Plot an argoFloats Object

Source: R/plot.R

The action depends on the type of the object, and this is set up by the function that created the object; see “Details”. These are basic plot styles, with somewhat limited scope for customization. Since the data with argoFloats objects are easy to extract, users should not find it difficult to create their own plots to meet a particular aesthetic. See “Examples” and Kelley et al. (2021) for more plotting examples.

# S4 method for argoFloats
plot(
  x,
  which = "map",
  bathymetry = argoDefaultBathymetry(),
  geographical = 0,
  xlim = NULL,
  ylim = NULL,
  xlab = NULL,
  ylab = NULL,
  type = "p",
  cex = par("cex"),
  col = par("fg"),
  pch = par("pch"),
  bg = par("bg"),
  eos = "gsw",
  mapControl = NULL,
  profileControl = NULL,
  QCControl = NULL,
  summaryControl = NULL,
  TSControl = NULL,
  debug = 0,
  ...
)

Arguments

x

an argoFloats object.

which

a character value indicating the type of plot. The possible choices are "map", "profile", "QC", "summary" and "TS"; see “Details”.

bathymetry

an argument used only if which="map", to control whether (and how) to indicate water depth. Note that the default was TRUE until 2021-12-02, but was changed to FALSE on that date, to avoid a bathymetry download, which can be a slow operations. See “Details” for details, and Example 4 for a hint on compensating for the margin adjustment done if an image is used to show bathymetry.

geographical

flag indicating the style of axes for the which="map" case, but only if no projection is called for in the mapControl argument. With geographical=0 (which is the default), the axis ticks are labeled with signed longitudes and latitudes, measured in degrees. The signs are dropped with geographical=1. In the geographical=4 case, signs are also dropped, but hemispheres are indicated by writing S, N, W or E after axis tick labels, except at the equator and prime meridian. Note that this scheme mimics that used by oce::plot,coastline-method().

xlim, ylim

numerical values, each a two-element vector, that set the x and y limits of plot axes, as for plot.default() and other conventional plotting functions.

xlab

a character value indicating the name for the horizontal axis, or NULL, which indicates that this function should choose an appropriate name depending on the value of which. Note that xlab is not obeyed if which="TS", because altering that label can be confusing to the user.

ylab

as xlab, but for the vertical axis.

type

a character value that controls the line type, with "p" for unconnected points, "l" for line segments between undrawn points, etc.; see the docs for par(), If type not specified, it defaults to "p".

cex

a character expansion factor for plot symbols, or NULL, to get an value that depends on the value of which.

col

the colour to be used for plot symbols, or NULL, to get an value that depends on the value of which (see “Details”). If which="TS", then the TSControl argument takes precedence over col.

pch

an integer or character value indicating the type of plot symbol, or NULL, to get a value that depends on the value of which. (See par() for more on specifying pch.)

bg

the colour to be used for plot symbol interior, for pch values that distinguish between the interior of the symbol and the border, e.g. for pch=21.

eos

a character value indicating the equation of state to use if which="TS". This must be "gsw" (the default) or "unesco"; see oce::plotTS().

mapControl

a list that permits particular control of the which="map" case. If provided, it may contain elements named bathymetry (which has the same effect as the parameter bathymetry), colLand (which indicates the colour of the land), and projection (which may be FALSE, meaning to plot longitude and latitude on rectilinear axes, TRUE, meaning to plot with oce::mapPlot(), using Mollweide projection that is suitable mainly for world-scale views, or a character value that will be supplied to oce::mapPlot(). If a projection is used, then the positions of the Argo floats are plotted with oce::mapPoints(), rather than with points(), and if the user wishes to locate points with mouse clicks, then oce::mapLocator() must be used instead of locator(). If bathymetry is not contained in mapControl, it defaults to FALSE, and if projection is not supplied, it defaults to FALSE. Note that mapControl takes precedence over the bathymetry argument, if both are provided. Also note that, at present, bathymetry cannot be shown with map projections.

profileControl

a list that permits control of the which="profile" case. If provided, it may contain elements named parameter (a character value naming the quantity to plot on the x axis), ytype (a character value equal to either "pressure" or "sigma0") and connect (a logical value indicating whether to skip across NA values if the type argument is "l", "o", or "b"). If profileControl is not provided, it defaults to list(parameter="temperature", ytype="pressure", connect=FALSE). Alternatively, if profileControl is missing any of the three elements, then they are given defaults as in the previous sentence.

QCControl

a list that permits control of the which="QC" case. If provided, it may contain an element named parameter, a character value naming the quantity for which the quality-control information is to be plotted, and an element named dataStateIndicator, a logical value controlling whether to add a panel showing this quantity (see Reference Table 6 of reference 1 to learn more about what is encoded in dataStateIndicator). If not provided, QCControl defaults to list(parameter="temperature",dataStateIndicator=FALSE).

summaryControl

a list that permits control of the which="summary". If provided, it should contain an element named items, a character vector naming the items to be shown. The possible entries in this vector are "dataStateIndicator" (see Reference Table 6 of reference 1, for more information on this quantity), "length" (the number of levels in the profile), "deepest" (the highest pressure recorded), "longitude" and "latitude". If summaryControl is not provided, all of these will be shown. If all the elements of x have the same ID, then the top panel will have ticks on its top axis, indicating the cycle.

TSControl

a list that permits control of the which="TS" case, and is ignored for the other cases. If TSControl is supplied, then other arguments that normally control plots are ignored. TSControl may have elements named drawByCycle, cex, col, pch, lty, lwdandtype. The first of these is a logical value indicating whether to plot individual cycles with different aesthetic characteristics, and the other elements describe those characteristics in the usual way. By default, typeis"p", and the point characteristics are given by the other parameters, which are passed to [rep()] to have one value per cycle. For example, TSControl=list(drawByCycle=TRUE,col=1:2,type="l")` indicates to draw lines for the individual cycles, with the first being black, the second red, the third black, etc. FIXME: THIS IS NOT CODED YET.

debug

an integer specifying the level of debugging.

...

extra arguments passed to the plot calls that are made within this function.

Value

None (invisible NULL).

Details

The various plot types are as follows.

  • For which="map", a map of profile locations is created if subtype is equal to cycles, or a rectangle highlighting the trajectory of a float ID is created when subtype is equal to trajectories. This only works if the type is "index" (meaning that x was created by getIndex() or a subset of such an object, created with subset,argoFloats-method()), or argos (meaning that x was created with readProfiles(). The plot range is auto-selected. If the ocedata package is available, then its coastlineWorldFine dataset is used to draw a coastline (which will be visible only if the plot region is large enough); otherwise, if the oce package is available, then its coastlineWorld dataset is used. The bathymetry argument controls whether (and how) to draw a map underlay that shows water depth. There are three possible values for bathymetry:

    1. FALSE (the default, via argoDefaultBathymetry()), meaning not to draw bathymetry;

    2. TRUE, meaning to draw bathymetry as an image, using data downloaded with oce::download.topo(). Example 4 illustrates this, and also shows how to adjust the margins after the plot, in case there is a need to add extra graphical elements using points(), lines(), etc.

    3. A list with items controlling both the bathymetry data and its representation in the plot (see Example 4). Those items are:

      1. source, a mandatory value that is either (a) the string "auto" (the default) to use oce::download.topo() to download the data or (b) a value returned by oce::read.topo().

      2. contour, an optional logical value (with FALSE as the default) indicating whether to represent bathymetry with contours (with depths of 100m, 200m, 500m shown, along with 1km, 2km up to 10km), as opposed to an image;

      3. colormap, ignored if contour is TRUE, an optional value that is either the string "auto" (the default) for a form of GEBCO colors computed with oce::oceColorsGebco(), or a value computed with oce::colormap() applied to the bathymetry data; and

      4. palette, ignored if contour is TRUE, an optional logical value (with TRUE as the default) indicating whether to draw a depth-color palette to the right of the plot. Note that the default value for bathymetry is set by a call to argoDefaultBathymetry(), and that this second function can only handle possibilities 1 and 2 above.

  • For which="profile", a profile plot is created, showing the variation of some quantity with pressure or potential density anomaly, as specified by the profileControl argument.

  • For which="QC", two time-series panels are shown, with time being that recorded in the individual profile in the dataset. An additional argument named parameter must be given, to name the quantity of interest. The function only works if x is an argoFloats object created with readProfiles(). The top panel shows the percent of data flagged with codes 1 (meaning good data), 2 (probably good), 5 (changed) or 8 (estimated), as a function of time (lower axis) and (if all cycles are from a single Argo float) cycle number (upper axis, with smaller font). Thus, low values on the top panel reveal profiles that are questionable. Note that if all of data at a given time have flag 0, meaning not assessed, then a quality of 0 is plotted at that time. The bottom panel shows the mean value of the parameter in question regardless of the flag value.

  • For which="summary", one or more time-series panels are shown in a vertical stack. If there is only one ID in x, then the cycle values are indicated along the top axis of the top panel. The choice of panels is set by the summaryControl argument.

  • For which="TS", a TS plot is created, by calling plotArgoTS() with the specified x, xlim, ylim, xlab, ylab, type, cex, col, pch, bg, eos, and TSControl, along with debug-1. See the help for plotArgoTS() for how these parameters are interpreted.

References

  1. Carval, Thierry, Bob Keeley, Yasushi Takatsuki, Takashi Yoshida, Stephen Loch, Claudia Schmid, and Roger Goldsmith. Argo User's Manual V3.3. Ifremer, 2019. doi:10.13155/29825

  2. Kelley, D. E., Harbin, J., & Richards, C. (2021). argoFloats: An R package for analyzing Argo data. Frontiers in Marine Science, (8), 636922. doi:10.3389/fmars.2021.635922

Author

Dan Kelley and Jaimie Harbin

Examples

# Example 1: map profiles in index
library(argoFloats)
data(index)
plot(index)


# Example 2: as Example 1, but narrow the margins and highlight floats
# within a circular region of diameter 100 km.
oldpar <- par(no.readonly=TRUE)
par(mar=c(2, 2, 1, 1), mgp=c(2, 0.7, 0))
plot(index)
lon <- index[["longitude"]]
lat <- index[["latitude"]]
near <- oce::geodDist(lon, lat, -77.06, 26.54) < 100
R <- subset(index, near)
#> Kept 315 cycles (31.2%)
points(R[["longitude"]], R[["latitude"]],
    cex=0.6, pch=20, col="red")

par(oldpar)

# Example 3: TS of a built-in data file.
f <- system.file("extdata", "SR2902204_131.nc", package="argoFloats")
a <- readProfiles(f)
#> Warning: Of 1 profiles read, 1 has >10% of BBP700 values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of chlorophyllA values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of oxygen values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of pressure values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
oldpar <- par(no.readonly=TRUE)
par(mar=c(3.3, 3.3, 1, 1), mgp=c(2, 0.7, 0)) # wide margins for axis names
plot(a, which="TS")

#> NULL
par(oldpar)
# \donttest{
# Example 4: map with bathymetry. Note that par(mar) is adjusted
# for the bathymetry palette, so it must be adjusted again after
# the plot(), in order to add new plot elements.
# This example specifies a coarse bathymetry dataset that is provided
# by the 'oce' package.  In typical applications, the user will use
# a finer-scale dataset, either by using bathymetry=TRUE (which
# downloads a file appropriate for the plot view), or by using
# an already-downloaded file.
data(topoWorld, package="oce")
oldpar <- par(no.readonly=TRUE)
par(mar=c(2, 2, 1, 2), mgp=c(2, 0.7, 0)) # narrow margins for a map
plot(index, bathymetry=list(source=topoWorld))
# For bathymetry plots that use images, plot() temporarily
# adds 2.75 to par("mar")[4] so the same must be done, in order
# to correctly position additional points (shown as black rings).
par(mar=par("mar") + c(0, 0, 0, 2.75))
points(index[["longitude"]], index[["latitude"]],
    cex=0.6, pch=20, col="red")

par(oldpar)

# Example 5. Simple contour version, using coarse dataset (ok on basin-scale).
# Hint: use oce::download.topo() to download high-resolution datasets to
# use instead of topoWorld.
oldpar <- par(no.readonly=TRUE)
par(mar=c(2, 2, 1, 1))
data(topoWorld, package="oce")
plot(index, bathymetry=list(source=topoWorld, contour=TRUE))

par(oldpar)

# Example 6. Customized map, sidestepping plot,argoFloats-method().
lon <- topoWorld[["longitude"]]
lat <- topoWorld[["latitude"]]
asp <- 1/cos(pi/180*mean(lat))
# Limit plot region to float region.
xlim <- range(index[["longitude"]])
ylim <- range(index[["latitude"]])
# Colourize 1km, 2km, etc, isobaths.
contour(x=lon, y=lat, z=topoWorld[["z"]], xlab="", ylab="",
    xlim=xlim, ylim=ylim, asp=asp,
    col=1:6, lwd=2, levels=-1000*1:6, drawlabels=FALSE)
# Show land
data(coastlineWorldFine, package="ocedata")
polygon(coastlineWorldFine[["longitude"]],
        coastlineWorldFine[["latitude"]], col="lightgray")
# Indicate float positions.
points(index[["longitude"]], index[["latitude"]], pch=20)



# Example 7: Temperature profile of the 131st cycle of float with ID 2902204
a <- readProfiles(system.file("extdata", "SR2902204_131.nc", package="argoFloats"))
#> Warning: Of 1 profiles read, 1 has >10% of BBP700 values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of chlorophyllA values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of oxygen values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of pressure values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
oldpar <- par(no.readonly=TRUE)
par(mfrow=c(1, 1))
par(mgp=c(2, 0.7, 0))                  # mimic the oce::plotProfile() default
par(mar=c(1,3.5,3.5,2))                # mimic the oce::plotProfile() default
plot(a, which="profile")

par(oldpar)

# Example 8: As Example 7, but showing temperature dependence on potential density anomaly.
a <- readProfiles(system.file("extdata", "SR2902204_131.nc", package="argoFloats"))
#> Warning: Of 1 profiles read, 1 has >10% of BBP700 values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of chlorophyllA values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of oxygen values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
#> Warning: Of 1 profiles read, 1 has >10% of pressure values with QC flag of 4, signalling bad data.
#>     The indices of the bad profiles are as follows.
#>     1
oldpar <- par(no.readonly=TRUE)
par(mgp=c(2, 0.7, 0))                  # mimic the oce::plotProfile() default
par(mar=c(1,3.5,3.5,2))                # mimic the oce::plotProfile() default
plot(a, which="profile", profileControl=list(parameter="temperature", ytype="sigma0"))

par(oldpar)

# }