HYSPLIT Forum: hysplitbbs.arl.noaa.gov

I am doing some comparison on hysplit trajectories generated from GDAS1 and GDAS0P5 data, and now I want a set of data which are in 0.5 degree resolution but with 23 vertical layers as GDAS1 data, how can I extract these data from GDAS0P5 data?

Thanks.

The HYSPLIT spatial extract program xtrct_grid, only extracts the lowest X number of levels, you cannot tell it which particular levels to get.

The GDAS1 levels are on pressure surfaces and the GDAS0P5 are on hybrid sigma-pressure, so except for the higher levels, the comparison would not be direct.

barbara.stunder wrote:The HYSPLIT spatial extract program xtrct_grid, only extracts the lowest X number of levels, you cannot tell it which particular levels to get.

The GDAS1 levels are on pressure surfaces and the GDAS0P5 are on hybrid sigma-pressure, so except for the higher levels, the comparison would not be direct.

Does it mean that the variables in GDAS1 and GDAS0P5 data may be very different since they are in different vertical coordinates? Because I find that the trajectories generated from GDAS1 and GDAS0P5 can be very different from each other, in this case, which trajectory should I trust?

And I don't understand as the GDAS1 data are on pressure surfaces, why there are still values on Tibetan Plateau on 1000 hPa level, where the surface pressure is much lower than 1000 hPa?

I want to know
Can we combine multiple GDAS0p5 files

Yes, with the 'cat' command.
cat day2 >> day1

dolces wrote:

barbara.stunder wrote:The HYSPLIT spatial extract program xtrct_grid, only extracts the lowest X number of levels, you cannot tell it which particular levels to get.

The GDAS1 levels are on pressure surfaces and the GDAS0P5 are on hybrid sigma-pressure, so except for the higher levels, the comparison would not be direct.

Does it mean that the variables in GDAS1 and GDAS0P5 data may be very different since they are in different vertical coordinates? Because I find that the trajectories generated from GDAS1 and GDAS0P5 can be very different from each other, in this case, which trajectory should I trust?

And I don't understand as the GDAS1 data are on pressure surfaces, why there are still values on Tibetan Plateau on 1000 hPa level, where the surface pressure is much lower than 1000 hPa?

HYSPLIT linearly interpolates the variables in the horizontal and vertical, so in general interpolating between two levels that are close together should be better than two that are far apart. Pressure vs. hybrid doesn’t matter so much as how many levels there are and how well they represent the wind gradients in the atmosphere.

Remember a trajectory shows one air parcel’s transport by the mean wind, it does not account for turbulence in the atmosphere; HYSPLIT dispersion accounts for turbulence. Methods such as running a HYSPLIT trajectory ensemble, or manually running trajectories varying the starting height somewhat can give an indication of the gradients. See trajectory error information given at one of the workshops http://www.arl.noaa.gov/HYSPLIT_workshop.php

Yes, the 1000 hPa level there is underground, and that is one benefit of using the data on hybrid levels. Nevertheless, the NCEP models output data at the fixed pressure levels. With the pressure level output, NCEP also outputs surface pressure and/or surface height, which is input to HYSPLIT.

How can I extract specific vertical layers from GDAS0P5 and re-write into a new ARL format data?

Are ARL data in ASCII format? Can I use MATLAB to read and re-write them?

dolces wrote:Are ARL data in ASCII format? Can I use MATLAB to read and re-write them?

Only the header for each record is in ASCII format. See the information and the sample FORTRAN program chk_data.f at the bottom of http://ready.arl.noaa.gov/archives.php

I am not familiar with MATLAB.

dolces wrote:How can I extract specific vertical layers from GDAS0P5 and re-write into a new ARL format data?

Can MeteoInfo do this? Link to MeteoInfo at http://ready.arl.noaa.gov/HYSPLIT.php