Global SPEI database


The Global SPEI database, SPEIbase, offers long-time, robust information about drought conditions at the global scale, with a 0.5 degrees spatial resolution and a monthly time resolution. It has a multi-scale character, providing SPEI time-scales between 1 and 48 months. Currently it covers the period between January 1901 and December 2014.

The SPEIbase can be accessed from digitalCSIC.

The SPEIbase is based on monthly precipitation and potential evapotranspiration from the Climatic Research Unit of the University of East Anglia. Currently the version 3.23 of the CRU TS dataset has been used. The SPEIbase is usually updated as soon as new data becomes available.

The SPEIbase consists of standardized values over the emerged land pixels. No land pixels are assigned a value of 1.0x10^30. In some rare cases it was not possible to achieve a good fit to the log-logistic distribution, resulting in a NAN (not a number) value in the database.

The SPEIbase is based on the FAO-56 Penman-Monteith estimation of potential evapotranspiration. This is a major difference with respect to the SPEI Global Drought Monitor, that uses the Thornthwaite PET estimation. The Penman-Montheith method is considered a superior method, so the SPEIbase is recommended for most uses including long-term climatological analysis.

Version history

  • SPEIbase v2.4: 1) Based on the CRU TS 3.23 dataset, extending the temporal range of the SPEIbase up to December 2014.
  • SPEIbase v2.3: 1) Based on the CRU TS 3.22 dataset, extending the temporal range of the SPEIbase up to December 2013.
  • SPEIbase v2.2: 1) The CRU TS 3.2 dataset has been used, extending the data range of the SPEIbase up to December 2011. 2) Potential evapotranspiration data from the CRU TS 3.2 dataset has been used instead of computing our own.
  • SPEIbase v2.1: 1) The revised dataset CRU TS 3.10.01 for precipitation is used. 2) Data on surface pressure and wind from 20th Century Reanalysis v.2 until December 2009 has been used for computing Penman PET. 3) An error while reading some data sources that yielded no data at longitudes 179.25 and 179.75 has been corrected.
  • SPEIbase v2.0: 1) Data has been extended to the period 1901-2010. 2) The FAO-56 Penman-Monteith's method has been used for computing PET instead of Thornthwaite in SPEIbase v1.0. 3) Unbiased probability weighted moments (ub-pwm) method has been used for fitting the log-Logistic distribution, instead of the sub-optimal plotting-position pwm method used in version 1.0. 4) The whole World is put in one single netCDF file.
  • SPEIbase v1.0: The global gridded SPEI dataset is available at time scales between 1 and 48 months, with spatial resolution of 0.5º lat/lon and temporal coverage between January 1901 and December 2006.

(Use of the newest version is recommended. Older versions are still available to allow replicability.)

Single location

In addition to accessing the complete dataset, time series of the SPEIbase at specific locations can be downloaded as a .csv (comma-separated values) files containing the time scales between 1 and 48 months. Instructions:

  • Enter the coordinates of the desired location in -180 to +180 degrees (longitude) and +90 to -90 degrees (latitude) format and clic on 'download' to get the data from the closest grid cell.
  • Or clic on 'map' to interactively choose the location on a map and press the button to have the coordinates pasted automatically to the form, then clic 'download'.


SPEIbase details

To calculate the SPEI we used the CRU TS 3.23 dataset. This is the most complete and updated dataset of gridded precipitation and potential evaporation at the global scale, has a spatial resolution of 0.5°, and covers the period 1901−2014.

Details of the SPEIbase and the comparison with other available drought data sets can be consulted in:

A number of users of our SPEIbase dataset worried because they could not read our netCDF files in their software. Indeed, some commercial softwares such as Matlab, ArcGIS and Idrisi do not have support for NetCDF4 or did not have it until recently. The main reason for us to use netCDF4 for the SPEIbase is because it allows for data compression, resulting in large file size and internet traffic saving. We did not use any of the other fancy options implemented in netCDF4 such as grouping, compound types or multiple unlimited dimensions. This means that our netCDF4 files conform to the classic model, so they can be converted back to the netCDF3 legacy format.

One way of converting to netCDF3 is by using the nccopy program by unidata. For example, to convert a netCDF-4 classic-model format file to a classic format file

nccopy -k classic

Another way would be to install the netCDF operators (NCO) toolset from unidata, and then use the ncks utility:
ncks -3

In fact, ncks allows for much greater functionality. For example, if one wants to extract the first 100 times from an SPEIbase file:

ncks -d time,0,100

would generate a (smaller) netCDF file with only those timesteps. In a similar fashion, it is possible to use ncks to select a specific geographical region.


The Global 0.5° gridded SPEI dataset is made available under the Open Database License. Any rights in individual contents of the database are licensed under the Database Contents License.

What follows is a human-readable summary of the ODbL 1.0 license. Please, read the full ODbL 1.0 license text for the exact terms that apply.

Users of the dataset are free to:
  • Share: copy, distribute and use the database, either commercially or non-commercially.
  • Create: produce derivative works from the database.
  • Adapt: modify, transform and build upon the database.
Under the following conditions:
  • Attribution: You must attribute any public use of the database, or works produced from the database, by citing one or more of the papers referenced in the References section below. For any use or redistribution of the database, or works produced from it, you must make clear to others the license of the original database.
  • Share-Alike: If you publicly use any adapted version of this database, or works produced from an adapted database, you must also offer that adapted database under the ODbL.



  • longitude: geographical coordinates;
  • latitude: geographical coordinates;
  • time: months.


  • long_name: "Standard Precipitation-Evapotranspiration Index";
  • units: "z values";
  • type: float;
  • range: -3.0E32 to 3.0E32;
  • missing value: 3.0E33f;
  • not a number: nan;
  • long_name: "Longitude values";
  • cartesian_axis: "X";
  • units: "degrees_E";
  • type: float;
  • interval: 0.5;
  • range : -179.75, 179.75;
  • long_name: "Latitude values";
  • cartesian_axis: "Y";
  • units: "degrees_N";
  • type: float;
  • interval: 0.5;
  • range : -89.75, 89.75;
  • long_name: "Time coordinate values";
  • cartesian_axis: "T";
  • units: "Months since 1901-01";
  • calendar: "Gregorian";
  • type: double integer;
  • interval: 1;
  • range: January 1901, December 2011;

Format: netcdf

The SPEIdata is provided as single files comprising the whole World for each SPEI time scale. The data format is netCDF.


All relevant metadata is self-contained in the .nc files, and can be read and displayed by any software allowing netCDF manipulation.

Suggested tools

Manipulation of netCDF data is usually done at the programming level, and there are several libraries with functions for reading and manipulating netCDF data for a diversity of languages. For more information on the netCDF format and suggested tools, see

Panoply is a cross-platform netCDF viewer developed in Java ( There are other netCDF viewers, such as ncBrowse, ncview, and nCDF_Browser.

netCDF data can be accessed and manipulated in the (open source) statistical analysis system R using the package netcdf.

Some commercial GIS packages allow reading netCDF data, such as ArcGIS (from version 9.2 on) and IDRISI Taiga.