Contemplations over the sinking Bakken
A month ago I published an article on the future perspectives of the shale oil and found them not too bright. I couldn’t go into detail in terms of that article, so I had to leave this task for later. The time has come and today I intend to focus on the biggest shale oilfield Bakken. But first I’d like to make a remark that I am writing for the general audience of Russian readers, and so I ask knowledgeable specialists not to reproach me for mentioning some truths well known to them.
1. RESERVES AND PRODUCTION
In terms of territory (520 thousand square kilometers) Bakken is the second largest oil field in the world, being just slightly smaller than the territory of the mainland France or Khanty-Mansiisk district. Oil-bearing formations are found in the territories of the American states of North Dakota and Montana as well as in the Canadian Saskatchewan and Manitoba. However, even in the better times the production in Canada did not exceed 9% of the total, and in the state of Montana the reserves have long ago been depleted and currently produce less than 300 t/d. Over 90% of the current production is accounted for by North Dakota; the oil-bearing formations in its territory cover 68 thousand square kilometers, which is 13% of the total oilfield’s territory.
The cross-section of Bakken formation shows three productive layers (pic.1). The main layer, the Middle Bakken is up to 40 m thick and consists of very tight sandstone with shale and dolomite inclusions. Above and behind it there are shales 5 to 15 m thick.
Pic.1. Bakken formation schematic cross-section
Due to this structure, which looks like a mince pie, WONDERFUL CONDITIONS for multistage hydraulic fracturing developed in Bakken. The thick bed between two plastic layers is a perfect place for a horizontal well. In the plots where the thickness reduces to 10-15 m and the inflow is small no hydraulic fracturing would help. Picture 2 shows that the formations more than 18 m thick take less than a half of the territory, the East and the South parts are thinner. Running ahead, I’ll say that these parts are hardly drilled.
Pic.2. Isopach map, lines of equal formation thicknesses
Below the main formation there are partially oil-bearing layers Sanish and Three Forks, but they contain much less reserves.
The initial Bakken reserves of oil in place are huge. Even though oil occupies just 5% of the rock volume, there are 22 billion of it in North Dakota alone. Over the past years 240 mln tons were recovered, so the oil recovery factor is (note!) 1.1%.
On the other hand, the rock permeability varies within 0.02 – 0.6 mD, which is hundreds and thousands times less than in rich oilfields. As I wrote earlier, most of the wells are drilled in the zone of natural fracturing, approximately in the center of Bakken. Now let’s add some figures to support this statement.
The North Dakota authorities publish detailed reports on all the wells for each county separately. For a start, let’s calculate how much oil is produced from a square kilometer daily in different counties. The results are shown on the map in the picture 3, the other details are presented in the table 1.
Pic.3. Specific oil recovery per unit area
Oil and gas production indicators
It turns out that 91% of all the oil is produced in four counties: McKenzie, Mountrail, Dann and Williams. 11,271 producing wells (73% of the well stock) are drilled here and they operate with a good average production rate of 8-13 t/d. Note that the total territory of these counties covers 23.4 thousand square km. This is a lot, but still it is 22 times less than the total Bakken. It’s not as big as France, it can only compare to Israel. 5 more counties account for 8.2% of the production, while in the other 7 counties it is insignificant. Almost 32% of the wells there are idle.
2. GAS INSTEAD OF OIL
In-situ under the abnormally high pressure of 430-460 bar every ton of Bakken oil contains 70-150 cubic meters of associated gas. After a well is put into operation, the pressure around it drops, and when it reaches 180 bar, the gas starts to separate. Since the viscosity of gas is lower than that of liquid, its share in the production grows and the share of oil declines.
The dynamics oil and gas production as well as the change of gas-oil ratio are shown in pic.4. One can see that as long as 10 years ago three times over the required amount of gas was produced there. It’s no wonder that the average oil production rate amounted to the modest 4 t/d back then.
Pic.4. Oil and gas production change
Intensive drilling in 2009-2014 added 8 thousand new wells, which led to the gas-oil ratio decline, followed by 60% increase in the last 3 years. Its peaks (450 cub.m/t) were reached in McKenzie and Williams. Further increase of gas production is restricted by the fact that when the gas-oil ratio reaches 700-1000 cub.m/t, the oil production rate is very low and the wells are usually shut down.
The gas-oil ratio increase is associated with the oil density decrease. Heavier components are suspended in the rock pores, while lighter ones are transported by the gas. This is the reason why the American WTI is 5-7% lighter than the other brands. Bakken oil is volatile and explosive, so it has to be carefully separated and allowed to settle before transporting. Violating these requirements may have been one of the reasons for the explosions during the railroad accident in Lac-Megantic, Canada. All these factors make Bakken oil considerably cheaper; in December a barrel of it cost $40.5, which is 20% cheaper than a barrel of WTI.
Though, all these are, so to say, minor difficulties, the major trouble is ahead. Judging by the increase in gas production it is possible to roughly estimate the near wellbore pressure; I estimate it at 40-60 bar. With these pressure values at the depth of 2500-3000 m the rocks compressive strength gets exceeded, followed by the distortion in the cement bond and the surrounding formations. Fractures developing around the wellbore often cause cross-flows of the formation water coming from other layers.
3.WATER FLOODING FROM ALL SIDES
Bakken is being flooded. Over 98% of the wells there produce water with the oil. The average watercut in the field amounts to 61%.
At first I couldn't believe it. Since there is no aquifer in the main bed, the Middle Bakken. Of course, a quarter of the wells produce from other layers, which contain water, but they make no difference. Obviously I have seen faulty wells with formation water cross-flows many times before. But I can hardly imagine that all the 10 thousand wells have such failures. I double-checked the tables and found no mistake. Last year 66 mln tons of water were produced for 50 tons of Bakken oil. And there are three serious reasons in favour of its external origin.
Firstly, even with the normal frac jobs approximately half of the wells experience water flows from the surrounding formations in a month or two. And here the formations are fractured dozens times before putting the wells into operation.
Secondly, the low bottomhole pressure favours water cross-flows. The pressure in the target formation is 50 bar, but in the overlaying it is 300 bar. Huge differential appears along the casing string and the rocks often can’t sustain it.
Thirdly, it is Bakken, where water flows remain stable for many months, as if the devil opened a tap for an inch. Only it’s not the devil, but another layer, which is connected to the well via small defects in the cement behind the casing. It may get bigger with time.
I’ll use this example to show the wells water flooding (pic.5).
Pic.5. Oil and water production dynamics in the wells of McKenzie county
The first well (the top chart) was put into operation in August 2012 with a wonderful oil production rate of 95 t/d. In three months' time a stable water flow of 15-30 t/d established. The oil production rate dropped to 6-7 t/d during the following 1.5 years.
In April 2014 the owner stimulated the well. As a result, the oil production grew twice, but the water production – 8 times, to 116 t/d. During the following 2.5 years the well operated with 60-90% watercut. By now the oil flow has run out and amounts to about 4 t/d.
Remarkable is the fact that the second well drilled in the same block (bottom chart) got 95% watercut at once. The water production reached 150-175 t/d, while the oil production did not exceed 5 t/d. Statoil drilled the total of 4 wells in this small block, all of which started producing 60% water at once; two of them have became completely water cut and are currently idle.
If you think about it, this is something to be expected. McKenzie county is drilled though the length and breadth, there are 4,157 wells there, one per every 1.78 square kilometres. They are horizontal wellbores up to 3 km long. Each has 20-30 artificial fractures up to 100 m long around it. The deposit has a huge net of high permeability with the pressures varying from 20 to 460 bar. With such gradients water can flow there easily and freely.
I won’t be surprised to find out that the formation water, which is injected into the Sanish or Longpole layers on the one bank of Missouri, is then produced from the Middle Bakken on the other bank. I happened to see even trickier zigzag courses of water. In the large Vatinskoe field near Samotlor geologists once decided to trace the injected water using a color indicator. They were extremely surprised to find it in two days’ time (!) in six other layers (!!!). Though it was never found in the target formation…
The producers in Bakken are slowing down the recovery because of the fast watercut. Last year the oil production declined by 15%, water production by 7%, the average watercut rose from 59.3 to 61.3%. To conclude this brief analysis we have to answer one last question:
4. WHERE SHALL WE DRILL FURTHER?
Look at pic.6. It is very representative, and it is not the first time I demonstrate it. This is the well-spacing map for a small area in Montrail county. Where can you drill here?
Ðiñ.6. Well-spacing map for an area near Parshall city
In the right part the zone seems to be non-productive. The only area, which has not been drilled yet is the territory of the big running-water lake, but the permission to drill there is extremely hard to get. Besides, this is the territory, where Indians live.
If we go back to the table 1, drilling in the Divide county will look tempting. There are 824 wells there, 90% of which operate with a reasonable production rate of 5.4 t/d. But after I calculated that the average watercut has already reached 70%, the idea to drill there lost its appeal at once. The American producers don’t seem to be eager to drill there as well: the active rig count grew by just 7 units in the recent months and now amounts to 37. This is 6 times less than 5 years ago during the shale gale.
There are no reasons to suspect that drilling the flanges of Bakken will provide more oil than the center. On the contrary, the entire data available (table 1) indicate that the oil-bearing capacity there is 3-5 times, in some places dozens times less. Producing that oil may become possible many years from now, using some new yet-to-find technology.
Bakken oilfield with its huge initial reserves has nevertheless reached its final development stage. There is every indication of the process: the decrease in oil and fluid production, the increase in the watercut and the idle well stock.
The main reason for the depletion is the progressing water flooding. The bottom-hole pressure decrease caused external formation water break-thought in all the wells. Stopping this process doesn’t seem possible with the current level of technology.
The rise in the oil prices that has happened has only slightly influenced the scope of drilling. The most productive zones of Bakken in the four counties have already been drilled to the full, and drilling outside them is fraught with the risks of low production rates and accelerated water flooding.
6. SOME SAD CONTEMPLATIONS IN CONCLUSION
The information I quoted here is available in generally accessible sources. The depletion of Bakken wells due to gas production growth was known 10 years ago. The progressing water flooding was identified in 2012. How come nobody speaks about it? Three years ago at R. Patterson’s forum there was a little discussion and that’sit. At that time this problem looked like young piggy, but now it has grown to be an enormous boar.
It’s time to act. The injectors need to be checked for leaks, the formation water flows need to be traced. The connection between the hydraulic fracturing and the cement distortion has to be defined at last. The service companies will surely raise a cry about it and will try to prove that it wasn’t their fault. It is their main source of income and they are going to fight for dear life. But one can collect non-biased statistics data.
Why aren’t these questions discussed by the federal geologists, the API, the SPE engineers? Why is Art Berman the only one to speak about the obvious failures in the shale technologies? Is it because it’s easier (and more profitable) to run with your eyes closed after the Red Queen or to praise the naked Emperor’s clothes then to object and defend your opinion? If so, I am ready to show the way out of this sad situation. All it takes is to blame it all on the Russians.
The Russians did it!
It was the Russian hackers who got into the well managing computers and prevent them from working for the good of America!
It is the Russians who pour water in the wells!
And all this dirt about Bakken is written by a Russian, and not for the sake of the petroleum science, not for the verity, but for the purpose of propaganda by the special order!
You are not happy with this option? Then all joking aside, there is a lot of work ahead.