ensemble.loadGrid

Load a gridded variable from a saved state vector ensemble

Syntax



[X, metadata] = obj.loadGrid(variableName)


[X, metadata] = obj.loadGrid(v)


[X, metadata, members, evolvingLabels] = obj.loadGrid(variable)


… = obj.loadGrid(variable, labels)


… = obj.loadGrid(variable, e)


… = obj.loadGrid(variable, -1)

Description



[X, metadata] = obj.loadGrid(variableName)


[X, metadata] = obj.loadGrid(v)

Loads a variable from a saved ensemble into memory. Regrids the state vector for the variable into a gridded dataset. The regridded variable will be an N-dimensional array. The leading dimensions of the array will be the state dimensions of the variable. The next dimension holds the different ensemble members. Finally, the last dimension holds the values for different ensembles in an evolving set. If the ensemble object implements a static ensemble, then this final dimension will be a singleton.

The returned metadata is a gridMetadata object that describes the regridded state dimensions of the variable. This method does not remove singleton state dimensions, so the metadata will hold information for all state dimension. If the variable implements a mean over a state dimension, the metadata for that dimension will have one row per element used in the mean. The metadata does not hold information on the different ensemble members or different ensembles in an evolving set.

The specified variable must be a variable used by the ensemble object, and the method will only load used ensemble members. By default, ensemble objects use all variables and ensemble members saved in the .ens file. See the “useVariables”, “useMembers”, and “evolving” commands to use different subsets of variables and ensemble members.

The “ensembleMetadata.regrid” command also regrids state vector variables and provides additional options for regridding variables. For example, customizing the order of regridded dimensions and removing singleton dimensions from the regridded dataset. For a memory efficient way to apply the “regrid” command to a saved ensemble: 1. Use “ensemble.useVariable” to select only the variable of interest, 2. Use “ensemble.load” to load the variable and an associated ensembleMetadata object, then 3. Use the ensembleMetadata object and its “regrid” command to regrid the variable.



[X, metadata, members, evolvingLabels] = obj.loadGrid(variable)

Also returns the ensemble members and labels of any evolving ensembles in the regridded dataset.



= obj.loadGrid(variable, labels)


= obj.loadGrid(variable, e)


= obj.loadGrid(variable, -1)

Loads specific ensembles in the evolving set. If the second input is -1, selects all ensembles in the evolving set. The dimension of the regridded dataset that corresponds to evolving ensembles is the dimension following the ensemble members. The length of this evolving ensemble dimension will match the number of requested ensembles, and the elements of this dimension will follow the order in which the ensembles were requested.

Input Arguments

variableName

string scalar
The name of a variable being used by the ensemble object.

v

logical vector | scalar linear index
The index of a variable being used by the ensemble object. If a logical vector, must have a single true element.

e

-1 | vector, linear indices | logical vector
Indicates the ensembles to load. If -1, loads all ensembles in the evolving set. If a logical vector, must have one element per ensemble in the evolving set. Otherwise, use the linear indices of ensembles in the evolving set.

labels

string vector
The labels of specific ensembles in the evolving set. You can only use labels to reference ensembles that have unique labels. Use linear indices to reference ensembles that share the same label.

Output Arguments

X

numeric array [? x nMembers x nEvolving]
The regridded loaded variable. Has one dimension for each state dimension, followed by ensemble members, and then evolving ensembles.

metadata

scalar gridMetadata object
A gridMetadata object for the regridded variable. Has information for each state dimension of the regridded variable. Metadata fields are in the same order as the leading dimensions of the output X. If the state vector implements a mean along a state dimension, the metadata for the dimension has a row for each element used in the mean.

members

array, positive integers, [nMembers x nEvolving]
The ensemble members for the loaded ensembles. Each element of “members” lists the index of an ensemble member saved in the

.ens file. The columns of “members” list the ensemble members for different loaded evolving ensembles.

labels

string vector [nEvolving]
The labels of the loaded ensembles in the evolving set.