Abstract
<title>Abstract</title> <p>The livestock sector is the cornerstone of Somalia’s economy and food security, contributing approximately 80% of total export earnings. However, the sector is increasingly vulnerable to climatic variability, structural shocks, and socio-political instability, creating a critical "forecasting gap" for national planning. This study develops a comprehensive predictive framework for Somalia’s meat production (1961–2024) by conducting a large-scale comparative analysis of 19 distinct time-series architectures. The evaluated suite includes seven single models (ARIMA, ETS, BATS, TBATS, Theta, ARFIMA, and NNAR) and twelve hybrid ensemble configurations designed to integrate linear stochastic processes with non-linear learning capabilities. Methodologically, the data were partitioned into a training set (1961–2012) and a validation set (2013–2024). Model performance was rigorously assessed using Mean Absolute Percentage Error (MAPE), Mean Absolute Scaled Error (MASE), and Theil’s statistics. Empirical results demonstrate that the ARFIMA model was the most precise single forecasting tool (MAPE: 2.17%, Theil’s U: 0.64), effectively capturing the long-memory persistence inherent in production cycles. Among the hybrid configurations, the ARIMA–NNAR ensemble emerged as the superior ensemble (MAPE: 8.10%), successfully bridging the gap between linear trend extraction and non-linear adaptability. Long-term projections for the 2025–2036 horizon suggest that meat production is likely to stabilize or marginally decline, ranging between 171,794 and 182,522units—remaining well below the 2005 historical peak. These findings indicate that the traditional pastoral system may have reached an environmental capacity ceiling, necessitating a strategic shift toward shock-responsive, data-driven livestock management. This study provides a replicable blueprint for agricultural forecasting in climate-vulnerable and data-constrained regions, offering essential evidence for formulating sustainable food security policies in the Horn of Africa.</p>