[pyicon_dwd] Use ERA5 data as reference for atmospheric plots: part 1

('part 1': all atmospheric plots created by qp_driver.py)

More details:

*) Tools are added to download and prepare necessary ERA5 data.
   tools/get_era5_pyicon.sh
   tools/prep_era5_pyicon.py
   The latter is written in xarray and does the following:
   -- interpolates the 0.25x0.25 data to arbitrary lat-lon
      resolution
   -- calculates time-mean maps for arbitrary time period
      for 2D fields and for 3D fields on selected pressure
      levels
   -- calculates zonal-mean time-mean lat-pressure
      cross-sections for arbitrary time period
   -- calculates global-mean (cell-area weighted) times
      series of selected variables

*) pyicon/quickplots/qp_driver.py is updated so that ERA5 is
   used as reference for bias plots. Many new plots are added
   as compared to the previous case, where only a limited set
   of variables were available from ERA Interim the format that
   was eatable by pyicon.

*) pyicon/quickplots/pyicon_quickplots.py is updated in order
   to show ERA5 data also in times series plots for the atmosphere.
   The changes are done under 'if use_tave_int_for_ts:' statements
   so that it does not change anything for ocean plots.

*) tools/run_qp_driver_*_dwd.sh are updated by the path and
   file name of the ERA5 reference data.

*) config_ckdtree/config_ckdtree_r2b4_atm_*_dwd.py are extended
   with a 1.5x1.5 grid. This is the resolution we use with R2B4.
   (also for a validation of R2B4 simulations, ERA5 data are
    interpolated to a 1.5x1.5 grid by tools/prep_era5_pyicon.py)

Additionally a bugfix is also included:

*) in qp_timeseries (pyicon/quickplots/pyicon_quickplots.py)
   the case have been fixed when use_tave_int_for_ts=True and
   vars_plot is a list of variables (and not just a single variable).
   times_plot had to be redefined within the loop over the vars_plot
   loop.