Purpose Global fluctuations in freight rates during crises – such as the COVID-19 pandemic and the Red Sea shipping disruptions – introduce significant uncertainty for logistics planners, shippers and policymakers. This study aims to develop an adaptive forecasting model specifically designed to support short-term decision-making in maritime logistics and supply chain management. The objective is to create a model that remains reliable in both stable and highly volatile market conditions while producing accurate predictions of the Shanghai Containerized Freight Index (SCFI). Design/methodology/approach A Modified Multivariable Prophet model is proposed to forecast the SCFI. The model incorporates three methodological advances: adaptive trend boundaries that recalibrate growth components with recent market data to prevent unrealistic projections, a sliding window retraining strategy to capture evolving patterns and structural breaks and integration of exogenous variables – such as bunker fuel prices, vessel transit volumes, schedule reliability and vessel delays – selected through Boruta and Maximal Information Coefficient methods. The model is empirically evaluated using data from 2019 to 2024. Findings The proposed model demonstrates strong predictive accuracy, achieving Root Mean Squared Error = 147.5, Mean Absolute Error = 93.6, Mean Absolute Percentage Error (MAPE) = 4.48% and Mean Squared Error = 21,765. It outperforms benchmark models, including Multivariate Long Short-Term Memory (MAPE = 7.25%) and Seasonal Autoregressive Integrated Moving Average (MAPE = 8.8%), by effectively capturing both volatility and seasonality while maintaining resilience against abrupt market disruptions, such as those triggered by COVID-19 and the Red Sea shipping crisis. Originality/value This study extends Prophet modeling by introducing adaptive mechanisms and robust feature integration tailored for volatile maritime markets. The proposed framework offers a practical and resilient forecasting tool for industry stakeholders, enabling more informed strategic planning and operational efficiency in the face of global shipping uncertainties.
Hartono et al. (Fri,) studied this question.