Experience from 10 Years of Emission Forecasting in Norway

Abstract

Abstract Statoil has forecasted CO2 emissions from their offshore fields for many years. This paper sums up 8-10 years history of forecasting the emissions from 6 offshore oil and gas fields with 8 producing facilities. The accuracies of the early forecasts for these facilities have been analyzed with focus on the forecasts made in 2001 and 2004. The emission forecasts are made for each producing facility using a field specific forecasting model. These models are calibrated annually by comparing calculated fuel demand with reported consumption. The analysis shows that Proper calibration of the forecasting model over several years improves the model significantly and is necessary in order to keep control of the uncertainties in the forecasts.Through annual calibration of the forecasting model over several years, accurate fuel and emission forecasts within an annual uncertainty level of +/- 4 % is achievable for a short term forecast, provided good production, injection and activity input forecasts. A short term forecast in this context is the next two to five years.Historical data show that the input data (production-, injection- and activity profiles) represent an uncertainty element of the same order and in some cases larger than the uncertainty in well calibrated forecasting models.Future modifications to cope with new and adjusted oil and gas production conditions will add to the uncertainty, in particular for the long term forecast.Operational flexibility will add uncertainty to the forecasts (example gas turbine operation).The uncertainties increases for long term forecasts (< 5 years ahead), due to increased uncertainties in the input profiles as well as in the model itself due to future changes in the production facilities.