A 30-Year, 500-m Resolution Daily Snow Water Equivalent Dataset for Croplands in China (1995–2024)

Overview

This dataset provides a high-resolution (500 m), long-term (1995–2024) daily Snow Water Equivalent (SWE) product specifically targeting cropland areas in China. Accurately mapping SWE in plain and agricultural regions has long been a challenge due to the frequent underestimation of shallow snow by traditional passive microwave remote sensing. This dataset is designed to fill this critical gap, providing robust data support for agricultural disaster assessment, particularly for investigating the exacerbation of winter wheat freezing injuries under climate change.

Methodology

The dataset was generated by integrating multi-source remote sensing data and meteorological observations. A Random Forest (RF) machine learning algorithm was coupled with a dynamic Sturm snow density model to achieve spatial downscaling and accuracy enhancement. This approach effectively mitigates the "shallow snow underestimation" problem prevalent in plain croplands, significantly improving the spatiotemporal continuous monitoring capability of SWE.

Data Format and Usage Notes

 To optimize storage and access efficiency while maintaining data fidelity, the daily spatial data are packaged into annual NetCDF (.nc) files using advanced lossless compression strategies (Zlib complevel=7 with Shuffle filter).

Spatial Resolution: 500 m

Temporal Resolution: Daily

Time Span: January 1, 1995 – August 31, 2024. (Note on 2024 Data: The dataset concludes on August 31, 2024. The data for this final year is provided in the file named Farmland_SWE_2024_JanAug.nc.)

Variables: SWE (Snow Water Equivalent, unit: mm)

Data Type & Scaling: The data are stored as Int16. A scale_factor of 0.1 is applied. (e.g., A stored pixel value of 125 represents an actual SWE of 12.5 mm). Most modern GIS software and Python libraries (like xarray) will apply this scale factor automatically upon reading.

NoData Value / FillValue: -9999

Application Potential & Utility Demonstrated

This open-access dataset contributes to a deeper understanding of snow hydrology in agricultural ecosystems. Notably, it effectively rectifies the systematic underestimation of shallow snowpacks over plain agricultural regions (e.g., the Huanghuaihai Plain) common in existing reanalysis data. By successfully capturing detailed spatial patterns—such as the severe fragmentation of continuous snow cover in primary winter wheat-producing regions—this dataset demonstrates its robust utility for investigating climate-induced agricultural risks (e.g., overwintering freezing injury). It serves as a highly reliable benchmark for regional climate modeling, spring drought monitoring, and crop risk assessments (SDG 2).