I cannot put geom_abline in the plot legend

I struggle with putting the abline in the legend. It works fine with the points and their colour, but as far as I succeed is the six lines overlapping with the points... Can I just somehow add another legend element with the black line labeled as "LMWL"? I know that I can post-edit this in Corel or any software but I would like to do it in the code...

ggplot(df_merge_iso, aes(x = d18O, y = d2H, color = category, shape = category, label = month)) +
  labs(
    x = expression(paste(delta^{18}, "O (\u2030)")),
    y = expression(paste(delta^{2}, "H (\u2030)"))) +
  geom_point(size = 1, stroke = 1) +
  theme_minimal() +
  theme(
    axis.text = element_text(size = 12, color = "black"),
    axis.title = element_text(size = 12, color = "black")
  ) +
  scale_color_manual(
    name = "",
    values = c(
      "A" = "#0015f0",
      "B" = "#6AC4FF",
      "C" = "#4B8CEB",
      "D" = "#FF7E5F",
      "E" = "#FFB300",
      "F" = "#E75007"
    )) +
  scale_shape_manual(
    name = "",
    values = c(
      "A" = 2,  
      "B" = 1,            
      "C" = 21,            
      "D" = 3,            
      "E" = 4,            
      "F" = 6  
    )) +
  geom_text_repel(
    vjust = -1, hjust = 0.5, size = 2, show.legend = FALSE,
    box.padding = 0.2
  ) +
  geom_abline(slope = lmwl_slope_wgt, intercept = lmwl_intercept_wgt, 
              colour = "black", linetype = "solid", size = 0.5)

If you want a legend you have to map on aesthetics. In case you want to add a legend entry to the already present color and shape legend you could do by mapping e.g. the constant "LMWL" on color and shape in geom_abline which also requires to move intercept and slope inside aes(). Additionally you have to provide a color and a shape for LMWL in the scales.

Using some fake example data:

library(ggrepel)
#> Loading required package: ggplot2
library(ggplot2)

df_merge_iso <- data.frame(
  d18O = 1:6,
  d2H = 1:6,
  category = LETTERS[1:6],
  month = letters[1:6]
)

lmwl_slope_wgt <- .4
lmwl_intercept_wgt <- 2

p <- ggplot(df_merge_iso, aes(x = d18O, y = d2H, color = category, shape = category, label = month)) +
  labs(
    x = expression(paste(delta^{
      18
    }, "O (\u2030)")),
    y = expression(paste(delta^{
      2
    }, "H (\u2030)"))
  ) +
  geom_point(size = 1, stroke = 1) +
  theme_minimal() +
  theme(
    axis.text = element_text(size = 12, color = "black"),
    axis.title = element_text(size = 12, color = "black")
  ) +
  scale_color_manual(
    name = "",
    values = c(
      "A" = "#0015f0",
      "B" = "#6AC4FF",
      "C" = "#4B8CEB",
      "D" = "#FF7E5F",
      "E" = "#FFB300",
      "F" = "#E75007",
      LMWL = "black"
    )
  ) +
  scale_shape_manual(
    name = "",
    values = c(
      "A" = 2,
      "B" = 1,
      "C" = 21,
      "D" = 3,
      "E" = 4,
      "F" = 6,
      LMWL = 1
    )
  ) +
  geom_text_repel(
    vjust = -1, hjust = 0.5, size = 2, show.legend = FALSE,
    box.padding = 0.2
  )

p +
  geom_abline(
    aes(
      color = "LMWL", shape = "LMWL",
      slope = lmwl_slope_wgt, intercept = lmwl_intercept_wgt
    ), 
    linetype = "solid", size = 0.5
  )

If you want a separate legend you could fake one by mapping e.g. on the linetype aes:

p +
  geom_abline(
    aes(
      linetype = "LMWL",
      slope = lmwl_slope_wgt, intercept = lmwl_intercept_wgt
    ),
    size = 0.5
  ) +
  scale_linetype_manual(name = NULL, values = "solid")