MPS-based spatial precipitation downscaling

A 2-dimensional precipitation downscaling method is suggested by using the multiple-point geostatistics (MPS) model (Gravey and Mariethoz, 2020) with a precipitation coefficient of variation adjustment. An example of an application in Beijing (China) is provided in the code. The MPS-based downscaling approach can be easily modified and applied to other rainfall products and regions.

An example of downscaling the hourly rainfall field recorded by CMPAS (a) at 10 am, on the 20th of July 2016 over Beijing. The rainfall field was upscaled to 8 km (b), downscaled by the MPS-QS (c), and adjusted for the rainfall CV (d). (e) and (f) present the absolute difference between the original and downscaled images (i.e., |d-a|) and their variograms (respectively).


Zou, W., Hu, G., Wiersma, P., Yin, S., Xiao, Y., Mariethoz, G., Peleg, N., Multiple-point geostatistics-based spatial downscaling of heavy rainfall fields. Journal of Hydrology, 632, 130899.


Zou, W., & Peleg, N. (2024). Mutiple-point geostatistics (MPS) based spatial downscaling of heavy rainfall fields —– An example of Beijing city. Zenodo.