# Master Thesis Topic

## Introduction
Due to the increasing need for heat for greenhouses in the area, the Straelen region is interested in geothermal potential. The goal of this project is to better characterize the fault zone in the area as a possible pathway for geothermal development. There are two 3D seismic lines available that cross the region of interest. Their interpretations display a typical fault zone with signs of structural complexity as features disappear within a distance of 400 meters. Consequently, a better understanding of this area is required. The gravity method would help image the density contrasts between the opposing fault blocks. They are composed of the Massenkalk and Kohlenkalk formation. Regional gravity data are available, however to specify the area of interest, a more complete survey with closer-spaced sta-tions and greater sampling is required. This project would involve acquiring gravity stations in the Straelen region at the beginning of November. The acquisition is then followed by processing using Fatiando (Uieda et al., 2013) an open-source python library to process, model and invert gravity data. More geophysical methods can be considered to image the complexity of this fault zone at different depths for a joint interpretation.
There are other geologic features of interest to further develop geothermal in the area. The Mas-senkalk is known to be more permeable formation in the area and could contain hydrothermal fluids associated to the fault zone. The magnetotelluric method would be adequate geophysical method to image changes of conductivity at approximately 1 km. However, its acquisition is challenging due to the presence of power-lines, train tracks and cultural noise that cause interference with the useful signal. Geophysical forward modeling would be necessary to determine whether this method is feasi-ble in the region. The candidate would use GemPy (de la Varga et al., 2019) to model the geologic features of interest and use SimPEG to create synthetic data and see different scenarios of signal-to-noise ratios.
The work packages include:
- Literature review, getting acquainted with python, Fatiando, GemPy, SimPEG, version control, documentation.
- Understanding implicit geological modeling and different geophysical methods.
- Understand the sensitivity of the starting model given the number of layers and features.
- Good practice of geophysical acquisition. (Organization, taking notes, maintain the equipment)
- Process gravity data and understand noise estimates.
- Write thesis and potentially write or contribute to a scientific publication.

## Meeting Notes
## Thursday, 10 Nov. 2022

#### Agenda
- General scope of the project.
- Thesis goals
- Schedule organization
- Miro board https://miro.com/welcomeonboard/dFJ6ZHVYa1VGTFJjd2MzWEsyTExXQmhzcVFXMDFIR0d1MHlQQXUwS0pUdGRTZjBDUTk2d2NnWnZ1WWtOZ1Q4S3wzNDU4NzY0NTIwNTgzMjM4NjM5fDI=?share_link_id=258660699861
- HackMD
- Weekly meetings
- Sciebo
- Operational Plan Gravity Survey
- GemPy Model

#### References
https://www.gempy.org/tutorials
https://github.com/leouieda/gravity-processing

#### Gravity Reference stations
http://bgidata.get.obs-mip.fr/cre_sketch_station.php?numref=020511&ordre=1&g=981184480

http://bgidata.get.obs-mip.fr/cre_sketch_station.php?numref=101601&ordre=1&g=981097076

## Thursday, 22 Nov. 2022

### Research Module Final Report Discussion and Outline

- Motivation (why is geothermal development important in this region)
- Geologic background (lithologies of interest )
- Available information (seismic and well log data)
- Gravity method theory ( why gravity would be useful for this problem/case)
- Instrumentation (describe the gravity meter)
- Procedure in the field ( how to read a gravity measurement)

*In case insurance does not get approved in time, we can take some test measurements and include in this report*

- GemPy model (ask Florian)

## Tuesday, 10 Jan. 2023

- Process the gravity data using Leo tutorial
- Update on GemPy model
- Research module report

### Research Module Final Report Discussion and Outline - Edit

- Motivation (why is geothermal development important in this region)
[1 page]
- Geologic background (lithologies of interest ) [3-4 pages]
- Regional geology
- Venlo block
- Available information (seismic and well log data)
- Gravity method theory ( why gravity would be useful for this problem/case) [1 page]
- Instrumentation (describe the gravity meter)
- Data Acquisition [Procedure in the field ( how to read a gravity measurement)] [1 page]
- Data Processing [Corrections][2-3 pages]

### To-Do:
Tomasz
- Process gravity data using Leo tutorial
- Work on Report
- Try the changes mentioned in the GemPy Code
- Push the GemPy code to the join repo

Andrea
- Push the comments to the processing code to repo
- Add the .tiff information to the code

## Tuesday, 17 Jan, 2023
Tomasz
- GemPy model uploaded on github. Model made smaller, increased number of points around fault. Makes geological sense, however no satysfying geometry obtained yet.

## Friday, 03 Feb, 2023

### To-Do:
Tomasz
- Add the relay ramp to GemPy model
- Interpolation of Gravity points

Andrea
- Ask Sofia how to add this to the fault
- Integrate Tomasz into Fraunhofer team

## Tuesday, 14 Feb, 2023

Tomasz
- Add more surface points to faults
- Transfer to jupyter notebooks
- Look at order of fault

Andrea
- Link numpy array to verde format for gridding.
- Grade report

## Wednesday, 15 March, 2023
Tomasz
- Add uncomformities to the GemPy model
- 3rd version with 3 main densities (tertiary group, middle group, massenkalk)
- Forward model Gravity
- https://www.fatiando.org/verde/latest/gallery/spline_cv.html#sphx-glr-gallery-spline-cv-py
- Check seequent license

Andrea
- Look for GemPy forward model (add_surfaces_values)

## Wednesday, 29 March, 2023

Tomasz
- Add contour lines to the spline interpolation,
- Try different density values for the Bouger correction,
- GemSys -> seismic tutorials at Seequent platform,
- Plot the gravity stations into the 3D models
- Examine different forward signal depending on the fault dip change,
- Try to take GemPy model into PyGimli,
- Look for the original seismic lines from NLog,
- Look for the Petrel model by Phillip

Velocity model from the seismic could be perhaps used for densities (relationship with impedence)

Note for later stages of the work:
attempt horizon constrained inversion for densities in the south of the model, and then using those refined values attempt density constrained inversion for the structure -> get more information about the divergence of the fault.

## Wednesday 05 April, 2023

Tomasz
- Quantify the results from dip test
- Coordinate system revision (try using oasis montaj for this as well)
- continute Oasis montaj gmsys. Import seismic SEGY and create gravity stations
- import gravity data ( first correct coordinates )

## Wednesday 19 April, 2023

Tomasz
- Work on GM Sys model, observe changes of the forward gravity, trend fitting.
- Grid the data and work with stations.
- Import southern seismic line in good resolution, possibly use SeiSee/Geosoft SEGy import.
- Leave fixing the 3D GemPy model for later in may (location of the fault vs location of the stations, pick better 'arbitrary points')

## Monday 8, May, 2023
Tomasz
- Forward model in GM Sys for the north line with new densities, pay attention to the fault angle
- Save maps as good quality figures for presentations and thesis : grid the data, for almost every column of the dataset
- Prepare for the next week PyGIMLI and GemPy reinstallation

## Monday 19, June, 2023
Way forward
### 2.5D Oasis montaj GM-SYS (methodology)
- input (segy reader)
- re-digitise
- assign density values
- calculated vs observed traces
### Gempy -> PyGimli
- data observed
#### 3 possible scenarios:
### Inversion - 3D
- observed data
-Error estimation
-mesh
-inversion parameters -> regularisation
- Resulting model vs gempy
### MT - forward model
- another method
- fluid occurence
- chasing resistivity
- resistics
### uncertainty quantification
- fault angle study on steroids
- how much is geophysics helping
##

### Plan for the presentation
- motivation
- geology
- field campaign
- processing
- oasis montaj
- gempy
- gempy-pygimli

### Presentation feedback 21, June, 2023
- Norbert: take into the considerations the uncertainties in the densities assigning
- Florian, Miguel and George: double check the gravity script and weird behaviour of the forward model, making sure the transition is smooth (sigmoidal function)
- Gabriella: think about the uncertainty of the depth of top Kohlenkalk
- Miguel: plot forward modelled gravity at stations just as points, over the interpolated graph to double check
- Tomasz to himself... maybe also plot exact values, but then need to take the account for the fact that Gempy is not doing moho so would need to somehow normalise

### Working on the 3D grav inversion
- import survey data for d_obs
- create a test cube to check if d_obs needs to be in every cell of the mesh
- error population, for now artificial as no repeated measurements
- forward operator -> defined function from dev version of Pygimli

## Inversion continues
things to consider:
upper and lower limits of the desired inversion - transcol(-2,2)

correlation lenghts - geostatistical constraints tutorial,
higher the value, the smoother the result in each direction

Florian geostatistical inversion paper

cType2 - can do first and second order smoothing - > regularisation parameters from pygymli

inv. - > inv things to plot
ex, data mistfit, chi - two things, side by side next to inversion result

density distribution, not too drastic values to be detected, try different magnitudes

could be error, that;s why no inversion in distributed inv

error required for each of data points, not just noise lvl

- prepare

GEMPY MODEL
PG MESH
DATA
raw data points
nbouger grav
interpolated bouger anomlay