sebvi
commented
3 years ago Hello @richardweedon
I would use 0-17-4 (note that you are reporting the wrong units for that one, it is km-2 day-1) with a statistical process set to "accumulation" that would change the units to km-2 (but you need to be careful if you accumulate for less than 24h or a day, you will need a conversion day -> hours)
The reason why the units are km-2 day-1 is because m-2 s-1 would give you extremely small values: 1 km-2 day-1 (1 lightning per km2 per day) would be 1.157*10^-11 m-2 s-1
amilan17
Currently we use two parameters in our models – o lightning_flash_accumulation-PT01H o lightning_flash_accumulation-PT03H
Both are supplied in NetCDF and have units of m-2. The lightning is both cloud to cloud and cloud to ground. We need to map these across to the GRIB2 tables but couldn’t find an exact mapping. The WMO Code tables (version 26) provides the following – 0 Lightning strike density,,m-2 s-1 1 Lightning potential index (LPI),,J kg-1 2 Cloud-to-ground Lightning flash density,,km-2 day-1 3 Cloud-to-cloud Lightning flash density,,km-2 day-1 4 Total Lightning flash density,,km-2
The question is should these parameters be mapped in a similar way to precipitation accumulation i.e. use a rate code and then use the statistical process to sum it across a time period. The tentative suggestion is to choose the WMO GRIB2 parameter code 0-17-0 Lightning Strike Density. In the above code table, the units stated are m-2 s-1. So could the statistical code be then used to then indicate it is a sum over 1 hour or 3 hours. a) Is this the right approach b) that the actual gridpoint values would then not need to be translated since they would remain as being in m-2? c) Is the use of statistical code of 1 or 3 in units of hours ok to use when the WMO GRIB2 parameter code units are in m-2 s-1