wrong units for deposition in FLEXPART-WRF?

[ikovalets]

Dear colleagues,
It seems that deposition units in output NETCDF files of FLEXPART-WRF are defined erroneously (pg·m-2) while It should be (ng·m-2)

This problem was first noted because of extremely high underestimation of deposition following release scenario as compared to results of simulations of other models. E.g. dry deposition calculated by the RODOS system for the same site, same release scenario and using the same meteorological data was on average by about 3 orders of magnitude higher as compared to FLEXPART results.

The source of potential error was identified by analysis of the FLEXPART code. The variable drygrid2 which is written to output netcdf file is defined in subroutine nc_write_output (source file netcdf_output_mod) as:

drygrid2(ix,jy,kp,nage)=1.e12*drygrid(ix,jy)/area(ix,jy)

Thus converstion to pico- units is done by multiplication on 1E12 factor here. Further inspection of code showed that the variable drygrid is in case of 1 age class and single specie simply equal to the variable drygridunc2, while the latter is defined in time loop when running model in the subroutine drydepokernel:

drygridunc2(ix,jy,ks,kp,nunc,nage)= &

drygridunc2(ix,jy,ks,kp,nunc,nage)+deposit(ks)

Regarding the variable deposit the comment in the same subroutine states that:

! deposit amount (kg) to be deposited
Thus final deposition should be in pKg·m-2=ng·m-2.

Attached is file with formulas used for conversion from kg to Bq.

[pesei]

Thank you for reporting. However, this is already mentioned in #237.

[ahilboll]

It's true that this is mentioned in #237, but #237 is about FLEXPART10, and this bug seems to be about FLEXPART-WRF. To me as an outsider, these seem to be different models (well, not really, but they live in two different repositories -- is there even a repository for FLEXPART-WRF?), so maybe bugs should be tracked independently?