The examples below are motivated by a set of frequently asked questions (FAQs), in turn highlighting common errors seen in forecasting, and are summarized by learning points that demonstrate why a consistent forecast definition is a pre-requisite for a lean forecasting process, applicable to resource estimation, business planning and decision making. It is not a requirement to use these definitions or the proposed forecasting principles but it is considered best practice; the examples will show that, the closer a company applies these definitions and principles, the leaner the overall forecasting, resource estimation and business planning process will be. Lean in this context means "getting it right the first time" and avoiding waste and unnecessary re-work. There are many situations, where the model objectives dictate another objective function than ultimate recovery; however, the forecaster should always plan for making a P10/P50/P90 forecast that is consistent with the resource estimates in addition to the primary objectives of the study. This should be done whether the customer asks for it or not. A reservoir engineer was requested to provide forecasting support for an exploration lease sale. A number of offshore blocks were on offer and he made a Monte Carlo simulation, based on seismically derived volumes, reservoir property trends, range of well count, development/operating costs and infrastructure requirements to point of sale. The objective functions were NPV and EMV for a significant number of prospects in these offshore blocks. This was exactly the information the exploration department had requested to determine the optimal bid value of these block.

A sound production forecast is the basis for any project-based resource estimate, and the same production forecast is also the basis for any business or development decision. With these principles, forecasting is part of asset management throughout the year. It may be envisaged as a continuous loop through the whole upstream lifecycle Fig 1. Forecast updates are triggered by reserves and corporate planning, but also by ad hoc changes and events, such as studies and subsurface information and development plan updates. Note that for the official forecasts (reserves and corporate planning), reasonable freeze dates should be agreed upon for input data and should be adhered to. Subsequent changes due to new wells, workovers or wells failing should be reflected in the ad hoc updates, The forecast model is kept up-to date and consistent with the latest surveillance data and development assumptions and when reserves or corporate forecast need to be updated, it may simply be derived from the latest model.

Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content.

Hydrocarbon production potential is often limited by constraints, and it is important that these constraints are understood and correctly represented when generating a realistic set of production profiles. The focus of this section is physical constraints in the system through which the fluid flows, but constraints applied because of reservoir management, contractual terms and economics are also highlighted. A production system includes the reservoir, wells, facilities and export system. Constraints within the system can be associated with any of the produced fluids (oil, gas or water) or a specific combination of them. For example, important factors to consider beyond the base deliverability of the reservoir are potential near-wellbore formation damage (skin), well tubing constraints, artificial lift availability, shared gathering system back pressures, flow line erosion velocity limits and facility capacities.

It follows that forecasts that deviate from the actual production with hindsight can still be good forecasts if the uncertainty range is properly defined, justified and documented as will be shown by Example 6 on the Production forecasting FAQs. However, forecasts that contradict each other even though they are based on the same information cannot be good forecasts, even if they are made for different purposes see Example 3 of the Production forecasting FAQs. It is customary in the industry to describe this uncertainty in terms of a low (P90)/high (P10) range. This is consistent with both the Petroleum Resource Management System (PRMS) [1] and the Securities and Exchange Commission (SEC) [2]. For volume estimates, a low (P90)/high (P10) range is thus unambiguously defined by statistics.

Given the important role production profiles play in key decision making and the inherent uncertainty and bias existing in the process of building them, it is essential that the QA/QC process is sufficient to test the robustness of the forecast range and ensure there is a shared understanding and buy-in of the assumptions behind it across the stakeholders involved in an asset. Getting other qualified professionals not directly involved in the development of this particular forecast to review inputs, methodology and outputs of the forecast. This is a powerful tool, particularly in mitigating the effects of human bias. Each reservoir and development is unique so there are likely to be different approaches that should all be considered. This should include plateau length and production rates as well as individual well rates, decline, response to injection and recovery factors.

A production forecast is an estimate of gross and net volumes (oil, gas, and water) expected to be produced from a hydrocarbon accumulation over its remaining life time, as well as fluids injected to produce these volumes (water, gas, and steam). The forecast must be aligned with the hydrocarbon lifecycle development strategy and plans. Imagine you had a time machine and you could travel 30 years into the future. You could observe the production of your asset to the end of its field life and know exactly what your field has produced over time. You could come back to the present and make the perfect forecast.

A good forecast combines all six production forcasting building blocks in a numerical model. If this is a green field model, similar steps should be carried out to calibrate the model with known analog performance. Always start by agreeing on the objectives of the forecast and by deriving a forecasting strategy based on these objectives. Once the objectives are clear, the appropriate input data must be gathered. These steps of history matching and model calibration can be carried out in parallel and potentially by different discipline engineers.

Fig 1 shows the building blocks for a production forecast which are further discussed throughout this document. Key reservoir characteristics are remaining hydrocarbon volumes, structure, architecture, phase contacts, rock and fluid properties, sand connectivity, heterogeneity, faulting, fracturing, drive mechanism, and aquifer pressure support to determine the future capability of the field to deliver production. The inflow and tubing lift performance of the production conduits impact the surface flow rates from the wells and their expected decline. Key aspects to consider are well design, completion efficiency, lift curves, impairment, sand production, well integrity, coning/cusping potential, liquid loading, interference and availability of artificial lift, and in the case of injection, whether this is done above or below fracture propagation pressure. Future production profiles and ultimate recovery from a field are affected by the envisaged development plan in terms of number, type (e.g., horizontal or smart wells), density and phasing of wells, reservoir management policy, project timing, displacement process, suitability of improved recovery techniques and injection volumes.

Whenever we make a production forecast, we use a model of the reservoir and production system, whether it be in the form of a simple mathematical equation (e.g. decline curves) or a full-field 3-D simulation model with surface networking. But no matter how complex the model, it is always a simplification of reality. The aim of the forecaster, in making reliable predictions, is to use a model that is sufficiently representative of the physical processes and constraints, and that adequately allows, on a resource-benefit basis, for treatment of uncertainty in the input and output of the model. Somewhere in the spectrum of modelling techniques, material balance plays an important role; either in supporting a more complex model (e.g. Material balance treats the reservoir as a tank (or limited number of tanks) with uniform properties.