Patterns and probabilities of dry spells and rainfall for improved rain-fed farming in Northwestern Ethiopia

Abstract

Dry spells and rainfall variability significantly impact rain-fed agriculture in Ethiopia, necessitating targeted adaptation strategies. This study used rainfall data from Ethiopia's Meteorological Institute (1992–2021) to evaluate the likelihood of rainfall and dry spells lasting more than 7, 10, 15, and 20 days in northwestern Ethiopia. Markov chain modeling assessed the probabilities of these events, while Modified Mann–Kendall tests examined trends in dry days at a significance level of P < 0.05. The area has a Kiremt-dominated monomodal rain, with 80–100% rainfall probability following dry days in July–August. June (40–67%) and September (37–60%) experienced high dry days with moderate to high variability (CV: 26–45%), underscoring the need for monitoring dry day risks during planting and crop maturation. In contrast, July (10–19%) and August (10–29%) had lower frequencies of dry days. The Belg (77–89%) and Bega (79–95%) seasons showed high dry day frequencies with low variability, indicating that rain-fed agriculture is impractical during these times. Trend analysis revealed significant increases in dry days at Ebenat and decreases at Simada, with most stations showing rising dry day patterns. Dry spell analysis found minimal (0–10%) likelihood during Kiremt, but a 100% risk from late September to February. Spatial analysis revealed a high risk of extended dry spells: 20–40% in August, 40–60% in July, 100% in October, and 75–100% in September-June. May has a 60–80% chance of a 7-day dry spell, likely delaying the growing season to June and reducing yields for rain-fed crops, suggesting targeted interventions, like using drought-tolerant crops or mitigation strategies.