ESO754

Homework #2

Please download the following data:

· Monthly Sea Surface Temperature (SST) indices from

http://www.cpc.ncep.noaa.gov/data/indices/sstoi.indices (for information on the data, see http://www.cpc.ncep.noaa.gov/data/indices/)

and perform the following actions (for time period from January 1950 to December 2006 only):

 
1. Display NINO3.4 SST as time series. (The plot should have labels for the time and values. The temporal resolution should be high enough so that one can tell the corresponding values for any given month.) (5 points)

2. Compute NINO3.4 SST climatological values (averages for specific time stamps, say, average over all January, …, December) for each month and plot the climatological result (i.e., average values of NINO3.4 SST for the 12 months). (5 points)

3. Compute the SST Anomaly (SSTA) by subtracting the climatological values from the corresponding original data values and display the computed SSTA data again overlaid with the SSTA from the original file (the ANOM values in Column 10) as in Step 1. (5 points)

4. Plot scatter plot by using the anomaly data you computed from the previous steps and those given in the original data file. (5 points)

5. (optional) Use STL procedure to decompose NINO3.4 SST data. After obtaining the trend, seasonal, and remainder components, add the trend and seasonal components together to get another version of NINO3.4 SSTA. Repeat Step 4 with this anomaly and the anomaly given in the data set. (10 points)

[Note: In the following questions, you should use the anomaly values from the original data file.]

6. Compute the Running Sum and display the result. (5 points)

7. Do you think this time series is stationary? (You may discuss this based on only the running sum results and the time series plot.) (5 points)

8. Use Fourier Transformation to compute the spectrum of the data. Plot the spectrum. Identify the first two dominant components in the spectral space. What the periods represented by the two dominant components? (10 points)

9. Use wavelet analysis to get the time-frequency (period)-energy information. Display the result and discuss it. (Suggested discussion topics: What new information can you find in wavelet analysis which is not in the Fourier analysis?) (10 points)

Submission:

Please submit your results by email to yang@yang.gmu.edu (in one file if you have ZIP feature). The file name should start with yourLastName_firstName (failure to include identifier for your electronic submission may result in point loss). Your submission should have one file for all answers and your programs. If you used a GUI tool without programming, you should give a short description of the tool (if not a common tool) and your major steps to get the results. Programs and tool descriptions may be shared with other students after the due day. If you have any concern on the distribution, please let the instructor know.

Post Date: March 19, 2008

Due Date: April 2, 2008

Grade: 50+10 points total and 15% contribution to the final score.