Almost nothing about 2026 can be read off the figures alone. The District did not set the supply, the limit, the clock, or the order in which water had to be served. Those were fixed elsewhere — by Reclamation, by biological opinions, by adjudicated water rights, by a 1954 contract, and by the State of Oregon — and the District operated inside them. They are collected here so that every number in this record can be checked against the rule that produced it.

1. The supply was deemed available by Reclamation, not allocated by the District

On April 6, 2026 Reclamation issued its water availability determination for the Klamath Project: 221,000 acre-feet of Project Supply from Upper Klamath Lake and the Klamath River for the 2026 irrigation season. That figure is Reclamation’s. K.I.D. did not request it, agree to it, or divide it, and operating within it is not an endorsement of it.

For scale against the plan it came out of: Reclamation’s 2026 Annual Operations Plan set a goal of storing up to 350,000 acre-feet for agriculture, depending on hydrology. The hydrology delivered 221,000 — about 63% of a full supply.

The 221,000 acre-feet is Upper Klamath Lake and the Klamath River only. The Project’s other two reservoirs are accounted separately, and in 2026 both filled their allotments: Gerber and Clear Lake each received a full 35,000 acre-foot supply. Water from those reservoirs is not charged against the 221,000, which is why a dry year on the lake and a full year on the east side can be true at the same time.

Under its 1954 Operations Contract, K.I.D. is a distribution agent — it moves water to the lands entitled to it. The 1954 contract contains no apportionment clause. There is no provision in it that authorizes K.I.D., or anyone else, to divide a short Project Supply among the districts, and no agreed method for doing so exists. What happened in 2026 is that each district distributed and managed its own diversions inside a single Project-wide figure.

2. One supply, two clocks — and they do not end on the same day

This is the single most common source of confusion in the 2026 record, and it is worth stating plainly.

• The water year ran October 1, 2025 to September 30, 2026. Inflow, storage, outflow and the long-term rankings are kept on this clock. It closed September 30, 2026.
• The 2026 irrigation season ran March 1 to November 30, 2026, under one 221,000 acre-foot supply. That clock crossed the water-year boundary on October 1 and kept running, still bound to the same limit.

So deliveries made in October and November 2026 fall in water year 2027 but are charged against the 2026 supply. Through September 30 the Project had delivered 198,673 acre-feet; the balance of the season adds to that same total rather than starting a new one.

3. How Reclamation classified the year — and what it classifies on

Reclamation does not classify a Klamath Project water year on its annual total. It classifies on April–September net inflow to Upper Klamath Lake — the window that actually fills the lake and supplies the crop.

Measured that way, 2026 was the 4th driest of 65 years. Measured as a water-year total it was the 8th driest of 46 — an ordinary dry year. Both are correct, and the gap between them is the finding of 2026: roughly three-quarters of the year’s water arrived before April, when it does the irrigation season no good.

4. The biological opinions and the annual operations plan set the river first

Project operations in 2026 ran under the 2024 Biological Opinions and Reclamation’s 2026 Annual Operations Plan. Together they fix the Upper Klamath Lake elevations that must be held for the endangered suckers, the flow targets that must be met below Keno Dam for coho salmon, and the accounts that hold water for specific purposes. In practice those requirements are satisfied first, and Project Supply is what the rules leave over.

Three accounts shaped the 2026 figures and are often mistaken for farm water:

• Project Supply — the 221,000 acre-feet. This is the only account charged for irrigation delivery.
• Deferred Project Supply (DPS) — return flow through the Lost River Diversion Channel and F/FF pumping that the districts move and send back downriver to help meet the Keno targets. 47,430 acre-feet in 2026, with 4,113 acre-feet of balance still available on October 8.
• Flexible Flow Account (FFA) — water held for timed river releases. A specified proportion of the calculated Keno releases between October 1 and March 1 is held back and stored in Upper Klamath Lake, for use between March 2 and June 30. The WY2027 account had accrued 557 acre-feet by October 8.

How that reads on a single day. Reclamation’s Keno release target for October 8, 2026 was 760 cfs — the 750 cfs base flow, adjusted by the river model and reduced by about 29 cfs to fill the Flexible Flow Account. That is the mechanism in miniature: the river target is set first, and water held back for the spring comes off the top of it.

The boat dance flow is the clearest example of the distinction. Reclamation released about 7,000 acre-feet below Keno on August 28–30 for the Yurok Tribe’s biennial World Renewal ceremony. That volume was reserved earlier in the year under the biological opinions and the operations plan, held separately, and did not come out of the 221,000 acre-feet.

5. What counted against the limit — and what did not

The 221,000 acre-foot figure is an accounting rule as much as a volume, and the accounting changed during the season as Reclamation and the districts reconciled. For this record, deliveries charged against the supply are A Canal + Miller Hill + Station 48 + Klamath Drainage District (North and Ady), with three adjustments:

• Ady-to-Refuge deliveries are deducted — refuge water is not Project irrigation supply.
• September Anderson-Rose Dam releases are counted as refuge deliveries, not Project deliveries, and are deducted. This is the change that moved the reported September 30 total from 202,543 to 198,673 acre-feet; it is a change of basis, not a correction of the gage record.
• Klamath Drainage District came off the cap on October 1 and moved to its fall and winter water rights. Its diversions after that date are not charged against the 221,000 acre-feet.

On this basis the District’s tracker and Reclamation’s own Total Ag figure agree to within 25 acre-feet through October 7 — 205,025 against Reclamation’s rounded 205,000. The two also agree on what is left: Reclamation’s October 8 projection assumes roughly 16,000 acre-feet of remaining Project Supply, against the tracker’s 15,975. That agreement is the reason this basis is the one published here.

6. Water rights set who is served, and in what order

K.I.D. delivers to 73,650.6 contracted acres through the A Canal and the Miller Hill pumps, and its works can reach a further 60,964.1 acres under Article 14 of the 1954 contract — 134,615 acres in all, against contract volumes totalling 223,822 acre-feet. Three separate limits apply to those acres, and where they disagree the lowest one governs: the 3.5 acre-feet per acre adjudicated duty, each district’s own contract volume, and the 0.75 acre-feet per acre Reclamation directed for Warren Act contractors in 2026.

Seniority sets the order of service. It does not exempt anyone from a wait list — it decides where on the list they sit. The order in K.I.D. is VBDC (1883), then Henley-Ankeny (1884), then K.I.D. 13(a) “A contract” holders (1905), then Warren Act 13(b) and (c) “B contract” deliveries (1911).

Seasons of use are adjudicated too, and they are why the irrigation season outlasts K.I.D.’s own deliveries: Tulelake Irrigation District keeps diverting at Station 48 after K.I.D. has shut down, to the close of the season on November 30.

Every district and contract, with acreage, volume, priority date and 2026 limit →

7. The rules themselves were in motion during 2026

The 2026 season was operated under the 2024 Proposed Action and biological opinions, but the framework around them did not stand still. Reclamation completed an ESA reassessment in early 2026 and began a new consultation. In June 2026 a Ninth Circuit ruling led Reclamation to reexamine its approach to consultation and to long-term operations. Reclamation is planning near-term operations for the 2027 water year while that work proceeds, with the water users’ associations and the districts participating.

In Congress, H.R. 8259, the Federal Water Projects Consultation Improvement Act of 2026, was introduced in April by Rep. Cliff Bentz, with Rep. Russ Fulcher as cosponsor; it would give Reclamation contractors more opportunity to participate in consultations.

These are recorded here as facts about the year, not as the District’s reading of them. Questions about the litigation or what it requires should go to K.I.D.’s counsel, Nathan Rietmann.

8. The State of Oregon regulated the lake, and that governed the refuges

Diversion from Upper Klamath Lake is also subject to Oregon Water Resources Department regulation for the Klamath Tribes’ adjudicated water-right elevations. In 2026 that rule, not a District decision, set when water could move to the refuges: OWRD lifted regulation and water moved through the districts to the Tule Lake and Lower Klamath refuges August 14–24; on August 24 the lake reached the elevation at which regulation resumes and deliveries stopped.

9. The District’s own rules, which it does control

Inside all of the above, these are K.I.D.’s own rules, adopted by its Board and enforced under the K.I.D. Water Delivery Policy:

• Order water on, and order it off. Every time. Water orders run on a minimum 12-hour set and require 24-hour prior notice, because from Upper Klamath Lake water can take more than 96 hours to reach the far end of some canals — and it moves slower at low flow. Once water is on its way, it cannot be put back in the lake.
• Use water only on permitted acres, in a beneficial manner. Water running down roads, into drains, or sitting unused in ditches is a violation.
• Waste has consequences. Under Section 3 of the policy they escalate: written warning, then fines up to $5,000, then shut-off of delivery.
• Inter-account transfers were capped at 0.5 acre-feet per acre per month for 2026.
• SWAP, the District’s own idling program, closed enrollment July 1 with 1,022 acres off Project surface water for the season.

The Klamath Project Drought Response Agency offered voluntary idling programs this year (No Irrigation, Limited Irrigation, and a Five-Year option). The 2026 application window closed June 15.

KWUA reports that the DRA's 2026 program exceeded its enrollment goal — the largest in the agency's history — with more than 50,000 acres enrolled — an estimate the DRA last updated July 28. The DRA is still reviewing applications, verifying enrollment, and finalizing acreage on its online public map, where you can check the parcels conditionally approved for your account (parcels not yet shown may still be pending review). Enrolled patrons will receive response letters directly and can review their submitted acreage once the map is updated. For details, see the KWUA July 2026 WaterWorks newsletter.

Through June, before either program's acreage was known, K.I.D. held diversions low and ran a wait list that peaked at 57 orders on June 23 — a hedge against over-committing the supply while demand was unsettled. As SWAP (1,022 acres, closed July 1) and DRA enrollment firmed up, the wait list was worked down. Enrollment concentrated in the one-year options rather than the Five-Year program — growers intend to keep farming, treating idling as a short-term bridge.

Program details remain posted at klamathwaterbank.com; for current status, contact the DRA at (541) 630-0752 · info@klamathwaterbank.com.

Reclamation forecast the 2026 Project water supply from Upper Klamath Lake and the Klamath River at 221,000 acre-feet — well below the Project's full irrigation demand. That shortfall is why deliveries are managed the way they are this year.

What the basin produced this water year. The shortage begins in the mountains. Peak snowpack reached only 4% of median, water-year precipitation 78% of median, and the water that actually arrived in Upper Klamath Lake — 805,415 acre-feet of net inflow for the full water year — came in at about three-quarters of the 65-year mean. Two independent measures of the watershed, the same answer. Rain on the valley floor was not the problem. Reclamation's AgriMet stations finished the water year above their long-term medians — Lorella 140%, Beatty 123%, Klamath Falls 115%, Agency 112% — and the Klamath Falls airport gage recorded 10.25 inches against an 11.27-inch average. Valley rain is not what fills the lake; mountain snow and runoff are, and the snow did not come.

The state recognized the same conditions. On May 12, 2026, Governor Kotek declared a drought emergency in Klamath County (Executive Order 26-08), citing basin snow water equivalent at 4% of normal.

The Klamath River Basin went from almost no drought to half the basin in drought in a single water year. The U.S. Drought Monitor tracks the basin as its own unit. On September 30, 2025 — the first day of this water year — 54% of the basin was drought-free and only 0.32% was in drought at all. As of August 18, 2026: no part of the basin is drought-free, 50.5% is in drought (D1–D4), and 15% is in severe or extreme drought (D2–D3) — extreme drought having more than doubled in the past week alone. Klamath County as a whole is 88.4% in drought, affecting 58,657 people, with its 16th-driest January through July in 132 years, about 5.17 inches below normal.

Sources: U.S. Drought Monitor, Klamath River Basin (released August 20, valid August 18) and NIDIS Klamath County conditions. NOAA's seasonal outlook has drought persisting through November.

Where that water went. These are water-year figures from K.I.D.'s daily tracking file and the federal gages, current to September 16. The water year runs October 1 to September 30 — a different clock than the irrigation season.

Read those together. Of the water that reached Upper Klamath Lake across the full water year, about 76% passed over Keno Dam and went down the Klamath River. Project diversions over the same twelve months came to about 238,000 acre-feet — the 198,673 charged against the irrigation season inside the water year, plus roughly 35,000 diverted outside it on fall and winter water rights. That is about 30% of the year's inflow, and it means more than two and a half acre-feet went down the river for every one diverted by the Project — 611,270 against about 238,000. And 47,430 acre-feet of what went down the river was Project return flow: water the districts moved and sent back to help meet the Keno flow targets.

Net inflow and storage are Reclamation's measured record; Keno outflow is U.S. Geological Survey gage 11509500. Deferred Project Supply is Reclamation's accounting of Lost River Diversion Channel and F/FF pumping returns; the 47,430 acre-foot split between the two is as of August 13, and the balance remaining stood at 5,537 acre-feet on September 30. The District's Water Dashboard carries these figures live, with the full daily record and the long-term context behind them.

The Klamath Reclamation Project is operated by several irrigation and drainage districts, and each district distributes and manages its own share of Project water under its own authorities. K.I.D. cannot control every basin-wide program or every diversion in the Project. What K.I.D. can do is manage its own operations, reduce avoidable demand, account for every drop put to beneficial use, and enforce its Water Delivery Policy against waste.

Operating within the supply figure Reclamation announced is not an endorsement of it. K.I.D. is operating within the constraints placed on the 2026 supply while continuing to press — through its contract and the proper channels — for the water its patrons are owed. The contract basis is in the rules.

Beyond managing wait lists, K.I.D. used every Board-authorized tool to reduce demand and prevent waste. The District’s standing rules are in the rules; these were the 2026-specific measures:

• Self-regulated the daily diversion from Upper Klamath Lake, holding the A Canal headworks and Miller Hill to a K.I.D.-specific average rate, coordinated with the other diversion districts.
• Lowered the A Canal operating level by about 6 inches where feasible, cutting seepage through the canal banks.
• K.I.D. evaluated whether future inter-account transfers should be ineligible for DRA validation.
• Ran every recirculation pump upgraded by K.B.I.D., which invested nearly $1 million; those pumps target 0.75 AF/acre.
• Applied periodic return-flow adjustments to Diversion Works volumes through the season.

In short, three things at once: cut the demand K.I.D. must divert, get more value from every acre-foot, and eliminate waste.

Surface Water Abatement Program (SWAP) — 1,022 acres idled (enrollment closed July 1)

SWAP is K.I.D.’s own idling program. It paid patrons who could take enrolled acres completely off K.I.D. surface water for the rest of 2026 using an alternate source they control — freeing up wet water, not paper water, so the District could keep irrigating later into the season.

Regardless of how the year unfolded, the districts worked together through 2026 toward a single shared vision — and it held:

• Keep water moving as long as we can. The District made every effort, and used every available resource, to keep deliveries going for as long as the supply lasted. With September rains, demand fell faster than the supply did, and deliveries carried past September — K.I.D. reached October 15, the goal its Board set in April, with the available supply committed in full.
• Stay within the announced supply. The Project has operated within the 221,000 acre-feet Reclamation made available for 2026, and is projected to deliver it in full by the close of the season on November 30.
• Use every tool we have. With the idling enrollment windows closed by July, how far the supply carried turned on disciplined daily diversion management and the acres idled.
• We stretched every drop. The math did not work on paper. But the District used every resource it had to squeeze that math toward a better result for the Project — and that work showed in the numbers.

Two things hit the refuges at once this summer. An outbreak of avian botulism at Tule Lake got much worse in the heat. At the same time, Lower Klamath and Tule Lake had their biggest waterbird hatch in more than ten years. Water was the tool for both — it dilutes the disease, and it keeps young birds alive until they can fly. The water that moved through the districts in August did both jobs.

Avian influenza — suspect H5N1 in the basin

Tule Lake Refuge — botulism outbreak, responding to water

Water remains the primary tool for both botulism management and wildlife survival, and this summer demonstrated it: moving water through the wetlands raised sump levels and cut daily mortality within days. A flow-through arrangement is a concept, not a finished plan. K.I.D. proposed such an approach in 2023, as the Section 7(a)(3) Applicant for the 2024 Proposed Action, and it was not adopted by the agencies. It would need engineering, accounting, and agreement among the districts, Reclamation and the wildlife agencies before it could be operated — work that has not been done. In the meantime the refuges depend on what the irrigation system returns to them.

Lower Klamath and Tule Lake National Wildlife Refuges — the largest hatch in a decade

Both refuges hosted the largest waterbird hatch our basin has witnessed in over ten years, and thousands of young birds were on the water when the sumps were at their lowest. Unit 2 reached its required elevation in time. Refuge crews banded more than 3,000 juvenile and flightless adult birds over two nights — birds that would very likely have been lost without water under them. At Lower Klamath, Unit 3 went nearly dry earlier in the season and habitat there was lost; return flows are now being moved to Unit 3 through the D plant to raise its elevation ahead of winter. The immediate priority on the refuge is securing and keeping enough water on the ground to sustain these nesting areas and ensure this historic generation of birds can successfully fledge.

Fall migration has begun, and early migrants are arriving and feeding in Klamath Drainage District and on leased fields. Unit 3 is being topped off with deep-plant water from Tulelake Irrigation District, and Unit 2 is holding steady. Sump 1A and 1B elevations have leveled off, with deep-plant pumping continuing.

Lower Klamath is open for waterfowl hunting — made possible by the water held in Unit 2 — and the California youth waterfowl season this weekend is bringing significant recreation to the basin. Return flows from District operations continue to move to Unit 3 through the D plant to raise its elevation ahead of winter — again, ordinary operation of the system rather than a special delivery. Unit 2 levels are declining again, because evapotranspiration is drawing more off the unit than the water now reaching it.

The deficit is the whole story: the Thomas Transfer's 12 cfs delivers roughly 24 acre-feet a day, while evapotranspiration is estimated to take about 30 acre-feet a day off Unit 2 alone. The transfer is a helpful tool, but at that rate it does little more than hold a single unit even — it leaves nothing to raise water levels, nothing for Unit 3, and nothing for the rest of the refuge. The 30 acre-feet per day figure is a conservative estimate; actual losses across the refuge are likely higher, which widens the gap further. That gap is the urgent case for more robust water solutions on the refuges. K.I.D. continues to work with its partners toward solutions that maintain a healthy ecosystem.

What worked in 2025 — and why it matters now

Last season showed what proactive water management can do. After the 2024 botulism outbreak, refuge staff used new elevation data to identify which wetland areas were most at risk at which water depths, and alerted Tulelake Irrigation District and Reclamation early. Operations were adjusted, and daily mortality at Tule Lake fell from 100–200 birds a day to one or two. It is the same principle K.I.D. has pressed for here: move water through the wetlands before the disease takes hold, rather than after. Ducks Unlimited documented that effort in Engineering a Comeback: How Science and Collaboration Offer Hope for Klamath Basin.

Two different diseases hit two different life stages of Klamath salmon in 2026 — juveniles in the spring, returning adults in late summer. Both track the same variable: water temperature.

Spring — Ceratonova shasta in the juveniles

C. shasta is a parasite that lives in the main river. It gets worse as the water warms. It first showed up on April 2, and by late April 90 to 100% of weekly samples carried it. By then most wild fish were already gone: of about 580,000 young wild Chinook counted this season, an estimated 80% had left before that first detection, while the water was still cold.

The early-June die-off that drew coverage fell almost entirely on hatchery fish — part of the 675,000 smolts released May 14 from the Fall Creek Hatchery, about 200 river miles from the sea. Roughly 2,170 of 3,500 juveniles caught at the Iron Gate and I-5 traps June 2–5 were dead or dying, nearly all of them hatchery-origin. Spore densities in the upper corridor were already above the 10 spores per liter the monitoring program ties to a 40% juvenile mortality threshold when those fish went in.

Coho salmon — the listed species the spring flow rules exist to protect — were essentially absent from the mainstem traps all season. They are a cool-tributary fish, not a warm summer mainstem fish. Spore levels dropped off at every index site by early August. K.I.D. has set out the full monitoring record, the temperature data and the sources behind this on its 2026 C. shasta and Klamath water temperatures page.

How the year finished on the parasite. Sampling continued through the fall at five mainstem index sites — near the Shasta River, near Beaver Creek, Seiad Valley, Tully Creek and Orleans. Across the five weeks from August 31 to September 28, C. shasta was detected at every site, every week — it is a natural resident of the river, and its absence was never the expectation. What matters is the level: no site exceeded 5 spores per liter at any point in those five weeks, against the 10 spores per liter the program ties to a 40% juvenile mortality threshold. The upper two sites ran higher than the lower three, as they do. Turbidity rose at the end of August behind the higher Boat Dance flows, and again at the end of September with the first seasonal rain.

Late summer — Ich in the returning adults

Adult salmon coming back from the ocean piled up in the lower Klamath through August, and the conditions turned against them. Reclamation measured the water at 23.2°C — about 74°F — with flows near 2,220 cfs. Every fish sampled had Ich, mild in most and severe in some, with signs of columnaris too. Ich is the parasite behind the 2002 fish kill. It spreads when crowded fish sit in warm, low water.

On August 27 Reclamation approved a Flow Augmentation Release under the Long Term Plan. Flows below Lewiston Dam on the Trinity River rose from about 450 cfs to as much as 1,200 cfs from August 28 through September 21, using approximately 35,000 acre-feet from Trinity Reservoir — a reservoir sitting at 109% of its historical average for the date.

Reclamation stated that this release does not affect water supply to agriculture, communities, or other environmental obligations. That statement applies to the Trinity release. It does not describe the Klamath side of the system: flows past Keno Dam are supplied from Upper Klamath Lake, including released stored water, from the same lake that supplies the Project.

The scale is worth noting. The 35,000 acre-feet released from Trinity for fish health is four times the 8,684 acre-feet remaining for every farm in the entire Klamath Project — and Trinity sits at 109% of average while the Project was held to about 44% of its full contract entitlement.

The fall run came in — and the quota filled

Against the disease picture, the fishery itself did well. Anglers met the Lower Klamath adult fall-run Chinook quota below the Highway 96 Bridge at Weitchpec on September 13, which under California's rules closes the adult Chinook fishery from that bridge to the mouth. The mouth itself closed September 2 and stays closed for the rest of the year.

Below Highway 96 the river remains open for jack Chinook — two-year-olds of 23 inches or less — with all adults released. Adult fishing continues above Weitchpec and on the whole Trinity River. Quota status is on the California Department of Fish and Wildlife hotline, (800) 564-6479.

A quota filled is a run that arrived. It does not undo what the warm water did to the fish holding below Weitchpec in August, but it is worth recording alongside it: the 2026 fall run supported its full recreational allocation on the lower river.

Where the fall run went — the tributary counts

The quota tells you the run arrived. The weirs tell you where it went. California Department of Fish and Wildlife counts through October 15:

Zero coho were counted at any of the five weirs. Coho are the listed species the spring flow rules exist to protect, and they were absent from the mainstem juvenile traps all season as well — they are a cool-tributary fish, and the mainstem is not where they are found.

The small tributaries filled in late. Between October 8 and October 15 the Shasta added 40 fish, but Scott River went from 10 to 40, Shovel Creek from 12 to 30, Bogus Creek from 4 to 11 and Jenny Creek from 1 to 10. Two of those weirs are new ground: Shovel Creek and Jenny Creek sit above the former dam sites and are counted now because the dams are gone. Counts are preliminary and do not include spawning observed downstream of the stations; final escapement follows once the data is complete. Figures from the California Department of Fish and Wildlife in-season update of October 16, 2026.

Cyanotoxins — low in the river, high in the sump

Blue-green algae toxins are often raised alongside fish health on the Klamath, so it is worth recording what the monitoring actually showed through August. Tribal programs sampled the lower river and the Environmental Protection Agency's Region 8 Laboratory ran the analyses.

The river ran low. The sump ran high. Across twenty-five river samples in August, most were below the reporting limit and the highest was 1.36 micrograms per litre. Twelve more samples in September came in lower still — the highest was 0.73, and every one was below the 0.8 microgram-per-litre level California uses to post even a Caution advisory. At the same time Tule Lake Sump 1A measured 8.84, carrying a California WARNING advisory that has been posted since May.

That contrast is the lesson, and it is the same one the botulism taught. The toxin problem was not in the moving water; it was in the warm, shallow, stagnant sump — the same conditions, in the same place, that produced the botulism outbreak. Water that moves does not grow these problems. Water that sits does.

For the fish themselves, the finding stands: the harm to salmon holding in the lower river came from warm water, low flow and the parasites that thrive in those conditions, not from cyanotoxins in the river.

A second kind of algae — and it is not in the Project

The microcystin figures above come from water-column grab samples, which pick up toxins from planktonic blooms drifting down from upstream lakes and reservoirs. A separate program run by the California Water Boards with the Klamath Tribal Water Quality Consortium monitors something different: benthic cyanobacteria — mats that grow attached to the streambed, whose toxin of concern is anatoxin-a rather than microcystin, and which concentrate in the mat material rather than in the water.

In September that program recommended yellow-bannered “Toxic Algae Alert” signage at three locations. The passive-sampler results were all well under the 100 nanogram-per-gram threshold; it was the estimated percent coverage of the mats that triggered the recommendation.

All three are tributaries — the Trinity, the Scott, and a Scott tributary. None is on the Klamath mainstem, and none is in the Klamath Project or in the reach the Project delivers to. The Quartz Valley Indian Reservation separately posted a Toxic Algae Alert on Kidder Creek based on its own grab samples. The Yurok Tribe holds posting authority on its own lands and its postings are not included here.

A posting can be lifted when visual surveys or mat samples show no toxigenic mats or toxins across two consecutive sampling events, or when cooling weather and flushing rain break the mats up and conditions stop new growth.

Cyanotoxin results: U.S. EPA Region 8 Laboratory for the R9 Tribal HAB 2026 program (Yurok, Karuk, Hoopa Valley and Pulikla monitoring), and refuge and Gerber sampling reported to the Klamath blue-green algae listserv. Benthic cyanobacteria monitoring and posting recommendations: California North Coast Regional Water Quality Control Board with the Klamath Tribal Water Quality Consortium. Other sources: Reclamation news release and Lower Klamath River Flow Augmentation Release technical recommendation, August 27, 2026; Oregon State University water-column spore monitoring; USFWS Arcata outmigrant monitoring; Karuk Tribe temperature record, used with permission. Figures are provisional.

Across the 2026 irrigation season, 121,235 acre-feet left the Project at Keno Dam beyond what the watershed delivered into Upper Klamath Lake. That water did not come from runoff. It came out of water already in storage.

From May 1 through September 30 — 153 days — net inflow to Upper Klamath Lake totaled 130,083 acre-feet. Measured releases over Keno Dam totaled 251,318 acre-feet. The river carried away 121,235 acre-feet more than the lake received, and on 125 of those 153 days outflow exceeded inflow.

For scale: 121,235 acre-feet is more than half the entire 221,000 acre-foot supply the Klamath Project was permitted for 2026, and more than K.I.D.'s whole-season deliveries through the A Canal and Miller Hill combined.

Month by month, the gap grew as the season ran

May is the month that looks normal, and it is the one that explains the rest. Spring runoff was still arriving, the lake nearly kept pace, and only 24% of what left at Keno came out of storage. Then the runoff stopped.

June, July and August are the case. Inflow collapsed to 15,384 acre-feet, then 7,051, then 6,027 — while releases held at 51,896, 41,018 and 46,754. In August the lake received 6,027 acre-feet and sent 46,754 down the river: nearly eight acre-feet out for every one in. Across those three months roughly four of every five acre-feet sent down the Klamath came out of storage rather than out of the watershed — and those are precisely the months when Project farms were being told there was not enough water, when the wait list ran to 57 orders, and when 1,022 acres sat idle under SWAP.

September is the control. Fall inflow arrived, the lake took in 51,250 acre-feet against 45,495 released, and for the first month all season the lake took in 5,755 acre-feet more than the river carried away. Storage still fell — diversions continued — but the deficit against the river closed. Nothing about the operation changed; the hydrology did. That is what the other four months look like when the watershed is actually producing water — and it is why the four shortage months total 126,990 acre-feet beyond inflow while the season nets to 121,235.

The lake is the evidence

The inflow-and-outflow comparison is an accounting argument. The lake itself is the physical one. Upper Klamath Lake lost 331,262 acre-feet between April 30 and September 30 — from 551,100 down to 219,838. That is not a model or an estimate. It is Reclamation's own daily measurement of how much water was in the reservoir.

Read the two charts together and the argument closes. The river was running a deficit against inflow in all but one month; the lake was falling the whole time; and the lake is where the deficit was made up. The water that left at Keno beyond what came in had only one place to come from.

Where the lake's water went

Upper Klamath Lake held 551,100 acre-feet on April 30 and 219,838 acre-feet on September 30 — a drawdown of 331,262 acre-feet across the irrigation season.

Put plainly: for every four acre-feet the Project delivered to a farm this season, nearly three acre-feet of stored water went down the river. Project farms operated under a hard 221,000 acre-foot ceiling, enforced to the acre-foot and accounted for daily. The releases at Keno operated under no equivalent limit.

Why 121,235 is the floor, not the ceiling

The obvious objection to a two-point comparison is that water enters the river between the lake and Keno — so some of the gap might be local runoff rather than storage. In 2026 the reach ran the other way. Reclamation's own accounting of net accretions in the Keno impoundment totals −33,447 acre-feet between May 1 and September 30: across the irrigation season the reach was a net loss, not a net gain.

The same point shows up at the lake itself. Link River Dam released 383,764 acre-feet between May 1 and September 30 against 130,083 acre-feet of net inflow — 253,681 acre-feet more than the lake took in. Much of that is re-diverted to Project use below the lake, which is why the figure measured at Keno is less than half of it. Either way, the 121,235 acre-feet measured at Keno understates what left storage; it does not overstate it.

The lake's own storage record closes the balance independently: net inflow less Link River releases less A Canal diversions accounts for the measured 331,262 acre-foot drawdown to within 435 acre-feet over 153 days — about one-tenth of one percent.

Two numbers that sound contradictory — and are not

Measured across the full water year, outflow over Keno (611,270 AF) came in 194,145 acre-feet below net inflow to Upper Klamath Lake (805,415 AF). Read alone, that looks like the river never drew on storage at all. It is a true figure, and it is the one most often cited.

It is also the wrong clock. An annual total nets the spring fill against the summer draw and reports only the remainder. The lake fills in March and April and is drawn down from May forward, so the annual figure averages away the entire period under dispute. The irrigation season is when the shortage is felt, when crops are lost, when wait lists run, and when water left the lake for the river faster than it arrived. On that clock the answer does not net out: 121,235 acre-feet.

The Project is adding water to the river, not taking it

Across the water year, measured outflow over Keno was 611,270 acre-feet against an estimated natural outflow — what the river would have carried with no Project in place — of 570,291 acre-feet. Project operations put 40,979 acre-feet more down the Klamath than the unaltered basin would have delivered, 107% of natural. Of that, 47,430 acre-feet is Deferred Project Supply: return flow through the Lost River Diversion Channel (31,440 AF) and F/FF pumping (15,990 AF) that the districts moved and sent back to help meet the Keno targets. Agriculture drew 41,990 acre-feet of that credit back into use, leaving a balance of 5,440 acre-feet when the water year closed — the recirculation was not a paper entry.

The Klamath Project is not a net drain on the river below Keno. It is a net contributor — and it contributed while operating at 221,000 acre-feet against an estimated demand of 390,460.