-
Notifications
You must be signed in to change notification settings - Fork 22
/
dineof.init
122 lines (81 loc) · 3.38 KB
/
dineof.init
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
!
! INPUT File for dineof 3.0
!
! Lines starting with a ! or # are comments
!
! gappy data to fill by DINEOF. For several matrices, separate names with commas
! Examples:
! data = ['seacoos2005.avhrr','seacoos2005.chl']
! data = ['2Dbelcolour_region_period_anomaly.gher']
data = ['2Dbelcolour_region_period_dat.gher']
!--------------MASK FILE----------------------------------------------
! Land-sea mask of the gappy data.
! Several masks can be especified, separated by commas:
! Example :
! mask = ['seacoos2005.avhrr.mask','seacoos2005.chl.mask']
!
! When no mask is especified, no land points are present in the initial file
mask = ['seacoos2005.avhrr.mask']
!---------------------------------------------------------------------
! time vector (necessary if B_DIFF option is activated in ppdef.h
! necessary files:
! time: time vector, indicating the increment between time steps in data file (must have same time dimension as data file)
! alpha: parameter specifying the strength of the filter
! numit: number of iterations for the filter
! See http:https://www.ocean-sci.net/5/475/2009/os-5-475-2009.pdf for more information
time = 'dates.dat'
alpha = 0.01
numit = 3
!
! Sets the numerical variables for the computation of the required
! singular values and associated modes.
!
! Give 'nev' the maximum of number of modes you allow to compute,
! Dineof will stop after computing the optimal number of modes + 3 extra modes
! for robust optimum detection.
nev = 100
! Give 'neini' the minimum number of modes you want to compute
neini = 1
! Give 'ncv' the maximal size for the Krylov subspace
! (Do not change it as soon as ncv > nev+5)
! ncv must also be smaller than the temporal size of your matrix
ncv = 106
! Give 'tol' the treshold for Lanczos convergence
! By default 1.e-8 is quite reasonable
tol = 1.0e-8
! Parameter 'nitemax' defining the maximum number of iteration allowed for the stabilisation of eofs obtained by the cycle ((eof decomposition <-> truncated reconstruction and replacement of missing data)). An automatic criteria is defined by the following parameter 'itstop' to go faster
nitemax = 300
! Parameter 'toliter' is a precision criteria defining the threshold of automatic stopping of dineof iterations, once the ratio (rms of successive missing data reconstruction)/stdv(existing data) becomes lower then 'toliter'.
toliter = 1.0e-3
! Parameter 'rec' for complete reconstruction of the matrix
! rec=1 will reconstruct all points; rec=0 only missing points
rec = 1
! Parameter 'eof' for writing the left and right modes of the
!input matrix. Disabled by default. To activate, set to 1
eof = 1
! Parameter 'norm' to activate the normalisation of the input matrix
!for multivariate case. Disabled by default. To activate, set to 1
norm = 0
! Output folder. Left and Right EOFs will be written here
!
!Output = './'
Output = 'Output/'
!
! user chosen cross-validation points,
! remove or comment-out the following entry if the cross-validation points are to be chosen
! internally
!
! clouds = 'crossvalidation.clouds'
!
! "results" contains the filenames of the filled data
!
!results = ['All_95_1of2.sst.filled']
results = ['Output/F2Dbelcolour_region_period_datfilled.gher']
EOF.U = ['eof.nc#Usst','eof.nc#Uchl']
EOF.V = 'eof.nc#V'
EOF.Sigma = 'eof.nc#Sigma'
! seed to initialize the random number generator
seed = 243435
!
! END OF PARAMETER FILE
!