Enno Peters
This is an older blog post, you will find one on more recent data here
These interactive presentations contain the latest oil & gas production data from all 14,469 horizontal wells in North Dakota that started production since 2005, through January.
January oil production in North Dakota was unchanged from the month before, at 1.4 million barrels of oil per day. In January, which is typically a slow month, just 85 wells started production.
The growth in natural gas production has been steeper in the past few years. Compared with January 2015, natural gas production rose by 88%, versus 18% for oil. The reason for this is that almost all wells experience a rising gas oil ratio, and even stronger for newer wells.
In the ‘Well quality’ tab, you’ll find the production profiles for all these wells. After several years of improving initial well productivity, the 2018 vintage eked out another small gain.
All 5 leading operators in North Dakota started the year at a higher production level than a year earlier (“Top operators”). Continental Resources was the first operator in the history of the state to reach 200 thousand barrels of oil production capacity in January. It doubled its output in the past 2 years.
From our analytics service (Professional), we can see how Continental Resources has changed its completion practices in the last couple of years:
In this dashboard, we can see that Continental Resources did not change the length of its laterals by much since 2013 (yellow curve), but it did almost quadruple the amount of proppant used, from 3 million pounds per completion in 2013, to 12 million pounds in 2017/2018 (shown by the pink curve). The impact that this had on the amount of oil recovered in the first 12 months is shown in the plots on the right side; the bottom plot shows the same information, but now normalized by lateral length (1,000 feet).
The ‘Advanced Insights’ presentation is displayed below:
This “Ultimate recovery” overview shows how all these horizontal wells are heading towards their ultimate recovery, with wells grouped by the year in which production started.
The almost 1,800 horizontal wells that started in 2012 have now recovered just above 200 thousand barrels, and are now producing at a rate of 40 bo/d, on average. The 971 wells that started 5 years later (2017) are, with an average recovery of 175 thousand barrels of oil after 14 months on production, not far behind, and they are still operating at a rate of 227 bo/d.
Early next week we will have an update on gas production in Pennsylvania, which just released January production data as well (already available in our subscription services!). It just set another record at over 18 Bcf/d.
For these presentations, I used data gathered from the following sources:
- DMR of North Dakota. These presentations only show the production from horizontal wells; a small amount (about 40 kbo/d) is produced from conventional vertical wells.
- FracFocus.org
====BRIEF MANUAL====
The above presentations have many interactive features:
- You can click through the blocks on the top to see the slides.
- Each slide has filters that can be set, e.g. to select individual or groups of operators. You can first click “all” to deselect all items. You have to click the “apply” button at the bottom to enforce the changes. After that, click anywhere on the presentation.
- Tooltips are shown by just hovering the mouse over parts of the presentation.
- You can move the map around, and zoom in/out.
- By clicking on the legend you can highlight selected items.
- Note that filters have to be set for each tab separately.
- The operator who currently owns the well is designated by “operator (current)”. The operator who operated a well in a past month is designated by “operator (actual)”. This distinction is useful when the ownership of a well changed over time.
- If you have any questions on how to use the interactivity, or how to analyze specific questions, please don’t hesitate to ask.
News, trends and data for the US upstream industry
Novi Energy Newsletter
30 Comments
Bakken will be limited in production with this high NG production. The NG capture rate is well below 88% which is State mandate. Lynn Helms says this.
Gas capture, workforce, and competition with the Permian and Anadarko shale oil plays for capital continue to limit drilling rig count
Interesting to see the continued improvements in well quality internal to the Bakken.
Lateral length (2MI) and spacing (660ft) have been tied in for a while… so we’re seeing outright improvements due to well design. It also doesn’t look like newer wells are systematically stealing from older wells (we see gorgeous transient behavior on the older well vintages. If these wells were losing reserves/resource, you’d be seeing much higher declines).
I would have thought that the play would be in it’s later innings at this point. This is usually associated with diminishing well performance and capital efficiency. Doesn’t appear to be the case.
I would somewhat disagree. There are couple of signs the new wells are not really better than the old ones. Gas/oil ratio is rising much faster (sign of faster depletion) with newer wells (2016 onward), and indeed the oil decline rates are steeper. So newer wells might not recover any more oil than the old ones, despite having larger fracs and higher IP rate.
That’s a valid item. Although, I’ll add that my interpretation of well quality wasn’t simply rooted in IPs or 1st year CUMs.
The log-oil vs cumulative plot is a very valuable diagnostic plot and the one from which I was inferring behaviors.
— Observe how well performance (by vintage) has shifted up and to the right. The slope doesn’t appear to be materially steeping or creating “concave down” behavior.
— This character is very important because it can be related back to flow regimes. If the operators were simply getting higher IPs without associated EURs, the slope would be steepened to trend to the same final CUM. Concave down behavior is particularly important because it implies that the wells have transitioned away from transient and into boundary dominated flow (terminal decline).
— The newer vintage wells appear to be exhibiting similar transient flowing attributes. GORs are higher, but this makes complete sense in the context of more aggressive artificial lift practices. Older wells were barely lifted, while newer wells often have aggressive gas lift setups or even EPSs. This aggressive behavior may result in accelerated recovery (versus outright increases in recovery). However, such changes are not yet indicated in any fashion in the log oil vs. cum curve.
The continued strong transient behavior in the Bakken puzzles me and makes me wonder how much oil they’re “stealing” from the other horizons. The Bakken member is relatively thin and there’s only so much oil.
For reference, Eagle Ford wells often hit terminal decline 36-48 months into their well lives. This compares to the Bakken where wells are still expressing strong transient (“infinite acting”) behavior 5-10 years into their well lives.
From the Enno’s graph it is easy to see that bo/d from recent vintages falls to levels close to older wells at around 36 months of production. It is interesting to see if the production rate actually falls below those levels in the future.
What I haven’t seen discussed before, is if you fit the natural logarithm curve a*i*ln(1+b*t), where i is initial production (I used 12M cumulative) and a & b are coefficients, to actual cumulative production curve, it fits very nicely and you get interesting results. Surprisingly, you can use the same coefficients to model Bakken and Permian wells (but not EF wells). You will also discover that new Bakken & Midland county wells do not fit the same curve. They decline faster.
?????
Looks like totally inverse of what you say to me.
And here the flow vs cum chart for the same wells, now by quarter to see more granularity and recent data.
Mikko, Stefan, Jim,
Interesting discussion.
Here is a comparison of the 2009, 2016 & 2017 vintages. I’ve removed wells that have been refrac’ed, as they introduced an upward bias to the 2009 wells.
First flow vs. time (download to see full resolution):
If we assume the Bakken is a 9 Billion oil field which it looks more and more likely that it is. The Bakken has produce around 3 billion barrels to date so we should start to see well degradation after the Bakken reaches the half way point.
So in 2019 and 2020 the Bakken will probably average around 1.45 million barrels a day. That is 1.1 billion barrels.
So we should start to see well degradation around the end of 2021 when half the 9 billion has been produced.
If the producers want to hold the Bakken plateau around 1.6 million barrels a day after 2021 they will have to spend a lot more money after 2021.
Oct. 30, 2018
Continental doubles its estimate for oil recovery from North Dakota’s Bakken shale, claiming 30B-40B barrels of the 250B barrels of oil in place will be recovered instead of the 20B barrels it estimated in 2011.
I have never understood how Continental could make the 30 billion barrel claim.
If we assume 40,000 locations then the average well would have to produce 750,000 barrels of oil. I just don’t see that.
My question is where is the SEC when we know these claims just can’t be true.
Look at the first 10,000 wells drilled in the basin there is no way they will get to 750,000 barrels oil.
That is why I think it is more like a 9 to 12 billion field with the average well producing like 300,000 barrels over their life. Because we know the last 20,000 wells will not be as productive as the first 20,000
Alden, how do you get to 40,000 locations?
That is what the North Dakota oil and gas commission thinks operator will drill in the Bakken.
This screenshot shows the performance of the 1,147 horizontal wells in the two fields where it all started in North Dakota : Sanish & Parshall (since 2007). It’s interesting to see that here all the improvements in well design never got well performance back to the original level.
Enno,
Don’t understand why the Apple of Shale oil, EOG did so poorly after 2010 in the best and oldest play, Bakken? or should I ask why EOG did so well before 2010 in Bakken?
CLR drilled very few wells in Sanish and almost zero in Parshall?
CLR’s post 2013 wells in Sanish are decent and passing 250/300K at close or over 100BOPD still decent rate.
It’s indeed remarkable how well these early wells performed in this area. You will not find many areas in the US where a whole vintage has recovered on average close to 500 thousand barrels (or is on track to do so), even now.
Also interesting is to see how the GOR has changed here in the past few years (see bottom right plot, which shows the GOR vs. cumulative production)
This one is for only EOG in this area.
Aldon, maybe CLR includes NG?
No they said 30 to 40 billion barrels of oil. I can’t believe they made that claim. It is very clear it will never get to 30 billion barrels unless there is another technology leap.
The maths I haven’t seen anyone doing is how many wells are required to drain the attractive parts of these plays (besides the oil producers themselves, but their incentive is to be optimistic). Well saturation of sweet spots will obviously have major consequences. It’s tough as an outsider, because some wells will be drilled too close, leaving more space for others, but I think estimation is worthwhile.
I’m definitely not a geologist, so I invite anyone with better insights to improve on the following, but this would be my approach for the Bakken. I’ve been told that it takes one well to drain about one square kilometer (about 250 acres) in the better areas of the Bakken. The sweet spot appears to be within Dunn, McKenzie, Mountrail and Williams as all but 2.7% wells were completed there last year. Furthermore, the area that was drilled was very roughly the size of McKenzie county in the Three Forks and, being generous, 50% bigger for the Middle Bakken. (see cumulative production ranking > show ranking by formation > select year 2018 > see map). McKenzie county is about 7500 square kilometers. That means 7500 spaces in the Three Forks layer and 11250 in the Middle Bakken.
These four counties currently have 4036 wells in the Three Forks and 8000 in the Middle Bakken. That leaves about 6750 locations. Last year 1233 wells were added in these counties (though I’m guessing the rate will increase this year). At that rate there are five and a half years of completions remaining, likely skewed towards operators that have been less aggressive/have more land. This is inherently imprecise, but I doubt it is off by some large factor. Especially given that activity is still increasing, I’m comfortable betting that the economics will look significantly worse in ten years. That is not priced into stocks (e.g. Whiting, >18x FCF) or reflected by low international activity, which seems to assume plays like these will be offsetting declines elsewhere.
Aldon, the SEC would not otherwise be responsible for confirming or denying Power Point (4P) reserves spewed forth in press releases or investor presentations. There is nothing to stop these shale guys from saying pretty much whatever they want to say. Not even pride. Same with 750K EUR’s.
There was an interesting tweet today from an old SEC hand (responding to Berman) who said when impairments required 30% or more cuts he was never allowed more than 5%. We are all being duped, badly, regarding shale oil reserves. “Up spacing” in the Permian SHOULD knock proximity based PUD reserves into the dirt. Betcha it won’t. Its pretty tough to believe anything these days, except realized production data. Thanks for that, Enno!!
shalers have to pray that similar miracles happened in Marcellus which defied Art’s prediction to happen again in Permian shale oil.
talking about heavier proppant loading improving production, it is actually the opposite that started the shale revolution. Conventional wisdom still can not understand it, and only spending billions to beat the odds the opposite way that roughnecks helped to discover.
Sometime in 1992, Ray Walker with Union Pacific Resources operating in Austin Chalk ( Giddings, TX) have been using gel-sand fracing (they used the correct spelling then which does not sounds too franKenstein). They pumped in mostly gel with only 10~20% water and sand that is 3 to 10 times more than today’s heaviest sand load in laterals. The cost is about $400K/300ft then, or about $3million/300ft in 2019 dollars. There is no way it will work out as today’s fracking. One night, Ray’s roughneck overlooked a faulty water valve/meter, and before the roughneck discovered and reported to Ray, the water is all pumped down with little sand and gel. Ray had to go ahead with flow back and was shocked to see the amazing results! This is the start of “slick water” revolution. It later helped reducing completion cost in Barnnet shale by 80~90% while increasing production 2~5x. To emphasize the wildcat nature of the discovery and also the contribution of roughnecks, they gave the famous SPE paper “Proppant, we don’t need no Stinkin’ Proppant”, a quote from 1948 movie “The Treasure of Sieera Madre”.
Ray Walker who later was COO for Range Resources and was flew to Pittsburgh to start the grand Marcellus revolution, and he said he felt it’s stupid to fly to Pittsburgh in TX.
here is a comparison of the fluid and sand load in Nick Steinsberger/Mitchell Energy and today’s GENERATION 1-3…
you could read more here,
https://www.linkedin.com/pulse/shale-revolution-arent-when-you-judge-conventional-wisdom-sheng-wu/?published=t
Sjoerd, great explanation of recoveries and remaining inventory in the Bakken. Makes total sense to me–how you arrive at the five year inventory and explains why rig count has been less upwardly mobile than in other basins. Engineered completions that use diverters and vary application of sand across the length of the lateral seem to be what all the public companies are pointing to as means to expand the economics of tier 1 acreage. Northern mentioned some strong recent results in Billings recently–not much has been going on there recently. Enno’s viz on EOG’s acreage probably explains why they drilled 7 wells in October, one in November and none since then in the Bakken. The stocks work in a rising price environment with new pipes in place but big public players except for WLL all moved out of the basin for growth. What wonder is when some of these privates get taken out, Kracken, Bruin, Slawson, Nine Point are drilling some very decent wells–not as good as Marathon, Hess and WPX…but pretty close.
http://www.clr.com/operations/bakken
Continental claims that they could drill 20 wells in 2 square miles unit.
With 4 in middle Bakken and 12~16 (3 benches actually) in the Three Forks, this really assumes the great potential of Three Forks, and so far from Enno’s plots, Three Forks have been underperforming compared to MB.
Thanks Ray. @Nuassemby, do you have any insight into which benches of three forks are actually being completed/produced? Easy for oil companies to just multiply by 3 for their presentations but I’m not sure to what extent that is representative of reality.
ran into this presentation here about total oil and gas in place and recoverable in Bakken
It seems that the work ignored the fact that in 2017 when growth is already pickup again for Bakken, and close to 2Billion BO already produced, and they gave a P50 number at high porosity of only 4Billion recoverable oil.
Based on the data here, it would be reasonable to assume that 7~10Billion of recoverable oil is quite possible.
Continental’s proven reserve in Bakken is about 1Billion BOE, and it produces about 1/7th of the total Bakken BOPD right now. As proven reserve is more reliable and strict than PPT reserve, it’s safe relatively to assume there is another 5~7 Billion of recoverable oil besides the 2 Billion more BO already produced.
Also, Continental has about 1million acres in Bakken, and if every 2 square miles has 10 wells (MB+TF123) at EUR of 300K BO, then we are looking at 2.5Billion BO; or if EUR is close to 500K BO, then we are looking at over 4 Billion BO. Again, if multiply by 7, we have recoverable oil at 17 Billion and 28Billion respectively.
Looking into the detailed numbers in the report above, it seems that they have a oil in place of only 59Billion, while Continental above is talking about 250Billion BO in place — even larger than the total oil generated that the report above used.
Many good points. thanks. what have been said in PR materials, including in result conference call slides are quite different from 10-Ks. Take CLR, I plucked the data concerning North Dakota Bakken only.