- February 28, 2017
- Posted by: Art Berman
- Category: The Petroleum Truth Report
It’s the beginning of the end for the Bakken Shale play.
The decline in Bakken oil production that started in January 2015 is probably not reversible. New well performance has deteriorated, gas-oil ratios have increased and water cuts are rising. Much of the reservoir energy from gas expansion is depleted and decline rates should accelerate. More drilling may increase daily output for awhile but won’t resolve the underlying problem of poorer well performance and declining per-well reserves.
December 2016 production fell 92,000 barrels per day (b/d)–a whopping 9% single-month drop (Figure 1). Over the past two years, output has fallen 285,000 b/d (23%). This was despite an increase in the number of producing wells that reached an all-time high of 13,520 in November. That number fell by 183 wells in December.
Well Performance Is Declining
Well performance was evaluated for eight operators using standard rate vs. time decline-curve analysis methods. These operators account for 65% of the production and also 65% of producing wells in the Bakken play (Table 1).
Estimated ultimate recovery (EUR) decreased over time for most operators and 2015 EUR was lower for all operators than in any previous year (Figure 2). This suggests that well performance has deteriorated despite improvements in technology and efficiency.
Figure 3 shows Bakken EUR and the commercial core area in green. The map on the left shows all wells with 12-months of production history and the map on the right, all wells with first production in 2015 and 2016.
Most 2015-2016 drilling was focused around the commercial core area. The fact that EURs from these core-centered locations were lower than earlier, less favorably located wells indicates that the commercial core is showing signs of depletion and well interference.
Well-level analysis indicates a fairly systematic steepening of decline rates over time. Figure 4 shows Continental Resources wells with first production in 2012 and 2015. 2012 wells have a shallow, super-harmonic (b-exponent = 1.3) decline rate but 2015 wells have a steeper, weakly hyperbolic (b-exponent=0.2) decline rate.
Oil reserves for 2012 wells averaged 343,000 barrels but only 229,000 barrels for 2015 wells–a 33% decrease in well performance. Steeper decline rates result in lower EURs.
In fact, a successive increase in oil production decline rates can be seen for all of the major operators evaluated in this study. Decline rates for 2014, 2015 and 2016 are higher than for previous years for these operators despite higher initial rates (Figure 5).
Gas-oil ratios (GOR) for most operators increased from 2012 through 2014 and then, decreased for wells with first production in 2015 (Figure 6).*
Changing GOR is important because it suggests decreasing reservoir energy. The Bakken has a solution gas drive mechanism. Initially, oil is produced by liquid expansion across the pressure drop from the reservoir to the well bore. Later, gas dissolved in the oil expands and this is the mechanism that lifts oil to the surface.
Rapidly increasing GOR in the Bakken probably indicates partial reservoir depletion and subsequently decreasing GOR suggests more advanced depletion accompanied by declining reservoir pressure, declining oil production and increasing water cut (Figure 7).
The sequence of events summarized in Figure 7 is demonstrated in Bakken field production shown below in Figure 8. Gas increased before oil production peaked in December 2014 and continued increasing through March 2016, and then declined.
Water cut—water as a percent of total liquid produced—has increased for most operators over time (Figure 9) and this provides additional support for progressive Bakken depletion.
Company Performance, Break-Even Prices and Future Drilling Locations
Well performance for the 8 key operators shown above in Table 1 provides a framework for company performance and break-even prices for the Bakken play.
Reserves were estimated for more than 4,400 wells with first production in 2012 through 2015 using standard rate vs. time methods. Decline-curve analysis (DCA) was used to evaluate wells with at least 12 months of production history for key operators. Production group DCA was done separately by operator and year of first production for oil, gas and water.
Results are summarized in the following tables.
None of the key operators’ average well breaks even at current Bakken wellhead prices of $42.50 per barrel although ConocoPhillips ($43.08 break-even price) is very close. EOG, XTO and Marathon all break even at prices less than $50 per barrel but other operators need higher oil prices to break even. It is worth noting that Bakken wellhead prices are about $10 per barrel less than WTI benchmark prices.
Current well density was calculated by measuring the area of the $50 commercial area (406,000 BOE cutoff) and dividing by the number of horizontal wells within that area. There are 5,500 producing wells within the 1.2 million acre commercial area shown in Figure 10. That equates to a current well density of 215 acres per well.
Tight oil operators describe infill spacing of 40 to 120 acres per well favoring the lower end of that range. Current well density in the Bakken core of 215 acres per well suggests substantial infill locations remain yet declining EURs, increasing water cut and falling GOR do not support further infill drilling.
The Bakken is unique because of the extraordinary lengths of lateral wellbores compared with other tight oil plays. Laterals are commonly more than 10,000 feet in length and often approach 12,000 feet.
Figure 11 shows lateral lengths in the Bakken. It is clear that within the commercial core area, most laterals exceed 8,000 feet. Available evidence suggests that current well density is sufficient to fully drain reservoir volumes. That implies that further drilling will not result in producing new oil volumes but will interfere with and cannibalize production from existing wells.
The Downside of Technology
The Bakken play represents the fullest application of modern horizontal drilling and hydraulic fracturing technologies. The Middle Bakken and Three Forks reservoirs are tight, naturally fractured sandstones that respond exceptionally well to long laterals and multi-stage fracture stimulation. Field rules allowed long laterals well before these were feasible in other plays.
The downside of efficiency and technology is that depletion has accelerated. Resulting higher initial rates masked underlying field decline that is becoming apparent only in wells with first production in 2015. The evidence for depletion is compelling but pressure data is not publicly available and is needed to complete the case.
The most appealing aspect of resource plays is their apparent lack of risk. Source rocks are the drilling target so finding oil and gas is given. Because the plays are continuous accumulations, there is no need to map and define a trap. Since the reservoirs are tight, seals are not an issue either. But commercial risk should be more of a concern for investors than it seems to be so far.
The downside is that there is no way to stay away from water and it is produced from day one in large volumes. The Bakken has produced 1.5 billion barrels of water along with its 2.2 billion barrels of oil over the decades. Where are they putting it and what does that cost?
Investors should be worried. As analysts cheered the resilience of shale plays after the 2014 price collapse, nearly a billion barrels of Bakken oil were produced at a loss–about 40% of total production since the 1960s.Vast volumes of oil were squandered at low prices for the sake of cash flow to support unmanageable debt loads and to satisfy investors about production growth. The clear message is that investors do not understand the uncertainties of tight oil and shale gas plays.
And all major Bakken producers continue to lose money at current wellhead prices. If observations presented here hold up, there may be nowhere for the Bakken to go but down. Higher oil prices may not help much because the best days for the play are behind us. Future profits were sacrificed for short-term objectives that lost the companies and their shareholders money.
The early demise of the Bakken should serve as a warning about the future of other tight oil plays.
*Statoil and Marathon depart somewhat from this general observation. GOR for these companies is lower than average and peaked earlier than most operators although Marathon’s GOR has been relatively flat.
Sincere thanks to Lynn Pittinger for his many useful comments during research for this post.
No need for the obsolete DAPL pipeline. Total waste of Taxpayers Money.
1. The reason for the drop in producing wells was weather (snow). Companies decided to shut in certain pads because the cost of keeping the road plowed did not justify keeping it online. This was well explained in the FEB NDIC webinar. You can see it clearly if you look at the production profile and see wells shut in. Those wells will come back on in May.
2. Comparing the cum curves, we don’t see wells getting worse. 20 month cums, thousands of barrels of oil:
2012: 106
2013: 116
2014: 120
2015: 135
For 2016, it is even better, although we don’t have 20 month cum yet. (but at 12 months were already at 106 M BO and whole cum curve was above the others.)
For 2016 (and even 2015), there is some high grading. Signified by the low number of completions. But still the most you can do is throw caveats to someone who touts the improvement. But you don’t get evidence of getting worse from that.
3. GOR is known to be locationally dependent. The center of the basin, which has the best rock on a BO basis, is also the gassiest. Looking at rig counts or well maps it is clear that we collapsed drilling into the core (which is economically rational). Perhaps completions are driving more gas, but you would need to do some analysis that holds location constant to prove that. And still, some of the GOR increase would be from the high grading. Age of wells also affects GOR, but of course this is irrelevant for generational comparisons at same time in life.
4. Water cut overall in the Bakken is good, even if it gets worse relative to previous. Much better than many other plays. Also, don’t ignore the impact of the higher volume completions affecting produced water in the future. Based on the rock they are drilling, given it is collapsed into the core, it is the opposite of moving into marginal acreage. Perhaps downspacing could have an impact. But they are not moving to the frings of the field, which is more watery. The opposite.
5. The main reason the Bakken is down is less drilling, because of low prices. It works a lot better at $100 WTI than at $50. So far, the arms race between “drill the best first” and “learning by doing” has been a standoff. Actually slightly won by the improvements.
6. Of course the Bakken will eventually be exhausted. That is the nature of mining. But still, who anticipated it 10 years ago doing what it did. And we have had continued development of new plays over the years. EF came in later than the Bakken. Permian later than those two. SCOOP/STACK have been relatively late. On the gas side, the Utica is now a substantial Ohio play and came in relatively late.
Source Shaleprofile.com
Anonymous,
The problem is in the rates of decline. Even though 2014, 2015 and 2016 have higher 12-month cumulative production, their respective EUR is lower every year because of steepening decline rates.
That’s why we do reserve forecasts and don’t just rely on cumulative production numbers.
Your optimistic views of the future of the Bakken is not based on deductive, systematic analysis of data. I wrote an entire data-rich post that you dismissed because you believe in a different outcome and found superficial data to support your belief.
All the best,
Art
Oh and DAPL will improve basis differential (and safety).
Rates seem to converge late in life. If you want to make the case that higher initial rates, don’t prevail later in life, then you have grounds.
But I don’t see clear evidence of crossing over. Which would be definitely required if EURs were t0 be getting worse, while initial production is higher and current cums are higher.
year rate-10 rate-20 rate-30 rate-40 rate-50 rate-60
2012 153 99 75 60 49 45
2013 176 104 76 60 NM NM
2014 173 103 73 NM NM NM
2015 209 112 NM NM NM NM
2016 196 NM NM NM NM NM
Source: Shaleprofile.com
Science is not about forcing data to fit a preconceived outcome.
The Bakken is following the same pattern that every oil field on earth has followed for the last 150 years. It is depleting. That is physics.
The data are very clear. Oil production is declining faster for each successive year after 2013.
Producers are using technology to increase initial rates but these methods also result in faster depletion.
Gas-Oil Ratios Are Increasing.
Water cuts are increasing.
I welcome your comments. However, if you want to argue that the earth physics of oil-field production and decline have changed and have made an exception for tight oil plays, please don’t do that here.
No one else in these comment pages presents themselves as “Anonymous.”
All the best,
Art
Art,
Good work. I am a retired Reservoir Engineer and have followed your work for some time now. Good Work. I am impressed when a Geologist crosses over and takes a keen interest in the (typically) Engineering side of things as it usually bodes well for encouraging the Engineers you work with to take a keen interest in the Geology which in my view is important to Reservoir Engineering and essential to Reservoir Modelling and Simulation.
Allow me to knit-pick a little bit. Arps showed fairly conclusively that if you have decline exponents going above one something may be wrong. In my work in various areas around the world, low-perm areas will often exhibit this, caused by transient flow. Decline curve analysis works best if only psuedo-steady state, boundary dominated flow is included in the analysis. Without a well validated model, identifying where one leaves off and the other begins is difficult and in multi-layered with highly-contrasting perm, in these reservoirs, fully boundary dominated flow may never completely occur. This all becomes somewhat academic as you further and further down the decline curve, as the true nature of the depletion reveals itself…..completely at abandonment.
You mentioned the pressure history is unavailable which is unfortunate as this would be very helpful. It should be possible to infer pressure performance by constructing a psuedo-PVT set using the API gravity, initial GOR, initial Reservoir Pressure, and whether or not it was undersaturated/saturated. But with the high API and probable volatile nature of this oil and if it was undersaturated ….. that may or may not work well. What may be more interesting is a 3-phase Relative Permeability analysis looking for a critical gas saturation at which point instead of a declining GOR(your fig. 7) later in the depletion performance, oil production falls off a cliff (becomes a non-moving phase) and gas production sky-rockets. Again, the very light and possibly volatile nature of the oil may dampen this.
Are there any lab generated Sor values available? In conventional Reservoirs, Sor is the key to roughly understanding EUR and ultimate recoveries less than this represent uncontacted/undrained reservoir rock (assuming we are willing to think of source rock in terms of reservoir rock).
Since the majors have to a large degree been absent these shale plays, I would think some similarities to the truly brilliant individuals (Bennet, Jones, Argawal, etc.)that I had the luxury of working with at Amoco Research may not have been applied to these reservoirs yet. Nonetheless, it appears to me you are on the right track with your analysis of the Bakken. There looks to now be enough performance history to Model this and validate the model. Do you know of any such modelling efforts?
Cheers,
I never argued that oil wells do not deplete. You are fighting a straw man there.
As far as the overall basin, it grew from the early 2000s through 2014, with a small hiccup when prices dropped in 2009. After 2014, it dropped because less wells were drilled. Less wells were drilled because world price had crashed. (Something shale oil helped cause.)
Of course it will eventually exhaust all economic resources even at $100. But as of now, there is no reason to think that the field declined from being drilled up. It was growing at over 200 Mbpd/year before the price crash. All kinds of credible sources, predicted further growth (absent the price crash).
Of course even in a $100 world, the field would eventually be drilled up. But there is no proof that this is why production dropped. Occam’s Razor says price drove the production down.
Oh and there is also some correlation of higher decline rates from higher initial production wells (has to do with over pressure, correlated to center of basin). However, this does not mean the EURs will be less.
If you want to criticize someone asserting higher IP leads to consistent higher production throughout the curve, feel free. But lower EURs? Burden of proof is on you, if you want to go there. (And the land prices in the Bakken sure don’t back you up–there is a reason people pay more for the center of the basin than the fringes.)
L Thurman:
The USGS has a good 2013 paper on their own method of decline curve analysis, behind their overall resource estimates.
Also, Google Scholar shows many academic papers written. I would suggest in particular the 2011 paper: “Impact of Geological Variation and Completion Type in the US Bakken Oil Shale Play Using Decline Curve Analysis and Transient Flow Character”.
I’m not sufficiently technically competent to judge the different efforts. But I do know that it is not some brainstorm to try a bunch of different models on the data. And even if we had the brightest guy in the room, we would still have to wait to see what happens in the end.
Art maybe the water cut is increasing generationally because they are pumping heavier fracs? I mean it is well explained that they put more water into the formations. I would think just having sent more water down the pipe would impact the long term flowback, no? Although I guess it could just be coming from more “rubbleization” also.
…. The modelling I was referring to at the end of my note was a gridded finite-difference IMPES numerical model, validated via a history match to well performance to date
Fantastic (and thorough) analysis. We’ve seen the same story play out in the Saskatchewan Bakken years ago. Once the treadmill starts, it’s impossible to keep up. Need secondary recov methods like WF just to moderate the decline (let alone grow production). Curious if the Author has looked at the EF or the Permian and if there are any distinct geological differences there that would make those two areas ‘buck’ this trend. In other words, is it only a matter of time before data like this starts to show the same conclusion in the Texas shales?
Dean,
I have evaluated the Eagle Ford and Permian many times. I suspect that the Eagle Ford is behaving similarly to the Bakken but that play is complicated by multi-well lease reporting so it is often impossible to separate EUR by year of first production. The Permian is in an earlier phase of maturity for tight oil plays.
All the best,
Art
L Thurm:
So? Why not look at the various models that have been attempted? Maybe it shows you something.
Great article! Sure seems like $50 is the new line in the sand for selling production. The last time the commercials were this short oil was in triple digits. I think everything oil is too bullish.
Technically, the oil sector has lagged the overall stock market quite a bit. Despite all the hard work the Saudis can’t get oil to budge towards $60. My guess is volatility will soon pick up for crude!
Art- Interesting analysis. Given the increasing water cut in these wells, is there a water table near the producing reservoir? Just wondering where the water is coming from if the depletion mechanism is solution gas drive (could there be some weak water drive also?). Could it possible they have some mechanical issues with some or many of these wells where casing leaks are causing some water encroachment? Given the extremely long laterals in these producers, it could be a challenge to effectively evaluate the integrity of the completions in these wells over time. One has to wonder if some of these long producing laterals haven’t collapsed especially if reservoir pressure has dropped significantly. Are these wells on artificial lift, either pumping unit or gas lift? If reservoir pressure is dropping, it seems these operators would be looking at a secondary recovery to maintain reservoir pressure and prolong production. Especially given the investment they have tied up in existing wells and infrastructure. Thanks
Jeff,
I don’t know where the water is from.
Most of the Bakken-Three Forks acccumulations have no traps (Sanish and Parshall were the initial Three Forks discoveries and are stratigraphic traps). After fracture stimulation, there is no way to segregate the produced fluids in the reservoir itself and it is pure speculation how much fluid contribution comes from adjacent horizons.
I doubt that secondary recovery on a micro-darcy, fractured reservoir would do much outside the fracture systems.
All the best,
Art
Great article art tho a lot of technical data that is difficult to follow. Any thoughts on the Permian? Success in the Permian is all you ever hear of these days. Also is deepwater oil dead or is that a viable answer to declining shale plays? Right now so much negative news on the offshore sector which has a fairly well established history. So much positive news for shale production.
Thanks for the comments! Love your articles
James,
The Permian basin tight oil plays are relatively less developed than the Eagle Ford and Bakken so there is more room to drill undrained locations. I wrote a fairly comprehensive post on these plays last year called Saudi Permian: A Race To The Bottom For Tight Oil.
All the best,
Art
Art-
It would be interesting to see pressure data from some of these infill wells to see if they are reaching undrained areas. It seems with this super tight rock, the effective drainage area would be somewhat limited beyond the reach of the laterals and the propped fracture extending from those laterals. Makes one think that the higher initial rates seen in the newer wells could be a function of advancements in the completion process these operators and service companies have most likely experienced over time. This is some what speculative, but if improvements in the completion process allowing for more effective fractures and less damage to the rock near these fractures, along with the limited drainage areas given the limits of oil flow through the tight rock beyond the reaches of the fractures, it may explain why at the higher initial rates, production decline is steeper for these newer wells. Would have to have the reservoir pressure data to know for sure. Am sure these operators are gathering pressure data in these new wells to better insure they are needed given the high cost of drilling and completing one of these monster drain holes. It is pretty amazing to think about fracturing and then propping that fracture along a 1.5 mile horizontal wellbore. Amazing stuff.
Jeff,
It would indeed be helpful to add pressure data to this study but it is not publicly available.
The higher initial rates in the newer wells appear to also have much steeper decline rates. This reinforces my finding that EURs are declining in more recently drilled wells which does not bode well for the future of the plays.
As I wrote in this post, the down side of improvements in technology and efficiency is that depletion proceeds faster.
All the best,
Art
[…] The Beginning of the End For The Bakken Shale Play […]
Hi Art,
Great to see your article on oil. Since your last article on oil, it seems we are getting closer to $50 and below every day. Luckily I’m not the ones that are holding their breath for $70+ oil :). Thanks for your insight. Have a great one.
Tanv
Tanv,
Thanks for your comments. Yesterday, prices fell a lot and are now below $50. I just posted a new article on that.
All the best,
Art
North Dakota was up 38,000 bpd in JAN. (Director’s Cut came out today.)
Director’s Cut webinar (only the audio is up now, hopefully slides up soon) has some response to this article. Worth checking out to get a different perspective than Art’s.
Art, thank you for this very thoughtful and informative article. I am a faithful listener of Macrovoices and look forward to your unbiased interviews (especially in a sector full of misinformation).
Although I am sophisticated accreditated investor, I am NOT a geologist or engineer. So forgive me if my question is light in tech-speak but more geared to the financial sell-side…but to me it sounds like by increasing IP at the expense of EUR the oil companies can increase immediate production while limiting the damage to the NPV 10 (or 15 or whatever discount rate we choose for time value of money) per well. Thus, they can increase immediate cash flows so they can sell equity or bonds to finance future production growth. This Ponzi scheme would then continue until the buyer of the new equity issue asks about actual profits rather than current production growth. (Just imagine what my company’s top line sales growth would be if I sold $1 for $0.95). Is this your take as well? Would love to know your break evens for Permian and EF (but I also understand that this is proprietary ?). Cheers
[…] Apparently, investors in American shale deposits have very short memories or they have not had enough punishment. They continue pour money into the Permian Basin located in Texas and New Mexico. The Permian is likely to be the only U.S. shale oil deposit that will see growth in oil production this year as low prices continue to take their toll on other shale plays such as the Bakken in North Dakota. […]
[…] oil production this year as low prices continue to take their toll on other shale plays such as the Bakken in North […]
[…] oil production this year as low prices continue to take their toll on other shale plays such as the Bakken in North […]
[…] oil production this year as low prices continue to take their toll on other shale plays such as the Bakken in North […]
Art,
In fig 5 you are saying: Decline rates for 2014, 2015 and 2016 are higher than for previous years for these operators despite higher initial rates… In my view, the strong decline of 2016 wells is exactly why they have high initial rates. There seems to be a trade off between high initial production rates and decline of these wells when they become older. So, in the media it sounds good when shale companies advertise higher performance of new wells, yet the statistic shows that exactly the high performance wells are the ones who decline faster in the future.
Heinrich,
That is exactly how I interpret the phenomenon. It is the downside of technology–better technology is a bigger spigot. It allows higher rates and faster depletion.
All the best,
Art
[…] Resource Art Berman […]