There was a total of 17 discoveries made from 57 exploration wells drilled on NCS in 2019, with the mature North Sea yielding 10 finds, according to NPD (The Shelf 2019).
The average exploration success ratio for 2019 is 30%, which is close to the historic global average. However, the size of the NCS drilled structures is on a declining trend, which means that an improved success ratio is needed for NCS to remain internationally competitive.
An alternative approach
There are several technologies available for reducing exploration risk, including seismic AVO (including rock physics), EM and petroleum systems analysis are three of the most used. All these techniques have been actively used the last 10 years, but the exploration success has not increased considerably.
It is the opinion of this author that exploration during the last few years has been largely driven by the improvements in seismic imaging technologies. Many prospects are defined and evaluated from seismic data only. In particular, the petroleum system is frequently not described when these prospects are presented at conferences. Here we evaluate whether a stochastic approach to petroleum systems modeling can improve the exploration success ratio.
The stochastic petroleum systems modeling approach discussed here uses a “full” 3D petroleum geomodel, which is simulated through geological time within a MonteCarlo loop, through thousands of simulations, incorporating several hundred parameters. The difference between the modeled and observed oil and gas column height in exploration wells is used to compute the weight of each simulation run, which can be used to select the “best” runs.
Testing the validity
The method was developed around 2010, and a Norwegian North Sea study with data from PGS using the new technology was completed in 2013 (reference). Therefore, all exploration wells within the study area after 2013 can be used to test the model. The first test was presented at NCS Exploration – Recent Discoveries 2016 and used a “Flowrisk” map to describe the risk for Mid Jurassic prospects.
The Flowrisk map shows the percentage of the best 1000 simulation runs with hydrocarbons migrating through every location within the modeled area. The model is run with a 1 x 1 km grid resolution. Dark green areas (see figure) have more than 75% of the simulation runs with hydrocarbon migration. Several source rock levels are into play, focusing on reservoirs of the Mid-Jurassic as this level is a well-established play model in the North Sea.
Figure: Modeled Charge probability (% Flowrisk) and drilling results (discoveries) presented in 2016. The discovery wells 35/8-6 ‘Vikafjell N/ Robbins’ and 30/11-12 S ‘Askja South East’ were drilled after the “prediction” was given at the conference.
From the analysis of the 2016 results, we proposed that only participating in the wells with modeled Flowrisks greater than 75% would have increased the success ratio with approximately 13%.
In the period from mid-2016 to late 2019, a total of 37 new exploration wells were drilled (see table). Of these, 22 contained Jurassic reservoirs (Mid/Upper Jurassic), while 15 wells missed Jurassic reservoirs. The most interesting observation from the table is that all but one well with Flowrisk values above 75% contained hydrocarbons in the Jurassic sequence, while only 2 of the wells with lower values than 75% contained hydrocarbons.
Table: Norwegian North Sea wells drilled down into the Jurassic between 2016 and 2019.
One use of the data in the table is to split the data into 3 classes, after drilling:
- Class A: Wells with Flowrisk > 75%
- Class B: Wells with Flowrisk < 75%
- Class C: Wells with missing Jurassic reservoirs
The success ratio of Class A wells is 93% (13 of 14 wells), while the success ratio of Class B wells is 25% (2 of 8 wells). The overall success ratio of the Jurassic wells with a proven reservoir is 68% (15 of 22 wells).
Our interpretation is that by only drilling Class A wells, the success ratio of the Jurassic reservoir play can be increased with 25%.
The “missing reservoir” wells must also be included in the full risk analysis, and for Class A wells the resulting success ratio increases from the total risk of 40% (15 of 37 wells) to 45% (13 of 29 wells).
Needs more testing
The stochastic petroleum systems modeling approach holds much potential, but the method still needs to be further validated.
There are two main reasons for this: (1) the database is still small, and (2) there is some subjective interpretation in determining the Flowrisk values for each exploration well. The Flowrisk values are picked at the well location. Still in cases where there is no trap modeled at this well location, an alternative location within the closure of the trap may be more appropriate to identify the proper Flowrisk value for the trap. Hence, a validation of the method should include publishing of pre-drill Flowrisk values.
It is now our plan to publish pre-drill estimates of Flowrisk for all wells spudded within the North Sea during 2020.
You can follow these predictions here at geoexpro.com, this article is also presented in GEO magazine (GEO no. 1, 2020).
Text: ØYVIND SYLTA
Migris
“Our company describe and reduce geological and economical exploration risk by simulating geological processes. Our unique technology describes geological risk faster and more accurately than ever. We combine a better description of the geology with next generation solution techniques and process descriptions.“
