This post is part of the GeoConnect³d blog.

Last week we had all our attention at the GeoERA Webinar Series, where updates from all the themes in the research programme were discussed. It was especially interesting to see how projects have come a long way with intra and cross-thematic collaborations.

About that, you might remember that back in June we had an interesting discussion with invited speaker Koen Beerten, working in the VoGERA project, about groundwater resource in our GeoConnect³d webinar series. If you missed it, you can watch the video here. A few interesting questions were left unanswered due to time constraints, so we invited him once again, along with Alberto Casillas, PhD student at Ghent University working on the Roer Valley Graben model presented, to discuss them.

Can hydraulic properties change along the fault? It seems from the model that only the northern part gives higher heads along the fault.

Alberto: Yes, the properties can be changed (and has been changed (tested) along the fault) with properties from both sides of the formation where it is located. This was the first step when including faults using the horizontal-flow barrier (HFB) MODFLOW package. And yes, all the faults mentioned by Kris Piessens have been included as well as the new H3O model from VMM. The hydraulic head observations only show higher heads along that area on the region, with a step of around 1.5 meters. By checking the sensitivity of the faults, Rauw fault has been the most sensitive. Then, the properties of the fault were optimized (standard particle swarm optimization, spso11). However, by assuming homogeneous hydraulic properties on the faults the hydraulic heads were the most closely modelled.

It seems that the fault is quite close to the boundary of the model. How does the boundary condition influence the results?

Alberto: Yes, it is relatively close to the boundary. The model boundary is a no flow boundary from the Nete catchment. Actually, the particle tracking shows that most of the flow comes from the recharge area (the eastern corner of the model). This is shown by previous and current NAM models.

What techniques were used to model or analyse the groundwater flows?

Alberto: Techniques: groundwater flow and transport modelling, and groundwater sampling and temperature-depth measurements. The area has been studied extensively in the past years. Currently, most (if not all) of the data existing in the area is being used, retrieved from VMM and gathered from previous projects in the area.

Using the HFB package, you have to simulate the fault as vertical.  Do you have some thoughts on how to include the actual geometry of the fault?

Alberto: Yes, the fault is simulated as vertical. The model cells are 400×400 metres, however, the Rauw fault for instance has a horizontal displacement – the lower boundary of the groundwater flow model (the Boom Clay) in respect to the upper part, of around 150 meters. Given this, discretizing the fault as vertical seems a good approach (displacement<model cell dimension). However, including the actual geometry of the fault is something I have not explored but I would think that having smaller cells in a very local model, and include the fault properties as a formation itself might be an idea but this would be a very expensive model. (maybe also an unstructured grid might do the job).

Kris: It may be useful to point out that the structural framework that we use to link, in this case the geological 3D and groundwater 3D models, allows that these models are upgraded, or improved ones are added. In this case, the Rauw fault is a defining element, linking the geological and hydrogeological models, but it is obvious that e.g. due to modelling constraints the geometry of that fault will be different in both models. If those differences (simplifications) are significant can be judged from the uncertainty envelop around the fault (the buffer). A geometrically more detailed groundwater model, would result in a better match.  

Are you also analysing groundwater quality in the region? If yes, is it showing any interesting results?

Alberto: Yes, there are a few interesting results to be published in the future, but it is still work in progress.

The Roer Valley Graben may be the most important cross-thematic pilot area in GeoERA and besides activities in GeoConnect³d and VoGERA, I believe there are also activities in HIKE, RESOURCE and HOVER. I look forward to seeing the final combined products for the GeoERA Information Platform / the European Geological Data Infrastructure (EGDI).

Koen: Yes, I agree. The products that are being produced in VoGERA include for instance the hydraulic properties of normal faults. This knowledge can directly be transferred to the H3O-PLUS workpackage in RESOURCE. Besides this, the fault zone is a good example of a geomanifestation, which is one of the themes covered in GeoConnect³d.

Renata: We were pleased to have Koen and Alberto, who are involved in VoGERA, joining us to discuss groundwater in the context of subsurface use and interactions. We hope to promote more actions like this where we can get together, discuss common areas and challenges with other GeoERA projects too. We believe this helps promoting lasting cross-thematic collaboration within the research programme.

Compiled by Renata Barros and Kris Piessens
Geological Survey of Belgium (GSB)

Note: Some questions were rewritten for clarity

Note 2: This blog is optimized for viewing in Chrome or Firefox. Click on images to display in full resolution.