| |
 |
|
|
Science Forum Index » Environment Forum » Auditing the Auditors #12
Page 1 of 1
|
| Author |
Message |
| David Ball |
 Posted: Wed Dec 17, 2003 12:09 pm |
|
|
|
Guest
|
I had a look at M&M's number 12....
Quote: 12. Did you commence your calculation of principal components
before the period in which all dataset members were available for the
following series: #72-80, #84-90? If so, please describe your
methodology for carrying out these calculations in the presence of
missing data and your justification for doing so?
If you look at the fortran code at:
ftp://holocene.evsc.virginia.edu/pub/MBH98/TREE/STAHLE/SWM/pca-stahleswm.f
you will find a block of text where the complex matrix is created
ipc = 10
im1 = iproxmin --> COMMAND LINE INPUT, PRESUMABLY 1700 FOR
SERIES 72-80
im2 = iproxmax --> 1980
N0 = iproxmax-iproxmin+1 --> 281
jj = 0
do j=1,nproxy --> NPROXY IS THE NUMBER OF DATAFILES BEING
READ IN THIS CASE 22
if (istart(j).le.iproxmin) then --> IF THE START YEAR
OF EACH FILE IS < 1700 CONTINUE => DO NOT CONSIDER ANY SERIES THAT
DOES NOT START BEFORE 1700.
jj = jj+1
i = 0
do i2=im1,im2 --> i2 GOES FROM 1700 to 1980
i = i+1
iyear = i2-nlow+1 --> INDEX IN THE INPUT ARRAY TO FIND
THE DATA
adum = aprox(iyear,j) --> DATA VALUE
AA(i,jj)=dcmplx(adum,zero) --> CREATE THE COMPLEX
ARRAY WITH ZERO AS THE COMPLEX PART
end do
end if
end do
M0 = jj
NU0 = M0
NV0 = M0
In short, the principle components are determined only for those
series spanning the period input from 1700 to 1980. If the start data
for the series starts after 1700, it is not put into the matrix fed to
the SVD subroutine.
Exactly the same block appears in the program:
ftp://holocene.evsc.virginia.edu/pub/MBH98/TREE/ITRDB/NOAMER/pca-noamer.f
used to determine the PC's for series 84-90.
In short, the calculation of the PC's starts at 1700 and 1750
respectively and does not consider any of the series that start after
that point. |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Sun Jul 06, 2008 10:28 pm
|
|
