Are You Learning From Your Forecasts? – Improving Resource Estimation Over Time
When we forecast oil and gas production and estimate Reserves and Resources, how often do we go back and review our historical estimates?
In larger companies there is usually a centralised Reserves group or technical discipline functions that should fulfil this role. However, there is an opportunity within other organisations, or indeed within investors and governments to perform this process and improve their own predictions.
Production forecasts and Reserves and/or Resources estimates are generated and used for various reasons. This includes:
- Short-term operational planning
- Longer term budgeting
- Development planning and scheduling
- Field life planning for facilities
- Company decision making
- Fundraising or informing stakeholders
Across these areas there are multiple stakeholders who rely on forecasts and estimates which include employees, management, JV partners, lenders, investors, analysts and regulators.
There are multiple reasons why production forecasts and estimates end up being different to the actual outcomes. These may be incorrect measurements, interpretations and assumptions in four different categories. These categories are listed below with some examples of corresponding issues:
- Earlier water breakthrough with corresponding reduction in oil production rate
- Greater compartmentalization reducing lateral reservoir connectivity and well drainage area
- Mechanical failure of downhole equipment (sliding sleeve) preventing opening of undrained reservoir zones
- Poor downhole pump performance reducing well uptime
- Increased number of weather events causing facility shutdowns
- Poor compressor performance resulting in higher downtime
- Wells shut in due to uneconomic production from lower oil price
- Lower gas demand from customers
Source: Shutterstock.com/Dashu Xinganling
So why do we get it wrong? Firstly, it’s human nature to think we know more than we actually do. Overconfidence bias is our inclination to view our own skills and abilities to be greater than they are. For prediction, this means that we are predisposed to believing our result is correct. A typical outcome of this is that the uncertainty ranges applied to our estimation results are too narrow. Additionally, there are uncertainties that we are unaware of and thus are not incorporated into our ranges or estimates. Trying to allow for these unknowns is the key challenge of forecasting and estimation.
Our prediction results are also impacted by our preferred outcome. Confirmation bias is the tendency to view and interpret data in a manner that confirms or maintains a prior belief or expectation. For example, if we have worked on a development project for a number of years, we’re more inclined to bias our estimates based on that historical experience, rather than the current field performance.
One way we can overcome these behaviours is by looking back at our historical forecasts and estimations to identify and understand where key differences arise. By evaluating these over longer periods, we can identify trends in our assumptions and interpretations across the four areas of reservoir, wells, facilities and market. Based on the differences identified and trends that are followed we can amend our future assumptions to improve our predictions. This is not only for currently producing fields, but we can also translate these lessons to new discoveries or prospects. Further, instead of just looking at one single field, by looking at a portfolio of assets we can identify more generally if interpretation or analysis decisions are leading to regularly skewed results.
For example, is simulation of heterogenous reservoirs leading to optimistic forecasts of water breakthrough and water cut development? Once trends and tendencies are detected then estimation and forecasting methodologies can be improved to account for these. These improvements should then be documented and implemented in company systems and processes.
Source: Shutterstock.com/Foxy burrow
Review of forecasts and recoverable estimates can be performed across the full field life cycle from exploration, appraisal, development and to production. The SPE Petroleum Resource Management System is designed to track Reserves and Resources estimation throughout a project’s life from Prospective Resources to Contingent Resources and finally to Reserves. This allows different classifications of resource type depending on risk and a range of outcomes depending on the uncertainty. By following the PRMS and documenting the results a company can track how recoverable volumes have changed over time and the basis for these changes. This analysis can be furthered by comparing not only recoverable estimates, but also forecasts against actuals over project life.
At ERCE we have performed Reserves and Resources audits and assessments since the 1970s. We are experts at both top-down and bottom-up methodologies for recoverable estimation and forecast generation. Our highly qualified staff and wide range of global experience has provided us with a deep understanding of different methodologies and techniques. We have also worked in, for and with a wide range of companies including Majors, NOCs and independents from which we have acquired knowledge of industry best-practice for forecasting and estimation. We can assist you in reviewing your company’s historical estimates along with your processes and methodologies, and provide key insights on how to improve them to get better predictive outcomes. ERCE can also support development of company Reserves and Resources systems and documentation aligned with the SPE PRMS and regulatory requirements for different jurisdictions.
To learn more, please get in touch with ERCE office in your region.