This repository represents the "Investigating SWE Predictive Capability Using GOES Bands and Convolutional Neural Network" project. This project aims to leverage Geostationary Operational Environmental Satellite (GOES) data to make spatial predictions of snow water equivalent (SWE) in the southern Sierra Nevada mountains. By employing Convolutional Neural Networks (CNNs), we analyze visual and infrared satellite imagery to generate these predictions. This repository includes all necessary code for downloading and processing data, model development, and evaluation.
To use this repository, please clone it and install the required packages using the requirements.yml file with the code below:
git clone https://github.com/NWC-CUAHSI-Summer-Institute/SI24_GOES-SWE.git
cd SI24_GOES-SWE
mamba env create -f requirments.yaml
Then, you should install the ipykernel and connect it to jupyter notebook.
pip install --user ipykernel
python -m ipykernel install --user --name=goes_kernel
Southern Sierra Nevada snowpack is a major source of spring streamflow in California. Precise measurements of SWE are critical for making accurate streamflow predictions year to year, which assist water management agencies in their decision making. Ground observations, such as those provided by snow telemetry (SNOTEL) and California Date Exchange Center (CDEC) stations, are limited to a small number of point source locations. This creates a need for more representative spatial predictions of SWE in areas without in-situ measurements.
Previous studies have developed models using observational data and remote sensing imagery to predict SWE in various regions of the United States, however, data from the GOES series satellites has been underutilized in hydrologic research and remains a promising source of information. This project aims to assess the feasibility of assimilating GOES imagery (visual and infrared) across the contiguous United States (CONUS) as inputs to a CNN model for predicting SWE in high elevation mountainous regions.