🎯 CryoGrid Optimization Automatization 🎯
Welcome to the documentation of CryoGrid Optimization Automatization.
This project provides a fully automated MATLAB workflow to calibrate and optimize CryoGrid simulations using
measured near-surface temperature for modelling mountain permafrost temperature.
As the variability of snow cover pattern in mountain environment makes snow depth one of the most challenging parameter to calibrate, this workflow is particularly designed to retrieve snow characteristics, especially snow fraction (i.e. the proportion of deposited snow when a snowfall occurs).
It automates the entire process of:
Recovering sensor characteristics measured temperature in an Excel file gathering sensors metadata
Updating CryoGrid input files based on sensor characteristics
Detecting snow periods from daily temperature measurements
Running Bayesian optimization (
bayesopt) to infer physical parametersGenerating statistics and plots for model performance evaluation
🛠️ Key Features
Automated Calibration: No manual parameter tuning is required.
Snow Detection Module: Analyzes daily temperature variations to detect snow cover periods.
Optimization method: Global optimization via Bayesian inference.
Modular MATLAB Workflow: Organized in four stages for clarity and reusability.
Scalable to Multiple Sensors: Adaptability to different sensor configuration.
Parallelized code: Parallelization to reduce optimization time. Workflow with lock mechanisms to prevent conflicts.
📂 Project Structure
CryoGrid_Optimization_Automatization/src/→ MATLAB scriptsdata/→ Sensor data files (CSV) and global excel file with all the sensors metadataforcing/→ Sensor forcing data files (.mat) and automatization forcing scriptsCryoGrid/→ All the CryoGrid scripts and file,CryoGridCommunity_results/is the optimization output folderdocs/→ Documentation (Sphinx)
📑 Documentation Overview
This documentation is organized into several sections: