Add QC filtering to the WSPD variable for AMSR2 L2 GAASP

satpy/readers/amsr2_l2_gaasp.py

@@ -35,6 +35,21 @@
 reader appends a `_NH` and `_SH` suffix to all variable names that are
 dynamically discovered from the provided files.
 
+Data Quality Filtering
+^^^^^^^^^^^^^^^^^^^^^^
+
+Some variables can be filtered based on Quality Control (QC) variables in the
+data files. At the time of writing the only supported filtering is the "WSPD"
+variable from the "OCEAN" files using the "WSPD_QC" variable. Where the QC
+variable is non-zero the WSPD data is set to NaN. By default no filtering is
+applied. To enable filtering pass the keyword arguments to the ``Scene``::
+
+    scn = Scene(reader="amsr2_l2_gaasp", filenames=[...],
+                reader_kwargs={"filter_wind_speed": True})
+
+Note that many variables in the GAASP data files already have additional
+quality filtering applied by the algorithm.
+
 """
 
 import datetime as dt
@@ -74,6 +89,11 @@ class GAASPFileHandler(BaseFileHandler):
         "Number_of_low_rez_FOVs": 10000,
     }
 
+    def __init__(self, filename, filename_info, filetype_info, filter_wind_speed=False):
+        """Initialize file reading and store filter keyword arguments."""
+        super().__init__(filename, filename_info, filetype_info)
+        self.filter_wind_speed = filter_wind_speed
+
     @cached_property
     def nc(self):
         """Get the xarray dataset for this file."""
@@ -129,7 +149,8 @@ def _scale_data(self, data_arr, attrs):
         add_offset = attrs.pop("add_offset", 0.)
         scaling_needed = not (scale_factor == 1 and add_offset == 0)
         if scaling_needed:
-            data_arr = data_arr * scale_factor + add_offset
+            new_dtype = np.float32 if np.issubdtype(data_arr.dtype, np.integer) else data_arr.dtype.type
+            data_arr = data_arr * new_dtype(scale_factor) + new_dtype(add_offset)
         return data_arr, attrs
 
     @staticmethod
@@ -165,6 +186,7 @@ def get_dataset(self, dataid, ds_info):
         attrs = data_arr.attrs.copy()
         data_arr, attrs = self._scale_data(data_arr, attrs)
         data_arr, attrs = self._fill_data(data_arr, attrs)
+        data_arr = self._filter_by_qc(orig_var_name, data_arr)
 
         attrs.update({
             "platform_name": self.platform_name,
@@ -180,6 +202,13 @@ def get_dataset(self, dataid, ds_info):
         data_arr.attrs = attrs
         return data_arr
 
+    def _filter_by_qc(self, orig_var_name: str, data_arr: xr.DataArray) -> xr.DataArray:
+        if orig_var_name == "WSPD" and self.filter_wind_speed:
+            wspd_qc = self.nc["WSPD_QC"]
+            data_arr = data_arr.where(wspd_qc == 0)
+        return data_arr
+
+
     def _available_if_this_file_type(self, configured_datasets):
         for is_avail, ds_info in (configured_datasets or []):
             if is_avail is not None: