FFMP

This page is currently hosting FFMP information relevant to some of the recent AWIPS build changes (e.g 16.2.1 MRMS/ARI and 17.1.1 HPN), and it will be updated to be a more general FFMP reference in the future.

New:

 

In 16.2.1, MRMS was added as a 5th precip source to the SCAN FFMP menu:

Note that the ARIs and MRMS implementation in 16.2.1 will not be visible unless the ITO works with the hydro focal point to follow the configuration instructions provided with the 16.2.1 Installation Instructions. See WDTD's ARIs and MRMS FFMP page for configuration support.

FFMP Precip Sources:

  1. DHR (Digital Hybrid Reflectivity) - "Legacy" based single radar
  2. DPR (Digital Precipitation Rate) - Dual Pol single radar
  3. HPE (High-Resolution Precipitation Estimator) - Mosaic that is DPR-based by default but can be configured to be DHR locally
    • DPR was set to be the default source for HPE in 14.3.1 for everyone. There is a new bug in 16.2.1 where the HPE source label in the upper-left part of the FFMP display incorrectly says it is not using Dual Pol when it really is (will be fixed in 16.4.1). The local site would need to change HPE tokens for it to use DHR instead of DPR, and that is not likely to happen at most sites.
  4. Bias HPE - Mosaic of HPE that has been bias corrected with mean field biases for each radar (RFC provides biases by default) 
    • The same Dual-Pol source label bug applies to Bias HPE as HPE (see HPE above)
  5. MRMS (Multi-Radar Multi-Sensor) - Mosaic that is MRMS Surface Precipitation Rate (SPR) based

 

Here is the table from the 16.2.1 Informational Overview that summarizes the characteristics of each precip source:

Table Notes:

  • MRMS *
    • MRMS only uses Dual Pol for reflectivity QC and does not use Dual Pol explicityly in the Z-Rs
    • MRMS has bias products, but they are not timely for flash flood decision making and are not applied to SPR and other radar-only accumulations.
  • HPE/Bias HPE **
    • DPR is the default precip source used in the HPE and Bias HPE mosaic, but that can be changed to DHR locally.
  • DPR FFMP bugs
    • DPR in FFMP (only) contained bugs in calculating basin averages that were fixed in 16.2.1 patch 1 that was released after 16.2.1.
  • Bias HPE bugs
    • The default bias factors used in Bias HPE are from the RFCs (can be configured locally to be different), and these have had some bugs in them. WDTD recommends only using Bias HPE on a case-by-case basis only if the biases look reasonable with local gauge comparison.

FFMP Usage Tips

  • The reflectivity-based MRMS mosaic and the Dual-Pol-based HPE mosaic are probably the best sources to start with on a given day due to the spatially varying Z-Rs and the convenience of the mosaic source.
  • Only use Bias HPE if the biases being use appear to be consistent with gauges and reports. To see the bias factors for each radar used in Bias HPE, see the upper-left corner of the FFMP Bias HPE display (load Bias HPE from SCAN FFMP menu) or the Bias HPE from the Volume Browser.
  • Routinely compare QPE sources to surface observations and reports duraing an event to identify the best precip source to use in FFMP on a given day.
  • WDTD advises using QPF in FFMP only for more advanced users. The split window (breaks up configured duration into 1hr QPF and remaining hr QPE) is more complicated to use. There are also bugs in FFMP where ratio and diff do not use the QPF when it is configured (attributes button in FFMP table).
    • HPE Nowcast is the best 1hr QPF source to consider using in steady state situations. This uses a sophisticated radar extrapolation algorithm that can be decent at times of more steady state convection. Bias HPE Nowcast could also be used if the biases look reasonable, though there are bugs with the RFC bias at times.
    • Do not use the MRMS-radarOnlyQPE01H QPF source in FFMP as it is not QPF.
    • HRRR QPF can be off significantly at times due to the spinup in the model, so only use if it looks reasonable.
  • Only use ARIs in FFMP after taking the ARIs training. See the ARIs page
  • When RFC FFG is updated during a precipitation event, it is becoming more common to see isolated spots of extremely low FFG in some rainfall areas (e.g. 0.15" dor 6hr duration) that causes extreme values of ratio and difference in FFMP.  Watch out for these updates (every 6-12 hrs) as FFMP can change immediately once they are ingested, and sometimes these areas are not in areas of flash flooding.
    • Be careful when using FFG in FFMP with long durations that cross times when FFG has been updated during an event. FFMP in AWIPS-1 used to account for FFG updates by delaying FFG to prevent using the reduced FFG with the same QPE that was used to reduce the FFG ("double counting" in some respects). AWIPS-2 has not implemented this, so if you configure a long duration that overlaps a time when you sample the QPE that was used in the FFG reduction, you will likely see extreme ratios and diffs that are not meaningful.

Questions?

  • For questions about FFMP contact Michael.A.Magsig@noaa.gov.