Kasper D. Fischer
d0292ce28d
Refines presentation of model uncertainty by adding error bars to the resistivity and thickness values in the output plot. Integrates a new table displaying layer resistivity, uncertainty, and thickness with corresponding uncertainties to enhance clarity for users. Updates environment requirements to include `pandas` for improved data manipulation capabilities.
2.3 KiB
2.3 KiB
resistivityVES
Binder environment using pyGIMLI to do a VES (Vertical Electrical Sounding) inversion.
Overview
This repository contains a Jupyter notebook (VES.ipynb) that demonstrates how to perform 1D DC resistivity inversion using pyGIMLI's built-in VES forward operator. The notebook uses real field data from Bausenberg to invert for a layered earth model.
Notebook Contents
The VES.ipynb notebook includes the following workflow:
1. Setup and Imports
- Imports necessary libraries:
numpy,matplotlib,pygimli, and theVESManagerfrompygimli.physics
2. Field Data from Bausenberg
- Uses real VES measurements with AB/2 distances ranging from 1.0 to 100.0 meters
- Apparent resistivity values (
rhoa) showing variations from ~64 to ~672 Ωm - Error estimates set at 2% for most measurements, increasing to 5% for deeper soundings
- MN/2 spacing fixed at 0.5 meters
3. Inversion Setup
- Configures a 3-layer earth model (
nlay=3) - Uses regularization parameter
lam=1000with a reduction factor of 0.8 - Inverts the apparent resistivity data to determine layer thicknesses and resistivities
4. Visualization
The notebook provides comprehensive visualization including:
- Model plot: Displays the inverted resistivity model as a function of depth (up to 50m)
- Data fit plot: Compares measured apparent resistivity data with the model response
- Both plots are displayed side-by-side for easy comparison
5. Uncertainty Analysis
- Computes model covariance matrix to assess parameter uncertainties
- Displays correlation matrix showing interdependencies between layer parameters
- Generates error bars for both resistivities and layer thicknesses
- Visualizes uncertainties at layer midpoints and boundaries
Requirements
See environment.yml for the complete list of dependencies. Main requirements:
- pyGIMLI>=1.5.0, which requires at least:
- numpy
- matplotlib
- suitesparse=5.10.1
- jupyterlab
Usage
Click the Binder badge above to launch an interactive session, or run locally with:
jupyter notebook VES.ipynb