Multimodel forecasting of non-renewable resources production

The article addresses the complexities of modelling and forecasting of non-renewable resources production (oil, gas, coal, etc.), by means of combining five production trend models with “custom” asymmetry, as well as with six models of fluctuation components: harmonic, independent from the trend; harmonic, proportional to the trend; simultaneous presence of the first and second models of fluctuation components; harmonic with “weighted amplitude”; “frequency-weighted” harmonics. The purpose of this research is to increase the production forecasting accuracy, by considering the fluctuation components models and by monitoring the models' evolution and fluctuation. The offered methods provide a production forecasting accuracy increase for oil in the U.S. - by 3.2%, for coal in Germany - by 5%, and gas in the Volgograd region (Russia) - by 25%. •We review well-known life cycle models for non-renewable resources production.•We propose combining these models with fluctuation component.•Piecewise models with fluctuations allow to decrease forecasting errors.•We find shares of the best trend and fluctuation models for gas production.•We study production of oil in the US, coal in Germany and gas in Volgograd Region.