A Warning in the Pacific Is Not a Forecast for Your Farm

NOAA’s March El Niño watch was a probabilistic climate signal; the useful response was preparedness, not a claim about any one harvest.

Retrospective written 16 October 2026; focus month: March 2026.

A climate forecast is not a rain forecast for a particular village. In March, NOAA’s Climate Prediction Center moved from a La Niña advisory to an El Niño watch. Its 12 March discussion expected a transition toward neutral conditions, with a 62% chance that El Niño would emerge in June–August. The probability was a statement about a large-scale climate pattern developing, not a promise about rainfall in a named Kenyan county.

That distinction is easy to miss because climate signals are often compressed into a single phrase: “El Niño is coming.” The actual notice was more careful. NOAA described the observed ocean-atmosphere state, the expected transition, and odds for a future period. Probability leaves room for the event not to occur. Even if it does, regional rainfall effects are shaped by season, geography and other climate drivers.

Kenya had a more local forecast available as the long-rains season approached. The Kenya Meteorological Department’s February outlook expected near- to above-average rainfall in several western and highland zones, while forecasting near- to below-average conditions in parts of the southeast, northeast and coast. It warned of uneven distribution, dry spells and occasional heavy downpours. This regional outlook was a more actionable guide for Kenyan planning than the Pacific ENSO probability alone.

What ENSO can and cannot tell us

El Niño and La Niña describe recurring changes in tropical Pacific ocean temperatures and atmospheric circulation. These patterns can tilt the odds of weather outcomes in different regions. They do not determine each storm, dry spell or harvest. For a farmer, useful guidance must translate the broad signal into a local outlook at the relevant planting horizon, then into decisions about seed, water, labor and market risk.

The World Meteorological Organization’s March update similarly reported a transition toward neutral conditions and rising El Niño chances. The agreement between the WMO and NOAA adds confidence that the climate pattern deserved monitoring. It still does not provide a county-level rainfall total. For that, users need national meteorological services, seasonal forecasts, local observations and frequent updates.

This is the difference between a hazard signal and an impact forecast. The physical signal may be reasonably clear while its local consequences remain uncertain. Whether a farm loses yield depends on crop type, soil, planting date, irrigation, seed variety, pests, access to labor, and the timing and distribution of rain. Whether a household experiences food insecurity depends on income, storage, market access, prices and support systems as well as harvest size. The WMO’s ENSO outlook advises combining ENSO information with regional and local factors; NOAA’s explainer on seasonal climate outlooks explains why these are probabilities, not promises.

Why probabilities should change decisions

Suppose a farmer has two possible seed choices. One matures quickly but yields less in a favorable season; another offers higher yield when conditions are good but needs more water. A probabilistic climate outlook can help compare the choices, but only if the farmer also knows the crop’s response, local water availability and the cost of failure. A forecast is not a decision by itself. It updates the odds attached to a decision.

An illustrative example makes the logic clear. Assume, purely for teaching, that one seed choice produces KSh 20,000 of net value in a wet season and KSh 8,000 in a dry season. If a farmer assigns a 60% chance to wet and 40% to dry, the expected net value is (0.60 × 20,000) + (0.40 × 8,000) = KSh 15,200. If new, credible local information changed the assumed wet-season probability to 40%, expected value would be KSh 12,800. These are invented inputs, not observed Kenyan farm returns or NOAA-derived probabilities. Real decisions require real local costs and crop data.

Preparedness can be valuable even when forecasts are imperfect. Water harvesting, storage, crop diversification and timely planting information can reduce exposure across several plausible conditions. The right investment often has benefits under both drought and heavy rainfall. It is more robust than betting the season on a single prediction.

The countercase: false alarms have costs too

Every preparedness measure has costs. A farmer may spend scarce money on irrigation equipment that is not used; a county may redirect resources from other needs; a trader may overstock an input. Overreacting to a probabilistic signal can be harmful, especially when the event has not yet formed. The balance depends on the cost of preparedness, the cost of being unprepared and how early action must occur.

That is why NOAA’s 62% figure should not be turned into “rainfall is 62% more likely” or “the harvest will fail.” It was an outlook for ENSO emergence. The local rainfall response needs a separate forecast. Nor should a global pattern be used as an excuse to explain every later weather event. Attribution requires evidence specific to the event and region.

A practical East African lens

For East Africa, the useful response to March’s watch was to build an information chain: national meteorological updates, county extension services, water and reservoir monitoring, crop calendars, and communication that reaches farmers in forms they can use. Health systems also have a reason to follow seasonal conditions. Water and food stress can affect nutrition and disease risk, while floods can disrupt roads, clinics and cold chains. Those are plausible pathways to monitor, not quantified consequences established by the March watch itself.

Digital agriculture and health tools can help translate new information into practical choices, but only when they use credible data and show uncertainty plainly. An app that presents a broad probabilistic signal as certainty can cause users to make brittle plans. Better tools explain the forecast horizon, local limitations and what new information would change the recommendation.

The forecast register published on 16 October has an analogous discipline: an explicit probability must later be judged against an explicit outcome. Its Brier score penalizes overconfidence more than a cautious forecast when events fail to occur. Climate forecasters face the same public challenge: retain the uncertainty and evaluate the product over many forecasts, not one memorable season.

From October, there is no claim here about what the March El Niño watch ultimately produced in any Kenyan locality. The practical outlook is conditional: if regional seasonal guidance confirms elevated rainfall risk, counties can prioritize drainage and access to care; if dry conditions become more likely, water planning and drought-tolerant inputs deserve attention. Either way, routine monitoring has value. March’s signal mattered because it gave institutions a lead time to prepare—not because it told every farmer what would happen.

Part 03b of The World Ledger: 2026, Read Backwards.

Intelligence has to survive the harvest

In the founder’s Kenya Vision 2060, agricultural intelligence belongs alongside storage, irrigation, genetics, logistics and settlement. A better weather signal has little economic value if a farmer cannot finance a response or move the harvest. Over the next season, compare actual spoilage, water use, transport reliability and net farm income against an explicit alternative. A forecast that earns attention but changes none of those outcomes has not yet become agricultural productivity.

Read the World Ledger

All 30 articles. Previous: A Strait Priced in Fuel. Next: The Price Notice and the Price of Time.

Mathematical methods: 45 parameters, assumptions and limits. Download the research and executable model.

AI-assisted research and writing using parallel Luna agents; editorial synthesis under the APEX lens. Historical scenes are illustrative. Sources support the stated observations; forward scenarios and inherited probabilities remain conditional, uncalibrated judgments. Kenya Vision 2060 is a founder-authored thesis, not an adopted government plan.

Cover image: Sea surface temperature anomalies from March to August 1997 showing evolution of major El Nino (2268-446).jpg — NOAA · Public domain. Shows the evolution from March through August 1997; this is a historical climate visualization, not a map of 2026 conditions. Resized for web display.