Assessment of Precipitation and Temperature Shifts in Hiiraan Region of Somalia Using a Var Model

The study analyzed the long-term variability of precipitation and temperature in the Hiiraan Region, Somalia, spanning 1981 to 2024, using a Vector Autoregression (VAR) model. Diagnostic indices indicated a statistically significant shift toward a more volatile climate regime. A dominant and increasing trend of strongly irregular rainfall was observed in the Precipitation Concentration Index (PCI), with 16 years of the 1981–2000 period and 23 years of the 2001–2024 period falling into the strongly irregular category (PCI > 20). The Seasonality Index (SI) indicated a shift from a moderately seasonal to markedly seasonal and extreme rainfall regime, where the frequency of extreme seasonality (SI ≥ 1.20) has doubled in the latter period. The analyses of Standardized Anomalies (SAI) confirmed the persistent warming across seasons, more marked for minimum temperatures. A stationary VAR (3) model was able to successfully capture the dynamic feedback between rainfall, maximum temperature, and minimum temperature. Through the Impulse Response Function analysis, it revealed an asymmetric relationship between them, maximum temperature shocks depressed rainfall in the short term, while rainfall shocks delivered only transient cooling. But, more importantly, evidence of a self-reinforcing warming loop was critical. The model showed high forecasting accuracy, besides passing all diagnostic checks, and thus was statistically robust. The findings underpin a transition to a climate of intense rainfall pulses and persistent warmth that requires strong adaptation strategies centered on water harvesting, flood preparedness, and climate resilient livelihoods.