如何绘制含两组分组与误差棒的堆叠柱状图？

问题说明 #

需基于土壤数据集生成满足以下要求的可视化图表：

• X轴为Time（包含May 21、July 21、July 22三个水平）
• 每个Time刻度下，按Treatment（T1-T4）分组并排展示柱子
• 每个柱子内部按Fraction堆叠，要求Fine earth堆叠在Plant remains之上
• 展示误差棒（均值±标准误）

步骤1：加载依赖包并预处理数据 #

使用dplyr计算分组统计量，同时生成堆叠总高度数据（用于总误差棒展示）：

library(dplyr)
library(ggplot2)

# 计算分组统计：均值、标准差、样本量、标准误
summ_data <- datas %>%
  group_by(Time, Treatment, Fraction) %>%
  summarise(
    mean_corg = mean(Corg, na.rm = TRUE),
    sd_corg = sd(Corg, na.rm = TRUE),
    n = n(),
    se_corg = sd_corg / sqrt(n),
    .groups = "drop"
  ) %>%
  # 确保Fine earth堆叠在Plant remains上方
  arrange(Time, Treatment, desc(Fraction)) %>%
  # 计算每个Time-Treatment组合的堆叠总高度
  group_by(Time, Treatment) %>%
  mutate(total_mean = cumsum(mean_corg)) %>%
  ungroup()

步骤2：绘制单块误差棒的堆叠分组图 #

此方案展示每个堆叠部分的独立误差棒：

ggplot(summ_data, aes(x = Time, y = mean_corg, fill = Fraction)) +
  # 实现分组并排+内部堆叠的柱子
  geom_col(aes(group = Treatment), 
           position = position_dodge2(width = 0.8, preserve = "single"), 
           width = 0.7) +
  # 添加堆叠块的误差棒，位置与柱子对齐
  geom_errorbar(aes(ymin = mean_corg - se_corg, ymax = mean_corg + se_corg),
                position = position_dodge2(width = 0.8, preserve = "single"),
                width = 0.2, color = "black") +
  # 使用指定配色方案
  scale_fill_manual(values = c("Fine earth" = "darkgoldenrod4", "Plant remains" = "darkolivegreen4")) +
  # 设置坐标轴标签与主题
  labs(x = "采样时间", y = "Corg含量", fill = "组分") +
  theme_classic()

步骤3：绘制总误差棒的堆叠分组图 #

若仅需展示每个Treatment柱子的总误差（堆叠顶部），使用以下代码：

ggplot(summ_data, aes(x = Time, fill = Fraction)) +
  # 分组并排+堆叠柱子
  geom_col(aes(y = mean_corg, group = Treatment), 
           position = position_dodge2(width = 0.8, preserve = "single"), 
           width = 0.7) +
  # 在堆叠顶部（Fine earth）添加总误差棒
  geom_errorbar(data = summ_data %>% filter(Fraction == "Fine earth"),
                aes(y = total_mean, ymin = total_mean - se_corg, ymax = total_mean + se_corg, group = Treatment),
                position = position_dodge2(width = 0.8, preserve = "single"),
                width = 0.2, color = "black") +
  scale_fill_manual(values = c("Fine earth" = "darkgoldenrod4", "Plant remains" = "darkolivegreen4")) +
  labs(x = "采样时间", y = "Corg含量", fill = "组分") +
  theme_classic()

关键说明 #

• 使用position_dodge2(preserve = "single")实现：同一Time下的不同Treatment柱子并排，每个柱子内部按Fraction堆叠
• 误差棒的position参数需与柱子完全一致，确保位置对齐
• 通过arrange(desc(Fraction))强制Fine earth处于堆叠上层

内容的提问来源于stack exchange，提问作者marinaaa2809