GDAS0P5 are converted from the same GDAS dataset with GDAS1 data, right? Why GDAS1 data have w volocity while GDAS0P5 data don't?
When generating trajectories with GDAS0P5 data, HYSPLIT has to calculate w velocity from diversity, is it possible that the w velocity field calculated from diversity is very different from that in GDAS1 data?
Why GDAS0P5 data have no w velocity?
-
- Posts: 451
- Joined: November 9th, 2012, 4:23 pm
- Registered HYSPLIT User: Yes
Re: Why GDAS0P5 data have no w velocity?
GDAS0P5 are NOT converted from the same GDAS dataset with GDAS1 data.
For ARL's GDAS1, NCEP post-processes NCEP GDAS output into 1-degree, pressure level data which includes vertical velocity. ARL takes that and processes it.
For ARL's GDAS0P5, ARL post-processes NCEP GDAS output into half-degree, hybrid level data which DOES NOT include vertical velocity. Why does it not include vertical velocity? Because the program we are using (developed by someone at NCEP) does not output vertical velocity. In our long-range plans (several years) we want to add vertical velocity to the GDAS0P5.
This difference of processing has been clarified at http://ready.arl.noaa.gov/gdas1.php and ftp://arlftp.arlhq.noaa.gov/archives/gd ... 5_info.txt.
Yes, vertical velociy in HYSPLIT between GDAS1 (using model vertical velocity) and GDAS0P5 (divergence) can differ. HYSPLIT can be forced to use the divergence method with the any meteorology dataset, though I don't think this option is available through the ARL web.
For ARL's GDAS1, NCEP post-processes NCEP GDAS output into 1-degree, pressure level data which includes vertical velocity. ARL takes that and processes it.
For ARL's GDAS0P5, ARL post-processes NCEP GDAS output into half-degree, hybrid level data which DOES NOT include vertical velocity. Why does it not include vertical velocity? Because the program we are using (developed by someone at NCEP) does not output vertical velocity. In our long-range plans (several years) we want to add vertical velocity to the GDAS0P5.
This difference of processing has been clarified at http://ready.arl.noaa.gov/gdas1.php and ftp://arlftp.arlhq.noaa.gov/archives/gd ... 5_info.txt.
Yes, vertical velociy in HYSPLIT between GDAS1 (using model vertical velocity) and GDAS0P5 (divergence) can differ. HYSPLIT can be forced to use the divergence method with the any meteorology dataset, though I don't think this option is available through the ARL web.
Re: Why GDAS0P5 data have no w velocity?
Can I have the programs to convert 1-degree and 0.5-degree GDAS data into ARL data?barbara.stunder wrote:GDAS0P5 are NOT converted from the same GDAS dataset with GDAS1 data.
For ARL's GDAS1, NCEP post-processes NCEP GDAS output into 1-degree, pressure level data which includes vertical velocity. ARL takes that and processes it.
For ARL's GDAS0P5, ARL post-processes NCEP GDAS output into half-degree, hybrid level data which DOES NOT include vertical velocity. Why does it not include vertical velocity? Because the program we are using (developed by someone at NCEP) does not output vertical velocity. In our long-range plans (several years) we want to add vertical velocity to the GDAS0P5.
This difference of processing has been clarified at http://ready.arl.noaa.gov/gdas1.php and ftp://arlftp.arlhq.noaa.gov/archives/gd ... 5_info.txt.
Yes, vertical velociy in HYSPLIT between GDAS1 (using model vertical velocity) and GDAS0P5 (divergence) can differ. HYSPLIT can be forced to use the divergence method with the any meteorology dataset, though I don't think this option is available through the ARL web.
-
- Posts: 451
- Joined: November 9th, 2012, 4:23 pm
- Registered HYSPLIT User: Yes
Re: Why GDAS0P5 data have no w velocity?
A general program to convert from 1-degree grib format is in the HYSPLIT PC distribution. c:\hysplit4\data2arl\metprog.zip, gbl_grib. The programs for the 0p5 degree are not available.dolces wrote:Can I have the programs to convert 1-degree and 0.5-degree GDAS data into ARL data?barbara.stunder wrote:GDAS0P5 are NOT converted from the same GDAS dataset with GDAS1 data.
For ARL's GDAS1, NCEP post-processes NCEP GDAS output into 1-degree, pressure level data which includes vertical velocity. ARL takes that and processes it.
For ARL's GDAS0P5, ARL post-processes NCEP GDAS output into half-degree, hybrid level data which DOES NOT include vertical velocity. Why does it not include vertical velocity? Because the program we are using (developed by someone at NCEP) does not output vertical velocity. In our long-range plans (several years) we want to add vertical velocity to the GDAS0P5.
This difference of processing has been clarified at http://ready.arl.noaa.gov/gdas1.php and ftp://arlftp.arlhq.noaa.gov/archives/gd ... 5_info.txt.
Yes, vertical velociy in HYSPLIT between GDAS1 (using model vertical velocity) and GDAS0P5 (divergence) can differ. HYSPLIT can be forced to use the divergence method with the any meteorology dataset, though I don't think this option is available through the ARL web.
Re: Why GDAS0P5 data have no w velocity?
Hi there, I've an enquiry about 0.5 GDAS
I've run many trajectories with the 0.5 GDAS, using the divergence vertical velocity calc. and in comparison with all other meteo data sets I've used (with w included), the amount of vertical motion (lots) is surprising. Infact, in many instances the vertical velocity for identical trajectory start points is the opposite for GDAS 0.5 vs Others, with GDAS0.5 showing ascent to the point of trajectory initialisation and other met. data sets showing descent... as a result the two trajectories end up continents apart.
I'm coming at trajectories with a aerosol background, so not got a great head for meteo data sets, but would the use of 0.5 GDAS for source identification therefore be risky? Are the errors likely to be much greater using 0.5 GDAS, than say 1-Deg GDAS or ERA-Interim data?
If so, it's a bit risky having it on the hysplit ftp without a note of caution given that many (like me) will assume it is the highest resolution data-set with such an ease of access and use with hysplit?
Thanks in advance
I've run many trajectories with the 0.5 GDAS, using the divergence vertical velocity calc. and in comparison with all other meteo data sets I've used (with w included), the amount of vertical motion (lots) is surprising. Infact, in many instances the vertical velocity for identical trajectory start points is the opposite for GDAS 0.5 vs Others, with GDAS0.5 showing ascent to the point of trajectory initialisation and other met. data sets showing descent... as a result the two trajectories end up continents apart.
I'm coming at trajectories with a aerosol background, so not got a great head for meteo data sets, but would the use of 0.5 GDAS for source identification therefore be risky? Are the errors likely to be much greater using 0.5 GDAS, than say 1-Deg GDAS or ERA-Interim data?
If so, it's a bit risky having it on the hysplit ftp without a note of caution given that many (like me) will assume it is the highest resolution data-set with such an ease of access and use with hysplit?
Thanks in advance
-
- Posts: 451
- Joined: November 9th, 2012, 4:23 pm
- Registered HYSPLIT User: Yes
Re: Why GDAS0P5 data have no w velocity?
Have you tried running the trajectory ensemble for a starting point/time where you get a large difference and a case where they are similar? I wonder if the case with the w-value in the model output would be within the envelope of the ensemble with the gdas0p5? The ensemble provides an estimate of the uncertainty.
Please give me the trajectory inputs (date-time, latitude-longitude, altitude) and the trajectory duration for a couple cases in which the gdas0p5 and gdas1 give large and small differences, and I'll look at them.
Trajectories do not show the effects of turbulence in the atmosphere, so any single trajectory will not necessarily show the actual transport path of an air parcel.
Please give me the trajectory inputs (date-time, latitude-longitude, altitude) and the trajectory duration for a couple cases in which the gdas0p5 and gdas1 give large and small differences, and I'll look at them.
Trajectories do not show the effects of turbulence in the atmosphere, so any single trajectory will not necessarily show the actual transport path of an air parcel.
-
- Posts: 14
- Joined: November 8th, 2012, 3:41 pm
- Registered HYSPLIT User: Yes
Re: Why GDAS0P5 data have no w velocity?
When there are strong vertical motions, the vertical velocity using the integrated divergence and the vertical velocity generated by the meteorological model will give similar results. An example from the 1-degree GDAS is shown. For situations where the motions are weaker, the divergence calculation may be driven by the uncertainties of small differences in the horizontal velocity gradients.