Finding the right stuff – Oil vs. gas and the perils of pre-drill prediction
OR
About the Course
PESA members, view the course for free!
In 1980 BP drilled the Phoenix-1 wildcat well in the Bedout Sub-Basin, a hitherto unexplored area of the North West Shelf (NWS). Although the province is famously gas prone, the target was oil. Phoenix-1 found some gas shows in tight reservoir, discouraging further exploration until Apache, Finder and Carnarvon Energy drilled Phoenix South-1 some thirty-four years later. The pre-drill expectation for hydrocarbon phase was gas but instead the well found under-saturated light oil. This led to headlines of a “new oil province” – but only until Quadrant Energy and Carnarvon Energy drilled Roc-1 up-dip from Phoenix South-1. The expected phase for Roc-1 was oil but it found rich gas-condensate. This led to the next prospect up dip – Dorado – being predicted to contain gas condensate.
As everyone now knows, Dorado-1 turned out to be the largest discovery of oil on the NWS in 30 years. With uncanny timing, Santos acquired Quadrant Energy just before Dorado-1 commenced. The story did not end there though: Figure 1 (title image) shows the hydrocarbon distribution in the Dorado field as reported before and after drilling of the first appraisal well.
Can we really predict what phase of hydrocarbons a wildcat prospect will contain or even whether it will be single phase or dual phase? This talk examines the factors that control hydrocarbon type and phase at the prospect level and the attendant uncertainties. It will be shown that, while prediction is possible at the whole of system (or play) level, uncertainties multiply when it comes to prospect level prediction. Part of the reason for this is that phase, migration and trapping are inextricably linked. Furthermore, there is a fundamental “blind spot” in traditional petroleum geochemistry practice – one which has caused misunderstanding of the dynamics of numerous petroleum systems around the world (including the Cooper-Eromanga Basin).
But it’s not all bad news: I will finish by describing (a) a template that uses the main factors controlling oil vs. gas to form a “Bayesian Prior” prediction (no basin model required) and (b) a new tool for probabilistic prediction of prospect phase linked to column heights, trap fill and volumes (Figure 2).
Figure 2: Probabalistic trap fluid distribution incorporating fluid, trap and seal rock
properties (proprietary to Zetaware)
Your Instructor
Following an early career in analytical chemistry and marine science/oceanography, Andrew was a senior research scientist with the Australian Geological Survey (now Geoscience Australia). In 1998 he joined Woodside Energy, retiring 18 years later as principal petroleum systems advisor. This period included a secondment to the Shell international research centre. He now runs his own company (Murray Partners PPSA) providing consulting, training and support to exploration and appraisal/development geoscientists. Clients have included 11 integrated exploration companies and several governments and academic institutions. Andrew holds M.Sc. and Ph.D. degrees in petroleum geochemistry and has 32 years of experience spread across most regions of the world, across government, academia, and industry, and across the basin modeling, geochemistry and reservoir engineering disciplines.