Snow Depth Prediction using L-Band InSAR Products

Overview

This work serves as a proof-of-concept that InSAR products can be used to estimate total snow depth through Machine Learning. This repository contains the code used to generate all results in our paper "Advancing Terrestrial Snow Depth Monitoring with Machine Learning and L-band InSAR data: A Case Study Using SnowEx 2017 Data" (doi: https://doi.org/10.3389/frsen.2024.1481848).

Dataset

This work uses dataset from the 2017 NASA SnowEx Campaign. The InSAR data is from NASA JPL's UAVSAR sensor, and the Lidar Depth is from ASO.

• Link to the remote sensing data we used: link.
• Link to the in situ depth measurements: NSIDC.

Software and hardware list

| Software used | Link to the software | Hardware specifications | OS required | |:---: |:---: |:---: |:---: | | Python 3.11.5 | https://github.com/pyenv/pyenv | This code should work on any recent PC/Laptop | Linux (any), MacOS, Windows|

Usage

To use the codes in this repository, you need to have Python installed. There are several ways to install Pyton. I will take you through how to install Python using the pyenv utility program. pyenv allows you to install different versions of Python and to seamlessly switch between them.

1. Install pyenv

Please follow the instructions here to install pyenv for your operation system (OS).

2. Setup the environment

2.1 Mac and Linux

Open the terminal and type the following commands to setup your virtual environment.

bash pyenv install 3.11.5 mkdir insar_project cd insar_project pyenv local 3.11.5 python -m venv .venv source .venv/bin/activate pip install --upgrade pip

2.2 Windows

pyenv does not officially support Windows OS. Hence, you should install Python directly for your own version of Windows either by using the Microsoft Store or by navigating to https://www.python.org/downloads/windows/ in your browser and following the download link. During installation, be sure to check the box that adds Python to your PATH environment variables. To run Python in a virtual environment, open the PowerShell program and type the follow commands.

bash mkdir insar_project cd insar_project python -m venv .venv .venv/Scripts/activate pip install --upgrade pip

If these instructions doesn't work for you (Windows users), you may need change the execution policy of your computer. To do this, run the PowerShell program as an administrator and run Set ExecutionPolicy Unrestricted. Afterwards, exit and reopen the PowerShell program (not as an administrator) then type commands in 2.2.

Note: the Windows installation guide was not tested and the entire codebase was not tested on a Windows machine!

3. Clone the Repository

bash git clone git@github.com:cryogars/uavsar-lidar-ml-project.git cd uavsar-lidar-ml-project

4. Install Source Code

bash pip install .

5. Directory Setup

Create the data folder using the command below.

bash mkdir data

6. Download Data

1. Download the remote sensing dataset into the data folder using this link.

2. Download the in situ data following these steps:

* Create the folder to put the *in situ* measurement and change directory:
```bash
mkdir data/SNOWEX
cd data/SNOWEX
```

* Dowlond the *in situ* measurement from [NSIDC](https://nsidc.org/data/snex17_sd/versions/1). When you land at the page navigate to `Data Access Tool` -> `Get Data` and click `Download Script`. At the time of writing,  `Download Script` is on the botton right of the page and the downloaded script is named `nsidc-download_SNEX17_SD.001_2025-01-07.py`. Note that you will need  Earthdata username and password to download the *in situ* measurements. Click [here](https://urs.earthdata.nasa.gov/users/new) to register.

* Run the Python script to download the *in situ* data using:
```bash
python nsidc-download_SNEX17_SD.001_2025-01-07.py
```

* Move back to the parent directory and run the codes. Note that you will need `Jupyter Notebook` installed.
```bash
cd ../..
```

7. Deactivate Environment

After running the experiments, you can deactivate the virtual environment by running the command below.

bash deactivate

Acknowledgements

• We would like to express our gratitude to the NASA Terrestrial Hydrology Program and all participants of the SnowEx campaign for providing the Lidar and in situ data.

• We also thank Yunling Lou - the UAVSAR Project Manager at NASA JPL - for the UAVSAR data.

• This work was partially funded by the U.S. Army Cold Regions Research and Engineering Laboratory (CRREL) under contract W913E523C0002.

Citation

We would appreciate citations if you use our code in your work.

bash @ARTICLE{alabi5advancing, AUTHOR={Alabi, Ibrahim Olalekan and Marshall, Hans-Peter and Mead, Jodi and Trujillo, Ernesto }, TITLE={Advancing terrestrial snow depth monitoring with machine learning and L-band InSAR data: a case study using NASA’s SnowEx 2017 data}, JOURNAL={Frontiers in Remote Sensing}, VOLUME={5}, YEAR={2025}, URL={https://www.frontiersin.org/journals/remote-sensing/articles/10.3389/frsen.2024.1481848}, DOI={10.3389/frsen.2024.1481848}, ISSN={2673-6187} }

Questions?

Ibrahim Alabi (Email: ibrahimolalekana@u.boisestate.edu)