Past trends, future insights of malaria incidence in threeselected states of Nigeria: ARIMA and Wavelet Approaches

Abstract

<title>Abstract</title> <bold>Background:</bold> Malaria, transmitted by infected female Anopheles mosquitoes, remains endemic inNigeria despite sustained control efforts. Transmission patterns vary considerably across Nigeria’s36 states, posing challenges for timely and effective intervention. State-specific analysis of spatialand temporal dynamics is therefore essential to strengthen early warning systems and guide targetedcontrol strategies. This study aims to forecast monthly malaria incidence for two years and examinetemporal dynamics of malaria transmission in three ecologically diverse Nigerian states, focusing onannual and seasonal cycles, synchrony, and lead–lag relationships between states from 2015 to 2024. <bold>Methods:</bold> Monthly uncomplicated malaria case data (2015–2025) were obtained for Oyo, Kebbi,and Plateau States. Seasonal autoregressive integrated moving average (SARIMA) models were usedto forecast malaria incidence, while wavelet analysis was applied to characterise underlying seasonaland multi-annual cycles. Bivariate wavelet analysis was further employed to explore synchrony andlead–lag relationships between pairs of states. <bold>Results:</bold> SARIMA forecasts identified distinct seasonal peaks: June–August in Oyo, August–Octoberin Kebbi, and October in Plateau. Univariate wavelet analysis revealed consistently strong annualseasonality in Oyo, fluctuating seasonal strength in Kebbi, and largely stable but weakening annualcycles in Plateau toward the end of the study period. Bivariate wavelet analysis showed thatOyo consistently led Kebbi, Oyo also led Plateau, while Plateau led Kebbi, indicating clear temporaloffsets in peak transmission across states. Synchrony analysis of malaria incidence across Oyo,Plateau, and Kebbi shows that transmission is more strongly aligned with shared climatic and ecologicalfactors than with geographic distance, with Plateau–Kebbi exhibiting the highest synchrony(r = 0.57 overall, r = 0.74 annual cycle). <bold>Conclusion:</bold> Forecasting results highlight the importance of state-specific intervention timing, withinterventions most effective if targeted in June–August for Oyo, August–October for Kebbi, and Octoberfor Plateau. The persistence of strong annual and multi-annual cycles underscores predictabledynamics that can support early warning systems and the need for adaptive strategies. The observedlead–lag relationships suggest that shared seasonal drivers of transmission operate on slightlydifferent schedules, underscoring the need for strengthened surveillance and tailored, state-specificmalaria control strategies.