Three-state Hidden Markov Model of Air Temperature Regimes over the Middle Belt States of Nigeria

This research applies a three-state Hidden Markov Model (HMM) to thirty (30) years (1991-2020) daily air temperature data over the Middle Belt states (Abuja, Benue, Kogi, Kwara, Nasarawa, Niger and Plateau) of Nigeria to identify the unique temperature regimes, characterize their distributional properties, and examine their temporal dynamics. Based on the state-dependent mean temperatures and standard deviation, the model divided the daily air temperature data series into three statistically different temperature regimes: regime 1 (Cool), regime 2 (Moderate), and regime 3 (Warm). Results show that the Warm (above average) regime dominates the majority of the time period, while the moderate (in-between) regime typically represents changes between transitional periods and the cool (below average) regime is associated primarily with the Harmattan (Northeast Trade Wind) Season. Transitions from Regime 1 (Cool) to Regime 2 (Moderate) and from Regime 2 to Regime 3 (Warm) happen more often than direct transitions from Regime 1 to Regime 3 as revealed by the Transition Probability Matrix. Warm regime occupancy reaches 54.4% in Benue and 45.7% in Kogi, indicating that high air temperature conditions dominate mainly in such area. The overall structure of the HMM used in this work provides a physical description of the temperature variability experienced by the Middle Belt States, Nigeria. Hence, creating a useful opportunities for assessments of climate variability, supporting heat-risk mapping and adaptation planning.