如何在R语言地图指定经纬度网格位置绘制风玫瑰图？

Absolutely, this is totally achievable! The trick is combining base R's map plotting with grid graphics (since openair::windRose uses grid under the hood) to overlay wind roses at specific geographic coordinates. Here's a step-by-step solution using your test data:

Step 1: Load Required Packages #

Make sure you have these installed and loaded first:

library(maps)
library(openair)
library(grid)

Step 2: Prepare Your Data (Matching Your Test Code) #

# Geographic coordinates
lats <- seq(18, 30, 3)
lons <- seq(-98, -86, 3)
lonlats <- expand.grid(lons, lats)

# Simulate current data (add set.seed for reproducibility)
set.seed(123)
arr.dir <- array(sample(1:360, 1250, replace=T), dim = c(5,5,50))
arr.spd <- array(sample(0:140, 1250, replace=T), dim = c(5,5,50))

Step 3: Plot the Base Map #

First, draw your base map and keep the plot active so we can convert geographic coordinates to device coordinates:

plot(lonlats$Var1, lonlats$Var2, xlab="Longitude", ylab="Latitude", asp=1, pch=19, col="blue")
map(database = "world", add=T, fill=T, col="gray"); box()

Step 4: Define a Helper Function to Plot Wind Roses on the Map #

Create a reusable function that handles coordinate conversion and wind rose plotting:

plot_wind_rose_on_map <- function(lon, lat, dir, spd, rose_size = 0.15) {
  # Convert geographic coordinates to normalized device coordinates (NDC)
  x_dev <- grconvertX(lon, from = "user", to = "ndc")
  y_dev <- grconvertY(lat, from = "user", to = "ndc")
  
  # Create a grid viewport at the target position
  vp <- viewport(x = x_dev, y = y_dev, width = rose_size, height = rose_size, just = c("center", "center"))
  pushViewport(vp)
  
  # Plot a compact wind rose (adjust parameters as needed)
  windRose(
    mydata = data.frame(ws = spd, wd = dir),
    paddle = FALSE,
    legend = FALSE, # Remove legend to save space
    axes = FALSE, # Hide outer axes
    border = NA, # No border around the rose
    main = NULL, # No title
    col = openair::openColours() # Use openair's default color scheme
  )
  
  popViewport()
}

Step 5: Plot Wind Roses at Target Coordinates #

Single Wind Rose (Your Test Point: -86, 30) #

# Extract data for the target coordinate
target_lon <- -86
target_lat <- 30
dir_data <- arr.dir[which(lons == target_lon), which(lats == target_lat), ]
spd_data <- arr.spd[which(lons == target_lon), which(lats == target_lat), ]

# Plot the wind rose at this location
plot_wind_rose_on_map(target_lon, target_lat, dir_data, spd_data)

Multiple Wind Roses (e.g., All Grid Points) #

Loop through your coordinate list to add roses at every grid point (adjust rose_size to avoid overlap):

for (i in 1:nrow(lonlats)) {
  lon <- lonlats$Var1[i]
  lat <- lonlats$Var2[i]
  
  # Get the corresponding array indices
  lon_idx <- which(lons == lon)
  lat_idx <- which(lats == lat)
  
  # Extract direction and speed data
  dir_data <- arr.dir[lon_idx, lat_idx, ]
  spd_data <- arr.spd[lon_idx, lat_idx, ]
  
  # Plot the wind rose
  plot_wind_rose_on_map(lon, lat, dir_data, spd_data, rose_size = 0.12)
}

Key Customization Tips: #

• Rose Size: Adjust rose_size (0 to 1, where 1 = full plot area) to fit your map density.
• Speed Binning: Add breaks = seq(0, 140, 20) to windRose to set custom speed intervals.
• Color Scheme: Replace openColours() with your preferred palette (e.g., viridis::viridis(5)).
• Add Legends: If you want a shared legend, plot it separately using drawOpenKey() from openair.

This method works by converting geographic coordinates to grid-compatible normalized device coordinates, then drawing each wind rose in its own isolated viewport over the base map.

内容的提问来源于stack exchange，提问作者MilloMarinE