← ENSO country deep dives

ENSO country deep dive

Zimbabwe

How bad is the 2026/27 season likely to be, how confident are we, and how much of it is El Niño?  ·  ERA5 0.25° 1981–2026  ·  Niño3.4 (NOAA PSL)  ·  578 grid cells, 578 with a DJF season

1. How bad is the coming season?

The figure reads the seas5-skill app's per-pixel skill cube for the 1 September 2026 issuance, summarised as the national median pixel. Top: the rainy season, November to March, unimodal and the same everywhere. Middle: how much skill SEAS5 has for each window from September — moderate for NDJ (0.48) and DJF (0.42), the two windows that carry the season, low for OND and JFM (0.27). Bottom: what the issuance forecasts — a dry season throughout, at return periods of 5 years for SON, 23 for OND and NDJ, 19 for DJF and 15 for JFM. Under the app's rule (return period ≥ 3 years at moderate skill or better) NDJ and DJF would raise an alert.

What the Sep 2026 issuance forecasts. Windows at or beyond the app's 3-year return-period threshold (in-season windows excluded): SON dry 5 yr (low skill), OND dry 23 yr (low skill), NDJ dry 23 yr (moderate skill), DJF dry 19 yr (moderate skill), JFM dry 15 yr (low skill). Under the app's rule — an alert needs the return period and at least moderate skill — this issuance would raise an alert for NDJ, DJF.

Each cell: median pixel r, its bin, the share of the zone's cells at moderate-or-better skill, and the median return period of the Sep 2026 forecast anomaly (dry = forecast below its hindcast median). Skill source: skill_stats_grid_detrended.nc (seas5-skill, DEV blob), the same cube behind the app's pixel skill map. Median of pixel correlations, not the correlation of the area mean, so it is a conservative summary for a coherent area.

Where food insecurity already sits (FEWS NET)

FEWS NET's August 2026 outlook update, at FEWS NET's own livelihood-zone × district units and not aggregated here. The June current situation and the August–September projection are Phase 1 and 2 throughout after the 2025/26 harvest. The medium-term projection for October 2026 to January 2027 moves 57 of 203 units to Phase 3 (Crisis) across Matabeleland North and South, Masvingo, southern Manicaland and the Zambezi valley — the lean season and the first half of the forecast rains. The last panel overlays the FEWS NET unit outlines on the El Niño drought hit-rate from the next section: the units projected into Crisis are also where El Niño seasons land in the driest third most often.

2. How confident are we? What El Niño seasons have done before

Fourteen DJF seasons since 1981 had El Niño conditions (Niño3.4 ≥ +0.5). Ten fell in Zimbabwe's driest third and seven in the driest fifth; none reached the wettest third. La Niña seasons ran the other way: one of twenty in the driest third, eleven in the wettest. The El Niño list is the country's disaster calendar — 1991/92 (driest in the record), 2015/16 (second), 1982/83, 1986/87, 1994/95, 2002/03, 2004/05, 2006/07, 2018/19 and 2023/24, when the season collapsed in February and a state of disaster followed in April. The misses are 1997/98 (near normal despite the strongest El Niño of the century; a positive Indian Ocean Dipole overrode the Pacific), 1987/88, 2009/10 and 2014/15.

For the current season the relevant comparison is the strong events. Of the six El Niños that reached +1.5 °C in DJF — 1982/83, 1991/92, 1997/98, 2009/10, 2015/16, 2023/24 — four produced a driest-fifth season; two (1997/98, 2009/10) were near normal. That is the historical basis for the confidence: El Niño roughly doubles the odds of a driest-third season and all but rules out a wet one, but a strong El Niño is not a guarantee, and a miss rate of one in four is why an ENSO-only trigger is not enough on its own.

ENSO phase (concurrent)	Seasons	Mean anomaly (SD)	In driest third	In driest fifth	In wettest third
El Niño	14	−0.81	71%	50%	0%
Neutral	11	−0.06	36%	18%	36%
La Niña	20	+0.60	5%	0%	55%

Every El Niño DJF season since 1981

Year	Niño3.4 (DJF)	Rainfall anomaly (SD)	Rank (1 = driest)	Outcome
1982	+2.16	−1.29	4 of 45	driest fifth
1986	+1.13	−1.21	5 of 45	driest fifth
1987	+0.71	+0.35	29 of 45	near normal
1991	+1.77	−2.06	1 of 45	driest fifth
1994	+1.00	−1.06	7 of 45	driest fifth
1997	+2.23	−0.62	16 of 45	near normal
2002	+0.87	−1.04	8 of 45	driest fifth
2004	+0.59	−0.76	13 of 45	driest third
2006	+0.66	−0.69	15 of 45	driest third
2009	+1.50	−0.18	21 of 45	near normal
2014	+0.55	+0.34	28 of 45	near normal
2015	+2.50	−1.36	2 of 45	driest fifth
2018	+0.75	−0.78	11 of 45	driest third
2023	+1.79	−1.00	9 of 45	driest fifth

The 15 driest-third DJF seasons: 1981 (Neutral), 1982 (El Niño), 1983 (La Niña), 1986 (El Niño), 1990 (Neutral), 1991 (El Niño), 1993 (Neutral), 1994 (El Niño), 2001 (Neutral), 2002 (El Niño), 2004 (El Niño), 2006 (El Niño), 2015 (El Niño), 2018 (El Niño), 2023 (El Niño). Phase split: El Niño 10, Neutral 4, La Niña 1.

Season by season: rainfall, El Niño, CERF and FEWS NET

Every season since 1981 on one line: how dry it was, whether it was an El Niño season, whether CERF funded a drought response for it, and what FEWS NET was projecting for October–January before that season began — the same product as the projection we have now, so the rows are comparable with the top one. Phase 4 (Emergency) has appeared once in FEWS NET's Zimbabwe record — the October 2016 outlook projected twelve southern units into Emergency for February–April 2017, the lean season after the 2015/16 drought — and never in a current-situation classification; Phase 5 never. The El Niño seasons that became disasters started from very different places: 2015/16 from 26% of units projected in Crisis, 2023/24 from 17%; the seasons after them (2016/17, 2024/25) began from 60% and 100%. 2026/27 begins from 28%.

Rainfall: ERA5 area mean over the rain-fed cells, standardised over 1981–2026. ENSO phase: concurrent Niño3.4 (≥ +0.5 El Niño, ≤ −0.5 La Niña; pinned NOAA series). CERF: Rapid Response / Underfunded Emergencies applications with emergency type “Drought” from the team's OneGMS mirror, attributed to the rainy season named in the CERF drought-period supplement — when that period spans two failed seasons the allocation sits under the later one and says so — or, failing that, to the season whose harvest year the allocation fell in (“by date”); “+N mo” is the approval month counted from 1 December of that season. FEWS NET: the outlook for October–January issued between June and September of that year — the same product as the current medium-term projection, so each row shows the pre-season picture that season started from — as shares of FEWS NET's classified units in Phase 3, 4 and 5 (unit counts, since older unit vintages carry no geometry; FEWS NET classifies areas and publishes no population by phase). “Issued in October” marks a year with no June–September outlook. Cells are shaded darker for worse outcomes; most recent season first, the forecast season on top.

3. How much of this is El Niño?

Latest Niño3.4 in NOAA PSL's current series: +1.89 °C for August 2026 (three-month mean +1.73), i.e. El Niño conditions by the ±0.5 threshold used on this page. The historical analysis below uses the survey's pinned Niño3.4 series (NOAA PSL, ERSST v5 basis), which runs about 0.2 °C cooler than the current ERSST v6 series.

Two things establish how much of the forecast is El Niño: the current state of the Pacific, and how tightly Zimbabwe's rainfall has followed it. Everything below is computed from the ERA5 monthly grid and the NOAA Niño3.4 index used by the global survey, restricted to cells inside the country.

Seasonal cycle

One rainy season, November to March, with December–February carrying about 55% of the annual total. The three climatology bands in the chart and the locator map are there only to show that the season's timing is the same across the country, which is why the rest of this page treats Zimbabwe as one unit.

Pixel-level correlation with Niño3.4

The DJF correlation with Niño3.4 is −0.68 nationally, the strongest in the survey; every cell is past −0.3 and 84% past −0.5. An r of −0.68 means ENSO accounts for about 46% of the year-to-year variance in DJF rainfall. The signal is already present in OND (−0.5), peaks in DJF and holds through JFM, with the same sign everywhere: there is no part of the country and no part of the season that responds the other way. This is the textbook southern-Africa teleconnection, quantified for Zimbabwe since Cane, Eshel and Buckland (1994) showed Pacific sea-surface temperature explained more than 60% of the variance in the national maize yield.

Season	Cells analysed	Share of annual rain	Median r	Range	Significant (p<0.05)	|r| ≥ 0.30	|r| ≥ 0.50
OND	508 / 578	30%	−0.41	−0.61 to −0.20	95% negative	94%	14%
NDJ	578 / 578	47%	−0.48	−0.68 to −0.30	100% negative	100%	38%
DJF	578 / 578	55%	−0.58	−0.73 to −0.39	100% negative	100%	84%
JFM	578 / 578	47%	−0.46	−0.59 to −0.28	100% negative	99%	26%

The same, by province

From the team's per-province ERA5 raster stats: concurrent DJF correlations run from −0.71 in Mashonaland East to −0.52 in Matabeleland South, and in every province at least eight of the fourteen El Niño seasons were in the province's own driest third against fewer than one in six La Niña seasons. The coupling is a little weaker in the far south-west; otherwise the country moves together.

Province	ERA5 pixels	r (concurrent)	El Niño mean (SD)	El Niño seasons in driest third	La Niña in driest third	El Niño driest-third seasons
Mashonaland East	1096	−0.71	−0.82	64% of 14	5%	1982, 1986, 1991, 1994, 1997, 2002, 2015, 2018, 2023
Manicaland	1231	−0.69	−0.79	64% of 14	10%	1982, 1986, 1991, 1994, 1997, 2002, 2004, 2015, 2018
Harare	33	−0.68	−0.78	71% of 14	5%	1982, 1986, 1991, 1994, 1997, 2002, 2004, 2015, 2018, 2023
Mashonaland West	1958	−0.65	−0.78	64% of 14	5%	1982, 1986, 1991, 1994, 2002, 2004, 2015, 2018, 2023
Midlands	1700	−0.64	−0.75	71% of 14	5%	1982, 1986, 1991, 1994, 1997, 2002, 2006, 2015, 2018, 2023
Matabeleland North	2586	−0.62	−0.78	71% of 14	5%	1982, 1986, 1991, 1994, 2002, 2004, 2006, 2015, 2018, 2023
Mashonaland Central	958	−0.61	−0.66	64% of 14	15%	1982, 1986, 1991, 1994, 1997, 2002, 2015, 2018, 2023
Masvingo	1960	−0.59	−0.72	57% of 14	15%	1982, 1986, 1991, 2002, 2004, 2009, 2015, 2018
Bulawayo	17	−0.59	−0.69	71% of 14	10%	1982, 1986, 1991, 1997, 2002, 2004, 2006, 2015, 2018, 2023
Matabeleland South	1873	−0.52	−0.61	64% of 14	10%	1982, 1986, 1991, 2002, 2004, 2006, 2009, 2015, 2023

Caveats

• The September 2026 SEAS5 issuance sits in the dry tail of its hindcast far more often than a calibrated forecast should, across the whole globe, not only here; a strong El Niño can produce that, but so can a normalisation problem, and the check has not been completed. Read ceiling return periods as “beyond the record”.
• Fourteen El Niño seasons, six of them strong, is a small sample; the hit rates carry wide uncertainty.
• ERA5 is a reanalysis and Zimbabwe's gauge network has thinned since 2000, so recent years lean more on the model background.
• FEWS NET classifies areas, not people, and publishes no population-in-phase figures; its analysis is IPC-compatible but independent of the IPC/CH consensus.

References

• Cane, M. A., Eshel, G. & Buckland, R. W. (1994). Forecasting Zimbabwean maize yield using eastern equatorial Pacific sea surface temperature. Nature 370, 204–205. link
• Unganai, L. S. & Mason, S. J. (2002). Long-range predictability of Zimbabwe summer rainfall. International Journal of Climatology 22, 1091–1103. link
• Manatsa, D., Chingombe, W. & Matarira, C. H. (2008). The impact of the positive Indian Ocean dipole on Zimbabwe droughts. International Journal of Climatology 28, 2011–2029. link
• Mason, S. J. & Goddard, L. (2001). Probabilistic precipitation anomalies associated with ENSO. Bulletin of the American Meteorological Society 82, 619–638. link
• UNICEF Zimbabwe (2024). 2024 El Niño induced drought disaster: domestic and international appeal for assistance (state of disaster declared 3 April 2024). link; OCHA (2023). Zimbabwe: El Niño Anticipatory Action Plan, Oct 2023 – Mar 2024. link
• FEWS NET Data Warehouse, Zimbabwe IPC-compatible classification, via ds-fewsnet-mirror. Source: FEWS NET. SEAS5 skill and forecast return periods from the seas5-skill app's detrended skill cube. Niño3.4 from NOAA PSL.

Generated by enso_deep_dive.py from deep_dives/zwe.toml. Method and data as in the global survey.