Re: [ferret_users] RE: Error in plotting time series data with 6.1?

Hi Mike,
There is a bug in the code that detects whether an axis is irregular! I also ran into this, just this week. We'll get it fixed in the next release. Unfortunately the code does a check for "is it actually regular" whether or not you have a point-spacing = regular attribute. I will change that also, so that if the file says it's irregular it will be treated as irregular no matter what.

What would help as a workaround is to make the time origin closer to the start of the data, say t0=1-jan-2000 if that's possible. The code is looking at the size of delta-t relative to the time coordinates themselves. Comparing the size of delta-t of seconds out of 1162774813 total seconds since the start of 1970 is more difficult than comparing the size of delta-t relative to a smaller number of seconds since year 2000. Again, we'll fix it so this WILL work, in the next version.

Ansley

McCann, Mike wrote:

Hi again,

I generated another netCDF file with fewer points (10-minute interval sampling from the original data) and Ferret properly recognizes the time axis as irregular:
> ferret
        NOAA/PMEL TMAP
        FERRET v6.1
        Linux(g77) 2.4.21-32 - 03/25/08
        26-Aug-08 13:11
yes? use "http://dods.mbari.org/cgi-bin/nph-nc/data/ssdsdata/deployments/mse2006/200609/mse2006_ctd_20061106_original.nc"
yes? show grid GAF1
    GRID GAF1
 name       axis              # pts   start                end
 LONGITUDE LONGITUDE            1mr   122.89W(-122.89)     122.89W(-122.89)
 LATITUDE  LATITUDE             1 r   36.237N              36.237N
 NOMINALDEPTH DEPTH (m)         1 r-  3450                 3450
 ESECS     TIME             14760 i   06-NOV-2006 01:00    27-APR-2007 23:50
Ferret thinks that the 2-minute interval data is regular:
yes? use "http://dods.mbari.org/cgi-bin/nph-nc/data/ssdsdata/deployments/mse2006/200609/mse2006_ctd_20061106_original_120.nc"
yes? show grid GMH1
    GRID GMH1
 name       axis              # pts   start                end
 LONGITUDE LONGITUDE            1mr   122.89W(-122.89)     122.89W(-122.89)
 LATITUDE  LATITUDE             1 r   36.237N              36.237N
 NOMINALDEPTH DEPTH (m)         1 r-  3450                 3450
 ESECS1    TIME             68919 r   06-NOV-2006 01:00    27-APR-2007 23:48
Seems like Ferret is making a decision to make the time axis regular even though it's not. Should I provide some help by adding an attribute of 'point_spacing: "uneven"' to the time axis definition in the netCDF file?

-Mike
From: McCann, Mike
Sent: Tuesday, August 26, 2008 10:32 AM
To: ferret_users@xxxxxxxx
Subject: Error in plotting time series data with 6.1?

Hello,

We just upgraded to Ferret 6.1 on a Linux host and I'm getting strange results in plotting an irregular time series.

Here's the plot with an older version running on my Windows box (plots correctly):
bash-2.05$ ferret
        NOAA/PMEL TMAP
        FERRET v5.51
        cygwin 1.3.x - 02/28/03
        26-Aug-08 10:06
yes? use "http://dods.mbari.org/cgi-bin/nph-nc/data/ssdsdata/deployments/mse2006/200609/mse2006_ctd_20061106_original_120.nc"
yes? plot PRESSURE
Here's the plot with the current release running on my RH Linux (seems to treat the time series as regular) :
> ferret
        NOAA/PMEL TMAP
        FERRET v6.1
        Linux(g77) 2.4.21-32 - 03/25/08
        26-Aug-08 10:07
yes? use "http://dods.mbari.org/cgi-bin/nph-nc/data/ssdsdata/deployments/mse2006/200609/mse2006_ctd_20061106_original_120.nc"
yes? plot PRESSURE
    
Any help or suggestions are appreciated.

-Mike

--
Mike McCann
Software Engineer
Monterey Bay Aquarium Research Institute
7700 Sandholdt Road
Moss Landing, CA 95039-9644
Voice: 831.775.1769 Fax: 831.775.1736 http://www.mbari.org