When will global oil supply stall and start to decline? An important part of the answer to this frequently asked question comes down to oil field decline rates - that is, the annual rate at which production from existing fields goes down. The decline rate is a key link in the chain of factors needed to understand the future of the world oil supply.

If we want a picture of likely oil production capacity in coming years, getting the decline rate right is critical - at least to understand the below ground potential. Of course, the decline rate does not address the important factors above ground - ranging from geopolitics to investment decisions - that will also affect future production.

Although the decline rates for individual fields can be ascertained, a view of the global decline rate has been notably missing in the past. The absence of a definitive, comprehensive analysis of decline rates has generated much speculation and a range of widely varying estimates.

The assumption that production from existing fields is declining rapidly - 8 percent or more is the figure one often hears - has contributed to pessimistic, "peak oil" views about the future.

In an effort to find that missing link and answer the question about the global decline rate, Cambridge Energy Research Associates (CERA) launched a detailed, in-depth research project that culminated in our new study - Finding the Critical Numbers: What are the Real Decline Rates for Global Oil Production?

To do the research, CERA undertook a substantive analysis drawing on the most extensive field production database in the world, developed and maintained by IHS, that includes tens of thousands of fields.

The particular data set used in the study covers the production characteristics of 811 separate fields. These are the biggest producers. Combined together, these fields account for over 60 percent of the world's current conventional oil production and half of total proved plus probable reserves.

The CERA analysis shows the aggregate global decline rate to be 4.5 percent, rather than that often cited 8 percent. This finding provides additional support for CERA's view that oil production capacity, at least in terms of the physical resource, can grow enough over the next decade to meet the anticipated increase in demand.

Forecasting how much existing oil fields will produce in the future is challenging because the amount that can be produced from a field in any given year varies over time (see figure, "Oil Field Production Profile"). Production typically rises during a field's early years. This is the Buildup Phase, which lasts from the onset of production until the field reaches 80 percent of its maximum potential production. After Buildup, a field goes into the Plateau Phase, when it is continuing to produce at 80 percent or more of its maximum potential.

A field then reaches the Decline Phase, which lasts from the time production first falls below 80 percent of maximum until, after a number of years, production declines to a low enough level that the field is no longer economic.

An individual field's production profile is determined by two things - the inherent reservoir physics and the investment strategies of the operators. Fields of different sizes tend to have different production profiles. Large fields - defined in the CERA study as those with more than 300 million barrels of originally present reserves - typically build up over an average of six years, produce on plateau for seven years and decline on average for more than 20 years.

By contrast, small fields - those with less than 300 million barrels of original reserves - build up over an average of three years, produce on plateau for five years and decline on average over more than 14 years.

Future production from existing fields depends on where along the spectrum - from buildup to plateau to decline - the world's fields are in aggregate. Thus, the global decline rate is at the center of any estimates about how much new oil must be discovered and brought on line in the future to replace capacity lost through the decline in production from existing fields, as well as to meet anticipated demand growth.

A small difference in assumptions about the decline rate can make a huge difference in projections about how much new oil production will be needed in the future. For example, the difference between assuming a 10 percent global decline rate, as opposed to a 5 percent rate, translates into a swing in which the world would need to discover and bring on line an extra 18.5 million more barrels of oil per day by the year 2017 - an amount equal to nearly one quarter of total oil production today.

Out of the 811 fields analyzed in the CERA study, 400 are large (more than 300 million barrels of originally proved-plus probable reserves) and the remainder, small (less than 300 million barrels of originally proved-plus probable reserves). They can be broken down into two groups: 375 fields are still in the buildup or plateau phases, with 436 in the decline phase.

When the CERA team analyzed recent production trends in its data base, it found that the aggregate global decline rate for fields currently in production is approximately 4.5 percent per year. This is far lower than the 8 percent figure used by many studies. One reason for this difference, we conclude, is that some analysts focus only on the fields that are in decline, not taking into account the production from the many fields in the world that are still in their buildup or plateau phase.

In the database, 46 percent of the 811 fields are in buildup or plateau. More importantly, these fields account for 59 percent of the current production and 63 percent of the reserves represented by the data set used in the study. Only 41 percent of production is from fields in the data set that are into the decline phase of their production lives.

In the fields that are in actual decline, the production weighted decline rate is 6.1 percent. Growing production from fields that are in buildup or plateau is thus able to offset in part shrinking production from fields in decline.

Some analysts have also focused on particular subsets of oil fields that have higher than average decline rates. Small fields, which are increasingly being developed in mature non-OPEC countries, and deepwater projects, which tend to flow at high rates as a requirement of commerciality, both tend to decline more rapidly.

Another important finding is that annual field decline rates are not increasing with time. Large fields in particular have benefited from increased investment, as well as improved planning and technology. Given adequate investment, low decline rates can be maintained in many fields for prolonged periods, and field life is very often longer than originally projected.

It is likely, according to CERA's analysis, that improved understanding of giant fields' geology and reservoir models over the course of long life cycles has allowed late field expansion that has arrested decline and, in many cases, allowed production to increase significantly.

Focusing on the subset of 436 oil fields that are past the buildup and plateau phase and actually in decline, the CERA research team found notable differences. Production from individual large fields declines at 5.8 percent on average, while in individual small fields it declines at 8.9 percent.

Because large fields account for 86 percent of production - and 95 percent of reserves - in the data set, the average production weighted decline rate for fields overall is, as noted above, 6.1 percent. The average for onshore fields is six percent, compared with 10 percent for shallow water offshore fields and 18 percent for deepwater fields.

It turns out that other factors - in addition to field size and onshore vs. offshore - also influence post-plateau decline rates, including reservoir characteristics, development location, regional setting and operational tactics. Limestone reservoirs, for example, which are more prevalent in OPEC countries, tend to deplete more slowly than sandstone reservoirs.

A key conclusion of the study: There is no evidence that oilfield decline rates will increase suddenly.

The study of decline rates allows for better and more reliable projections about future oil supply. The new study reinforces CERA's parallel research, that global liquids capacity - with "liquids" defined here as conventional oil, as well as unconventional liquid fuels such as gas-related liquids, extra-heavy oils like tar sands, ultra deepwater oil and biofuels - which stood at approximately 91 million barrels per day (mbd) in 2007 - could climb to as much as 112 mbd by 2017.

This outlook is supported by a key conclusion of the study: There is no evidence that oilfield decline rates will increase suddenly.

The CERA study is a signpost that shows we are gaining a better understanding of the below-ground factors, such as decline rate, that will shape the future of world oil supply.

We welcome the further research and discussion that will continue to improve that understanding. At the same time, it needs to be recognized that a range of above-ground factors - geopolitics, investment patterns, rising costs, government decisions and environmental concerns - will continue to have a major impact.

These above-ground factors remain even more difficult to parse than the complex geological and engineering factors that underlie the decline rate of the world's oil fields.