A GANs-based Deep Learning Framework for Calculating Climatological Monthly Rainfall Erosivity

Description

This dataset provides the deep learning source code for calculating climatological monthly rainfall erosivity based on GANs, along with the optimal pre-trained rainfall erosivity model. Additionally, it includes the model-predicted rainfall erosivity data for mainland China at a 1km resolution under three climate scenarios (SSP1.RCP1.9, SSP2.RCP4.5, and SSP5.RCP8.5) across two future climate periods (2041-2070 and 2071-2100). The model code heavily references the pix2pixhd code (https://github.com/NVIDIA/pix2pixHD). The attention mechanism reference code can be found at EMA (https://github.com/YOLOonMe/EMA-attention-module). The final data is based on this version.

Categories

Funders

• National Natural Science Foundation of China

  Beijing, China

  Grant ID: 42377332

Licence