here’s the decision nobody understood: I was a ranked member of the United States Ski Team. The best program in the world. Nobody walks away from that.

I walked away from it.

I teamed up with coach Carlos Guisado and built Jamaica’s ski program from scratch. No federation. No funding. No coaching staff. No equipment warehouse. You know Cool Runnings? The fun part is Hollywood. The reality is a hundred phone calls, letters to people who don’t write back, and figuring out FIS certification for a country that’s never had a ski team.

I earned every FIS qualifying point on courses in the Alps and Japan, sleeping in budget hotels, running a three-person operation while the Americans and Austrians showed up with fifteen-person crews. Tenth at the Skicross World Championships in Inawashiro, Japan. For Jamaica. A country where it’s never snowed.

Vancouver 2010

February 12, 2010. BC Place Stadium, Vancouver. Sixty thousand people. The Opening Ceremony of the XXI Winter Olympic Games.

I was the sole athlete representing Jamaica. Every other country walks in with a delegation. Jamaica walks in with one guy. Me. And they hand me the flag.

Standing in that tunnel, waiting for them to call “Jamaica,” I lost it. Everything I’d fought for. Every door that got slammed. Leaving the U.S. Ski Team. My father, who never got to see any of it. All of it at once.

Then the announcer says it. “JAMAICA.” And you walk out.

Sixty thousand Canadians losing their minds for Jamaica. I walked slow. Moments like that don’t come back. You hold the flag high and you walk tall.

Nine days later. Cypress Mountain. The first ski cross event in Olympic history — the sport had never been in the Games before. Eight seconds in the start gate. Pure adrenaline and focus. Then the gate drops, everything goes quiet in your head, and your body takes over. React. Absorb. Attack.

I finished ninth. Out of thirty-three of the best ski cross racers on the planet. The best result by a Caribbean athlete at any Winter Olympics. Ever. A kid from Flatbush, Brooklyn, who grew up on the Squaw Valley Ski Team, made the U.S. Ski Team, walked away to represent his father’s island, and finished ninth in the world.

The Mountain Doesn’t Stop Giving

I live in Truckee now. I run excavation crews — heavy equipment, underground utilities. Everything the mountain taught me goes straight into the work. Preparation. Composure under pressure. Respect for the terrain. The mountain can end you if you don’t respect it. A trench can do the same. First one there, last one to leave. That’s not a bumper sticker. That’s how you build something real.

I still instruct at Palisades Tahoe in the winters. The mountain that raised me as a racer, I give back to by putting the next generation of kids on snow. Lindsay and I are building our life here. A baby on the way. The same mountains that raised me, turned me into a racer, then an Olympian, then a foreman, and now a father.

There’s no golden ticket. There never was. You earn everything — every turn, every race, every flag you carry, every trench you dig. The mountain doesn’t care where you came from. It cares if you show up.

I keep showing up.

So, a TV crew came up this past winter to do a story on snow removal at the Village and found out the guy running one of the loaders used to race ski cross for Jamaica at the Olympics. That guy's me. The KCRA segment aired February 24, 2026, and the video above is the full piece. But the real story is what happens out there every time the Sierra decides to bury us.

I work snow removal at Palisades Tahoe Village for Ruppert Inc. When a big storm rolls in, and up here that can mean two to four feet overnight, we go to 24/7 operations. There's no waiting until morning. There's no "we'll get to it tomorrow." The Village has to be open. Guests are arriving, employees need to get to work, and the storm doesn't run on anyone's schedule.

The operation runs loaders, plows, and hauling trucks around the clock during storm cycles. We push snow out of the pedestrian areas, clear the access roads, open up parking structures, and haul the excess out to designated snow storage areas. At 6,100 feet, the snow is heavy and it piles up fast. You're not just plowing a dusting off the sidewalk. You're moving walls of snow taller than the loader cab.

The nights are where the real work happens. Most of the heavy clearing runs between about 10 PM and 6 AM, when foot traffic is lowest and we can move freely through the Village without dodging pedestrians. You're out there in the dark, running under lights, snow still coming down, pushing load after load. The loaders are running constant cycles, and the hauling trucks are making trips out and back all night.

What people don't realize about snow removal at elevation is the volume. A three-foot storm across the Village footprint is thousands of cubic yards of snow that has to go somewhere. You can't just push it to the side and call it done. It stacks up. After a few storms the piles are 15, 20 feet tall and you're running out of room. That's when the hauling trucks earn their keep, moving snow off site so you have somewhere to put the next round.

The cold is a factor too. We're regularly working in single digits, sometimes below zero with wind chill. Equipment gets sluggish. Hydraulics are slower to warm up. Your hands stop working right after a while even with good gloves. But the work doesn't stop. The storm doesn't stop, so we don't stop.

I've been doing snow removal up here for several seasons now, and every winter is different. Some years we get hammered early and it never lets up. Some years it's dry until January and then the sky opens up. But the job is the same every time: when the snow falls, we move it. Before the sun comes up, before the Village opens, before anyone walks out their front door, the crew has already been out there for hours making sure the place is passable.

That KCRA piece was a good look at what we do. Most people see the finished product, the cleared walkways and the open roads. They don't see the 14-hour overnight shifts, the frozen equipment, or the crew grinding it out in a blizzard. Now they have a better idea. And yeah, the Olympian thing makes for a good headline. But the work is the work regardless of what you did before you got here. The mountain doesn't care about your resume. It just wants the snow moved.

Compaction is the thing that separates a backfill job that holds up from one that settles, cracks the pavement, and turns into a callback six months later. I've seen crews dump material in a trench, run a plate compactor across the top, and call it good. That's not compaction. That's a warranty claim waiting to happen.

So here's what compaction actually means. You dig a trench, you put material back in, and that material is loose. Air voids everywhere between the particles. Compaction is you mechanically forcing those particles closer together, squeezing the air out, making the soil denser. Denser material is stronger material. It stays where you put it.

The number you hear on every spec is 95% compaction. That comes from a lab test called a Proctor test. A geotech takes a sample of your backfill material, adds different amounts of water, compacts it in a mold with a standardized hammer, and gets two numbers: maximum dry density and the optimum moisture content to reach it. When the spec says 95%, it means the soil in your trench needs to hit at least 95% of the max density the lab got under controlled conditions. Miss it and you're digging it back up.

There are two versions of the Proctor test. Standard Proctor uses a lighter hammer and is typical for utility trench backfill. Modified Proctor uses a heavier hammer and produces a higher maximum density. Make sure you know which one the spec references. Hitting 95% of Modified Proctor is significantly harder than 95% of Standard Proctor for the same material.

Nuclear Density Gauge is the most common field test. The gauge contains a small radioactive source that emits gamma rays into the soil. A detector measures how many rays pass through versus how many are absorbed. Denser soil absorbs more radiation. The gauge gives you wet density, dry density, and moisture content in about 60 seconds. It's fast, accurate, and the industry standard. The catch is the paperwork. You can't just buy one — the source is tracked, it gets leak-tested every six months, and in California the license runs through the state's radiologic health program rather than the NRC directly. That's why the testing firm owns the gauge and you don't.

Sand Cone Test is the manual alternative. You dig a small hole in the compacted lift, weigh the soil you removed, then fill the hole with calibrated sand from a jar and measure how much sand it took. The volume of sand equals the volume of the hole. Weight of soil divided by volume gives you density. It's simple, accurate, and doesn't need a radioactive source. The downside is that it takes 15 to 20 minutes per test, and you need a flat, clean surface to get a good seal with the base plate.

Plate Load Test is less common for trench work but shows up on large earthwork projects. A hydraulic jack applies load to a plate on the soil surface and measures deflection. It tells you the bearing capacity rather than density directly, but it gives a real-world measure of how the soil will perform under load.

The variable most crews underestimate is moisture content. Soil compacts best at its optimum moisture content, the number the Proctor test identifies. Too dry and the particles won't move together no matter how many passes you make with the compactor. Too wet and you're just pushing mud around. In Truckee, the native decomposed granite is usually dry enough that we need to add water. I run a hose into the trench between lifts and let the material soak for a few minutes before compacting. In spring, when snowmelt is running, the problem is the opposite. The material is too wet, and we have to let it dry out or swap it for imported granular fill.

Lift thickness is the other critical factor. A lift is a single layer of material that you compact before adding the next layer. For most trench backfill, the spec calls for 6 to 8 inch loose lifts compacted to about 4 to 6 inches. If you try to compact a 12-inch lift with a plate compactor, the bottom half won't reach density. The compaction energy dissipates before it gets to the bottom of the lift. Larger equipment like a sheepsfoot roller can handle 12-inch lifts, but you're not fitting a sheepsfoot in a 3-foot-wide trench.

I've seen compaction tests fail for three reasons, and it's almost always one of these: the lift was too thick, the moisture was wrong, or the operator didn't make enough passes. A jumping jack or plate compactor typically needs 3 to 5 passes over the same area to reach density. One pass isn't enough. I tell my crew to count their passes and make at least four on every lift. It takes an extra couple of minutes per lift, and it saves you from ripping the trench back open when the test fails.

When the testing firm shows up and your compaction fails, you have two options: add water and recompact, or remove the material and start the lift over. If it's a moisture issue, adding water and hitting it again usually works. If the lift was too thick, you're pulling material out. Either way, it's time and money. Get it right the first time.

So, every operator hits water eventually. You dig down four feet, five feet, six feet, and suddenly the bottom of the trench starts filling up. In Truckee the water table swings wildly — where you are, what month it is, how close you sit to the river or the creeks that feed it. I've dug trenches in July where the ground was bone dry at eight feet. Dug trenches in May where water was seeping in at three. You don't get to choose. You deal with what the ground gives you.

The simplest dewatering method is a sump pump in a pit. You dig the trench to grade, then dig a small pit at the low end, about two feet deeper and two feet wider than the trench. Water flows to the low point by gravity and collects in the pit. A submersible trash pump sits in the pit and pushes water out through a discharge hose. For low to moderate groundwater inflow, this is all you need. I keep a 2-inch and a 3-inch trash pump on the truck at all times during spring work. The 2-inch handles most seepage. The 3-inch is for when things get serious.

The key to sump pumping is keeping ahead of the water. If the pump can't keep up with the inflow, the water level in the trench rises, the bottom turns to soup, and you can't compact or lay pipe. I size the pump based on how fast the pit fills. A 2-foot by 2-foot by 2-foot pit holds about 60 gallons, so if it fills in 5 minutes, that's about 12 gallons a minute of inflow. A standard 2-inch trash pump moves 100 to 150 GPM, so it'll handle that with room to spare. If the pit fills in under a minute, you're looking at 60-plus GPM of inflow — and once hose runs and lift eat into pump capacity, that's when the 3-inch comes off the truck.

Wellpoint Systems are what you use when sump pumping isn't enough. A wellpoint system lowers the water table around the excavation before you dig. You drive a series of small-diameter well screens into the ground along the trench alignment, connect them with a header pipe, and attach the header to a vacuum pump. The pump pulls water out of the ground through the wellpoints, lowering the water table below your excavation depth. When you dig, the trench is dry.

Wellpoint installation is specialized work. The screens are typically 2 to 3 inches in diameter and 3 to 5 feet long, driven 10 to 15 feet deep depending on the required drawdown. Spacing depends on soil permeability. In sandy ground like the decomposed granite we see around Truckee, wellpoints every 3 to 6 feet will do the job. In silty soil the spacing gets tighter, not wider — the water moves slower, so each point only drains a small circle around itself. The header pipe connects all the wellpoints to a central pump, usually a diesel-powered vacuum unit that runs 24 hours until the excavation is complete and backfilled.

To size for flow, start with the permeability of the soil. Sandy gravel can transmit water at 1,000 gallons per day per square foot. Fine sand is closer to 100. Silt is 1 to 10. Clay is essentially zero, which is why you rarely need dewatering in clay. Multiply the permeability by the cross-sectional area of the excavation face that's below the water table, and you get the approximate inflow rate. It's not exact, but it gets you in the right ballpark for pump sizing.

Environmental Discharge is the part most crews don't think about until the inspector shows up. You can't just pump groundwater into the street. In California, construction dewatering discharge is regulated by the Regional Water Quality Control Board. If the water is clean groundwater with no contamination, you can usually discharge to the storm drain system with a permit. If the water has sediment, which it almost always does when you're pumping from a sump pit, you need to run it through a sediment control system before discharge. That can be as simple as a straw wattle filter or as complex as a settling tank and filter bag system.

On our jobs, I run the discharge hose into a sediment bag or a straw bale check dam before the water reaches any drainage inlet. The goal is to keep sediment out of the storm drain. If you're on a site with a SWPPP, the dewatering plan should already be in the document. If it's not, you need to add one before you start pumping. The fines for discharging turbid water to a waterway in California can run into the thousands per day. It's cheaper to buy ten sediment bags than to pay one fine.

Timing matters too. If you know the water table is high, plan your dewatering before you dig. Set up wellpoints or have pumps staged and ready. Don't wait until the trench is full and the pipe crew is standing around. On one job near Donner Lake, we started wellpoints two days before excavation. By the time we cut the trench, the ground was dry enough to work in without any sump pumping at all. That two-day lead time saved us three days of fighting water during the pipe installation.

The other option is to rethink the schedule. If you're doing sewer work in an area with a seasonal water table, and you have any flexibility on timing, wait until late summer when the water table drops. In Truckee, the difference between May and August water table levels can be three feet or more. That three feet can be the difference between a dry trench and a full-time dewatering operation. I plan wet-area work for July and August whenever the schedule allows.

Dewatering isn't complicated in concept. Water goes in, pump takes it out. But managing it well — sizing the right equipment, staying ahead of the inflow, keeping the discharge clean, planning for it before you're standing in a flooded trench — is what separates a crew that handles groundwater from one that drowns in it.

Every excavation job starts the same way. Before I fire up the machine, before the crew even grabs a shovel, I walk the site and read the paint. Those spray-painted lines on the ground aren't decoration. They're a map of everything down there, and if you can't read them, you have no business running a bucket.

The system is called the APWA Uniform Color Code, and it's the same across the entire country. Every color means a specific utility. Learn them once and you'll never second-guess what you're digging over.

Red means electric. Power lines, conduit, cables. Hit a live electric line with a steel bucket and the best case is a blown fuse on a transformer. The worst case is someone doesn't go home. I treat every red mark like it's hot until the utility company confirms it's de-energized. No exceptions.

Yellow means gas. Natural gas, oil, petroleum, steam. A gas line strike can evacuate an entire block. On our jobs in Truckee, the gas lines are usually polyethylene pipe buried about 18 inches. You won't see it until your bucket is already through it. That's why you hand-dig within 24 inches of any yellow mark. Every time.

Blue means potable water. Drinking water mains and service lines. Reclaimed water, irrigation, and slurry lines get purple — you'll see it on landscape jobs. A water main break isn't as dangerous as gas or electric, but it'll flood your trench, shut down the street, and cost you a day of production plus the repair bill. On mountain jobs, water lines can be deeper than you expect because of frost depth requirements.

Green means sewer. Sanitary sewer, storm drain, drain lines. Sewer is usually the deepest utility in the trench, often six to eight feet or more. It runs by gravity, so the alignment follows the slope of the ground. Puncture a sewer main and you're dealing with a hazmat situation, plus the repair, plus the fine from the district.

Orange means telecom. Fiber optic, cable TV, communication lines, conduit. Telecom damage is expensive in a different way. A single fiber cut can knock out service for thousands of customers. The carriers will send you a bill that makes your eyes water. Fiber is fragile and often in shallow conduit, sometimes only 12 inches deep.

White means proposed excavation. This is your outline, your dig boundary. The white paint shows the locator where you plan to work so they know which utilities to mark. If you call 811 and they ask where you're digging, the white paint is the answer.

Pink means survey. Temporary survey markings, property lines, section corners. These are reference points, not utilities. But don't blast through a survey monument. Those are legal markers and replacing them requires a surveyor and a lot of paperwork.

Before every dig, you call 811. That's the law in every state. You call, you give them the address and the scope, and the utility companies send locators to mark their lines. It takes two to three business days in most areas. I've seen guys try to skip this step to save time. It never saves time. One unmarked gas line hit will cost you more than every 811 call you'll ever make combined.

What happens when you hit an unmarked line? First, you stop. Don't move the bucket, don't pull it out, don't try to patch it yourself. For gas, evacuate the area and call 911. For electric, stay on the machine and wait for the power company. For water and sewer, call the utility district. For telecom, call the carrier.

In every case, document everything. Photos, measurements, the 811 ticket number, the location of the markings relative to the strike. If the line wasn't marked or was marked in the wrong location, the locating company bears the liability, not you. But only if you followed the rules. No 811 ticket, no protection.

I've been running excavation in Truckee for years now, and I still walk every set of markings before the first cut. I've caught mismarks, missed lines, and markings that were three feet off. The five minutes it takes to walk the site and compare the marks to the plans has saved me from strikes that would have shut down the job. Read the paint. Every single time.

I've watched operators burn an entire shift fighting the wrong bucket. A 36-inch GP bucket on a trenching job where they needed an 18-inch. A smooth-edge cleanup bucket trying to rip through decomposed granite. The bucket is the business end of the machine. Pick wrong and the machine fights you on every cut.

The GP bucket, general purpose, is what most machines wear as their daily driver. All-rounder. Medium width, usually 24 to 36 inches on a mid-size excavator, with teeth for penetration and a curved profile for good rollback. General site work, loading trucks, digging footings, moving mixed material, the GP handles it. Not the best at anything but good enough at everything. On my JD 35G, the 24-inch GP lives on the machine about 60 percent of the time.

The trenching bucket is built for one thing: cutting a narrow, clean-walled trench. Usually 12 to 18 inches wide, deeper profile than a GP, with a flat bottom and straight sides. When I'm laying pipe, running conduit, or setting utilities, this is the bucket. The narrow width means less material to move, less spoil piled up on the side, and a tighter trench that's easier to backfill and compact. On utility work, a trenching bucket can double your production compared to digging the same trench with a GP.

The cleanup bucket, also called a ditch cleaning bucket or grading bucket, is the wide, flat, smooth-edge bucket you use for finish work. No teeth, just a bolt-on cutting edge. Widths run from 36 inches up to 60 inches or wider. This is what you use to grade the bottom of a trench to a precise elevation, clean out a ditch, or do final grading on a pad. The smooth edge gives you a flat, even cut. On GPS-controlled machines, a cleanup bucket paired with the design surface is how you hit grade to a tenth of a foot without a single grade stake.

The rock bucket is the heavy hitter. Reinforced sidewalls, heavy-duty teeth, sometimes ripper shanks on the back edge. When I'm digging in the Sierra granite, the decomposed stuff that turns into solid rock about four feet down, this is what goes on the machine. The teeth are heavier and sharper so they bite into rock instead of bouncing off it. A rock bucket weighs significantly more than a GP bucket of the same width, so you give up some lift at full reach, but you gain the ability to rip through material that would fold a standard bucket.

The V-bucket, or trapezoidal bucket, is purpose-built for digging V-shaped drainage ditches and channels. The angled sides match the designed slope ratio of the ditch, usually 2:1 or 3:1. Instead of cutting the ditch in multiple passes with a standard bucket, one pass with a V-bucket gives you the full profile. We use them on road drainage work and anywhere the plans call for a shaped swale.

Bucket width drives production more than most operators realize. Go wider and you move more material per cycle, but you need more breakout force and the machine works harder. Go narrower and each cycle is smaller, but the machine cuts easier and you use less fuel. The sweet spot depends on the material. In soft clay, go wide. In rocky ground, go narrow. On a typical utility trench job, I match the bucket width to the pipe diameter plus clearance. Eight-inch pipe gets an 18-inch bucket. Twelve-inch pipe gets a 24-inch. No point digging a trench wider than you need to, because you just have to backfill and compact all that extra.

Quick-couplers changed the game. Before couplers, changing a bucket meant pulling pins with a hammer and a pry bar, 15 to 20 minutes of crawling around under a boom. Now I swap buckets in under a minute without leaving the cab. Hydraulic quick-couplers let you go from a trenching bucket to a cleanup bucket between cuts. That flexibility means I carry three or four buckets on the trailer and match the tool to each phase of the job. Mass excavation with the GP, trench with the trenching bucket, grade with the cleanup, all in the same shift.

One thing I see guys overlook is bucket maintenance. Teeth wear down and lose their self-sharpening profile. A dull tooth means the machine works harder for every cut, burns more fuel, and puts more stress on the linkage. I check teeth daily and replace them before they're worn to the adapter. Cutting edges on cleanup buckets take a beating too. A rolled or chipped edge means you can't hold grade. Replacing a cutting edge costs a couple hundred dollars. Re-grading a pad because your edge was shot costs a couple thousand.

Pick the right bucket, keep the teeth fresh, and the machine does the work. Fight the wrong one and you'll wonder why the job's behind schedule.

So, I've been building at altitude for years now. Running excavation jobs in Truckee at 5,800 to 6,200 feet, and the elevation touches everything. Equipment. Materials. People. Schedule. If you've only built at sea level, mountain construction will humble you fast. The rules are different up here, and the mountain doesn't care about your timeline.

Start with the iron. A diesel engine loses roughly three percent of its power for every thousand feet of elevation above sea level. At 6,000 feet, that's close to 18 percent. A 75-horsepower excavator at sea level is running about 62 horses up here. You feel it in the hydraulics. Breakout force drops. Cycle times slow down. The machine works harder for every bucket.

Newer Tier 4 engines with turbochargers compensate better than the older naturally aspirated stuff, but you still notice it. I plan for about 15 percent less production on any machine compared to its sea-level specs. If you're bidding a mountain job using flatland production rates, you're going to lose money.

The thin air hits the crew too. Workers coming up from Reno or Sacramento, even just from 4,500 feet, can feel the difference at 6,000. The first few days are the hardest. Guys get winded faster, dehydrate quicker, and fatigue hits earlier in the shift. I keep extra water on the jobsite and watch for signs of altitude sickness in new crew members. Headaches, nausea, dizziness. It's rare at 6,000 feet but it happens, especially when guys are working hard in the sun. Acclimation takes three to five days. By the second week, most people have adjusted.

Concrete is its own challenge at elevation. The lower air pressure means water evaporates faster from the surface, so plastic shrinkage cracking is a real concern. You have to cure more aggressively up here. Wet curing, curing compound, or both.

Temperature is the other factor. Truckee can see 90 degrees in July and freezing temps by September. Concrete needs to stay above 50 degrees for at least 48 hours to develop adequate strength. Pour too late in the season and you're running ground heaters and insulated blankets to protect the slab. I've seen October pours in Truckee where the crew was heating the ground for a week before the trucks arrived and heating the slab for a week after. That's expensive, but a frozen slab is more expensive.

Weather windows are everything. The building season in Truckee runs roughly May through November, and even that's optimistic. Snow can fly in October, and the ground might not thaw until late April. That gives you maybe six solid months to get your dirt work done, foundations poured, and the building dried in. Compare that to the Central Valley, where you can work 11 months out of 12. Every day of production matters more up here because there are fewer of them.

Summer isn't a free pass either. Afternoon thunderstorms roll in almost daily from mid-June through August. By 1 PM, the cumulus clouds start building over the Sierra crest. By 3 PM, you can have lightning, hail, and an inch of rain in 20 minutes. I run my crews hard in the morning because I know the afternoon might get taken from me. When I see the CAPE values climbing above 500 on the morning forecast, I mentally cut the workday to six hours and plan accordingly.

Frost depth drives foundation design. In Truckee, the code-required frost depth is 36 inches. That means every footing, every pier, every foundation wall has to extend at least 36 inches below finished grade to get below the frost line. In Sacramento, it's 12 inches. That difference means deeper excavation, more concrete, more rebar, more time, and more cost for every single structure. A simple deck footing that takes 30 minutes to dig and pour at sea level is a full afternoon up here.

Access is the invisible cost. Mountain roads are narrow, steep, and sometimes unpaved. Getting a concrete truck up a 15 percent grade on a dirt road in spring mud is an adventure. Getting a 50-ton excavator delivered on a lowboy trailer around a switchback with six feet of clearance is a puzzle.

I've had jobs where the access road was the first thing we built, just to get the equipment to the pad. And then in November, the road closes for snow season and your jobsite is locked until May. If you left material or equipment up there, it's sitting under eight feet of snow until spring.

Snow load is a design factor you don't think about at sea level. Truckee gets 200 to 400 inches of snow in a normal winter. Roofs up here are designed for 150 to 200 pounds per square foot of snow load. That means heavier framing, steeper pitches, and engineered trusses on everything. A roof design that works in Sacramento will collapse in Truckee. Period.

None of this means you can't build up here. Some of the most beautiful structures I've worked on are in the Sierra. But the mountain demands respect. Plan for less production, shorter seasons, deeper foundations, aggressive curing, and weather that changes faster than you can react. The crews that succeed are the ones that adapt to the mountain instead of fighting it.

So, GPS machine control changed everything about how I run a grade. I still remember the first time I saw it work on a job. Watching the bucket track a design surface in real time on the in-cab screen, and thinking, this is going to make a lot of grade checkers nervous. It did.

But here's the thing nobody tells you: the system is only as good as the setup. Bad setup means you're chasing hundredths all day and blaming the satellites for your own mess. Here's how I set it up every morning before the bucket hits dirt.

It starts with the base station. The base is a GPS receiver that sits on a known survey point, a benchmark or a control point that's been shot in by a surveyor with a known coordinate and elevation. You set a tripod over that point, level it, plumb it, and power it on. The base locks onto satellites and starts broadcasting correction signals.

It knows exactly where it is, so it can calculate the error in the GPS signal and send that correction to the rover on the machine. Without the base, your GPS is only accurate to about three to ten feet. With it, you're down to about an inch. That correction is called RTK, real-time kinematic, and it's what makes machine control possible at construction-grade accuracy.

The base broadcasts corrections over a radio link, usually UHF, to the rover on the machine. On a typical excavator setup, the GNSS antennas don't ride the boom — they mount up on the machine body, out of the way of the digging, and give the system position and heading. The boom, stick, and bucket each carry a sensor that reads its angle.

Those angles, plus the machine's measured-up geometry — boom length, stick length, bucket dimensions — let the system calculate exactly where the cutting edge is in three-dimensional space. Every time you move the stick or curl the bucket, the system recalculates. It's doing that math dozens of times per second.

Before you start digging, you run a calibration routine. This is where you check the system against a known point. You set the bucket tooth on a survey hub with a known elevation, tell the system to compare, and see what the offset is. If you're within a tenth, you're good. If not, you troubleshoot.

Common issues are a loose mount on the receiver, a bad satellite constellation that morning, multipath interference from nearby buildings or rock faces, or a base station that didn't initialize properly. I've seen guys skip calibration and end up half a foot off by lunch. Don't skip it.

The in-cab display is where it all comes together. On the Trimble system I run, the display shows a cross-section of the design surface and where your bucket is relative to it. There's a color-coded indicator: red means you're above grade and still have material to cut, blue means you've gone below grade and owe fill, and green means you're on grade within the tolerance you set.

Most finish grading work is set to a half-inch tolerance. Rough grading might be two inches. The display also shows your plan view, so you can see where you are on the site relative to the design, property lines, utilities, and any other layers the surveyor loaded into the model.

If you saw my vlog on the Cat 335 running Trimble GPS on a guided dig near Lake Tahoe, that's exactly this system in action. That Cat 335 ran dual GNSS antennas, a Trimble TD520 in the cab, and a design surface loaded for a commercial pad. Watching the operator work that machine was like watching someone play a video game, except the stakes are real and the tolerances are tight. The bucket held grade inside a half inch on a pad a hundred feet wide. No grade checker needed. No stakes to chase. Just the machine, the satellites, and the design.

Typical accuracy specs on a well-set-up RTK system are plus or minus one-tenth of a foot horizontal and plus or minus two-tenths of a foot vertical. In practice, with good satellite geometry, a clear sky, and a solid base station setup, I consistently see better than that. But accuracy degrades in certain conditions. Heavy tree canopy blocks satellite signals. Working in a narrow trench between tall buildings creates multipath errors. Rain and heavy cloud cover can weaken the correction link. And if your base station is more than about six miles from the machine, the correction quality starts to drop.

For most jobsite work, none of that is an issue. But when you're pushing the limits of the system, you need to know where those limits are.

One thing people don't talk about enough is maintaining the mounts. Those antenna and sensor mounts take a beating. The machine vibrates, the boom slams, dirt and debris hit everything. A loose mount means a moving receiver, and a moving receiver means your position is bouncing around even when the bucket is still.

I check the mounts every morning as part of my walkaround. If a bolt is loose or a bracket is cracked, it gets fixed before the machine starts. Ten minutes of maintenance saves you a day of chasing bad readings.

GPS machine control isn't magic. It's a well-engineered system that requires disciplined setup and regular calibration. But when it's running right, there's nothing faster or more accurate for moving dirt to a design grade. It cut my staking time to nearly zero, reduced rework by at least half, and let me run grades that pass the first time the surveyor shoots them. If you're running excavation and you're not on GPS, you're working harder than you need to.

So, every time I put a bucket in the ground up here in the Sierra, the dirt tells me a story. It's never the same story twice. You can be digging clean decomposed granite one minute and hit a clay pocket or a boulder the size of a washing machine the next. What's down there decides your bucket, your pace, your compaction plan — and whether you make schedule or spend the day fighting the ground.

Decomposed granite — DG — is the most common material we dig in the Truckee area. It's exactly what the name says: granite that's broken down over millions of years into coarse, sandy, gritty material. Dry, it's dream dirt — cuts clean, doesn't stick to the bucket, and compacts beautifully with a plate compactor or a roller. The grains interlock and you get excellent density numbers without a lot of water or passes.

Wet is a different animal. DG turns into a heavy, soupy mess that sticks to everything and won't release from a smooth bucket. I run a toothed bucket in it almost exclusively — the teeth break it up and the material falls out cleaner than a smooth edge that just smears it.

Volcanic rock is the wildcard. The Sierra Nevada sits on volcanic and granitic geology, and depending on where you're digging you might hit basalt, andesite, or welded tuff. Some of it breaks up with a standard bucket if it's weathered enough. Some of it laughs at you.

I've hit ledges of andesite that stopped a 35,000-pound excavator cold. When you hit solid rock you're looking at a hydraulic breaker, a rock saw, or — if the project allows it — controlled blasting, none of them cheap or fast. The geotech report should tell you what's down there, but I've seen plenty that said "anticipated rippable rock" and then we hit solid ledge at four feet. The ground doesn't read geotech reports.

Clay pockets show up where you least expect them. The Truckee basin is mostly granitic, but clay hides in the low spots — near meadows and old lake beds — and the Martis Valley area is notorious for it. Clay is the opposite of DG in almost every way: it holds water, swells, shrinks, and has almost no bearing capacity when wet.

If you're putting a foundation on clay, you're overexcavating and replacing it with engineered fill. Period. In the trench, clay sticks to everything — the bucket, the teeth, the walls. You'll spend as much time cleaning your bucket as you do digging, so I switch to a clean-out bucket or a wider ditching bucket to move more volume and live with the sticking.

Glacial till is what the glaciers left behind when they carved these valleys — sand, gravel, cobbles, boulders, clay, and silt all jumbled together with no sorting. Digging it is unpredictable by definition. You might pull a bucket of clean gravel followed by a three-foot boulder wrapped in clay. I quote till work at about 60% of my normal DG production rate, and even that's optimistic if the boulders run dense.

Elevation changes the ground itself. At 6,000 feet in Truckee, winter frost can drive 36 inches deep or more, and that frozen layer digs like concrete until it thaws. Even in late spring you'll hit frost pockets on north-facing slopes that never see direct sun.

The thaw cycle is the sneaky part — it leaves a wet, unstable layer between the frozen ground and the surface that fails slopes and collapses trench walls. I won't dig a trench on a north-facing slope in April without probing the thaw with a bar first. Bar stops at 18 inches? You're still frozen. Wait or plan for it.

Water table depth is the other elevation factor. In the Truckee basin it can be surprisingly shallow, especially in spring when snowmelt is recharging the aquifer. I've hit groundwater at three feet on sites where the geotech said eight — because the report was done in September and I'm digging in May.

When you hit water, everything changes. You need a dewatering plan — pumps, maybe a sump or a wellpoint system. You can't compact saturated soil, you can't hold a trench wall in saturated sand, and you definitely can't pour a footing in standing water. Know your water table before you break ground.

Soil drives your grade control too. DG holds a grade once it's compacted — cut it to a tenth with GPS and it stays where you put it. Clay moves: it shrinks when it dries and swells when it gets wet, so a grade you cut today might be a half inch off next week.

Till with cobbles is the hardest to cut. The edge bounces off rocks and the grade jumps around on the display, so in rocky till I'll cut rough grade with the machine, then check with a laser and a rod and hand-trim the last couple inches. Slower, but it's accurate.

Reading the ground is a skill that takes years. The color, the texture, the smell, the way it sounds when the bucket hits — all of it tells you something. Dark soil means organics you're going to overexcavate. Red or orange streaks mean iron, which means water's been moving through.

Smooth, shiny trench walls mean clay. Gray, gritty material that crumbles in your hand — that's your DG. Learn to read it and the ground tells you everything before the lab results come back.

So, every set of construction plans has a grading plan in it, and if you're running excavation you'd better know how to read one. Nobody teaches this well. Most guys learn it by osmosis, one stake at a time. Here's how I break it down with the plans rolled out on the hood of the truck.

First thing you look for is the bench mark. That's your reference point for every elevation on the sheet — a fixed, known point like a survey monument, a manhole rim, the top of a fire hydrant. Something that won't move. It'll have an elevation listed next to it, something like BM = 5,842.50, and every other number on the plan is relative to that point. Lose the bench mark or knock out the survey hub and you're flying blind. Protect your bench marks like they're made of gold.

Contour lines are the curved lines that show you the shape of the ground. Each line represents a specific elevation. When the lines are close together, the ground is steep. When they spread apart, the ground is flat. Existing contour lines are usually dashed. Proposed contour lines, the ones you're building to, are solid. The interval between lines is noted in the legend, usually one foot or two feet depending on the scale. Read that legend before you start interpreting anything.

Cut and fill is where the money is. Cut means you're removing material. Fill means you're adding it. The engineer will often show cut areas with one hatch pattern and fill with another. Some plans use color, red for cut and blue for fill, or they'll just call out spot elevations and let you do the math. Existing grade minus proposed grade: positive number means cut, negative means fill. When I walk a site before we start, I'm doing that math in my head at every stake, building a picture of where the dirt's going to move and how much of it there is.

Slope ratios show up as numbers like 2:1 or 3:1. That's horizontal to vertical. A 2:1 slope means for every two feet you go out horizontally, the ground drops one foot vertically. Steeper slopes like 1.5:1 or 2:1 need compaction and might need erosion control. Flatter slopes like 4:1 are usually for areas that get mowed or driven on. If the plan calls for a 2:1 and you cut it at 1.5:1, that slope's going to fail, and it's going to be your problem.

Utility callouts are the lines and symbols that show you where the underground infrastructure goes. Water, sewer, storm drain, gas, electric, communications. Each one has its own line type. Solid, dashed, dot-dash, each means something different. The plan should have a legend telling you which is which. Pay attention to the invert elevations — that's the elevation at the bottom of the pipe, not the top. When you're digging a trench for a sewer line, the invert is what matters because that's what sets your grade for flow. Miss it by two inches and the pipe doesn't drain. Get it right and the inspector signs off the first time.

Spot elevations are individual points on the plan with a specific elevation called out, like FF = 5,844.00 for a finished floor or TC = 5,843.25 for top of curb. These are your targets. When I'm running GPS on the machine, those spot elevations are loaded into the design surface and I can see them on my in-cab display in real time. But even without GPS, you need to know where those numbers are on the paper and how they relate to each other.

The last thing I check is the drainage pattern. Water flows downhill, and the grading plan should show you exactly where it's going. Look for the drainage arrows, the valley lines between high points, and the locations of catch basins, area drains, and outfalls. If the finish grade doesn't drain the way the plan shows, you're going to have standing water, erosion, and a callback. Grade for drainage first, everything else second.

Reading a grading plan isn't hard once you know what to look for. But you have to actually look. I've seen crews start cutting without reading the plan and end up three feet off because they assumed instead of checked. Don't assume. Find the bench mark, check the contours, run the cut/fill numbers, then start digging. The plan is the map. Treat it like one.

So, every construction crew deals with weather. But running excavation in the mountains is a completely different game. Down at sea level you check the forecast, see rain, make a call. Up here at 6,000 feet the forecast might say 20% chance of showers and you end up under a full thunderstorm by 2 PM with zero warning from the morning sky. That's why we built a system for it. No more guessing.

We run an 8-model forecast ensemble that pulls data from the major weather models: GFS, ECMWF, NAM, HRRR, NBM, and a few others. Each model has strengths and blind spots. The GFS is good for broad patterns but misses mountain-scale detail. The HRRR updates every hour and handles convective storms better than anything else out there. The ECMWF is the best global model running but sometimes lags on fast-developing Sierra systems. By stacking them together and looking at where they agree and disagree, you get a much clearer picture than any single forecast gives you.

The GO/NO-GO call happens before the crew leaves the yard. I'm looking at precipitation probability, wind speeds, lightning risk, and ground conditions from the previous 48 hours.

Here's what triggers an automatic NO-GO: sustained winds over 35 mph, which is a crane and rigging safety issue and makes dust control impossible. Any lightning in the forecast within the work window. More than half an inch of rain expected during the shift, because at that point you're not grading, you're making mud. And if the ground is already saturated from recent storms, even a light shower can turn a jobsite into a skating rink for tracked equipment.

The tricky calls are the ones where the models disagree. Three models say dry, two say afternoon showers, three are somewhere in between. That's where experience matters more than data.

I look at the upper-level pattern. Is there a trough moving through? What's the freezing level doing? Is there moisture streaming in from the Pacific or are we under a ridge? If the HRRR is firing up convective cells over the Sierra crest by 10 AM in the model run, I take that seriously even if the GFS says clear skies. The HRRR sees what the bigger models miss at mountain scale.

We also watch soil temperature and recent precipitation totals. After a wet period, even a sunny day might be a NO-GO if the subgrade hasn't had time to dry. You can compact wet soil all day and it'll fail the density test every time. Better to wait a day and do it right than burn fuel and hours on work that won't pass inspection. The moisture content of the soil matters as much as what the sky is doing.

Mountain weather is different from valley weather for one simple reason: terrain forces air up. When moist air hits the Sierra, it rises, cools, and condenses. That's why Truckee can be sunny at 8 AM and getting hammered by 1 PM on a day the Sacramento forecast called partly cloudy. The mountains create their own weather. If you're not accounting for that in your planning, you're going to get caught.

The system we use gives the crew a morning briefing with a confidence rating. High confidence GO means every model agrees, skies are clear, ground is dry, we're running all day. Marginal GO means conditions are workable but we might need to pull early. Weather hold means we're watching conditions and will reassess by a certain time. Full NO-GO means the machines stay parked.

The crew knows the system, trusts it, and doesn't waste time arguing about weather at 6:15 in the morning. The data makes the call.

Ski racing taught me this. You don't send a racer down a course in conditions you haven't assessed. You inspect, you measure, you check the snow, the wind, the visibility, and then you make a call. Same discipline applies to running a jobsite. The forecast is your course inspection. Treat it like one and you'll keep your crew safe, your schedule on track, and your grade intact. Run your own GO/NO-GO check on the crew weather dashboard, or dig into the full 8-model Sierra weather platform for the raw data.

So, if you run a crew in Truckee between June and September, summer thunderstorms are part of the job. They're not valley rain. They build fast, hit hard, drop lightning on ridgelines and open ground, and disappear in an hour — jobsite soaked, sky blue again like nothing happened. Knowing how they work isn't optional up here. It's a safety issue.

The thing driving Sierra summer storms is something called CAPE. Convective Available Potential Energy. Sounds like a weather nerd term and it is, but it matters on the jobsite. CAPE measures how much energy is loaded in the atmosphere for air to rise fast and build thunderstorms. Under 500 J/kg, you're usually fine. Stable air. Once it climbs above 1,000, storms can fire. Above 2,000 and you're looking at severe cells, heavy rain, hail, dangerous lightning. I check CAPE every morning. It's one of the first numbers I look at before the crew leaves the yard.

The pattern in Truckee is predictable in its general shape even if the specifics change daily. Morning starts clear and cool. The sun heats the ground, the ground heats the air above it, and that warm air starts rising. By late morning the cumulus clouds start building over the ridgelines. By early afternoon those clouds are towering, and if there's enough moisture and instability in the atmosphere, they go vertical and become cumulonimbus, full thunderstorm cells. The typical window for summer thunderstorms in the Truckee area is between noon and 5 PM, with the peak risk around 2 to 3 PM.

Lightning is the primary danger for construction crews. When you're standing on open ground next to a steel excavator, you're the tallest conductive path in the area. OSHA doesn't have a specific lightning standard, but the National Weather Service 30-30 rule is the baseline: if the time between a flash and the thunder is 30 seconds or less, the storm is within six miles and you get off the ground. Wait 30 minutes after the last flash before going back. That's the rule and we follow it. No exceptions. No waiting to see if it passes. When I hear thunder, the crew's already moving to vehicles.

The tricky part in the Sierra is that storms can develop directly overhead with very little advance warning. In flat terrain you can see a storm building 30 miles away and watch it approach. In the mountains, a cell can form on the backside of a ridge and be on top of you before you've heard a single rumble. That's why the morning forecast matters so much. If the models show CAPE above 1,000 and the HRRR is firing convective cells in the afternoon run, I build that into the day's plan. We front-load the critical work. Grade checks, compaction tests, anything that needs people on the ground in open areas, that happens before noon. Afternoon work shifts to tasks that can be paused quickly or done near cover.

Rain itself is a separate problem. A summer thunderstorm in Truckee can dump half an inch in 20 minutes. That's enough to turn a freshly graded pad into a rutted mess if equipment is still tracking across it. Trenches fill with water. Slopes that were just compacted start sheeting runoff. SWPPP controls get tested in a hurry. If you have silt fencing and wattles in place, great. If you don't, you're going to have a sediment problem and a conversation with the inspector you don't want.

Flash flooding is real in the Sierra even in summer. The soil up here is a thin layer of decomposed granite over bedrock. DG drains fine in a normal rain, but a cloudburst outruns it — the ground can't take water in that fast, so it sheets downhill into whatever low point exists. If your trench is the low point, you've got a problem. We always check drainage paths before a storm window and make sure the site can handle a burst of heavy rain without trapping water where people are working.

The best thing you can do as a foreman is build storm awareness into the crew culture. Everyone should know the signs: towering cumulus that starts going dark on the bottom, a sudden wind shift, a drop in temperature, the smell of rain in the dry air. By the time you hear thunder it's already close. The guys who've been through a few Sierra summers know the drill. The new guys need to learn it fast. One close lightning strike on a jobsite is usually enough education for a lifetime, but I'd rather they learn from the briefing than from the bolt.

So, this is what it looks like when an excavator knows exactly where it is in three-dimensional space — down to about an inch — in real time while it digs. If you've never seen a guided dig before, watch the video, then let me walk you through what's actually happening in that cab.

A guided dig means the excavator is equipped with a GPS/GNSS machine control system that shows the operator exactly where the bucket teeth are relative to the design grade surface at all times. The in-cab display gives you a live cross-section of the ground, the design surface from the engineered plans, and your bucket position overlaid on top of it. You can see exactly how much material needs to come out, where you're cutting into grade, and where you still owe fill. No guessing. No eyeballing it. The machine tells you precisely where you are relative to where you need to be.

The way Trimble sets this up on an excavator like the 335 is with dual GNSS antennas mounted on the machine body, up out of the digging, while the boom, stick, and bucket carry angle sensors. The antennas talk to GPS and GLONASS satellites overhead, and a base station set up on a known survey point nearby sends real-time kinematic corrections — what we call RTK — to tighten the accuracy from a couple of feet down to about an inch. The system knows the exact geometry of the machine — boom length, stick length, bucket dimensions, all of it — so when you combine the satellite position, the RTK correction, and the angle sensors, the computer can calculate exactly where the tip of your bucket is at any given moment.

The in-cab display is a Trimble unit, usually a TD520 or similar, mounted right where you can see it without taking your eyes off the work for more than a second. It shows you a color-coded cut/fill map: red means you're above grade and still cutting, blue means you've dug below grade and owe fill. When the screen goes green you're on grade. That color feedback is instant. You move the bucket a quarter inch and the display updates. It's like having a grade checker standing in the trench with a laser level at all times, except this one never takes a break and never misreads the rod.

The Cat 335 is a solid mid-size excavator for this kind of work. You're looking at roughly 80,000 pounds operating weight, a reach that handles most utility trench depths and foundation work without issue, and enough power to move material in the kind of ground we deal with around Tahoe. And up here, the ground is no joke. You hit decomposed granite, you hit rock, you hit volcanic material that doesn't want to move. The 335 has the muscle to work through that while the GPS keeps every cut precise.

What a guided dig changes on the job is significant. Without GPS, you're relying on grade stakes set by a survey crew, a laser level in the trench, and a grade checker calling out readings while the operator adjusts. That works, and it's how we did it for decades, but it's slow. Every time you need a grade check the work stops. The surveyor has to come back out to re-stake if anything gets disturbed. And the margin for human error — misreading a cut sheet, pulling a measurement off the wrong benchmark — is always there.

With the guided dig, the operator is self-sufficient on grade. I can see the design surface on my screen, I can see my bucket position, and I can cut to within a tenth of a foot on the first pass. That means fewer re-digs, less over-excavation, less wasted material that has to be hauled off or backfilled, and faster production overall. On a job where you're moving thousands of yards of material, the savings in time and backfill costs add up fast. It also means less fuel burned because you're not making unnecessary passes, and less wear on the machine.

Running GPS at elevation around Lake Tahoe adds its own variables. The terrain is steep, the ground conditions change every fifty feet, and you're working at 6,000-plus feet where ridgelines and tree cover can eat into your sky view. But a clean base setup handles that. As long as the base sits on a known point with good sky visibility, the corrections stay tight and the system performs. I've run guided digs in conditions where you couldn't see the bottom of the trench from the cab and the GPS still had me within spec.

Full PPE on site as always. Hard hat, high-vis vest, steel toes, safety glasses. The GPS doesn't change the safety requirements, it just makes the finished product more precise. That's what a guided dig looks like in the field — the Cat 335, Trimble GPS, and a design surface the machine follows in real time. This is how modern excavation gets done. See the full fleet and specs on the equipment page, and browse the field-tested tools that run alongside the iron.

Dug this one out of the archive. Winter X-Games, Aspen, January 2010. Ski cross. Four of us in the gate and about ninety seconds to figure out who wants it more.

Ski cross is the only discipline where you race wheel to wheel. No time trials, no split screens. You're right there next to three other riders coming off the start, jockeying for position into the first turn, making decisions at 60 mph you can't take back. Contact happens. Crashes happen.

The course is built with rollers, jumps, banked turns, and flat sections designed to separate riders and punish anyone who isn't precise. You can be the fastest skier on the hill and still lose if your line is off by a foot going into a compression.

What people don't see on TV is how physical it is. SX is a full-body sport. Your legs are burning from absorbing terrain at speed, your core is locked in because any time you get loose you lose time, and your arms are driving the poles and managing balance through features that are throwing you around. A 90-second race run feels like a three-minute sprint when your quads are on fire at 9,000 feet in Aspen.

I was representing Jamaica at this point. Had been on the FIS circuit, qualified through the rankings, earned my spot. People always want the novelty angle on that, the Jamaican on snow. But there was nothing novelty about the training.

I came up through the U.S. Ski Team development system, trained on the same mountains as the Americans, the Canadians, the Swiss. The flag was Jamaica but the preparation was world class. By the time I got to Aspen for X-Games I had hundreds of race starts behind me and I knew exactly where I stood in the field.

This footage is raw. You can hear the crowd, you can see the course, and you can watch how a ski cross heat actually unfolds from start to finish. The start is everything. If you're a half second late out of the gate you're chasing for the rest of the run, and in SX there's no catching up once someone has the inside line into the first bank. I studied those gates like I study grade stakes now. Every detail matters when the margin is that small.

Two weeks after this I was in Vancouver carrying the Jamaican flag in the opening ceremony. Finished ninth in the Olympic ski cross event. Not a medal, but top ten in the world in a discipline making its Olympic debut. I'll take that every day of the week.

I keep this footage around because it reminds me that speed without control is just noise. Same lesson applies on a ski course at 60 mph and on a jobsite running a track loader at 6,000 feet. Read the terrain, hold your line, stay precise. The arena changed but the mentality didn't.

So, the solstice is this week, which means Truckee is running on about fifteen hours of daylight right now. Sun comes up around 5:30, doesn't fully quit until after 8:30 at night. For most people that means longer evenings on the patio. For a construction crew it means the clock just got a whole lot more generous, and we're going to use every minute of it.

There's a window in the Sierra that only opens once a year. The ground is finally dry from snowmelt, the frost is out of the soil, permits are moving, and now the sun is giving you the most working light you'll get all season. Mid-June through mid-September is when everything real gets built up here. Miss it and you're scrambling against short days and early snow before you know what happened. So when the window is wide open like it is right now, you run.

My alarm goes off at 5:30. In the truck by 6:15, coffee in hand, running through the day's plan in my head. On site by 7:00, walking the grade before the full crew gets rolling, checking what the overnight temps did to the ground, making sure nothing shifted. From there it's a full sprint. Normal day we're wrapping up by 3:30. On a push, maybe 5:30 if the job demands it and the crew has the legs for it. People ask me how long my workday is and I tell them it depends on the job, not the clock.

The heat is real but different up here. We sit at almost 6,000 feet, so the air is thin and the UV is no joke even when the thermometer only reads 82. Sea-level guys visit a Truckee jobsite in July and can't figure out why they're cooked by noon. It's the altitude. The sun hits harder and the thin air doesn't cool you the same way.

Hydration isn't optional — it's infrastructure. I go through a gallon of water before lunch most days, and if a guy on the crew isn't drinking, I notice and I say something. Heatstroke at elevation sneaks up on you faster than people think.

But here's the thing about these long days that makes all of it worth it. You can actually see the progress. In winter, between the short light and the weather delays, a week of work might move the needle an inch on the schedule. In June, a single good day can change the shape of a jobsite.

Trenches get cut, pipe goes in the ground, backfill gets compacted, and by the time that 8:30 sunset rolls around you're standing on a site that looks fundamentally different than it did at sunrise. That feeling never gets old. Sixteen years of running equipment and I still get a kick out of looking back at a day's work and thinking yeah, we moved some earth today.

The other thing the long days do is compress your personal life into these weird little pockets. Lindsay has figured out the rhythm by now. Dinner is late. Weeknight plans don't exist between June and October. Saturdays I try to protect but the job doesn't always cooperate.

It's the trade-off for living in a place where the off-season actually means something. When November hits and the snow flies, the machines get parked and the pace changes completely. But right now, in the thick of it, the days are long, the work is good, and the Sierra is letting us build.

I remember racing ski cross in June. The season was over, the competitive circuit was done, but we'd still be on snow somewhere in the southern hemisphere or up on a glacier doing camp turns and staying sharp. The mentality was the same: when conditions are right, you train. You don't waste good snow and you don't waste good weather. The mountain gives you a window, you use it. That lesson didn't come from a textbook. It came from showing up every day the conditions allowed and not quitting until they didn't.

So if you're driving through Truckee this week and you see a crew still out at five in the afternoon, that's not overtime. That's a push day, getting it done while conditions allow. The longest days are the most valuable days, and we've got about three months of them left before the Sierra starts reminding us who's really in charge. I plan to make every one of them count.

So, if you live in the Sierra and you ski, your garage is a disaster. I know because mine was. Skis leaning against the water heater, poles tangled behind the snowblower, boots in a pile by the door where they'd freeze solid overnight. Every morning before first chair was a ten-minute excavation project just to find matching gear. Finally got tired of it and built something.

The goal was simple. Every pair of skis in the house up and out of the way, and the floor completely clear so I can still park and still get to the workbench without climbing over a yard sale of Völkls.

The racks are steel rails on strut hangers, mounted overhead in the dead space above the garage door tracks. Every garage has that space and almost nobody uses it. Everything is lagged into solid framing — not drywall, not toggle bolts. A loaded rack is heavy, and anything that lets go up there lands on something expensive. Pre-drilled, sank 3/8-inch lags, loaded it up. That rack isn't going anywhere.

Each pair rides flat across two rails, bindings down, edges separated so they don't chew each other up. The skis stay off the ground, dry, and out of reach of everything that moves down below.

The rails aren't the whole system either. Overhead mesh platforms took everything that used to colonize the floor — bins, boxes, the stuff you touch twice a year. Same rule up there: the ceiling works for a living, the floor is for machines.

What I'd do differently next time: wider spacing between pairs. I packed them tight to maximize capacity and it works, but pulling a pair out of the middle is a two-hand operation. Another inch per slot would make the grab-and-go smoother. Also, I'd add a dedicated wax station. Right now I do edge work on the regular workbench, which means clearing tools every time. A fold-down table on the opposite wall would fix that.

Total build time was about six hours across two weekends. Materials ran around $180 in rails, lags, and hardware. The most expensive part of the whole thing was the beer afterward. If you have a drill, a stud finder, and a level, you can do this in a day. No welding required, though I did think about fabbing the whole thing from scratch just because I could. Lindsay talked me out of it. She was right.

The real payoff is the morning routine. Walk in, grab skis, grab poles, grab boots, walk out. No digging, no untangling, no knocking over the snowblower. That alone was worth every lag bolt.

People think a machine runs the same everywhere. It doesn't. Truckee sits at almost 6,000 feet, and every engine I've ever put to work up here gives back a little less than the spec sheet says. Thin air is the tax you pay for the view, and the dirt doesn't grade itself any easier because the scenery is nice.

Start with power. Less oxygen up here means less burn, and a naturally aspirated diesel loses roughly three percent of its output for every thousand feet of elevation. Do that math on a small loader or a mini ex and you're already down close to a fifth of the muscle the sticker promised before you ever touch a lever. Turbos claw most of it back by force-feeding air, but they work harder to do it, and harder means hotter. Keep that in your head when you're sizing a job or wondering why the machine feels lazy pushing up a grade.

Then there's the cold, even in June. Summer mornings in the Sierra still drop into the 30s. Cold diesel is thick, cold hydraulic fluid is thicker, and dropping a stiff machine straight into a hard cycle is how you cook seals and shorten a pump's life.

Let it idle and come up to temp. Cycle the hydraulics with no load a few times until everything moves smooth and quiet. Five minutes of patience beats a five-figure pump every single time.

Cooling is the one that bites people. Thinner air carries less heat away, so your radiator is already at a disadvantage before Sierra dust and pollen pack the fins solid. Blow them out, then check them again that afternoon. A clogged cooler at altitude on a hot day is how you end up watching a machine steam in the middle of a dig instead of finishing it.

Same story with air filters out here. This dry, dusty ground eats them faster than anywhere I've ever worked, so carry spares and check them more than you think you need to.

Fuel and batteries close it out. With our temperature swings, condensation is real, so top the tank at the end of the day, not the start. Less empty space means less room for water to collect overnight in the bottom of your tank.

And a battery that limps through a cold morning is telling you it's done. Swap it before it strands you on the busy day, not after.

None of this is fancy. It's grease guns, clean filters, and giving the machine a minute before you ask it for everything. The altitude is going to take its cut no matter what you do. Your only job is to make sure you don't hand it any extra. Treat the iron right up here and it'll put down clean work all season long. Treat it like sea level and the mountain will find the weak point for you, usually at the worst possible time.

Up here at 5,820 feet, you don't pick the day. The mountain picks it for you. June in Truckee is when the snow finally lets go and the ground turns workable, and every spring I get the same reminder. The calendar on the wall means nothing next to the one written in the dirt.

I learned this long before I ever ran a machine. Ski cross taught me that you can't force a start. You can be the strongest guy in the gate and still lose if you commit half a second before the course is ready for you. Timing beats power almost every time. The racers who lasted were the ones who knew how to wait for the right moment, then take it without hesitation.

Excavation runs on the same rule. You can't grade saturated ground. You can't trust a slope that's still holding meltwater. Push a job before conditions are right and the Sierra makes you do it twice. So you read the sky, you check the soil, and you let the work tell you when it's ready instead of the other way around. That isn't slacking. That's the discipline nobody puts in a highlight reel.

People mistake patience for sitting still. It isn't. Patience is staying ready while you wait. All winter, with the snow on the ground, I was in the shop welding, fixing what broke last season, getting gear set so the day the ground opened up we could move without scrambling. The waiting is where the real work hides. By the time conditions line up, you've either prepared or you haven't, and the mountain can tell the difference.

There's a bigger version of this lesson coming for me. Lindsay and I have a boy on the way later this year, and if anything's going to arrive on its own schedule instead of mine, it's a kid. I can't rush that either. All I can do is be ready, the same way I get ready for dig season. Sharpen what I can control. Make peace with what I can't.

That's the trade you make living in the Sierra. The weather's the boss. It doesn't care about your timeline, your invoices, or your plans for the weekend. But if you respect it, if you learn to read it and move when it gives you a window, it'll let you build things that last. The structures that survive up here get built by people who waited for the right day and then didn't waste it.

So I'm not in a hurry. The ground's finally ready, the machines are running, and the season's open. I spent the winter getting set for exactly this. Now it's time to put down clean work while the window's good, because up here you never know how long it'll stay open.

Some people know me from the 2010 Vancouver Olympics — Jamaican flag bearer, ninth in ski cross, the best Caribbean finish at a Winter Games. Others found me on the cover of Ski Magazine, or in the Warren Miller films. All of that still matters. None of it is the finish line.

What ski cross taught me is that speed without control is just noise. You read the terrain, you commit, and you hold your line when the course gets rough. That lesson followed me off the mountain and into the work I do now, running heavy equipment and building things that have to stand up to real weather and real load.

The Olympics gave me a stage. The films gave me a story. The cover gave me a moment. The discipline behind all of it is what I actually carry forward. Every day I get to point that same focus at a different kind of slope.

I'm still the kid who moved to Truckee and decided the impossible was worth chasing. The gear has changed. The drive hasn't. If you've followed any part of this ride, thank you. The best chapters are the ones still being written, and I plan to keep building them one clean line at a time.