The Oceanic Niño Index for May to July 2026, higher at the same point of the year than in 1997 (+1.13), 2015 (+1.19) or 2023 (+0.73), the episodes forecasters use as analogues and benchmarks.
On 1 September 2026 UNICEF warned that more than 162 million children, over 2 in 3 children in Eastern and Southern Africa, live in areas exposed to the impact of the strengthening 2026-27 El Niño, and named 13 countries facing urgent and overlapping risks: Djibouti, Eritrea, Ethiopia, Kenya, Madagascar, Malawi, Mozambique, Somalia, South Sudan, Sudan, Uganda, Zambia and Zimbabwe (UNICEF, 1 September 2026). The warning rests on a Pacific that is already running ahead of its analogues. NOAA's Climate Prediction Center holds an El Niño Advisory, gives a greater than 90% chance of a very strong event this autumn and winter, and puts the chance that October-December 2026 exceeds every El Niño since 1950 at 69% (CPC ENSO diagnostic discussion, 13 August 2026). The Oceanic Niño Index reached +1.39°C for May-July 2026 (CPC ONI table), and the weekly Niño-3.4 anomaly stood at +2.6°C (OISSTv2.1 weekly index, 1991-2020 base) in the week of 26 August, after a peak of +2.7°C on 12 August (CPC weekly indices).
PRISM's record for the 13 countries shows what the last 2 episodes cost. During 2023-24, EM-DAT counts 8.4 million people affected by floods in the 8 Eastern African countries and 26.8 million affected by drought in the 5 Southern African countries: 4.0 and 2.6 times the 2015-16 totals, and 3.5 and 4.2 times the annual mean of the 2017-19 baseline. Disasters caused 7.4 million new internal displacements across the 13 countries over the same 2 years (IDMC). The geography follows the teleconnection the literature describes as a dipole: a wet east in the October-December short rains and a dry south through the austral summer, a pattern Cai and colleagues link to interactions with the Indian Ocean Dipole (Cai et al., 2025).
The forecasts for the season ahead reproduce that geography. ICPAC gives a 90% probability of above-normal October-December rainfall over southern Ethiopia, central to southern Somalia and north-eastern Kenya (GHACOF 74 technical statement, 18 August 2026); SARCOF-33 favours below-normal rainfall across Zimbabwe, Mozambique and southern Zambia with high confidence (SADC, 28 August 2026). FEWS NET already counts at least 51.5 million people in IPC Phase 3 or worse across 11 of the 13 countries in its July 2026 briefs, rising to at least 57.5 million in its most-likely projections (FEWS NET data warehouse, read 2 September 2026). UNICEF asks for US$173 million within existing appeals and notes that less than 1% of international humanitarian assistance is available for anticipatory action.
The piece in 90 seconds. Every figure on screen comes from the dataset below. Also on YouTube.
Each of the last 2 El Niño episodes left a larger mark on the 13 countries than the one before it. EM-DAT records 8.4 million flood-affected people in Eastern Africa and 26.8 million drought-affected people in Southern Africa during 2023-24, against 2.1 million and 10.1 million in 2015-16, even though the 2015-16 event was the stronger of the 2 in the Pacific (peak ONI +2.59 against +1.99). CPC gives the 2026-27 event a 95% probability of very strong status in October-December, level with the 2015-16 class on the ONI and above the 2023-24 event, and a 69% chance of exceeding every El Niño since 1950.
The figures marked as PRISM's come from a curated panel of the 13 countries UNICEF named, 2000 to 2026, assembled on 3 September 2026 (local time) from PRISM's production APIs (EM-DAT disaster counts and affected totals, GDACS alerts, IDMC displacement, HDX HAPI dekadal rainfall, EC JRC ASAP hotspots, INFORM Risk and INFORM Climate, UNICEF child indicators, OCHA FTS plan funding), from NOAA CPC's ONI table, from FEWS NET's public data warehouse (read 2 September 2026, 22:18 UTC) and from World Bank population data. EM-DAT appears as event counts and affected totals only, under its licence.
FEWS NET publishes ranges; every FEWS NET figure here is the low end of its range and reads as "at least". El Niño episodes follow NOAA's definition (5 consecutive overlapping seasons at or above +0.5°C). The panel starts in 2000, so the 1997-98 episode appears through its ONI peak and through the primary reports of the time.
Children under 18 are estimated as World Bank ages 0-14 plus 60% of ages 15-19, a different quantity from UNICEF's "children in exposed areas". The full dataset is deposited on Zenodo as the PRISM El Niño 2026-27 dataset, version 1.0 (DOI 10.5281/zenodo.22272966), under CC BY-SA 4.0 with the EM-DAT columns excluded, and will be listed on PRISM's climate and disasters data catalogue.
What the record says
Since 1996 NOAA's Oceanic Niño Index has registered 9 completed El Niño episodes and the developing 2026 event, 4 of the completed ones strong or very strong: 1997-98 (peak +2.37), 2009-10 (+1.50), 2014-16 (+2.59) and 2023-24 (+1.99). The 2026 event stands at +1.39 after 2 seasons above the threshold, already 5th among those peaks while still developing. The region's 2 largest disaster years in PRISM's panel fell outside El Niño: 2022, when 41.3 million people across the 13 countries were affected by drought or flood in a La Niña year, and 2021 (26.9 million, also La Niña); 2015 (26.6 million) was the El Niño peak year and 2024 (26.1 million) carried the Southern African drought that followed the 2023 Pacific peak. Averaged over 2000-2025, El Niño-labelled years carried 7.8 million affected people against 14.4 million in other years.
For planners the record carries a lesson about timing. The Pacific peaks in November-January; the Eastern African floods land in the October-December short rains of the same year; the Southern African drought registers in the harvest of the following calendar year, which is why the record places the 9.8 million drought-affected in Zambia and the 7.6 million in Zimbabwe in 2024, the year after the 2023 Pacific peak. The strip below carries the whole period on one line: the 4 strong or very strong episodes as coloured segments, the drought-and-flood-affected curve for the 13 countries ghosted behind, and 14 event dots for the largest country-years in the record. Click a dot to read it.
1997-98. FAO/GIEWS put Somalia's flood toll at an estimated 2,000 deaths and 250,000 displaced, and counted some 10 million people across Eastern Africa in need of emergency assistance (FAO/GIEWS, 5 February 1998). Rift Valley fever followed the floods: 478 unexplained deaths in North Eastern Province, Kenya, and southern Somalia were reported in December 1997 (CDC MMWR, 1998).
2015-16. Ethiopia's 2016 Humanitarian Requirements Document counted 10.2 million people needing emergency food assistance and a US$1.4 billion requirement (Government of Ethiopia and partners, December 2015); SADC declared a regional emergency on 26 July 2016 for 41 million people (RIASCO, 2017).
2023-24. Somalia's Deyr floods affected 2.48 million people, displaced 1.2 million and killed about 118 (SoDMA figures, OCHA, 24 December 2023). Zambia declared a national disaster on 29 February 2024 (DMMU, April 2024), Malawi in March 2024, in 23 of its 28 districts (OCHA flash appeal) and Zimbabwe on 3 April (IFRC, May 2024); SADC appealed for at least US$5.5 billion for over 61 million people (SADC, 20 May 2024).
CPC (13 August 2026). Greater than 90% chance of a very strong El Niño in Northern Hemisphere autumn and winter; 69% chance that October-December 2026 exceeds every event since 1950 on the relative ONI (ENSO diagnostic discussion).
IRI (19 August 2026). 25 of 26 models place October-December in the very strong category and 15 forecast Niño-3.4 at or above +3.0°C (CCSR/IRI technical update).
ICPAC (18 August 2026). 1997 and 2023 selected as analogue years for October-December, both wetter than normal across most of the Greater Horn (GHACOF 74 statement).
OCHA (August 2026). A 70% chance that October-December rainfall deficits in Southern Africa exceed those of any El Niño year since 1950 (Southern Africa preparedness alert); the 2026-27 event "could be equally severe, or worse" than 2023-24, when 61 million people across Southern Africa, the Democratic Republic of the Congo included, required assistance (OCHA, 7 August 2026).
- FAO/GIEWS, Heavy rains attributed to El Niño cause extensive crop damage in parts of Eastern Africa (February 1998)
- OCHA/RIASCO, Report on the RIASCO Action Plan for the El Niño-induced drought in Southern Africa 2016/2017
- OCHA, Eastern Africa El Niño impact snapshot (December 2023)
- OCHA, The humanitarian impacts of El Niño in Southern Africa (September 2024)
- Cai et al., Climate impacts of the El Niño-Southern Oscillation in Africa, Nature Reviews Earth & Environment (2025)
The dipole
The mechanism is well established. Cai and colleagues identify a strong dipole in rainfall anomalies between eastern and southern Africa during October-December and December-February as the defining feature of ENSO's African footprint: El Niño brings floods to East Africa in the October-December short rains and drought to southern Africa in the December-February summer (Cai et al., 2025). ENSO and the Indian Ocean Dipole together exert a dominant influence on the year-to-year variability of Eastern African rainfall (Palmer et al., 2023), and southern Africa is dry in El Niño austral summers (Steinkopf and Engelbrecht, 2025).
PRISM's panel records the same split in the affected totals. The chart plots each country's annual mean of flood-affected people (Eastern Africa, above the axis) or drought-affected people (Southern Africa, below the axis) in each episode, against the same measure for the 2017-19 comparison years.
In 2023-24 the split matches the mechanism: Somalia 1.78 million flood-affected per year, Ethiopia 0.96 million, South Sudan 0.70 million and Kenya 0.59 million above the line; Zambia 4.9 million, Zimbabwe 3.8 million, Malawi 3.05 million and Mozambique 1.65 million drought-affected per year below it. In 2015-16 drought dominated on both sides of the divide: Ethiopia's 2015 drought alone affected 10.2 million people and sits in the flood-plotted half of the chart, which is why the eastern bars are short in that episode. Use the toggle to switch between the strength classes.
Chart unavailable; the figures are in the text and in the dataset workbook.
2023-24, Eastern Africa. 19 flood events and 8.43 million flood-affected people across the 8 countries, 4.0 times the 2015-16 total and 3.45 times the 2017-19 annual mean. Somalia's 3.33 million in 2023 is the largest flood country-year in the panel since Mozambique's 4.5 million in 2000.
2023-24, Southern Africa. 4 drought events and 26.8 million drought-affected people across the 5 countries, 2.64 times 2015-16 and 4.19 times the 2017-19 annual mean. Zambia 9.8 million, Zimbabwe 7.6 million, Malawi 6.1 million, Mozambique 3.3 million.
2015-16. Eastern Africa recorded 2.11 million flood-affected and 24.7 million drought-affected people; Southern Africa 10.14 million drought-affected across Malawi (6.7 million), Mozambique (2.3 million) and Madagascar (1.14 million). FAO/GIEWS put the food-insecure caseload at an estimated 6.7 million in Malawi and just over 4 million in Zimbabwe in 2016 (FAO, 2018).
Wet east. ICPAC's Zone I carries a 90% probability of above-normal October-December rainfall over southern Ethiopia, central to southern Somalia and north-eastern Kenya, and a greater than 90% chance of more than 400 mm over central to southern Somalia, central Kenya and the Lake Victoria basin (GHACOF 74). IRI's Indian Ocean Dipole forecast turns positive from August and stays strongly positive through December (IRI, 19 August 2026).
Dry south. SARCOF-33 favours below-normal October-December rainfall across Zimbabwe, Mozambique and southern Zambia with high confidence, and keeps the dry signal through November-January, December-February and January-March across southern Zambia, most of Zimbabwe, central Mozambique and southern Malawi (SADC, 28 August 2026). FEWS NET expects widespread Crisis (IPC Phase 3) outcomes from October 2026 to January 2027 across southern and central Malawi, southern Mozambique and Madagascar's Grand Sud (FEWS NET, June 2026).
Where the season stands
The dry half of the dipole is already in place across the northern sector. ICPAC's Climate Watch of 11 August recorded rainfall deficits from May through July across Djibouti, Eritrea, Ethiopia, Somalia, South Sudan, Sudan, Uganda and western Kenya, with widespread soil-moisture deficits and vegetation stress (ICPAC, August 2026). The regional Food Security and Nutrition Working Group found that central and northern Ethiopia, Sudan's pastoral belt, most of Eritrea and Djibouti, Karamoja and western and north-western Kenya received less than 60% of normal rainfall between 1 June and 25 August, and in the worst-affected areas the season ranks among the driest on record (FSNWG special report, 19 August 2026). FEWS NET reports Ethiopia's June-July as the driest on record across much of the country, with deficits of roughly 40% to 70% in the north, east and centre, and a below-average meher harvest as the most likely outcome (FEWS NET, 7 August 2026).
PRISM's dekadal rainfall reading (HDX HAPI, read 3 September 2026) puts 5 of the 13 countries below their long-term average and 8 above, with Kenya at 130% of average, Mozambique at 116% and Madagascar at 85%. A single dekad at the tail of the dry season carries limited weight for the south and for the Horn's eastern lowlands, where the rains that matter open in October; the reading matters most for the unimodal north, where the June-September season is ending. The EC JRC's ASAP system flags 8 of the 13 countries: Somalia, South Sudan and Sudan as major hotspots, and Kenya, Eritrea, Ethiopia, Madagascar and Mozambique as hotspots. GDACS has issued 237 alerts across the 13 countries in 2026 to date, most of them wildfire alerts in Sudan (107), Zambia (65) and Ethiopia (23).
Chart unavailable; the figures are in the table below and in the dataset workbook.
| Country | Rainfall vs avg | ASAP status | GDACS 2026 | FEWS NET areas of highest concern |
|---|---|---|---|---|
| Somalia | 96% | Major hotspot | 2 | Bay, Bakool, Gedo and Hiiraan agropastoral areas; riverine areas; IDP settlements |
| South Sudan | 107% | Major hotspot | 1 | Jonglei and southern Upper Nile |
| Sudan | 96% | Major hotspot | 109 | North Darfur; Greater Kordofan; areas of Greater Darfur with high numbers of IDPs |
| Kenya | 130% | Hotspot | 5 | Refugee camps; northern and eastern pastoral areas |
| Ethiopia | 101% | Hotspot | 28 | Northern and eastern cropping areas (meher deficit); southern and south-eastern lowlands (flood) |
| Eritrea | 99% | Hotspot | 1 | Gash Barka and Debub crop areas (ASAP) |
| Madagascar | 85% | Hotspot | 6 | Atsimo-Andrefana, Androy, Anosy and Atsimo-Atsinanana |
| Mozambique | 116% | Hotspot | 10 | Cabo Delgado and southern Mozambique |
| Uganda | 107% | No hotspot | 3 | Refugee settlements and Karamoja |
| Malawi | 103% | No hotspot | 1 | Southern Malawi |
| Zambia | 98% | No hotspot | 67 | Southern and south-western Zambia |
| Zimbabwe | 108% | No hotspot | 4 | Deficit-producing areas in the south, east, west and extreme north |
| Djibouti | 102% | No hotspot | 0 | Refugee camps; Obock, Ali Sabieh and Tadjourah (peak needs record) |
Acute food insecurity, before the season turns
FEWS NET's July 2026 briefs count at least 51.5 million people in IPC Phase 3 or worse across 11 of the 13 countries, with Sudan at 22 million, Ethiopia and South Sudan at 8 million each and Somalia at 5 million (Ethiopia's brief is dated January 2026; FEWS NET data warehouse, read 2 September 2026). Its most-likely projections sum to at least 57.5 million: Sudan 19 million in January 2027, Ethiopia 15 million in July 2026 and South Sudan 7 million.
The largest of those figures rest on war before they rest on weather. In Sudan, FEWS NET attributes the outcomes to the fighting between the Sudanese Armed Forces and the Rapid Support Forces, now in its 4th year, with drone strikes, sieges and displacement blocking access to fields, and adds that El Niño-linked below-average rains are expected to limit harvests where rainfall in North Darfur and parts of North Kordofan already ranks as the driest in the 40-year record (FEWS NET Sudan, July 2026 to January 2027). South Sudan stays among the most food-insecure countries in the world through January 2027 because of conflict and El Niño-driven drought, seasonal floods and a collapsing economy, with a credible risk of Famine in Akobo, Nyirol, Ulang and Nasir (FEWS NET South Sudan, August 2026). In Somalia, clan conflict in the central regions and fighting in parts of Lower Shabelle restrict grazing, markets and farming, and more than 738,000 people were displaced between January and July 2026, before the Deyr floods arrive (FEWS NET Somalia, July 2026).
Somalia rises from 5 million to 6 million, and FEWS NET assesses a credible risk of Famine (IPC Phase 5) in the Bay, Bakool, Gedo and Hiiraan agropastoral areas, all riverine areas except Lower Shabelle, and the Baidoa, Jowhar, Beletweyne and Dolow IDP settlements between October and January in the event of extreme, widespread flooding (FEWS NET Somalia outlook, June 2026). Uganda rises from 0.75 million to 2 million, Zimbabwe from 1 million to 2 million and Zambia from 0.25 million to 0.5 million.
The IPC's own Zambia analysis in PRISM's archive moves from 697,809 people in Phase 3 or worse for April-September 2026 to 1,253,124 for October 2026 to March 2027, a rise of 79.6% (PRISM country profile API, Zambia). Save the Children puts the rise at 72%, from nearly 700,000 people to about 1.2 million (Save the Children, 27 August 2026).
Chart unavailable; the figures are in the text and in the dataset workbook.
Late 2023, the last wet Deyr. Across Eastern Africa at least 479 people were killed and more than 5.2 million affected by heavy rains and flooding between September and mid-December 2023, with nearly 2 million displaced (OCHA, December 2023). Kenya recorded at least 174 deaths and more than 500,000 people displaced by 18 December (OCHA, 20 December 2023); Ethiopia an estimated 1.5 million affected, more than 632,700 displaced and 57 deaths (OCHA, 1 December 2023). Somalia's post-disaster assessment put damages and losses at US$193.4 million (SoDMA, UN, World Bank and EU, 2024).
Somalia. FEWS NET sees strong odds that cumulative September-November precipitation exceeds 120% of the historical average, with assistance needs peaking at 6.0 to 6.99 million people in late 2026 (FEWS NET, June 2026); FAO-SWALIM expects a dry period until October followed by elevated risk of extensive flooding along the Juba and Shabelle (FAO Somalia, 2 June 2026).
Displacement. The Danish Refugee Council's scenarios put flood-related displacement across Somalia, Kenya, South Sudan and Uganda at 750,000 in the less severe case and up to 3 million in the most severe case, with Somalia accounting for 60% and Kenya 33% of the worst case (DRC, 25 August 2026). Kenya's Interior Ministry has flagged 18 high-risk counties (Kenyans.co.ke, 10 August 2026).
Who is exposed
PRISM estimates 215.9 million children under 18 in the 13 countries, 46.7% of a combined population of 462.1 million: 154.1 million in the 8 Eastern African countries and 61.9 million in the 5 Southern African countries. Ethiopia alone holds 61.4 million, followed by Uganda (25.7 million), Kenya (24.8 million) and Sudan (24.1 million); the 3 largest account for 51.9% of the region's children. UNICEF's figure of more than 162 million counts children in exposed areas, a narrower quantity than a whole-country total, and the 2 numbers are not interchangeable. The unit chart below gives each million children a square.
The systems that would absorb the shock are already under strain. OCHA's Eastern Africa alert counts 48.5 million people needing humanitarian assistance across its 11 countries, more than 40 million facing acute food insecurity and 8.3 million children acutely malnourished (OCHA, 14 August 2026). Among the 9 countries reporting stunting, 5 sit at or above 30%, led by Madagascar (39.8% in 2021), Ethiopia (36.8% in 2019) and Mozambique (36.7% in 2022); Djibouti reports the highest wasting prevalence at 12.1% (2023). Safely managed drinking water reached at most 27.5% of the population in 2024 in the 6 countries that report it, and 13.5% in Ethiopia.
The 10 OCHA-coordinated plans across the countries carry 108.9 million people in need, 30.4 million of them in Sudan, 28.6 million in Ethiopia and 10 million in South Sudan. UNICEF's country lines are specific: flooding in Somalia could directly affect up to 2.5 million people under the most severe scenario mapping; tens of thousands of children in Ethiopia are likely to be affected in areas already facing severe food insecurity; in South Sudan, drought, extreme heat and localised flooding threaten to deepen food insecurity and malnutrition and compound famine risks (UNICEF, 1 September 2026).
| Country | Children <18 | Stunting (year) | Wasting (year) | Safe water (year) | INFORM Risk |
|---|---|---|---|---|---|
| Ethiopia | 61.4M | 36.8% (2019) | 6.8% (2019) | 13.5% (2024) | 7.1 very high |
| Uganda | 25.7M | 24.4% (2022) | 3.2% (2022) | 18.1% (2024) | 6.6 very high |
| Kenya | 24.8M | 17.6% (2022) | 4.5% (2022) | · | 6.1 high |
| Sudan | 24.1M | · | · | · | 7.8 very high |
| Mozambique | 18.1M | 36.7% (2022) | 3.8% (2022) | 27.5% (2024) | 7.1 very high |
| Madagascar | 14.8M | 39.8% (2021) | 7.2% (2021) | 22.1% (2024) | 5.7 high |
| Malawi | 10.5M | 34.9% (2021) | 2.1% (2021) | 18.2% (2024) | 4.7 medium |
| Zambia | 10.5M | 34.6% (2018) | 4.2% (2018) | · | 5.8 high |
| Somalia | 10.4M | · | · | · | 8.2 very high |
| Zimbabwe | 8.0M | 25.9% (2024) | 5.1% (2024) | 25.5% (2024) | 4.9 medium |
| South Sudan | 5.6M | · | · | · | 8.3 very high |
| Eritrea | 1.6M | · | · | · | 4.4 medium |
| Djibouti | 0.4M | 25.5% (2023) | 12.1% (2023) | · | 5.2 high |
What the forecast says
The Pacific forecast leaves little room for a weak outcome. CPC's official probabilities, issued in August 2026, give El Niño a 100% probability for every season from July-September 2026 to January-March 2027, 97% for February-April and 82% for March-May 2027, with La Niña at 0% throughout (CPC ENSO probabilities). The strength table puts a very strong event (relative ONI at or above +2.0°C) at 93% for September-November, 95% for October-December and 90% for November-January, before the odds fall through the spring (CPC strength probabilities).
IRI's 26-model plume agrees: July's monthly Niño-3.4 anomaly of +2.03°C, a Southern Oscillation Index of -29.1 and subsurface temperatures more than 8°C above climatology in places, with 100% model-based El Niño probability through February-April 2027 (IRI, 19 August 2026). WMO's seasonal update forecasts a seasonal-average SST anomaly of about 2.9°C in the El Niño region for August-October, with the intensification peaking in November, and a positive Indian Ocean Dipole reaching a seasonal mean of 0.6°C (WMO GSCU, 31 July 2026). CPC adds that an event of this magnitude raises the chances of El Niño-consistent impacts without guaranteeing them.
Chart unavailable; the probabilities are in the text and in the dataset workbook.
Eastern Africa: the October-December season
ICPAC's outlook for October-December 2026, issued at GHACOF 74 in Kigali on 18 August, gives a 90% probability of above-normal rainfall over southern Ethiopia, central to southern Somalia and north-eastern Kenya, 65% and 55% zones across most of the rest of the Greater Horn, and a 45% below-normal zone limited to cross-border south-eastern Sudan and north-eastern South Sudan. ICPAC expects an early to normal onset in the east and a delayed onset in the west, where El Niño suppresses September rainfall before its regional influence turns wetter; warmer-than-average temperatures are most likely region-wide, with an 85% zone over eastern Kenya, Ethiopia, Eritrea, Djibouti and much of Somalia (GHACOF 74 technical statement). The season delivers up to 70% of annual rainfall in parts of Ethiopia, Kenya and Somalia (ICPAC press release, 18 August 2026).
ICPAC's September-November product expects above-normal rainfall over most of the IGAD region and below-normal rainfall over central to north-eastern South Sudan, central and western Ethiopia and central to southern Sudan (ICPAC seasonal forecast). Kenya's Meteorological Department expects above-average rainfall in more than 40 of Kenya's 47 counties, with onset in the 2nd to 3rd week of October (The Kenya Times, reporting KMD's 26 August outlook); Uganda's Ministry of Water and Environment expects near-normal to above-normal rains, established mid to late September and peaking from the end of October (Uganda Radio Network, 1 September 2026).
Southern Africa: the 2026-27 rainy season
SARCOF-33, held in Swakopmund on 25-28 August, expects the bulk of the SADC region to receive below-normal rainfall in October-December 2026, driven by an El Niño peaking around that season and a strong positive Indian Ocean Dipole likely by September. For the 5 countries UNICEF named, SARCOF-33 expects below-normal rainfall with high confidence across Zimbabwe, Mozambique and southern Zambia; normal to above-normal rainfall over northern Zambia, central Malawi and northern Mozambique; above-normal rainfall with high confidence over extreme northern Malawi and Zambia; and normal to below-normal rainfall over western and south-western Madagascar. The dry signal persists into January-March 2027 across the bulk of Zambia, central and southern Malawi, central Mozambique and the bulk of Zimbabwe, with above-normal temperatures from October to January outside a cooler central to north-eastern band. SADC cautions that the outlook must not be read at sub-regional, country or local scale and points users to their national services (SARCOF-33 statement).
Those services are aligning: Malawi's DCCMS gives a high chance of below-normal rainfall over most of the Southern and Central regions, with normal to above-normal in the North and a significantly delayed onset in the South (Maravi Express, 2 September 2026); Zimbabwe's Meteorological Services Department issued a preliminary warning in May of an 88% to 94% chance of El Niño and a 65% historical likelihood of below-normal rainfall (MSD Zimbabwe, 8 May 2026); Mozambique's forecasts presented to the Council of Ministers expect below-normal rainfall in the South and Centre and above-average rainfall in the North (AIM, 19 May 2026); Madagascar's meteorological service expects normal to below-normal September-November rainfall in Melaky, Menabe, Atsimo Andrefana, Anosy, Androy and Ihorombe (DGM Madagascar, SON 2026 bulletin).
| Country | Signal | Outlook | Issuing body |
|---|---|---|---|
| Somalia | Wet, OND | 90% above normal in the centre and south; >90% chance of more than 400 mm | ICPAC, GHACOF 74 |
| Kenya | Wet, OND | 90% above normal in the north-east; KMD: above average in more than 40 of 47 counties | ICPAC; KMD |
| Ethiopia | Mixed | 90% above normal in the south; below normal in the centre and west in SON; kiremt deficits already recorded | ICPAC; FSNWG |
| Uganda | Wet, SOND | Near to above normal; delayed onset in the west of the region; Karamoja near to slightly above normal | ICPAC; UNMA |
| South Sudan | Mixed | Above normal over most of the country; 45% below-normal zone in the north-east | ICPAC |
| Sudan | Dry, SON | Below normal over central to southern Sudan in SON; OND is climatologically dry over most of the country | ICPAC |
| Djibouti, Eritrea | Warm | Above-normal temperatures in the 85% zone; Djibouti in the above-normal rainfall statement for SON, Eritrea not named | ICPAC |
| Zimbabwe | Dry, OND to JFM | Below normal with high confidence in OND and through the bulk of the country to JFM 2027 | SARCOF-33; MSD |
| Mozambique | Dry south and centre | Below normal with high confidence except the north (normal to above); central Mozambique dry through JFM | SARCOF-33; Government of Mozambique |
| Zambia | Dry south | Southern Zambia below normal with high confidence; northern normal to above; bulk of the country below normal by JFM | SARCOF-33 |
| Malawi | Dry south from NDJ | SARCOF: central normal to above, extreme north above in OND, southern Malawi below normal from NDJ; DCCMS: high chance of below normal over most of the Southern and Central regions, onset significantly delayed in the South | SARCOF-33; DCCMS |
| Madagascar | Mixed | West and south-west normal to below; north and east normal to above; Grand Sud household stocks likely to last 2 months against a typical 4 | SARCOF-33; DGM; FEWS NET |
1997. Somalia's worst floods in decades followed torrential rains from mid-October; crop losses reached around 80% in the worst-affected riverine districts along the Juba and Shabelle (FAO/GIEWS, February 1998). In Garissa District an estimated 27,500 people were infected with Rift Valley fever in the largest outbreak then recorded in East Africa (Woods et al., 2002).
2023. FEWS NET's own benchmark for Somalia's Deyr: 1.2 million people displaced, 8,000 livestock lost, 11,000 hectares of cropland destroyed and about US$193.4 million in damages (FEWS NET, June 2026). In the FSNWG's count, EM-DAT recorded 2,823 flood deaths across the region in the 1997 analogue and 705 in 2023 (FSNWG, 19 August 2026).
2015-16 and 2023-24 in the south. The 2015-16 event was stronger as a climate event, while the 2023-24 impacts were more significant and reached more countries (OCHA, September 2024); Zimbabwe's 2024 maize harvest of around 635,000 tonnes ran about 60% below its 10-year average (FEWS NET, June 2024) and Zambia's estimated 1.6 million tonnes almost 50% below its 5-year average (FAO/GIEWS, October 2024).
Strength. CPC's relative ONI reached +1.0°C for May-July 2026; the 2015 series peaked at +2.3 and the 2023 series at +1.4 on the same table (CPC, 31 August 2026), and CPC's historic threshold for October-December is +2.5 (CPC ENSO diagnostic discussion, 13 August 2026). The EC JRC calls an El Niño this year virtually certain, very likely very strong and possibly unprecedented, with the caveat that in the most extreme scenario the models are extrapolating beyond any historical precedent (JRC, 15 June 2026).
Timing. WFP expects the event to peak between September and December 2026 with effects felt throughout 2027 (WFP, 5 August 2026); JRC expects extreme heat to peak from December 2026 to February 2027 (JRC, 15 June 2026); IRI's ensemble weakens sharply in spring 2027, consistent with strong events decaying quickly after their peak (IRI, 19 August 2026).
Uncertainty. ICPAC records that Indian Ocean Dipole forecast skill has historically been low (GHACOF 74); SARCOF notes that areas of weak or mixed model probability may see rainfall closer to normal (SARCOF-33); WFP's Somalia analysis states that the forecasts say the season will be wet without saying how severe (WFP, 27 August 2026).
- NOAA CPC, ENSO Diagnostic Discussion (13 August 2026)
- IRI, ENSO forecast and CCSR/IRI technical update (19 August 2026)
- ICPAC, Technical statement from GHACOF 74 (18 August 2026)
- SADC, Statement from SARCOF-33 (28 August 2026)
- WMO, El Niño impacts Greater Horn of Africa (19 August 2026)
- FSNWG, Special report on El Niño (19 August 2026)
The decision window
The calendar compresses the choices. The Eastern African short rains open in October and deliver their floods before the end of December; the Southern African planting window opens in the same weeks and its harvest, or the lack of one, registers from January. Every forecast update between now and then narrows the range, and every week that passes shortens the lead time for pre-positioning. The band below places the checkpoints that will update this picture against the seasons they concern.
- 10 September 2026. CPC's next ENSO diagnostic discussion; IRI's CCSR/IRI forecast follows on the 19th of each month (CPC; IRI).
- Mid to late September. Uganda's steady rains establish, peaking from the end of October and running to early January (Ministry of Water and Environment, via URN); ICPAC expects a delayed onset in the western Greater Horn and an early to normal onset in Somalia, central to eastern Kenya and southern Ethiopia (GHACOF 74).
- 2nd to 3rd week of October. Kenya's short-rains onset in Nairobi and the central highlands, 3rd to 4th week in the south-east (KMD, via The Kenya Times). OCHA reports Somalia's partners pre-positioning supplies before October, with peak flood risk anticipated between September and December (Xinhua, citing OCHA, 18 August 2026).
- October 2026 to March 2027. Southern Africa's rainy season and lean season; the IPC's Zambia projection window (1,253,124 people in Phase 3 or worse) and FEWS NET's expectation of widespread Crisis outcomes across southern and central Malawi, southern Mozambique and the Grand Sud from October to January (FEWS NET).
- Mid-December. Malawi's rains establish, with onset significantly delayed in the South and a season of 100 to 120 days there (DCCMS, via Maravi Express). JRC expects extreme heat to peak from December to February (JRC).
- January to March 2027. FEWS NET's peak-needs windows for Madagascar, Malawi and Mozambique, and Zambia's from December to February; the FAO-WFP joint appeal and FAO's Ethiopia window both close in March 2027 (FAO, 18 June 2026; FAO Ethiopia, 10 August 2026).
Whether the response is ready
The plans that would carry the response are 44.8% funded. Across the 10 countries with an OCHA-coordinated plan, OCHA FTS records US$14.52 billion in requirements against US$6.50 billion in funding, a gap of US$8.02 billion as read on 3 September 2026. Of the 10 plans, 6 sit under 50%: Zimbabwe (36.6%), Ethiopia (36.8%), Malawi (39.6%), Sudan (40.1%), Mozambique (42.7%) and Somalia (47.3%); Kenya (82.3%) and Madagascar (76.2%) are the best covered. Ethiopia carries the largest gap at US$2.52 billion and Sudan the largest requirement at US$4.16 billion.
The plans carry different cycle years in their names, from 2023 flash appeals to the 2026 Mozambique and 2025 Sudan plans, so the total mixes cycles; the chart labels each one. Djibouti, Eritrea and Uganda have no OCHA-coordinated humanitarian plan on FTS.
Chart unavailable; the coverage figures are in the text and in the dataset workbook.
Anticipatory action and the money behind it
UNICEF requires US$173 million within its existing 2026 Humanitarian Action for Children appeals to reach children and families with anticipatory and life-saving assistance across the 13 priority countries, and notes that less than 1% of international humanitarian assistance is currently available for anticipatory action (UNICEF, 1 September 2026). FAO and WFP's joint appeal seeks US$202 million to protect 8.8 million people in 22 countries between June 2026 and March 2027; the 2 agencies state they are already positioned to provide anticipatory action for 1.2 million people and seek a further US$167 million to reach 7.6 million more, which leaves a US$35 million difference by PRISM's arithmetic; 10 of the 13 countries are covered, with Djibouti, Eritrea and Zambia absent (FAO, 18 June 2026). OCHA stands ready to disburse up to US$100 million from CERF, with more than US$20 million already allocated for anticipatory action in 6 countries (OCHA, 13 July 2026). WFP has triggered anticipatory action plans in 6 countries since May, South Sudan and Uganda among them, releasing over US$14 million for 500,000 people, and projects that this El Niño could push at least 49 million more people into acute food insecurity by the end of 2027, more than 18 million of them in East and Southern Africa (WFP, 5 August 2026).
National systems have moved. Ethiopia's Disaster Risk Management Commission has triggered anticipatory action in 114 woredas where an estimated 9.3 million people live, 84 of them at priority levels 1 and 2 in Afar, Amhara, northern Somali and Tigray, and FAO asks for US$50 million to assist 4.66 million people as reported by Addis Standard (FAO, 10 August 2026; Addis Standard via allAfrica, 11 August 2026). Somalia's cabinet approved El Niño preparedness measures on 8 July (Dawan Africa, 8 July 2026); Kenya's Deputy President urged counties on 24 August to activate disaster plans and clear drainage (The Star, 24 August 2026); the SADC Summit of 17 August urged member states to enhance preparedness through early warning, drought-preparedness measures and drought-tolerant crop varieties (SADC communiqué). The Danish Refugee Council has activated drought anticipatory action in Afar for 700 households and cash assistance for 1,228 households in the Karamoja cluster (DRC, 25 August 2026).
2016. Ethiopia's revised requirement of US$1.6 billion was 70% funded, with more than 10.2 million people targeted at the April peak (OCHA EHF annual report, 2017); the RIASCO action plan for Southern Africa received US$900 million of US$1.3 billion, also 70% (RIASCO, 2017).
2024. The flash appeals for Malawi (US$137 million), Zambia (US$228 million), Zimbabwe (US$429 million) and Mozambique (US$222 million) were all under 18% funded by mid-September 2024; CERF allocated US$27.5 million and ARC drought insurance paid US$53.7 million to Malawi, Zambia and Zimbabwe (OCHA, September 2024).
2023 to 2025. Anticipatory action reached at least 3.1 million people across 8 Southern African countries with US$65.4 million in early financing between 2023 and 2025, and, between January 2024 and September 2025, 5.5 million people across 10 Eastern African countries with US$71.2 million (OCHA, 7 August 2026).
Scale. The FAO-WFP appeal's 8.8 million people and UNICEF's 13-country requirement sit against FEWS NET's most-likely count of at least 57.5 million people in Phase 3 or worse across 11 of the countries, and against plans that are 44.8% funded.
Return. UNICEF cites up to US$15 saved in future recovery costs for every US$1 invested in disaster risk reduction; WFP cites US$3 to US$7 in avoided losses per US$1 of anticipatory programming; the FAO-WFP appeal cites up to US$7 (FAO, 18 June 2026). The ratios rest on different evidence bases and carry their own attributions (UNICEF; WFP).
Timing. OCHA's Eastern Africa alert calls for anticipatory action and flexible financing to pre-position assistance before the October-December season (OCHA, 14 August 2026); FAO-SWALIM's meteorologist states that preparedness in Somalia should begin now (FAO Somalia, 2 June 2026).
The 13 countries, tracked live
Each card carries the GDACS alert count for 2026 to date and the EC JRC ASAP hotspot status as read from PRISM on 3 September 2026, and opens the country's public profile, where the alerts, rainfall anomaly, INFORM scores, plan funding and displacement figures update from source. GDACS counts cover every hazard type; the large Sudan, Zambia and Ethiopia counts are mostly wildfire alerts.
The El Niño season is tracked live on PRISM. The climate dashboard carries GDACS alerts, rainfall anomalies, ASAP hotspots and INFORM Climate for every country; Somalia's profile carries the Deyr flood risk, IPC classification and plan funding as they update.
4 decisions for September
UNICEF's US$173 million, the FAO-WFP appeal's further US$167 million and CERF's US$100 million standby together come to US$440 million, about 3% of the US$14.52 billion the region's plans already request, and every agency's return estimate, from US$3 to US$15 per dollar, rests on money arriving before October. Donors should release flexible allocations in September against the CPC and ICPAC triggers already published.
The eastern flood peak runs October to December and the southern lean season runs January to March. Regional coordination should hold the 1997 and 2023 Deyr floods and the 2016 and 2024 southern droughts as named scenarios, pre-position river-basin supplies in Somalia, south-eastern Ethiopia and Kenya's 18 flagged counties now, and move drought-tolerant seed and livestock protection into southern Zambia, Zimbabwe, southern Malawi and central Mozambique before planting.
Stunting is at or above 30% in 5 of the 9 reporting countries and safely managed water reaches at most 27.5% of people in the 6 that report it. Country teams and donors should sequence UNICEF's 5 asks, from pre-positioned nutrition supplies and early screening to disease surveillance and shock-responsive cash, by these baselines, putting Madagascar, Ethiopia, Mozambique, Malawi and Zambia first for nutrition and the flood corridors first for water and cholera surveillance.
Of the 10 plans, 6 are under 50% funded, and Ethiopia's gap alone is US$2.52 billion. Djibouti, Eritrea and Uganda have no OCHA-coordinated plan, and Djibouti and Eritrea are also absent from the FAO-WFP appeal, yet Uganda holds 25.7 million children and Djibouti the highest wasting prevalence in the group. The 2027 planning cycle should carry an explicit El Niño line for each of the 13 countries.
Watchlist
- 10 September 2026
CPC's ENSO diagnostic discussion, with the first official probabilities issued after the August weekly peak of +2.7°C; the ONI value for June-August 2026 follows in the same table.
Updated by NOAA CPC - 19 September 2026
IRI's CCSR/IRI forecast and model plume, including the count of models at or above +3.0°C for October-December and the Indian Ocean Dipole outlook.
Updated by IRI - September 2026
National outlooks to check for the 2026-27 season from Zimbabwe's MSD, Zambia's ZMD, Mozambique's INAM and Ethiopia's EMI (Bega); each sharpens SARCOF's and ICPAC's regional zones to district scale.
- October 2026
Onset of the short rains against ICPAC's early-to-normal expectation in the east and delayed expectation in the west; ICPAC's floods and heavy-rainfall products and its Climate Watch advisories.
Updated by ICPAC - October to December 2026
FEWS NET's Somalia outlook updates against its Famine-risk statement for Bay, Bakool, Gedo and Hiiraan, the riverine areas and the 4 named IDP settlements; the IPC's post-Deyr analysis for Somalia.
- October 2026 to March 2027
Southern Africa planting and the lean season: IPC and FEWS NET updates for Zambia (1.25 million projected), Malawi, Mozambique, Zimbabwe and Madagascar, and OCHA ROSEA's preparedness alerts.
Updated by OCHA ROSEA and FEWS NET - Continuous
GDACS alerts, dekadal rainfall anomalies, ASAP hotspot status and FTS plan coverage for each of the 13 countries on PRISM, alongside the IPC archive that holds the Zambia projection cited here.
Updated by PRISM climate dashboard and the climate and disasters catalogue