Monday, July 6, 2015

Oil Shock Models with Different Ultimately Recoverable Resources of Crude plus Condensate (3100 Gb to 3700 Gb)

The post that follows relies heavily on the previous work of both Paul Pukite (aka Webhubbletelescope) and Jean Laherrere and I thank them both for sharing their knowledge, any mistakes are my responsibility.  

In a previous post I presented a simplified Oil Shock model that closely followed a 2013 estimate of World C+C Ultimately Recoverable Resources (URR) by Jean Laherrere of 2700 Gb, where 2200 Gb was from crude plus condensate less extra heavy oil (C+C-XH) and 500 Gb was from extra heavy (XH) oil resources in the Canadian and Venezuelan oil sands.

In the analysis here, I use the Hubbert Linearization (HL) method to estimate World C+C-XH URR to be about 2500 Gb.  The creaming curve method preferred by Jean Laherrere suggests the lower URR of 2200 Gb, if we assume only 200 Gb of future reserve growth and oil discovery.  

Previously, I have shown that US oil reserve growth (of proved plus probable reserves) was 63% from 1980 to 2005.  If we assume all of the 200 Gb of reserves added to the URR=2200 Gb model are from oil discoveries and that in a URR=2500 Gb, oil discoveries are also 200 Gb, then 300 Gb of reserve growth would be needed over all future years (we will use 90 years to 2100) or about 35% reserve growth on the 850 Gb of 2P (proved plus probable) reserves in 2010.  I conclude that a URR of 2500 Gb for C+C-XH is quite conservative.
A problem with the Hubbert Linearization method is that there is a tendency to underestimate URR.

HL Analyses from 1998, 2005, and 2015 shown in the chart below give URRs of 1600 Gb, 2000 Gb, and 2500 Gb respectively, so this leads one to suspect the HL estimate is the minimum of future URR.

Figure 2

The United States Geological Survey gives an optimistic World C+C-XH URR estimate of 3100 Gb, (600 Gb of reserve growth and 500 Gb of discoveries) I believe this is too high and think the average of the HL and USGS estimate may be reasonable at 2800 Gb for World C+C-XH URR.

A second reason that a World C+C-XH URR may be higher than the HL estimate is to consider the United States at its peak in 1970.  The HL estimate for 1950 to 1970 pointed to a C+C URR of 208 Gb.  In 2007, the HL estimate for 1950 to 2007 pointed to a URR of 226 Gb.  After 2007, the large increase in US light tight oil (LTO) output no longer follows the linear trend and the US C+C URR is likely to be higher than 226 Gb.  

US HL Chart below:

Figure 3

Simply using the 226/208 ratio indicates that the URR estimate at peak output is likely to underestimate URR by at least 9%, so if we increase the World URR HL estimate by 9% (2500*1.09) we get 2700 Gb for the World C+C-XH URR estimate.

Another perspective is to consider that at the 1970 peak for US C+C output, the cumulative C+C output (Q=96 Gb) was about 42% of the current (1950-2007) HL URR estimate (96/226).  If we assume 2014 is the peak for World C+C-XH output and that the current cumulative C+C-XH of 1240 Gb is 42% of the eventual URR, then the World C+C-XH URR would be (1240/0.42) 2950 Gb.

As before, I would choose to average these two estimates of World C+C-XH URR to 2825 Gb and then round to two significant digits, the final estimate is 2800 Gb.

What about extra heavy (XH) oil from the oil sands of Canada and Venezuela?  Jean Laherrere‘s estimate for the XH URR is 500 Gb and the USGS estimate is 1000 Gb.  Initially, I was tempted to estimate the XH URR at 750 Gb, but conversations with Fernando Leanme (who is familiar with production problems in the Orinoco belt) led me to reduce the XH URR estimate to 600 Gb.  

The XH output scenario is shown below, it is assumed that this scenario is unchanged by changes in the C+C-XH scenario. This scenario is more conservative than previous XH scenarios I have used, which were more similar to those of Jean Laherrere.  Peak output is about 10.3 Mb/d in 2056 vs earlier scenarios which peaked at 14 Mb/d in 2070.

 Figure 4

For comparison Jean Laherrere’s XH scenario is in the chart below, peak of 15 Mb/d in 2070.

Figure 5

Jean Laherrere has provided the best data I have seen for World proved plus probable (2P) reserves for C+C-XH.  His Feb 2013 estimate for the end of 2010 for C+C-XH 2P reserves was 850 Gb.  The Oil Shock model estimates proved producing reserves for the World.  By using United States data from the EIA and assuming that 2P reserves are 1.5 times higher than proved reserves, we find that over the 1996 to 2008 period in the US, 52% of 2P reserves are producing reserves.  For the World, I have assumed the ratio of producing reserves to 2P reserves is 50 to 51%, so in my Oil shock models I aim for this ratio in 2010 where 2P reserves are well estimated at 850 Gb.

I present 3 different scenarios below where the XH output scenario is the same (600 Gb), but the C+C-XH URR varies from 2500 Gb to 3100 Gb, with an intermediate scenario of 2800 Gb.  In an earlier post, I presented a simplified oil shock model where the maximum entropy probability distribution was defined by a constant “k”, where 1/k is the mean time between oil discovery and first production for the average oil field.  For the models presented below, 1/k is different for each of the models in order for the ratio of producing reserves to 2P reserves in 2010 to remain similar as we change the URR of the C+C-XH models over a range of 600 Gb.

URR of Model    1/k
2500 Gb                29 years
2800 Gb                34 years
3100 Gb                38 years

For all three models the percentage of producing reserves to 2P reserves at the end of 2010 is between 50 and 51%.  The extraction rate was chosen so that the model matched EIA C+C output from 1960 to 2014, for years earlier than 1960 a single extraction rate was chosen so that the 1959 output level was consistent with 1960 output.  We do not know what extraction rates will be after 2014, we can only guess. 

I used the increasing trend of extraction rate over the 2009 to 2014 period and fit a line to the extraction rates for these 5 years, I assumed the trend would continue unless output was above 80 Mb/d, or the extraction rate was above the previous maximum rate in 1979, or output had peaked.  If any of these three things occurred the extraction rate would gradually level off and then would begin to decline, these are guesses and are merely suggestive of what could occur. 

I expect that future high oil prices will cause economic growth to slow and oil demand will decrease and thereby reduce the extraction rate.  There will also be some substitution away from oil use for transport as oil prices rise which will also reduce oil demand.  It is not possible to forecast these changes with any degree of accuracy.

The Scenario with URR=3400 Gb (including 600 Gb of XH oil) is below, peak is 79 Mb/d in 2018.  Annual decline rates are 1% or less until 2032 and remain below 1.5% until 2040 and stay between 1.5% and 1.83% until 2100.  Cumulative C+C output is 2900 Gb in 2100.

Figure 6

The Scenario with a World C+C URR of 3100 Gb (600 Gb of XH oil included) is presented below, the peak is in 2014 at 78 Mb/d and annual decline rates remain below 1.75% from 2015 to 2100 and are less than 1.5% until 2028.  Cumulative C+C output is 2600 Gb in 2100.

Figure 7

The final scenario has a World C+C URR of 3700 Gb (with 600 Gb of XH oil), peak is 80 Mb/d in 2023.  Annual decline rates remain below 1% until 2042, under 1.5% until 2047 and remain under 1.62% until 2100.  Cumulative C+C output is 3000 Gb in 2100.

Figure 8

The three scenarios can be compared in the chart below.

Figure 10 

If there were no future oil shocks due to economic crises or major wars and World C+C URR is in fact between 3100 Gb and 3700 Gb, then future C+C output could follow an output path between the green and red curves in the chart above.  It is unlikely that there will be no severe oil shocks in the future, though predicting the timing of such shocks would be difficult.

The three excel files with the three scenarios presented above can be found at the links below.  By adjusting the extraction rates to higher or lower values than I have chosen you can create any scenario you choose.  Optimists can have extraction rates go to 20% if they choose and pessimists can make the extraction rates go to zero (which would mean no output) if they wish.  One would have to explain why either of these would occur.  We don’t want to assume a collapse in order to have a collapse in oil output, and we shouldn’t assume that resources are infinite to prove that there will be no peak in oil output.  Each spreadsheet is about 3.5 megabytes (the order is 3400 Gb, 3100 Gb, and 3700 Gb).

A pessimistic scenario is presented below with an “oil shock” beginning in 2035 and similar in magnitude to the 1980-5 oil shock (extraction rate falls by a similar percentage), we could envision a war in the middle east where oil supplies from Iran, Iraq, Kuwait, and Saudia Arabia are reduced to zero and the resulting recession reduces oil demand even further, the URR is assumed to be 3400 Gb for this scenario.  Peak is 79 Mb/d in 2018, annual decline rate is less than 1.1% until 2035.

 Figure 11

An optimistic scenario is presented below where extraction rates increase at the 2009-2014 trend until they are 10% higher than the highest rate previously reached in 1979, then extraction rates remain at this level until 2100.  The C+C URR is 3400 Gb.  Peak is 81 Mb/d in 2023 and annual decline rate remains below 1.1% until 2035.

 Figure 12

A final “realistic”scenario has a less severe oil shock in 2035 with extraction rates decreasing only half as much as the “pessimistic” scenario and with extraction rates continuing to decline, but less steeply after 2045.  This might result from a severe financial crisis in 2034 with a depression following.  Peak is 79 Mb/d in 2018 and annual decline rates stay below 1.1% until 2036. Chart below.

Figure 13

The chart below compares the optimistic and pessimistic scenarios with my “realistic” scenario for a World C+C URR of 3400 Gb.  Cumulative output of the optimistic scenario is about 2900 Gb in 2100 and the pessimistic and realistic scenario has lower cumulative output at about 2800 Gb in 2100.

Figure 14

Excel files for realistic, pessimistic, and optimistic scenarios respectively are at the links.


  1. Quick comments:

    1. Not sure how useful the "estimate the total resource and then how we eat it up" is so useful. History has shown that it has pretty much underperformed because it doesn't count on new resources that are enabled with higher prices or with technical advances.

    2. Seems like there is a tendency for your models to always show a relatively soon peak after what looks like a history of going up. I just wonder if someone had done the similar models, 5 or 10 or 20 or 30 years ago, if they would also see a near term peak. And that wasn't what happened of course.

  2. Hi Nony,

    It depends on if the guesses about URR are correct and how fast the resource can be extracted and if it can be done so profitably.

    The scenarios are a guess about how supply and demand will play out over the future. Most scenarios of the future prove to be incorrect. Note that a model done 20 years ago with the same assumptions that I have used about the C+C URR would only be changed by not having the EIA output data from 1994 to 2014.

    The oil price 20 years ago was also very low and if a person believed that the futures strip gave the best indicator of future oil prices they would have expected oil prices to remain low an would have predicted the extraction rate to remain constant (and if they had been right about oil prices this might have occurred, except that demand may not have been met as those oil prices).

    The scenario peaks in 2023 at 77.2 Mb/d and actually does fairly well through 2013, I have used the 3100 Gb scenario because you are an optimist. Also note that if oil prices follow the futures strip as you believe there will be little incentive for extraction rates to increase in the future. Chart at link below:

    1. Hi Nony,

      So an optimist using a World C+C URR of 3700 Gb in 1996 (as I used the 1995 extraction rate) would have predicted a peak in 2023 if they had used the Oil Shock model, if you consider 27 years to be "near term", then they would have predicted a near term peak.

      Note also that the 3700 Gb scenario is only different from the USGS URR prediction by have ing 400 Gb less output from Oil Sands, the USGS expects about 500 Gb from Canada and 500 Gb from Venezuela, where I expect about 425 Gb from Canada and 175 Gb from Venezuela.

      Fernando Leanme is quite familiar with what is going on in the Orinoco belt and pointed out that these fields are being produced quite badly, to the point that water issues will make future SAGD production uneconomic. If the 1000 Gb estimate for extra heavy oil by the USGS proves correct that will have little effect on the scenario prior to 2100 because these reserves take a long time to ramp up.

      There are limits to how high oil prices can go and to how quickly oil resources can be produced.

  3. I think the whole approach of "guess the URR" and then assume that is what is used up is flawed. Not your implementation, but your whole method with implicit assumptions.

    Guess the URR and use it up is what Hubbert did. What Campbell did. It had no connection to economics. A geological assessment. Yours does allude to economics...but very weakly. The actual URR will depend on future price (with behind that, global demand and growth, as well as substitutes and possible GW regulations), and it will depend on future extraction technology. Evolution of EOR, evolution of shale drilling [it is early days after all], deepwater, potentially even kerogen recovery schemes. And...of course...depletion. But that is just one dog in the pack.

    1. Hi Nony,

      You are correct that economics will enter into it, find me an economic model that accurately predicts future supply and demand for oil and I will use it. I don't buy the idea that EOR advances will be all that great, shale drilling is really not anything new if you talk to the oil guys, the technology advances at this point are very incremental. Note that the USGS looks at what is technically possible without including the economics. You are constantly cautioning that oil prices are likely to be lower rather than higher, this will tend to reduce profitability and output.

      Also note that my Bakken models include the economics and when I use the low oil prices that match the futures strip, output is much lower. Then you say well if prices had not fallen output would be higher.

      You cannot have it both ways. There is no model that tells us what future oil prices will be. You can think of the URR that I use as the maximum that is possible under oil price assumptions that are reasonable. If oil prices rise to $175/b in 2015$ and stay there it is unlikely that we would see a World C+C URR above 3700 Gb. Kerogen? are you serious? Kerogen will always be the future of oil, a future that never arrives, not enough water for output to ever make a dent, especially at low oil prices.

      Don't you think oil will be replaced by other forms of energy? I expect as oil prices go up as oil depletes substitutes will eventually take over, this is part of why the 3400 Gb scenario seems more reasonable than either the 3100 Gb scenario or the 3700 Gb scenario.

      You do realize that EOR and ultra deep water oil are very expensive, yes?

  4. I mean consider this. What if US LTO goes global? Yeah, yeah, they don't have our service sector or mineral laws. and the Euros are a bunch of Greenies. But we are talking time frames of decades. To think that Bakken and EF are the only places with LTO is just unlikely. You have to think the ME, Russia, etc. are sitting on lots of LTO. So that is in the ultimate resource. For that matter supposedly the Paris Basin has some promising shale.

    1. Hi Nony,

      If LTO goes global, it adds maybe 50 Gb to the URR, the US will probably be 50 Gb tops, optimistically we might see that from the rest of the World at high oil prices, but that will be decades into the future and would problably be covered by my high case already. At some point oil demand will not be able to support a high enough oil price to make much of the URR producible at a profit. Sustained oil prices at $150 to $200 per barrel will make all kinds of substitutes competitive and a lot of oil and coal will be left in the ground.

  5. 1. Proving the counterproposition is not my point. I'm just saying include it in the solution space. As a possibility. After all if solar could become more competitive since the 1970s, perhaps Carter's synfuels might too.

    2. If shale oil extraction was such a no brainer, than why was it not in all the deemed POer models? Why was it not referred to as an expected supply source in mid 2000s?

    3. I would be pretty leery of "the Rockman" or some of the other commenters on the blogs here. You are looking at a very small subset and those who happen to be interested in coming to peaker sites. For one thing, look how shale gas exploded despite all the Berman comments in 2009-2010 that "Marcellus would disappoint" or that shale gas would need north of $8 or $10 to continue development. (The Marcellus is now ~15MCF/day and selling for $1-2/mcf!) There are a huge number of industry professionals out there as well as articles that are both scholarly and accessible. Don't automatically assume that they are blowing smoke either--remember the Simmons/Staniford fussing about Saudi Arabia peaking? Just regarding the govt statements as truth would have given you better insights than following the peaker meme. Rockdoc is pretty sound and the Cuban is interesting. But even there...ask for sources, citations, mathematics. Rockman in particular has a very limited experience in modern shale exploration. the industry has learned (iteratively, sure) over the last 5 years and his "Austin Chalk" (a carbonate with oil within the fractures rather than the tight matrix) is pretty weak stuff. At least read the real stuff. Not just comments and posts on blogs. (even if you are least read up on the conventional media.) The Cuban is kind of interesting and a cool guy...but even here, don't take some Internet commenter as the be all, end all. Read widely. Oh...and Jeff Brown is a freaking idjit with the cut and pasted ELM crap. and with his "40-50 API oil is bad" schitck.


    EIA estimate method. People can kvetch if they want to, but guys like Ron have themselves had to gradually be dragged into understanding how significant the revisions are over time. (At one point, he said they were only relevant 1 month back!) I also see a strong tone of conspiracy thinking from people wanting to say EIA is cooking the numbers.

    The big drop and return around January was related to the contango trade (and a bit of weather).

  7. As oil was $60 in June as well as May, I expect another moderate jump in that month. Chance that noses the number over the DEC14 record.

    Maybe a drop in July given the price drop at beginning of the month. Still looking at a year that basically goes sideways. No more rocking the party with growth. But not falling off a cliff either.

  8. The Bakken, I meant, above. Supposedly US overall dropped in May and June, from the 9.7 high.

  9. Pretty measured piece on the Bakken by John Kemp.

  10. Interesting presentation on shale decline curves:

    Note how much better the Bakken does far out. I wonder if that is from it not being a true shale (middle Bakken is a reservoir rock). Kind of implies that we will have stripper wells far out in the future for Bakken and that EF/Wolfcamp will have more plug and abandon.

  11. Dennis:

    If you haven't done so already, recommend to follow John Kemp's twitter and to sign up for his daily email. He gets some interesting stories in the email that I don't always get from google searches. On the twitter, he has occasional light analyses or links to those of others (e.g. traders) that are not in his columns. Think you/Enno/Alex/Dean would appreciate the content. Not saying in any way that it is earth shattering or end all be all, but interesting and helpful even with caveats, etc.

  12. EIA monthlies are out. They have the US doing 9.6 MM bpd C&C in June. Although the numbers do bounce around strangely from JAN to JUN. And there are obviously some estimates in there (footnote makes that explicit)

  13. Interesting listening to CLR and EOG earnings calls. First time I saw CLR come off better. I think CLR said they are already consuming DUC and EOG will do so next year (they really had no good rationale for not doing it immediately). Both companies looking for better oil prices to resume drilling, so implicitly given futures strip, probably won't see EOG return to mich drilling and will see CLR do further reductions. Production declines in late 2015, but pretty flat in 2016 YOY.

  14. Not sure when it went up, but there is finally a recorded webinar for JUL NDIC director's cut. Some good Q&A.

    They think they can stay stable for 2 years with 75 rigs and depleting the inventory. Thinks June will be big month but then drops after that because of price. Big jumps in DEC and JAN up, because of need to complete wells by deadlines.

    About 1% lower volume because of well conditioning. Basically taking propane etc. out of the oil barrel and moving it into the gas handling plants. Doesn't include any drop because of the hurdle of the processing.

    Few other items.

  15. For Enno, re oil/gas/water ratios:

    1. Conditioning rule is moving some volume from liquid volume into the gas stream. (I think I've heard a couple percent of the oil volume, but not sure. In any case, we know the direction. When we move propane out of the oil and into the gas, it makes one go down and the other go up.)

    2. Also possible is the geography shift: I think the center, deep part of the basin has more natural gas and less water.

    3. Nice discussion with coffee on choking. Assuming that producers are going full bore is just the opposite of the choking argument, though. If you had some doing one and others doing the opposite, it could even average out. Yes, near term cash desires could cause some producers to prioritize immediate cash over long term NPV. In addition, some of them may be implicitly speculating on future oil prices--not making their judgments off the strip but assuming prices will recover more, etc. (based on their statements). This could vary form company to company.

  16. I am really interested to follow these kind of topics. so please keep writing and sharing with us. Thank you so much.

    inexpensive essays service

  17. For your recent Bakken columnist. Skimmed the Bakken valuation model post. Didn't go into details but looked like decent effort, kudos.

    And surely most wells are not positive NPV at $37 WTI. (Even with an option to refrack!)

    One major mistake in his model is that he includes a borrowing cost in OPEX. He needs to strip that out. Cost of borrowing is already encompassed by the WACC (the 10% rate of return). You have to return on the money employed regardless of the capital structure. [This is a fundamental concept of valuations (see texts like Copeland et al).] The fellow is double counting the hit of financing.

  18. Looked at it a bit more. Certainly a lot of work and some of it decent. My comments more to help the author fix some things:

    1. (as before) including charges for interest in OPEX AND requiring 10% return on investment is double counting.

    2. Even if you did some form of equity DCF NPV (and don’t, calculation should be more impartial to capital structure), you still shouldn’t have the interest cost as part of a “per barrel” flat $12. After all, the amount of financing is the same but is distributed over more barrels for the prolific wells. [fixed with correcting item 1, but think this is an important teaching point.]

    3. The whole variable cost $12 looks fishy. Need to show that buildup (sources, assumptions, breakdown). It’s a big part of what is driving the result here. So go into more detail.

    4. $2.25 for gas (going forward for next several years) seems low. HH shows strip climbing to ~3 in a couple years.

    5. Discussion needs to address the price of oil by year (after all, we have a structure that shows different amounts of production of oil per year). There is a strip. Could consider to just copy it (don’t have to go month by month, just approximate). Or at least make some simple structure like a ramp or the like.

    6. Yes, the IBs predict low oil prices for 2016:

    but the strip is even lower (~$41 average price). In any case, don’t make the mistake of thinking only of 2016 though. After all you have a model that goes 20 years out. And producers make decisions based on the total NPV, not the initial year.

    I would be more inclined to use the 2015 or 2014 first year productions. After all, the cums do seem to have gone up. What matters is immediate decisions, not old well choices or situation years from now when the rock runs out.

    The EOG results are interesting. Wonder if they throttled production or did something strange with pad developments in 2014-5. They have some of the best rock in the basin (at the anticline in Montrail), so was expecting them to look better than they did.


    Again nice work and extensive. Take the comments as areas for improving something salvageable.

  19. "I agree with Ron that their [EIA] US estimates are somewhat better (though still on the optimistic side in my opinion.)"

    EIA consistently underestimated the US LTO rise every year as it went up. They also predicted a faster dropoff than has occurred so far (and that is with price surprising to the low side).

    1. Hi Nony,

      Yes the EIA and most others underestimated LTO expansion, the DPR estimates for the drop in shale output are not very good. I expect the decline in LTO output will not be that big 500 to 1000 kb/d, depending on oil prices.
      If the futures strip proves to be a good oil price forecast, then 1000 kb/d drop in LTO output from the peak is conservative in my opinion.

    2. I expected it to drop MUCH faster than it did. Expected even DEC14 or JAN15 drops. Thought the lags were overrated. Turns out I was wrong. Think rig contracts make a big difference (cause lags).

      At this point, sure, agreed on the decline. Yergin agrees also:

      I do think the stuff about shale running out of sweet spots or being NPV negative at $100 was pretty silly though. We were on track to do 10.5-11 MM bpd if oil had not dropped. Even that little blip up to 60s last JUN-JUL (prompt, strip ran to 70s) showed that shale has "legs". Rigs picked up at that point. I think that pretty much shows that you can't expect something sustained that is far higher than 65 (because shale turns on again).

  20. The stupid, it hurts:

    1. Totally ignores price and low drilling rigs. Not to mention the low STRIP for prices.

    2. Ignores the curious stabilization in the Haynesville decline after 2014 (the anti-Hubbard, anti-Seneca shape)

    3. Ignores the Utica growth (which has massive recent growth). Completely. Despite it being from same DPR source and a gas play.

    4. Ignores HOW Marcellus grew while the other plays declined (and price did): gas on gas competition.

    5. Ignores low App pricing (even lower than strip) along with limits on transport out of basin.

    6. Ignores the low heating demand this winter along with the flush storage. Nat gas is known to have weather variable demand. It is a basic feature of this commodity. Yet the peaker here decides to try to draw a long term inference from a few months slowdown.

    7. Mixes in tight oil at the end for a general rant, when title is on gas.

    8. Some confused "Empire" ranting. Didn't we have a decaying empire before the shale boom too?

    9. Amens from the commentariat. Ron distances himself from the nutter Empire stuff, but doesn't see the flaws in the analysis itself.