1 – Geomodeling = integrating interpretations. Geologists’ data are outcrops, rock samples, core and logs among other things. Geophysicists have seismic data while petrophysicists have raw logs to work on. Engineers deal with dynamic data (pressure, production data…). Geomodelers have no raw data per se. Instead, their input is the interpretation each other team member has built about the reservoir. The geomodeler is here to understand these interpretations and integrate them into the geomodels.

2 – Geomodeling requires collaboration with the whole team. Limiting communication with the team to data transfer is risky because, in many teams, an interpretation is more than data themselves. It might be a concept never clearly shared with the team, a drawing on a white board, a set of cross-sections or maps done by hand. Only through discussion with each team member can a geomodeler really comprehend how the team pictures the reservoir.

3 – Pure data-driven geomodels are problematic. A data-driven geomodel is built from the data alone and from their analysis in the geomodeling project through data analysis tools (histograms, cross-plots, variograms…). A data-driven geomodel doesn’t try to capture the intuition of the team about the reservoir or the interpretation done by others. It simply does the full analysis on the fly by assuming there are enough data to represent correctly the reservoir. Unfortunately, it’s never the case. A good geomodel will come from using analysis tools to reproduce or approximate mathematically what the team has interpreted.

4 – Geostatistics represent only a fraction of any geomodeling project. Most of the time of a geomodeling project is spent on tasks that are not related to geostatistics. It might be about cleaning data, spending time discussing with the team or running tools that are not geostatistical tools (like building the 3D-grid, upscaling or property modeling through functions such perm=f(poro) or SW=f(distance-to-contact)). From there, you can’t expect to train a new geomodeler by sending him/her only to pure geostatistics course.

5 – Building a geomodel often lead to refining interpretations. In many teams, it is assumed that interpretations are finalized before the geomodeling work start. In that approach, geomodeling is just there to blindly create an object (the 3D grid) which exactly follow the provided interpretation. In that approach, geomodeling is seen more as a technician’s work than a real important piece of the team’s work. Reality is often different. Bringing an interpretation done on 2D displays (maps, cross-sections, log display) into a 3D environment might show small mistake here and there. Combining the vision (=interpretation) of the geologist and of the other team members might show some contradictions between them. For example, maybe that permeability was studied on rock types which can’t overlap with the facies described by the geologist. Building a geomodel is often the occasion to validate, correct and extend interpretation. It’s like writing a report: often, it’s while writing a formal report that gaps in the thought process are spotted and require additional work. It’s the same for geomodeling and interpretations.

6 – If you can, let a geologist be the geomodeler. The main aspect about any geomodel is probably how the facies are represented in the reservoir. The facies distribution will highly influence the distribution of all the rock properties (NTG, porosity, permeability, SW…). For that reason, if your team has to train someone to become a geomodeler, let it be the geologist as he/she will be best equipped to model facies. Of course, this is not an absolute: any team member can become the geomodeler, as long as he/she can place geology and facies modeling at the core of his/her geomodeling work.

7 – If you can, let someone open-minded be the geomodeler. As geomodeling is about integration of the work of all the other team member, the geomodeler must have a mind open to what other do. You can’t afford to have a geologist do the geomodel if his only interest in life is to make well correlation. Or a pure geostatistician who only trust data-driven analysis. Or an engineer who will focus only on adjusting dynamic parameters to get an history match. Your geomodeler can have any of these backgrounds, but also, he/she must show great respect for what each other specialist can bring to the team. He/she must be able to put aside his/her instincts to be able to listen to someone’s else different approach to the problem at hand.

8 – A good geomodel is one in which each team member finds his/her interpretation of the reservoir. Non-geomodelers in your team don’t need to master the intricacies of your geomodeling workflow. What they need is to be able to look at the final product (a 3D grid with properties modeled in it) and recognize in it their interpretation.

9 – While geomodels are 3D by nature, sharing the results with the team requires more than 3D displays. As they work in a 3D window all the time, geomodelers tend to document their work by simply taking screen capture of their geomodels in these 3D cameras. While such displays can be elegant, they might not be the best for sharing a geomodel with the team. Firstly, looking at a 3D picture can be tricky and the 3D perspective, while here of course, might be difficult to grab on a static 3D picture. Secondly, most specialists have specific displays that they are familiar to use to understand their data. For efficient team work, it might be good for the geomodeler to master these displays. It can be maps and cross-sections to convince the geologist. It can be some objects exported into the geophysical package to help the geophysicist see that their seismic interpretation was properly used. It can be some histogram or cross-plots for your petrophysicists.

10 – Train your people in soft skills to make your team more efficient. Because it requires working efficient with everyone, no good geomodel can be built if people don’t know how to communicate between efficiently. Building geomodels can be testing for your team because it can’t afford to let people work in silos. As such, weaknesses in your team members might show up such as lack of interest for anyone else’s work, hiding information on the pretext that information is power, lack of respect for new technology like those in geomodeling packages, lack of respect for traditional interpretation technics because of too much trust put into algorithm and shiny software interfaces. All such problems won’t be solved by sending your team to more technical courses. But they might if you get them train in project management, coaching, anger management… or whatever else you need them to become for the team to work better as a cohesive entity.

Bonus – Have fun building geomodels. Well done, building a geomodel can become the moment when all the different interpretations take shape into a single object, the 3D-grid, which your team endorse as a great representation of everything they understand about your asset. Keep this in mind while you see your team struggling at some steps. Well done, it’s ultimately all worth it. So, try having fun seeing geomodels built in your team.