Global 0.25-degree gridded Snow water equivalent data derived from machine learning using in-situ measurements.

Area of Science:

• Earth and Environmental Sciences
• Climate Science
• Hydrology

Background:

• Accurate snow water equivalent (SWE) data is crucial for water resource management and climate studies.
• Existing SWE datasets have limitations in spatial coverage, temporal resolution, or accuracy.

Purpose of the Study:

• To develop a high-resolution, machine learning-based global daily SWE product (SWEML) spanning 1980-2020.
• To evaluate the performance of SWEML against diverse reference datasets.

Main Methods:

• Utilized k-means clustering to group in-situ SWE measurements into 14 clusters.
• Trained a random forest model using meteorological forcing and terrain attributes on 11,687 grid points.
• Compared SWEML with ten diverse reference datasets, including reanalysis, snow models, and remote sensing products.

Main Results:

• Achieved an overall root mean square error (RMSE) of 10.33 mm and a bias of -7.13 mm.
• Demonstrated high accuracy in high-elevation regions like the Rocky Mountains (RMSE = 7.30 mm, R=0.98).
• Showed strong agreement with airborne SWE data over North America and similar patterns to the Andes Snow Reanalysis.

Conclusions:

• The developed SWEML product offers a robust and accurate daily global SWE dataset.
• SWEML performs well in both data-rich and data-scarce regions, highlighting its broad applicability.
• This dataset can significantly enhance hydrological modeling, climate research, and water resource management.