How We Read and Predict the San Juan River: The Story Behind Pagosa’s River Desk and Black Hat Hydrology Lab

I have been in the river business long enough to learn two things.

First, the San Juan River will do whatever it wants.

Second, everyone assumes I know what that will be.

For more than 20 years, I could not walk into a Pagosa coffee shop, sit down on a bar stool or show my face in public between the first winter storm and the end of spring runoff without somebody asking me to predict the river.

“How’s the snowpack?”

“Are we going to have a good rafting year?”

“When will the river peak?”

“When will tubing start?”

“Will there still be enough water when my family gets here?”

Occasionally, somebody would want to know the exact flow on a particular Tuesday 73 days in the future. Because I owned the river company, I was apparently assumed to be part outfitter, part hydrologist and part waterproof fortune teller.

The uncomfortable truth was that I usually did have a reasonably good answer. Not a certain answer—anyone claiming certainty about a free-flowing mountain river should probably be moved a few stools farther down the bar—but after decades of watching the same watershed, I could often tell whether we were building toward a strong season, a normal season or one of those mild years when everybody needed to quit waiting for warm June weather and get on the rafts while the water was still there.

Those answers came from experience, but they were never purely “gut instinct.”

My gut had browser tabs.

The San Juan is a particularly difficult river to predict

Some river outfitters work below dams. They may not love every release decision, but at least somebody upstream is deciding how much water to send and publishing a schedule.

The San Juan through Pagosa Springs does not offer that courtesy. It is a free-flowing mountain river driven by snow, temperature, rain, dust, terrain and human water use. There is no upstream dam release giving us tomorrow’s operating flow, and there is no useful chain of upstream gauges providing hours of warning as a rise moves toward town.

Wolf Creek is famous for receiving the highest average snowfall in Colorado, and Pagosa Springs sits only about 25 miles from the top of the watershed. That sounds like a river forecaster’s dream until you meet the South San Juan Mountains.

Weather here is highly regionalized. One drainage can be wet, shaded and lush while the next drainage—separated by a ridge, a slightly different aspect and a few miles—holds a dry creekbed. A storm can bury one headwater basin and mostly miss another. A thunderstorm over downtown can soak the patios without contributing much water to the river at all.

Even our spring runoff pulse arrives with poor manners. Heat from the previous afternoon can become a midnight flush through Pagosa Springs, long after the sun has gone down and most reasonable people have stopped staring at snow temperatures.

We also lack an upstream gauge that would warn us the flush was coming. The primary gauge sits in the middle of the river sections we operate and historically reported with about a 17-minute delay. By the time the change appeared on the screen, that water was already in town and had passed through part of our operating corridor.

In other words, our “live” gauge was sometimes less of an early-warning system and more of a confirmation that the thing we were wondering about had already happened.

Then there are irrigation diversions. During high spring runoff, a diversion may represent a relatively small percentage of the river. During lower flows, that same amount of water can become a meaningful piece of what remains.

When ranchers begin irrigating hay fields, we can watch the river bump down. When it is time to stop watering and cut hay, the drop becomes smaller or the river holds a little better. Those are legitimate, longstanding uses of the water, but they add another local variable that cannot be explained by snowmelt and weather alone.

So when somebody asks, “What will the river be doing next week?” the honest answer requires more than checking one gauge. The San Juan is the combined result of several mountain drainages receiving different weather, snow melting at different rates, nighttime temperatures, dust, irrigation demand and a gauge that tells us what reached town about 17 minutes ago.

That is why a generic river app was never going to answer the questions we were being asked.

Before the River Desk, the forecast lived in my head

The old morning ritual began with the river gauge. Then came the Upper San Juan and Wolf Creek Summit SNOTEL stations, the official river forecast, the weather in town, the weather in the headwaters and a comparison with whichever previous season looked least unlike the current one.

After that came the information that did not fit neatly into a government chart. Had the nights stopped freezing at 10,000 feet? Was dust darkening the snow? Was the river rising faster every afternoon? Had irrigation demand changed? How much snow remained in the shaded timber? Were the jumping trout on Blackhead Peak and the geese on Quartz Ridge beginning to appear?

Eventually, I combined the numbers, forecasts, field observations and more than 20 years of memory and made a call. It sounds charmingly old-school now. At the time, it mostly meant too much coffee, 14 open browser tabs and a room full of employees waiting for me to say something useful.

The answer mattered far beyond bar-stool bragging rights. Our longer-range river calls affect when we hire and train guides, how many rafts we prepare, when we open reservation calendars, how we schedule drivers, what gear we purchase and when the business begins shifting from rafting toward tubing and lake rentals.

A mild rafting year means encouraging guests to come in May instead of waiting for their traditional June vacation. A strong season means preparing more guides, boats and shuttle capacity—and reminding locals that this may be the year they should not miss.

Guess too high and you prepare for business that never arrives. Guess too low and the river arrives before you do.

For years, all that information went into my head, bounced around with a few decades of memories and came back out as an answer. Eventually, the obvious question became:

Why had nobody built a single Pagosa-specific tool that did this?

Building the river tool we could never find

Pagosa Outside has always evolved by looking for the missing piece. If a better piece of river gear does not exist, we find one or convince somebody to make it. If a trip can run more smoothly, we change the logistics. If guests keep struggling with the same confusing question, we try to build a clearer answer.

The Pagosa Outside River Desk grew from that habit.

I knew the questions it needed to answer. I knew the local flow ranges, the watershed, the seasonal patterns and the decisions we had been making behind the scenes for decades. AI helped turn those ideas into formulas, code, graphs and a live tool that could update itself without requiring me to stand behind the website turning a crank.

Several thousand edits later, the River Desk became a digital version of our old morning routine—only more organized, more consistent and considerably less dependent on whether I had finished my second cup of coffee.

The gauge gives you a number. The Desk explains the river.

The USGS river gauge is one of the most important ingredients behind the River Desk. It tells us how much water is passing through Pagosa Springs in cubic feet per second.

What it does not tell an unfamiliar visitor is whether that number means rafting, tubing, shallow water, high water or finding something entirely different to do.

Most people are not searching for “instantaneous discharge at the San Juan River gauge.” They are searching for:

• Is the Pagosa River running?

• Is the river open in Pagosa Springs?

• Can we float today?

• Is the river high enough for rafting?

• Are tubes running?

• Is the water too low?

• How deep is the river?

• Should we reserve a raft or rent a tube?

• Can we raft Pagosa Springs in July?

• What is the river level right now?

Some of those questions are technically wrong. There is no “Pagosa River.” The San Juan does not open and close like a ski lift, and one CFS reading does not describe the depth of every rapid, pool and channel through town.

They are still perfectly reasonable questions. The River Desk’s job is to understand what people mean and give them a useful answer.

Rafts and tubes are mostly different seasons

One of the best parts of the River Desk is its concentric flow gauge. It shows the live river reading against Pagosa’s rafting and tubing ranges and makes one confusing idea immediately visible:

Rafting and tubing are mostly different seasons, with a small and occasionally rowdy overlap.

Pagosa Outside generally uses these flow references during our operating season:

• DIY River Tube Rentals: 30–399 CFS

• Deluxe Tubing Trips: 60–399 CFS

• Splash & Dash rafting: generally 300 CFS and higher

• Mesa Canyon rafting: generally 600 CFS and higher

• West Fork rafting: generally 800 CFS and higher

Those numbers are useful ranges, not giant automatic switches mounted beside the gauge. Three hundred fifty CFS might be the fading end of rafting season, lively high-water tubing or the awkward handoff when both answers are partly correct. Cold May water and warm July weather can produce very different experiences at the same flow.

Weather, water temperature, debris, access, channel conditions and the River Crew still get a vote. The Desk reads the number together with its direction, season and surrounding conditions. That is what turns a gauge into an answer.

The river is not standing still because the gauge has a number

A gauge reading tells us what has already arrived in Pagosa Springs. On this river, where the gauge sits in the middle of our operating corridor and reports after the water gets there, that distinction matters more than it might elsewhere.

The next question is whether more water is coming.

The River Desk’s River in Motion graph connects seven days of measured flow to the official seven-day river forecast. Beyond that, it widens into a weather-aware outlook through day 10.

Those pieces look different because they are different. Yesterday is a measurement. Tomorrow is a forecast. Day 10 is Mother Nature watching us get a little too confident.

During spring runoff, the Desk watches warm headwater days, nighttime refreezing, remaining snow and how quickly the river is responding. It tries to anticipate whether yesterday afternoon’s heat may become tonight’s rising water—even though we do not have an upstream gauge watching that pulse begin its trip toward town.

During summer, the Desk looks for a different story. A thunderstorm over downtown may soak the tourists, empty the patios and do very little to the river. Widespread rain over the East Fork, West Fork and surrounding headwaters has a much better chance of producing a measurable rise later.

The weather cards separate morning and afternoon conditions because “sunny with thunderstorms” is not a contradiction in the San Juan Mountains. It is a fairly normal Tuesday.

The River Desk changes jobs with the seasons

A useful river dashboard in January should not behave like the same dashboard in May, July or September. The river is not asking the same question all year.

In fall, the Desk looks back at the completed water year and begins watching for the first mountain snow that does more than pose for a photograph before melting. A snowflake in the forecast is not enough. The Desk waits for measured snow to settle into the two headwater SNOTEL stations before declaring that the new mountain water reserve has begun.

During winter, the clearest measurement is percent of average. Fifteen inches of snow-water equivalent can be encouraging on one date and concerning on another. Comparing the current total with what is normally present on that same date tells us whether winter is keeping pace, falling behind or quietly assembling something interesting.

After peak snowpack, the language changes. Then the useful question is not merely how the snow compares with average. We want to know how many inches of stored water remain, how quickly those inches are disappearing and how much rafting-season firepower is still sitting in the mountains.

In spring, the Desk watches stored snow become river. By summer, when the monitored snow has usually melted out, native flow, irrigation demand and widespread headwater rain take over the story.

The Desk changes its emphasis because the watershed changes its answer.

Then somebody asked us to predict a family vacation

The River Desk answered much of the “What is the river doing now?” problem. It did not solve the toughest question our front desk receives:

“What will the river be doing when we get there?”

Someone calls in March with a trip planned for June 21. They do not care what the gauge reads today; they want to know whether to reserve a raft. Someone else is arriving in July and wants to know whether there will be enough water to tube.

A wedding party is coming in six weeks. A family reunion has one free afternoon. Twelve people have already booked flights and a vacation rental, and Grandma has been put in charge of finding something everyone will enjoy.

Our staff could explain normal seasonal patterns, but eventually everyone looked toward me. Apparently, 30-plus years in the river business meant I should know what a free-flowing, ungauged-from-above, heavily diverted mountain river would be doing on a particular Tuesday 53 days from now.

That is how a respectable river company ends up wearing a black hat.

Questionable hats. Serious data. Experimental forecasts.

The Black Hat Hydrology Lab began as an attempt to turn our longer-range river guesses into repeatable, testable formulas.

The name is partly a joke about treating public river data like classified intelligence. It is partly permission to experiment without pretending we have finally defeated hydrology. It also provides plausible deniability for an older Pagosa forecasting system involving groundhogs, a snowfield that sometimes looks like a jumping trout and measuring the height of the high country skunk cabbage.

The project began the way all sensible hydrology programs begin: I asked AI whether it could help organize some river data and build the tool I had been carrying around in my head.

Several thousand edits later, we had seasonal forecasting models, an interactive 90-day river map, a public scoreboard and enough JavaScript to make backing a raft trailer through downtown Pagosa feel relaxing.

The Lab now answers three different kinds of questions.

What is the river question right now?

Right Now in the Lab follows whichever seasonal question is currently unfolding. It starts by watching for the first lasting headwater snow, then follows peak snowpack, the first return to 400 CFS, the developing spring crest and the strength of the May–June rafting season.

After peak runoff, it watches the river’s transition toward tubing. Once tubing opens, it becomes a Monsoon + Bonus Rain-Flow Watch, estimating whether headwater rain may boost the river during the next seven days, when that water might reach Pagosa and whether the rise is more likely to remain in tubing range or climb back above 400 CFS.

This estimate is allowed to change. New snow, updated weather forecasts and fresh river readings should improve the answer. Right Now in the Lab is the chalkboard: it gets erased and updated as the lesson develops.

What might the river be doing on my date?

Ask the Black Hat exists because “Well…normally” is not especially helpful when somebody has already booked airfare.

Choose any date during the next 90 days and the Lab returns:

• A center flow estimate

• A wider experimental range

• The probability of tubing-range flow

• The probability of reaching 300 CFS

• The probability of reaching 600 CFS

• The probability of reaching 800 CFS

That date is also highlighted on the larger 90-day river curve. Visitors can work in the other direction, too: tap a point on the graph and the Lab turns that day into an Ask the Black Hat result.

The goal is not to promise exactly 427 CFS at 10:00 on a Tuesday morning 67 days from now. The goal is to give guests and our staff a more informed answer than a shrug, a historical average and whatever mood I happen to be in that morning.

It is the function I wish our front desk had possessed for the past two decades.

Were we actually right?

Live estimates are useful, but there is a problem with predictions that are allowed to keep improving: eventually, everybody remembers being right.

The Black Hat vs Mother Nature game prevents that convenient little miracle.

The Lab makes eight official predictions covering the full water year:

1. When will the first snow stick on the mountains?

2. How much water will the snowpack hold at peak?

3. When will the San Juan first reach rafting levels?

4. Will it be a good rafting season?

5. When and how high will spring runoff peak?

6. How long will rafting season last?

7. When will tubing season start?

8. How long will tubing season last?

These were not selected from an academic hydrology textbook. They are the coffee-shop questions, bar-stool bets, front-desk phone calls and staff-meeting mysteries that everybody expected me to answer because I was the person in town with the most to gain—or lose—from getting them right.

Each prediction has an issue point and a measurable finish line. Once the answer locks, it cannot move.

Right Now in the Lab can keep learning, but the official prediction sits there like an old bar bet written in permanent marker. A hit earns a point for the Black Hat. A miss earns a point for Mother Nature. A near miss remains a miss because Mother Nature has declined every invitation to negotiate the scoring rules.

Why some predictions wait for the river

The first six official predictions use established calendar issue dates. Lead time is part of the test; predicting the spring peak while standing beside it would be considerably less impressive.

Beginning with the 2027 game, the final two predictions wait for meaningful river events. The tubing-opening prediction issues after spring peak, when the falling river makes its transition below 800 CFS. The remaining-tubing-days prediction issues once tubing opening has been confirmed.

That allows the Lab to keep concentrating on rafting while rafting is still the question people care about.

The archived 2026 cards still show their original May 10 and May 15 issue dates. They should. Those were the rules when those bets were made, and improving the next game does not give us permission to walk backward through the Film Room and move the old goalposts.

The 2026 river punched the Black Hat directly in the hat

The Lab’s first season was an excellent stress test. Unfortunately, “excellent stress test” is usually another way of saying the river company would have preferred a different year.

Record rain and flooding arrived in October 2025. Then winter failed to build much snow.

The average of the two monitored headwater stations peaked at only 14.3 inches of SWE on March 6. The San Juan reached its spring crest at 903 CFS at 1:15 a.m. MDT on May 15.

Runoff was mild and short, and summer brought very little help from the monsoon. Some predictions handled the strange season surprisingly well. Others did not.

The original peak-snow model predicted 37.3 inches. Mother Nature delivered 14.3 inches.

That is not a rounding error. That is a public flogging.

But it exposed exactly the kind of problem the Lab was built to find. The original model had not seen enough winters as poor as 2026. Faced with an extreme year, it kept dragging the answer back toward the more normal seasons it understood.

We did not erase the prediction or quietly change the formula after the fact. The miss remains on the scoreboard, and the model went back to practice.

The Black Hat has a practice squad

Not every clever new formula gets to touch an official prediction. The Lab continuously tests challengers: new weather signals, additional stations, high-country refreeze information, long-range outlooks, different flow relationships and alternative ways of estimating how many useful days remain.

Some produce prettier answers. Some sound more sophisticated. Some manage to be wrong with exceptional mathematical confidence.

A challenger earns a starting job only if it beats the existing formula across historical seasons without peeking at the year it is supposed to predict. That rule has taught us that more data is not automatically better.

Adding more snow stations did not automatically improve every prediction. Broad climate outlooks did not help everywhere. A complicated spring-peak timing adjustment failed to beat Pagosa’s surprisingly persistent near-midnight peak pattern.

The goal is not to build the formula with the most ingredients. It is to build the simplest formula that consistently earns a better answer.

After the season, we watch the film

Pagosa Outside’s guest-facing river season runs from May 1 through the Sunday of Labor Day weekend. Labor Day Monday is when we close, deflate and begin packing everything away.

The Black Hat game reaches its final finish line at the same time, but the next season does not immediately replace it. September belongs to the Black Hat Film Room.

That is where the completed season stays together:

• The final score

• All eight original predictions

• The biggest hits and misses

• The complete snowpack and river story

• What fooled the formulas

• Which challengers improved the answer

• Which models kept their jobs

• Which ideas were sent back to the parking lot

The Film Room keeps an unusual season from being reduced to one peak-flow number. It also prevents our memories from becoming suspiciously flattering over time.

On October 1, the next water-year game begins. The old season remains in the Film Room while the Black Hat starts making fresh mistakes.

The unmarked door at the back of the Lab

The River Desk was built for guests and locals. The Black Hat Lab was built for people curious about what may happen next.

The Pagosa Outside Control Room was built because somebody still has to run the river company.

It is the compact, behind-the-scenes operating dashboard that gathers the gauge, snowpack, melt energy, weather, forecasts, activity thresholds and the current Black Hat experiment in one place.

The River Desk takes time to explain the river. The Control Room assumes the person looking at it already knows why a freezing night at Wolf Creek matters and would simply like all six weather feeds to stop hiding on different websites.

It helps us make the daily and seasonal calls that affect guide schedules, shuttle capacity, trip availability, guest communication and the transition among rafting, tubing and lake rentals.

 

Most visitors do not need it. Guides, hydrologists, emergency managers and the sort of people who consider six weather feeds a pleasant breakfast may disagree.

Consider it the unmarked door behind the front counter labeled:

Probably more information than you wanted.

The mountain still gets a vote

Turning experience into formulas did not mean throwing away what we learned by looking outside.

Local educator and river observer Josh Kurz has spent years studying the relationships among snowpack, dust, temperature and San Juan River runoff. His annual predictions and South San Juan Streamflow Reports helped frame many of the same questions the Black Hat now attempts to test.

Josh has also documented how dust can accelerate spring runoff. Dark dust absorbs more solar energy than clean snow, allowing a seemingly healthy snowpack to disappear faster than its stored water alone might suggest.

Then there are the snowfield indicators. As snow melts on Blackhead Peak, a shape resembling a jumping trout emerges from the mountain. Nearby, snow geese begin appearing on Quartz Ridge.

Local observers have long used those shapes as clues that the seasonal melt is progressing and the San Juan may be moving beyond its spring crest. The trout and geese do not replace SNOTEL stations; they are visible evidence of what is actually happening across the mountain landscape.

The tools are strongest when the formulas remain connected to the river and mountains they are trying to describe.

Why make our river brain public?

The River Desk is my public brain dump. It is also an affirmation of what Pagosa Outside has been doing behind the scenes for decades.

We have always gathered the best information we could find, added what we were observing on the river and in the mountains, and tried to give guests an honest answer. The Desk makes that process visible. The Black Hat Lab turns our more questionable future calls into public, testable experiments. The Control Room helps us use the same information to operate more consistently.

As far as we know, nobody else has assembled this exact combination of live data, locally specific activity ranges, seasonal river interpretation, future-date forecasting and publicly scored predictions for the San Juan River—or perhaps any river.

We could have kept it all inside the shop, but the questions were never limited to our paying customers. Locals want to know when to raft. Visitors want to know whether to pack for tubing. Anglers watch temperature and flow. Property owners watch high water. Ranchers, outfitters and other businesses all pay attention to what the watershed is doing.

The river runs through the middle of town. The useful information about it should not remain trapped inside my head.

That reflects a larger Pagosa Outside philosophy:

When the tool we need does not exist—or all the useful pieces are scattered across the internet—we try to invent a better one.

Then we keep learning from it.

Which river tool should you use?

Use the River Desk when you want to know what the San Juan is doing now, what the current flow means and what could change during the next 7–10 days.

Use the Black Hat Hydrology Lab when you want to ask about a future date, explore the next 90 days, follow the current seasonal experiment or watch Mother Nature grade our official predictions.

Use the Control Room when you want all the operational information at once and promise not to complain that it looks like a river company exploded across your screen.

The shortest explanation is:

The River Desk explains what the river is doing. The Black Hat Lab shows what we are still trying to learn. The Control Room helps us decide what to do about it.

Frequently Asked Pagosa River Questions

Is the Pagosa River running right now?

The river through Pagosa Springs is the San Juan River, and it flows year-round. What changes is the amount of water and which activities fit the conditions. Check the River Desk for the current CFS, direction and river story.

Is the river open in Pagosa Springs?

The San Juan River does not have one universal open or closed status. Pagosa Outside’s activities depend on the operating season, flow, water temperature, weather, debris, access and current channel conditions.

Are tubes running in Pagosa Springs today?

Pagosa Outside generally uses 30–399 CFS for DIY Tube Rentals and 60–399 CFS for Deluxe Tubing Trips. Flow is only part of the operating decision, so check the River Desk and current trip availability.

Is the river high enough to float?

That depends on what you mean by “float.” Rafts and tubes require different conditions. The River Desk translates the current flow into the activities and general river experience that fit.

Why does the San Juan River level change so quickly?

The San Juan through Pagosa Springs is a free-flowing mountain river without a predictable upstream dam release. Snowmelt, regional storms, nighttime temperatures and irrigation diversions can all cause noticeable changes. Because the primary gauge sits in town rather than far upstream, some changes are already in the operating corridor before they appear online.

What flow is needed for rafting in Pagosa Springs?

Splash & Dash generally begins around 300 CFS, Mesa Canyon around 600 CFS and West Fork around 800 CFS. These are useful flow references, not automatic guarantees that a trip will operate.

When does rafting season begin?

Pagosa Outside’s operating season begins May 1, but useful rafting water depends on snowpack, temperature and runoff timing. Some seasons arrive early. Others make everybody stare at the gauge for another week.

When does tubing season open?

Tubing generally follows the post-peak transition below approximately 400 CFS, after the river has warmed enough and other conditions fit. The river—not a fixed calendar date—provides the final answer.

Can we raft in Pagosa Springs in July?

Sometimes. A strong snow year or widespread summer rain can maintain or briefly restore rafting flow. Many seasons have transitioned toward tubing by July.

Will there be enough water during my vacation?

Nobody can guarantee the future flow of a free-flowing mountain river. Ask the Black Hat provides an experimental estimate for dates up to 90 days away. The River Desk becomes more useful as your date approaches.

When will the San Juan River peak?

Peak runoff depends on snowpack, temperature, nighttime refreezing, dust, wind and spring weather. Right Now in the Lab updates the developing estimate, while the annual game separately scores the official prediction locked before the crest.

The mountain gets the final vote

The San Juan will never become completely predictable. Late snow can rescue a disappointing winter. Dust can accelerate the melt. Cold nights can hold runoff back. Several warm nights can release it in a hurry. Irrigation demand can take a noticeable bite from low flow. A summer storm can miss the headwaters entirely—or bring the river back to life before breakfast.

For decades, I made the best calls I could with gauges, forecasts, irrigation patterns, local observations, historical memory and experience. The River Desk, Black Hat Lab and Control Room organize that process, make it useful to more people and preserve enough evidence to find out whether we were actually right.

They are not the end of our education. They are the latest step in Pagosa Outside’s continuing effort to find, develop and invent the tools that help us do our jobs better and give our guests a better day on the water.

We will keep refining the formulas, publishing the predictions and learning from the misses. We will also keep watching for the jumping trout on Blackhead Peak and the geese on Quartz Ridge.

Some parts of the old forecasting system are simply too Pagosa to automate.