How can I connect two groups of lines in ggplot2, when one group is a single black line and the other group is multiple coloured lines?

I am trying to create a plot that looks like this (mockup in Powerpoint):

Unfortunately, I can only manage to get this (A):

Or this (B):

Or this (C): What I've tried:

   data < - examp = tribble(~scenario, ~years, ~value,  ~dummy,
   'Normals',' 1961-1990',  0.0000000, 'Normal',
   'Normals',' 1961-1990',  0.0000000, 'Normal',
   'Normals',' 1991-2020',  0.1965065, 'Normal',
   'SSP126 ','2011-2040 ', 0.1366428 , 'Normal',
   'SSP126 ','2041-2070 ', 0.1727662 , 'Normal',
   'SSP126 ','2071-2100 ', 0.1866775 , 'Normal',
   'SSP245 ','2011-2040 ', 0.1334094 , 'Normal',
   'SSP245 ','2041-2070 ', 0.1865576 , 'Normal',
   'SSP245 ','2071-2100 ', 0.2203371 , 'Normal',
   'SSP370 ','2011-2040 ', 0.1315721 , 'Normal',
   'SSP370 ','2041-2070 ', 0.2000836 , 'Normal',
   'SSP370 ','2071-2100 ', 0.2675299 , 'Normal',
   'SSP585 ','2011-2040 ', 0.1351562 , 'Normal',
   'SSP585 ','2041-2070 ', 0.2060075 , 'Normal',
   'SSP585 ','2071-2100 ', 0.3025003 , 'Normal')

    data %>% 
      arrange(years) %>% 
      mutate(scenario = fct_reorder(scenario, value)) %>% 
      ggplot(aes(x = years, y = value)) +
      geom_path(aes(group = scenario, colour = scenario), linewidth = 1) +
      scale_colour_manual(name = "Climate scenario",
                          values = col,
                          breaks = c("Normals", "Normals", "SSP126", "SSP245", "SSP370", "SSP585"),
                          labels = c("Normals", "", "SSP126", "SSP245", "SSP370", "SSP585")) +
      scale_x_discrete(name = "Years",
                       breaks = c("1961-1990", "1991-2020", "2011-2040", "2041-2070", "2071-2100"),
                       labels = c("1961-1990", "1991-2020", "2011-2040", "2041-2070", "2071-2100")) +
      theme(axis.text.x = element_text(angle = 35, hjust = 1))

This code snippet produces plot (A). Including group = dummy produces plot (B), and removing arrange(years) at the beginning produces plot (C). As you can see I have tried reordering the years so that I don't get the paths closing on each other, like in plot (C). I think (C) is closest, but I just need to prevent the coloured paths from closing on each other. Any assistance would be much appreciated!

Solution

You need an extra path that joins the two x axis categories "1991-2020" an "2011-2040". You can do this on the fly using the data argument of a new geom_path layer

col <- c("black", "black", "#1e73b5", "#e3923d", "#328e5c", "#7b78b2")

data %>% 
  arrange(years) %>% 
  mutate(scenario = fct_reorder(scenario, value)) %>% 
  ggplot(aes(x = years, y = value)) +
  geom_path(aes(group = scenario, colour = scenario), linewidth = 1) +
  geom_path(aes(group = 1), colour = "black", linewidth = 1,
            data = . %>% filter(years %in% c("1991-2020", "2011-2040")) %>%
              group_by(years) %>% summarise(value = mean(value))) +
  scale_colour_manual(name = "Climate scenario",
                      values = col,
                      breaks = c("Normals", "Normals", "SSP126",
                                 "SSP245", "SSP370", "SSP585"),
                      labels = c("Normals", "", "SSP126", "SSP245", 
                                 "SSP370", "SSP585")) +
  scale_x_discrete(name = "Years",
                   breaks = c("1961-1990", "1991-2020", "2011-2040", 
                              "2041-2070", "2071-2100"),
                   labels = c("1961-1990", "1991-2020", "2011-2040", 
                              "2041-2070", "2071-2100")) +
  theme_classic() +
  theme(axis.text.x = element_text(angle = 35, hjust = 1),
        panel.background = element_rect(fill = "white", color = "black"),
        legend.position = "bottom")