In this post I present an update to my previous posts over at The Oil Drum (The Red Queen series) on developments in tight oil production from the Bakken formation in North Dakota with some additional estimates, mainly presented in charts. The expansion is much about the differences between wells capable of producing, actual producing wells and idle wells (here defined as the difference between the number of wells capable of producing and the number of actual producing wells).
Figure 01: The chart above shows monthly net additions of producing wells (green columns plotted against the rh scale) and development in oil production from Bakken (ND) (thick dark blue line, lh scale) as of January 2000 and as of October 2013. The 12 Month Moving Average (12 MMA) is also plotted (thick dotted dark red line, lh scale).
There is still noticeable growth in tight oil production from an accelerated additions of producing wells.
- For October 2013 North Dakota Industrial Commission (NDIC) reported a production of 877 kb/d from Bakken/Three Forks.
- In October 2013YTD production from Bakken/Three Forks (ND) was 775 kb/d.
(It is now expected that average daily production for all 2013 from Bakken (ND) will become around 800 kb/d. - The cash flow analysis now suggests less use of debt for manufacturing wells for 2013.
Major funding for new wells now appears to come mainly from from net cash flows.
kb; kilo barrels = 1,000 barrels
MODELLED VERSUS ACTUAL PRODUCTION
Figure 02: The colored bands show total production [production profile for the “2011 average/reference” well multiplied by the net number of producing wells added during the month] added by month and its projected development (lh scale). The white circles show net added producing wells by month (rh scale). The thick black line reported production from Bakken (North Dakota) by NDIC (lh scale).
The chart also shows forecast developments for total oil production with respectively 1,500 (red dotted line) and 1,800 (light green dotted line) reference wells added annually through 2013 and 2014.
The model was calibrated to start simulations as from January 2010.
NOTE: The chart shows the models forecast towards 2025 from the population of producing wells as of October 2013. Producing wells will continue to be added, thus actual future production will be higher.
Any divergences developing between modelled and actual production may over a period of months give early indications about directional changes to average well productivity.
The forecasts with 1,500 and 1,800 “2011 reference” wells added through 2013 and 2014 (with baselines of Janaury 2013) also serves as additional references that may be indicative of directional changes in average well productivity.
If the model over time develops a growing deficit against actual reported production, this could suggest that newer wells have an improved well productivity relative to the reference well and vice versa.
The chart shows a deficit between modelled and actual production during 2010 which also demonstrates higher average well productivity in 2010.
The model estimated a decline in total production from the wells capable of producing in October to November to around 40 kb/d and that it now takes the net additions of 120 – 125 reference wells per month for the next few months to sustain the October production level. This is exclusive of any seasonal effects of increase in the number of wells shut in (idled) due to weather related causes.
For what it is worth the model estimated that total production from the producing wells (Bakken (ND)) as of October 2013 to October 2014 (year over year) would decline by 318 kb/d or 37%.
POSSIBLE EXPLANATIONS FOR DEVIATIONS BETWEEN MODELLED AND ACTUAL PRODUCTION
Figure 03: The chart above shows on a relative basis (blue columns) how monthly modelled production compared to actual (100% is a perfect match between modelled and actual).
To smoothen any seasonal swings from seasonally (temporarily) shut in wells a 12 Month Moving Average (12 MMA) has been added (dark red line).
How to read the chart: If modelled total production is higher (above 100%) than actual, this may suggest that newer wells added on average has a “poorer” productivity than the “2011 reference” well and vice versa.
A likely explanation for the recent deviations is that during winter a relatively higher number of wells get shut in and are not brought back to flow before the spring. The model does not recognize these events and continues to add flow from these wells wherever they were expected to be on their flow curve. This may cause wells added during winter to appear as slightly poorer than they really are as the model sees less flow from the added wells relative to the reference well. Conversely as the situation allows (coming out of winter), the shut in wells are gradually brought back to flow, the model will not recognize this, and new wells added may therefore look somewhat better than the reference well.
Wells in conventional oil reservoirs tend to flow somewhat better for some time when started up after being shut in (rested) for some time. Wells in tight oil formations should be expected to exhibit a similar behavior and thus have temporarily improved flows (gush) when restarted.
It is difficult to predict at what point in a well’s productive life it becomes temporarily shut in, when this happens and how long the shut in will last. It is possible to use actual data to make an approximate model that describes the distribution of wells in time and durations and thus get an estimate of how this affects total production.
Winter, due to accessibility issues, appears to cause an increase in the number of wells shut in and shut in periods could be as long as 3 to 4 months. This winter 4 – 500 wells are expected to become subject to seasonal shut ins. With time and as the total number of wells capable of producing grows, it should be expected that a higher number of wells may be subject to seasonal shut ins, ref also the section “IDLE WELLS” below.
From GrandForksHerald:
“It’s pretty significant when you think about 400 or 500 wells being shut off for three to four months,” Helms said.
Other effects may be that in some months the average for the added wells are somewhat better than the reference well, and other months somewhat poorer. However if a trend in any direction develops over several months, this may suggest a change in well productivities relative to the reference well used for the model.
ESTIMATES ON NET CASH FLOWS FROM WELL ACTIVITIES
Figure 04: The chart above shows an estimate of cumulative net cash flows post CAPEX of tight oil from Bakken (ND) as of January 2009 and as of May 2013 (red area and rh scale) and estimated monthly net cash flows post CAPEX (black columns and lh scale).
The assumptions for the chart are WTI oil price (realized price), average well costs starting at $8 Million in January 2009 and growing to $10 Million as of January 2011 and $9 Million as of January 2013. All costs assumed incurred as the wells were reported starting to flow (this creates some backlog for cumulative costs as costs in reality are incurred continuously as the wells are manufactured) and the estimates do not include costs for completed non- flowing and dry wells.
Economic assumptions; royalties of 15%, production tax of 5%, an extraction tax of 6.5%, OPEX at $4/Bbl, transport (from wellhead to refinery) $12/Bbl and interest of 5% on debt (before any corporate tax effects).
Estimates do not include any effects of hedging, dividend payouts, retained earnings and income from natural gas/NGPL sales (which now and on average grosses around $3/Bbl).
Estimates do not include investments in processing/transport facilities and other externalities like road upkeep etc.
The purpose with the estimates presented in the chart is to get an approximation of net cash flows and development of total debt used.
The chart suggests that so far in 2013 capital expenditures (CAPEX) for added producing wells has mainly been financed from the net cash flows from the existing population of producing wells. Use of debt appears to have noticeably slowed. As these estimates are for all added wells in the Bakken (ND) it should be expected that there are differences between the operating companies.
In 2013 an estimated $20 Billion may be used for manufacturing tight oil wells and construction of associated facilities in North Dakota.
THE 4 COUNTIES WITH THE BIGGEST PRODUCTION
Here follows a closer look at tight oil production developments from Dunn, McKenzie, Mountrail and Williams which now are the 4 counties in North Dakota with the biggest tight oil production. NDIC is now reporting oil production from 17 counties and the 4 counties with the biggest production now has around 85% of the total oil production in North Dakota.
- As from 2008 and through 2012 Mountrail was the county that provided the biggest portion of the production from Bakken (ND).
McKenzie is now the county with the biggest production portion from Bakken (ND), refer also figures 05 and 06. - The major portion of growth in tight oil production from the Bakken is now expected to come from McKenzie county as the other counties now have slower growth, refer also figures 05 and 07.
- Mountrail with prolific fields/pools as Alger (refer also figure 09), Parshall, Reunion Bay, Sanish (refer also figure 10) and Van Hook appears to have a slow down in production or are in decline.
- Mountrail took over from McKenzie as the county with the highest well density in mid 2010 and still keeps this position, refer also figure 15.
- McKenzie now provides for around half of the annualized production growth from these 4 counties.
Figure 05: Chart above shows developments in reported tight oil production from from the 4 counties with the biggest production (Dunn, McKenzie, Mountrail and Williams). A 12 Month Moving Average (12 MMA) smoothing has been added to better identify underlying trends in oil production for these 4 counties.
Figure 06: The chart above with the stacked area shows development in (total) reported oil production for the 4 counties with the biggest production .
Figure 06 shows that the growth in oil production from Williams and Mountrail has slowed down while there is good growth in oil production in Dunn and McKenzie.
To better visualize the underlying trends for oil production from these 4 counties the developments described by 12 MMA (annualized) was applied as shown in figure 07 below.
Figure 07: The chart above shows the development in oil production from the 4 counties with the biggest production by using 12 Months Moving Averages (12 MMA, annualized).
By using 12 MMA it becomes clearer that growth in production in Williams and Mountrail is slowing, while growth is lead by McKenzie followed by Dunn.
THE OTHER COUNTIES IN NORTH DAKOTA
Figure 08: The chart above shows development in NDIC reported oil production from the other counties in North Dakota. The black line shows the oil production development for these counties expressed by 12 MMA.
For the other counties in North Dakota three counties, Divide, Stark and Burke saw some growth in oil production starting back in 2010. Recently only Divide continues to show growth.
DEVELOPMENTS FOR 2 POOLS IN MOUNTRAIL
In figures 6 and 7 it is shown that in recent years Mountrail was the county that lead the growth in oil production starting back in 2008. Note the decline in production following the collapse of the oil price in 2008. More recently the growth in production in Mountrail has slowed down. Mountrail is where prolific pools like Alger, Parshall, Reunion Bay, Sanish and Van Hook are found.
Figure 09: The chart above shows the development in total tight oil production for the Alger pool split between the company with highest production (Statoil, dark green area), and others, pink area all rh scale. The chart also shows the development of the number of wells split on Statoil with the highest number of wells (white circles connected by grey lines) and total number of wells (yellow circles connected by black lines) both plotted against lh scale. The black columns at the bottom shows a month over month changes in wells (lh scale).
NOTE: The chart does not include wells and production from wells on confidential list.
During the last year and despite a high number of wells added, production in Alger has been in decline.
Figure 10: The chart above shows the development in total tight oil production for the Sanish pool, dark green area, rh scale. The chart also shows the development in the total number of wells (yellow circles connected by black lines) plotted against lh scale. The black columns at the bottom shows a month over month changes in wells (pH scale).
NOTE: The chart does not include wells and production from wells on confidential list.
A FEW NOTES ABOUT WELLS
The monthly statistics in the public domain from NDIC on changes in the total number of producing wells does not distinguish between flows from wells flowing for the first time (“virgin” wells) and wells that were capable of flowing and for some reason and duration was idle and brought back to flow.
- North Dakota Industrial Commission (NDIC) reports show that the number of net added producing wells in the Bakken, Sanish, Three Forks and Bakken/Three Forks pools increased by 195 from September to October.
- NDIC reports show that the number of wells actually producing increased by 199 in North Dakota from September to October of which 154 was in the 4 counties with the biggest production.
- NDIC reports show that the number of wells capable of producing increased by 161 in North Dakota from September to October, of which 122 was in the 4 counties with the biggest production.
- NDIC reports show that at anytime 9 – 10% of wells capable of producing are idle.
Idle wells are here meant to describe the difference between wells capable of producing and wells actually producing. - McKenzie now has around one third of the rigs drilling in Bakken (North Dakota) and provides more than half of the growth in tight oil production for the 4 counties with the biggest production, refer also figures 06 and 07.
The numbers of wells capable and actually producing includes wells on confidential list.
How many new (“virgin”) wells were brought in during October 2013?
WELLS CAPABLE OF PRODUCING
Figure 11: Chart above shows developments in reported number of wells capable of producing for the 4 counties with the biggest oil production in North Dakota.
In the chart above note the recent acceleration in wells additions for McKenzie.
Figure 12: Chart above shows developments in month over month changes in wells capable of producing for the 4 counties with the highest oil production. The chart also include 12 MMAs for the same counties.
In Williams there is a decline in added wells capable of producing, this is also reflected by slower growth in production as shown in figures 06 and 07.
Dunn and Mountrail see a relatively constant addition of wells reflecting a slower growth in production.
Mckenzie has good growth in wells capable of producing which is also reflected by growth in actual production.
WELLS ACTUALLY PRODUCING
Figure 13: Chart above shows developments in reported number of actual producing wells for the 4 counties with the biggest oil production in North Dakota.
Figure 14: Chart above shows developments in month over month changes of actual producing wells for the 4 counties with the highest oil production. The chart also include 12 MMAs for the same counties.
Data from NDIC for McKenzie shows that in September 2013 was an increase of 32 actual producing wells and an increase of 60 wells capable of producing.
How should the difference between the increase in the numbers of actual producing and capable of producing be interpreted?
WELL DENSITY (Wells capable of producing)
For what it is worth a chart showing the development of well density for the 4 counties with the biggest production has been included.
Figure 15: Chart above shows developments in well densities for the 4 counties with the biggest production.
1 square mile = 640 acres
Highest well density is now in Mountrail followed by McKenzie which appears to catch up. McKenzie had the highest well density until Mountrail took over in mid 2010 and Mountrail still has the highest well density despite the lowered drilling activity.
All the areas in the 4 counties with the biggest tight oil production should not be expected to have an ubiquitous productivity. There will also be a limit to how close the wells may be drilled without causing interference.
Promoters of the oil industry who claim that production in the Bakken region will continue to increase for many years to come appear to assume that the area outside of the 4 main counties (Dunn, McKenzie, Mountrail and Williams) will be as productive as within these 4.
IDLE WELLS
Idle wells are here defined as the difference between wells capable of producing and wells actually producing. At any time a number of wells capable of producing will be shut in for some reason, ref above the discussion about deviations between modeled and actual production.
As the total number of wells with time increases it could be expected that an increasing number of wells will become subject to seasonal effects, making it more demanding to bring these back to flow as condition permits.
Figure 16: Chart above shows developments in number of idle wells for the 4 counties with the biggest production.
Figure 17: Chart above shows developments in the portion of idle wells relative to the number of wells capable of producing for the 4 counties with the biggest production.
Mckenzie has the highest number and biggest portions of idle wells, and as of October – 13 around 11% of the wells were idle. McKenzie has been and is the county with the highest portion of idle wells which may be a seasonal accessibility issue.
Effects from wells temporarily shut in (idled) will reduce cash flows and returns.
—-
Red Queen posts on The Oil Drum:
Is Shale Oil Production from Bakken Headed for a Run with “The Red Queen”? (Sep 25 2012)
Reposted (Jan 01 2013)
Is the Typical NDIC Bakken Tight Oil Well a Sales Pitch? (Apr 29 2013)
Maugeri Misses Bakken “Red Queen” (Aug 05 2013)
