| preferAdjusted {oce} | R Documentation |
Set Preference for Adjusted Values
Description
argo data can contain "adjusted" forms of data items,
which may be more trustworthy than the original
data, and preferAdjusted lets the user express a preference
for such adjusted data. This means that using
[[,argo-method on the results returned by preferAdjusted
will (if possible) return adjusted data, and also use those adjusted
data in computations of derived quantities such as Absolute Salinity.
The preference applies also to units and to data-quality flags,
both of which can be returned by [[,argo-method, as
discussed in “Details”.
Usage
preferAdjusted(argo, which = "all", fallback = TRUE)
Arguments
argo |
An argo object. |
which |
A character vector naming the items for which
(depending also on the value of |
fallback |
A logical value indicating whether to fall back
to unadjusted values for any data field in which the
adjusted values are all |
Details
preferAdjusted() merely sets two items in the metadata slot of the
returned argo object. The real action is carried out by
[[,argo-method but, for convenience, the details are explained here.
Consider salinity, for example.
If which equals "all", or if it is a character
vector containing "salinity", then using
[[,argo-method on the returned object
will yield the adjusted forms of the salinity data,
its associated flags, or its units. Thus, in the salinity
case,
-
argo[["salinity"]]will attempt to returnargo@data$salinityAdjustedinstead of returningargo@data$salinity, although if the adjusted values are allNAthen, depending on the value offallback, the unadjusted values may be returned; similarly -
argo[["salinityFlags"]]will attempt to returnargo@metadata$flags$salinityAdjustedinstead ofargo@metadata$flags$salinity, and -
argo[["salinityUnits"]]will attempt to returnargo@metadata$units$salinityAdjustedinstead ofargo@metadata$units$salinity.
The default value, which="all", indicates that this
preference for adjusted values will apply to all the
elements of the data slot of the returned object, along
with associated flags and units. This can be handy for quick
work, but analysts may also choose to restrict their use of
adjusted values to a subset of variables, based on their own
decisions about data quality or accuracy.
The default value fallback=TRUE indicates that later calls to
[[,argo-method should return unadjusted values for any
data items that have NA for all the adjusted values. This
condition is rare for core variables (salinity, temperature and
pressure) but is annoyingly common for biogeochemical variables; see
e.g. Section 2.2.5 of Reference 1 for a discussion of
the conditions under which Argo netcdf files contain
adjusted values. Setting fallback=FALSE means that adjusted
values (if they exist) will always be returned, even if they
are a useless collection of NA values.
Error fields, such as salinityAdjustedError, are returned
as-is by [[,argo-method, regardless of whether
the object was created by preferAdjusted.
It should be noted that, regardless of whether preferAdjusted
has been used, the analyst can always access either unadjusted
or adjusted data directly, using the original variable names stored
in the source netcdf file. For example, argo[["PSAL"]]
yields unadjusted salinity values, and
argo[["PSAL_ADJUSTED"]] yields adjusted values (if they exist, or
NULL if they do not).
Similarly, adjusted value can always be obtained by using a form
like argo[["salinityAdjusted"]].
Value
An argo object its metadata slot altered
(in its adjustedWhich and adjustedFallback elements)
as a signal for how [[,argo-method should
function on the object.
Author(s)
Dan Kelley, based on discussions with Jaimie Harbin (with
respect to the [[,argo-method interface) and Clark Richards
(with respect to storing the preference in the metadata slot).
References
Argo Data Management Team. "Argo User's Manual V3.3." Ifremer, November 28, 2019. doi:10.13155/29825
Examples
library(oce)
data(argo)
argoAdjusted <- preferAdjusted(argo)
all.equal(argo[["salinityAdjusted"]], argoAdjusted[["salinity"]])
all.equal(argo[["salinityFlagsAdjusted"]], argoAdjusted[["salinityFlags"]])
all.equal(argo[["salinityUnitsAdjusted"]], argoAdjusted[["salinityUnits"]])