GPC/m: Global Precipitation Climatology by Machine Learning; Quasi-global, Daily, and One Degree Spatial Resolution
Description
A precipitation dataset, Global Precipitation Climatology by Machine Learning (ML), GPC/m, is released.
This new precipitation dataset has been produced by machine learning, which is daily from 1979 to 2020 (will be to present), 1° × 1° spatial resolution. Three ML methods are used. Data is produced from outgoing longwave radiation (OLR) and atmospheric circulation from reanalysis. You can download this with DOI.
This daily precipitation dataset has been produced by machine learning (ML) methods using satellite observations and atmospheric circulations from reanalysis. The quasi-global daily precipitation dataset has been around for 42 years from 1979 to 2020, which will be updated to the present. The spatial resolution is 1° × 1° zonally global and from 40°S to 50°N. The ML methods are supervised learning, and the reference data are estimated precipitation datasets from 2001 to the present. The input data are somewhat modified based on knowledge of the climatological background. Using the trained statistical models, we predict back to 1979, when daily precipitation data was almost unavailable globally. For now, this GPC/m precipitation dataset version is GPC/m-v1-2024. This data will be updated in the future with added value. The purpose of this dataset is a challenge to produce a climatological dataset by reducing artificial gaps as much as possible for discussion of climatology, climate variability, and climate change. This dataset is very useful for statistical analysis, such as composite analysis and correlation analysis. Disadvantages should also be understood in the description paper (Takahashi, 2024c). Also, I hope that this dataset can contribute to improving the current precipitation datasets, which are based on physical or researcher-explaining algorithms.
To facilitate analysis of the dataset, it is distributed in Network Common Data Form (netCDF) format and the Grid Analysis and Display System (GrADS) format (with control file). If you would like recently updated data, please contact the creator. If it has already been created, it can be distributed.
Added on September 18, 2024.
More details are in the preprint paper at this link (Takahashi, 2024, https://doi.org/10.48550/arXiv.2409.09639).
Files
Files
(4.0 GB)
Name | Size | Download all |
---|---|---|
md5:ebb00ba5080b159775aa4bea6e018bd3
|
2.0 GB | Download |
md5:c4386a4d5e718d3135792714efc9984f
|
279 Bytes | Download |
md5:62cef408d5f8f4810f12b96c96850c8b
|
2.0 GB | Download |
Additional details
Funding
- Japan Society for the Promotion of Science KAKENHI grant 21K18403
- Japan Society for the Promotion of Science
- Japan Society for the Promotion of Science KAKENHI grant 22H00037
- Japan Society for the Promotion of Science
- Japan Society for the Promotion of Science KAKENHI grant 24K21389
- Japan Society for the Promotion of Science
- 3rd Earth Observation Research Announcement (EORA-3) ER3GPF012
- Japan Aerospace Exploration Agency
References
- 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