sounding does not show full wind profile

Errors and unexpected results
Post Reply
User avatar
hproe
Posts: 504
Joined: Sat Nov 27, 2010 3:46 pm

sounding does not show full wind profile

Post by hproe »

Hi -
In recent nightlies, using GFS analyses data, the formula '3D grid > Temperature > Derived > Sounding Data (with true winds)' does not show the full wind profile, but only some winds in the lower troposphere (see attached screen capture). Version 1.7u1 (still installed on my notebook) shows the complete wind profile. Can you reproduce this issue?
cheers, HP
Attachments
skewT.png
User avatar
bobc
Posts: 988
Joined: Mon Nov 15, 2010 5:57 pm

Re: sounding does not show full wind profile

Post by bobc »

Hi HP,

I'm not replicating your exact problem, but rather using the nightly build I'm not able to get any soundings plotted from this GFS 0.5degree analysis data. There's an error at Create Display complaining about units not matching up. Where things are going awry on my end is that the dewpoint isn't being derived correctly. With the formula that calculates dewpoint as-is McV picks a temperature field with only a couple vertical levels ("Temperature_height_above_ground" - units of meters) and relative humidity with many vertical levels ("Relative_humidity_isobaric" - units of Pa).

I'm testing out some changes where I updated the dewpoint (and mixing ratio) formulas to use a parameter group to make sure that the T and RH variables passed through the formula have common levels/units. Thus far, things seem to be working well on my end. I'll let you know when the nightly gets updated to these changes.

Could you please try starting up McIDAS-V with -tempuserpath ("runMcV.bat -tempuserpath" from a command prompt) and see if you run into the problem I mentioned (error at Create Display), or if you're still seeing the problem you reported where the T/Td profiles look fine, but not the wind? I'm curious if you may have made modifications to the formulas or parameter aliases/groups on your end that are making things work this way. For example, maybe Temperature and Relative Humidity @ isobaric levels are being used, but the incorrect u/v wind fields are being used.

Thanks for reporting this,
Bob Carp
User avatar
hproe
Posts: 504
Joined: Sat Nov 27, 2010 3:46 pm

Re: sounding does not show full wind profile

Post by hproe »

Hi Bob -

Sorry for the delay. I have tested with McV starting from scratch and see the following things (nightly of 29 Oct 09:30):

1. There are 2 instances of 'Sounding Data (with true winds)'. In my configuration I have 4 (presumably the 2 repeated).

2. Using the first instance I get the error message 'Unknown vertical coordinate with Unit null, using linear scale to Altitude', but giving the OK processing goes on and the complete sounding is shown, without dew points though.

3. Using the second instance no error message appears, only the 2 lowest levels are shown all data present (hard to see in the graph but well present in the table).

Going back back to my configuration (same nightly) I see the same behaviour as above, with the difference that in the complete sounding also the dew points are present.

cheers, HP
User avatar
bobc
Posts: 988
Joined: Mon Nov 15, 2010 5:57 pm

Re: sounding does not show full wind profile

Post by bobc »

Hi HP,

Thanks for following up on this. It's still a work in progress on getting my code changes into the nightly. For the time being, if you would like, you could give a plugin I created a try. If you are interested, you can pick it up from our anonymous ftp server using the following commands from a terminal/command prompt window:

ftp ftp.ssec.wisc.edu
user: anonymous
password: your email address
binary
cd pub/incoming
get tempRh.jar
bye

This will remain on ftp for about a week before it is automatically removed. If you want the plugin but don't get a chance to grab it before it's removed, let me know and I can re-post it. Once you install this plugin and restart McV, add your GFS data source. I'm not exactly sure of what your setup looks like, but after I install this plugin and add a data source there are 2 "Sounding Data (with true winds)" derived fields listed. The first one plots everything correctly, with the full T, Td, and wind profiles. The second one gives an error and then only plots windbarbs at the two lowest levels.

This plugin added a parameter group for "Temperature and Relative Humidity". The plugin also modified a few of the derived formulas to use the parameter group, including Dewpoint, Mixing Ratio, and Equivalent Potential Temperature. Since the formula now derives Dewpoint correctly, the sounding works better. The reason that this isn't in the actual nightly yet is that the edited formulas from the plugin as well as the the IDV's version of the formulas are both being used. This means that several derived fields will be listed multiple times, where some work and some don't. We will implement a fix so that only these new formulas using the parameter groups will be used which should remove any of the derived fields that currently error.

When these code changes are added you can uninstall the plugin. Hopefully this can at least get you up and running in the nightly for now. Please let me know if you have any problems/questions related to the plugin. I'll keep you updated on when this code gets added to the nightly.

Thanks,
Bob
User avatar
hproe
Posts: 504
Joined: Sat Nov 27, 2010 3:46 pm

Re: sounding does not show full wind profile

Post by hproe »

Hi Bob -

Sincere thanks for your effort. I got your plugin working nicely with the first instance of the formula, be it in my configuration or when starting from scratch. The second instance delivers now the 3 lowest levels in the table without plottable data.

However, I have discovered another problem. When using 'Match Display Region' the profile is constructed only at 0deg lon/0deg lat. When I move the pointer off this point into the ROI the graph and table are void. Selectng manually the region with 'Select A Region' all is well.

HP
User avatar
bobc
Posts: 988
Joined: Mon Nov 15, 2010 5:57 pm

Re: sounding does not show full wind profile

Post by bobc »

Hi HP,

I'm glad that the plugin is working for you.

I'm somewhat replicating your problem with "Match Display Region", but not exactly. Things differ dependent on where things are centered. For example:

  1. Display centered over Wisconsin, USA from the World projection, zoomed in pretty far: Match Display Region works fine. The sounding seems to work everywhere in this domain.
  2. Display in the Europe projection without adjusting the zooming: Using Match Display Region, I can only get the sounding to work if the probe is to the right of the prime meridian. If I move the probe to the left, then the sounding goes away.

There are known issues with grids that straddle the prime meridian related to the netCDF version that we package in McIDAS-V. I think that this problem that I'm seeing is related to those issues. Just as an FYI, the IDV group mentioned that they're going to begin trying to update to netCDF/java 5.2. One major improvement that we're hopeful will come from this work is that these dateline issues with grids could all be resolved, including subsetting across the dateline in the Region tab of the Field Selector. It will be a lot of work to move to this new netCDF/java version, so it will take take some to get done. I'll let you know how things progress with this.

Since I'm not replicating your exact problem, could you please let me know what your display window looks like when you try using Match Display Region? For example, what projection are you in and did you adjust the zoom level at all?

Thanks again,
Bob
User avatar
hproe
Posts: 504
Joined: Sat Nov 27, 2010 3:46 pm

Re: sounding does not show full wind profile

Post by hproe »

Hi Bob -

First of all, I am sure that Europeans will be thrilled once the 0deg-meridian issue will be gone.
I am working in the Mercator projection. In my most recent case, I have been zooming on the Azores (spent gorgeous hiking holidays in September), i.e. zoomed west of deg.

best, HP
Post Reply