Fig. 5: GSPIC-RT precipitation correction and gap-filling model schematic. | Communications Earth & Environment

Fig. 5: GSPIC-RT precipitation correction and gap-filling model schematic.

From: Regional-scale intelligent optimization and topography impact in restoring global precipitation data gaps

Fig. 5

a Workflow of the GSPIC-RT precipitation correction and gap-filling model. b Schematic of terrain factor selection, including slope and aspect calculation formulas. Input dimensionality was limited to 5 or fewer to mitigate potential issues such as feature extraction challenges, increased training difficulty, overfitting, and reduced computational efficiency. Prior to model training, Z-Score normalization was applied to standardize raw data and eliminate scale differences. c Evaluation metrics for the inversion accuracy of precipitation products, with classification metrics (POD, FAR, CSI) used to assess the model’s predictive capability for precipitation events.

Back to article page