rvad

The goal of rvad is to approximate the horizontal components of the wind from radial wind measured by Doppler radar using the Velocity Azimuth Display method from Browning and Wexler (1968).

Installation

You can install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("paocorrales/rvad")

Example

rvad estimates the horizontal components of wind from radial velocity measured with Doppler radar. It takes data on a tidy format –one value of radial velocity for each azimuth, elevation angle and distance to the radar (range):

library(rvad)
str(radial_wind)  # sample dataset
#> Classes 'data.table' and 'data.frame':   2076960 obs. of  4 variables:
#>  $ range      : int  250 750 1250 1750 2250 2750 3250 3750 4250 4750 ...
#>  $ radial_wind: num  NA NA -4.189 0.419 -4.556 ...
#>  $ azimuth    : num  214 214 214 214 214 ...
#>  $ elevation  : num  0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 ...
#>  - attr(*, ".internal.selfref")=<externalptr>

This is how the data looks like for one particular elevation angle.

library(ggplot2)
one_elevation <- subset(radial_wind, elevation == unique(elevation)[3] &
                          !is.na(radial_wind))
ggplot(one_elevation, aes(azimuth, range)) +
  geom_point(aes(color = radial_wind, size = range^2)) +
  scale_radius(range = c(0, 0.05), guide = "none") +
  scale_color_gradient2(low = "blue", high = "red") +
  scale_y_continuous(limits = c(0, 100000)) +
  coord_polar()
#> Warning: Removed 259 rows containing missing values (geom_point).

Red means that the wind is moving away from the radar and blue is wind moving towards the radar. So this plot shows that, overall, the wind is comming from the northeast.

vad_fit() does some quality control and then fits a sinusoidal model to the data and returns zonal and meridional wind for each elevation angle, and range. It also approximates the propagation of the radar beam to get the height of each observation.

VAD <- with(radial_wind, vad_fit(radial_wind, azimuth, range, elevation))
str(VAD)
#> Classes 'rvad_vad' and 'data.frame': 5760 obs. of  7 variables:
#>  $ height   : num  2.19 6.58 11 15.45 19.93 ...
#>  $ u        : num  NA NA NA NA NA NA NA NA NA NA ...
#>  $ v        : num  NA NA NA NA NA NA NA NA NA NA ...
#>  $ range    : int  250 750 1250 1750 2250 2750 3250 3750 4250 4750 ...
#>  $ elevation: num  0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 ...
#>  $ r2       : num  NA NA NA NA NA NA NA NA NA NA ...
#>  $ rmse     : num  NA NA NA NA NA NA NA NA NA NA ...
#>  - attr(*, "rvad_raw")= logi TRUE

This function has a plot method.

plot(VAD)

The optimal quality control parameters (minimum r^2, maximum missing values, etc…) can be case-dependant so it’s advisable to test different values.

vad_regrid() aggregates the result into a single wind profile defined in a regular grid using local regression.

wind_profile <- vad_regrid(VAD, layer_width = 100)

plot(wind_profile)