idw360 {aire.zmvm} | R Documentation |
Inverse Distance Weighting with Directional Data
Description
Function for inverse distance weighted interpolation with directional data. Useful for when you are working with data whose unit of measurement is degrees (i.e. the average of 35 degrees and 355 degrees should be 15 degrees). It works by finding the shortest distance between two degree marks on a circle.
Usage
idw360(values, coords, grid, idp = 2)
Arguments
values |
the dependent variable |
coords |
the spatial data locations where the values were measured. First column x/longitude, second y/latitude |
grid |
data frame or Spatial object with the locations to predict. First column x/longitude, second y/latitude |
idp |
The inverse distance weighting power |
Value
data.frame with the interpolated values for each of the grid points
Examples
library("sp")
library("ggplot2")
## Could be wind direction values in degrees
values <- c(55, 355)
## Location of sensors. First column x/longitud, second y/latitude
locations <- data.frame(lon = c(1, 2), lat = c(1, 2))
coordinates(locations) <- ~lon+lat
## The grid for which to extrapolate values
grid <- data.frame(lon = c(1, 2, 1, 2), lat = c(1, 2, 2, 1))
coordinates(grid) <- ~lon+lat
## Perform the inverse distance weighted interpolation
res <- idw360(values, locations, grid)
head(res)
## Not run:
df <- cbind(res, as.data.frame(grid))
## The wind direction compass starts where the 90 degree mark is located
ggplot(df, aes(lon, lat)) +
geom_point() +
geom_spoke(aes(angle = ((90 - pred) %% 360) * pi / 180),
radius = 1,
arrow=arrow(length = unit(0.2, "npc")))
library("mapproj")
## Random values in each of the measuring stations
locations <- stations[, c("lon", "lat")]
coordinates(locations) <- ~lon+lat
crs_string <- "+proj=longlat +ellps=WGS84 +no_defs +towgs84=0,0,0"
proj4string(locations) <- CRS(crs_string)
values <- runif(length(locations), 0, 360)
pixels <- 10
grid <- expand.grid(lon = seq((min(coordinates(locations)[, 1]) - .1),
(max(coordinates(locations)[, 1]) + .1),
length.out = pixels),
lat = seq((min(coordinates(locations)[, 2]) - .1),
(max(coordinates(locations)[, 2]) + .1),
length.out = pixels))
grid <- SpatialPoints(grid)
proj4string(grid) <- CRS(crs_string)
## bind the extrapolated values for plotting
df <- cbind(idw360(values, locations, grid), as.data.frame(grid))
ggplot(df, aes(lon, lat)) +
geom_point(size = .1) +
geom_spoke(aes(angle = ((90 - pred) %% 360) * pi / 180),
radius = .07,
arrow=arrow(length = unit(0.2,"cm"))) +
coord_map()
## End(Not run)
[Package aire.zmvm version 1.0.0 Index]