A strong El Niño is building for autumn and winter 2026-2027. Its effects on food security could become noticeable as early as late 2026 or during 2027 in several African regions, with a peak expected for southern Africa between November 2027 and March 2028. This delay between the climate shock and its effects on food prices makes reactive interventions ineffective at protecting rural livelihoods: late aid often arrives after households have sold part of their livestock or consumed survival resources. FEWS NET, the early warning network created by USAID, identifies a foreseeable window between November 2027 and March 2028, a window that justifies pre-financing today’s social safety nets, food stocks, and recovery seeds.
The essentials
- The NOAA Climate Prediction Center estimates a probability greater than 90 percent of a very strong El Niño in 2026-2027, according to FEWS NET. A food security peak is expected in southern Africa between November 2027 and March 2028, while other sub-regions of sub-Saharan Africa will experience different peak timelines.
- The structural delay between climate shock and food crisis often exceeds one season: interventions that wait for prices to spike arrive after the liquidation of rural assets.
- Pre-financing social safety nets and seed stocks before the peak, based on probabilistic forecasts, can protect household recovery capacity.
- Mechanisms for anticipated financing triggered by probability thresholds exist in prototype form; they have not yet been deployed at the scale needed.
- The window for action is open now: each month of delay reduces the margin to build up stocks and reach vulnerable households before the at-risk season.
El Niño builds long before hunger arrives
The El Niño phenomenon disrupts rainfall patterns across much of southern and eastern Africa. Rainy seasons shorten or disappear, harvests collapse, prices rise. But this sequence takes time. The effects of a poor rainy season can be transmitted through several channels: household stocks, income, production, markets, and prices, depending on local context. This delay, which can exceed twelve months, has been documented by the FEWS NET network for decades.
The October 2026 edition of FEWS NET reports a probability greater than 90 percent of a very strong El Niño for autumn and winter 2026-2027, according to the NOAA Climate Prediction Center. This probabilistic forecast makes it possible to identify a food security window at risk more than a year in advance. Between November 2027 and March 2028, the most exposed zones of southern Africa should experience an increase in humanitarian needs and Crisis-level outcomes, according to the IPC framework.
This temporal precision changes the nature of the problem. It transforms it from an emergency to be managed into a risk to be covered.
Household losses before aid arrives
When a food crisis breaks out, humanitarian agencies mobilize, media cover it, donors release funds. But at that stage, rural households have already gone through several steps of degradation that nobody documents in real time.
The first is meal reduction: families go from three meals to two, then to one. The second is the sale of non-productive assets, jewelry, utensils, furniture. Then comes the liquidation of productive assets: livestock sold in emergency at fire-sale prices, the consumption of seeds meant for the next season, the disposal of tools. A household that has liquidated its productive assets can survive the crisis, but it no longer has the means to restart agricultural activity when the rains return.
It is this mechanism—not visible starvation, but the silent destruction of recovery capacity—that reactive funding, often too slow, can scarcely prevent. Aid arrives when the crisis is declared. A famine declaration is made on the basis of an analysis bringing together thresholds of food deprivation or livelihood collapse, acute malnutrition and mortality, not just the breach of a threshold of severe malnutrition. The sale or depletion of productive assets can, however, precede a famine by several phases; famine classification also depends mandatorily on thresholds of consumption, malnutrition, and mortality.
The causal chain is known. It is documented. And it repeats.
Pre-financing before the spike: what the FEWS NET window makes possible
The operational value of a forecast more than a year out does not lie in its meteorological precision. It lies in the time it gives to prepare a non-reactive response.
Three types of intervention can be pre-financed starting now. Social safety nets first: conditional or unconditional cash transfers to rural households identified as vulnerable in at-risk zones. These programs exist in most of the countries concerned—Kenya, Ethiopia, Mozambique, Zambia—but their funding is annual and often insufficient to absorb a shock of the expected magnitude. Supplementing them before the peak, based on the probabilistic forecast, would make it possible to maintain household consumption without them having to liquidate their livestock.
The stockpiling of recovery seeds second. A corn or sorghum seed purchased today, stored in appropriate conditions, and distributed to households after the poor season costs a fraction of what an additional season of delay costs. Organizations like the WFP and the Ethiopian and Kenyan governments experimented with emergency seed reserves after the 2015-2016 El Niño, with strong results on the speed of agricultural recovery. Building these reserves in 2027, before seed markets are disrupted, is a decision that can be made today.
The prepositioning of food stocks finally, in isolated areas that take three to four weeks to resupply by land in normal times and much longer in crisis situations. This prepositioning exists in the humanitarian system, but it is often decided after the crisis is declared, which reduces its effectiveness for households that have already liquidated part of their livelihoods.
The logic is that of insurance. You don’t wait for a fire to take out a policy. On this point, African states already have parametric insurance instruments against drought that show that activating coverage before the risk occurs is feasible; the challenge is to generalize this logic to social safety nets and stocks.
The mechanisms missing to act on forecast
The lack of pre-arranged financing is a major documented obstacle, compounded by institutional gaps in preparedness and triggering, as well as legal and operational barriers. Some humanitarian budgets remain primarily oriented toward response to declared crises, but humanitarian mechanisms also finance anticipatory and preventive action. A donor that releases funds in 2027 for a crisis whose peak is expected in 2028 steps outside the usual framework of its emergency procedures. This action is only prevention or anticipatory action if the funds actually finance measures implemented before the shock or its acute impacts. Its accounting rules, its outcome indicators, and sometimes its legal mandates are not calibrated for this logic.
Instruments nonetheless exist in prototype form. The risk-based emergency financing mechanism established by the United Nations Office for the Coordination of Humanitarian Affairs allows in certain cases for anticipated disbursements based on forecasts. The African Risk Capacity, which insures African governments against climate disasters, offers automatic payments triggered when meteorological models cross certain thresholds, without waiting for field assessment. These mechanisms exist. They are under-capitalized and underused.
The question posed by the FEWS NET 2027-2028 window is whether donors—the World Bank, the European Union, USAID, major donor countries—can release preventive funds based on a 90 percent probability, without waiting for probability to become certainty. This is a political question as much as a technical one.
Thermal protection as urban infrastructure in Africa followed a comparable path: adaptation investments that seemed premature when risks were probabilistic became urgent when temperatures crossed critical thresholds. Climate adaptation works better when it precedes the threshold.
Two trajectories for 2027-2028
The window identified by FEWS NET outlines two distinct trajectories, depending on decisions made in the coming months.
In the first, the governments of at-risk countries and their partners mobilize starting now. Existing social safety nets are supplemented. Seed stocks are built up. Cash transfers to vulnerable households are activated before local food prices start climbing. Such anticipated mobilization would make it possible to significantly reduce livelihood losses and accelerate recovery after the crisis.
The total cost of the response would likely be lower than that of a purely reactive response.
In the second trajectory, an early response is launched from the alert onward to prevent high-risk adaptation strategies and major livestock losses. Donors wait for the crisis declaration. Governments wait for the price spike. In the late scenario, aid arrives after dangerous adaptation strategies have been adopted and when acute malnutrition is high. At that point, households have already ceded part of their productive assets and entered a spiral of impoverishment whose recovery is measured in years.
The humanitarian and economic cost of this trajectory is documented by previous crises: it is systematically higher than the cost of prevention.
The two scenarios are not symmetric in their uncertainty. The NOAA Climate Prediction Center estimates the probability that El Niño will occur at greater than 90 percent, according to FEWS NET. Reactive interventions historically arrive after households have liquidated a substantial part of their productive assets: this is the documented functioning of previous food crises.
What remains uncertain is the political and institutional will to act on a forecast rather than on a declared crisis. This lock is the only one that is not climatic.
Two signals to track if action precedes the crisis
The deployment of preparations can be tracked through two concrete indicators.
The first is the share of rural households in zones identified by FEWS NET as most exposed that are covered by an active social safety net before the start of the at-risk season, before October 2027. If this coverage progresses significantly by then in the countries concerned, it is the signal that governments and donors have incorporated the probabilistic forecast into their budgetary decisions. If it stagnates, the signal is the opposite.
The second is the delay between the issuance of a FEWS NET alert and the first actual disbursement of food or social assistance in the countries concerned. This delay generally stretches over several months. Shortening it by drawing on anticipated financing mechanisms is technically possible. Organizations like the WFP and the African Risk Capacity have instruments to achieve this. The issue is to activate them before November 2027, not after.
The IPCC, in its sixth assessment report, underscores the importance of climate services, social safety nets, and contingent financing to improve adaptation in countries with high climate exposure. Adapting agricultural and social policies based on probabilistic forecasts, rather than waiting for declared crises, is a key vector of this adaptation. The 2027-2028 horizon is an opportunity to test whether this deficit can be filled.
The window is open. It will not remain so.
Sources
- FEWS NET, 2026-2027 El Niño food security impacts, October 2026
- GEOGLAM Crop Monitor, special El Niño report (no guaranteed URL)
- WFP (World Food Programme), reports on post-2015-2016 El Niño emergency seed reserves (no guaranteed URL)
- African Risk Capacity, parametric insurance mechanisms against climate disasters (no guaranteed URL)
- IPCC, Sixth Assessment Report, Working Group II, Climate Change 2022: Impacts, Adaptation and Vulnerability (no guaranteed URL)



