If "chunks=None" is set in open_mfdataset, it is changed to "chunks={}" before being passed to "_dataset_from_backend_dataset"

What happened?

Using the grib2io engine, but have to use on a system that currently doesn't allow dask to be installed. Looking through the code I think that setting "chunks=None" would work to not use dask, but on https://github.com/pydata/xarray/blob/0f4e99d036b0d6d76a3271e6191eacbc9922662f/xarray/backends/api.py#L968 "chunks=None" is converted to "chunks={}".

This means that at this test https://github.com/pydata/xarray/blob/0f4e99d036b0d6d76a3271e6191eacbc9922662f/xarray/backends/api.py#L339 for "chunks is None" will never be true and the dask code path will always run.

The example below uses the rasterio engine because I could open publicly available files from S3. The rasterio engine gives the same error as the grib2io engine.

What did you expect to happen?

Expected open_mfdataset to work without dask installed.

Minimal Complete Verifiable Example

# Have to create an environment that doesn't include dask.  For example:
#     conda create -n xarrayenv -c conda-forge xarray rioxarray
#     conda activate xarrayenv

import xarray as xr
import os

os.environ["AWS_NO_SIGN_REQUEST"] = "YES"

ds = xr.open_mfdataset(
    [
        "/vsis3/noaa-nbm-grib2-pds/blend.20230401/02/core/blend.t02z.core.f003.co.grib2",
        "/vsis3/noaa-nbm-grib2-pds/blend.20230401/02/core/blend.t02z.core.f004.co.grib2",
    ],
    engine="rasterio",
    chunks=None,
)

# Traceback (most recent call last):                                                                                                                                         
#   File "/home/tim/test.py", line 6, in <module>                                                                                                                            
#     ds = xr.open_mfdataset(                                                                                                                                                
#          ^^^^^^^^^^^^^^^^^^
#   File "/home/tim/anaconda3/envs/xarray/lib/python3.11/site-packages/xarray/backends/api.py", line 982, in open_mfdataset
#     datasets = [open_(p, **open_kwargs) for p in paths]
#                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
#   File "/home/tim/anaconda3/envs/xarray/lib/python3.11/site-packages/xarray/backends/api.py", line 982, in <listcomp>
#     datasets = [open_(p, **open_kwargs) for p in paths]
#                 ^^^^^^^^^^^^^^^^^^^^^^^
#   File "/home/tim/anaconda3/envs/xarray/lib/python3.11/site-packages/xarray/backends/api.py", line 531, in open_dataset
#     ds = _dataset_from_backend_dataset(
#          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
#   File "/home/tim/anaconda3/envs/xarray/lib/python3.11/site-packages/xarray/backends/api.py", line 342, in _dataset_from_backend_dataset
#     ds = _chunk_ds(
#          ^^^^^^^^^^
#   File "/home/tim/anaconda3/envs/xarray/lib/python3.11/site-packages/xarray/backends/api.py", line 302, in _chunk_ds
#     from dask.base import tokenize
# ModuleNotFoundError: No module named 'dask'

MVCE confirmation

• Minimal example — the example is as focused as reasonably possible to demonstrate the underlying issue in xarray.
• Complete example — the example is self-contained, including all data and the text of any traceback.
• Verifiable example — the example copy & pastes into an IPython prompt or Binder notebook, returning the result.
• New issue — a search of GitHub Issues suggests this is not a duplicate.

Relevant log output

No response

Anything else we need to know?

No response

Environment

Start in xarray/backends/api.py at open_mfdataset, _dataset_from_backend_dataset, and the chunks handling around the referenced lines; reproduce the MVCE in an environment without dask. Trace why chunks=None becomes chunks={} and verify that open_mfdataset with chunks=None works without importing dask, including coverage for this path.