Developing an Original Probabilistic Heatwave Impact Model for Public Health Emergency Planning in Northwest Nigerian Cities

Heatwave mortality and morbidity risk in Northwest Nigeria's major cities, including Kano, Sokoto, and Kaduna, represent an expanding public health challenge under observed and projected warming trends, yet probabilistic modelling frameworks linking meteorological heatwave characteristics to health outcomes are absent from Nigerian scientific literature. This study develops an original probabilistic heatwave impact model for public health emergency planning across Northwest Nigerian cities, constituting both a methodological and applied scientific contribution. The modelling framework integrates extreme value statistical modelling of heatwave intensity and duration, epidemiological exposure-response functions estimated from hospital admission records in Aminu Kano Teaching Hospital and Usmanu Danfodiyo University Teaching Hospital, and probabilistic seasonal heatwave forecasts derived from NiMet and ECMWF extended-range ensemble products. A Bayesian hierarchical exposure-response model is developed to estimate heatwave mortality risk as a function of maximum temperature, duration, nighttime minimum temperature, relative humidity, and timing within the calendar year. Original methodological contributions include adaptation of the heat vulnerability index framework to the socioeconomic and built environment context of Northwest Nigerian cities, incorporating variables for housing density, access to cooling infrastructure, occupation, and age structure from the 2023 population census. Retrospective model validation against documented heatwave-associated excess mortality events in 2022 demonstrates that the model correctly estimates excess mortality within plus or minus 18 percent at 90 percent confidence intervals. Recommendations include embedding the model within Kano State's emergency response trigger system. Keywords: heatwave, probabilistic model, public health, Northwest Nigeria, Bayesian.