HYSPLIT Forum: hysplitbbs.arl.noaa.gov

Hello,

I'm a post doc working on dispersion modelling projects with HYSPLIT. I'm still relatively new to the field and working on legacy code that handles data retrieval and modelling. I use data from the ECMWF-IFS as my HYSPLIT input which is converted with grib2arl. As part of this legacy pipeline, the precipitation of the initial GRIB1 file is re-binned as the format of ECMWF precipitation as the "Total Precipitation" needs to converted to Precipitation per time bin before being converted to ARL. This is an update coming that will mean the ECMWF-IFS products will be GRIB2 rather than GRIB1 so I have a chance to review some of the legacy code.

My question is, is the precipitation re-binning still necessary when converting GRIB2 to ARL via api2arl or is this handled by the conversion tool? I apologise if this is clarified in literature but I have been unable to find a definitive answer.

The re-binning may be necessary if the ECMWF-IFS precip is summed from the simulation's initial time and the converter is configured to process, for example, 3 hourly accumulated precip (TPP3). Please check what precip is available in the ECMWF-IFS file and api2arl.cfg used for the converter.

For reference, HYSPLIT currently supports the following precip variables:
TPPD - accumulation daily over 24 hours
TPPT - accumulation over 12 hours
TPP6 - accumulation over 6 hours
TPP3 - accumulation over 3 hours
TPP1 - accumulation over 1 hours
Or
PRT3 or PRT6 - precipitation rate over 3 hours and 6 hours (m/min)

Ahh I see, that explains it. Indeed to my knowledge the ECMWF-IFS doesn't have any precip variables that match the representation expected by HYSPLIT so this re-binning step makes sense. Thank you very much for the explanation.