Gas production from gas wells and gas production from oil wells are shown in figure 11 and to study in detail the evolution of the associated gas production from oil wells is compared to oil production in figure 13.

Figure 13: Gas production from oil wells and petroleum production.

It is striking that associated gas production was in line with oil production from 1950 to 1979 but from 1979, oil declines when gas increases.

In his book "Alternative energy resources" in 1976, James Hartnett gave on page 35 his forecasts on oil & gas US production until 2020. He forecasted for oil a continuous decline of the Lower 48, a new cycle with Alaska Prudhoe Bay since 1975 and another cycle due to EOR (Enhanced Oil Recovery), expecting 2.1 Gb/a in 2000 for conventional oil and another 1 Gb/a for EOR. In fact oil US production was at 2.1 Gb/a, but the dreams of EOR did not come true. But for natural gas he forecasted also Alaskan production in addition to declining Lower 48 with a total of 15 Tcf/a in 2000 and an additional 12 Tcf/a for new techniques (27 Tcf/a in 2000), but all declining since 1985. In fact North Slope gas is not there, total production for 2000 was about 18 Tcf (dry), but rising since 1985. Hartnett was all wrong.

Why has gas production been increasing since the trough of 1985? It comes in part from gas wells but also surprising in part from oil wells, despite the decline of oil. As shown in figure 13, gas production from oil wells rises from 1985 when oil declines sharply: Why?

It is well known that at the end of an oilfield associated gas production rises. The increase of gas production from oil wells seems to indicate the near end of most US oilfields. This behaviour was not forecast by most of the experts.

One good example of increasing associated gas with declining oil is shown with Brent oilfield in UK and Parentis oilfield in France.

Figure 14: Brent oilfield production of oil and gas.

From 1996 oil declines and gas rises sharply. GOR (gas oil ratio) jumps from 2,5 to 8 Gcf/Mb.

Figure 15: Parentis oilfield production of oil and gas.

From 1993 GOR jumps from 0,05 Gcf/Mb to 0.25, when the oil production keeps declining.

Reserves

The remaining reserves (which are future production) vary by author. USDOE data shows a change in value in 1979 when they took over the duty of reporting from API (American Petroleum Institute) and AGA (American Gas Association).

Figure 16: Remaining US reserves.

In the US to comply with the SEC (Securities & Exchange Commission) listed companies have to report only proved reserves (estimated with reasonable certainty to be commercially recoverable,), omitting probable reserves. In the rest of the world reserves are reported as proven + probable. Under the SPE/WPC/AAPG rules, proved reserves are defined as "a high degree of confidence" under the deterministic approach or with a "90% probability to be equalled or exceeded" under the probabilistic approach. "Probable reserves are those unproved reserves which analysis of geological and engineering data suggests are more likely than not to be recoverable. In this context, when probabilistic methods are used, there should be at least a 50% probability that the quantities actually recovered will equal or exceed the sum of estimated proved plus probable reserves". There is a contradiction in the deterministic approach where probable being as likely or unlikely i.e. 50%, when in the probabilistic approach it is proved + probable being 50%.

The poor US practice on proved reserves leads to a very strong revision and the concept of reserve growth, whereas using the mean (or expected value) reserves for proven + probable results statistically in no growth.

The experts are clear:

Ross 1998 wrote: "The term "reserves" often is treated as if it were synonymous with "proved reserves". This practice completely ignores the fact that any prudent operator will have, at least internally, estimates of probable and possible reserves."

DeSorcy 1993 for the Canadian Standing Committee on Reserve Definition wrote: "There are currently almost as many definitions for reserves as there are evaluators, oil and gas companies, securities commissions and government departments. Each one uses its own version of the definitions for its own purposes."

Capen, one of the best US experts in reserve definition, wrote in 1996: "An industry that prides itself on its use of science, technology and frontier risk assessment finds itself in the 1990s with a reserve definition more reminiscent of the 1890s" "illegal addition of proved reserves."

It is amazing to find in the oil and gas industry so much advanced technology such as the deepwater and so much obsolete practice such as using the old term of bbl (being the blue barrel: why blue?) or refusing to use the system of international units (SI) (it was the cause of the lost of the Mars Climate Orbitor which crashed when NASA sent instructions of thrust in newtons when Lockheed has built it in pounds), still using M for thousand when the layman uses Y2K and not Y2M. Most of US gas producers refuse the probabilistic approach in estimating the reserves, as they use only area (spacing) in acre, net pay in foot and recovery as cubic foot per acre-foot from the closer field, or often ten times the annual production!

One of the biggest mistakes in dealing with proved reserves is that aggregation of the proved field reserves does not give the proved reserves of the country or a basin, it underestimates it. But every official agency or media does it. Only the addition of "mean" field reserves corresponds to the "mean" reserves of the country.

The comparison between the current proved remaining reserves and the backdated mean (proven + probable) reserves (corrected using MMS reserve growth model) shows a large difference in volume and in decline rate. The "mean" reserves have declined since 1965 and the slope is constant since 1980 at 2.5%/a, whereas the current proved stays flat since 1988.

Figure 17: US remaining gas reserves: current proved versus backdated mean.

The biggest problem in forecasting US production is the poor quality of the reserves data because of the SEC rules and the conservatism of the gas industry in reporting data.

Gulf of Mexico

Data is difficult to gather onshore US because of the number of producers and confidentiality of the data - gas is owned by the landowners. It seems that the situation should be better in the federal waters of the Gulf of Mexico = GOM OCS (Outer Continental Shelf) where oil and gas is owned by the Federal Administration (selling leases to private companies) and productions are managed by the USDOI MMS (Minerals Management Services).

It is surprising to see the increase of the federal lands in the production of fossil fuels from 3% in 1950 to more than 30% in 1998 and it is likely to continue in the future and to increase more if the access to federal lands is more open to exploration and production.

Figure 18: Percentage of production in federal lands.

Reserves have been changed in the 80s because of the change in definition from API (American Petroleum Institute) to SPE (Society of Petroleum Engineers) rules.

Figure 19: GOM OCS: evolution of original reserves and production.

The incomplete MMS data gives a distribution of gas reserves and production versus water depth, which shows that gas is mainly in the shallow waters.

Figure 20: GOM percentage of gas discovery and production versus water depth from MMS.

The MMS data displays a peak in the annual number of discoveries (in the graph the number is multiplied by 10) around 1984 (70) and that the average size of the original reserves which was around 0.6 Tcf in the 50s has sharply declined since 1973. The average water depth of the discoveries, which was around 200 ft in the 70s, is around 150 ft in the 80s and the first half of the 90s.

Figure 21: GOM OCS: annual average size and number of gas discoveries from MMS.

The cumulative gas discoveries (from MMS) shows that the main part comes from water depth less than 1000 ft, but this part decreases since 1980.

Figure 22: GOM OCS cumulative gas discoveries detailed by water depth from MMS.

The MMS proved gas field (984 fields) reserves grown to 2P (or mean) with MMS reserve growth factor (the first year estimate is multiplied 50 years later by 4.5) are plotted with a fractal display (rank of increasing size versus field size in a log-log graph) for each decade. This graph shows that the largest fields are found first and that an ultimate curve can be drawn.

Figure 23: GOM gas field size from MMS: fractal display.

MMS provides on the web the data up to1998 giving the detail of the evolution of every field since 1975 in the GOM OCS, totalling 984 fields. Unfortunately the huge database (197 pages) on reserve history shows on the first page for the first field cumulative oil declining at the fifth line (for 1979), meaning that the annual oil production was negative (which is impossible!). It is obvious that operators sent flawed data (change in grouping?) to MMS, which did not bother to check if it is correct.

Figure 24: GOM OCS: MMS proved & industry unproved discoveries.

We have compared for the deepwater of the GOM the data given as oil equivalent (initial) reserves from MMS, Infields and Wood Mackenzie (WM). The comparison for the same field between two files X and Y is given in percentage of change ((X-Y)/Y*100) and the percentage is ranked by increasing percentage. The plot shows that the change varies from —80% to +150%. The total of the reserves for about 50 fields is 7.5 Gboe for Wood Mackenzie and 7.4 Gboe for Infields, being close, but the difference was larger when compared with MMS. For 20 fields MMS total is 2.3 Gboe when WM is 3 Gboe. For 27 fields MMS total is 3.3 Gboe when Infields is 3.6 Gboe.

Figure 25: GOM deepwater: comparison of field reserves from Infield, Wood Mackenzie and MMS.

When these databases are plotted as cumulative discovery versus time, the discrepancy is huge. For deepwater (>1000 ft) the cumulative discovery up to 1998 is 15 Tcf for Infield, but 8 Tcf for MMS.

Figure 26: Cumulative gas discovery up to 1998 versus water depth from Infield and MMS.

It is obvious that MMS database is incomplete having only 39 fields against 132 fields for Infields. The plot versus time shows that the discrepancy starts in number since 1985 and in discovery since 1988.

Figure 27: Cumulative gas discovery versus time from Infield and MMS.

MMS is incomplete and late to report, now only up to 1998.

To get updated data, it is necessary to rely on published papers such as Dodson in Offshore January 2001. Gas production is flattened since 1990 with less than 5 Tcf/a and oil has risen since 1990 because of the deepwater.

Figure 28: GOM oil & gas production up to 2000.

Figure 29: Decline of GOM production from gas wells from 1972 to 1998.

< http://www.eia.doe.gov/oiaf/servicerpt/depletion/table_g1.html >.

Table G1.  Average Production from Wells in the Federal Offshore Gulf of Mexico, 1972 to 1996

Source: Energy Information Administration, Office of Oil and Gas, Reserves and Production Division (Dallas, TX).

Reserve growth

In 1981, faced with the unreliability of the so-called proved reserves, the USGS (in an assessment of undiscovered conventional oil and gas as of end 1979), called them measured reserves, and called inferred reserves the expected additions on these measured reserves, which is now called reserve growth by the new team.

Reserve growth is the most important problem in assessing the US oil & gas potential, as a proper approach needs to deal with "mean" values, called "proven + probable" in most of the countries outside the US and Canada.

Technology allows cheaper and faster production, seldom to increase the conventional reserves (but in contrast to unconventional oil and gas). The best proof is that the depletion rate of gas wells has increased drastically over the last 20 years, being now over 50%/a in the GOM.

What is wrong is the US practice of estimating reserves because the SEC rules prohibit reporting probable reserves as is done in the rest of the world. It is why from 1988 to 1999 new gas field discoveries represent only an average of 5% of the total reserves additions, and the revisions of the past discoveries represents 95%, meaning that the past estimates are pretty lousy. US practice for reporting reserves is very poor (as obsolete as the punched cards of the last presidential elections!). This poor practice leads to a strong reserve growth wrongly attributed to new technology when in fact it is due to poor methodology. But such practice allows oil & gas companies to report growth even without any discovery and they love it.

The evolution of the percentage of the original reserves versus the reported value in 1998 for 12 largest gas fields discovered from 1948 to 1969 (average year of discovery 1955 and average water depth 60 ft) shows a large range of variation but the average rises sharply from 1975 (first year of data) to 1985, then the estimate went down following the counter shock of 1986 to return in 1998 to the 1985 value.

It means that the evolution is stabilised and no significant reserve growth should be expected statistically, but despite that some fields can still vary up or down.

Figure 30: GOM gas fields: evolution of original reserves with time.

The deepwater production has really only started since 1990 as shown in figure 28 and the behaviour in these new conditions is still immature. Some deepwater development have shown some disappointment as the Shell Macaroni field (developed with subsea equipments peaking at 15 000 b/d instead of the 35 000 b/d expected).

The study of the production of individual mature fields shows a different story from what is reported as reserves. The three largest GOM gasfields (all found before 1958) are TS000 (Tiger Shoal), VR014 (Vermilion) and VR039 with about 3 Tcf each.

The evolution of original reserves given by MMS since 1975 shows that for TS000 (found in 1958) by 13 ft of water) the ultimate was about 3.7 Tcf around 1985 to go down to 3.2 Tcf in 1998. The evolution of the ultimate with up and downs is not smooth, likely coming from goals varying with years.

Figure 31: GOM Tiger Shoal gas field: reserves evolution.

The analysis of the annual production versus cumulative production shows a decline from 1975 to 1990 giving an ultimate around 3.5 Tcf (explaining the 3.7 Tcf in 1985) but since 1991 the depletion has been accelerated and the present ultimate is around 3.3 Tcf against MMS estimate of 3.18 Tcf.

Figure 32: GOM Tiger Shoal gas field: production decline.

The fourth largest gas reserves Eugene Island 330 oilfield (found in 1971 in 246 ft of water and in 1998 the largest oil and gas original reserves with 760 Mboe) is interesting as this field was taken by the Wall Street Journal (Cooper 1999) as the example of refilling of reserves and to wonder about a deeper origin for oil, and to explain the huge increase in the Middle East reserves in the second half of the 80s. In fact, because of the large depletion of pressure, this field seems to have been charged again from the source-rock, by one of the largest (the Red Fault) and best faults in the GOM studied by many university seismic surveys in 4D (http://www.ldeo.columbia.edu/4d4/talks/expl/index.html). But again the first oil estimates reported to MMS seem odd and different from OGJ estimates and the increase in1995 due to an increase in annual production is minor compared to the decrease of 1988. An analysis of the pressure should explain the recharge of the reservoir.

The difference between MMS and OGJ on the cumulative production seems to be mainly discrepancies between 1977 and 1979.

Figure 33: GOM Eugene Island 330 field: gas reserves evolution.

The new increase in annual oil production starting in 1993 peaked in 1996 and declined again. The last value plotted for 2000 is only an estimate from the month of May in MMS last report. The pre 1993 ultimate estimate from the decline was about 350 Mb, after the recharge the estimate is now about 400 Mb (416 MMS & 389 OGJ).

As for oil production, the EI 330 associated gas production declined since 1977 to 1993 to rise until 1996 and declined again. Before 1993 the ultimate could have been estimated at about 1.75 Tcf, after the recharge the present decline rate suggests (questionable value for 2000 as only for one month) the ultimate is about 2 Tcf (1.93 Tcf MMS).

Figure 34: GOM Eugene Island 330: gas production decline.

In the DOE/EIA annual report 1997 the reserve growth on federal offshore is given as 33 Tcf.

We believe that it is completely wrong to measure the reserve growth as a ratio of the estimate of proved reserves at discovery year. The estimate is only reliable after a few appraisal wells. Furthermore the proved reserves depend on the number of wells drilled and it is better to wait for the complete development of the field to assess the real value of the proved reserves. In the past when API was in charge of reporting the reserves, the first six years after discovery were considered as unreliable. Offshore the development is decided only after appraisal wells and the only reliable estimate should be the first one at the decision of development which in fact is decided not on "proved" reserves but on "mean" reserves, if not the size of the development is underestimated costing money later.