CODEWITHSUNDEEP
rplotweather
Badger
Badger

Reputation: 1053

Plot Wind Barb in R

This is more a question to see if anyone has seen anything like this in their travels. I am working with a lot of weather data and I would like to plot wind based on wind barbs.

I have looked into the package RadioSonde however its plotwind() function is not doing the job I had anticipated. It does have a good example of the type of data data(ExampleSonde)

Arguably I can use TeachingDemos in conjunction with my.symbols() to create these wind barbs. I was just curious if anyone has found (or created) a way to plot wind barbs. Otherwise my.symbols() it is.

Thanks,

Badger

Upvotes: 2

Views: 2633

Answers (1)

user20650
user20650

Reputation: 25864

Another way is to create the wind barbs using grid graphics.

First step is to calculate how many, and what type of barb is needed. As described here, I created three types, that represent 50, 10, and 5 knots - I round down the speed to the nearest five.

The function below wind_barb generates a new grob for each wind speed it is given. Using an idea from Integrating Grid Graphics Output with Base Graphics Output - Murrell (pg4) you can plot the grobs easily and represent the wind direction by rotating the viewport.

An example

Create some data

set.seed(1)
dat <- data.frame(x=-2:2, y=-2:2, 
                  direction=sample(0:360, 5), 
                  speed=c(10, 15, 50, 75, 100))
#    x  y direction speed
# 1 -2 -2        95    10
# 2 -1 -1       133    15
# 3  0  0       205    50
# 4  1  1       325    75
# 5  2  2        72   100

Plot

library(gridBase)
library(grid)

with(dat, plot(x, y, ylim=c(-3, 3), xlim=c(-3, 3), pch=16))

vps <- baseViewports()
pushViewport(vps$inner, vps$figure, vps$plot)
# Plot
for (i in 1:nrow(dat)) {
    pushViewport(viewport(
                    x=unit(dat$x[i], "native"),
                    y=unit(dat$y[i], "native"), 
                    angle=dat$direction[i]))
        wind_barb(dat$speed[i])
    popViewport()
  }

popViewport(3)

Which produces

enter image description here

wind_barb function to create barbs (please simplify me!). You can change the height and width of the barb by adjusting mlength and wblength arguments respectively.

wind_barb <- function(x, mlength=0.1, wblength=0.025) {

  # Calculate which / how many barbs
    # any triangles (50)
    fif <- floor(x /50)
    # and then look for longer lines for remaining speed (10)
    tn <- floor( (x - fif* 50)/10)
    # and then look for shorter lines for remaining speed (5)
    fv <- floor( (x - fif* 50 - tn* 10)/5)

    # Spacing & barb length
    yadj <- 0.5+mlength
    dist <- (yadj-0.5) / 10    
    xadj <- 0.5+wblength
    xfadj <- 0.5+wblength/2        

  # Create grobs
    main_grob <- linesGrob(0.5, c(0.5, yadj ))

    # 50 windspeed
    if(fif != 0) {
      fify <- c(yadj, yadj-dist*seq_len(2* fif) )
      fifx <- c(0.5, xadj)[rep(1:2, length=length(fify))]
      fif_grob <- pathGrob(fifx, fify, gp=gpar(fill="black"))
    } else {
      fif_grob <- NULL
      fify <- yadj+dist
    }

    # Ten windspeed
    if(tn != 0) {
    tny <- lapply(seq_len(tn) , function(x) min(fify) - dist*c(x, x-1))  
    tn_grob <- do.call(gList, 
                      mapply(function(x,y) 
                             linesGrob(x=x, y=y, gp=gpar(fill="black")),
                                      x=list(c(0.5, xadj)), y=tny, SIMPLIFY=FALSE))
    } else {
      tn_grob <- NULL
      tny <- fify
    }

    # Five windspeed
    if(fv != 0) {
    fvy <- lapply(seq_len(fv) , function(x) min(unlist(tny)) -dist* c(x, x-0.5))
    fv_grob <- do.call(gList, 
                        mapply(function(x,y) 
                              linesGrob(x=x, y=y, gp=gpar(fill="black")),
                                      x=list(c(0.5, xfadj)), y=fvy, SIMPLIFY=FALSE))
    } else {
      fv_grob <- NULL
    }    

    # Draw    
    #grid.newpage()
    grid.draw(gList(main_grob, fif_grob, tn_grob, fv_grob))
}

-------------------------------------

comment from sezen below

The plotted wind direction is wrong. To have right meteorological wind direction, use angle = 360 - dat$direction[i]. See http://tornado.sfsu.edu/geosciences/classes/m430/Wind/WindDirection.html

Upvotes: 5

Related Questions