How to draw a shapefile with ggpolar

I want to use ggpolar to plot a polar projection map of raster data, but the outermost pixels tend to obscure the land boundaries. Therefore, I want to add a local shapefile, but I failed to draw it using geom_sf in ggpolar. Even after converting the shapefile to the appropriate projection, it still couldn't be plotted. Could you please suggest a way to solve this problem? Below is my code:

continent <- st_read('F:/data/vector/seven_continent/continents/continent_30_90.shp')
ggpolar(pole = "N", max.lat = 90, min.lat = 30, 
    land.fill.colour = "#D9D9D9", land.outline.colour = "#343434", 
    plt.lat.labels = FALSE,size = 0.8) +
ggtitle("A") +
theme(plot.title = element_text(hjust = 0, family = "Arial", size = 18,face = "bold")) 
+
geom_tile(data = change_ratio_df, aes(x = lon, y = lat, fill = change_ratio)) + 
geom_point(data = selected_points, aes(x = lon, y = lat), color = "#343434", size =0.5) +
scale_fill_stepsn(colors = custom_colors,
                breaks = break_points, 
                labels = c("-5", "-4", "-3", "-2", "-1", "0", "1", "2", "3", "4", "5"),
                #labels = c("-0.005", "-0.004", "-0.003", "-0.002", "-0.001", "0", "0.001", "0.002", "0.003", "0.004", "0.005"),
                limits = c(-6, 6),
                #limits = c(-0.006, 0.006),
                name = expression(AE~('%'~ decade^{-1})),
                #name = NULL,
                na.value = "transparent", 
                guide = guide_colorbar(barwidth = 0.8, barheight = unit(10, "cm"), direction = "vertical",
                                       title.position = "right", title.hjust=0.5, label.position = "right", 
                                       title.theme = element_text(family = "Arial", size = 16, angle = 90),
                                       label.theme = element_text(family = "Arial", size = 12)))+
geom_sf(data = continent, fill = NA, color = "#343434") +
coord_sf(crs = sf::st_crs("ESRI:102016"))+
theme(
plot.background = element_rect(fill = NA, color = NA),  
panel.background = element_rect(fill = NA, color = NA), 
legend.position = "none",
legend.key.width = unit(2, "line"),
legend.key.height = unit(0.5, "line"),
legend.title = element_text(family = "Arial", size = 12),
legend.text = element_text(size = 12),
axis.title = element_blank(),
axis.text = element_text(family = "Arial", size = 12),
#panel.grid.major = element_line(size=0.25,linetype='dashed',colour="black"),
axis.ticks = element_blank()
)

Here's my shapefile: https://drive.google.com/drive/folders/1tuxes8oPd_F2OmWYfPyqPjuITE0bbBBp?usp=drive_link

The following figure is without added native shapefile：

The following figure is drawn by the untransformed projection of the shapefile： Code Error：Coordinate system already present. Adding new coordinate system, which will replace the existing one.

The following figure is plotted after the shapefile transforms the projection： Code Error：Coordinate system already present. Adding new coordinate

I think the easiest way to handle this issue is to convert your continent sf object to a dataframe and use ggplot2::geom_path() to plot it.

First, load required packages, create some example raster data, and load shapefile. You can safely ignore the warning that this code produces in this use case.

# remotes::install_github("EarthSystemDiagnostics/grfxtools")

library(geodata)
library(terra)
library(sf)
library(ggplot2)
library(grfxtools)

# Sample SpatRaster
r <- worldclim_global("tavg", 10, ".")
r <- r[[1]]
r <- crop(r, ext(-180, 180, 30, 90))

# Convert r to df
r_df <- as.data.frame(r, xy = TRUE)

# Load continent shapefile (assumed to be WGS84/EPSG:4326), and
# convert to multilinestring
continent <- st_read("F:/data/vector/seven_continent/continents/continent_30_90.shp") |>
  st_cast("MULTILINESTRING")

# Convert to continent sf to df
line_df <- as.data.frame(st_coordinates(st_cast(continent, "LINESTRING")))

Plot:

ggpolar(pole = "N",
        min.lat = 30, 
        max.lat = 90,
        land.fill.colour = "#D9D9D9",
        land.outline.colour = NA, 
        plt.lat.labels = FALSE,
        size = 0.8) +
  geom_tile(data = r_df,
            aes(x, y, fill = wc2.1_10m_tavg_01)) +
  geom_path(data = line_df,
            aes(X, Y, group = L1),
            colour = "#343434") +
  theme(
    plot.background = element_rect(fill = NA, color = NA),  
    panel.background = element_rect(fill = NA, color = NA), 
    legend.position = "none",
    legend.key.width = unit(2, "line"),
    legend.key.height = unit(0.5, "line"),
    legend.title = element_text(family = "Arial", size = 12),
    legend.text = element_text(size = 12),
    axis.title = element_blank(),
    axis.text = element_text(family = "Arial", size = 12),
    axis.ticks = element_blank()
  )

In the event that the linked data in the question are deleted, to make this answer a reprex here is an equivalent option using rnaturalearth data:

library(rnaturalearth) # You may be prompted to install rnaturalearthdata
library(dplyr)

# Sample sf
sf_use_s2(FALSE)

poly_sf <- ne_countries(scale = 50) |>
  summarise(geometry = st_union(geometry))

# Ensure polygon sf cropped to intended plot extent, convert to MULTILINESTRING
line_sf <- poly_sf |>
  st_crop(ext(-180, 180, 30, 90)) |>
  st_cast("MULTILINESTRING")

sf_use_s2(TRUE)

# Convert line_sf to df
line_df <- as.data.frame(st_coordinates(st_cast(continent, "LINESTRING")))