El Niño Hazard Board: North & Central America
What to Know and How to Prepare (2026-2027)
A record-strength El Niño does a different thing in every region, and the impacts arrive staggered across two years — a wet winter that suppresses fire in one place while growing next summer's fuel, a dry winter that empties the snowpack in another.
This board maps which hazards are likely to intensify where, and when across the winter, and the 2027 fire and water year where most of the consequences will land.
Each cell carries a confidence mark, and each hazard column shows the kind of evidence behind it, so you can tell a well-forecast hazard from a plausible one.
Reviewed and updated weekly. Free to use. Not an agency product. Verify with sources before making operational decisions.
Reviewed daily · 12 September 2026
What a historic El Niño makes worse, and where
The 2026–27 event is on track to be the strongest in the instrumental record. This board maps the hazards that intensify by region across two horizons: the winter itself, and the lagged consequences that arrive in the 2027 fire and water year.
- Niño-3.4 anomaly
- +1.8 °CAugust monthly mean, CPC 10 Sep 2026
- Very strong, fall & winter
- > 90%CPC, 10 Sep 2026. NIFC cites >95% by November; the CPC figure is used here.
- Record-exceeding event
- 75%Oct–Dec 3-month RONI ≥ +2.5 °C, beyond anything since 1950. Raised from 69% on 10 Sep.
- Still El Niño Apr–Jun 2027
- 78%decay begins in spring
- What this is
- An editorial synthesis of published agency forecasts and peer-reviewed findings, assembled to support planning. Every cell is sourced, confidence-marked and evidence-tiered.
- What it is not
- Not an agency product and not an operational forecast. It carries no authority. For decisions, use the agency products it cites and the live monitoring linked below.
- Maintenance
- Reviewed daily against the source products. Corrections are logged in full, with dates, under Revision log at the foot of the page rather than quietly amended.
- Turning this into action
- Every region name in the matrix links to that region's hazards and the actions that serve them. Seven printable briefs sit behind that index, with named counties and agency-attributed steps: El Niño Action Briefs.
ENSO diagnostics current through the CPC discussion of 10 Sep 2026 · WMO ENSO Update, 3 Sep 2026 · NIFC outlook, 1 Sep 2026 — next NIFC outlook 1 Oct 2026, next CPC discussion 8 Oct 2026
Composite hazard load by region
Each tile carries the highest hazard level in that region for the selected horizon, and names the hazard driving it. Geography is schematic, not to scale — the tiles are arranged to echo the continent, not to draw it. Hawaii sits as an inset below, far off its true position.
Region × hazard matrix
Happening now, and what it does and doesn't confirm
The events of the first half of September sit on opposite sides of the attribution line. One is the board's cleanest confirmation. The other is the trap. Each entry carries the dates it refers to, so it can be read correctly after the event has passed.
Hurricane Lowell · on signal
This is the teleconnection working exactly as documented. Lowell passed northwest of Kauai as a weakening major hurricane — 85 mph gusts at Waimea, nearly 18 inches of rain in Kauai's mountains, 90% of the island without power, 30-foot surf, and Hawaii's first known tornado watch. NOAA had already called a 70% chance of an above-normal Central Pacific season, 5 to 13 storms against a normal 4 to 5, on the strength of El Niño alone.
Kiritimati · the ocean is already there
At the geographic center of the event, the reef is bleaching before the winter peak. Researchers working on Christmas Island in the Central Pacific in August reported sea surface temperatures of 30 to 30.5 °C, at least two degrees above average, with active coral bleaching and rainfall heavy enough to cut the island's dirt roads. The winter peak is still months out. This is the marine heat and ecosystems column arriving early at the one place the index is actually measured, and it is the strongest live confirmation on this page that the ocean is behaving as forecast.
The Central Pacific timing gap
The Central Pacific hurricane season peaks August through October, before the winter panel opens. The hazard El Niño raises most reliably in the Pacific therefore falls between the two horizons shown here. Track it directly through the Central Pacific Hurricane Center rather than reading it off the winter panel.
Southern California's early-September heat · off signal
88 to 108 °F under an NWS extreme heat warning across Los Angeles County from the morning of 9 September to the evening of 10 September 2026, and El Niño was not why. The warning has since expired. Early-September heat is climatological in Southern California. This board puts the region's winter warm anomaly at level 1 — the wet signal is the strong one there, not the hot one. Attributing an episode of this kind to El Niño is the exact error the confidence tiers exist to prevent.
Why it still matters operationally
Wrong cause, real consequence. That heat landed inside the window NIFC still flags for elevated fire potential, on fuels that have been curing all summer, weeks before the storm door opens. The sequence that matters is heat first, ignition risk alongside it, then a wet winter that grows next year's grass — and the same warmth is what holds sea surface temperatures up off the coast. The attribution is wrong; the operational exposure is not.
Sea level and ice: one fast hazard, one slow one
These two behave differently from everything else on the board. Sea level responds within weeks and gives the height back when the event decays. Ice responds a season late and never gives it back.
Coastal sea level · weeks
Weakened trade winds launch downwelling Kelvin waves that travel east and then poleward along the Americas, warming and expanding the coastal ocean. The measured 2015–16 anomaly was +11 cm averaged along the US West Coast, +17 cm in Oregon, and roughly 3–5 cm per 1 °C of equatorial warming. Figures approaching a foot are circulating for California; they exceed anything in the tide-gauge record by roughly a factor of three. Plan against the measured range.
The flood-day numbers exist. Use them.
NOAA's 2026–27 High Tide Flooding Outlook projects a national record of 7–12 flood days. Mid-Atlantic 15–20, Pacific Northwest 7–18, Northeast 5–11, Southern California 3–6, with Toke Point WA at 17–29 and Duck NC at 22–27. NOAA names El Niño as the driver and the Mid-Atlantic and Pacific coasts as most susceptible.
Global mean sea level · temporary
Expect a spike in the global record, then a partial retreat. The 2014–16 event added 9.16 mm and 2023–24 added 7.70 mm. Ocean mass — not thermal expansion — was 68% and 81% of it, driven mainly by terrestrial water storage depletion in the Americas and Africa. Read a 2026–27 jump as the event, not as an acceleration in the trend.
Glacier mass balance · a year late, permanent
The mechanism is accumulation, not melt. A warm dry winter starves the snowpack, the transient snowline runs high the following summer, bare ice and firn are exposed, and albedo feedback finishes the job. Snowline rise across western North America has already gone fourfold, from 2.1 to 8.9 m per year since 2010.
Where the ice signal concentrates
Alaska, the Coast Mountains, the Cascades and the Canadian Rockies sit under the warm-dry pole all winter, and they are already the world's fastest-losing region — 22.2 Gt a year across western Canada and the conterminous US from 2021 to 2024, double the prior decade, and 12% of remaining volume gone since 2020. Alaska glaciers lose up to three additional weeks of melt per 1 °C of summer warming. The ENSO correlation is positive but often not statistically significant; temperature and snowpack do most of the work.
Fire and ice are now coupled
Wildfire soot is a named driver of the recent acceleration in western North American glacier loss, alongside reduced winter snowfall and early heat waves. A season that burns hard darkens the ice that a snow-short winter already exposed. That is a feedback, not a coincidence, and it runs in the direction this event pushes both variables.
The trap for water managers
Ice melt masks the snow deficit, then stops. A long, hot melt season can hold late-summer streamflow up in glacier-fed basins even as snow water equivalent collapses, so gauges look tolerable in the year the storage is actually being spent. Plan on the snowpack number, not the hydrograph.
Heat and wind in 2027
These two columns are the least alike on the board. The heat signal is among the most confident things anyone can say about 2027. The wind signal is the least — and it is still the one that decides fire outcomes.
Heat runs a season behind the ocean
Global temperature peaks three to six months after the ENSO peak, so an October–December 2026 crest lands on 2027, not 2026. WMO puts an 86% chance that some year in 2026–2030 beats 2024, and names 2027 as the leading candidate on the strength of this event.
Where the heat concentrates
The warm-dry pole carries it. Environment and Climate Change Canada’s CanSIPSv3 forecast issued 21 August 2026 puts western Canada at at least a 90% probability of above-normal temperatures for March to May 2027, and the same signal runs south through the Northern Rockies and Great Basin, and separately across interior Mexico and Central America. Heat on a depleted snowpack is a different hazard from heat alone.
Wet winter, then flash drought
The Southwest's relief is not durable. Saturated soils moderate early-summer heat, then dry down fast once the profile is spent. A soaked winter followed by a record-warm summer is the recipe for rapid onset drought over a landscape carrying an unusually heavy grass crop.
Wind has almost no seasonal forecast
No skillful seasonal wind outlook exists at this range, so most wind cells here stay low confidence. The Southern Plains is a moderate-confidence call because its spring wind season is climatological — it arrives whether or not El Niño has an opinion. Southern California is moderate for a different reason, below.
Santa Ana
Three independent studies — an observational pressure-field climatology, a 6 km downscaled simulation and a 65-year downscaled reanalysis — all point the same way: El Niño winters run about 10% stronger in seasonal Santa Ana intensity, La Niña winters about 10% weaker, with the signal strongest when PDO is in phase. Modest, seasonal-aggregate, and silent about any individual event.
The decay scenario
Roughly 70% of strong events transition to La Niña within four to six months. That autumn would be dry with heavy fuels — but on the wind literature above, offshore wind intensity would run lower, not higher. The compound worth planning against is a cured landscape with reduced ignition-to-conflagration wind.
Where the signal is weak
Snowpack
The least reliable line on the board. Western sites are roughly equally likely to land above or below average snowfall in El Niño years, and the Central Sierra Snow Lab has come in above average in only half of them. Precipitation totals carry the signal; snow water equivalent does not.
The Great Basin hinge
The dividing line sits further south than the region's name suggests. Within Utah the practical boundary runs near 37–38°N, so the reliable El Niño wet signal is a southern Utah phenomenon. CPC's composite gives northern Utah near-normal precipitation at best in moderate-to-strong events, and Wasatch snowfall correlates weakly negative while southern Utah correlates strongly positive. For this event CPC calls northern Utah and north-central Colorado a toss-up on precipitation and warmer than normal — which is why this column stays low confidence.
Strength ≠ certainty
CPC says so explicitly: with an event of this magnitude the chances of El Niño-consistent impacts are larger, but they are not guaranteed, and a very strong event will not produce the expected effects everywhere.
Does magnitude even matter?
Three defensible positions, and they disagree. One line of modeling says only strong events shift California's odds at all. A second finds the southern dipole scaling continuously with amplitude. A third argues amplitude is the wrong variable entirely — that the pattern of warming decides the response, which is why 2015–16 matched 1997–98 on Niño-3.4 and delivered near-average California precipitation. That counterexample is the one every skeptic will raise, and they are entitled to.
The models are not independent
Every outlook on this board runs downstream of an overlapping model pool. NMME and the Copernicus multi-system share members; the Met Office system is itself inside Copernicus. Agreement between them is weaker evidence than it looks — it partly reflects shared initialization and the sheer size of the ENSO forcing. This is the argument for checking against observations rather than against another forecast. There is a second split worth knowing about, because it runs the other way. Statistical models learn from past events; dynamical models solve the physics from current conditions. With ocean heat outside the range of anything in the training record, the statistical approach has the weaker footing this year — and the dynamical models are the ones forecasting the more disruptive event. Where the two disagree, the historical-analog answer is the one carrying the larger unstated assumption.
Fuel is not fire weather
The 2027 fire column and the fine-fuel column now point in opposite directions across the southern tier, and that is deliberate. A 1984–2022 analysis of ENSO against area burned across the nine US coordination centers — Hoell et al., JGR Atmospheres, 2026 — finds autumn El Niño reduces the chance of extensive area burned in the Southern, Southwest and Rocky Mountain regions through the following spring and summer, while raising it in the Eastern and Northern Rockies regions. Fuel still accumulates. What the event suppresses is the weather that converts an ignition into a large fire — and the same work finds the suppression lifts in the subsequent summer and early autumn where an El Niño persists into spring. Read this as deferred, not removed. Two things to hold alongside it: this is a climatological base rate drawn from 39 years of observed events, not a forecast for this one, and no operational product yet covers spring or summer 2027 — NIFC's outlook horizon reaches March 2027 with its 1 December issuance and June 2027 with its 1 March issuance.
A warm winter is not a cold-free winter
The warm northern-tier signal is a seasonal mean, and planning against the mean is how a cold snap catches an agency short. El Niño winters carry an increased chance of the stratospheric polar vortex breaking down: planetary wave activity reaches the stratosphere more often, the late-winter vortex runs weaker, and major disruptions become more frequent. A breakdown can shift surface temperatures for days to weeks afterwards, pushing Arctic air far south of where a seasonal outlook would put it. Two things bound this. It is not forecastable at seasonal range — NOAA puts the limit at 10 to 15 days even for the stratospheric event itself, let alone its surface consequences. And the relationship is loose: there were no major sudden stratospheric warmings during 1997–98, the closest analog this event has. Warm on average, with cold outbreaks that arrive on two weeks' notice, is the shape to staff for.
Alaska is not an El Niño story
USGS states the linkage plainly: weak for most of Alaska. The odds of a warmer-than-normal winter shift by 5–10%, strongest in Southeast Alaska and unchanged in the northwest, with a small increase in the odds of storms in the south. Autumn sea ice decline is overwhelming whatever influence the event carries. Alaska's coastal erosion and storm exposure are among the most severe hazards on this continent — and this is the clearest case on the board of a place where a low signal level says nothing about how much is at stake.
Plains wind in 2027
No seasonal wind forecast exists, and no published ENSO signal in spring Plains wind speed was located. Spring wind on the Southern Plains is driven by lee cyclogenesis and the dryline, both robust climatological features of March through May in any phase. The fine-fuel pathway is real but conditional: it requires a wet winter followed by a spring drying, and the direct ENSO record argues against the second half. The operational answer is a measurement, not a forecast — assess the actual cured grass load in February 2027 and re-rate from it.
The Colorado River
The wet signal falls where the water is used, not where it is made. Roughly 90% of the river's flow originates in headwaters with limited ENSO connection, and the Upper Basin signal is not merely weak but sign-unstable — it flips between sub-basins and between indices. Reclamation's own August 24-month study applies no ENSO adjustment at all. Its 2027 projection is a statistical ensemble, not an El Niño forecast.
Great Salt Lake
One winter cannot move this lake. Its elevation lags watershed precipitation by roughly three years and tracks a decadal Pacific cycle rather than individual events. The 1982–83 analog is over-attributed: that El Niño produced the initial rise, but the lake climbed for three further years to its 1986 record through a wet cycle that was not a strong event, and CPC notes 1982–83 was atypical for northern Utah.
The spring predictability barrier
March through June 2027 are scenarios, not forecasts. Model runs at that range spread from −0.5 to +1.5 °C in Niño-3.4, and seasonal outlooks for key agricultural regions sit at roughly equal odds across all three terciles. Everything in the 2027 panel is a conditional read on how the decay actually goes.
The 2027 tropical season
Atlantic suppression is the safest call for 2026 and the least safe for 2027. El Niño still holds a 78% probability through Apr–Jun 2027, but a decaying event, or a swing toward La Niña, would remove the shear that is currently holding the basin down.
Evidence base, by hazard
Not every column on this board rests on the same kind of evidence. Tier A is an operational agency forecast for this event. Tier B is a peer-reviewed teleconnection with this event's conditions applied to it. Tier C is mechanism and inference, where this board extends published findings rather than citing a source that makes the claim directly.
| Hazard | Tier | What it rests on |
|---|---|---|
| Flood & debris flow | A | CPC seasonal outlooks and the USGS Southwest CASC multi-model synthesis, with the ENSO precipitation dipole (wet south of ~35°N, dry north of ~40°N) as the published mechanism. |
| Snowpack & water deficit | B | Strong agency agreement on precipitation, weak published agreement on snow water equivalent. The most-cited line and the least-supported one. |
| Cool-season fire | A | NIFC significant wildland fire potential outlook, 1 Sep 2026, and the trinational North American assessment co-authored with Natural Resources Canada and Mexico's SMN. |
| Severe storm & ice | B | Cold-season tornado climatology, which runs opposite to intuition — ENSO explains only about 10% of variance and the Arctic Oscillation more. The ice-storm signal for eastern Canada rests on the 1998 analog, not on a forecast. |
| Sea level & tidal flooding | A | The strongest-sourced column here. NOAA's 2026–27 Annual High Tide Flooding Outlook gives per-station flood-day projections, backed by tide-gauge measurement of the 2015–16 event. |
| Marine heat & ecosystems | C | Mechanism and prior-event analogy. No operational marine heatwave outlook for this event is cited here, and none should be inferred. |
| Warm anomaly · Extreme heat | A | CPC and ECCC seasonal temperature outlooks, plus the WMO decadal update for the global 2027 framing. The most confident row on the board. |
| Wildfire potential (2027) | B | No agency has issued a 2027 outlook yet, so this column rests on a peer-reviewed 1984–2022 analysis of ENSO against area burned by US coordination center — Hoell et al., JGR Atmospheres, 2026 — rather than on any forecast. The nearest current corroboration is NIFC's call of near-normal potential for southern California through December on explicit El Niño reasoning. The analysis's own stated limits: it measures area burned, not severity, ignition count or extreme single events, and excludes wind, terrain and vegetation type. |
| Fine-fuel loading | B | The best-documented lag here: one to three years in open ponderosa systems, within the following year in low-elevation grassland, and about eight months in the Southern Great Plains. CONAFOR makes the same argument for Mexico independently. |
| Drought & water shortfall | A | US Drought Monitor and CPC drought outlooks, the Canadian Drought Monitor, and the Central American climate forum with FEWS NET food-security classifications. |
| Glacier & icefield loss | B | Strong peer-reviewed measurement of the trend and its drivers. The ENSO-specific attribution is the weak part — the published correlation is positive but often not significant. |
| Wind potential | C | No seasonal wind forecast exists. Southern California is the exception, where three independent studies agree on a modest ENSO modulation of Santa Ana intensity. |
| Tropical cyclone | A | NOAA's Central Pacific outlook put this season at a 70% chance of above normal on El Niño grounds, and Hurricane Lowell delivered. Atlantic shear suppression is textbook and well supported for 2026. Only the 2027 season, which depends on decay timing, drops to inference. |
Check it against reality
Everything above is a forecast. These are observations, and they will tell you within weeks whether this board is holding. If the snowpack column is going to fail, SNOTEL will show it long before anyone reissues an outlook.
Snow
NRCS SNOTEL interactive map — real-time snow water equivalent as percent of the 1991–2020 median. California B-120 forecasts run February through June, so the first Sierra verification point is Feb 2027.
Drought
US Drought Monitor, released Thursdays. NOAA NIDIS for the portal view. Canadian Drought Monitor, monthly. CPC seasonal drought outlook for the forward view.
Coastal water levels
NOAA Coastal Inundation Dashboard and the monthly high tide flooding outlook. Station-level observed flood days against the projections above — the cleanest scoreboard on this page.
ENSO itself
CPC diagnostic discussion, second Thursday monthly — next 10 Sep 2026. WMO ENSO Update. IRI plume for the model spread.
Fire
NIFC outlooks, reissued the 1st of each month. CWFIS for Canada. Mexico's SNIF for CONAFOR fire statistics.
Tropical cyclones
Central Pacific Hurricane Center and the National Hurricane Center. The Central Pacific count against NOAA's 5-to-13 forecast is the fastest-settling verification on this board — the season closes 30 November.
Seasonal outlooks
CPC monthly and seasonal outlooks, reissued around mid-month. ECCC probabilistic seasonal forecast for Canada — maps only, no written narrative.
Updated 13 September 2026
What to do, by region and by hazard
Start with By region to see which hazards the board rates level 3 where you work, and the actions that serve them. The seven hazard briefs behind it carry the full sequence for each. Every action is attributed to the agency that recommends it — and where no agency has published guidance, this says so rather than filling the gap.
What this is.
An editorial synthesis of published agency forecasts and peer-reviewed findings, assembled by Wonder Labs to support planning across regions that are normally forecast separately. It draws on NOAA's Climate Prediction Center, the IRI plume, the WMO ENSO update, NIFC's fire potential outlooks, NOAA's high tide flooding outlook, the US and Canadian Drought Monitors, Environment and Climate Change Canada, CONAGUA and CONAFOR in Mexico, and CRRH-SICA with FEWS NET for Central America — alongside the peer-reviewed literature on ENSO teleconnections, glacier mass balance, cold-season severe weather and offshore wind.
What it is not.
Not an agency product, not an operational forecast. For operational decisions, use the agency products it cites and the live monitoring it links to. Every source is named on the page.
How to read the levels.
Levels describe the strength of the El Niño signal for that hazard, not absolute risk. A level 3 where nobody lives is not the same problem as a level 2 where they do. Alaska is the clearest case: its coastal erosion exposure is among the most severe on the continent and its El Niño signal is close to zero, and the board only measures the second. Dotted cells are low confidence. Blue cells mark hazards El Niño turns down — a quieter Atlantic hurricane season, drought relief in the Southwest — which is planning information too.
How to get from a rating to an action.
Every region name in the matrix links to that region's entry in the action briefs, which lists its level 3 hazards and the specific steps that serve them, each attributed to the agency that recommends it. Where no brief covers a region's hazard, the entry says so.
How to check it.
The board links the observational products that will confirm or contradict it as the season unfolds: SNOTEL for snow water equivalent, the Drought Monitor, NOAA's coastal inundation dashboard, the hurricane centers.
How to cite it.
Wonder Labs, El Niño Hazard Board: North & Central America 2026–27, accessed [date]. wonder-labs.org
Spotted an error, or know a better source?
This board is reviewed daily and every material change is logged with the date it was made. Corrections raised here go into that log.
Most useful: a regional detail that is wrong or out of date, an agency product this should be citing and isn't, a program deadline that has moved, or a hazard in your region the board doesn't cover. A link to the source helps.
If you would rather not post publicly, say so in your message and we’ll keep your comment confidential.