How to fill missing 0.5deg data with another data source?

[hbarbosa]

I am trying to use the GDAS 0.5deg data set to do my Hysplit simulations. Its a 5day backward simulation starting on 7 march 2014, that means I need these files:

Code: Select all

20140308_gdas0p5
20140307_gdas0p5
20140306_gdas0p5
20140305_gdas0p5
20140304_gdas0p5
20140303_gdas0p5
20140302_gdas0p5

The problem is the missing files. At ARL repository I can only find this:

Code: Select all

20140308_gdas0p5
20140307_gdas0p5.missing 
20140306_gdas0p5
20140305_gdas0p5
20140304_gdas0p5
20140303_gdas0p5
20140302_gdas0p5

Reading the "readme_gdas0p5_missings.txt", I found the following:

Code: Select all

NOTE: GDAS1 or Reanalysis can be used to fill-in for the missings when running HYSPLIT.

But how is one supposed to do that? The GDAS1 files have one week in each file, while I'm missing just 1 day!

1) Should I just add the 1-week 1deg file together with the other 0.5 files? What will hysplit do when it sees two met-fields for the same time but with different resolution?

2) Or do I need to convert from 1week-1deg to 1day-0.5deg? if so, with which program?

3) Or is it something else?

Henrique

[ariel.stein]

Just add the 1 week gdas 1 degree to the control file. Hysplit will try to use the highest horizontal resolution (gdas 0.5), when the data is not available it will use the gdas 1 degree.

[hbarbosa]

Hi Ariel,

Just add the 1 week gdas 1 degree to the control file. Hysplit will try to use the highest horizontal resolution (gdas 0.5), when the data is not available it will use the gdas 1 degree.

Thanks! One more question: I'm runing Hysplit on a linux server, so it would be very easy to just make a symbolic link from the 1-week-1-deg file to the filename of my missing day(s) in the 0.5deg... Like this:

Code: Select all

hbarbosa@aerossol:/LFANAS/hbarbosa/arlftp.arlhq.noaa.gov/pub/archives/gdas0p5> ln -s ../gdas1/gdas1.mar14.w1 20140307_gdas0p5
hbarbosa@aerossol:/LFANAS/hbarbosa/arlftp.arlhq.noaa.gov/pub/archives/gdas0p5> ls -l 2014030*
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  1  2014 20140301_gdas0p5
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  2  2014 20140302_gdas0p5
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  3  2014 20140303_gdas0p5
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  4  2014 20140304_gdas0p5
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  5  2014 20140305_gdas0p5
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  7  2014 20140306_gdas0p5
lrwxrwxrwx 1 nobody nogroup        23 Aug 18  2015 20140307_gdas0p5 -> ../gdas1/gdas1.mar14.w1
-rw-rw-r-- 1 nobody nogroup         8 Aug 19  2013 20140307_gdas0p5.missing
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  8  2014 20140308_gdas0p5
-rw-rw-r-- 1 nobody nogroup 723756480 Mar  9  2014 20140309_gdas0p5

Would that work?

I mean, does HySplit relies on the file name to decide what is inside? or does it find it out by reading the grib descriptors?

[ariel.stein]

That should work. Hysplit reads the header of the file to see the horizontal resolution and the time coverage.