Change 'np.NaN' to 'np.nan' re: Numpy 2.0

'numpy.nan' is the correct spelling in Numpy 2.0. All others are deprecated.

Missed in PR #272.

nowcast/figures/shared.py

@@ -193,7 +193,7 @@ def correct_model_ssh(ssh_model, t_model, ttide):
 
 ## TODO: This should probably be in a salishsea_tools module.
 def interp_to_model_time(t_model, values, t_values):
-    """Interpolate a an array of values to model output times.
+    """Interpolate an array of values to model output times.
 
     Strategy: Convert times to seconds past a reference value and use those
     as the independent variable in interpolation.
@@ -215,7 +215,7 @@ def interp_to_model_time(t_model, values, t_values):
     t_values_wrt_epoch = np.array([(t - epoch).total_seconds() for t in t_values])
     t_model_wrt_epoch = np.array([(t - epoch).total_seconds() for t in t_model])
     return np.interp(
-        t_model_wrt_epoch, t_values_wrt_epoch, values, left=np.nan, right=np.NaN
+        t_model_wrt_epoch, t_values_wrt_epoch, values, left=np.nan, right=np.nan
     )