ATMO/GEOS 595c Patterns and Mechanisms of Decadal Climate Variability
Data Exercise 3: Construct and analyze a proxy-based index of the Indian Ocean Dipole
Exercise Questions
 Introduction
 Help
 Part I.  Proxy IOD index development
Part II.  Interpretation

Questions for discussion Monday, Feb 18th, as a result of this data analysis exercise:
Saji et al., Nature 401: 360-363 (1999), Fig. 3
Fig. 2 from Saji et al. 1999: A canonical dipole event (SST and wind anomalies, colors and bolded vectors significant effects) in the Indian Ocean sector. 
 
Introduction
In Data Exercise 1 and 2, you were introduced to timeseries filtering and to paleoclimate reconstructions.  In this Data Exercise we will develop a proxy index for tropical climate varaibility in the Indian Ocean observed and modeled by Saji et al. (1999), Webster et al. (1999) and many others.  Bring your answers to class for discussion Monday, March 10, 2008; you can also upload results to share into our course server directory.

Help 

Feel free to employ any computational environment you prefer.  I will give some direction on how to do this in matlab, which is currently sitelicensed at the University of Arizona (freeware Octave workalike is here).  A short introduction to operation, nuts and bolts coding, and displaying graphics in matlab from another course I offer is here; a rough set of computing principles is here.  See also the links in this assignment to further explanation of terminology and concepts. 

A good general reference for time series analysis is Chatfield, Analysis of Time Series: An Introduction (6th ed., 2003).

More introduction to coral paleoclimatology is here.  A nice report on the Annual Records in Tropical Systems (ARTS) Initiative is here.
Part I.  Develop a coral-based paleoproxy index of tropical Indian ocean-atmopshere variations.

Based on our discussion Monday March 3 of Indian Ocean Dipole observations (and see Fig. 2 from Saji et al. 1999, reproduced above), what would be the best locations and climatological variables to use to develop an index of the Indian Ocean Dipole (hereafter, IOD)?  Why?

Many proxy climate data are available via internet through the National Geophysical Data Center's (NCDC) World Data Center-A for Paleoclimatology, physically located in Boulder, CO USA, and virtually located here.  If you are not familar with this website, spend some time surfing around and learning about the various sources of proxy data (tree ring, coral, borehole, lake, model, historical, cave, ice core. ...).  You are welcome to proceed with any proxy data you deem appropriate.  A useful tool at the WDCA-P is the ArcIMS data mapper tool which you can access from any of the proxy data type web pages, for instance, in the middle right hand side of the page on lake sediments here.  

WDCA-P main page

Based on our discussion Monday March 3 of Indian Ocean Dipole observations (and see Fig. 2 from Saji et al. 1999, reproduced above; and see Charles et al. (2003) as well), why should coral paleoproxy data be useful for reconstruction of past variations in the tropical Indian Ocean?  Can you think of other proxy data that might also be useful?  What are the pros and cons of using these data and sampling sites?  Are there other considerations?

I think the coral records from Bali, Indonesia and Mahe, Seychelles might provide a simple IOD index.  (Why?)   If you agree with me, you can go to the WDCA-P coral and schlerosponge Data Search page (click on the coral picture from the main page, then click on "Coral Data Search Engine" to get there).  Select the Investigator "Charles, C.D." and the variable "delta18O".  Click Start search.  Click on the "Lombok Strait" choice to get the Bali dataset README file for important background information and the data; click on the "Mahe" choice for data and README files.

WDCA coral data search engine


According to Saji et al. (1999) and their Fig. 2 above, what season should give the least ambiguous IOD index at Bali?  How about for Mahe, Seychelles?   Form your proxy IOD for either this season or the annual average, based on the difference between the two coral data series (or your chosen proxy series).

Part II.  Interpretation.

Plot the correlation of your IOD index with the index found here for 1856-2007, based on the SST analysis of Kaplan et al. (1998).  (This index is SST anomaly for 50-70E, 10S-10N) minus SST anomaly for (90E-110E,10S-0N); I made it averaged for the May-October season).  For each data series, plot its power spectrum.  How do they compare?  Make a low pass filter and apply it to the data series.  How good is your proxy IOD index on interannual and decadal time scales?  Is there decadal variability in the IO Dipole?  How would you characterize it?

Now go back to the top of this page and prepare your thoughts on the answer to our discussion questions.


back to Syllabus/Schedule.  Last updated 3 Mar 2008.