Marine Data Literacy 2.0

Providing instruction for managing, converting, analyzing and displaying oceanographic station data, marine meteorological data, GIS-compatible marine and coastal data or model simulations, and mapped remote sensing imagery

 

 

 

 

Home > 9. Operational/Synoptic > 9.20 CORIOLIS Profiles

9.20 Visualizing Operational Subsurface Ocean Data in ODV & IDV: CORIOLIS

1.  Much of the beginning of this exercise parallels the steps in these ODV exercises:

If your intention is only to work with ODV, then we recommended you drop this exercise and follow those exercises, because they provide more details relating specifically to ODV, where the subsurface grids can better be created and displayed.

2.  Deprecated panel.
3.  Open the CORIOLIS Data Selection page to see thousands of recently collection profiles and underway data points.
4.  Below the map there are 2 types of data to request:  VERTICAL PROFILES and time-series.

Select all the vertical profiles

5.  De-select the time-series this time by removing the check marks.
6.  Make these other settings:
  • Start Date:  2 weeks before today
  • End Date:  Today's date
  • Geographic coordinates:  Enter the Liberia area of interest values.  Don't use plus/minus signs, use NSEW convention.
5.  Make these selections.
6.  Click the REFRESH control to make your selections active.
7.  You should see this map.  If the number of stations is far fewer, then increase the time window to one full month (or more!).
8.  The vertical profile inventory shows 52 Argo profiles, which is a good number.  Try to get at least this number.
9.  Click on DOWNLOAD just above where NetCDF Argo is the data choice.
10.  You'll see this form to fill out.  Supply the needed information, then click VALIDATE.
11.  This message will come immediately to your mail box.
12.  Here is the later message announcing your data file.  Notice that it is a TGZ zipped file.
13.  Use SAVE TARGET AS and navigate to the folder DATA > OCEAN > CORIOLIS and save the file with the existing filename, DataSelection_2012069_134041_44741.tgz
14.  Unzip the file in place.  Inspect the results with Windows Explorer.  You'll find a folder with the same name as the TGZ file, containing many individual NC profiles, plus an explanatory TXT file.
15.  Every attempt by the author to analyze/display these multi-profile data in IDV have failed (except the extremely limited exercise Visualizing Real-Time Temperature/Salinity Profiles in IDV:  GTSPP), even from alternate sources.  Current work to update the data sources and software platforms (including IDV) look very good.  It is likely that in the near future, direct connection between these data and IDV will be possible (via OPeNDAP?).  At present however it is necessary to go through a helper application, ODV, to make a necessary format change before IDV analysis is possible.
15.  Run ODV.
16.  In ODV select FILE > NEW to create a new collection.  Navigate to PRODUCTS > ODV > COLLECTIONS > PROFILES > CORIOLIS and enter the collection name profiles_20120615_20120629_liberia_coriolis
17.  For the collection variables, select ARGO PROFILE
18.  This map opens, showing no data yet loaded.
19.  Select IMPORT > ARGO FORMATS > FLOAT PROFILES.
20.  Navigate to the data folder DATA > OCEAN > CORIOLIS > DataSelection_2012069_134041_44741 where the NetCDF files were unzipped.
21.  Block all the NC files.  Then click OPEN.
22.  This window opens, showing all the incoming data are recognized and assigned to variables in the empty collection.

Click OK.

23.  This message indicates the size of the new collection.

The author does not know why it doesn't match with the CORIOLIS inventory above.

Click OK.

24.  This station map appears.  You could use AUTO-ZOOM a few times to see the stations better.
25.  The Argo variables include PRESSURE, but not DEPTH.  Select VIEW > DERIVED VARIABLES.
26.  Select DEPTH FROM PRESSURE [M] and click ADD to move it to the left side.  Then click OK.

This adds the depth variable to the collection, automatically making the necessary calculations.

27.  IDV cannot handle most filtering and data management, as we can do with ODV.  So we'll have to do a little preparation before we can send the data on.
28.  Right-click on the map, and select SAMPLE SELECTION CRITERIA.
29.  Select QUALITY and then any variable.  For that variable, select GOOD DATA, and then APPLY TO ALL VARIABLES.

Then click OK.

30.  If you're familiar with ODV, then you can view any scatter plot to see the data.  Here is temperature versus depth, a typical scatter plot.
31.  But we need to take the criteria selection process a little further.  Because IDV cannot discriminate between data values from different depths, we must export separate data subsets from ODV for desired depths.

Back in the SAMPLE SELECTION CRITERIA control, select RANGE.

Then select VARIABLE = DEPTH FROM PRESSURE [M].

And set the ACCEPTABLE RANGE to be from 0 to 5.

Then click OK.

32.  You now have only data in that interval.
33.  Select EXPORT > ODV SPREADSHEET FILE
34.  Navigate to PRODUCTS > ODV > SUBSETS and enter the filename data_from_profiles_20120615_20120629_0m_5m_liberia_coriolis.txt
35.  Here you are asked which variables to export.  We know from earlier experiments that the "ADJUSTED" values of the variables include all the values, original or appropriately changed.  So we select these 3, plus the depth.

You can add any others you want; it's up to you.

Then click OK.

36.  Make these selections, then click OK.
37.  This message indicates success.
38.  You can examine the exported file with Excel to see its contents.  Notice that the first 10 lines are metadata, the 11th line is column labels.  The data begins on line 12.

Also, the data fields are separated by TABS.

39.  Also export the data between 350-450m depth to the file data_from_profiles_20120615_20120629_350m_450m_liberia_coriolis.txt
40.  Now run IDV.
42.  On the dashboard, select DATA CHOOSERS > GENERAL > FILES.

Set DATA SOURCE TYPE = TEXT POINT DATA FILES.

 

43.  Navigate to the folder PRODUCTS > ODV > SUBSETS, and select the spreadsheet data_from_profiles_20120615_20120629_0m_5m_liberia_coriolis.txt and click ADD SOURCE
44.  This page opens, allowing us to parse the spreadsheet, line by line and item by item.
45.  Change the DELIMITER to TAB, and then scroll downward with the DOWN ARROW control on the left.

 

46.  When you get to the 11th row, the data begin to be read by IDV.  On this row you have only labels, but it's a good place to get the data field names.

Identify these 5 variables, as you see here, add their NAMES, UNITS/FORMATS and MISSING VALUES. 

TIME (the field just above LONGITUDE) is usually required for IDV, but it can be ignored here.  If you do use TIME, then use this UNIT/DATE FORMAT:

yyy-mm-dd hh:mm:ss.sss

This has been found to work well with the ODV export.

47.  After you add these items, remember to scroll one more row downward to the first data line.
48.  To avoid all this work in the future, click on PREFERENCES to save all these settings for this particular type of incoming data.  You can create any name you want for the settings; here the author has chosen "CORIOLIS>ODV"
49.  Now, finally, you can click OK at the bottom to import the spreadsheet.
50.  Now IDV loads the spreadsheet, and you see these choices.

Double-click on POINT DATA PLOT.

51.  Here is a plot of the data points in IDV.
52.  Compared to the other functions that IDV performs well, the gridding function we're about to use is not as good as typical GIS grid utilities.  It is not very flexible, and has built-in default settings that can only partially be improved.  For training purposes, and until the situation has been improved, we'll go ahead and use it.
53.  Make these selections to begin gridding the temperature.
54.  On the GRID PARAMETERS tab, uncheck the USE DEFAULT box.

You can read about the gridding in the IDV help text, but at this time, please make these changes.  We invite you to try to find a combination of settings that is better than what you see below.

Click on CREATE DISPLAY.

55.  This initial view of the grid shows that not all of the stations have been included.  Apparently IDV has an internal algorithm that sets the grid domain, and it is only a partial horizontal band through the middle of the point array.
56.  To improve the coloration, you can select DISPLAYS > COLOR TABLE > TEMPERATURE and select RADAR > DbZ to get a usual rainbow palette.
57.  You can also select DbZ > CHANGE RANGE
58.  The default values are -90 to 45, which is completely too large.

You can replace these with values you have selected for this area, based on your knowledge of regional climate.

Or, you can simply select USE PREDEFINED > FROM ALL DATA.

59.  This shows us that the current range is 24.6-28.2, which will be mapped to the rainbow palette.

Click OK if this is what you want.

60.  Now we have a more readable figure.
61.  Here the author has moved out from the gridded area to the usual Liberia area of interest to show you how much of the data has not been gridded.

Many attempts and experiments have not yielded any solution to this shortcoming in IDV.

62.  Here is the same figure, with the color legend on the right.

A comparison with the surface data from the World Ocean Atlas shows almost exactly the same range of temperature values, but they are distributed spatially different.  This is not surprising, since synoptic data can be different from climatologies, due to meso-scale motion.

 

63.  We know that many errors can occur in surface measurements.  So to avoid noise in the data, here is the parallel grid of the data from 350-450 m, or a "400-m grid".
64.  For comparison purposes, here is the graphic of the same area from the World Ocean Atlas.

The range of temperature values here is almost exactly the same as the graphic above in Panel 63.  Once again the spatial distribution, however, is different.  So meso-scale motions are probably still important even at this depth.

65.  You can use FILE > SAVE AS and navigate to the folder PRODUCTS > IDV to save this analysis as data_from_profiles_20120615_20120629_350m_450m_liberia_coriolis_txt.xidv for quick recall later on.