Skip to content
GitLab
Menu
Projects
Groups
Snippets
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
mpim-sw
cdo
Commits
9eaaeada
Commit
9eaaeada
authored
Jan 05, 2011
by
Uwe Schulzweida
Browse files
eof docu update
parent
dc7dea0e
Changes
6
Hide whitespace changes
Inline
Side-by-side
OPERATORS
View file @
9eaaeada
...
...
@@ -208,6 +208,7 @@ Operator catalog:
Ensstat2 enscrps Cumulative Ranked Probability score
Ensstat2 ensrkhistspace Ranked Histogram averaged over time
Ensstat2 ensrkhisttime Ranked Histogram averaged over space
Ensstat2 ensroc Ensemble Receiver Operating characteristics
Fldstat fldmin Field minimum
Fldstat fldmax Field maximum
Fldstat fldsum Field sum
...
...
config/default
View file @
9eaaeada
...
...
@@ -171,6 +171,7 @@ case "${HOSTNAME}" in
# powerpc-ibm-aix5.3.0.0
blizzard
*
)
${
CONFPATH
}
configure
--prefix
=
$HOME
/local
\
--with-jasper
=
/sw/aix53/jasper-1.900.1
\
--with-grib_api
=
/sw/aix61/ECMWF
\
--with-netcdf
=
/sw/aix61/netcdf-4.1.1-rc2
\
--with-hdf5
=
/sw/aix61/hdf5-1.8.4-patch1-threadsafe
\
...
...
doc/tex/bib.tex
View file @
9eaaeada
...
...
@@ -82,6 +82,10 @@
{
Max Planck Institute for Meteorologie
}
\bibitem
[Peisendorfer]
{
Peisendorfer
}
\ \\
Rudolph W. Peisendorfer: Principal Component Analysis, Elsevier (1988)
\bibitem
[PROJ.4]
{
PROJ.4
}
\ \\
\href
{
http://trac.osgeo.org/proj/
}{
Cartographic Projections Library
}
,
originally written by Gerald Evenden then of the USGS.
...
...
@@ -96,4 +100,8 @@
\href
{
http://www.hdfgroup.org/doc
_
resource/SZIP
}{
Szip compression software
}
,
developed at University of New Mexico.
\bibitem
[vonStorch]
{
vonStorch
}
\ \\
Hans von Storch, Walter Zwiers: Statistical Analysis in Climate Research, Cambridge University Press (1999)
\end{thebibliography}
doc/tex/mod/EOFs
View file @
9eaaeada
...
...
@@ -29,9 +29,9 @@ Missing values are not fully supported. Support is only checked for non-changing
masks of missing values in time. Although there still will be results, they are
not trustworthy, and a warning will occur. In the latter case we suggest to
replace missing values by 0 in @file{ifile}.
@EndModule
@BeginOperator_eof
@Title = Calculate EOFs in spatial or time space
@Parameter = neof
...
...
@@ -59,11 +59,13 @@ replace missing values by 0 in @file{ifile}.
@EndOperator
@EndDescription
@BeginParameter neof
@Item = neof
INTEGER Number of eigen functions
@EndParameter
@BeginExample
To calculate the first 40 EOFs of a data-set containing anomalies use:
@BeginVerbatim
...
...
@@ -78,6 +80,7 @@ and use:
@EndVerbatim
@EndExample
@BeginEnvironment
@Item = CDO_SVD_MODE
Is used to choose the algorithm for eigenvalue calculation. Options are 'jacobi' for
...
...
@@ -97,4 +100,3 @@ a sufficient level of convergence. If all annihilation-pairs of vectors have
a norm below this value, the computation is considered to have converged
properly. Otherwise, a warning will occur. The default value 1e-05.
@EndEnvironment
doc/tex/ref_intro_eofs.tex
View file @
9eaaeada
This section contains modules to compute Empirical Orthogonal Functions and
- once they are computed - their principal coefficients.
An introduction to the theory
of principal component analysis as applied here can be found in:
\\
~~Rudolph W. Peisendorfer:
Principal Component Analysis
, Elsevier (1988).
\\
- once they are computed - their principal coefficients.
\\
An introduction to the theory
of principal component analysis as applied here can be found in:
\\
\hspace*
{
1cm
}
Principal Component Analysis
\cite
{
Peisendorfer
}
\\
Details about calculation in the time- and spatial spaces are found in:
\\
~~Hans von Storch, Walter Zwiers: Statistical Analysis in Climate Research,
Cambridge University Press (1999).
\hspace*
{
1cm
}
Statistical Analysis in Climate Research
\cite
{
vonStorch
}
EOFs are defined as the eigen values of the scatter matrix (covariance matrix) of
the data. For the sake of simplicity, samples are regarded as
{
\bf
time series of anomalies
}
...
...
src/operator_help.h
View file @
9eaaeada
...
...
@@ -1353,7 +1353,7 @@ static char *EnsstatHelp[] = {
static
char
*
Ensstat2Help
[]
=
{
"NAME"
,
" ensbrs, enscrps, ensrkhistspace, ensrkhisttime - "
,
" ensbrs, enscrps, ensrkhistspace, ensrkhisttime
, ensroc
- "
,
" Statistical values over an ensemble"
,
""
,
"SYNOPSIS"
,
...
...
@@ -1362,7 +1362,7 @@ static char *Ensstat2Help[] = {
"DESCRIPTION"
,
" This module computes statistical values over the ensemble of ensfiles using"
,
" obsfile as a reference. Depending on the chosen operator the Brier score, "
,
" Cumulative ranked probability score
or a
ranked Histogram over all Ensembles ensfiles"
,
" Cumulative ranked probability score ranked Histogram
or a roc-curve
over all Ensembles ensfiles"
,
" with reference to obsfile is written to ofile. "
,
" The date and grid information of a time step in ofile is the date of the "
,
" first input file. Thus all input files are required to have the same structure in "
,
...
...
@@ -1403,6 +1403,7 @@ static char *Ensstat2Help[] = {
" o(t,x) = crps( {i1(t,x)}, {i2(t,x),...,in(t,x)} )"
,
" ensrkhistspace Ranked Histogram averaged over time"
,
" ensrkhisttime Ranked Histogram averaged over space"
,
" ensroc Ensemble Receiver Operating characteristics"
,
NULL
};
...
...
@@ -2747,24 +2748,6 @@ static char *EOFsHelp[] = {
" masks of missing values in time. Although there still will be results, they are"
,
" not trustworthy, and a warning will occur. In the latter case we suggest to "
,
" replace missing values by 0 in ifile. "
,
" "
,
" Environmental settings are available to customize the way the eigenvalue computation"
,
" is done:"
,
" CDO_SVD_MODE can be used to set the algorithm which is used to calculate the"
,
" eigenvalues. Available Options are "
,
" (1) 'jacobi' for a one-sided parallel jacobi-algorithm (only executed in parallel if -P "
,
" flag is set."
,
" (2) 'danielson_lanczos' for a non-parallel d/l algorithm."
,
" Default: 'jacobi'"
,
" "
,
" MAX_JACOBI_ITER Is the maximum integer number of annihilation sweeps that is executed"
,
" if the jacobi-algorithm is used to compute the eigen values. Default: 5"
,
" "
,
" FNORM_PRECISION Is the Frobenius norm of the matrix consisting of an annihilation pair"
,
" of eigenvectors that is used to determine if the eigenvectors have reached a sufficient level"
,
" of convergence. If all annihilation-pairs of vectors have a norm below this value, the computation"
,
" is considered to have converged properly. Otherwise, a warning will occur. Default: 1e-05 "
,
" "
,
""
,
"OPERATORS"
,
" eof Calculate EOFs in spatial or time space"
,
...
...
@@ -2774,6 +2757,25 @@ static char *EOFsHelp[] = {
""
,
"PARAMETER"
,
" neof INTEGER Number of eigen functions"
,
""
,
"ENVIRONMENT"
,
" CDO_SVD_MODE "
,
" Is used to choose the algorithm for eigenvalue calculation. Options are 'jacobi' for "
,
" a one-sided parallel jacobi-algorithm (only executed in parallel if -P "
,
" flag is set) and 'danielson_lanczos' for a non-parallel d/l algorithm."
,
" The default setting is 'jacobi'."
,
" "
,
" MAX_JACOBI_ITER"
,
" Is the maximum integer number of annihilation sweeps that is executed"
,
" if the jacobi-algorithm is used to compute the eigen values. "
,
" The default value is 5."
,
" "
,
" FNORM_PRECISION"
,
" Is the Frobenius norm of the matrix consisting of an annihilation pair"
,
" of eigenvectors that is used to determine if the eigenvectors have reached "
,
" a sufficient level of convergence. If all annihilation-pairs of vectors have "
,
" a norm below this value, the computation is considered to have converged "
,
" properly. Otherwise, a warning will occur. The default value 1e-05."
,
NULL
};
...
...
Write
Preview
Supports
Markdown
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment