GPC/m: Global Precipitation Climatology by Machine Learning; Quasi-global, Daily, and One Degree Spatial Resolution

GPC/m (Global Precipitation Climatology by Machine Learning) is a quasi-global
daily precipitation dataset at 1° × 1° spatial resolution. It covers the
zonally global domain from 40.5°S to 49.5°N for 1979-2020 and uses a 365-day
no-leap calendar.

The dataset was produced using supervised machine-learning methods trained with
satellite-based precipitation estimates and atmospheric predictors, including
outgoing longwave radiation and circulation fields from reanalysis. It was
developed to provide a long, spatially consistent daily precipitation record
for climatological analyses, including composite analysis, correlation
analysis, climate variability, and climate-change studies. The strengths and
limitations of the dataset are described in the accompanying paper.

Version v1.1-2024 is a distribution and accessibility update to the GPC/m
v1-2024 dataset. The scientific precipitation-field values and existing data
file names are unchanged. The data are distributed as 42 independent yearly
NetCDF files so that users can download only the years required. The version
number v1.1-2024 refers to the Zenodo distribution release.

The main variable is `pr` (mm day-1), and the coordinate variables are `lat`
and `lon`. The files use `calendar = "365_day"`; February 29 is not included.

This release also provides GrADS descriptor files, SHA-256 checksums, expanded
documentation, and a Bash download script. The yearly files can be downloaded
either from Zenodo or from the direct Tokyo Metropolitan University mirror:

https://camo3climate.github.io/gpcm-precipitation-data/

https://camo.fpark.tmu.ac.jp/gpcm.html (v1-2024; old version, tentative)

Users should cite the Zenodo dataset DOI and the description paper even when
the files are obtained from the external mirror.

Description paper:

Takahashi, H. G. (2024), GPC/m: Global Precipitation Climatology by Machine
Learning; Quasi-global, Daily, and One Degree Spatial Resolution.
https://doi.org/10.48550/arXiv.2409.09639