Ace - thanks for much additional insight. I'll try and answer the main points:

However, the IEA average observed decline rate for post peak fields is 6.7% which is about 10% higher than CERA's estimate of 6.1% (Table 1 CERA report Oct 2007). This should imply that the IEA's decline rate for FIP should be about 5%, higher than the year old CERA figure of 4.5%, but it's not.

Everyone needs to be wary of comparing segments of IEA with CERA. They use the same terminology as each other, but apply different definitions to that terminology. CERA Figure 5 shows their definition of "build up" - zero to 80% of peak, and plateau - 80% of peak either side. CERA Table 1 is for Post Plateau - so that is all fields that have already declined beyond 80% of peak. The IEA definition is given in Box 10.3, page 235. Their definition of plateau is 85% of peak. But they start to measure decline from peak - Decline phase 1 is form peak to 85% of peak. CERA decline phase 1 is from 80 to 50% of peak.

But this doesn't explain your observation since the IEA definition is more conservative and should be lower than CERA - but as you point out, its not. From memory, CERA Table 2 provides an excellent summary of their findings that is lacking in the IEA report. The latter contains so many different definitions, for me it is near impossible to follow.

On the same theme, the IEA say this, page 221.

The decline rates for fields not included in our data set are, on average, likley to be at least as high for the large fields in our database. On this basis, we estimate that the average production-weighted observed decline rate world wide is 6.7% for post-peak fields

On the same page they note that the post-peak decline rate is 5.1% for their data set, and so they are adjusting this upwards by 1.6% to account for higher decline rates in the myriad smaller fields not included in their data base. This seems a very reasonable thing to do. I don't believe that CERA made such adjustment, and so on this basis one may expect the IEA decline figure to be higher than CERA - but its not.

My cynical view of the 2030 IEA 105 mbd liquids forecast is that it had to meet two key criteria to satisfy political objectives. First, the forecast had to be at least 100 mbd in 2030. Forecast production below 100 mbd would be seen as too pessimistic, although more realistic. Second, no peak in the production could be shown. Both these criteria were met by the IEA's forecast.

IMO it is a pretty straight forward exercise to calculate decline rates from this data set of 800 fields and to then apply the results consistently. I'd estimate its 1 to 4 weeks work and the results could be summarised in a 2 page report. So I understand your cynicism born out of reading hundreds of pages of technically detailed prose that do not stand up to scrutiny and cross examination.

In deep water:

Table 10.8 p 238

Post-plateau average estimate for deep water from the IEA is 11.2%. CERA table 2 quote 17.9% decline for deep water fields. The IEA post-plateau figure should be broadly equivalent to the CERA decline figure.

Last time I looked at UK North Sea decline rates (Nov 2006) the underlying decline rate was 13% moderated to an observed decline rate of 7.6% by new field developments. The underlying decline rate incorporates operating activities like in fill drilling and EOR, thus the natural decline rate will be somewhat higher than 13%. But for the purpose of production forecasting it is a figure close to 8% that should be applied.

http://www.theoildrum.com/story/2006/11/19/135819/75