ST 445 -- SUMMER I, 2012

HOMEWORK #4 -- DUE WEDNESDAY 13 JUNE 2012

J. K. Angell compiled radiosonde measurements for many years to examine the 
patterns of temperature change in the atmosphere.  One compilation has
five regions of the earth (and associated files):

   north polar          npolar.txt
   north temperate      ntemp.txt
   tropical             trop.txt
   south temperate      stemp.txt
   south polar          spolar.txt
Measurements are in degrees C, deviations from climatological mean.

The first 3 files have a 24 line header; headers in the last two are 
26 lines long.  One year corresponds to a record, with four seasons 
(winter, spring, summer, autumn) for each year, with four levels of 
the atmosphere (surface, troposphere (850-300mb), tropopause (300-100mb),
lower stratosphere (100-50mb), delineated by pressure (mb=millibars).  
So on each record is the year, the four seasons and annual average for 
surface, four seasons and annual for troposphere, etc, finishing with 
four seasons and annual for lower stratosphere:
    year win spr sum aut ann win spr sum....sum aut ann
so a total for 1+4*(4+1) = 21 columns of data.

For further information see the site
  http://cdiac.ornl.gov/trends/temp/angell/angell.html

Read in the data from ONE of the five regions and perform the following tasks:
 a) Create a dataset for the SURFACE ONLY with each observation corresponding 
to a season/quarter, with variables YEAR, QTR (numeric)(quarter/season), and 
SURF.  Also create a time or date variable that properly indexes the data.
(You can use something simple here; we will cover date variables soon enough.)
 b) Plot surface temperature deviations versus time/date.
 c) Check whether given the annual averages match (close enough) the sample 
mean from the quarterly values.  (Note the given annual average may be 
missing for a year if any quarterly value is missing.)
 d) What quarter/season has the highest variance?
 e) Using a simple linear regression model with year and annual averages, 
estimate the annual rate of change (slope) (and give a standard error).  
(You may want to use a different dataset to do this part.)

 *Bonus* Do (a) and (b) for another level of the atmosphere (troposphere,
tropopause, stratosphere).

Clearly say which file you used.

FOR ALL EXERCISES: Hand in (paper) 
 (1) The answers to any questions asked. 
 (2) Your program * With clear dicumentation using comments * 
 (3) Your output