El Niño and its potential impacts on agrifood systems in Latin America and the Caribbean

Forecasts for the 2026–27 El Niño point to a high probability of strong warming in the tropical Pacific, with potential implications for harvests, food prices and rural incomes in Latin America and the Caribbean (LAC). The forecast from the National Oceanic and Atmospheric Administration (NOAA) cited below assigns an 83% probability to the Relative Oceanic Niño Index (RONI)^[1] reaching +2.5 °C or higher in October–December 2026. This is a probability concerning ocean conditions, rather than a prediction of specific agricultural losses. A historical stress scenario built from past El Niño events, which combines for each product the strongest average deviation from trend observed, shows agrifood productivity around 4% to 6.5% below trend in several countries in Central America and the Caribbean, alongside gains in parts of the Southern Cone. These patterns provide a basis for assessing risks and considering anticipatory measures, although outcomes will vary across territories and production systems.

Understanding ENSO: El Niño and La Niña

The El Niño–Southern Oscillation (ENSO) is a natural climate cycle in the tropical Pacific Ocean, driven by the interaction between sea surface temperatures and the atmosphere. It has three phases. El Niño is the warm phase, when waters in the central and eastern equatorial Pacific become unusually warm and the trade winds weaken. La Niña is the cold phase, when those waters cool and the trade winds strengthen. In between lie neutral periods, when neither pattern dominates. Because the Pacific is so large, these shifts move the tropical rain belts and alter weather patterns around the world.

ENSO is irregular but recurrent. An El Niño or La Niña episode typically develops every two to seven years. Events usually begin to form around mid-year, peak between November and January, and last nine to twelve months, although some La Niña episodes persist for two or three years in a row. For this reason, ENSO years are usually counted from mid-year to mid-year, as in 1997–98.

Not all events are alike. In the historical analysis, their intensity is measured using NOAA’s Oceanic Niño Index (ONI), the three-month average sea surface temperature anomaly in the Niño 3.4 region of the central Pacific. In the historical classification used here, prepared by Golden Gate Weather Services on the basis of the ONI, an episode is counted when the index stays at or above +0.5 °C (El Niño) or at or below −0.5 °C (La Niña) for at least five consecutive overlapping three-month periods. Each episode is then classified by its peak as weak (0.5 to 0.9 °C), moderate (1.0 to 1.4 °C), strong (1.5 to 1.9 °C) or very strong (2.0 °C or more), provided that the threshold is sustained for at least three consecutive periods. Under this classification, 28 of the ENSO years between 1950–51 and 2025–26 were El Niño years, 25 were La Niña years and the remaining 23 were neutral. Stronger events tend to bring larger and more widespread impacts, although no two events are identical.

In LAC, El Niño has a well-known fingerprint. Central America and the Caribbean tend to receive less rainfall, with recurrent droughts in the Dry Corridor that stretches from Guatemala to Panama. Northern South America and the Amazon basin also tend to dry out, which lowers river levels and hydropower generation. By contrast, the coasts of Ecuador and northern Peru often face torrential rains and floods, while warmer seas disrupt fisheries. Further south, Argentina, Uruguay and southern Brazil usually receive more rainfall, which generally favors summer crops. The same phenomenon, therefore, can bring drought, floods and bumper harvests to different parts of the region at the same time.

Historical patterns and the intensity of El Niño

Of the 28 El Niño years recorded since 1950, nine reached strong or very strong intensity (Figure 1). The ONI value of +2.2 °C for July-September 2026 (the three-month average centered on August) reported in the data used here would place the current event in that group, above the peak of the 2023–24 episode. Comparisons with earlier events help characterize its potential intensity, but do not establish that its agricultural impacts will follow the same pattern.

Figure 1.

Forecast probabilities and the agricultural calendar

The forecasts cited here indicate a high probability that El Niño will remain strong to very strong through late 2026 and into the first quarter of 2027. In its October 2026 diagnostic discussion, NOAA’s Climate Prediction Center estimated an 83% probability that RONI would reach +2.5 °C or higher in October–December (Figure 2). This probability refers to a specific ocean temperature threshold and period; it does not imply the same probability of drought, flooding or productivity losses in any particular location.

Figure 2.

According to the October 19, 2026, forecast of the International Research Institute for Climate and Society (IRI), all 27 models considered (16 dynamical and 11 statistical) project a very strong El Niño, with sea surface temperature anomalies of at least +2.0 °C in the Niño 3.4 region from October–December 2026 through December 2026–February 2027, and a multimodel average of around +3.5°C in October-December and November-January. The forecast probability of El Niño conditions remains at 100% through March-May 2027 and then declines to 92% in April–June 2027, 60% in May-July 2027 and 37% in June-August 2027. These projections describe a range of possible developments and should be updated as new observations and forecasts become available.

For agriculture in LAC, the timing of a possible peak matters alongside its magnitude. The 2026–27 summer cropping season in the Southern Hemisphere, the postrera maize and bean season harvested toward the end of 2026, and the apante bean season that extends into early 2027 in parts of Central America could all overlap with strong El Niño conditions. The implications will depend on local rainfall and temperatures, crop calendars and production conditions. Seasonal forecasts can help guide preparation, provided they are combined with local monitoring.

How could agrifood productivity be affected?

To gauge how strongly El Niño could affect the region’s agriculture, we classified every year between 1961 and 2024 according to its ONI intensity. Using FAOSTAT data, we then measured how much the productivity of each product group in each country deviated from its long-term trend in El Niño years, compared with neutral years. Crop productivity was measured as yield per hectare and livestock productivity as output per animal.

The long-term trend was estimated with a Hodrick-Prescott filter applied to the logarithm of each productivity series, which separates gradual changes driven by technology and investment from year-to-year fluctuations. The annual deviation is the percentage difference between observed productivity and its trend. For each El Niño intensity category (weak, moderate, strong and very strong), we averaged the deviations of the years in that category and subtracted the average deviation of neutral years, so that the effect is measured against years without either El Niño or La Niña. Because harvests in the Southern Hemisphere occur after the event's peak, summer crops and livestock in those countries were matched with the following year's harvest.

We then identified the dominant direction of the effect for each product and country in three steps. First, we took the sign that prevailed among the strong and very strong categories. If they disagreed, we took the sign that prevailed across all four intensity categories. If a tie remained, we used the sign of the largest absolute effect between the strong and very strong categories. The stress scenario was then defined as the largest average effect in that direction across all categories. This scenario is therefore not the outcome of a single historical episode, but a stress scenario that combines, product by product, the most adverse average response observed (or the most favorable, where El Niño has tended to raise productivity). Results were aggregated using the value of production of each product as weights.

Aggregate patterns associated with El Niño can be more predictable than impacts in specific territories. As the analysis moves to smaller geographic scales, differences in rainfall, soils, elevation, water access, crop calendars and management practices make outcomes more heterogeneous and harder to anticipate. National averages may therefore conceal losses and gains within the same country. The results shown here should be interpreted with caution: they describe historical associations under a severe-event scenario, rather than causal estimates, local forecasts or losses with an assigned probability. The scenario based on the largest observed deviations is not an estimate of the most likely outcome. Local assessments and updated forecasts are needed to guide decisions.

The historical scenario illustrates different patterns across LAC (Figure 3). In Central America and the Caribbean, estimated productivity deviations range from around 4% to 6.5% below trend in several countries: Costa Rica (−6.5%), El Salvador (−5.2%), the Dominican Republic (−4.6%), and Panama, Honduras and Nicaragua (around −4.2%); the estimated deviation for Guatemala is −3.7%. The scenario also shows negative deviations in Ecuador (−2.2%), Peru (−2.1%) and Colombia (−1.2%). Positive deviations appear in Argentina (+4.1%), Chile (+2.8%) and Brazil (+1.7%), while Paraguay (−1.8%) and Uruguay (−1.2%) show negative deviations. These values summarize the historical scenario used in the analysis and should not be read as forecasts for the coming harvests.

Figure 3.

At the regional level, the estimated average deviation in this scenario is slightly positive, reflecting the large combined share of Brazil and Argentina in LAC’s agricultural output. This aggregate can conceal substantial differences among countries, territories and producers. Negative deviations in countries where households depend on rainfed smallholder agriculture may have significant implications for food security, even when regional output appears relatively stable.

This scenario also differs by product within the same country (Figure 4). In Central America, estimated yield deviations range from 5% to 18% below trend for maize and from 4% to 11% below trend for rice, with potential implications for smallholders’ food supplies and incomes. In Argentina, the scenario shows positive deviations for maize (+5.9%), oilseeds (+6.3%) and wheat (+13%). In Brazil, the positive aggregate coexists with negative estimates for wheat (−23%) and the group comprising coffee, cocoa and other crops (−13%). These differences reinforce the need to interpret country averages with caution. Livestock productivity, measured as output per animal, shows smaller deviations than crop productivity, although this indicator does not capture all potential effects, including drought-related distress sales and herd losses.

Figure 4.

Anticipatory action and regional cooperation

The forecasts available for the 2026–27 El Niño provide an opportunity to assess risks and prepare before potential impacts materialize. They indicate probabilities for the event’s development and intensity, while its timing and local consequences remain uncertain. Preparation can therefore proceed through proportionate measures informed by those probabilities, with decisions revisited as forecasts and local information are updated. This requires linking climate information with local exposure, vulnerability and response capacity.

Experience with previous El Niño events suggests that forecasts do not always translate into timely preparation. Anticipatory action can help protect production and livelihoods, although its benefits depend on the intervention, timing and local context. The Food and Agriculture Organization of the United Nations (FAO) and the World Food Programme (WFP) estimate that every dollar invested in anticipatory action can help avert up to seven dollars in losses and response costs (FAO, 2026). These estimates illustrate potential benefits rather than returns that can be assumed in every setting. Financing arrangements such as index insurance, contingency funds and credit lines with pre-agreed triggers can support earlier action when linked to relevant indicators and clearly defined responsibilities.

Regional cooperation can complement national and territorial preparation, as El Niño can affect multiple agricultural producing regions simultaneously. Applying a consistent methodology across countries and regional groupings worldwide reveals agrifood productivity below trend not only in Central America, the Caribbean and the Andes, but also in Australia and New Zealand, much of sub-Saharan Africa, South Asia, parts of Southeast Asia, China, Russia and Kazakhstan. These estimates indicate the possibility of simultaneous pressures on supply, rather than forecasting their occurrence in 2026–27. Such pressures could affect international prices, while export restrictions could increase costs for importing countries. Potential gains in parts of the Southern Cone could help offset shortfalls elsewhere, depending on actual harvests, logistics and trade conditions. Coordinating anticipatory financing, sharing updated information and maintaining predictable access to markets for food and agricultural inputs can help manage these risks.

About the authors

Valeria Piñeiro

Director of Directorate of Technical Cooperation (DCT) at IICA

She holds a PhD in Agricultural Economics from the University of Maryland and is also a faculty member in the Advanced Academic Programs at Johns Hopkins University.

Juan Pablo Gianatiempo

Research Analyst with the Markets, Trade, and Institutions Unit at IFPRI

He holds a Master's degree in Data Science from the Buenos Aires Institute of Technology (ITBA). He previously worked as a Senior Data Analyst at Kynetec and at the Buenos Aires Grain Exchange.