Each production forecast for a megaproject assumes simplistically that peak production starts on day 1 and decreases according to the corresponding annual decline rate.

Sorry, but what?

My understanding is that the nominal case is production takes a time to reach a value, maintains a rough plateau for a while, then declines at a rate determined by geology as much as economics.

Is the XLS file around anywhere so that we can see what would happen with more appropriate shapes?

And on a related matter, the megaprojects data is all very well, but by the nature of planning cycles you wouldn't necessarily expect to have any data on new megaprojects past the 2012 timeframe as yet. To me that puts a big hole in predictions of 2011+ timeframes - which is just where we would need good data to access decline rates. How can that data be found?

the nominal case is production takes a time to reach a value



It used to be that you would bring the first well into production and then drill step out wells, develop better understanding of the field and then bring additional wells into production so that the production profile showed a gradual increment toward a peak for that field.


It is current common practice to assess the field and establish an overall production strategy for the entire field and then drill all the required wells so that peak production is achieved much earlier in the life of the field.


The assumptions made by ACE are not necessarily incorrect.

Agreed, but it still takes time to drill those wells - they don't appear overnight. That's particularly true of a larger field (eg megaprojects). We are talking a minimum of 1-2 years to reach that plateau stage.

Compare the shape of these graphs with those others have provided for forecast production and you can see the impact the assumptions have on available production at any date. In essence its a convolution of "field coming onstream" events with curves for expected production volumes. Different convolved shapes will have a cumulative impact on total volume (eg not getting it right is a systematic rather than random error).