Oil and Gas Evaluation Report : Oil & Gas Evaluation Professionals : Obsidian Energy Company : · Seismic Interpretation, Petroleum Engineering

What is the hydrocarbon potential of a shale?  To answer this question you first have to determine if the shale contains sufficient amount of organic matter, and second, has the shale been subject to the geologic processes needed to convert organic material to oil and/or gas?

What is the total organic carbon (TOC) present in the shale?   TOC is essentially the total amount of organic matter (kerogen) in a given sample of rock.  Most shale plays have a TOC greater than 3%.   TOC alone is only a start in evaluating a shale play.

Rock-Eval Pyrolisis evaluates the type and the maturity of the organic matter as well as determining its petroleum potential.  Rock-Eval reveals the amount of generated oil and gas in the shale sample (S1), the amount of oil and gas generated through thermal breakdown of organic matter by heating (S2), the amount of CO2 released during pyrolysis(S3), and the temperature of maximum release of hydrocarbons (Tmax).  Tmax indicates the stage of maturation.  From this data the origin of the organic matter can be determined by the hydrogen index (HI).  The oxygen index (OI) measures the oxygen richness of the sample.  The production index (PI) is the ratio of generated hydrocarbons to potential hydrocarbons.  Low ratios indicate immature or postmature organic matter.

Vitrinite reflectance (VF) is used for determining the thermal maturity of the shale.  Vintrinite reflectance is sensitive to temperatures that correlate to hydrocarbon generation (60 to 120 degrees C).  Generally, onset of oil generation in an oil-prone shale is a VF of 0.6% and a VF of >1.35%, while the onset of generation in a gas-prone shale is a VF of 0.8% and a VF of >2%.

Harry Dembicki Jr. in his article "Three common source rock evaluation errors made by geologists during prospect or play appraisals" (AAPG Bulletin, v.93, no.3 (March 2009), pp. 341-356) discusses the pitfalls of relying on only one aspect of petroleum geochemistry.  Mr. Dembicki points out that it is necessary to "fully integrate TOC and Rock-Eval data with pyrolysis-gas chromatography, and using burial history diagrams to help interpret vitrinite reflectance."

Remember the petroleum geochemistry of any shale in not homogeneous vertically or horizontally in a particular unit.  This heterogeneity must be taken into account when evaluating a play before and after production has been commenced.  In addition, the shale must have adequate volume. That is it must be sufficiently thick and have an areal extent to generate producible hydrocarbons.

• Email This
• Print
• Comments
• Trackbacks
• Share Link

The concept of portfolio diversification has been a widely accepted method for reducing uncertainty (risk) in financial markets since Harry Markowitz's work in the 1950s.  In fact, one of the reasons financial institutions like oil and gas investments is, in part, due to the concept of portfolio optimization and the affect direct investment in these commodities has on their overall portfolio.  In general, oil companies, however, have been less receptive to implementing these concepts.

First some basic portfolio optimization concepts.  The basic idea is that while individual investments (properties, projects, etc) have their own unique set of parameters (ENPV-estimated net present value, risk, etc), when you combine these investments into a portfolio the interaction of these individual investments can alter the risk of the portfolio.  This is best demonstrated by an example used in Ball & Savage Associates 1999 paper where you have $10MM to invest and two distinct projects to choose from.  The first is a relatively "safe" investment and the second is relatively "risky," but the estimated NPV is the same for each.

ENPV(safe) = 60% * $50 + 40% * (-$10) = $26MM

ENPV(risky) = 40% * $80 + 60% * (-$10) = $26MM

The two projects are independent, that is to say the result of one doesn't affect the other.  In this example the authors go on to suppose your job is dependent on not losing money.  So, you can see with the "safe" project you only have a 40% chance of losing your job, but with the "risky" project you run a 60% chance of being fired.  Since both projects have the same ENPV, most people would correctly choose to invest in the "safe" project.

However, in this example if you were allowed to invest half your money in each project the only way you could lose money is by both projects being unsuccessful (if one is successful it pays for the other being unsuccessful), and since they are independent, the chance of both being unsuccessful is 40% * 60% = 24%.  So, by spreading your money between two projects instead of one "safe" project you have reduced your "risk" of unemployment from 40% down to 24%.  This is the power of the "diversification effect" and is not intuitively obvious to most of us.

The effects of diversification can be even greater when the projects are not completely independent as stated above.  If the projects are statistically dependent then the outcome of one will have an effect on the outcome of the other.  If the outcome of one project increases the chance of a similar outcome in the other, then the projects are Positively Correlated.  If, however, the outcome of one project reduces the chance of a similar outcome in the other, then they are Negatively Correlated. 

In the above example, if the projects were positively correlated, then the 50/50 portfolio would have a greater than 24% change of getting you fired, but if they were negatively correlated you would have an even less than 24% chance of being sacked.  The exact chance would depend on the degree of correlation.  The idea here is risk can be minimized by spreading your investment across many projects and trying to avoid positive correlations while looking for negative correlations. 

As I mentioned at the beginning, this is one reason institutional investors like direct investment in commodities such as oil and gas.  Historically, direct investment in oil and gas has had a relatively strong negative correlation to more traditional investments (stocks and bonds).  So by including some oil and gas investments in a traditional portfolio the "risk" of the entire portfolio can be reduced and moved closer the "efficient frontier" as advocated by Markowitz; of course there are other reasons to invest in oil and gas as Kathy Heshelow outlines on her web site.

These concepts are used by some of the larger oil companies and mostly in determining exploration programs, but my experience is that they are not being used by most of the smaller companies and rarely when looking at A&D programs.  While there are computer programs that can simulate this portfolio effect, most of the smaller companies I am familiar with shy away from this methodology and prefer more intuitive investment strategies; specifically new start-ups who prefer to get their capital invested quickly and would rather take 100% of fewer projects rather than "spread the risk" and delay getting their capital in play.

The general concepts of portfolio optimization can be applied without having to perform the rigorous calculations or build complex computer models simply by practicing diversification and spreading the risk around (several smaller deals instead of fewer big ones) and looking for negatively correlated projects.

• Email This
• Print
• Comments
• Trackbacks
• Share Link

The economics of a shale play are sensitive to certain criteria that may not be critical to a conventional type oil or gas play.  One important criterion is the Initial Potential (IP) and the shape of the hyperbolic decline production curve.  A hyperbolic decline curve is composed of an IP followed by an initial steep decline rate, transitioning into a later long term, shallow, stabilized decline rate (see the graph below from Range Resources).  Shale production is characterizes by a steep decline curve early in its productive life.  The more oil and/or gas that you can make up front the better the economics.  

We've heard about the impressive IP's coming out of the Haynesville Shale and Marcellus Shale plays.  Currently, there is a lot of discussion about the initial decline rate of the Haynesville.  Analysis of current producing wells indicates that the wells are stabilizing in about one year.  This rapid decline calls into question some of the large reserves and the economics being proclaimed by the operators in the play.  Allen Brooks discusses this in his article:  New Research Questions Haynesville Shale Economics.

Obviously, costs and prices are also important criteria affecting the economics.  However, these two factors are known early on in a shale plays life.  Costs are determined by:

• Depth of target shale and length of lateral
• Number and size of stimulations
• Lease prices
• Existing infrastructure

Finally, price (especially for gas) can make or break a play.  Gas price depends on proximity to demand.  A big play driver for the Marcellus Shale is the price that producers get for their product.  The northeastern U.S. has the highest gas prices in the nation.  With the current price disparity between oil and gas, an oil shale play such as the Bakken has better economics than the Barnett Shale.  Due to the amount of reserves produced early in the wells life, the price on day one may determine if you drill or not.  Testament to this fact is the decline in activity in the Barnett over the last year.

Time will sort out the economics of all shale plays.  Watch the decline curves and hope for higher gas prices.

• Email This
• Print
• Comments
• Trackbacks
• Share Link

Flooding an oil field with extraneous water has been a widely accepted method for increasing a reservoir's recovery since the 1950's, but to the uninitiated it may seem odd.  After all, water production is a bad thing; it increases lifting costs, puts more strain on equipment, and may even prevent flowing wells from flowing.  Plus, the produced water must be dealt with in an environmentally sound way, which also adds to the operating costs.

So why add water?  For two reasons: First, injecting anything into a reservoir will increase the reservoir pressure, and second, water and oil don't mix.  This second reason may again seem odd, but because they don't mix water, under higher pressure, will displace the oil it contacts.

So what does this mean?  First we need to understand that most oil reservoirs are solution gas drive reservoirs*.  This means as the oil is produced the reservoir's pressure is reduced and the gas that was held in solution begins to breakout and expand, thus "driving" the oil towards the producing wells.  This is a familiar process we see when opening a bottle of soda (Mentos added for emphasis).

The problem with a solution gas drive reservoir is when the gas breaks out of solution it is free to flow to the producing well and be produced, and once the gas is produced the reservoir's energy is lost.  Typically a solution gas drive reservoir will only recover 5-20% of the reservoir's original volume of oil leaving a large portion behind.

By injecting water in a controlled manner, the loss of reservoir pressure can be controlled and reversed.  Water is injected into dedicated injection wells strategically located throughout the reservoir, and the water itself can be used to displace the remaining oil towards the producing wells.  If properly designed and operated, a water flood can double the reservoir's oil recovery.

Even with double the recovery (10-40%), we are leaving large volumes of oil behind in these solution gas drive reservoirs, and with the ever-growing oil-thirsty economies around the world we need to do better.  This is where enhanced oil recovery (EOR) techniques come to play, but that discussion is for another day.

This all sounds great and water flooding has been used successfully for decades, however, it is important to take care to design and operate the flood appropriately, otherwise all the bad things we mentioned at the beginning may be all you get.  There are may factors to consider when designing a successful water flood including:

• reservoir permeability (both absolute and relative)
• beginning and ending fluid saturations (oil, water and gas)
• reservoir heterogeneity
• oil gravity and viscosity
• water source and compatibility
• formation clay content
• depth and lifting costs

But if done right, a well run water flood will significantly improve oil recovery and produce attractive returns for many years.

 *Most of the oil reservoirs considered for water flooding are solution gas drive; of course there are a great many oil reservoirs around the world that are not, but they are not the subject of this discussion.

• Email This
• Print
• Comments (3)
• Trackbacks
• Share Link

We will periodically feature a guest author and this post was contributed by Staffan Van Dyke 

Spectral decomposition is a novel seismic technique that was originally pioneered through research at BP and Amoco in the 1990’s. Spectral decomposition (spec-decomp) is an imaging innovation that provides interpreters with high-resolution reservoir detail for imaging and mapping temporal bed thickness and geological discontinuities within 3D seismic surveys by breaking down the seismic signal into its component frequencies.

A fully processed seismic survey contains all of the frequencies that are capable of being recorded by the geophones/hydrophones used for that particular survey (this is known as its “dynamic range”). After the seismic source has been “shot,” the energy propagates downward into the subsurface and at each geologic boundary (e.g., an unconformity, bed boundaries, etc.), the seismic energy is reflected, refracted, and/or absorbed.

As the wave front continues to propagate into the underlying sediments, it attenuates, causing the frequency content to decrease with depth, i.e., higher frequencies are better preserved at the top of the section. Due to this attenuation, the higher frequencies deeper in the seismic survey are “drowned” by the more dominant, lower frequencies. The purpose of spectral decomposition is to see the seismic response at different, discrete frequency intervals, as higher frequencies image thinner beds, while lower frequencies image thicker beds.

The concept behind spectral decomposition is that the seismic reflection from a thin bed has a characteristic expression in the frequency domain that is indicative of its thickness in time. For example, a simple homogeneous thin bed contains a predictable and periodic sequence of notches into the amplitude spectrum of the composite reflection (Praptono et al., 2003; Figure 1). However, typically a seismic wavelet contains the information from multiple subsurface layers and not just one simple thin bed. The combined seismic response from these multiple subsurface layers usually results in a complex tuned reflection which has a unique frequency domain expression; in order to help resolve these thin beds, spec-decomp can be used. 

 Figure 1: Spectral decomposition is used to identify thin beds through analysis of the frequency spectrum in a short window around the time of the bed (Partyka et al., 1999).

As stated before, spectral decomposition can be used to break down the seismic survey into its component frequencies (Figure 2).  When determining which frequency to extract from the dataset, it's best to use a non-standard or octave scale in order to avoid potential harmonics (seeing the same information at multiples of its base frequency).  Thus, multiple datasets are created at these pre-selected, discrete intervals, e.g. 15.3 Hz, 29.6 Hz, 44.4 Hz and so on.  After determining these frequency intervals, each subsequent dataset produced via spec-decomp manifests only that particular frequency.

After all datasets have been produced, the reservoir interval of interest can then be scrutinized in greater detail.  This is carried out by capturing the seismic response at each frequency subset (15.3 Hz, 29.6 Hz, 44.4 Hz, etc.) - essentially, a "screen-capture" of the seismic image for each of these intervals can be input into an animated sequence from lower frequencies to higher frequencies, thus revealing spatial changes in stratigraphic thickness otherwise impossible to ascertain from the full frequency dataset.  Spectral decomposition reveals details that no single frequency attribute can match.

Figure 2: Note the different seismic response at 40 Hz as compared to 20 Hz; much more detail can be ascertained with the 40 Hz wavelet, however, the 20 Hz wavelet still manifests information about temporal bed thickness and the stratigraphic nature of the deposit.

*Nexen Petroleum U.S.A. Inc., Dallas, TX, U.S.A. (email: staffan_vandyke@nexeninc.com)

• Email This
• Print
• Comments
• Trackbacks
• Share Link