If these are model output data, then there shouldn't be any missing values,
anyway. A possible source of small negative mixing ratios is advection errors.
The model might not make any corrections to variables that should be positive
definite. If you can visualize the data, see if the negative values have a
pattern around larger values, which would indicate undershoots that are more
likely in strong gradients. Anyway, if the values are small compared to typical
values, then I would just set them to zero. If they are large negative, then
maybe something is wrong with applying the scale factor and offset.
-- Ted
__________________________________________________________
| Edward Mansell <ted.mansell@xxxxxxxx>
| National Severe Storms Laboratory
|--------------------------------------------------------------
| "The contents of this message are mine personally and
| do not reflect any position of the U.S. Government or NOAA."
|--------------------------------------------------------------
> On Nov 4, 2017, at 3:18 PM, Richard Cloete <richard.cloete@xxxxxxxxxxx> wrote:
>
> Hi! Sure, here is the dump:
>
> S:\>ncdump -h -s aermr04_2017_1_1_0.nc
>
> netcdf aermr04_2017_1_1_0 {
> dimensions:
> longitude = 2880 ;
> latitude = 1441 ;
> level = 21 ;
> time = UNLIMITED ; // (1 currently)
> variables:
> float longitude(longitude) ;
> longitude:units = "degrees_east" ;
> longitude:long_name = "longitude" ;
> float latitude(latitude) ;
> latitude:units = "degrees_north" ;
> latitude:long_name = "latitude" ;
> int level(level) ;
> level:units = "millibars" ;
> level:long_name = "pressure_level" ;
> int time(time) ;
> time:units = "hours since 1900-01-01 00:00:0.0" ;
> time:long_name = "time" ;
> time:calendar = "gregorian" ;
> short aermr04(time, level, latitude, longitude) ;
> aermr04:scale_factor = 1.84160928156453e-012 ;
> aermr04:add_offset = 6.03421697197432e-008 ;
> aermr04:_FillValue = -32767s ;
> aermr04:missing_value = -32767s ;
> aermr04:units = "kg kg**-1" ;
> aermr04:long_name = "Dust Aerosol (0.03 - 0.55 um) Mixing
> Ratio" ;
>
> // global attributes:
> :Conventions = "CF-1.6" ;
> :history = "2017-10-30 22:48:34 GMT by grib_to_netcdf-2.5.0:
> grib_to_netcdf
> /data/data04/scratch/_mars-atls02-a82bacafb5c306db76464bc7e824bb75-6HA6Rj.grib
> -o
> /data/data03/scratch/_grib2netcdf-atls00-98f536083ae965b31b0d04811be6f4c6-AYhwQY.nc
> -utime" ;
> :_Format = "64-bit offset" ;
> }
>
> From: Ed Hartnett <edwardjameshartnett@xxxxxxxxx>
> Date: Saturday, 4 November 2017 at 20:13
> To: Richard Cloete <richard.cloete@xxxxxxxxxxx>
> Cc: Ketan Kulkarni <ketan.kulkarni@xxxxxxxxx>, Daniel Neumann
> <daniel.neumann@xxxxxxxxxxxxxxxxx>, "netcdfgroup@xxxxxxxxxxxxxxxx"
> <netcdfgroup@xxxxxxxxxxxxxxxx>
> Subject: Re: [netcdfgroup] Negative dust values
>
> Howdy!
>
> I suggest you run ncdump -h -s <FILENAME> and post the results to the mailing
> list.
>
> Thanks,
> Ed Hartnett
>
> On Sat, Nov 4, 2017 at 2:02 PM, Richard Cloete <richard.cloete@xxxxxxxxxxx
> <mailto:richard.cloete@xxxxxxxxxxx>> wrote:
> Thanks everyone!
>
> It turns out that there is an attribute called “_FillValue”. There are also
> other attributes such as “scale_factor”, “add_offset” and “missing_value”
> Curiously, FillValue == missing_value. The negative values are actually in
> the NetCDF file (I used Panoply to open it). The negative values found in the
> NetCDF file are different from the _FillValue, though, which is odd: there
> cannot be a negative Dust Aerosol (0.03 - 0.55 um) Mixing Ratio.
>
> Kind regards,
> Richard
>
> From: <netcdfgroup-bounces@xxxxxxxxxxxxxxxx
> <mailto:netcdfgroup-bounces@xxxxxxxxxxxxxxxx>> on behalf of Ketan Kulkarni
> <ketan.kulkarni@xxxxxxxxx <mailto:ketan.kulkarni@xxxxxxxxx>>
> Date: Saturday, 4 November 2017 at 15:26
> To: Daniel Neumann <daniel.neumann@xxxxxxxxxxxxxxxxx
> <mailto:daniel.neumann@xxxxxxxxxxxxxxxxx>>
> Cc: "netcdfgroup@xxxxxxxxxxxxxxxx <mailto:netcdfgroup@xxxxxxxxxxxxxxxx>"
> <netcdfgroup@xxxxxxxxxxxxxxxx <mailto:netcdfgroup@xxxxxxxxxxxxxxxx>>
> Subject: Re: [netcdfgroup] Negative dust values
>
> Check if the data is in packed netcdf format which has scale and offset
> values. I remember era interim data is in such format.
>
> On Nov 4, 2017 4:17 PM, "Daniel Neumann" <daniel.neumann@xxxxxxxxxxxxxxxxx
> <mailto:daniel.neumann@xxxxxxxxxxxxxxxxx>> wrote:
> Hi Richard,F
>
> If you find strange/unexpected values in your data you should contact the
> data creator. Possibly a contact person is noted in the global attributes of
> your netCDF file. I am not sure whether ECMWF provides only their own data or
> also data of third parties.
>
> What might be the reason for the negative values: If you look into the
> original netCDF file (e.g. with "ncdump -h FILENAME") you will possibly find
> an attribute "_fillvalue". Depending on the software you used to process the
> netCDF file, all values equal to "_fillvalue" are ignored or not (they should
> be ignored and kept untouched). Because _fillvalue is often negative, it came
> to my mind that your software possibly treats the _fillvalues like normal
> values. Because the atmospheric pressure is not constant, you will get
> slightly different negative values in space and time (where the _fillvalues
> have been before). At least this might be a reason for getting the negative
> values.
>
> Cheers,
> Daniel
>
> On 04.11.2017 15:07, Richard Cloete wrote:
> Hi all,
>
> Pretty new to NetCDF and ECMWF data.
>
> Just a quick (probably with an obvious answer) question: What do I do with
> negative values? For example, I am using aermr04, aermr05 and aermr06 (dust)
> data from ECMWF. I have noticed many negative values.
> What do I do with these data? Shall I just treat them as zeros? I ask because
> I am calculating the concentrations using the following
> formula:https://software.ecmwf.int/wiki/display/CKB/How+to+compute+concentration+from+a+mixing+ratio+field
>
> <https://software.ecmwf.int/wiki/display/CKB/How+to+compute+concentration+from+a+mixing+ratio+field>
>
> Many thanks,
> Richard
>
>
> _______________________________________________
> NOTE: All exchanges posted to Unidata maintained email lists are
> recorded in the Unidata inquiry tracking system and made publicly
> available through the web. Users who post to any of the lists we
> maintain are reminded to remove any personal information that they
> do not want to be made public.
>
>
> netcdfgroup mailing list
> netcdfgroup@xxxxxxxxxxxxxxxx <mailto:netcdfgroup@xxxxxxxxxxxxxxxx>
> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
> <http://www.unidata.ucar.edu/mailing_lists/>
>
> _______________________________________________
> NOTE: All exchanges posted to Unidata maintained email lists are
> recorded in the Unidata inquiry tracking system and made publicly
> available through the web. Users who post to any of the lists we
> maintain are reminded to remove any personal information that they
> do not want to be made public.
>
>
> netcdfgroup mailing list
> netcdfgroup@xxxxxxxxxxxxxxxx <mailto:netcdfgroup@xxxxxxxxxxxxxxxx>
> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
> <http://www.unidata.ucar.edu/mailing_lists/>
>
> _______________________________________________
> NOTE: All exchanges posted to Unidata maintained email lists are
> recorded in the Unidata inquiry tracking system and made publicly
> available through the web. Users who post to any of the lists we
> maintain are reminded to remove any personal information that they
> do not want to be made public.
>
>
> netcdfgroup mailing list
> netcdfgroup@xxxxxxxxxxxxxxxx
> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
