Climate change detection and attribution: Bayesian estimation of abrupt change, seasonality and trend model, and Mann–Kendall trend test approaches

ABSTRACT Climate change is projected to have adverse impacts on environmental sustainability. This research combines statistical analysis and Bayesian modeling for climate change detection and attribution in Kaduna, Northern Nigeria. The study combines the Bayesian estimation of abrupt change, seasonality, and trend model (BEAST) with the Mann–Kendall (M–K) trend test for detection and correlation analysis with optimum fingerprinting for attribution. The study used 122 years of climate data (1901–2022), focusing on average annual rainfall and average annual surface temperature for climate change detection, alongside a 30-year analysis of greenhouse gas (GHG) emissions (1990–2020) for climate change attribution. The result of the M–K test reveals a significant increasing trend in temperature (approximately 0.004 °C/year) and a decreasing trend in rainfall (approximately 0.756 mm/year), indicating a warming climate and potential drought conditions. The Bayesian approach further identified multiple changepoints in temperature data, highlighting years of significant climatic shifts. Correlation analysis demonstrated a weak positive relationship between temperature increases and GHG emissions with a correlation coefficient of 0.27. Optimum fingerprinting results show a non-statistically significant relationship between the variables with an R2 value of 0.071, indicating that only 7.1% of the variability in temperature can be explained by the model.