Now It's 75 Percent: The September El Niño Forecast and the Mokelumne Clock

Weather
Written By Don Bradford
Now It's 75 Percent: The September El Niño Forecast and the Mokelumne Clock

Now It's 75 Percent: The September El Niño Forecast and the Mokelumne Clock

Lodi411 / LodiEye — Civic Climate Brief — September 2026

Updated with NOAA CPC, ECMWF, EBMUD, and California Data Exchange Center data through September 21

Summary

This brief updates our August 20 report, From Very Strong to Historic, which put NOAA's chance of a record-setting El Niño at 69%. On September 10 the Climate Prediction Center raised that figure to 75% for October through December, and raised the chance of a very strong event in the same three months from 95% to 98%. The weekly warming-adjusted index crossed the very strong threshold in mid-September, which means the event is no longer forecast to be historically large — it is now measured as the strongest on record for this point in its life cycle, with its peak still ahead.

The California side of the forecast also firmed up, with an important local wrinkle. NOAA's September 17 seasonal outlook tilts October–December wet only across the southern half of the state; the signal does not reach the northern San Joaquin Valley until the November–January season and later, where it peaks at a 50–60% chance of above-normal precipitation. NOAA said plainly that it may raise those California numbers on October 15. The European ensemble is already further along: more than a 70% chance of a winter in the wettest fifth of the record, statewide.

The most concrete local development is not atmospheric. It is a deadline. Under the 1962 agreement between EBMUD and the Army Corps of Engineers, Pardee and Camanche Reservoirs must hold between 130,000 and 200,000 acre-feet of empty flood space by November 5. As of the latest state readings the two reservoirs hold about 72,000 acre-feet of space, and current releases are draining the system at roughly a tenth of the rate needed to close that gap. Lodi should expect substantially higher Mokelumne flows through October. That schedule, published and checkable, is the single most useful local indicator of how much room the river has when the storms arrive.

What Changed in September

NOAA Raised Both Probabilities

The Climate Prediction Center's September 10 diagnostic discussion kept the El Niño Advisory in place and moved both of the numbers this series has been tracking. The chance of a historic event during October–December — defined as a three-month RONI of +2.5°C or higher, a level that would exceed every El Niño in the record back to 1950 — rose from 69% to 75%. The chance of a very strong event in the same season rose from 95% to 98% (NOAA Climate Prediction Center, as tabulated by The Watchers). The headline synopsis is unchanged in wording and stronger in substance: greater than a 90% chance of a very strong event during Northern Hemisphere fall and winter 2026–27, with El Niño expected to keep strengthening through the end of the year.

Two terms, one report

RONI is the Relative Oceanic Niño Index. It measures how much warmer the central Pacific is than the rest of the tropical ocean, which strips out general global ocean warming and leaves the El Niño signal itself. It generally reads lower than the traditional index, so the 75% historic probability is calculated on the conservative scale, not the flattering one.

Very strong is the highest official strength category, set at a three-month index of +2.0°C or above. Historic is the threshold NOAA introduced in August at +2.5°C or above. "Super El Niño," which appears in much of the national coverage, is not an official category at all.

The ocean measurements behind the revision came from August. The monthly Niño index values were +1.8°C in Niño-3.4, +2.5°C in Niño-3, and +3.4°C in Niño-1+2, the coastal South American waters — each a further rise from July's +1.4°C, +1.7°C, and +2.9°C. Sea surface temperature anomalies exceeded +3.0°C across the eastern equatorial Pacific. Subsurface anomalies remained above +10.0°C at depth, the thermocline stayed deeper than average, and both Southern Oscillation indices remained negative. One index went the other way: Niño-4, in the western part of the monitoring region, fell to +0.1°C, leaving the warmth concentrated sharply toward the eastern Pacific and the South American coast.

On September 14, CNN reported that the latest weekly RONI value had reached 2.0°C above average, crossing the very strong threshold and making this the strongest El Niño on record for its stage of development — the first event ever tracked with the warming-adjusted index, which NOAA implemented in February (CNN). The distinction matters for how this report reads: in April this was a probability, in July a fact, in August a record forecast. In September it is a record measurement with its peak still months away.

The Event's Full Arc: NOAA's Very Strong Probability by Season

FIGURE 1. Official CPC probability that the three-month RONI reaches +2.0°C or higher — the very strong category — for nine overlapping seasons, from the September 2026 issuance. Season labels use first letters of the three months (OND = October–November–December). Source: NOAA CPC official ENSO strength probabilities, issued September 2026.

Read left to right, Figure 1 is the schedule of this winter. The event is near certain at very strong strength from September through January, holds a 75% chance through December–February, and then falls off quickly — 39% for January–March, 5% for February–April, and effectively zero by spring. Lodi's exposure is therefore front-loaded into the period that also carries the Mokelumne flood reservation and the winter's highest tides. The event does not linger into a second wet season; the pressure lands in a roughly four-month window.

The Model Ensemble Has Not Backed Down

As of publication, the IRI/Columbia plume still showed its August 19 issuance, so the standing model picture is the one described in our last report: 25 of 26 models forecasting a very strong event for October–December, 15 of them at or beyond +3.0°C — past the top of the IRI's own intensity scale, which stops at +2.0°C — and strong-to-moderate conditions persisting through March–May 2027 with little sign of a transition away from El Niño before spring (IRI). The World Meteorological Organization's late-August update put the likelihood of El Niño persisting through February 2027 at near 100% (WMO). Nothing in the September cycle walked any of this back.

The Official California Signal, and Where Lodi Sits in It

NOAA's Own Map Does Not Reach Lodi Until November

NOAA issued its updated seasonal outlook on September 17, and the fine print is the part that matters here. For October–December, above-normal precipitation probabilities are enhanced from the southern half of California east-northeastward across the Four Corners and Great Plains — a footprint that does not include the northern San Joaquin Valley. From November–January through February–April, the wet signal extends across California, reaching a maximum of 50–60% odds of above-normal precipitation near the coast and extending as far north as central Oregon. NOAA added that the California probabilities may be increased in the next seasonal outlook once the latest model runs are reassessed (NOAA CPC seasonal outlook discussion, September 17, 2026). The next issuance is October 15.

For readers in Lodi, that is a useful calibration rather than a letdown. The official federal forecast currently says the strongest wet tilt arrives here after the calendar turns toward the new year, not in October. It also says, in NOAA's own words, that the number is more likely to go up than down next month. The April report described the northern San Joaquin Valley as sitting near the hinge of El Niño's rainfall geography. Five months of forecast revisions later, the hinge is still visible — it has simply moved from a question of whether Lodi gets wet to a question of when in the winter it does.

The European Ensemble Is Further Out Front

The ECMWF seasonal ensemble updated for September and, according to UCLA and National Center for Atmospheric Research climate scientist Daniel Swain, continues to show a very strong tilt toward wet conditions across the entire U.S. southern tier, including all of California. The figure he highlighted: greater than 70% odds of a winter substantially wetter than average, meaning the 80th percentile or wetter. He called it the strongest seasonal-scale predictive signal he has seen in California (Daniel Swain, September 6, 2026).

The gap between a European model at greater than 70% for a top-quintile winter and a NOAA map at 50–60% near the coast is not a contradiction. They measure different things — one the odds of an unusually wet winter, the other the odds of any above-normal total — and NOAA's product is deliberately conservative and revised monthly. The practical reading is that the two are converging, and NOAA has told readers which direction its next revision is likely to move.

The Mokelumne Clock: A Published Deadline, Fifteen Days Past Its Start

The most actionable local development this month has nothing to do with sea surface temperatures. It is a flood-space schedule that began running on September 15 and expires on November 5.

Under the March 19, 1962 agreement between EBMUD and the U.S. Army Corps of Engineers, EBMUD must begin creating flood-control reservation — deliberately empty storage — in Pardee and Camanche Reservoirs in mid-September each year. The mandatory Rain Flood Reservation starts at zero on September 15, increases uniformly to a maximum of 200,000 acre-feet by November 5, and is held through March 15 before declining to zero by May 31. The requirement can be met in any proportion between the two reservoirs, and a portion is transferable upstream to PG&E's Salt Springs and Lower Bear reservoirs by formula, which is why EBMUD describes the November 5 obligation as a range: between 130,000 and 200,000 acre-feet, depending on how full the upstream reservoirs are at the time (EBMUD, "EBMUD's Mokelumne River Project in Perspective").

What "flood reservation" means in plain terms

A flood reservation is empty space held on purpose. The reservoir gives up water it could have kept in order to have room to catch a storm and release it slowly, instead of passing a surge straight down the river. On the Mokelumne, that empty space is the primary flood protection for Lodi, Woodbridge, and Thornton, because no levee-reconstruction capital project comparable to Stockton's is underway on this corridor. The trade is explicit: water released in October is water the system no longer has, purchased as insurance against January.

The Arithmetic as of This Week

State readings show Camanche holding 352,780 acre-feet on September 17 and Pardee holding 190,288 acre-feet on September 20 (California Data Exchange Center daily storage for CMN and PAR). Against maximum capacities of 417,120 acre-feet at Camanche and 197,950 at Pardee, the system currently carries roughly 72,000 acre-feet of empty space — against a November 5 requirement of 130,000 to 200,000.

Flood Space Available Now Versus Required November 5

FIGURE 2. Empty storage in the Pardee–Camanche system compared with the November 5 flood-reservation requirement. "Available now" is combined capacity (615,070 acre-feet) minus latest reported storage (543,068 acre-feet). The requirement is a range because a portion of the reservation is transferable to PG&E's upstream reservoirs by formula. Sources: California Data Exchange Center; EBMUD. LodiEye calculation.

That leaves a gap of roughly 58,000 to 128,000 acre-feet to shed in 46 days, or about 1,260 to 2,780 acre-feet per day. Camanche's outflow in mid-September ran between 295 and 315 cubic feet per second, and net storage was declining by only about 230 acre-feet per day. At one cubic foot per second equal to roughly 1.98 acre-feet per day, closing the gap on schedule implies releases rising to somewhere in the range of 800 to 1,600 cubic feet per second and holding there through late October — three to five times the current rate.

Camanche Storage Through 2026, Against the November 5 Target Band

FIGURE 3. Camanche Reservoir monthly storage for 2026 with the September 17 daily reading, shown against the Camanche storage levels that would leave 130,000 and 200,000 acre-feet of combined system space given Pardee's September 20 reading of 190,288 acre-feet, which leaves Pardee itself nearly full and only about 7,700 acre-feet of space. Because the reservation may be split between the two reservoirs and partly credited upstream, these bands are indicative of the system obligation rather than a Camanche-specific limit. Sources: California Data Exchange Center monthly storage; EBMUD. LodiEye calculation.

Three caveats belong with those figures. The reservation can be satisfied at either reservoir, so Camanche's own curve is only part of the picture. Upstream credit at Salt Springs and Lower Bear can pull the obligation toward the 130,000 end. And EBMUD operates to this schedule every year and knows it better than any outside observer — the drawdown may already be planned for October rather than September. What the numbers establish is not a failure but a magnitude: the Mokelumne has to move a lot of water past Lodi in the next six weeks, and the amount is published, dated, and checkable.

The agreement also defines what happens if storms fill that space anyway. When the reservation is encroached upon, releases from Camanche are to be limited to 5,000 cubic feet per second insofar as possible, and the rate of change is capped at 1,000 cubic feet per second per two-hour interval. Those two numbers are the practical ceiling and ramp rate on the Mokelumne during a flood fight, and they are worth knowing before a storm rather than during one.

What This Means Downstream

Higher October releases are not a hazard in themselves — they are the system working as designed. But they change what residents along the river will see. Expect noticeably higher flows through Woodbridge and past Lodi Lake during a month when the channel is usually quiet, and expect Woodbridge Irrigation District operations to interact with them: WID's principal diversion right of 300 cubic feet per second runs through October 31, after which its entitlement drops to 114.4 cubic feet per second for the November–January period (Woodbridge Irrigation District). The practical consequence is that the same reservoir releases will translate into more water in the lower river after October 31 than before it.

Everything Else That Moved This Month

Tenmile Slough Is on Schedule

The Tenmile Slough levee reach behind Brookside in west Stockton — the first constructed segment of the Lower San Joaquin River Project and the reach San Joaquin County has described as the most critically deficient among those slated for improvement in the Central Valley system — is targeting completion by the end of October. San Joaquin Area Flood Control Agency executive director Darren Suen's assessment: "It's looking good." The work, underway since April 2025, rebuilds roughly a mile of levee with an impermeable slurry wall and rock-armored slope. It is the leading edge of a program to rebuild 24 miles of Stockton levees protecting 71,000 acres, 264,000 residents, and about $21 billion in property, at an estimated $2 billion and a completion horizon of 2039 depending on funding (Stocktonia). For this winter, one reach finishes weeks before the wet signal arrives and the remaining 23 miles do not.

The King Tide Dates Are Now Published

Our August report said the winter's highest-tide calendar would publish this fall. It has. The California Coastal Commission lists statewide king tides for November 24–26, 2026, December 23–25, 2026, and January 21–22, 2027, noting that Northern California locations may also include December 22 and January 20 or 23 (California King Tides Project). These are the dates to cross-reference against storm forecasts. Tidal influence reaches well inland into the Delta, and elevated base water levels during a storm are part of the flood-risk arithmetic rather than a coastal curiosity. All three windows fall inside the period Figure 1 shows as near certain for a very strong event.

Winter 2026–27 dates to hold in one place
Date What happens Why it matters locally
October 8, 2026 Next NOAA CPC ENSO diagnostic discussion Whether the 75% historic probability holds, climbs, or is walked back as the peak window opens
October 15, 2026 Next NOAA CPC seasonal outlook NOAA has said California above-normal precipitation probabilities may be raised in this issuance
End of October 2026 Target completion, Tenmile Slough levee (TS30L) The one regional levee reach that finishes before the storm season
October 31, 2026 WID principal diversion right steps down from 300 to 114.4 cfs Same reservoir releases yield more water in the lower Mokelumne after this date
November 5, 2026 Full Mokelumne flood reservation due (130,000–200,000 acre-feet) The system's flood buffer for Lodi, Woodbridge, and Thornton
November 24–26, 2026 First statewide king tide window Elevated Delta base water levels; storm overlap is the compounding risk
November 30, 2026 Eastern Pacific hurricane season ends Until then, any system tracking toward the Southwest is a potential warm-rain event for the Sierra watershed
December 23–25, 2026 and January 21–22, 2027 Remaining statewide king tide windows Both fall inside the very strong probability peak shown in Figure 1
March 15, 2027 Mokelumne rain flood reservation may begin declining End of the mandatory flood-space period, subject to snowmelt requirements

The State's Preparations Are Split Between Concrete and Paperwork

Southern California's agencies are spending the pre-season on physical capacity. Los Angeles County is removing as much as 1.3 million cubic yards of sediment from mountainside reservoirs and debris basins this year to restore flood-catching room, work its assistant deputy public works director describes as critical ahead of the season (Los Angeles Times).

In Sacramento, the fight is procedural, and it is the one with the most direct bearing on San Joaquin County's long-term water position. The state wants to route storm flows into depleted aquifers rather than watch them run to the ocean, and permitting is the bottleneck. Groundwater supplies roughly 40% of California's water in normal years and closer to 60% in dry ones. After the 2023 executive order waiving some environmental rules, water managers diverted more than 400,000 acre-feet underground — and the state's groundwater still sits about 40 million acre-feet below its level two decades ago. Assembly Bill 2026, which would have extended streamlined recharge permitting to temporary floodwater diversions, passed the Assembly unanimously and was held in Senate Appropriations. A coalition of local officials from Orange to Tulare County has asked the governor for an urgent executive order (Politico California Currents).

For a county whose aquifer is in documented overdraft, that is the difference between a wet winter that is only a flood problem and one that is also a recharge opportunity. Karla Nemeth, the former Department of Water Resources director now leading the Association of California Water Agencies, offered the state's summary: California is "as ready as we can be," with more work to do to save the water for a drought. She also noted that DWR's next Central Valley Flood Protection Plan, due in 2027, will include a vulnerability index for the first time, on the premise that lower-income residents carry disproportionate flood risk (Politico California Currents).

The Calibrated-Caution Section, One Month On

The April report's core warning has not expired: a strength label describes the ocean, not California's rainfall, and 2015–16 remains the case where historic strength produced an ordinary northern-California winter. Three points still deserve weight against the forecast.

First, the geography in NOAA's own current outlook does not put the northern San Joaquin Valley in the enhanced-precipitation footprint until November–January. Second, most of California's costliest flood years since 1950 did not occur during El Niño events; flood disasters require storm timing and intensity, not a seasonal tilt. Third, the historical impact maps were calibrated against a cooler planet, and an event outside the observed envelope carries real uncertainty in both directions — the same point Swain has made repeatedly.

What September changed is the balance of what is knowable. The atmospheric forecast remains probabilistic. The Mokelumne flood reservation is not. One is a 75% chance; the other is a dated obligation with a published number attached, and it can be checked against state gauge readings any day of the week. When a forecast is uncertain, the most useful local reporting moves toward the parts of the system that are not.

The Lodi Focus

1. The Mokelumne System

The action item from this month's reporting is a single habit: read the storage numbers. Pardee and Camanche daily storage is published by the state and free to check. Between now and November 5, the question is whether combined empty space climbs from roughly 72,000 acre-feet toward the 130,000–200,000 acre-foot requirement, and how fast. If it does, the Mokelumne enters winter with its designed buffer. If it lags, the system carries less absorption room into exactly the months Figure 1 identifies as near certain for a very strong event. Either way, expect higher October flows past Lodi Lake than the month usually brings.

2. Vineyards

Harvest is underway under the conditions flagged in our last two reports: warm nights, elevated humidity, and an active Eastern Pacific tropical season that runs through November 30. Late-season rain during picking remains the classic risk to old-vine Zinfandel quality, and sorting capacity is the variable that decides how much of that risk turns into loss. Looking past this crush, the ECMWF winter signal points toward a 2023-style spring as the planning template for 2027: deep soil moisture, mildew pressure, and aggressive canopy work. Decisions about drainage and spring canopy plans are cheaper to make in October than in February.

3. Urban Flood Exposure and Insurance

Flood insurance carries a 30-day waiting period. A policy purchased in late September is active before the November king tides. One purchased when the first atmospheric river appears on the forecast map is not. With NOAA now at 75% on a record event and 98% on a very strong October–December, this is the decision that most rewards acting early and cannot be recovered later.

Stockton's structural exposure is unchanged: a single 100-year levee system protecting more than 129,000 residents and roughly $16 billion in property, with Delta islands sitting 10 to 25 feet below sea level behind aging levees, and 23 of 24 miles of planned improvement still ahead. What September adds is precision about timing. The king tide dates are known. The reservation deadline is known. Residents in lower-lying neighborhoods near the Mokelumne can put evacuation-route familiarity and an insurance review on the calendar this month rather than treating them as winter tasks.

Practical Steps for Lodi Residents

  • Check whether your property sits in a FEMA Special Flood Hazard Area at msc.fema.gov.
  • Decide on flood insurance now — the 30-day waiting period means a September policy covers the November tides and a November policy does not.
  • Know your evacuation route: San Joaquin County maintains maps through its flood protection information page.
  • Sign up for San Joaquin County Office of Emergency Services alerts; the county proclaimed September as National Preparedness Month and publishes advisories through its OES press-release index.
  • Watch the Mokelumne yourself: daily Pardee and Camanche storage and the Corps' Camanche release-change report are both public.
  • Mark the king tide windows — November 24–26, December 23–25, January 21–22 — and check the storm forecast against them.
  • Growers: confirm sorting capacity and harvest labor contingencies, and revisit drainage and spring canopy plans before winter rather than during it.
  • Clear gutters, downspouts, and yard drains before the first storms, not after.

LodiEye will continue tracking ENSO forecast updates, the Pardee–Camanche reservation schedule, and regional flood preparedness through the 2026–27 water year. Readers with local observations to share can reach us at editor@lodi411.com.

LodiEye is the original civic-reporting and analysis arm of Lodi411.com, a citizen-run civic data and transparency platform serving Lodi, California and San Joaquin County. LodiEye gathers information of public interest, applies editorial judgment to public records, meetings, and data, and publishes original explanatory reporting for its readers — the work of a newsroom, and a representative of the news media as that term is defined under federal law. Our reporting emphasizes primary sources, public data, and full source transparency so readers can check every claim. LodiEye complements, and does not replace, the other outlets covering this region; for additional reporting on Lodi, San Joaquin County, and the broader region, we also encourage readers to consult the Lodi News-Sentinel, Stocktonia, The Sacramento Bee, CalMatters, and other established news organizations. Our full editorial standards and news-media-status statement is published at lodi411.com/editorial-standards.

This LodiEye civic climate brief was produced using artificial intelligence tools under the direction and review of the founder. Lodi411 uses multiple AI platforms in its research and publication workflow, including Anthropic's Claude (primarily Opus and Sonnet models) and Perplexity AI across a variety of large language models offered by each. These tools were used in the following capacities:

Source Discovery: AI-assisted search and retrieval identified the primary and secondary sources for this update, including NOAA Climate Prediction Center's September 10 ENSO Diagnostic Discussion, the CPC official RONI strength-probability table issued September 2026, the CPC seasonal outlook discussion of September 17, the standing August 19 IRI/Columbia plume, Daniel Swain's September commentary on the ECMWF ensemble, EBMUD's published description of the 1962 Corps flood-control agreement, California Data Exchange Center gauge records for Pardee and Camanche, the California King Tides Project calendar, and reporting from Stocktonia, the Los Angeles Times, Politico, CNN, and WMO.

Credibility Validation: The strength forecast was cross-referenced across NOAA CPC, IRI/Columbia, WMO, and ECMWF via published expert analysis, prioritizing government forecast products, institutional analysis, expert commentary, and news reporting in that order. The September escalation was verified against CPC's own discussion and probability table, distinguishing the greater-than-90% fall-winter very-strong statement from the 98% October–December figure and the separately defined 75% historic probability at RONI +2.5°C. Reservoir storage figures were taken from state gauge records rather than third-party aggregators after two aggregator values were found to disagree with the official readings.

Analysis and Synthesis: The flood-reservation gap analysis in Figures 2 and 3 is a LodiEye calculation combining EBMUD's published reservation schedule with California Data Exchange Center storage and outflow records, with the method and its limitations stated inline. AI assisted in carrying the April report's three-system Lodi risk framework — the Mokelumne water supply chain, the vineyard economy, and urban flood exposure — forward against the September forecast, and in integrating the published king tide calendar and the state recharge-permitting question into the local picture.

Presentation: AI assisted in drafting, structuring, and formatting the report, including the three inline chart visualizations and the consolidated winter date table.

Final Review: Multiple AI models reviewed the completed draft for factual consistency, source attribution accuracy, arithmetic, and balanced presentation. Throughout the process, the editor sets the report's goals, scope, and tone; creates and shapes draft content; reviews and edits the report; integrates independent fact checks; and reviews the AI cross-checks and validations. Multi-tool cross-checking across independent models and sources is the primary error-reduction mechanism.

Lodi411/LodiEye believes that transparency about how our research is produced — including our use of AI under human direction — strengthens trust with readers and the broader information ecosystem. Readers who spot an error are encouraged to write editor@lodi411.com so we can correct it.

Sources

Contact: editor@lodi411.com

Don Bradford
Next

Lodi Site Plan and Architectural Review Committee - September 23, 2026