Guiding the South Fork of the Snake

Four hundred and fourteen cubic feet per second went into Jackson Lake on the evening of 27 July 2026. Four thousand nine hundred and fifty came out. Twelve times as much water left the reservoir as entered it, and the ground between the two gauges is 321 square miles of Wyoming that would have to be producing a water yield of 14.1 for that to be a river doing it. Nothing produces 14.1. The gauge below the dam is not measuring a river; it is measuring a delivery, and everything downstream of it belongs to the same system. Comparable water is covered throughout the iconic fisheries writing.

A yield that cannot be a river

6.134 cubic feet per second per square mile at Moran.

The station below Jackson Lake Dam was carrying 4,950 cubic feet per second on a published drainage area of 807 square miles. That works out at 6.134.

The highest natural figure this series has recorded anywhere in fifty states is 4.54, from a volcanic spring system in Oregon fed by an aquifer, and the next two are a Wind River snowmelt creek at 2.77 and a temperate rain forest at 2.51. Moran is thirty-five per cent above the record, in a dry state, at the end of July.

A number that far outside the distribution is not a discovery. It is a signal that the thing being measured is not the thing the number appears to describe.

What it describes is storage being released on a schedule. The reservoir behind that gauge holds water through the runoff and lets it out through the irrigation season, and the gauge reads the release.

Art of Fishing, MI

Do the subtraction and it becomes obvious

321 square miles would have to produce 4,536 cubic feet per second.

The station above the lake at Flagg Ranch drains 486 square miles and was carrying 414 cubic feet per second, a yield of 0.852. That is a perfectly ordinary mountain figure and it is what the country up there actually produces.

Moran drains 807 square miles. The difference between the two stations is 321 square miles of catchment, and the difference in flow is 4,536 cubic feet per second.

Divide one by the other and the intervening ground would need a yield of 14.1, which is three times the highest number ever measured in this series and roughly seventeen times what the station immediately upstream was managing.

That is the whole argument, and it takes two gauges and one division. Anyone can run it, and running it is a better habit than trusting a single reading, because a single reading has no way to tell you it is describing a valve.

This dam does not make cold water

19.1 degrees going in, 19.8 coming out.

The two live temperature readings on this page bracket Jackson Lake. Above it the Snake was 19.1 degrees Celsius; below it, 19.8. The reservoir returned the water seven tenths of a degree warmer than it received it.

That is worth stating plainly because the assumption running through most writing about dams is the opposite. On the Bighorn a reservoir released water seven degrees colder than its own tributary; on the Green in Utah it released at 15.4 in a desert.

The difference is where the outlet sits. A deep reservoir drawing from the bottom releases the cold layer; a shallower one, or one drawing nearer the surface, releases roughly what it took in and often a little warmer for having sat in the sun.

So a guide cannot reason from the word tailwater. Two structures that look identical on a map produce opposite results, and the only way to know which you have is to read the gauge above and the gauge below.

Two dams, and the river peaks at Heise

414, then 4,950, then 12,600, then 13,200.

Below Jackson Lake the Snake picks up the country between Wyoming and Idaho and then meets a second storage project. At Irwin it was carrying 12,600 cubic feet per second off 5,225 square miles, which works out at 2.412. At Heise the figures were 13,200 and 5,752, giving 2.295.

Both of those figures would be near the top of the national distribution if they were natural, and neither is. They are the combined output of two reservoirs operating for delivery in the middle of the delivery season.

Heise is the peak. It is also, not coincidentally, the bottom of the reach that this river is famous for, and the reach where most of the guiding happens.

Which places the well-known water in an exact position: it sits at the top of an irrigation distribution system, and its summer volume is the schedule of that system rather than the yield of its catchment.

The plain takes back fifty-seven per cent

13,200 at Heise, 5,650 near Blackfoot.

Between those two stations the river loses 7,550 cubic feet per second, fifty-seven per cent of everything it was carrying, while its drainage area grows from 5,752 to 11,310 square miles.

The catchment nearly doubles and the river falls to well under half. Yield collapses from 2.295 to 0.500.

The loss is in two steps. Heise to Menan takes out 3,720 cubic feet per second, about twenty-eight per cent. Menan to Blackfoot takes out another 3,830, about forty per cent of what was left.

This is the Snake River Plain, and it is one of the most heavily irrigated landscapes in the country. What arrives at Heise as a fishery leaves the plain as a series of headgates.

Peak flow and peak fishery are the same place by accident

The best-known reach ends exactly where the withdrawals begin.

It is easy to read the coincidence backwards and conclude that the fishing is good because the flow is high. The causation runs the other way round: the flow is high there because that is the last point before the water is spoken for.

That distinction matters for anyone planning a business rather than a trip. The reach is not protected by its quality; it is upstream of the demand, and its condition is set by how much has to move through it and when.

A year in which the system delivers early and hard looks different from one in which it holds back, and neither is predictable from weather. Both are visible in reservoir reports months in advance.

That is the same working habit this series has arrived at on every managed river. Read storage, then read the delivery schedule, then read the forecast, in that order.

A two-year rule cycle and seven regions

Idaho publishes 2025 to 2027 fishing rules, with the Upper Snake as its own region.

The state's fishing rules run on a cycle labelled 2025 to 2027, and the state divides its waters into seven regions: Panhandle, Clearwater, Southwest, Magic Valley, Southeast, Upper Snake and Salmon.

Each region carries its own maps and special-rules waters section, and the Upper Snake section is the one that governs this river.

A multi-year cycle is unusual and it cuts both ways for an operator. It means the rules are stable for longer than in states that publish annually, and it means a change, when it comes, arrives in a bigger block.

It also means special-rules waters are the operative document rather than the general rules, and the specific reach you sell has to be looked up rather than assumed. Confirm the current rules for your water with the agency before the season, because a cycle boundary is exactly where a stale assumption becomes an expensive one.

One station publishes no catchment at all

Menan reports flow and stage and no drainage area.

The gauge near Menan returned 9,480 cubic feet per second and a stage of 5.07 feet, both current to the minute, and no published drainage area. Without that denominator there is no yield figure for it, and the table above says so rather than estimating one.

It is a useful reminder that the gaps in a public dataset are not all the same kind of gap. A dark sensor is a missing measurement; a missing drainage area is a missing constant, and it silently removes a whole column of analysis.

Both are recoverable if you notice them and both are invisible if you do not, because a calculation that quietly skips a station looks exactly like one that includes it.

The habit that catches it is counting. Six stations went into this page and five yields came out, and the difference is the thing worth explaining rather than hiding.

Four thermometers, all dark since 2015

Irwin, Menan and Blackfoot last reported temperatures in October and November 2015.

Three of the four Idaho stations on this page served water temperatures stamped 2015, and the fourth served none at all. Not one Idaho reading on this river is usable.

Read together with the neighbouring system, where every station was similarly dark, the pattern is regional rather than incidental. Discharge and stage are maintained across this basin; temperature is not.

That absence has a real cost. On a river whose flow is entirely managed, temperature is the one variable that would tell an operator something the release schedule does not, and it is the one variable the network will not supply.

The response is the same as on the Henrys Fork: keep your own log, on named reaches, and cite it as your own rather than dressing it as public data.

Fish The Fly, WY

The reservoir is warmer than the river above it

Seven tenths of a degree added, not taken away.

Most of the temperature effects measured in this series run the other way, so the Jackson Lake figures are worth dwelling on. Water arrived at 19.1 degrees and left at 19.8.

A body of water sitting at altitude through a July afternoon absorbs heat at the surface. If the outlet takes from near that surface, the release carries the heat with it, and the reservoir behaves as a warming device rather than a cooling one.

Which means the two Wyoming stations on this page are, in temperature terms, effectively one reading. The reservoir is not creating a fishery here; it is creating a volume.

Separating those two effects is the single most useful thing to understand about managed water. A dam can give you cold, or water, or both, or neither, and only the gauges tell you which.

Volume without temperature is half a fishery

13,200 cubic feet per second says nothing about whether fish will move.

The Idaho reach has extraordinary volume by any national standard and no usable temperature data at all. That combination shapes what an operator can honestly promise.

Flow can be cited, checked and planned around. It determines access, boat handling, wading and safety, all of which matter. It does not determine the thing a client cares about most.

An operation that quietly lets high flow stand in for good conditions is making a claim it cannot support, and on a river this heavily managed the two genuinely come apart: a delivery peak is not a fishing peak.

The honest version separates them, and it is more persuasive for doing so. Saying that the water is abundant and that you measure temperature yourself is a stronger position than implying the first proves the second.

A basin measured in headgates

Two reservoirs above, and 7,550 cubic feet per second removed below.

Read the whole page as one system and it resolves into three parts: storage at the top, a short reach where the water is briefly a river, and a plain that consumes more than half of it.

The middle part is the entire commercial fishery, and it exists in the gap between two engineering functions rather than as a feature of the landscape.

That is not a criticism of the water, which is genuinely good. It is a description of why it is good, and it explains why its condition tracks a schedule rather than a season.

Anyone who intends to spend a career here is better served understanding that from the start than discovering it in a difficult year, and the arithmetic that shows it is all on this page.

Two states, two offices, one radar

Moran answers to Riverton, Heise to Pocatello, and KSFX covers both.

Pull the grid data for either gauge and the national forecast service puts Moran in Riverton office territory, grid 49 by 156, zone WYZ013, referenced to Kelly, Wyoming. Heise falls to Pocatello, grid 137 by 78, zone IDZ053, referenced to Ririe, Idaho.

Different states, different offices, different zones, and one radar at Idaho Falls doing the work for both of them.

What none of those products describe is the volume in the channel, which on this river is the variable that matters most and is not weather at all.

A client checking a forecast for a trip here is reading the least informative available input. Saying so, and pointing them at the gauge instead, is a small piece of service that almost no operation offers.

What a delivery system asks of an operator

Learn the schedule, because the schedule is the river.

On a natural river a guide's expertise is largely biological and observational. Here a substantial part of it is administrative: knowing what is stored, what is committed, and when it moves.

That is unglamorous and it is also the actual edge. Two operators with identical skill on the water will differ mostly in whether they can tell a client in March what August is likely to look like.

The information is public, it is published well ahead, and reading it is a habit rather than a talent. The barrier is that nobody in this trade thinks of it as their job.

Anyone building here should treat that as the core competence rather than a nice extra, alongside the ordinary groundwork of credentials and pricing, and rules move, so verify the current position with the agency each season. The floor is set out in the state rate benchmark and in what each state asks of a guide.

What the numbers say about where to sit

Above the withdrawals, and honest about why.

Every figure on this page points to the same conclusion. The stretch between the second dam and Heise carries the most water, sits above the demand, and is where the trade has concentrated for good arithmetic reasons.

The competitive consequence is that it is busy, and an operation entering now is entering a crowded market on a short reach. The alternative is not downstream, because downstream the water leaves.

So growth here comes from the calendar and from the trip rather than from new water, which is the same structural answer this series reached on another single-reach basin for entirely different physical reasons.

Longer and better-designed trips are the usual route, and they price differently. Revenue follows from how a multi-day trip is put together, and from how far ahead this kind of client commits.

The lesson generalises past this river

A number outside the distribution is a question, not a fact.

The most useful thing on this page is not the 6.134. It is the reflex of noticing that 6.134 could not be true and then finding out what it actually meant.

Across this series that reflex has caught a release, a diversion, a stale sensor, a no-data sentinel and a missing constant, all of which would otherwise have become confident sentences in somebody's marketing.

The check costs one division and a comparison against the range you have already seen. It does not require expertise so much as the willingness to distrust a convenient number.

For a guide, that habit is worth more than any single fact it produces, because the alternative is repeating something plausible to a client who can look it up. The route into fly-fishing guiding teaches almost none of this, and it should.