Error in Cluster Analysis - Cannot read tdump files

Topics about the trajectory clustering program for HYSPLIT.
Post Reply
cg gc
Posts: 5
Joined: March 28th, 2018, 1:18 am
Registered HYSPLIT User: Yes

Error in Cluster Analysis - Cannot read tdump files

Post by cg gc »

I am trying to run cluster analysis on a long stream of back trajectories. I have 5, 7, and 10 day lengths and at multiple heights. I have tried running the cluster analysis several times with my files and can't seem to get them to read. I have tried using a few different sets of trajectory files. The groups all start at the same height and within the same year. No matter which ones I use, I receive one of two error messages.

If the "Hours to cluster" is less than 5, I receive this error:

Model started ... it is slower at the beginning
Started Cluster (Version: April 2015)
Number of trajectories in INFILE: 31
clusread - should not be here; this traj has fewer endpts than needed

If I use 5 or more hours, I receive this error message.
Model started ... it is slower at the beginning
Started Cluster (Version: April 2015)
Number of trajectories in INFILE: 31
NOTICE - No trajectories to cluster. Check trajectory duration
in the files in INFILE and/or the Hours to cluster

I have checked the tdump files and they are complete to at least -120 hours. I have placed the files in the endpts file. I can create the INFILE, but not the clusters. When I copied my trajectory data onto the example files, they work fine. Normally this wouldn't be a problem, but I have a few hundred clusters that need completed.

The only thing I can think of that might be causing problems is the header on my files is larger than the header on the example files. Is there a limit to the header length that can be used with the cluster analysis?
barbara.stunder
Posts: 451
Joined: November 9th, 2012, 4:23 pm
Registered HYSPLIT User: Yes

Re: Error in Cluster Analysis - Cannot read tdump files

Post by barbara.stunder »

I'm not sure what you mean by the trajectory header. When you created the trajectories, is meteorology output along the trajectory?

Are the trajectory endpoints files each one trajectory only (one starting height)?
cg gc
Posts: 5
Joined: March 28th, 2018, 1:18 am
Registered HYSPLIT User: Yes

Re: Error in Cluster Analysis - Cannot read tdump files

Post by cg gc »

Yes, every file is exactly the same, minus changes in the met data at the top. I do have different heights, but each file only has one trajectory and starts at one height.
Here is an example.

8 1
ECMG 0 3 15 0 0
ECMG 0 3 22 0 0
ECMG 0 3 29 0 0
ECMG 0 4 1 0 0
ECMG 0 4 8 0 0
ECMG 0 4 15 0 0
ECMG 0 4 22 0 0
ECMG 0 4 29 0 0
1 BACKWARD OMEGA
0 4 2 6 -71.603 111.263 10.0
10 PRESSURE THETA AIR_TEMP RAINFALL MIXDEPTH RELHUMID SPCHUMID H2OMIXRA TERR_MSL SUN_FLUX
1 1 0 4 2 6 0 0 0.0 -71.603 111.263 10.0 691.4 252.0 226.7 0.0 49.9 96.3 0.1 0.1 2737.2 NaN
1 1 0 4 2 5 0 0 -1.0 -71.942 111.659 15.6 684.6 252.0 226.1 0.0 43.5 96.6 0.1 0.1 2796.5 NaN
1 1 0 4 2 4 0 0 -2.0 -72.258 112.096 18.6 681.0 251.7 225.5 0.0 26.9 97.3 0.1 0.1 2834.8 NaN
1 1 0 4 2 3 0 0 -3.0 -72.551 112.575 22.1 676.9 251.4 224.9 0.0 14.4 97.8 0.1 0.1 2878.4 NaN
1 1 0 4 2 2 0 0 -4.0 -72.824 113.101 26.1 673.7 251.1 224.3 0.0 9.0 98.3 0.1 0.1 2912.6 NaN
1 1 0 4 2 1 0 0 -5.0 -73.075 113.680 30.8 670.7 250.9 223.9 0.0 9.0 99.0 0.1 0.1 2945.4 NaN
...
Shortened for space.
cg gc
Posts: 5
Joined: March 28th, 2018, 1:18 am
Registered HYSPLIT User: Yes

Re: Error in Cluster Analysis - Cannot read tdump files

Post by cg gc »

Also, I am not sure if this matters, but my trajectories start every 6 hours.

example:
tdump00120100
tdump00120106
tdump00120112
tdump00120118
barbara.stunder
Posts: 451
Joined: November 9th, 2012, 4:23 pm
Registered HYSPLIT User: Yes

Re: Error in Cluster Analysis - Cannot read tdump files

Post by barbara.stunder »

There is a problem with the SUN_FLUX values; they show as NaN in the tdump file. Maybe that field is not in the meteorology files. Re-run the trajectories without the SUN_FLUX, then it should work in the cluster program.
cg gc
Posts: 5
Joined: March 28th, 2018, 1:18 am
Registered HYSPLIT User: Yes

Re: Error in Cluster Analysis - Cannot read tdump files

Post by cg gc »

That fixed one problem, but there is still one more. Or rather, a new problem.

I can create the clusters, but not the means. I get an error message with a day old time stamp on it. Can someone translate this?

I think it has to do with the new trajectory files themselves because it works fine with trajectories that only have pressure and not all of the extra met data attached.

This is the error message:

Started Trajectory Drawing
Last Changed Rev: 854
Last Changed Date: 2017-01-12 10:47:54 -0500 (Thu, 12 Jan 2017)

ERROR: no trajectories to plot
900
Started Trajectory Drawing
Last Changed Rev: 854
Last Changed Date: 2017-01-12 10:47:54 -0500 (Thu, 12 Jan 2017)

ERROR: no trajectories to plot
900
while executing
"exec $exec_dir/trajplot.exe $arg1 $arg2 $arg3 $arg4 $arg5 ${arg6} $arg7 $arg8 $arg9 $arga $argb $argc"
(procedure "run_means" line 91)
invoked from within
"run_means 1"
invoked from within
".clusdisp.bot.save invoke"
("uplevel" body line 1)
invoked from within
"uplevel #0 [list $w invoke]"
(procedure "tk::ButtonUp" line 24)
invoked from within
"tk::ButtonUp .clusdisp.bot.save"
(command bound to event)
barbara.stunder
Posts: 451
Joined: November 9th, 2012, 4:23 pm
Registered HYSPLIT User: Yes

Re: Error in Cluster Analysis - Cannot read tdump files

Post by barbara.stunder »

Use the set of trajectories with only the pressure, not the other meteorological variables. I need to see why that is not working.
Post Reply

Return to “Cluster Analysis”