geocontour.fill: geocontour.fill plots colored or black and white contours on...
In geo: Draw and Annotate Maps, Especially Charts of the North Atlantic

Description Usage Arguments Details See Also Examples

The program accepts data on a rectangular grid. Irregular data can be interpolated on the grid in a number of ways, using kriging, the Splus function interp, etc. NA's are allowed in the matrix but they are changed to 0 or the average value of z early in the program. The geo package contains two different contour plot programs, geocontour and geocontour.fill. Geocontour draws contour lines and geocontour.fill fills in between the lines. Hardcopy of the plot can be done on a color postscript printer or on a black and white postscript printer. In all cases the "postscript" driver included with Splus has to be used. There is another postscript driver included with Splus called "pscript". That driver is used when a plot is on the screen and the print button is pressed. That can not be done when the postscript driver is used as described here after.

geocontour.fill(grd, z, levels = NULL, nlevels = 0, cex = 0.7,
  digits = 1, col = NULL, working.space = 0, labels = 1, ratio = 1000,
  only.positive = FALSE, fill = 0, maxcol = 155, white = FALSE,
  label.location = 0, labels.only = FALSE, bordercheck = FALSE,
  maxn = 10000, bcrat = 0.05, limits = NULL, col.names = c("lon",
  "lat"), minsym = "<", label.resolution = 0, labtxt = NULL, boxcol = 0,
  first.color.trans = TRUE, mai = c(0, 1, 0, 1), leftrat = 0.1,
  labbox = TRUE, csi = NULL)

`grd`	List with components \$lat and \$lon (or \$x, \$y) defining the grid values. Can be made for example by lat <- list(lat = seq(60, 65, length = 100), lon = seq(-30, - 10, length = 200)). Also lat can be the outcome of the program grid and then the components lat\$grpt\$lat, lat\$grpt\$lon and lat\$reg are used.
`z`	Matrix or vector of values. Length(z) = nlat*nlon except when lat is a list with components lat\$grpt and lat\$xgr. Then the length of z is the same as the length of lat\$xgr\$lat and lat\$xgr\$lon. This is the case if the output of the grid program is used as input lat.
`levels`	Values at contourlines. Default value is zero. If levels is zero the program determines the contourlines from the data. If levels is of length 3 with levels = c(0, 1, 2) then there are 4 groups each with special color i.e. <0, 0-1, 1-2 and > 2.
`nlevels`	Number of contourlines. Used if the program has to determine the contourlines. Default value is 10. nlevels = 10 does not always mean 10 due to characteristic of the pretty command.
`cex`	Character size expansion. Size of letters in labels. Default value is 0.7.
`digits`	Number of digits in the labels. Default value is 1.
`col`	Color number for the contourlines. A vector one longer than the vector levels Default is blue-green- yellow-red from lowest to the highest values. The program assumes certain setup of the splus colors that is described later. Colors should not be specified explicitly except the contour lines are.
`working.space`	Size of working space. The program determines it from data and prints on the screen. If mysterious errors occur then it is likely that the program has not reserved enough working space. The program writes the used work space on the screen for use in similar situations. (saves time.)
`labels`	Type of labels, either 1 or 2 <s-example> labels = 1 means type of label used with few contourlines <20. labels = 2 means type of label used with many contourlines >20. can also use labels = 3 means only labels, no contour plot. </s-example> Labels can be inserted in two ways, by calling geoplot with cont = TRUE or by specifying label.location. In the first case the left part of the plot is reserved for labels while in the latter case the label is put in a place specified by the user. The latter method is recommended in most cases.
`ratio`	Factor used to avoid numerical problems. Default value 100. Lower values decrease numerical problems but can introduce bias.
`only.positive`	Logical value. If TRUE then negative values are not allowed else negative values are set to zero. Default value is FALSE.
`fill`	Determines whether NA's should be replaced with zeros (fill = 1) or mean(z) (fill = 2). Default value is fill = 1. Used when lat is a list with components lat\$xgr\$lat, lat\$xgr\$lon, lat\$grpt\$lat and lat\$grpt\$lon. Only used in special cases.
`maxcol`	Number of colors used (excluding #0). Default value 155.
`white`	If true the the first class is represented with white (color 0). Default value is FALSE but white = TRUE is also often used.
`label.location`	List with components \$lat and \$lon. (or \$x, \$y) Gives the lower left and upper right corner of the box where the labels are put. Default value is 0 that means no labels are put on the drawing (except when geopar\$cont = TRUE). l1 is best given by geolocator or directly by specifying label.location = "locator".
`labels.only`	If labels.only is true no contours are drawn but only labels. Default value is false. Order of the commands could be: <s-example> > geoplot(deg, plot = FALSE) > geocontour.fill(lat, z = z, levels = lev, reg = area) # Draw contours. > geocontour(lat, z = z, levels = lev, colors = FALSE, reg = area) # Only used with black and white printers to make distinction # between different levels clearer. (not used with color # printers). > geoplot(deg, new = TRUE) # Refresh gridlines. > geocontour.fill(lat, z = z, levels = lev, labels.only = TRUE, label.location = l1) # Add labels. </s-example>
`bordercheck`	if true the program checks if plot is outside border and does not plot what is outside border.
`maxn`	a parameter determing the resolution of the plot, unimportant.
`bcrat`	bordercheck ratio, how much outside the border we will allow to be plotted, default is bcrat = 0.05 meaning that we will allow the plot to go 5% off the border.
`limits`	To be described.
`col.names`	the names of the columns in grid, the first argument will be plotted on the x-axes and the second on the y-axes. Default is col.names = c("lat", "lon").
`minsym`	minimum symbol, default is "<", meaning that if levels = c(1, 2), labels will be presented as < 1, 1-2, 2 <, but if minsym = " " labels will be presented as 1, 1-2, 2. See also labels.resolution.
`label.resolution`	the resolution (precision) of the label numbers, default is 0, meaning that if levels = c(1, 2), labels will be presented < 1, 1-2, 2 <, if label.resolution = 0.1 labels will be presented < 1, 1.1-2, 2.1<, see also minsym. If label.resolution = "none", the labels will present the lowest number of the interval with each color.
`labtxt`	To be described.
`boxcol`	Colour of box around labels (legend?).
`first.color.trans`	To be described.
`mai`	`par` argument 'margin in inches'?
`leftrat`	To be described.
`labbox`	should a box be drawn around labels (legend?).
`csi`	Size of character. This parameter can not be set in R but for compatibility with old Splus scripts the parameter cex is readjusted by cex = cex*csi/0.12. Use of this parameter is not recommended. Default value is NULL i.e not used.

<s-example>

The program is based on making triangles out of a matrix of data nx * ny. The number of triangles is (nx-1)*(ny- 1)*4 and linear interpolation is used inside each triangle. The factor 4 comes from the fact that the number of points is doubled by interpolation. </s-example> <s-example>

Most of the calculations in the program are done by a program written in C. The Splus part of the program prepares the data for the C program, reserves memory and the user interface is in Splus. The C program is loaded automatically. </s-example> <s-example>

The program is currently based on the following color setup in Splus. </s-example> <s-example>

splus*color : white black blue 50 green 50 yellow 50 red </s-example> <s-example>

This gives 155 colors with color#0 white, #1 black, #2 blue, #53 green, #104 yellow, and #155 red. The parameter maxcol is set to 155 based on this setup but it can be changed if another setup is used. The vector postcol stores the mapping from the colors on the terminal to color postscript based on this setup. </s-example> <s-example>

If the openlook() or motif() windowmanager is used in Splus the colors can be changed inside Splus. In openlook it is best to make a colorscheme called geoplot: black blue 50 green 50 yellow 50 red. The background color, i.e. white is not in the definition here. </s-example> <s-example>

The maximum number of contourlines that can be used is currently 60. </s-example> <s-example>

The program requires that geoplot is called before to set up the drawing. geoplot is called by the parameter cont = TRUE. This is to reserve space for labels on the drawing. </s-example> <s-example>

If the drawing is written to a file geoplot should be called with plot = FALSE. Then it only sets up the drawing but does not plot anything so the size of the file will be reduced. </s-example> <s-example>

If the labels do not fit the relative space taken by labels and picture can be changed by the parameter lcont in geoplot. After geocontour.fill is called geoplot is called again with new = TRUE to get all kind of lines back but they were painted over by the program. </s-example> <s-example>

The program treats borders in a special way. It begins by making contour-lines over the hole area as if the borders did not exist. When that is finished the area outside the borders is painted white. </s-example> <s-example>

If good picture is needed geocontour should be used between levels to get sharper pictures. </s-example> <s-example>

If the matrix z is not full the program should be called by a list lat\$xgr\$lat, lat\$xgr\$lon, lat\$grpt\$lat and lat\$grpt\$lon. Then the length of the vectors z, lat\$xgr\$lat and lat\$xgr\$lon is the same. lat\$grpt\$lat and lat\$grpt\$lon is on the other hand the coordinates of the rows and columns of the matrix. The program grid makes a list with components with these names. </s-example> <s-example>

The functions geolines, geopoint, geopolygon, geotext & geosymbols can be used to add things to the contourplot. </s-example>

geoplot, geolines, geopolygon, geotext, geosymbols, geogrid, geopar, geolocator, geocontour, reitaplott, geodefine.

## Not run:   ######################################################
  # Example 1.                                         #
  ######################################################

  # Need the data.frame botnv.2004 to be able to compile this
  # example, if not attached use:
  # >attach("/usr/local/reikn/Splus5/Aflaskyrslur/Data")

  codgrd <-list(lat = seq(62, 68, by = 0.1), lon = seq(-30, -9, by = 0.25))
  # A grid is made.
  lab.loc<-list(lat = c(63.9, 65.6), lon = c(-20.75, -16.5))
  # Location of the label.

  tmp <- combine.rt(botnv.2004$lat, botnv.2004$lon,
                    botnv.2004$torskur, codgrd, fun = "sum", fill = TRUE)
  # The data is read into the grid with combine.rt.

  tmp$z <- tmp$z/(cos(tmp$lat*pi/180)*0.1*60*0.25*60)
  # Data changed.(from being in
  # kilos per box to kilos per square mile).

  vg <- list(nugget = 0.1, sill = 1, range = 50)
  # Parameters for the variogram

  z <- pointkriging(tmp$lat, tmp$lon, tmp$z, codgrd, vg,
                    maxnumber = 80, maxdist = 30, set = -1)
  # Dataset smoothened with pointkriging.

  geoplot(lat = c(63, 67.5), lon = c(-27, -11), grid = FALSE, axlabels = TRUE, type = "n")
  # Plot initialized
  level = c(160, 200, 320, 500, 700, 1000, 2000, 3000, 4000, 5000, 6000)
  # Levels for geocontour.fill

  geocontour.fill(codgrd, z, levels = level, white = TRUE      # Plot the data.
                , working.space = 300000)

  geopolygon(island)
  # Contourlines inside Iceland overwritten.
  geolines(island)
  # Iceland redrawn with geolines.

  geocontour.fill(codgrd, z, levels = level, white = TRUE,
                  label.location = lab.loc, labels.only = TRUE)
  # Call geocontour.fill again only to plot the labels.

  #########################################################
  # Example 2.                                            #
  #########################################################


  # Preperation for pointkriging
  # th4.2002 is the data used here, dataframe [lon, lat, mat].
  # >attach(?????)


  grd.smb <-list(lat = seq(62.8, 67.5, length = 80),   # Set up the grid.
                 lon = seq(-28, -10, length = 130))
  m.lev<-c(0, 0.1, 0.2, 0.3, 0.5)
  # Levels for the geocontour.fill.
  m.col<-c(0, 14, 59, 104, 119, 149)
  # Colors for the levels.
  lab.in.island<-list(lat = c(63.9, 65.6), lon = c(-20.75, -16.5))
  # Location of the Label

  vg <- list(nugget = 0.3, sill = 1, range = 50)
  # Initialize variogram parameters.

  zfj<-pointkriging(th4.2002$lat, th4.2002$lon, z = th4.2002$mat,
                  grd.smb, vg, maxnumber = 80, maxdist = 30, set = -1)
  # Smooth the data with pointkriging.

  #
  # Plotting
  #

  par(mfrow = c(1, 1),  mai = rep(0, 4))
  # Set up graphic parameters.
  geoplot(lat = c(63, 67.5), lon = c(-27, -11), grid = FALSE, axlabels = FALSE, type = "n")
  # Draw a background with Iceland with geoplot.

  geocontour.fill(zfj, levels = m.lev, col = m.col, working.space = 300000)
  # Plot the data with geocontour.fill.

  geopolygon(gbdypif.500, col = 0, exterior = TRUE, r = 0)
  # Remove contours outside gbdypif.500.
  geoplot(lat = c(63, 67.5), lon = c(-27, -11), grid = FALSE,          # Replot.
          axlabels = FALSE, type = "n", new = TRUE)

  geopolygon(island, col = 43)
  # Remove contours inside Iceland and color Iceland.
  geocontour.fill(zfj, levels = m.lev, label.location = lab.in.island,
                  labels.only = TRUE, csi = 0.1, col = m.col, working.space = 300000)
  # Call geocontour.fill to plot labels.
  geolines(island)
  # Redraw the lines of Iceland.

## End(Not run)