Opened 12 years ago

Closed 12 years ago

#213 closed help (fixed)

two questions for Jeff

Reported by: anmcr Owned by: jeff
Component: UM Model Keywords: ancillaries
Cc: Platform:
UM Version: 6.1

Description

Hi Jeff,

I have two questions for you, which I would be very grateful if you could consider.

1) I have run into the problem with the LNER function in deck RCF_IANCILA in which my modified seaice ancillary and dump grid spacing values do not match to 6 significant figures. I saw that you answered this problem in Ticket #109 by suggesting a reconfiguration modset.

I've implemented your suggestion but am having trouble getting it to work. I get the impression that the modset is not being read. One question I have is whether the modset should be labelled .mh, mf77, or .m90. This is experiment xdpxd, with the modset in /anmcr/work/UMAncil/20020221/N216.

Moreover, I don't know why I am getting this problem as in creating the seaice .nc ancillary file using IDL I was careful to use double precision and to use the 64 bit version of xancil when converting my .nc file.

Note that by preforming an additional interpolation step on my .nc file it works: from a 432x325 regular grid to a 432x325 grid with row spacing 0.555556. However I would rather not have to do this additional step.

2) In creating my own sst and seaice ancillary files, I am running into the problem of the data once it has been gridded to my UM grid having 'land' points which are not masked by the UM land sea mask. This is obviously a problem. Rather than change the UM land sea mask, this requires that these points are removed, requiring extrapolation. However, is a simpler method to set these points to NaN, and then use xconv to fill in the missing data?

Thanks for your help,

Andrew (anmcr)

Change History (2)

comment:1 in reply to: ↑ description Changed 12 years ago by jeff

  • Owner changed from um_support to jeff
  • Status changed from new to assigned

Hi Andrew

I have two questions for you, which I would be very grateful if you could consider.

1) I have run into the problem with the LNER function in deck RCF_IANCILA in which my modified seaice ancillary and dump grid spacing values do not match to 6 significant figures. I saw that you answered this problem in Ticket #109 by suggesting a reconfiguration modset.

I've implemented your suggestion but am having trouble getting it to work. I get the impression that the modset is not being read. One question I have is whether the modset should be labelled .mh, mf77, or .m90. This is experiment xdpxd, with the modset in /anmcr/work/UMAncil/20020221/N216.

It should be called .mf90, but it is still in fixed format so you need 6 extra spaces in front of the line.

Moreover, I don't know why I am getting this problem as in creating the seaice .nc ancillary file using IDL I was careful to use double precision and to use the 64 bit version of xancil when converting my .nc file.

Which netcdf files are you using to create the ancil files? Some are 32 bit and some are 64 bit, use ncdump -h ncfile to check, if the variables are float then they are 32 bit, doubles are 64 bit.

Note that by preforming an additional interpolation step on my .nc file it works: from a 432x325 regular grid to a 432x325 grid with row spacing 0.555556. However I would rather not have to do this additional step.

2) In creating my own sst and seaice ancillary files, I am running into the problem of the data once it has been gridded to my UM grid having 'land' points which are not masked by the UM land sea mask. This is obviously a problem. Rather than change the UM land sea mask, this requires that these points are removed, requiring extrapolation. However, is a simpler method to set these points to NaN, and then use xconv to fill in the missing data?

To get xconv to do this you need to set the points to the missing data value and each variable needs to have the attribute missing_value and/or _FillValue set to this value. Xconv will not work with NaNs?.

Jeff.

comment:2 Changed 12 years ago by jeff

  • Resolution set to fixed
  • Status changed from assigned to closed
Note: See TracTickets for help on using tickets.