GLM data not aggregating
GLM data not aggregating
I am trying to visualize some GLM data. These data are composed of several NetCDF files per minute, each with 20 sec of data so it doesn't make sense to load individually, but rather aggregate my time. I tried aggregating a bunch in Mc-V and I got an error. Are there any plans for supporting these data in Mc-V. Will the McIDAS-X GLM ADDE server be packaged up with Mc-V any time soon? I was hoping to use/view this data for my research sooner than later. At first glance the files do not seem to be particularly well formatted in terms of array structure and dimensions.
Re: GLM data not aggregating
Hi Kris -
Yuan, from the Unidata IDV group, put together a plugin that allows for aggregating these GLM netCDF files into one data source. You can pick up the plugin here:
ftp://ftp.ssec.wisc.edu/pub/incoming/GLM_L2.jar
Once you install this plugin (via Tools>Plugins>Manage from the Main Display window) and restart McV, you will see 3 new Data Types available through the General>Files/Directories chooser:
If you are interested in plotting EVENTS, you would choose the "GLM EVENTS Data files" Data Type, select your files, and click Add Source. Note that you will want to bin the times up a bit to create a usable loop. You can do the binning by right-clicking on the data source in the Field Selector and going to Properties. For my testing that I've done thus far, I've been setting Bin Size and Round To to 5 minutes. Choose the Point Data field, the Point Data Plot display type, and hit Create Display. Note that if you want to plot group, flash, and event, you will have to create 3 individual data sources for each one, using the different dataset types. It may also be beneficial to create layout models for group, flash, and event so you know what is what when it is plotted. For example, I plot all 3 with a small plus sign at the location, but use different colors for group, flash, and event.
Note that this plugin has gone through some pretty minimal testing on my end thus far. I haven't done too much side-by-side with McX to verify the output, but what I've seen thus far has been promising. I've given this plugin to a couple scientists in our building, but I haven't heard anything back from them yet. If you give this plugin a try and see any bugs or have comments/suggestions, please let me know.
As for getting GLM data working in McV, this is part of Inquiry 2377. There has been discussion on getting the GLM ADDE servers into McV, but we don't yet have an idea of the scope of work to get this done. Getting the servers themselves in may be easy, but it may require a new chooser, so I can't estimate on the timeline yet. However, this is set as a Critical priority inquiry and it is definitely on the programmer's minds.
I'll update you if there are any updates on getting GLM ADDE working in McV, or if there are any new versions of the plugin. FYI this plugin will remain on our ftp server for about a week before it will be removed. If you need it after it's been removed, let me know and I'll re-post it.
Thanks -
Bob Carp
Yuan, from the Unidata IDV group, put together a plugin that allows for aggregating these GLM netCDF files into one data source. You can pick up the plugin here:
ftp://ftp.ssec.wisc.edu/pub/incoming/GLM_L2.jar
Once you install this plugin (via Tools>Plugins>Manage from the Main Display window) and restart McV, you will see 3 new Data Types available through the General>Files/Directories chooser:
- GLM FLASHES Data files
- GLM EVENTS Data files
- GLM GROUPS Data files
If you are interested in plotting EVENTS, you would choose the "GLM EVENTS Data files" Data Type, select your files, and click Add Source. Note that you will want to bin the times up a bit to create a usable loop. You can do the binning by right-clicking on the data source in the Field Selector and going to Properties. For my testing that I've done thus far, I've been setting Bin Size and Round To to 5 minutes. Choose the Point Data field, the Point Data Plot display type, and hit Create Display. Note that if you want to plot group, flash, and event, you will have to create 3 individual data sources for each one, using the different dataset types. It may also be beneficial to create layout models for group, flash, and event so you know what is what when it is plotted. For example, I plot all 3 with a small plus sign at the location, but use different colors for group, flash, and event.
Note that this plugin has gone through some pretty minimal testing on my end thus far. I haven't done too much side-by-side with McX to verify the output, but what I've seen thus far has been promising. I've given this plugin to a couple scientists in our building, but I haven't heard anything back from them yet. If you give this plugin a try and see any bugs or have comments/suggestions, please let me know.
As for getting GLM data working in McV, this is part of Inquiry 2377. There has been discussion on getting the GLM ADDE servers into McV, but we don't yet have an idea of the scope of work to get this done. Getting the servers themselves in may be easy, but it may require a new chooser, so I can't estimate on the timeline yet. However, this is set as a Critical priority inquiry and it is definitely on the programmer's minds.
I'll update you if there are any updates on getting GLM ADDE working in McV, or if there are any new versions of the plugin. FYI this plugin will remain on our ftp server for about a week before it will be removed. If you need it after it's been removed, let me know and I'll re-post it.
Thanks -
Bob Carp
Re: GLM data not aggregating
Thanks, this mostly worked for me. I wanted 1-min bins, so I did the following in the attached image. I turned off the Objective Analysis Make Grid Fields because it was taking a very long time and I just wanted to see the point data. Despite my choosing 1 min bins, 20-sec data is still being displayed. Otherwise all is well
- Attachments
-
Re: GLM data not aggregating
Hi Kris -
I'm finding that I need to set both "Bin Size" and "Round To" to 1 minute to get a loop of 1 minute data. It looks like with your case (Bin Size 1; Round To 0) things are getting confused since the actual data is in less than 1 minute intervals (20 seconds). I noticed some similar oddities in my initial testing, and wrote up McIDAS-V Inquiry 2591 to better handle time binning in cases where there are multiple timesteps per minute. I've added this case to the inquiry so it can be investigated further.
Thanks for reporting this -
Bob
I'm finding that I need to set both "Bin Size" and "Round To" to 1 minute to get a loop of 1 minute data. It looks like with your case (Bin Size 1; Round To 0) things are getting confused since the actual data is in less than 1 minute intervals (20 seconds). I noticed some similar oddities in my initial testing, and wrote up McIDAS-V Inquiry 2591 to better handle time binning in cases where there are multiple timesteps per minute. I've added this case to the inquiry so it can be investigated further.
Thanks for reporting this -
Bob