Past years' forecasting performance has demonstrated that single ''most likely'' ͑deterministic͒ prediction of future production, without quantifying the associated uncertainties, is inadequate for management and planning purposes. Based on this an effort was initiated to improve the forecasting methodology and procedures with special emphasis on quantifying the uncertainty in short-term production forecasting ͑STPF͒.The Valhall field, offshore Norway, presented special problems for production forecasting because many unusual events affect production, even for STPF. The soft compacting reservoir chalk causes well collapses and chalk influxes and makes drilling difficult. Modeling of reservoir fracturing and therefore well performance is also uncertain.A probabilistic forecasting approach was adopted using a customized spreadsheet and commercially available statistical analysis add ins which allowed deterministic forecasting, partially probabilistic analysis, or fully integrated uncertainty analysis. Uncertainties were characterized by distributions based on historical data where possible.Communication and integration of knowledge were also key success factors since the process required input from several departments and many individuals to ensure that the ''company knowledge'' was fully reflected. A ''common language'' to communicate uncertainty and an auditable process to ensure buy in and consistency were also critical.

Accurate production forecasting is a critical factor in an oil company's management and decision making. The production forecast forms the basis for long term planning and decision making, investment decisions and to set performance targets for the organization. For the Valhall field, offshore Norway, forecasting has proved to be difficult as a single discrete event, like the failure of a key producer, can reduce annual average production by up to 10%. The current activity level at Valhall is at an all time high, with two drilling rigs occupied with drilling of new production wells, adding up to 12 new, long reach horizontal wells to the production potential yearly. Accurate prediction of drilling time and initial well rates has proved challenging, adding significant uncertainty to the production forecast. Past years forecasting performance has demonstrated that single most likely' prediction of future production is inadequate for management and planning purposes. Based on this an effort was initiated to improve the forecasting methodology and procedures with special emphasize on handling of uncertainty in Short Term Production Forecasting (STPF). With the STPF being an ongoing effort throughout the year, the preferred methodology should balance the need for using techniques that adequately handle complex relationships while still being efficient and relatively easy to apply in the daily work situation. A spreadsheet based approach to the problem is described, where the standard parameters that impact the future production of most oil fields are handled, as well as the special challenges faced in predicting the performance of the Valhall field. The proposed methodology shows how a customized spreadsheet and use of commercially available statistical analysis ad-ins allows for traditional deterministic production forecasting, partially probabilistic analysis or fully integrated uncertainty analysis. The resulting forecasts assign probabilities to all outcomes as well as a ranking of the uncertainties in the form of a tornado chart. The approach is not unique, as some companies are using techniques based on similar concepts. The specific challenges at Valhall adds additional layers of complexity, and requires extensive integration of the various risk elements. In addition to having an effective forecasting tool, communication and integration of knowledge are seen as key success factors in order to improve the short term production forecasting. The production forecast requires input from several departments and many individuals to ensure that the "company knowledge" is fully reflected. A common language's to communicate uncertainty, and an auditable process to ensure buy-in and consistency is critical for a successful outcome, and is equally important when communicating results. This paper discusses both the development and use of the forecasting tool, and the forecasting process. P. 149