MCIP/WRF Precipitation Definitions



I have run MCIPv4.4 to prepare meteorology for CMAQ from WRFv3.6.1.

The non-cumulus rain output from MCIP in the METCRO2D file (RN) appears to be quite different than the non-cumulus rain output from the wrfout file (RAINNC). In METCRO2D, RN ~0.03 cm, and only in a small localized area. In wrfout, RAINNC ~30 mm (0.3 cm) across a much greater area. Additionally, METCRO2D RN varies hourly, while wrfout RAINNC is constant each hour.

What is MCIP doing to the WRF precipitation output? Is there an issue with my MCIP results, or might I be interpreting them incorrectly?



The WRF precipitation variables represent accumulated precip since the start of a run. So, if it is constant from time t to time t+1, that indicates no rain fell during that interval.

MCIP computes the difference for each time step, so that RN and RC are the amount of precip over the output interval.


Thank you, that makes sense.

Would it be correct to say:
RAINNC(t+1) - RAINNC(t) {WRF} = RN(t+1) {CMAQ}


That’s more or less correct. WRF’s RAINNC has units of mm, while CMAQ/MCIP’s RN has units of cm, so there is a factor of 10 difference there. Also, for long WRF runs you can use a “tipping bucket” (typically 1000 mm, but also configurable) for maintaining precision, and MCIP needs to take that into account also.

This is all in rdwrfem.f90 in MCIP.


To add to what Chris mentioned here…

There is an error in MCIPv4.4 in the tipping bucket calculation. If you are using the tipping bucket in WRF (for long runs, for example), please write back and I’ll provide the fix. That option is off by default in WRF so you would have had to deliberately enable it.