The GPCP global monthly precipitation product merges rainfall estimates from several satellite-borne sensors and the gridded rain gauge-based product by the Global Precipitation Climatology Centre (GPCC). A respective list of input data is given in Table 1, based on Tables 1 and 2 in the C-ATBD [D1]. The methods used for generating the dataset are described in the C-ATBD [D1] and Adler et al. (2003). Recent changes related to the upgrade to latest version (v2.3) are discussed by Adler et al. (2018).

Table 1: Input data for GPCP monthly, see Tables 1 and 2 and algorithm description in [D1].

Table 2 provides an overview of achieved and targeted coverage and error-related properties of the TCDR-part of the dataset (until 12/2017). The target requirements (errors, stability) have been formulated in the context of monitoring the quality of precipitation products within C3S via Key Performance Indicators (KPI). They refer to the comparison of the dataset to a specific reference dataset. Here, the achieved specifications are computed in comparison with TMPA 3B43 (TRMM, 2011) which is available between 50°S and 50°N.

Table 2: Characteristics of the GPCP monthly mean precipitation rate v2.3 (CDR – Jan 1979 to Dec 2017)

The absolute random error percentiles are computed based on the uncertainties that are provided with the dataset at all grid cells and at all available time slices. The respective percentiles of the relative error are relative to the retrieved grid cell-wise precipitation estimates. The systematic error is computed per month as absolute bias (i.e. difference between GPCP monthly and TMPA 3B43 mean values in the TMPA window). The stability is the absolute of the slope of the bias time series of GPCP with respect to TMPA. 

Table 3 provides an overview for the continuously updated ICDR-part of the dataset (from 01/2018; see the General Definitions section for details). Stability is not evaluated over such short periods. The targets for the systematic error are based on the respective performance of the TCDR – here the 2.5- and 97.5-percentiles; see the PQAR [D3] for details. Due to the decommissioning of TMPA products in 12/2019, the KPI accuracy is evaluated with respect to precipitation in ERA5 (C3S, 2017) for the data from 01/2020.

Table 3: Characteristics of the GPCP monthly mean precipitation rate v2.3 (ICDR – Jan 2018 onwards)

The data format is NetCDF, compatible with conventions CF 1.6 and ACDD 1.3. The filenames are gpcp_v02r03_monthly_dYYYYMM.nc where YYYY specifies the year and MM specifies the month to which the monthly mean provided in the file refers.

The data are provided as monthly means of precipitation since Jan 1979 until Dec 2017 for the CDR, and are subsequently updated every three months for the ICDR. The two-dimensional spatial grid is equidistant in geographical latitude and longitude (2.5 degree), i.e., centre points are at 88.75 degree South, 86.25 degree South, ..., 88.75 degree North in latitude (72 nodes) and at 1.25 degree East, 3.75 degree East, ..., 358.75 degree East in longitude (144 nodes).

A list of known issues is included in the respective C-ATBD [D1], section 6 “Assumptions and Limitations”. 

When exploiting GPCP data you are kindly requested to acknowledge this contribution accordingly and make reference to the Global Precipitation Climatology Project, e.g. by stating “The work performed was done (i.a.) by using data from the Global Precipitation Climatology Project”. It is highly recommended to clearly identify the product version and temporal resolution (daily, monthly) used.

Please also include the following dataset and literature citations:

The GPCP global daily precipitation product merges rainfall estimates from several satellite-borne sensors. A respective list of input data is given in Table 4, based on the Table 1 in the C-ATBD [D2]. The methods used for generating the dataset are described in the C-ATBD [D2] and Huffman et al. (2001).

Table 4: Input data for GPCP daily, see Table 1 and algorithm description in [D2].

Table 5 provides an overview of achieved and targeted coverage- and error-related properties of the TCDR dataset. The target requirements (errors, stability) have been formulated in the context of monitoring the quality of precipitation products within C3S via Key Performance Indicators. They refer to the comparison of the dataset to a specific reference dataset. Here, the achieved specifications are computed in comparison with TMPA 3B42 (Goddard Earth Sciences Data and Information Services Center, 2016) which is available between 50°S and 50°N.

Table 5: Characteristics of the GPCP daily mean precipitation rate v1.3 (CDR)

The systematic error is computed per day as the absolute bias (i.e. difference between GPCP daily and TMPA 3B42 mean values in the TMPA window). The stability is the absolute of the slope of the bias time series of GPCP with respect to TMPA.

Table 6 provides a respective overview for the continuously updated ICDR-part of the dataset (from 01/2018; see the General Definitions section for details). Stability is not evaluated over such short periods. The targets for the systematic error are based on the respective performance of the TCDR – here the 2.5- and 97.5-percentiles; see the PQAR [D3] for details. Due to the decommissioning of TMPA products in 12/2019, the KPI accuracy is evaluated with respect to precipitation in ERA5 (C3S, 2017) for the data from 01/2020. 

Table 6: Characteristics of the GPCP daily mean precipitation rate v2.3 (ICDR)

The data format is NetCDF, compatible with conventions CF 1.6 and ACDD 1.3. The filenames are gpcp_v01r03_daily_dYYYYMMDD.nc where YYYY specifies the year, MM specifies the month, and DD specifies the day of the month to which the daily mean provided in the file refers.

The data are provided as daily means of precipitation since 1 Oct 1996 until 31 Dec 2017 for the CDR, and are subsequently updated every three months for the ICDR. The two-dimensional spatial grid is equidistant in geographical latitude and longitude (1.0 degree), i.e., centre points are at 89.5 degree South, 88.5 degree South, ..., 89.5 degree North in latitude (180 nodes) and at 0.5 degree East, 1.5 degree East, ..., 359.5 degree East in longitude (360 nodes). Note, however, that unlike in the case of the monthly data, the latitude and longitude variables in the NetCDF files specify the southeast corner of each grid cell.

The NetCDF files feature a valid_range attribute for the precipitation data, stating that values between 0 and 100 mm/d are valid. However, data points > 100 mm/d exist and are valid. Users are advised to not apply the upper boundary of 100 mm/d when using the data.

When exploiting GPCP data you are kindly requested to acknowledge this contribution accordingly and make reference to the Global Precipitation Climatology Project, e.g. by stating “The work performed was done (i.a.) by using data from the Global Precipitation Climatology Project”. It is highly recommended to clearly identify the product version and temporal resolution (daily, monthly) used.

Please also include the following dataset and literature citations:

The original repository for the data at the UMD can be found at http://gpcp.umd.edu/. Respective documentation at a separate NOAA repository (see URLs below) includes the C-ATBDs [D1, D2], acknowledgment guidelines, maturity matrices, and source codes.

• https://www.ncei.noaa.gov/products/climate-data-records/precipitation-gpcp-daily
• https://www.ncei.noaa.gov/products/climate-data-records/precipitation-gpcp-monthly

Within C3S, the distribution will be through the CDS (https://cds.climate.copernicus.eu/) where the C-ATBDs [D1, D2] will be uploaded, too. Additional documentation created for the inclusion of the dataset in the CDS, such as this PUGS document, will also be provided.