Dry Corridor rainfall outlook for 2027: Primera and Postrera

2026 was an El Niño developing year, and it behaved like one. The framework activated in all three countries in March 2026 on the SEAS5 forecast for May–August. Primera 2026 then came in at 48–71% of typical in CHIRPS (51–71% in ERA5) across the three countries and their framework departments, after a Primera 2025 that was below typical in most areas (77–101%). Early Postrera rain is less clear: September 2026 came in at 86–95% of typical in CHIRPS but 64–95% in ERA5, and the two disagree most in El Salvador (95% against 64%). ERA5 has September–October at 71–97%, lowest in El Salvador. FEWS NET expects food insecurity in Guatemala and Honduras to peak between April and August 2027, so the 2027 Primera harvest is the recovery point.

The survey grades the Dry Corridor's El Niño signal as robust: El Niño dries the rainy season while it is under way, most in July–October (Dai and Wigley 2000; our ERA5 correlation for August–October is about −0.6 in all three countries). In the July–August canícula the link is as strong (r ≈ −0.6). In May–June it is weak or absent. The year after a strong event is different: El Niño fades, and the tropical North Atlantic, which warms in the months after an El Niño peak, adds rain. What matters for Primera is the contrast between the two oceans.

Where El Niño is heading. Niño3.4 reached +2.2 °C in October 2026 (1991–2020 base). CPC (8 October) expects a strong to very strong event through January–March 2027. The IRI plume of October still gives El Niño a 60% chance in May–July and 37% in June–August. The NMME runs of November average about +2 °C in March–May, +1 °C in May–July and about 0 by September–October, with a spread from La Niña to El Niño. After the seven strongest past winters in their record, the same models were too warm in May–July five times, by about 0.4 °C on average, and over all years their May–July forecasts have overshot. We therefore scale them back to how their past forecasts verified: May–July then sits at +0.68 °C and September–October at +0.02 °C.

The tropical Atlantic was +0.64 °C in September (PSL) and the models warm it to about +1.0 °C by May–July 2027, near the record warmth of 2010, 2023 and 2024. Among those three, 2010 brought a wet Primera as La Niña set in, 2024 a mixed one (84–122% of typical across the three countries and framework departments) as El Niño faded, and 2023 a dry one as El Niño built.

Onset is the first day from 15 April with at least 1 mm that starts three days totalling 20 mm or more, as long as no dry spell of ten days or more (days under 1 mm) follows in the 30 days after those three. It is computed on each 5 km CHIRPS cell for 1981–2025 and then averaged by area. As a check, a simpler rule (the first five days from 15 April totalling 25 mm, with no dry-spell test) gives the same three clearly late years and a smaller average delay. The end of the rains is the last day between 1 September and 28 November that starts three days totalling 20 mm or more. The maps show the median shift over the six seasons after strong El Niño winters.

• A later start is the clearest timing signal. Onset came later than typical in 53 of 54 departments (median of the six seasons, by 6 to 21 days). Typical onset is about 1 May in El Salvador, 7 May in Guatemala and 15 May in Honduras. In 1983, 1992 and 1998 the rains started one to three weeks late across most of the region, under both onset rules. These are the three oldest of the six cases. 2024 was late across most of the region under the main rule but not the simpler one, 2016 was late in parts, and 2010 started on time (early under the simpler rule), which leaves the average delay at 8–14 days by country and framework area under the main rule (5–12 days under the simpler one). The share of the area with a late onset averaged 47–74% in those years, against roughly a quarter to a third in other years.
• How far it reaches. Seasons with El Niño under way were also late, by 4–9 days on average (five cases). Over all 45 years onset is later after a warmer December–February Pacific (correlations of 0.21–0.47), but most of that link comes from the six years above; without them it is weak. The NMME runs put March–May 2027 at about +2 °C, or about +1.5 °C after allowing for their warm bias after strong winters: warmer than in any of the six past cases (+0.4 to +1.0 °C), so 2027 is beyond the range these cases cover.
• The canícula (July–August) has not been the weak spot after strong events. July–August rain was at or above typical in three to six of the six years, depending on the area. Dry-spell counts from CHIRPS cannot be used here: it records rain on about 24–42 of the 62 days of July–August in a typical year, too many for spell lengths to mean much, so the canícula is judged on totals.
• No sign of an early end to Postrera. The last heavy rains came at about the usual time or later (median shifts of −3 to +15 days by department), even though October–November totals were below typical in four of six years in most areas.

The team's EM-DAT snapshot (records that start up to 2023) lists drought in Guatemala, Honduras or El Salvador starting in 14 different years since 1981. Nine of those years had El Niño conditions during the June–November rains (Niño3.4 of +0.5 °C or more in June–August or September–November): 1987, 1994, 1997, 2002, 2004, 2009, 2015, 2018 and 2023. Four had none: 2000, 2001, 2012 and 2014, the last just before the 2015 event, with a warm Pacific late in the year (+0.4 °C). One, El Salvador from May 1998, came the year after a strong El Niño, and fits the late, dry start of Primera described above. The 2014–2016 records span the 2015 event and the year after it.

Three of the framework's four reference drought years (2009, 2015 and 2018) were El Niño years, and the fourth (2014) led into one; 2026 fits too. So the conditions that the framework, and most past responses, were built around are those of a developing El Niño. 2027 is a different case: a year after, with less evidence and more uncertainty.

What would change this outlook: an El Niño that holds on into June–August, a cooler Atlantic than forecast, or SEAS5 runs from February to May that move Primera toward the dry end of the typical range.