Potential Impact of Cantarell's Decline on Mexico's Oil Production

In 2005, Cantarell produced more than 60% of Mexico's oil production. Therefore, Cantarell can be considered as the "Ghawar" of Mexico. As we can see on Fig. 1, Cantarell's output doubled within a few years thanks to Nitrogen injection. However, the oilfield production has started to decline rapidly since 2004:


Fig. 1- Cantarell Production (source: PEMEX)

We are proposing to analyze the Mexican production based on two separate Hubbert Linearizations. The resulting production decline is much more important than the decline forecasted by the application of one Hubbert Linearization on the total profile.
Acronyms:

URR= Ultimate Recoverable Ressource
HL= Hubbert Linearization
mbpd= Millions of barrels per day
Gb= Billions of barrels

Cantarell is in fact an oil field complex situated near the Yucatan peninsula (see Fig. 2) which has produced 11.674 Gb since 1979 (Mexico's cumulative production is around 34.0 Gb). The volume of oil in place (3P) is estimated to be around 33.7 Gb according to PEMEX. Click To Enlarge.


Fig. 2- Cantarell complex
. EIA. Click To Enlarge.



Fig. 3- Mexican main oilfields (Deutsche Bank). Click To Enlarge.

We estimate the logistic curve parameters on the total Mexican production (Fig. 4). The estimated URR is around 66 Gb which means that cumulative production will reach 52% of the URR in 2006.



Fig. 4- Hubbert Linearization applied on the Mexican Production (crude oil + NGL, source: BP). Click To Enlarge.

We don't need a complete production profile for Cantarell, we use only the data from 1992 to 2005 available on PEMEX website. The result of the HL technique applied on Cantarell is shown on Fig. 5, only the last two or three points can be used to perform the linear fit. The resulting decline is quite steep (K= 19.3%) and the URR is 17.14 Gb which means that 11.674 / 17.14= 68% of what is recoverable has been produced!


Fig. 5- Hubbert Linearization technique applied on the Cantarell oilfield (crude oil + NGL, data source: PEMEX). Cumulative production has been modified to match 11.674 Gb in 2005. Click To Enlarge.

We apply a second Hubbert Linearization on the rest of the production as shown on Fig. 6. The decline is comparable to the one derived from the total production.


Fig. 6- Hubbert Linearization technique applied on the total production excluding the Cantarell oilfield (crude oil + NGL, data source: BP). Cumulative production has been modified to get 22.0 Gb in 2005. Click To Enlarge.

Finally, we put together the three above results on Fig. 7. The red dotted curve is the sum of the two logistic curves estimated for Cantarell (in blue) and the rest of the production (in green). We can see that the resulting curve is strongly affected by Cantarell decline with a production down by 30% in 2010! The URR corresponding to the new curve is only 28 Gb + 17 Gb= 45 Gb (Close to the ASPO estimate at 50Gb).


Fig. 7- The three logistic-based forecasts corresponding to the total production (black), Cantarell production (blue) and the rest of the fields (in green). The red dotted line is the sum of the green and blue line. The red circle is an early estimation for 2006 from the IEA (3.7 mbpd). Click To Enlarge.

Note that Cantarell decline rate at 20% is not exceptional and that steeper decline rates (around 25-30%) have been observed on Norway's top fields. 2006 will definitively be a turning point for Mexico oil production (see also Dave's post for the political context). Of course, the red line represents the worst case scenario where no new developments can offset Cantarell decline. The reality will probably be between the black line and the red line. Even the IEA is not very positive, from the last IEA report:

Mexico– April actual: Official production data from state company Pemex showed a rise of 20 kb/d for crude and 10 kb/d for NGL versus March, reaching 3.37 mb/d and 440 kb/d respectively. However, this Report has trimmed expectations for crude production in 2006 overall by 30 kb/d, to 3.27 mb/d. It is increasingly difficult to see offsets for declining Cantarell field production before 2007. More immediately, the campaign for the 2 July Presidential election has seen oil sector reform become an issue. The reversal of production decline and falling reserves is expected to require tax reform in order to boost Pemex’s budget, and an opening of the upstream sector to foreign company participation in deepwater hydrocarbon developments.

I give also the official production forecast from PEMEX and Wood Mackenzie who are seeing growth in the Ku-Malloop-Zaap (KMZ) field.


Fig. 8- PEMEX production forecast (2004). PeakOil.nl. Click To Enlarge.




Fig. 9- Wood Mackenzie production forecast. OilOnline. Click To Enlarge.


[Update by Khebab on 07/15/06 at 6:30 PM EDT] Following the many comments about the HL using only two points, I added the Fig. 10 below that shows alternative curves for Cantarell's decline. The exponential solution follows the comment made by plucky underdog.


Fig. 10- Various decline estimate for Cantarell: 1) In red: HL result using two points (same result as Fig. 5); 2) In orange: HL result using three points; 3) In purple: exponential decline

References:

TOD: Trouble South of the Border -- Mexico's Oil Production
ASPO: newsletter #35
PEMEX: Liquid Hydrocarbons Production
PEMEX: Hydrocarbon reserves 2005
BP: Statistical Review of World Energy 2006
GraphOilogy; Mexico's Ability to Export Oil
In Saudi Arabia water is used to boost oil pressure and recoverable oil.  What is the advantage of using Nitrogen?  The results seem impressive.  According to Fig. 1, production increased by some 800,000 b/day.
Cantarell Nitrogen Injection Project (pdf)
The above considerations, along with other technical and economical studies led Pemex to conclude that pressure maintenance by nitrogen injection was required to maximize the economic value of Cantarell fields. Among gas injection technologies, nitrogen was selected by considering availability, cost, handling infrastructure, environmental, safety, and reservoir issues.

Nitrogen?  The results seem impressive.

What I though was impressive was how the Nitrogen used represented a large %age of total Nitrogen gas produced.

2-3 years ago a web page was claiming %40 (or was it 60%) of production was used.

Spectacular work!

Three Points:  

(1)  Based on the WSJ article (which was based on leaked internal Pemex reports), the remaining oil column at Cantarell--between the water leg and the gas cap--is about 825'.  It is thinning at the rate of about 300' per year.  The worst case internal (Pemex) decline rate estimate is about 40% per year.  

(2)  Virtually all reservoirs are more permeable to gas and water than to oil.   Heavier oil is more viscous than light oil.  So, one big difference between Ghawar (light oil) and Cantarell (heavy oil) is that it is easier for the gas and water to bypass the oil at Cantarell than it is at Ghawar.

(3) The easy oil is gone at both Cantarell and Ghawar. From this point forward, high production rates are the enemy of high (remaining) recovery rates.   So, they can have higher recovery rates, or higher short term production rates--but not both.

Thanks for the info!
Very good work, Professor, as usual.

Regarding getting more traffic to theoildrum.com, you guys should consider posting your charts, graphs, and diagrams to Flickr.com. Flickr is Web 2.0 for images. For my site something like 20% of my inbounds (exclusive of searches) come from Flickr.

interesting, the above images are already on my flickr account. However, they have not been tagged.
Can you provide a URL to your Flickr account?
It's not very well organised for now:

http://www.flickr.com/photos/jacobfoucher/


The old adage that when you hold a hammer every problem looks like a nail seems appropriate here.  The logistics curve is sometimes a useful forecasting tool but I think it should be used with discretion.  Figure 5 demonstrates how inapropriate the logistic curve can be at times.
   
Absolutely! The top half of figure 5 is the most unconvincing linear fit I've ever seen anyone dare publish. You've got dots  making a flattish trend and then you arbitrarily take the last two and draw a line through them. You call that a linear fit? Come on! This part of the analysis is totally off base. You could plausibly draw a million other lines on that graph and get different results.
Look, everything is highly speculative here! Cantarell production has just peaked in 2003 and we have only two years of decline. However, engineers on the ground are saying that it will be a sharp terminal decline so no, you cannot draw whatever lines you want based on this information. Note also that the production levels predicted by the logistic curve on Fig 5. are the same as those given by PEMEX (Fig. 8) and Wood Mackenzie (Fig. 9): around 1.2 mbpd in 2010.
Look, everything is highly speculative here! Cantarell production has just peaked in 2003 and we have only two years of decline.

But the argument is that Hubbert linearization, in isolation (i.e., no USGS studies, no CERA data) allows to predict the peak in advance, right?  

Two years ago would an HL fit have told us that Cantrell was at its peak?  From the data shown, I would guess no.

The HL is not used here to predict the peak production (i.e. when we gonna peak and at what rate?), Cantarell has already officially peaked. The HL is used here to estimate the URR and the logistic decline rate (K) (i.e. how fast we gonna lose Cantarell's production). The resulting logistic curve is simply modeling Cantarell's decline post-2005.
I understand this.  I am questioning whether the HL technique has any predictive value at all.  If it didn't predict Cantrell's peak, why would the technique suddenly become predictive immediately following the peak?

While I like the elegance and simplicity of the method, the shown fits do not convince me of its utility.

why would the technique suddenly become predictive immediately following the peak?
I see two main reasons:
  • we know that a fast and terminal decline is expected (see the official chart on Fig. 8).
  • the predicted URR (17 Gb) is close to the official number (20.5 Gb)

I could have used something else that a logistic curve: a gaussian or a constant R/P ratio used by economist. Any monotonic curve that is delimiting an area close to 9 Gb will have done the job.
Is there any reason to believe then that the HL method is better model than, for example, eyeballing the decline (based on the two post-peak points) and calculating the area under the graph to generate a URR figure?
You can find some background on the logistic model applied to the modeling of ressource depletion in the following post:

Links to tutorial material on Hubbert Linearization

It seems to work but no one knows why. I have a problem with understanding what the logistics K physically means. I think the real depletion rate for all Mexico is under 10% of available reserves using a more realistic rate model.

http://mobjectivist.blogspot.com/2006/03/mexico-oil-shock-model-part-2.html

Again, I don't think K makes any real physical sense.

"You could plausibly draw a million other lines on that graph and get different results."

This is true, so long as your "million other lines" fit within the constraints of an oil column of 825' thinning at the rate of about 300' per year.

In regard to the HL method, IMO the key point is that the HL plot of total Mexican oil production fits the same pattern that we have seen in other major oil producing regions--Texas; Lower 48; Total US; Russia; North Sea--and most recently, Saudi Arabia and the world.  

The big fields tend to peak at the same time as overall regional production--Texas & the East Texas Field; Mexico & Cantarell; Saudi Arabia & Ghawar.  

Based on Khebab's work, we published an article on Graphoilogy and the Energy Bulletin predicting that Saudi Arabia and the world were on the verge of the same kind of decline that Texas and the Lower 48 showed.  Guess what?  All of the available data to date show that our prediction was correct?  

Instead of hypercritical discussions of how to linearlize the production in a field that is known to be declining, how about an acknowledgment of the success of the HL method, when applied to large producing regions?

Should be:  "All of the available data to date show that our prediction was correct."
What happens with the Mexico data if we fit the HL data between 10 and 20 Gb cumulative production?  How predictive was an HL fit from 1980-1990 of 1990-2006 production in Mexico?
From my understanding the HL method does not predict anything until peak is reached and steady state or decline has started - basically when only geology is the limiting factor in extraction rates.  You can't use HL on a new developing field, or on a data that spans recovery methods (before and after N injection).

HL is used to predict URR after peak.

in 1956 hubbert used it to predict that us lower 48 would peak in 1971, which is what happened. Since then aspo et al have tried to extend this success to the world, so far have repeatedly been too early.
the HL technique gives a credible production profiles for very few regions (Lower-48, Norway that I know of).

The following posts tried to assess the predictive power of the logistic curve fitting:

How Reliable is the Hubbert Linearization Method?
Bootstrapping Technique Applied to the Hubbert Linearization
How Reliable is the Hubbert Lin. Method? the world case

The HL fails when there is an abrupt change in the logistic growth (or decline rate) K which is the case for Cantarell, see this post from duncanK:

The Yibal Model - Does it Apply to Ghawar, Burgan, etc.?

"Instead of hypercritical discussions of how to linearlize the production in a field that is known to be declining, how about an acknowledgment of the success of the HL method, when applied to large producing regions?"

FYI--Just to be clear, I was responding to Halfin.

/begin sarcasm alert
Yeah, Halfin, I guess the market would draw a rising vertical line through those two declining points and declare everything's ok, right?
/end sarcasm alert
I agree, "fitting" a line to two points is a bit of a joke.  What an easy way to get and r of 1.00

Its not that the data don't deserve speculation and discussion, but using linear fit to come up with "hard" numbers, even if speculative, seems a bit disingenuous.  Otherwise great post.

-Ptone

Ok, I surrender! The term "fitting" does not apply for two points, I should have called it "line drawing" or used three points instead :). Anyway, it would not have change the result significantly. Note also that the line gives an URR around 17 Gb consistent with the official estimate:
Cumulative: 11.6 Gb
Reserves:  8.9 Gb
URR= 20.5 Gb
src: PEMEX: Hydrocarbon reserves 2005 (Table A5, page 29)
Absolutely first rate work!
A+.
In The seattle Times yesterday. " Yergin said petroleum supply-demand fundamentals are improving, with global oil inventories and spare-production capacity rising, but clearly not enough to offset the unrest." I need to know where he gets his drugs. Interesting comment on "Yergin Day"
Also in the Washington Times. "The oil price has really become a Richter scale for geopolitical turbulence and upheaval," said Daniel Yergin, the head of Cambridge Energy Research Associates. "The market's fundamentals are actually getting a little better, but fear and uncertainty is mounting over Iran, the Middle East and Nigeria." If it wern't so serious an issue to the average sleepwalking american it would be funny.
This month's Harper's (Aug) features an essay about peak oil. (Not on news stands yet.) The piece just barely skews towards the skeptical side, but I think that that is designed to lend it a sense of balance.

It covers most of the main technical points though it does get a couple of things wrong.

The author attended a symposium in Yellow Springs, Ohio and reports on the people he met, examining the general political trend among attendees, calling his experience in the essay's subtitle: Scenes from a liberal apocalypse. He likens the people of the peak oil scene to the Left Behind series and several other "the end is nigh" movements, except that he does admit that we do live on a sphere and that oil is necessarily limited, which he concedes lends credibility to the peakers.

The info is nothing new, but it is compelling for what it represents in terms of nascent media coverage. Harper's is one of the nation's better, more intellectual magazines and often covers stories before they break into the mainstream media. He does mention the increasing number of cities that are examining their futures post oil.

A nice read, but not a must read for hard core peakers. It would be a good article to recommend to the uninitiated.

I'll look for that. I had been hoping to do a piece for Harpers but saw that they didn't take unsolicited manuscripts. So, my take on what you say is that they sent some naive friend of theirs to do a story instead. That's too bad and it just shows how the media, even the pretty good media like Harpers, fail us regarding Peak Oil.
I just got my Harper's.  The article, which I haven't read yet, is by Greg Palast!
No, it's not. Palast did not write the Harper's article.
Sorry, I picked up the wrong Harper's.  Palast wrote an article in I think the June issue on Chavez and the end of cheap oil.
Dave,
Except for peer-reviewed academic journals, you need an agent to get published.

A good agent is hard to find.

Been there.

Done that.

Oy.

Aye, this isn't entirely true, but there's an element of reality to the idea.

Often it is more difficult to get published without an agent (mainly because you have to act as your own agent while also writing, which means big distraction from the fun stuff), but it can be done.

And, if you're agentless, and break through with a novel (and I know several people who have done this), getting an agent is a piece of cake. And, at this stage, an agent is highly recommended, for they'll likely be able to get you a bigger advance, among other things.

-best

For one year an agent tried and failed to find a publisher for my textbook. He gave up.

A year later, I found a publisher and achieved wealth beyond the dreams of avarice;-)

(Don't I wish . . . .)

Agents are generally no help for textbooks, but for most other kinds of writing an agent is worth twice her weight in 24 karat gold.

From PEMEX comments, I think they know it is in decline and serious decline.  They have and will continue invest in expensive enhanced recovery methods, but they are betting on bringing new smaller fields online to counter the decline.

The last I heard is that new fields are falling behind schedule due to costs and resources.  As for resources, everyone is complaining about lack of people and equipment.  This is not going to change at all for the forseable future.  

I remember in 2000, Exxon officials were saying just replacing production of declining oil fields will be a challenge and they expected oil prices to firm up because of this.

Please forgive a new member a (possibly) foolish question:

There was a prior peak in the field just before Nitrogen injection was started. Could that prior peak data be used to predict URR?

From reading prior posts here, discussing why some fields fit the HL method better than others, EOR techniques seemed to skew the data by not showing production declines until the field was much older (had produced a greater % of it's URR). Seeing that first peak, I wonder if that was not the true 50% of the field? And if so, what the predicted URR would be?

Again, please forgive if this has all been discussed elsewhere. Still trying to catch up.

It's possible, I would have to complete my Cantarell production profile to verify your idea.
It would be interesting to see mexico's internal consumption superimposed on fig 9.
Consumption in magenta (from BP 2006):

Mexico would seem to be a potentially good case to examine Westexas' 'exportland' model.
I explored this approach a few months ago but I have to revisit that model:

Mexico's Ability to Export Oil

Hello Data Freak Supreme Khebab,

Super Work, big thxs.  Scary how much Cantarell resembles DuncanK's Scudia model.  Let's hope a Cat 5 hurricane never hits Campeche Bay.  Speaking of a potential hurricane....

Mexico's election is far from a done deal.  The wildcard, the NUCLEAR OPTION for AMLO to truly energize his poor supporters, is for him to disclose the PEMEX reports, and other research such as our vaunted Khebab's work.

If AMLO goes whole hog on Mexican depletion, and convinces his supporters that now is the ONLY TIME POSSIBLE for oil wealth redistribution if he becomes Presidente', then they can easily start a revolution:
--------------------------------
 Mexico's Tiger Stirs

Porfirio Díaz, Mexico's dictator from 1876 to 1910, always feared a popular revolution. "We must not awaken the tiger," Díaz famously declared. The revolution that erupted in 1910 cost more than a million lives. Mexican intellectuals have recently warned that the tiger is stirring once again in the wake of the country's contested July 2 presidential election, the initial results of which granted a razor-thin victory to Felipe Calderón of the conservative National Action Party (PAN).
-------------------
http://www.thenation.com/doc/20060731/editors

What does that one million in 1910 extrapolate to today's Mexico-- 5 million or more?  Our 1860s Civil War would look like a friendly neighborhood block party compared to the carnage that could erupt down south.  All AMLO has to do is tell the poor that they will never get a chance for a decent life because Cantarell is tanking and Calderon wants to export as much as possible to the US.  AMLO can drive the crowd into a frenzy by saying, 'I am your very last chance, the vote tampering was obvious--do you want to be forever screwed by Calderon?

Bob Shaw in Phx,Az  Are Humans Smarter than Yeast?

The issues on the latest graph is the actual decline rate of Cantarell and the direction of the internal usage curve for Mexican citizens.  Plug in the 40% thingy and 2009 looks grim for exports. Any uptick in local consumption, ditto.
Bob,

You may want to check out this article by Palast ( I know many here feel he is a dufus for his insane views on peak oil, but, when it comes to political reporting, he is usually accurate.)

http://www.commondreams.org/views06/0714-20.htm

The following quote gives some idea as to the article's content.

Lopez Obrador is of a different breed. At the rally last Saturday in Mexico City, he played video and audio tapes of the evidence of fraud on a screen eighty feet tall. Imagine if Gore had projected the "scrub sheets" of purged Black voters on a ten-story-high screen in front of the White House.

Lopez Obrador put political force behind his legal demands by calling on voters from every state in Mexico to march to the capital. Two million are expected to arrive this Sunday. The result: the word among the political classes is that the election may be annulled. Even the conservative Financial Times has warned Mexico's elite not to "fool itself" by ignoring the demand for a full vote count.

This has the potential to be a very big deal.  Think of the immagration pressure if there were fighting down there instead of just bad economic conditions.
Hello Twilight and Cherenkov,

Thxs for responding.  IMO, the best thing for Mexico to do is have a runoff election between AMLO and Calderon to help prevent violence.  For the next two months, they campaign, but are restricted to only debating Mexico's Peakoil.  AMLO gets full control of one of the major media sources, Calderon the other to foster the debate equally.

AMLO would emphasize Biosolar Powerup and Detritus Powerdown, and the creation of a Mexican Foundation to help direct this end of entropy planning.  The whole gamut of NAFTA withdrawal, curtailing oil exports to the US, vol. pop. cntrls., building mass-transit, education, the shift to alternative energies, economic justice, Earthmarine protection strategies... on and on.  AMLO merely has to be honest with his supporters.

Calderon would have to justify business as usual and continued exports to the US in the face of rapidly collapsing depletion.  He would have to explain how the detritovores will continue to polarize mexican society, and how privatisation of PEMEX will mean less future oil for Mexicans.  He would be confronted with claims that infinite growth is now impossible and to stop spewing false promises.  He would have to explain his employment strategy so that millions will not invade the US as they seek to follow the Mexican oil.  He would not be able to bullshit anyone with political lies if the debate is constrained to  the REAL FACT of Mexican Production Collapse.

The Mexicans could then choose the candidate that offers the most appealing decline path.  They have no other choice but to repeat the Mayan decline, but do it much better if they plan ahead.

Failing this hypothetical scenario I proposed--Mexico will eventually become like Nigeria, Zimbabwe, Haiti, Somalia, Iraq, etc.  Being a fast-crash doomer, I think the previous sentence is more likely than the paragraphs above because the US always gets involved in Mexican affairs.  But one can hope.

Bob Shaw in Phx,AZ  Are Humans Smarter than Yeast?

You seem to be looking for a way to make it fair.  Fair is not the goal - winning is.  At any cost.  Just like here.
Hello Twilight,

You are probably correct-- Mexico has a long history of assassinations where the winner [the US?] takes all:

http://www.meta-religion.com/Secret_societies/Conspiracies/Assassination/mexicos_fiesta_of_assassins .htm

http://en.wikipedia.org/wiki/Luis_Donaldo_Colosio

http://en.wikipedia.org/wiki/Special:Search?search=mexican+assassinations&fulltext=Search

Ruppert of FTW makes a pretty powerful case of world drug control by the US topdogs, and Afghanistan is expecting a record opium crop under US military control.  Gee....

Makes one wonder if AMLO is living on borrowed time?

Bob Shaw in Phx,AZ  Are Humans Smarter than Yeast?

Ruppert of FTW makes a pretty powerful case of world drug control by the US topdogs, and Afghanistan is expecting a record opium crop under US military control.

Catherine Austin Fitts makes the same pitch.
She's got her 'red button story'
http://www.scoop.co.nz/stories/HL0510/S00173.htm
And one of her sites sent me here:
http://www.larouchepub.com/pr/2001/2839grasso_factor.html

So Ruppert isn't the only voice.

Hello Eric Blair,

Yep, tax-free income at that super-profit level is a very unstoppable force.  I have read Fitts's writings before.

Bob Shaw in Phx,AZ  Are Humans Smarter than Yeast?

Hi Bob,

I've been reading a bit about the Neocon plan for Mexico, which, broadly, is to combine Mexico, the U.S. and Canada into one large "North American Union" with a super-interstate highway connecting all three countries.

A new currency would also be adopted, called the "amero."

I'm sure this idea would be real popular in Arizona.

It's being championed by the Council on Foreign Relations, which, I suppose, is just another name for the Trilateral Commission ...

Hello Don in Colorado,

Yep, I am familiar with that idea too.  I have no idea if it will be driven from the topdogs down or if Mexican and Canadian nationalism for hoarding their oil will predominate.  We appear to be nearing a 'balanced teeter-totter' with globalized growth on one side and complexity collapse & HELP on the opposite side.  

Bob Shaw in Phx,Az  Are Humans Smarter than Yeast?

All you folks giving Khebab a hard time about doing a fit using 2 (year) points of data should be glad he didn't use 5 (2006 to 2008). Then, knowing what we know from PEMEX and the geology of the field, you would be saying
"What are we going to do without imports from Mexico?!?!?!?
You take care, now ---

Dave

"We don't need no stinking oil!"

(Sorry, Dave, I'm an old movie buff ...)

So is Mexico going to have a smooth transition?  I mean its new president, not its new oil production status!
There's something else to consider with Cantarell. How will its decline affect the Mexican governmental budget? How will that, in turn, affect the illegal immigration situation. Expect Mexico to want to "export" even more of its "problem" in the future.
Thank You.

Though in the general public if its not something they can read in about 3 mins it does not get past their need to go do somthing else.  Graphs help to catch the eye and pull in a viewer.  I know this post was meant for the general readership of TOD which reads for the full length of most posts, so the above comment is not meant as a negative to you.   The issues at hand are so important that we really just don't want to be another sound bite news item.  The Hired guns like Yergin get longer sessions to shoot us down if we only have 3 mins to tell the story as it is.  Though as the peak gets closer and more things change for the worse even Yergin et al. can't still be standing there telling folks pumping gas don't worry,  we predict it'll get better honest, just give us one more chance!! As the gas station owner takes the title deed of the car owner's house in payment for the gas.

"Poor Mexico, so far from God and so close to the United States!"

-- José de la Cruz Porfirio Díaz Mori
[Mexican President 1876 until 1911 (with the exception of a single four-year period)].

WSJ and leaked reports be damned (sorry westexas): ditto HL (sorry Khebab - but read on and find out why). There is a great deal of information about Cantarell available to the public, more than enough to allow a competent reservoir engineer to make a reasonable forecast of the fate of the field.

I moused around a bit and found an excellent case study paper on the Society of Petroleum Engineers eLibrary website. If you really care about Cantarell (and a lot of folk on ToD seem to), it's well worth the $20 non-member fee to buy an electronic copy. If you are a member of the SPE, you can get it for $6. If you have academic or corporate access, use that. Start at http://www.spe.org/ and mouse around from there. The SPE has several other papers about Cantarell; I didn't get round to checking the AAPG or SEG, but if you spent a hundred bucks on data, you would know an immense amount about the field. And a very strange and wonderful reservoir it is, too.

The paper is SPE97385, "Nitrogen Injection in the Cantarell Complex: Results After Four Years of Operation", by Sanchez et al, 6 pp., 12 fig., presented at the SPE Latin American and Caribbean Petroleum Engineering Conference in Rio, 20-23 June 2005. So it's probably reasonably up to date.

Here's my own abstract of the paper.

GEOLOGY - STRUCTURE, ROCKS AND FLUIDS

Cantarell is a buried reef complex containing multiple oil reservoirs. The largest, Akal, contains 86% of the oil originally in place, and is probably what Khebab's graph refers to. The main productive formations are fractured, vuggy limestones with low porosity (8%), although up to 35% of the pore space is represented by vugs and fractures. The matrix permeability is 1mD but the fracture system permeability is an astonishingly high 3000mD (compare with ~600mD for Ghawar).

The crest of the Akal reef is some 3608 feet subsea, the initial oil water contact was at 10500ft. The maps and sections in the paper are difficult to interpret (no scale bars dammit), but it appears that the aquifer is perched on the southern side of the reservoir, and the oil column extends deeper to the north.

The reservoir fluid is a 22 degree API oil, so would be classed as heavy. Solution GOR is not quoted (there are hints that it might be about 370) but it seems the reservoir was initially just above bubble point i.e. no free gas in the reservoir.

DEVELOPMENT HISTORY AND PRODUCTION PROFILE

Production started in 1979 (see the graph) and ramped up to about a million barrels per day by 1982. Wells were drilled from multiple platforms (looks like about 12-15 on the map - not sure what all those symbols are). Initial oil rates averaged 29 Mb/d per well, and only 40 wells were producing in 1981. The plateau was maintained by drilling additional wells and by installing gas lift to supplement the energy provided by reservoir pressure. Initial pressure was 3840 psi, but by 1995, when the expansion project was conceived, 150 gas-lifted wells were producing an average of 7 Mb/d each.

The expansion project involved the drilling of 205 new producers and 9 injectors - THAT is the reason for the production increase from 1996. That sharp dip in production in late 1995 may represent a brief field shutdown to hook up new platform, manifolds etc. Some of the new wells may have been sidetracks of earlier gassed-out wells.

Nitrogen injection simply protects the reservoir from the collapse in pressure that would result from the increased liquid extraction. Target oil production rate for the new plateau was 2 MMb/d and target gas injection was 1.2 Bscfpd, which would represent approximately 100% voidage replacement and maintain reservoir pressure - as confirmed by the graph. This in turn would keep the flowrate per well approximately constant.

RESERVOIR RESPONSE AND WELL PERFORMANCE

By 1995, pressure had fallen to about 1820 psi and a large secondary gas cap had formed. Because of the very high long-range permeability, the contact was almost flat everywhere in the reservoir i.e. it was a horizontal surface moving down through the structure. One thing that is difficult to estimate from the paper is how much oil remains in the matrix above the contact, and how fast it is draining out. It might be possible to get a handle on this from the graph of gas-oil contact depth vs. time in the paper. In 1995 the GOC was at 5900 feet, an average speed of 140 feet per year since 1979. Nitrogen injection started in May 2000 and reached its maximum programmed rate in December of that year. The contact would accelerate in response to production, not injection, but over the period 2000-2004 the contact has moved at about 230 feet per year. The most recent values for OWC and GOC given in the paper are 8860 ft subsea and 7380 ft subsea respectively, an oil column of 1478 feet at the beginning of 2004.

MY PROGNOSIS

I would not expect the oil water contact to move very much further now that reservoir pressure has stabilized. If injection and production continue at the same rate, the GOC should keep moving at the same speed.

I would see one of two things happening as far as the wells and facilities are concerned:

(1) The GOC starts invading the perforated intervals and gas production starts to increase. This will only become serious when the gas production reaches plant capacity, when it will become necessary to shut in production wells to cut back gas production. At that point the field will enter a regime of exponential or hyperbolic decline (note to non-techies - exponential does NOT mean extremely fast, it means the same percent reduction in flowrate each year). This may be happening already. In this regime, production can be forecast quite easily by fitting an exponential curve to history - NOT by Hubbert Linearization.

(2) So far nitrogen has not broken through in a big way, implying that most of the productive perforations are still submerged in oil. Given the extremely high permeability, the wells could in principle produce at full rate from a very thin shrinking layer of oil between water and gas. If the operator can conduct enough gas and water shut-off operations with the platform rigs and/or cranes (depends on platform layout), the oil production rate could remain very high right up to almost the end, at which point it would fall off a cliff. It wouldn't go to zero, but I would expect an abrupt fall of several tens of percent ("abrupt" means over a year or two).

In either case, decline would eventually switch to a different mode driven by oil drainage out of the matrix above the gas oil contact. It is difficult to estimate what production rate would be sustainable under this regime, but it would probably be several tens to a few hundred thousand barrels per day (wild-ass guess).

Hope this helps.

Thanks for sharing your excellent analysis!

Following your comment, I've just added a new figure (Fig. 10) that shows your proposed exponential decline. The decline is less abrupt compared to the two logistic curves. The area (i.e. remaining reserves) under the exponential curve is 15.16 Gb instead of 5.46 Gb and 7.43 Gb for the two logistic curves.

Urrrggghhhh ... the point is a bit more subtle than that. Decline analysis only works when the physical mechanisms of final decline have set in. Whatever functional form you use - the exponential-parabolic-hyperbolic continuum beloved of reservoir engineers, Hubbert linearization, a straight line, a straight line on transformed axes or anything else - you can only extrapolate it as long as the underlying physics of the reservoir doesn't change. Increasing the number of wells by 150% from 1996 would count as changing the underlying physics, for example. It accounts for the sharp increase in production. Going from zero injection to 100% voidage replacement is another change in the physics. It has a countervailing effect against the increase in well numbers, namely arresting the decline in reservoir pressure.

The next step change in the physics of the system that I can foresee will occur (is occurring?) when free gas starts to break through into the production wells. At that point the operator can choose to react or not to react. In the former case the reaction itself will become part of the physics of the system. An appropriate reaction would be:

(a) shutting in wells to hold back the gas, in which case some form of steady decline will set in

(b) a lot of well repairs or sidetracks, or even new wells deeper in the structure -  which will hold back the tide until the oil layer has almost vanished. This depends on how agile the well service department is and how long it takes to intervene in each well and skid the rig to the next slot.

(c) expanding gas handling capacity, in which case production will decline until the new capacity is on line, at which point it will increase again - and then decrease under the radically different decline regime of falling reservoir pressure

(d) expanding gas INJECTION capacity, to offset the increased pressure decline due to excess gas production

I have listed these in approximately ascending order of difficulty, and PEMEX's response plan doubtless includes elements of all four. But we don't know what, or when, or how much, or whether it will be successful.

The decline we have seen so far is probably a combination of (a) with continued drilling and workover ops i.e. some (b). The field decline regime won't switch neatly overnight - something this size can take months to react.

Bottom line is that you should expect the decline curve to be "roll over" and remain indefinite for the next few years. If you revisit in 2010 then the decline mode should be amazingly obvious - in retrospect. But there are so many physical, economic and human factors interacting at the moment that you really don't have much chance of fitting anything simple to observed production history. I don't want to pour cold water, but unless I was a PEMEX insider, I wouldn't bother trying.

Thanks again for sharing your knowledge!

Your comment shows how complex the reservoir dynamic can be! Clearly, the curve fitting approach proposed here can be put in the "wild-ass guess" estimate category but what else can we do considering, as you said, the countless physical, economic and human factors involved.
 

I think what Khebab and others have tried to do (and thus have a significant advantage over "insider" reservoir engineers) is to use statistical methods of aggregate sets of data and to extract much better predictions than you can do in analyzing individual fields.

In other words, it might not work for an individual field like Canterell but you get enough of this data, you better watch out.

The "wouldn't bother trying" remark is what causes everyone to stick their head in the sand. We are certainly not going to get the bottiom-line from the oil companies, are we?  Over fifty years of prior marketing says no.

This is a bullshit comment. You are much more capable that this. Don't let your emotions get the better of you. Love.
It's kind of hard to do any real math on this bloggy-thing you know, what with the lack of good markup capabilities and all.
Hi WebHub,

Engineering and curve extrapolation aren't really in competition with each other. PEMEX will make their own decisions, actively or by default. Those decisions will modulate the physical evolution of the reservoir. I don't know what they will decide, and I don't think Khebab does either. Do you?

You're right in asserting that HL often works if you've got a hundred independent reservoirs, whatever the shape of the individual production profiles. HL on "Mexico apart from Akal" might work just fine - the Mexican oil industry goes back to the early years of the last century, and reservoirs are being discovered, developed, produced and abandoned all the time.

As you have also implicitly noted, the problem with HL on a single reservoir is that the laws of physics and the whims of oil companies usually don't conform to Hubbert's curves. Just look at the shape of the Akal profile pre-peak. Some of my decline scenarios could look surprisingly Hubbert-like; others won't.

The best forecast for Mexico would probably be an engineering calculation, updated by reference to the trade press, for Akal, added to a simple functional form (Hubbert, or maybe just hyperbolic decline) for the rest of the country.

Khebab - does your database of country production profiles include anything that you would point to and say "No, OK, Hubbert gets it wrong here"? I'm just wondering about your criteria for goodness of fit. How do you deal with multiple peaks, for instance?

Khebab - does your database of country production profiles include anything that you would point to and say "No, OK, Hubbert gets it wrong here"? I'm just wondering about your criteria for goodness of fit. How do you deal with multiple peaks, for instance?
There are many cases where the Hubbert approach fails especially when production is heavily constraint for political/economical reasons such as (e.g. Saudi Arabia, Iran, Russia) or when one field dominate the production with not many small fields bieng developped. Generally, it fails when the logistic growth rate (K) is not stationary and varies because of different production regimes. See the example proposed by DuncanK.
My answer is not complete:

How do you deal with multiple peaks, for instance?

Jean Laherrere has proposed a multi-peak approach:

Multi-Hubbert Modeling

If you look at the mathematics, Hubbert Linearization amounts to fitting the cusp of the curve to (1-kx<sup>2</sup>).