Irregular regridding gives cube which is not correctly read by iris (for specific dataset)

Describe the bug For a specific dataset (see recipe below), the regridded cube cannot be correctly read by iris.

Recipe


preprocessors:
  regrid:
    regrid:
      target_grid: 2x2
      scheme: linear

diagnostics:
  diag_x_zhai:
    variables:
      tos:
        preprocessor: regrid
        exp: historical
        ensemble: r1i1p1
        project: CMIP5
        mip: Omon
        start_year: 1980
        end_year: 2004
        reference_dataset: ''
    additional_datasets:
       - {dataset: inmcm4}
    scripts: null

Output

This results in the following output fiile:

netcdf CMIP5_inmcm4_Omon_historical_r1i1p1_tos_1980-2004 {
dimensions:
        time = 300 ;
        lat = 40 ;
        lon = 180 ;
        bnds = 2 ;
variables:
        float tos(time, lat, lon) ;
                tos:_FillValue = 1.e+20f ;
                tos:standard_name = "sea_surface_temperature" ;
                tos:long_name = "Sea Surface Temperature" ;
                tos:units = "K" ;
                tos:cell_methods = "time: mean (interval: 1 month)" ;
                tos:grid_mapping = "rotated_latitude_longitude" ;
        int rotated_latitude_longitude ;
                rotated_latitude_longitude:grid_mapping_name = "rotated_latitude_longitude" ;
                rotated_latitude_longitude:grid_north_pole_latitude = 70. ;
                rotated_latitude_longitude:grid_north_pole_longitude = 100. ;
                rotated_latitude_longitude:north_pole_grid_longitude = -80. ;
        double time(time) ;
                time:axis = "T" ;
                time:bounds = "time_bnds" ;
                time:units = "days since 1850-1-1 00:00:00" ;
                time:standard_name = "time" ;
                time:long_name = "time" ;
                time:calendar = "365_day" ;
        double time_bnds(time, bnds) ;
        double lat(lat) ;
                lat:axis = "Y" ;
                lat:bounds = "lat_bnds" ;
                lat:units = "degrees_north" ;
                lat:standard_name = "latitude" ;
        double lat_bnds(lat, bnds) ;
        double lon(lon) ;
                lon:axis = "X" ;
                lon:bounds = "lon_bnds" ;
                lon:units = "degrees_east" ;
                lon:standard_name = "longitude" ;
        double lon_bnds(lon, bnds) ;

// global attributes:
                :_NCProperties = "version=2,netcdf=4.6.2,hdf5=1.10.5" ;
                :associated_files = "baseURL: http://cmip-pcmdi.llnl.gov/CMIP5/dataLocation gridspecFile: gridspec_fx_inmcm4_historical_r0i0p0.nc areacello: areacello_fx_inmcm4_historical_r0i0p0.nc" ;
                :branch_time = 0. ;
                :cmor_version = "2.0.0" ;
                :comment = "\"this may differ from \"\"surface temperature\"\" in regions of sea ice.\"" ;
                :contact = "Evgeny Volodin, volodin@inm.ras.ru,INM RAS, Gubkina 8, Moscow, 119333 Russia,+7-495-9383904" ;
                :experiment = "Historical" ;
                :experiment_id = "historical" ;
                :forcing = "N/A" ;
                :frequency = "mon" ;
                :initialization_method = 1 ;
                :institute_id = "INM" ;
                :institution = "INM (Institute for Numerical Mathematics,  Moscow, Russia)" ;
                :model_id = "inmcm4" ;
                :modeling_realm = "ocean" ;
                :original_name = "tos" ;
                :parent_experiment = "pre-industrial control" ;
                :parent_experiment_id = "piControl" ;
                :physics_version = 1 ;
                :product = "output" ;
                :project_id = "CMIP5" ;
                :realization = 1 ;
                :references = "Volodin, Diansky, Gusev 2010. Climate model INMCM4.0. Izvestia RAS. Atmospheric and oceanic physics, V.46, N4, in print." ;
                :source = "inmcm4 (2009)" ;
                :table_id = "Table Omon (12 May 2010) f2afe576fb73a3a11aaa3cc8f2e62cf3" ;
                :title = "inmcm4 model output prepared for CMIP5 Historical" ;
                :Conventions = "CF-1.7" ;

When read with iris, this gives

sea_surface_temperature / (K)       (time: 300; -- : 40; -- : 180)
     Dimension coordinates:
          time                           x         -        -
     Attributes:
          Conventions: CF-1.7
          associated_files: baseURL: http://cmip-pcmdi.llnl.gov/CMIP5/dataLocation gridspecFile: gridspec_fx_inmcm4_historical_r0i0p0.nc...
          branch_time: 0.0
          cmor_version: 2.0.0
          comment: "this may differ from ""surface temperature"" in regions of sea ice."
          contact: Evgeny Volodin, volodin@inm.ras.ru,INM RAS, Gubkina 8, Moscow, 119333 ...
          experiment: Historical
          experiment_id: historical
          forcing: N/A
          frequency: mon
          initialization_method: 1
          institute_id: INM
          institution: INM (Institute for Numerical Mathematics,  Moscow, Russia)
          model_id: inmcm4
          modeling_realm: ocean
          original_name: tos
          parent_experiment: pre-industrial control
          parent_experiment_id: piControl
          physics_version: 1
          product: output
          project_id: CMIP5
          realization: 1
          references: Volodin, Diansky, Gusev 2010. Climate model INMCM4.0. Izvestia RAS. Atmospheric...
          source: inmcm4 (2009)
          table_id: Table Omon (12 May 2010) f2afe576fb73a3a11aaa3cc8f2e62cf3
          title: inmcm4 model output prepared for CMIP5 Historical
     Cell methods:
          mean: time (1 month)

As you can see, the latitude and longitude coordinate are not present in the cube. I think the problem in the netcdf file is the rotated_latitude_longitude variable, which somehow tells iris to use a rotated grid, but it fails to do so because only a regular grid is present.

This is not also not an issue of the original data, which lives on a rotated grid. Also, iris needs the rotated_latitude_longitude variable in the original file to correctly read it in the first place.

Thus, it is necessary to somehow get rid of this variable after the irregular regridding.

ESMValGroup / ESMValCore

Irregular regridding gives cube which is not correctly read by iris (for specific dataset) #479