Fix: wrong display of missing data in World plots

Author: heptaflar

README.md

@@ -133,5 +133,5 @@ deviations implied by constraints on the error probability distribution and the
 Examples above illustrate the mean, the standard deviation and the minimum-to-maximum
 range of normalized simulated errors for an ensemble of ten variants of global monthly
 sea surface temperature (January 2000). The standard deviation and range of normalized
-errors is reduced at the North Pole and in the Weddell and Ross Sea due to constraints
-imposed by the freezing point of seawater. 
+errors is reduced at the North Pole and in the Weddell Sea due to constraints imposed
+by the freezing point of seawater. 

doc/map-mean-Sea surface temperature.png

@@ -50,7 +50,7 @@ def plot(
         y: Literal["lat", "latitude"] | None = None,
     ) -> Figure:
         if projection is None:
-            projection = self.projection
+            projection = self.sinusoidal
         fig, ax = plt.subplots(
             figsize=plot_size,
             subplot_kw={"projection": projection},
@@ -99,15 +99,32 @@ def land(self):
         )
 
     @property
-    def projection(self):
-        """Returns the default projection."""
+    def interrupted_goode_homolosine(self):
+        """
+        Returns the Interrupted Goode Homolosine projection
+        for ocean.
+
+        Plots using this projection do not show missing data
+        transparent (white).
+        """
         from cartopy.crs import InterruptedGoodeHomolosine
 
         return InterruptedGoodeHomolosine(
             central_longitude=-160.0,
             emphasis="ocean",
         )
 
+    @property
+    def sinusoidal(self):
+        """
+        Returns the sinusoidal projection.
+        """
+        from cartopy.crs import Sinusoidal
+
+        return Sinusoidal(
+            central_longitude=0.0,
+        )
+
     @property
     def transform(self):
         """Returns the default transform."""