Clustering analysis with ensemble trajectory

[plastic_th]

Hi all,
I found some difficulties in clustering analysis. I used PC hysplit version
I have tried to run ensemble trajectory and got 27 multiple trajectories, and then I used simple script to divide into 27 tdump single trajectory files. ( it makes look like the single run file!)
Then I put these 27 tdump files into clustering analysis in endpts file. I tried too many times, the clustering analysis still failure! the simulation log message as showed in following:


Model started ... it is slower at the beginning
Started Cluster (Version: January 2010)
Number of trajectories in INFILE: 18
WARNING Should have at least 30 trajectories to cluster. INFILE, Skip:
18 1



I guess my tdump files have the same staring year , month, day or hour and which will cause the error message. So, I have a second trial, I run 27 single trajectory in different day and replace the tdump file in last 4 columns which are the position latitude, longitude, height and pressure level. but it got the same error message! Does anyone help me?
I hope I can calculate the cluster mean in 5 clusters, which can get the percentage of each cluster!

Thanks!

[plastic_th]

I tried to modify the tdump files, the clustering analysis show the following error message:

Model started ... it is slower at the beginning
Started Cluster (Version: January 2010)
Number of trajectories in INFILE: 27
WARNING Should have at least 30 trajectories to cluster. INFILE, Skip:
27 1
ERROR allocation time array: 1275068517


what is the problem of my tdump file? please help!