by Gary Citron, PhD.
While writing his historical novel ‘Roots’, Alex Haley travelled back to his ancestral homeland to learn about a certain aspect of his history. However, to get to that aspect, he was required by the tribal chief to listen to their entire history. While writing this blog contributing to our series on assurance, I have a better understanding of why that is.
It helped me recall the associated circumstances of my first experience as a consultant to relay a key message: When starting an assurance review, please remember to let the team get their story out, the first time, without interruption. In other words, listen first. Reflecting upon memories, I would like to share my experience learning that message.
In 1999, I was deployed to Tokyo by Pete Rose to consult with the Japanese NOC (now JOGMEC). That week in my hotel room I received a call from James Painter at Ocean Energy (OE) to help them implement a risk analysis system. This entailed (1) training their staff on assessment concepts, (2) demonstrating a prospect review process via our ‘surgical theatre’ template of reality checking a prospect characterization and (3) helping implement the central coordination process to compare the prospective apples of one group to the oranges of another via an assurance team.
For a company like OE this was particularly important because of their diverse portfolio, built organically in the USA (from their Flores and Rucks heritage) and from their international opportunities that came in part from their purchase of Meridian Oil.
Their impressive assurance team staff included a stratigrapher with vast international experience from Shell, two very talented reservoir engineers and a relatively senior exploration manager who had the ear of their CEO James Hackett. Mr. Hackett went on to distinguish himself further at Devon Energy, which purchased OE in 2003, and particularly later when he joined Anadarko as their CEO. With their highly trained technical staff, savvy assurance team and executive support, they should be well positioned to succeed. Prediction accuracy success (or failure) would largely depend on assurance execution, and I was hired to assist on assurance design and implementation matters.
Part of my remit was to attend several of their initial assurance reviews and advise the assurance team on best practice. I found an alarming pattern from those reviews. The technical staff were upset that these talented assurance guys were dissecting their prospects before even hearing about the play context. The feedback I heard from the technical teams was that with the current behaviors by the assurance team the technical teams would be hard pressed to solicit further engagement, much less participate in future reviews.
To paraphrase, ‘what was the point of having them if we could not explain the nature of the opportunity’. That feedback was easy to compile, but hard to share with the assurance team. I suspect they hadn’t heard much criticism before in their careers, where they advanced to key technical and managerial roles rather rapidly.
Fortunately, the feedback was readily accepted (after some disbelief stated along the lines of “really?”), and the team made sure they structured their sessions with the discipline to wait until the key points were established, which then provided a foundation for questions and a critical review of the uncertainty and chance factors associated with key parameters.
The take-away we apply for other clients: assurance teams should issue engagement guidelines to the teams that generate opportunities. Such guidelines describe what (and the sequencing for what) they need to see for a review, and in return, what the assurance team provides.
Overcoming the initial hiccup, OE established a state of prediction accuracy over the next three years that was enviable.
The percentile histogram (Enciso et al, 2003) represents virtually no estimation bias, perfection being a straight line across at 20%. As defined by Otis and Schneiderman (1997) in their landmark paper illustrating how an assurance team can make an impact, percentile histograms show the percent of discoveries that fall in prescribed buckets by measuring where the result for each discovery posts on each prospect file’s success case EUR distribution.
This type of graph facilitates benchmarking against other companies to illustrate thematic estimation problems across the industry. We’ve noted in our prospect risk analysis course that many companies have percentile histograms where more than 40% of their discoveries fall below the forecast P80 of the EUR distribution. The next blog will delve into percentile histogram construction and diagnosis in more detail, so stay tuned.
While there are many attributes to successful assurance efforts (which we will further explore in this blog series), active but patient listening in an assurance review is a good start.
Enciso, G., Painter, J. and Koerner, K.R., 2003, Total Approach to Risk Analysis and Post-audit Results by an Independent, Poster presentation to the AAPG International Convention, Barcelona, Spain.
Otis, R., and N. Schneidermann, 1997, A Process for Evaluating Prospects, AAPG Bulletin, v. 81, n. 7, pp. 1087-1109.
Gary served on Amoco’s assurance team from 1994 to 1999. In 2001 he became Pete Rose’s first partner in Rose & Associates. He helped create the Risk Coordinators Workshop series in 2008, which remains active.
by Peter Carragher
I have been engaged with teams in reviewing and assuring plays and prospects since the late 1990s. I think it’s important to keep the title of this brief note front and center as a constant reminder that there is always something to learn. Continuing to learn will improve your ability to be effective in prospect evaluation and remain a valued contributor. Here are four areas to consider.
A prerequisite for learning from failures is the absolute necessity for rigorous written documentation of the pre-drill assessment. That must include all the pre-drill geological and geophysical assumptions, as well as the actual parameters and chance assessment logic you and the team used to build the assessment. Particularly instructive is writing down the possible failure modes for the prospect so you can determine if the actual failure was a surprise or not. The assignment of chance to the individual elements should reflect these concerns. Over time, you can determine the dominant failure modes in your program and take steps to address those issues with technology, or with different decisions.
Over estimation of discovered resources is a common problem, sometimes offset in a portfolio by underestimation of the chance of success, a classic case of two wrongs sort of making a right. What is happening here is that the relative ranking of prospects is incorrect, and that has consequences in the management decisions about the portfolio. For example, a company decided to drill a prospect at 100% and reject a valid farm-in / interest swap offer because the prospect resource and value was estimated to be very large. Unfortunately, one of the major risks was not recognized pre-drill. Not only was the well a dry hole, but the counterparty well was a commercial discovery.
One of the benefits of positive interaction in an assurance review is that the experience of the assurance team can help the project team in their technical work, prospect description and communications. Less well appreciated, is the necessity that the assurance team learns from the project team. Project teams bring diversity of thinking, new technology, different visualizations, new interpretation models, and the details of local geoscience data to bear on prospects. Careful, respectful listening and seeking to understand that point of view is a prerequisite for long term success in assurance.
Technological advances in the geosciences contribute towards increased specialization. Special effort is required to really learn what is going on in these specialized sub-disciplines, and how they interact with each other. For example, a 3D oil and gas migration model is critically dependent not only on the geochemistry of the source rock, but also the crustal mode that drives the temperature profile, biostratigraphy determining the ages of the burial sequence, geophysics generating depth maps and sequence stratigraphy predicting the distribution of permeability that controls migration vectors.
There is also much to learn from the wider set of disciplines involved in subsurface projects. You will become a better geoscientist if you take the time and put in the effort to learn from your colleagues in reservoir engineering, reservoir modeling and simulation, drilling, and production engineering, to list some of the opportunities. Assurance teams are best served by individuals who have taken this step, integrating learning across disciplines.
Whether you are involved as an assurance team member, or as a project team member, assurance presents a series of great opportunities to learn something new. Take them!
Pete Carragher has been the Managing Partner of Rose & Associates LLP since July 2014. Prior to joining R&A in 2010, his last role at BP was VP of Geoscience and Exploration, working as a member of BP’s global exploration leadership. Pete and his team introduced systematic risk and volume assessment, and assurance, into Amoco in 1990. Since then, he has evaluated hundreds of prospects from most of the world’s productive and frontier basins.
by Juliet Irvin, Guest Contributor
As an experienced assurance professional, I have seen many different approaches to assurance over the years, some more successful than others. The first part of this article explores when and how assurance can be most effective, the role of the outsider, the importance of putting things into context and testing assumptions.
Early engagement between the assurance team and the technical team is absolutely key. For example, if the assurance team sees the prospect for the first time in the technical reviews (when usually a large proportion of the work has already been completed), it may be too late to steer the analysis and correct any errors, thus exposing a business decision to an incomplete or flawed evaluation.
Imagine an alternative approach, where a member of the assurance team is involved from the beginning of the prospect characterisation and can assist the technical team throughout. Together they can frame the business questions which must be addressed and design an evaluation that is fit for purpose and adds value. For example, there is little value to be had from a very detailed prospect assessment, separating out every fault and reservoir segment, when the business decision at hand is whether to bid a 2D seismic survey to secure a block in a frontier area. This would be a significant waste of time and resources that could have been easily avoided.
There have been many discussions about project teams conducting their own assurance process while evaluating prospects, vs. the need for external assurance. Which method works best? Having been involved with several hundred prospect assessments, I’m convinced that assurance works most effectively when an impartial and independent person is involved, and that person can contribute by:
Any prospect assessment is only as good as the inputs (“garbage in, garbage out”), so another key aspect of assurance is understanding what is behind the numbers. For example:
We will now take a closer look at each of these components to explore how impartial assurance, and hence an outsider view, can be an integrated and valuable part of project evaluation.
This may sound like an obvious question, but when this is not clearly understood, it can cause significant problems. To avoid different assumptions, it is important to describe explicitly the geologic model of the prospect being assessed. For example, are we assessing the chance of sand or the chance of a specific environment of deposition? These are not necessarily the same thing. Are there additional geologic models / scenarios, which could yield a success case? Do they need to be characterised at this stage, to maximise understanding of the value of the prospect? Drawing out this clear description up front ensures a common understanding for the technical, management, and assurance teams, and helps to mitigate erroneous assumptions.
When considering analogues, it is vital to consider how likely those scenarios are to occur. For example, if they are highly optimistic cases, management need to be aware of that for subsequent valuation and decision purposes.
When conducting a prospect assessment, the technical team will provide a range of values for each of the input parameters, e.g. porosity, net-to-gross (NTG), saturation, etc. Understanding the basis of these ranges is an important part of the assurance process.
For example: when characterising the NTG, we are usually looking to describe the full range it can average across the prospect, given success. Often the ranges in the assessment do not honour this concept – for example, we may get a maximum NTG value from a single well, or a minimum NTG value seen in a core sample, without considering if those values could represent the minimum to maximum average for the prospect as a whole.
In my experience, a useful way of testing for plausible ranges is to ask geoscientists to draw minimum and maximum NTG maps depicting the environment of deposition (Figure 1). This requires the interpreter to explicitly describe the geologic scenarios which would give rise to those extremes and consider if they are appropriate, and then modify their assessments if needed.
Figure 1. Reservoir geologic scenarios for a prospect (dashed outline) in the SE quadrant of the map
“Assess early and assess often” was a mantra often heard from one of my mentors, and it ties in directly with another one: “what business question are we trying to answer?”
Most technical evaluations have both a divergent ideas stage and a convergent stage. Initially the teams are using their creativity, seeking all available information and coming up with ideas of which areas and reservoirs may be prospective. When concepts start to converge, this is a useful stage to engage with the assurance team, to conduct an early prospect assessment, and explore which aspects of the evaluation have the biggest impact on the results. Figure 2, for example, shows that GRV has the greatest impact on the estimated resources with Recovery Factors influence being minimal.
Figure 2. Resource parameters and their influence on the resource assessment.
This can ensure the remaining work effort is focused on the most impactful inputs (in this case, GRV), thus avoiding spending time and resources on work that will make little difference to the business decision.
It is very easy to plug numbers into a volumetric software package and calculate a range of outcomes. But are the outcomes meaningful? Do they address the relevant business questions? This is where assurance can help, contributing with early and active engagement, an outside facilitator, an open listening mindset, and some techniques to practically engage with the technical team.
In my view, a key sign of success is when there are no major surprises at the final project reviews, because all relevant aspects of the assessment have already been tested, documented, and addressed along the way with an analysis that has directly addressed the business questions in order to support timely and appropriate decision making.
Management decisions are more informed when the assurance team participates as an advisor, participating in discussions to understand any concerns and provide an independent view of the technical inputs into the analysis and the associated justifications, identify any biases that may have been in play, and communicate the critical risks and uncertainties with the prospect. This can help ensure the conversation is focused on the geologic concepts underlying the analysis, rather than just the numbers coming out. If management concerns are still valid, the assurance team can work with the technical team to address and revise the evaluation if appropriate.
At the end of the day, all technical assessments should help the decision makers make better business decisions by having the right information and analysis available.
Juliet is an advanced Risk and Uncertainty Specialist and Geoscientist. She has 22 years’ experience with ExxonMobil (1998-2021) spanning Oil & Gas Exploration, Development, Production, Research and IT, and has been an Assessment and Assurance advisor / specialist for over 10 years. She also has more than 14 years’ experience running professional training classes including Risk & Uncertainty and Prospect and Play Assessment.
by Bruce Appelbaum (Mosaic Resources), Guest Contributor
My association with Texaco’s exploration program began in 1990 with the offshore Gulf of Mexico group in New Orleans. At the time, Texaco’s yearly exploration budget was a portion of the producing divisions’ stipend, with only informal communication among the exploration groups except at yearly budget meetings. Those meetings consisted of prospect presentations with little commonality and process largely independently determined by the division originating the prospects. The prospects were presented with the originator’s bias and inconsistent risking methodology. Gamesmanship was rampant as success was determined by the amount of budget committed from a pot controlled by the global exploration managers’ group. The process was akin to getting dinner at the boarding house table. Coal bed methane projects competed with rank offshore wildcats for limited capital. Given the bias and gamesmanship, global exploration success was predictably poor.
In New Orleans, portfolio analysis was taken seriously, and risk analysis training was brought in as a standard tool in the planning process. Pete Rose was invited in as the entire division was put through his flagship course on prospect evaluation. Concurrently, a process was emplaced which led to the formation of a small group of risk specialists that produced an independent view of prospect size and chance before the prospect was added to the portfolio. The group also reviewed and evaluated individual prospects being developed for OCS sales in the Gulf of Mexico. Metrics were developed to indicate what was needed to both replace our produced reserves and grow reserves at a significant rate. The increasing success of the division was noticed at our New York corporate HQ, and the credibility of the New Orleans offshore group grew in great measure.
One of the problems holding the offshore group back was a history of expensive dry holes in deeper Gulf waters and a fear at HQ of further deep-water drilling. This was during the period of initial success by Texaco’s major competitors. The company had to get in the deep-water game if it was to have a chance at elevating to top quartile performance among its peers. At the time, the perception was that HQ believed the division was throwing darts at a map of the Gulf to pick its drill sites!
To change this perception, we demonstrated our exploration process and risking technique to management in New York. A plot of our top six deep-water prospects was constructed. Given the individual risks of the prospects shown, we had a 60% probability of achieving at least one discovery with the portfolio presented. The group was duly impressed, we won the day, and the deep-water Gulf frontier was added to our purview. The revamped focus led to Petronius, Tahiti, Blind Faith and the Perdidio area discoveries, to name a few.
In 1996 a newly created global exploration job in Houston was assigned to me, with an ambition to elevate the company’s exploration success both domestically and internationally. As part of my preparation, I evaluated the more successful programs of our peers and interviewed several of our competitors. Amoco was an extremely successful exploration force at the time, with notable success in West Africa, the North Sea and other areas. I visited my Amoco counterpart David Work, and we chatted through Amoco’s process. It started with a global risk team that reviewed both pre-drill and post-drill the world-wide portfolio of projects.
I determined that a global exploration group separate from the producing divisions was necessary at Texaco. This would allow a focus to fund the most prospective projects within a single inventory of competitive, properly risked opportunities. Part of that effort was a fully dedicated Global Risk and Standards Team (GRST) charged with properly evaluating the risk attributes of all projects in the global exploration portfolio. Jim Mackay, Paul Haryott, Michael Joseph, Dave Taber and Jim Varnon were all part of iterations of the risk team. Ultimately, Paul led the team at Texaco and continued as the first manager of Chevron-Texaco’s assurance effort, called the Exploration Review Team.
Concurrent with the new portfolio methodology, Texaco evolved a focused approach concentrated on several basins which played to our strengths, and abandoned those plays that sapped our financial and technical resources. We allocated about three quarters of our resources to the Gulf of Mexico, offshore West Africa, and offshore Brazil, which had just opened to international investment. The remaining quarter was targeted at the Australian Northwest Shelf, the North Sea, Trinidad, and several emerging high potential arenas. We retreated from efforts in Italy, Poland, Thailand, Viet Nam and several other venues which could not meet our metrics. Though highly prospective, we chose not to enter Venezuelan exploration because of the restrictive licensing terms.
The period from 1997 to the Chevron merger in 2001 bore out our reconstituted exploration efforts and Texaco indeed achieved top quartile exploration metrics. Texaco added approximately four billion BOE net through exploration in this period, including giant discoveries at Agbami in deep water Nigeria as well as Janz on the Australian Northwest Shelf. Janz is part of the Gorgon LNG project and remains the largest gas discovery ever made in Australia. Many of the discoveries were in initial phases of development when the merger talks with Chevron heated up. I believe these resources were a very large driver for the ultimate combination of Texaco with Chevron.
Post-merger, I understand that Texaco’s exploration process was largely adopted by the combined company. While the terminology may have been different, there is no question that the assurance process, driven by consistent, dispassionate risk assessment by a great crew of geoscientists and engineers was a key component of a great company’s last independent successes. All who worked at Texaco during this period can be justifiably proud of what their efforts produced and its foundation for Chevron-Texaco’s continued success.
Dr. Bruce Appelbaum has enjoyed a long and successful career in the oil and gas industry, culminating in his being named a Vice President and corporate officer of Texaco Inc. He has served on the Board of Directors of the CQS Rig Finance Fund and Input/Output. Additionally, he is a Distinguished Trustee of the American Geosciences Institute Foundation. He is a Member of the AAPG Corporation and an AAPG Foundation Trustee Associate. His advisory positions include the School of Geosciences at Texas A&M, the Dean’s Advisory Council at the State University of New York at Buffalo, and formerly, the School of Earth Sciences at Stanford University. He is currently a member of the Baker Institute Roundtable at Rice University.
by Kenneth C. Hood, Guest Contributor
When geoscientists evaluate an opportunity, we tend to focus on the geological aspects we are most familiar with. This is human nature. Unfortunately, what we spend the most time on may not be the most important factor for understanding an opportunity’s potential. As an example, a team may spend many hours refining estimates of porosity or net-to-gross, when the impact of these parameters often pales in comparison to the impact of hydrocarbon column height (hydrocarbon-water contact depth).
In my assurance experience, different assessment teams have used very different representations of column height or hydrocarbon-water contact depth as part of opportunity evaluations. Often at the screening stage, they would simply evaluate an opportunity as being filled to spill. This, coupled with permissive structure sizes based on preliminary mapping of limited data, tends to result in exceptionally optimistic hydrocarbon volume estimates that almost always get reduced with additional work. Other assessors preferred an exponential decline of column heights from the assessment minimum down to synclinal spill. This would imply that every trap is almost always significantly underfilled, with essentially no chance of being filled to spill. Compared to a filled-to-spill scenario, this is a very pessimistic outlook indeed. While it is important to be as accurate as possible with column height estimates, it is also important to be consistent among opportunities, so they at least have a reliable basis for comparison.
As part of the assurance process, it is essential to verify and document assumptions about seal capacity as well as all potential geometric limits, aka leak mechanisms (including probability), and to ensure that the assessment analysis is created in a manner consistent with the geologic concept being evaluated. Experience has shown that the analyses frequently do not match the geologic description and the available constraints.
The details of how to best represent hydrocarbon column height in an evaluation will vary depending on software capabilities. Where possible, the workflow preferred here is to build background column heights and geometric limits (spill depth) as separate distributions. This approach enables the use of standard background column height distributions for families of related prospects while still honoring the unique configuration of each trap.
The background column height distribution is the column height supported by bed seal capacity, with a potential overprint of unresolved geometric spills (if based on analogs). Such distributions can be based on capillary constraints (e.g., capillary pressure data coupled with geologic models) or empirical column height data from local or analog discoveries (for a recent example, see Edmundson et al, 2021). Using empirical data can be challenging in that hydrocarbon pools controlled by geometric limits document the minimum column that the seal can support but not the upper limit. Figure 1 illustrates some representative background column height distributions applied to different families of structural prospects. Column height distributions should start at a consistent assessment minimum. The assessment minimum can be effectively linked to seal capacity if based on a minimum column height and not a minimum hydrocarbon volume.
The definable geometric limits on column height include controls such as synclinal spill, reservoir juxtapositions, fault intersections, and channels or scours, each with an associated probability of leak. Each probability is conditional on the success of shallower potential limits, such that the total must sum to 1.0.
Figure 2 illustrates the importance of using separate distributions for column height and geometric limits for an opportunity evaluation. Because the input column height distribution represents the capacity of the seal to support a hydrocarbon column, in many cases it will extend beyond the trap spill(s). During the Monte Carlo convolution, both the column height distribution and the spill depth distribution are randomly sampled. The output column height for each realization is the minimum of these two values, thus creating a mode at the spill limit where the magnitude is automatically scaled to and consistent with the seal capacity. This output distribution is this displayed in a frequency format. In this example, the base case is a 600 m closure with a regional column height distribution, resulting in a mean volume of 305 MOEB (lowermost table, Figure 2A). Now consider the addition of a potential leak, such as a fault intersection, over a 50 m interval centered at 500 m below the crest. This zone has a 0.5 chance of leaking. The realizations for which this leak controls the contact should produce a mode in the output column height distribution. If this mode is represented using a weighted input column height distribution (Figure 2B), the apparent prospect volume actually increases. The increase results because the weighted input distribution reduces outcomes from above the leak as well as below it. Such behavior is not geologic – the leak should only reduce realizations deeper than the fault intersection. By using the regional column height distribution coupled with a weighted spill-depth distribution (Figure 2C), the prospect volume decreases as expected. With the weighted spill distributions, realizations above the leak are unchanged from the base case. This example illustrates how combining the background column height and explicit geometric limit(s) into a single input distribution produces non-geologic and erroneous results in Figure 2B.
Part 2 continues with a discussion of the impact of column height on opportunity evaluation. The impact on the economic viability of an opportunity can be substantial.
I thank ExxonMobil for releasing this material. Many colleagues have contributed to this work.
Edmundson, I.S., Davies, R., Frette, L.U., Mackie, S, Kavli, E.A., Rotevatn, A., Yielding, G, and Dunbar, A., 2021, An empirical approach to estimating hydrocarbon column heights for improved predrill volume prediction in hydrocarbon exploration, AAPG Bulletin v105, n12, pp 2381-2403.
Hood, K., 2019, Hydrocarbon Column Height, Presentation at the 2019 Rose & Associates Risk Coordinators Workshop # 17, Houston, Texas.
Ken Hood holds a Ph.D. in Geology from The University of Kansas. Ken retired in 2020 after 31 years with ExxonMobil. Much of his career was spent working in assessment and assurance of conventional and unconventional resources at play and prospect scales.
