MSWEP Retrospective
Version: 2.2

    Multi-Source Weighted-Ensemble Precipitation (MSWEP) is a unique gridded precipitation dataset spanning 1979–2017. This product combines a variety of different data sources to create the most accurate gridded precipitation product available.

    Main Features/Advantages:
  • Full global coverage (all land and oceans)
  • High spatial (0.1°) and temporal (3 hourly) resolution
  • Merging of precipitation estimates based on rain gauges (WorldClim, GHCN-D, GSOD, GPCC, and others), satellites (CMORPH, GridSat, GSMaP, and TMPA 3B42RT), and models (ERA-Interim and JRA-55)
  • Correction of systematic terrestrial precipitation biases using river discharge observations from ~14,000 stations across the globe
  • Incorporation of daily observations from ~75,000 rain gauges worldwide
  • Better overall performance than any other precipitation dataset


    MSWEP opti­mal­ly merges a wide range of gauge, satel­lite, and reanaly­sis data to pro­vide reli­able pre­cip­i­ta­tion esti­mates over the entire globe. See the fol­low­ing paper for a detailed descrip­tion of the MSWEP V2 method­ol­o­gy:

    The fol­low­ing paper describes the MSWEP V1 method­ol­o­gy:

    The most impor­tant changes in V2 com­pared to V1 include: (i) the intro­duc­tion of cumu­la­tive dis­tri­b­u­tion func­tion and pre­cip­i­ta­tion fre­quen­cy cor­rec­tions, to account for spu­ri­ous driz­zle and atten­u­at­ed peaks evi­dent in V1; (ii) increas­ing spa­tial res­o­lu­tion from 0.25° to 0.1° to increase the local rel­e­vance of the pre­cip­i­ta­tion esti­mates (espe­cial­ly impor­tant for high water-yield moun­tain­ous regions); (iii) the inclu­sion of ocean areas to enable ocean­ic stud­ies and ter­res­tri­al hydrol­o­gy stud­ies for coastal areas and small islands; (iv) the addi­tion of pre­cip­i­ta­tion esti­mates derived from Grid­ded Satel­lite (Grid­Sat) ther­mal infrared imagery for the pre-TRMM era to sup­ple­ment the reanaly­sis and gauge data; (v) the use of a dai­ly (rather than month­ly) gauge cor­rec­tion scheme that accounts for region­al dif­fer­ences in report­ing times, to min­i­mize tim­ing mis­match­es when apply­ing the dai­ly gauge cor­rec­tions; (vi) the use of a large data­base of dai­ly gauge obser­va­tions com­piled from sev­er­al sources to replace the 0.5° CPC Uni­fied dataset; and (vii) exten­sion of the data record to 2017 (MSWEP V1 fin­ished in 2016).

For each catchment, the precipitation dataset that provided the best streamflow simulations. See Beck et al. (2017b) for details.