During the test phase with the Easy case, several questions popped up on the different
Mixing level height
The mixing level height can reach values that are very high for dispersion modelling. In
the ensemble data we have limited the value to 1500m.
The roughness length is not included in the standard Harmonie output. We extracted a
roughness-length field (times g
) from one of the analyses. Over land this can be
used throughout the period. Over sea the roughness length depends on the sea state, but
as the plumes will not go over sea, we propose to use this fixed field anyway.
Some of the models also need the surface height. We extracted a fixed geopotential height
field (surface height times g
) and included this in the ensemble data.
Over sea the same considerations hold as for the roughness length.
) is provided at surface level and
at hybrid levels. At surface level there are two precipitation fields. One is the one hour
accumulated precipitation in kg/m², which corresponds to mm/h. For most models this is the
precipitation you want. The other field is an indication of instantaneous precipitation,
converted to a mm/h amount.
At the hybrid levels log(precipitation)
(which can hence be negative) is given
in kg/kg (kg water per kg air).
Because these different precipitation variables gave rise to some confusion, and only the
accumulated preciptation at the surface is needed, we omitted the instantaneous parameters
at model levels and surface in the dataset for the Borssele-2 scenario.
GRIB coded data
In the GRIB coded data the parameter number is 181 and the accumulated and instantaneous
fields can be identified by the timeRangeIndicator
TR which is 0 for instantaneous
and 4 for accumulated quantities.
As an example for a +10 hour forecast the output of wgrib for the one hour accumulated
3:361224:d=17011106:LPSX:kpds5=181:kpds6=105:kpds7=0:TR=4:P1=9:P2=10:TimeU=1:0m above gnd:9-10hr acc:NAve=0
where P1=9 is the start of the accumulation and P2=10 the end.
The instantaneous surface field will show TR=0, P1=10 and P2=0.
The conversion from GRIB to NetCDF seems to have lost the information from the
In the NetCDF files which are provided in this project for the ensemble in the Borssele
case, the variables are called rain which is the 1 hour accumulated surface rain
with TR=4, rain_2 which is the rain at the hybrid levels, and rain_3 which
is the instantaneous surface rain with TR=0.
In the Borssele-2 dataset this variable is called just rain.