API¶
Source code: https://bitbucket.org/kul-reseco/dineof/src.
For issues, bugs, proposals or remarks, visit the issue tracker.
-
class
dineof.core.dineof.Dineof(data_path: str, output_folder: str, eof_max: int, krylov_max: int, mask_path: str = None, eof_min: int = 1, eof_max_it: int = 300, eof_it_stop: float = 0.001, lanczos: float = 1e-08, time_path: str = None, time_alpha: float = 0.01, time_it: int = 3, reconstruct_all: int = 1, save_eof_base: int = 0, normalize_input: int = 0, cv_seed: int = 243435, cv_clouds: str = None)[source]¶ Bases:
objectEOF-based method to fill in missing data from geophysical fields, such as clouds in sea surface temperature.
Parameters: - data_path – path to netCDF data with gaps to be filled
- output_folder – path to output folder where all results will be stored
- eof_max – the max number of EOF modes to compute; only the optimal number of EOFs + 3 will be computed for robust optimum detection must be <= temporal size of data - 5; start high and adjust down for subsequent reconstructions; if the max number of modes have been used, the algorithm probably needs more information
- krylov_max – the maximum size for the Krylov subspace; must be: >= eof_max + 5 and <= temporal size of data
- mask_path – path to land-sea mask; netCDF format (optional) without mask, all points are considered as ocean
- eof_min – the min number of EOF modes to computes; default = 1 (recommended)
- eof_max_it – set the maximum iterations for each EOF calculation; default = 300; eof decomposition <-> truncated reconstruction and replacement of missing data; use large number and only increase if this number is reached for each EOF
- eof_it_stop – stop EOF iterations once: rms(successive reconstruction) / std(existing data) < eof_it_stop
- lanczos – threshold for Lanczos convergence; default = 1.e-8 (recommended)
- time_path – path to time file; netCDF format; used for the diffusion of the covariance matrix specified as the relative increment in time between your individual images; example: [1,2,3,4,…] if all data is 1 day apart; [1,2,4,…] if day 3 is missing
- time_alpha – sets strength of the filter; 0 to deactivate
- time_it – number of iterations for the filter
- reconstruct_all – 1: reconstruct all points; 0: reconstruct only missing data
- save_eof_base – 1: save left and right EOFs; 0: not
- normalize_input – 1: normalize the input matrix for multivariate case; 0: not
- cv_seed – initialize random number generator, for selection of cross-validation points; repeating an experiment with the same seed will yield the exact same result while another seed will yield a slight variation
- cv_clouds – path to cloud file; netCDF format; mask with manual selection of cross-validation points cv_seed will be ignored
For more information on how DINEOF works, please refer to:
Alvera-Azcárate, A., Barth, A., Rixen, M., Beckers, J.M., 2005, Reconstruction of incomplete oceanographic data sets using empirical orthogonal functions: application to the Adriatic Sea surface temperature, Ocean Modelling, 9-4, p. 325-346
Beckers, J.M., Rixen, M., 2003, EOF calculations and data filling from incomplete oceanographic data sets, Journal of Atmospheric and Oceanic Technology, 20-12, p. 1839-1856
The multivariate application of DINEOF is explained in:
Alvera-Azcárate, A., Barth, A., Beckers, J.M., Weisberg, R.H., 2007, Multivariate reconstruction of missing data in sea surface temperature, chlorophyll, and wind satellite fields, Journal of Geophysical Research. Oceans, 112-C3, p. C03008
The error calculation using an optimal interpolation approach is explained in:
Beckers, J.M., Barth, A., Alvera-Azcárate, A., 2006, DINEOF reconstruction of clouded images including error maps – application to the Sea-Surface Temperature around Corsican Island, Ocean Science, 2, 183–199
For more information about the Lanczos solver:
Toumazou, V., Cretaux, J.F., 2001, Using a Lanczos Eigensolver in the Computation of Empirical Orthogonal Functions, Monthly Weather Review, 129-5, p. 1243–1250