Wrong units of accumulated precipitation in FCTABLES

[svianaj]

In Deode-Workflow, precipitation FCTABLES are extracted from pcp variables with a param_file like this for HARMONIE or AROME:
(...)
{ "harp_param":"Pcp",
"method":"nearest",
"grib_id":[ {"parameterNumber":"65"},
{"parameterNumber":"66"},
{"parameterNumber":"75"}
],
"common":{ "productDefinitionTemplateNumber":"8",
"typeOfStatisticalProcessing":"accum",
"parameterCategory":"1"
},
"function":"sum"
},
(...)

Currently in the output gribfiles the accumulated total rain seems to have some wrongly encoded info , including units:

#============== MESSAGE 104 ( length=1555857 ) ==============
GRIB {
Meteorological products (grib2/tables/15/0.0.table)
discipline = 0;
editionNumber = 2
(...)
Average, accumulation, extreme values or other statistically processed values at a horizontal level or in a horizontal layer in a continuous or non-continuous time interval (grib2/tables/15/4.0.table)
productDefinitionTemplateNumber = 8;
Moisture (grib2/tables/15/4.1.0.table)
parameterCategory = 1;
Rain precipitation rate (kg m-2 s-1) (grib2/tables/15/4.2.0.1.table)
parameterNumber = 65;
-READ ONLY- parameterUnits = kg m-2 s-1;
-READ ONLY- parameterName = Rain precipitation rate;
(....)
stepRange = 0-31;
shortName = tirf;
name = Time integral of rain flux;
cfName = unknown;
-READ ONLY- cfVarName = tirf;

The grib message contains accumulated rain (kg m-2) according to the data values, but the output FCTABLES has kg m-2 s-1 instead of kg m-2 due to this bad metadata in the grib file. This complicates things when using oper-harp-verif for running the verification. These scripts could handle a change of units through a suitable configured scale_fcst in the params_list (set_params.R), but we're already doing so for the references (IFS, Global_DT) where the precipitation comes in m.

Ideally, I would prefer a way to rename units in grib2sqlite instead. The wrong units might be fixed sometime in Deode-Workflow, but if e.g. we could have an optional "units" keyword in the param_file, this could help also e.g. to rename "percent" by % or mb to hPa...
@adeckmyn

[adeckmyn]

Ideally, of course, this should be handled in the [decoding of the] grib message. But accumulated precipitation has always been a nuisance. The different conventions (like storing it as the accumulation of a rain rate) make it particularly hard to deal with. So I agree that a manual override is probably the most sensible thing.

[svianaj]

On the other hand, I just noticed that DW OD runs don't currently have parameterNumbers 66 & 75 ( for AROME and HARMONIE) in the output. Meanwhile, for ALARO all pcp components are there. So I need to open a new issue now in Deode-Workflow.