RandomFields (version 3.0.10)

RMtrafo: Transformation of coordinate systems

Description

The functions transform a coordinate system into another. Currently, essentially only from the earth system to cartesian. RMtrafo is the internal basic function that also allows to reduce vectors to their norm.

Usage

RMtrafo(isotropy)
RFearth2cartesian(coord, units=NULL)
RFearth2dist(coord, units=NULL, ...)

Arguments

isotropy

one of the values ISOTROPIC, SPACEISOTROPIC, CARTESIAN_COORD,

coord

matrix or vector of earth coordinates

units

"km" or "miles"; if not given and RFoptions()$general$units != "" the latter is used. Otherwise "km".

...

the optional arguments of dist

Value

• The function RMtrafo returns a matrix, in general. For fixed column, the results, applied to each row of the matrix, are returned.

  The function RFearth2cartesian returns a matrix in one-to-one correspondance with coord assuming that the earth is an ellipsoid.

  The function RFearth2dist calculates distances, cf. dist, assuming that the earth is an ellipsoid.

Details

The functions transform between different coordinate systems.

References

For calculating the earth coordinates as ellipsoid:

• en.wikipedia.org/wiki/Geographic_coordinate_system
• nssdc.gsfc.nasa.gov/planetary/factsheet/earthfact.html

See Also

link{constants}, RMangle

Examples

data(weather)
(coord <- weather[1:5, 3:4])

(z <- RFfctn(RMtrafo(CARTESIAN_COORD), coord))
(z1 <- RFearth2cartesian(coord)) ## equals t(z)
z1 - t(z) ## 0, i.e., z1 and t(z) are the same
dist(t(z))


(d <- RFearth2dist(coord)) 
d - dist(t(z)) ## 0, i.e., d and dist(t(z)) are the same


\dontrun{


## RFearth2dist is more precise !!!
library("fields")
(unix.time(z2 <- rdist.earth(coord, miles=FALSE)))
z2

(z2 - as.matrix(d)) / z2 * 100 ## RFearth2dist is more precise


# 
n <- 10000
unix.time(for (i in 1:n) d <- RFearth2dist (coord))) ## the same as dist(t(z))

unix.time(for (i in 1:n)z2 <- rdist.earth(xx)))


data(weather)
xx <- matrix(weather[, 3:4], ncol=2) [ 1:5, ]

xx <- weather[, 3:4][ 1:5, ]
Print(xx)
Print(yy <- RFfctn(RMtrafo(iso=CARTESIAN_COORD), xx), units="miles")

(z1 <- as.matrix(dist(t(yy))))

library(fields)
(z2 <- rdist.earth(xx))

(z1-z2) / z1 * 100
z1
}


RFoptions(seed=NA)