Streamlines are paths that are always tangential to a vector field. In the case of a steady field, it's identical to the path of a massless particle that moves with the "flow".
geom_streamline( mapping = NULL, data = NULL, stat = "streamline", position = "identity", ..., L = 5, min.L = 0, res = 1, S = NULL, dt = NULL, xwrap = NULL, ywrap = NULL, skip = 1, skip.x = skip, skip.y = skip, n = NULL, nx = n, ny = n, jitter = 1, jitter.x = jitter, jitter.y = jitter, arrow.angle = 6, arrow.length = 0.5, arrow.ends = "last", arrow.type = "closed", arrow = grid::arrow(arrow.angle, grid::unit(arrow.length, "lines"), ends = arrow.ends, type = arrow.type), lineend = "butt", na.rm = TRUE, show.legend = NA, inherit.aes = TRUE ) stat_streamline( mapping = NULL, data = NULL, geom = "streamline", position = "identity", ..., L = 5, min.L = 0, res = 1, S = NULL, dt = NULL, xwrap = NULL, ywrap = NULL, skip = 1, skip.x = skip, skip.y = skip, n = NULL, nx = n, ny = n, jitter = 1, jitter.x = jitter, jitter.y = jitter, arrow.angle = 6, arrow.length = 0.5, arrow.ends = "last", arrow.type = "closed", arrow = grid::arrow(arrow.angle, grid::unit(arrow.length, "lines"), ends = arrow.ends, type = arrow.type), lineend = "butt", na.rm = TRUE, show.legend = NA, inherit.aes = TRUE )
mapping  Set of aesthetic mappings created by 

data  The data to be displayed in this layer. There are three options: If A A 
stat  The statistical transformation to use on the data for this layer, as a string. 
position  Position adjustment, either as a string, or the result of a call to a position adjustment function. 
...  Other arguments passed on to 
L,  typical length of a streamline in x and y units 
min.L  minimum length of segments to show 
res,  resolution parameter (higher numbers increases the resolution) 
S  optional numeric number of timesteps for integration 
dt  optional numeric size "timestep" for integration 
xwrap, ywrap  vector of length two used to wrap the circular dimension. 
skip  numeric specifying number of gridpoints not to draw in the x and y direction 
skip.x  numeric specifying number of gridpoints not to draw in the x and y direction 
skip.y  numeric specifying number of gridpoints not to draw in the x and y direction 
n, nx, ny  optional numeric indicating the number of points to draw in the
x and y direction (replaces 
jitter, jitter.x, jitter.y  amount of jitter of the starting points 
arrow.angle  parameters passed to grid::arrow 
arrow.length  parameters passed to grid::arrow 
arrow.ends  parameters passed to grid::arrow 
arrow.type  parameters passed to grid::arrow 
arrow  specification for arrow heads, as created by arrow(). 
lineend  Line end style (round, butt, square). 
na.rm  If 
show.legend  logical. Should this layer be included in the legends?

inherit.aes  If 
geom  The geometric object to use display the data 
Streamlines are computed by simple integration with a forward Euler method.
By default, stat_streamline()
computes dt
and S
from L
, res
,
the resolution of the grid and the mean magnitude of the field. S
is
then defined as the number of steps necessary to make a streamline of length
L
under an uniform mean field and dt
is chosen so that each step is no
larger than the resolution of the data (divided by the res
parameter). Be
aware that this rule of thumb might fail in field with very skewed distribution
of magnitudes.
Alternatively, L
and/or res
are ignored if S
and/or dt
are specified
explicitly. This not only makes it possible to finetune the result but also
divorces the integration parameters from the properties of the data and makes
it possible to compare streamlines between different fields.
The starting grid is a semi regular grid defined, either by the resolution of the
field and the skip.x
and skip.y
parameters o the nx
and ny
parameters,
jittered by an amount proportional to the resolution of the data and the
jitter.x
and jitter.y
parameters.
It might be important that the units of the vector field are compatible to the units
of the x and y dimensions. For example, passing dx
and dy
in m/s on a
longitudelatitude grid will might misleading results (see spherical).
Missing values are not permitted and the field must be defined on a regular grid, for now.
stat_streamline
understands the following aesthetics (required aesthetics are in bold)
x
y
dx
dy
alpha
colour
linetype
size
step in the simulation
dx at each location of the streamline
dy at each location of the streamline
Other ggplot2 helpers:
DivideTimeseries()
,
MakeBreaks()
,
WrapCircular()
,
geom_arrow()
,
geom_contour2()
,
geom_contour_fill()
,
geom_label_contour()
,
geom_relief()
,
guide_colourstrip()
,
map_labels
,
reverselog_trans()
,
scale_divergent
,
scale_longitude
,
stat_na()
,
stat_subset()
if (FALSE) { library(data.table) library(ggplot2) data(geopotential) geopotential < copy(geopotential)[date == date[1]] geopotential[, gh.z := Anomaly(gh), by = .(lat)] geopotential[, c("u", "v") := GeostrophicWind(gh.z, lon, lat)] (g < ggplot(geopotential, aes(lon, lat)) + geom_contour2(aes(z = gh.z), xwrap = c(0, 360)) + geom_streamline(aes(dx = dlon(u, lat), dy = dlat(v)), L = 60, xwrap = c(0, 360))) # The circular parameter is particularly important for polar coordinates g + coord_polar() # If u and v are not converted into degrees/second, the resulting # streamlines have problems, specially near the pole. ggplot(geopotential, aes(lon, lat)) + geom_contour(aes(z = gh.z)) + geom_streamline(aes(dx = u, dy = v), L = 50) # The step variable can be mapped to size or alpha to # get cute "drops". It's important to note that ..dx.. (the calculated variable) # is NOT the same as dx (from the data). ggplot(geopotential, aes(lon, lat)) + geom_streamline(aes(dx = dlon(u, lat), dy = dlat(v), alpha = ..step.., color = sqrt(..dx..^2 + ..dy..^2), size = ..step..), L = 40, xwrap = c(0, 360), res = 2, arrow = NULL, lineend = "round") + scale_size(range = c(0, 0.6)) # Using topographic information to simulate "rivers" from slope topo < GetTopography(295, 55+360, 30, 42, res = 1/20) # needs internet! topo[, c("dx", "dy") := Derivate(h ~ lon + lat)] topo[h <= 0, c("dx", "dy") := 0] # See how in this example the integration step is too coarse in the # western montanous region where the slope is much higher than in the # flatlands of La Pampa at in the east. ggplot(topo, aes(lon, lat)) + geom_relief(aes(z = h), interpolate = TRUE, data = topo[h >= 0]) + geom_contour(aes(z = h), breaks = 0, color = "black") + geom_streamline(aes(dx = dx, dy = dy), L = 10, skip = 3, arrow = NULL, color = "#4658BD") + coord_quickmap() }