Comparative Production Modeling and Forecasting of Crude Oil and Natural Gas in Nigeria

Abstract Nigeria's petroleum sector continues to face challenges related to production variability, reservoir maturity, and underinvestment, with important implications for export revenue, domestic energy supply, and national energy planning. Accurate forecasting of hydrocarbon output is therefore essential for effective field management, reserves assessment, and macro-level energy strategy. This study presents a comparative production modeling and forecasting framework for Nigeria's national crude oil and natural gas output using monthly production data obtained from the Nigerian Upstream Petroleum Regulatory Commission. The dataset covered January 2021 to November 2025 for crude oil and January 2021 to December 2025 for natural gas. Classical decline curve analysis based on Arps’ exponential, hyperbolic, and harmonic formulations was applied alongside statistical time-series modeling to evaluate suitability for aggregated national-level forecasting. Model performance was assessed using out-of-sample testing with Root Mean Square Error (RMSE) and symmetric Mean Absolute Percentage Error (sMAPE). The results showed that for crude oil forecasting, Exponential Smoothing with additive trend (ETS) significantly outperformed all decline curve models, achieving a test RMSE of 4.73×10⁶ and sMAPE of 4.40%, compared with sMAPE values between 12.78% and 14.99% for Arps-based models. For natural gas, ETS also demonstrated superior predictive stability, with a test RMSE of 1.83×10⁴ and sMAPE of 7.26% under structurally volatile conditions. Oil production forecasts indicated a moderate upward trend consistent with recovery-driven growth rather than depletion-dominated decline, while gas production projections suggested a mild downward trend over the forecast horizon. To evaluate structural hydrocarbon balance, a Production Transition Index (PTI), defined as Gas/(Gas + Oil), was introduced. PTI values remained within a narrow range of 0.0023–0.0034 throughout the study period, indicating no evidence of a macro-level transition toward gas-dominant production. Reduced volatility in monthly transition rates after 2023 suggested increasing structural stabilization of the national production system. Overall, the findings demonstrate that classical reservoir-level decline curve models are unsuitable for aggregated national production forecasting without structural modification, while statistical time-series methods provide superior accuracy and robustness. The study contributes a validated comparative forecasting framework and a quantitative transition metric to support national energy planning, production outlook assessment, and long-term hydrocarbon system analysis.