ticoi 0.1.0

A package to run Temporal Inversion using linear Combinations of Observations, and Interpolation (TICOI)

TICOI

TICOI is a tool to postprocess surface velocity time series estimated from remote sensing (e.g., ice flow, landslides). The method is based on the temporal closure principle. It fuses velocity measurements which are multi-temporal (with different temporal baselines) and multi-sensor (from different satellite images), and may have been computed by different processing chains. It takes as input NetCDF files containing the image-pair velocities, that you may have generated yourself, or natively supports data from the NASA ITS_LIVE project or from Millan et al. (2022).

The package is based on the methodological developments published in:

• Charrier, L., Dehecq, A., Guo, L., Brun, F., Millan, R., Lioret, N., ... & Halas, P. (2025). TICOI: an operational Python package to generate regular glacier velocity time series. EGUsphere, 2025, 1-40.

• Charrier, L., Yan, Y., Koeniguer, E. C., Leinss, S., & Trouvé, E. (2021). Extraction of velocity time series with an optimal temporal sampling from displacement observation networks. IEEE Transactions on Geoscience and Remote Sensing, 60, 1-10.

The main principle of TICOI relies on the temporal closure of the displacement measurement network. Measured displacements with different temporal baselines are expressed as linear combinations of estimated displacement (see the Figure below). The aim is to take advantage of different types of information (displacement measured using different temporal baselines, on images from different types of satellite) to extract glacier velocity time series, with a given temporal sampling. This enable the harmonization of various datasets, and the creation of standardized sub-annual velocity products.

INSTALLATION

With mamba

Clone the git repo and create a mamba environment (see how to install mamba in the mamba documentation):

mamba env create -f environment.yml  # Add '-n custom_name' if you want.
mamba install -c conda-forge ticoi

With pip

pip install ticoi

TUTORIALS

Basic examples

- notebook

• How to process one pixel of a NetCDF file
• How to process one pixel of ITS_LIVE data, stored on a cloud

- python_script

• How to process one cube
• How to process one pixel

Advanced examples

• How to process one ITS_LIVE cube directly from the cloud
• How to format several geotiff files into a netCDF file
• How to apply GLAFT on TICOI results
• How to compute statistics in static areas

TO USE YOUR OWN DATASET

You have geotiff files

You need to convert them into netcdf, by modifying this script.

You have netcdf files

If it is ITS_LIVE data, or Millan et al., 2022, you can directly use them! If not, you have to create your own dataloader, within which the dimension should be ("mid_date", "y", "x"), and the variables should be "vx", "vy", and should contain the projection information in the ds.proj4 attribute. You can add in this file.

HYPERPARAMETERS AND OUTPUTS

• to understand the output of pixel_demo please check README_output
• to understand the parameters you can change, please check README_possible_parameters

TO CONTRIBUTE

We welcome any contribution to this package! See guidelines here.