NASA Logo - GSFC + NASA Homepage
+ Goddard Homepage
NMQ: National Mosaic and Multi-Sensor QPE

Comparisons of rainrates from Q3 (MRMS) and GMI, DPR, CMB, KU

GPM surface rainrates

    Following three GPM surface rain products are compared with MRMS (now called "Q3").
  • GMI
  • GMI surface rainrates are generated on a pixel by pixel basis from radiometer brightness temperature data using the Goddard Profiling algorithm GPROF2014.

  • DPR
  • The DPR Level-2A product, 2ADPR, "DPR precipitation", is in a 3-swath structure. The swaths are normal scans (NS), matched scans (MS), and high sensitivity scans (HS).

  • CMB
  • The Combined Level-2 product, 2BCMB, "Level-2 DPR and GMI Combined", is in a two-swath structure. The first swath, NS, contains 49 rays that match Ku DPR. The second swath, MS, contains 25 rays that match Ka DPR.

  • KU
  • The Ku Level-2A product, 2AKu, "Ku precipitation", is written as a 1-swath structure (NS).

    Comparisons with Q3 (footprint-matched)

    The comparisons are conducted over the CONUS during GPM overpasses from Mar 2014 to most current dates dependent on the data availability.
        Vers 3,   Mar 2014 --> Feb 2016  (combined various V-3 versions)
        Ver 04A,  Mar 2014 --> Jul 2016
        ITE109,   Jun 2014 --> May 2016   KU, DPR only
        ITE110,   Mar 2014 --> Jul 2016   GMI only
        ITE111,   Mar 2014 --> Jul 2016   GMI only
        ITE113,   Jun 2014 --> Jul 2016   KU, DPR only
        ITE114,   Jun 2014 --> Jul 2016
        ITE117,   Mar 2014 --> Jul 2016   GMI only
        Ver 05A,  Mar 2014 --> Jun 2017
    The Q3 instantaneous rainrates are matched with GMI, DPR, CMB, KU rain products and resampled at the footprint resolution, then averaged the matched rainrates at 50km x 50km grids. Rainrates less than 0.2 mm/h from all products are not included in the analysis as the GMI gets a lot of false light rain. Different rainrate thresholds are used in the following comparison plots because many researches show the satellite rain bias is a function of rain intensities.

    For each matched pair of DPR, GMI or CMB with Q3 rainrates, following QCs/filters are applied:
    • QC0
      • Rainrates >=0.2 mm/h
    • QC1
      • Rainrates >=0.2 mm/h
      • RQI>=0.9
      • liquid precipitation only
    • QC2
      • Rainrates >=0.0 mm/h
      • set all values < 0.01 to “0”
    • QC3
      • Rainrates >=0.0 mm/h
      • set all values < 0.01 to “0”
      • RQI>=0.9
      • liquid precipitation only
    Note: QC2 and QC3 are only applied to V04A, ITE111 and later versions. For V04A, GMI rain is set to 0 when GMI POP <40%.

  • Scatter plot for different rainrate and rain type categories (Yearly, Mission To Date)

    PDF by rain occurrence and volume for different rainrate and rain type categories (Yearly, Mission To Date)

  • Density plot for different rainrate and rain type categories (Yearly, Mission To Date)

  • Error Map (Mission To Date)
    The mean rain maps for DPR, GMI, CMB or KU rainrates, error maps (mean relative error, mean absolute error, normalized root mean squre error for DPR, GMI, CMB or KU rainrates relative to Q3 rainrates), as well as the sample size, are based on the matched data at 50km resolution after applying the QC1 or QC0filter.
  • Rain map
    The rain map for GMI surface precipitation (SP), GMI most likely precipitation (MP), DPR near surface rainrate , CMB surface rainrate, KU near surface rainrate, or Q3 is based on its native spatial resolution (GMI/DPR/KU: ~5km; Q3: 0.01o).

FirstGov logo Privacy Policy and Important Notices NASA logo Web Curator: David A. Marks    NASA Official: David B. Wolff