A multi-indicator framework for assessing the likelihood of ENSO impacts across southern Africa

Tamuka Magadzire , Surekha Ramessur, Sunshine Mduduzi Gamedze, Nicholas Christopher Mbangiwa, Kgakgamatso Mphale, Gregory Husak, Laura Harrison, Chris Funk

The El Niño-Southern Oscillation (ENSO) is the dominant source of seasonal predictability for Southern Africa, yet operational guidance rarely conveys which agriculturally relevant variables it can skillfully predict, or where ENSO impacts are most likely. We present an operational analysis framework that converts an ENSO state, classified by phase and strength, into gridded, analogue-composite outlooks for 14 indicators spanning rainfall, evaporative demand and related standardized indices (SPI and SPEI), crop water balance, dry-spells, rainfall-season timing, and heat extremes. El Niño typically produces the adverse outcome across the central and southern interior, with Moderate-to-Strong events having more robust impacts. We assess skill with cross-validated ranked probability skill scores at the grid-cell scale and aggregated over the response domain, and phase-specific verification scores. Predictability is substantial but uneven: temperature and potential evapotranspiration are the most predictable. Rainfall is skillful across much of the core though spatially variable. Crop water balance, season length and a smoothed dry-spell metric are skillful as regional indices, while the onset and cessation dates carry little skill. Skill peaks at the height of the rains, is weaker for La Niña than El Niño, and increases with event strength. A single ENSO-based outlook can be misleading. Outlook guidance, including confidence should be set by location, season, indicator, phase and event strength, with strong or very strong El Niño events being most impactful. This is relevant to the very strong El Niño forecast for 2026/27. The framework is delivered openly through the Southern Africa ENSO Explorer (https://enso.ubramplab.org).