ggplot2绘制垂直廓线时x/y轴翻转及坐标匹配问题与对数轴需求

Hey there! Let's tackle two key problems you're facing: mismatched values and coordinates, plus setting up a logarithmic pressure axis for that classic meteorological profile look. Here's a detailed breakdown and corrected code:

What Went Wrong in Your Original Script #

• You used discrete scales (scale_x_discrete/scale_y_discrete) for continuous data (pressure levels and humidity bias), which broke the value-coordinate mapping
• After coord_flip(), the axis mappings got reversed but your scale settings didn't account for that, leading to further confusion
• 0 hPa is an invalid pressure level for this context, and including it would cause issues with logarithmic scaling

Corrected Code (Tested with Your Data) #

First, let's use the sample data you provided to verify the fix, then adapt it for batch processing:

library(ggplot2)

# Your test data
dat <- structure(list(Lev = c(1000L, 950L, 900L, 850L, 800L, 750L, 700L, 650L, 600L, 550L, 500L, 450L, 400L, 350L, 300L, 250L, 200L, 150L, 100L, 50L, 0L), 
                      April29 = c(NA, NA, NA, NA, 85, NA, NA, 39.5, 33, NA, NA, 56, NA, 57, 49, NA, NA, 42, 20, 15, NA)), 
                 class = "data.frame", row.names = c(NA,-21L))

# Build the plot with proper mappings
p <- ggplot(dat, aes(x = April29, y = Lev)) +
  # Draw line, skipping NA values to avoid breaks
  geom_line(aes(color = "April29"), na.rm = TRUE) +
  # Add red X markers for data points
  geom_point(shape = 4, size = 2, color = "red", na.rm = TRUE) +
  # Flip axes to get vertical pressure profile
  coord_flip() +
  # Clean theme matching your original style
  theme(panel.background = element_rect(fill = "white"),
        plot.margin = unit(c(1,1,1,1),"cm"),
        panel.border = element_rect(colour = "black", fill = NA, size = 1),
        axis.line.x = element_line(colour = "black"),
        axis.line.y = element_line(colour = "black"),
        axis.text = element_text(size = 15, colour = "black", family = "serif"),
        axis.title = element_text(size = 15, colour = "black", family = "serif"),
        legend.position = "") +
  # Set line color
  scale_colour_manual(name = "", values = c("April29" = "red")) +
  # Logarithmic pressure axis (y-axis after flip)
  scale_y_log10(
    breaks = c(100, 200, 300, 500, 700, 1000),  # Custom pressure ticks
    labels = c("100", "200", "300", "500", "700", "1000"),
    limits = c(100, 1000)  # Exclude 0 hPa (invalid for log scale)
  ) +
  # Continuous humidity bias axis (x-axis after flip)
  scale_x_continuous(
    limits = c(-30, 90),  # Match your original range
    breaks = seq(-30, 90, 10)  # Clear interval ticks
  ) +
  # Horizontal reference line at 0 bias (becomes vertical after flip)
  geom_vline(xintercept = 0, color = "black") +
  # Axis labels
  labs(x = "Relative Humidity Bias (%)", y = "Pressure Level (hPa)")

# Save the plot
ggsave("rh_profile_test.png", p, width = 5, height = 8)

Key Fixes Explained #

1. Clear Axis Mappings: I defined x = April29 (humidity bias) and y = Lev (pressure) first, then used coord_flip()—this makes it easier to track which scale applies to which axis after flipping.
2. Logarithmic Pressure Axis: scale_y_log10() creates the compact lower-atmosphere spacing you want, which is standard for meteorological vertical profiles. We exclude 0 hPa because log(0) is undefined.
3. Continuous Scales: Replaced discrete scales with continuous ones for both axes, ensuring your data values align perfectly with the plot coordinates.
4. NA Handling: Added na.rm = TRUE to geom_line and geom_point so missing values don't break the line or create empty points.

Batch Processing Adaptation #

To apply this to all your rh_profile.csv files, just swap the test data block with your file-reading loop:

library(ggplot2)

inDir <- ".."
imgDir <- ".."
inFiles <- list.files(inDir, pattern = "rh_profile.csv", full.names = TRUE)

for (inFile in inFiles) {
  dat <- read.csv(inFile, header = TRUE, sep = ",")
  
  p <- ggplot(dat, aes(x = April29, y = Lev)) +
    geom_line(aes(color = "April29"), na.rm = TRUE) +
    geom_point(shape = 4, size = 2, color = "red", na.rm = TRUE) +
    coord_flip() +
    theme(panel.background = element_rect(fill = "white"),
          plot.margin = unit(c(1,1,1,1),"cm"),
          panel.border = element_rect(colour = "black", fill = NA, size = 1),
          axis.line.x = element_line(colour = "black"),
          axis.line.y = element_line(colour = "black"),
          axis.text = element_text(size = 15, colour = "black", family = "serif"),
          axis.title = element_text(size = 15, colour = "black", family = "serif"),
          legend.position = "") +
    scale_colour_manual(name = "", values = c("April29" = "red")) +
    scale_y_log10(breaks = c(100,200,300,500,700,1000),
                  labels = c("100","200","300","500","700","1000"),
                  limits = c(100, 1000)) +
    scale_x_continuous(limits = c(-30, 90), breaks = seq(-30, 90, 10)) +
    geom_vline(xintercept = 0, color = "black") +
    labs(x = "Relative Humidity Bias (%)", y = "Pressure Level (hPa)")
  
  outImg <- paste0(imgDir, "/", gsub("\\.csv", "", basename(inFile)), ".png")
  ggsave(outImg, p, width = 5, height = 8)
}

This should give you perfectly aligned, logarithmically scaled vertical profiles as expected!

内容的提问来源于stack exchange，提问作者Lyndz