GRaD-GB (1H1K): A Gauge-Radar Precipitation Dataset for Great Britain (1 hour and 1 km), 2006-2023
===================================================================
If you have any questions, please send an email to Dr. Xiaobin Qiu (qxiaobin111@gmail.com and Xiaobin.Qiu@newcastle.ac.uk)
===================================================================
Overview
--------
GRaD-GB (1H1K) provides gridded, hourly precipitation estimates at 1 km
spatial resolution for Great Britain, covering 1 January 2006 to
31 December 2023. It combines quality-controlled radar rainfall estimates
with hourly rain-gauge observations using the Gauss Blending Method.
Dataset authors: Xiaobin Qiu, Amy Charlotte Green, Stephen Blenkinsop,
and Hayley J. Fowler (Newcastle University, UK).
The full methodological description, quality control, known limitations,
and guidance for interpretation are provided in
Sample/GRaD-GB-1H1K-Data_resource_supporting_doc.docx.

Files and folders
-----------------
GRaD_GB_1H1K/YYYY/ contains daily NetCDF files for year YYYY (2006-2023).
Each file is named ds_GRaD_GB_hourly_YYYY-MM-DD.nc, where the date identifies
the day covered by that file. The deposited daily collection contains 6,428
NetCDF files. Some dates have no file, and some files may have missing hourly
values. Do not assume a complete, uninterrupted hourly record.

Sample/ contains two example daily rainfall files, gauge and radar location
metadata in CSV format, and NetCDF masks for geographical features and
potential sources of radar artefacts. It also contains the detailed supporting
document named above and an explanatory presentation about the feature types
in the high-object mask.

Sample/Google_colab_link.txt gives the address of an optional Google Colab
notebook prepared by the dataset author. The notebook demonstrates how the
sample data look and can be explored. The online notebook is not needed to
understand or read the deposited files; the essential information is provided
here and in the accompanying supporting document.

The _logs folder and the separate GRaD_GB_1H1K_monthly folder are not part of
this deposit.

Reading the daily NetCDF files
------------------------------
The files can be opened with software that supports NetCDF, such as Python
with xarray. For example:

    import xarray as xr

    ds = xr.open_dataset("GRaD_GB_1H1K/2007/ds_GRaD_GB_hourly_2007-07-20.nc")
    print(ds)
    hourly_rainfall = ds["precip"]

The principal variables and coordinates are:

    precip          Hourly accumulated precipitation (time, y, x), in mm.
    classification  Rainfall estimation method flag (time, y, x).
    time            UTC timestamp at the end of each one-hour accumulation.
    time_bnds       Start and end of each hourly accumulation interval.
    x, y            Easting and northing in metres, respectively.
    lon, lat        Two-dimensional longitude and latitude coordinates.

The projected grid uses the British National Grid (EPSG:27700). Do not treat
x and y as longitude and latitude. In the example file, x increases from
west to east and y decreases from north to south. Check coordinate order
when combining these data with other grids or masks.

The classification values are 111 for Gauss blending of radar and gauge
rainfall, 110 for linear interpolation of gauge rainfall, and 100 for
nearest-neighbour interpolation of gauge rainfall. See the supporting
document for interpretation of these methods and the spatial coverage.

The precipitation unit stored in the NetCDF variable is mm: each value is
the depth accumulated over the preceding one-hour interval. For example,
the timestamp 10:00 UTC describes rainfall from 09:00 to 10:00 UTC.
When comparing with the source gauge observations, account for the known
one-hour radar-gauge timestamp offset described in the supporting document.

Reuse notes
-----------
Missing days or hours and changes in radar coverage can affect analyses.
Rainfall outside Great Britain may represent radar-only estimates. Residual
radar- and gauge-related artefacts may remain, especially when interpreting
localised extremes. Consult the supporting document before scientific reuse.

Getting Started with the GRaD-GB Dataset: Google Colab Tutorial
---------------------------------------------------------------
To help you understand how to read and visualise the dataset, I have created
a simple Python notebook in Google Colab.
You can access the notebook here:
https://colab.research.google.com/drive/16krPrDksAYI7U9o_bxZNVWma3y02EUzY?usp=sharing

Before running the code, please upload all the files from the GRaD_GB_sample folder
to the Google Colab session. The notebook provides a simple example of how to load, 
explore, and visualise the rainfall data.


How to cite this dataset and its methods
----------------------------------------
To help others find and credit this dataset, cite the published EIDC dataset
record and its DOI when these become available. Use the citation supplied by
the EIDC record; this README does not specify a dataset DOI.

Please also cite both papers below when using GRaD-GB (1H1K), especially when
describing the radar quality control and the radar-gauge blending method:

Qiu, X., Green, A.C., Blenkinsop, S., and Fowler, H.J. (2026).
Quality control of the composite radar quantitative precipitation estimation
product for Great Britain. Journal of Hydrology, 666, 134755.
DOI: 10.1016/j.jhydrol.2025.134755

Qiu, X., Green, A.C., Blenkinsop, S., and Fowler, H.J. (2026).
Developing a radar-rain gauge hourly blended precipitation dataset for Great
Britain using the Gauss Blending Method. Journal of Hydrology, 668, 134954.
DOI: 10.1016/j.jhydrol.2026.134954
