explicitly turn on flask sampling in ctdas

da/tm5/observationoperator.py

@@ -181,7 +181,8 @@ class TM5ObservationOperator(ObservationOperator):
                     'jobstep.length': 'inf',
                     'ct.params.input.dir': self.DaCycle['dir.input'],
                     'ct.params.input.file': os.path.join(self.DaCycle['dir.input'], 'parameters'),
-                    'output.flask.infile': self.DaCycle['ObsOperator.inputfile']
+                    'output.flask.infile': self.DaCycle['ObsOperator.inputfile'] ,
+                    'output.flask': 'True'
                     }
 
         if self.DaCycle['time.restart']:  # If this is a restart from a previous cycle, the TM5 model should do a restart