id,node_id,number,title,user,state,locked,assignee,milestone,comments,created_at,updated_at,closed_at,author_association,active_lock_reason,draft,pull_request,body,reactions,performed_via_github_app,state_reason,repo,type
732910109,MDU6SXNzdWU3MzI5MTAxMDk=,4554,Unexpected chunking of 3d DataArray in `polyfit()`,26591824,open,0,,,3,2020-10-30T06:07:34Z,2021-04-19T15:44:07Z,,CONTRIBUTOR,,,,"<!-- Please include a self-contained copy-pastable example that generates the issue if possible.

Please be concise with code posted. See guidelines below on how to provide a good bug report:

- Craft Minimal Bug Reports: http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports
- Minimal Complete Verifiable Examples: https://stackoverflow.com/help/mcve

Bug reports that follow these guidelines are easier to diagnose, and so are often handled much more quickly.
-->

**What happened**:
When running `polyfit()` on a 3d chunked xarray DataArray, the output is chunked differently than the input array. 

**What you expected to happen**:
I expect the output to have the same chunking as the input.

**Minimal Complete Verifiable Example**:
(from @rabernat in [https://github.com/xgcm/xrft/issues/116](https://github.com/xgcm/xrft/issues/116))

Example: number of chunks decreases
```python
import dask.array as dsa
import xarray as xr

nz, ny, nx = (10, 20, 30)
data = dsa.ones((nz, ny, nx), chunks=(1, 5, nx))
da = xr.DataArray(data, dims=['z', 'y', 'x'])
da.chunks
# -> ((1, 1, 1, 1, 1, 1, 1, 1, 1, 1), (5, 5, 5, 5), (30,))

pf = da.polyfit('x', 1)
pf.polyfit_coefficients.chunks
# -> ((1, 1, 1, 1, 1, 1, 1, 1, 1, 1), (20,), (30,))
# chunks on the y dimension have been consolidated!

pv = xr.polyval(da.x, pf.polyfit_coefficients).transpose('z', 'y', 'x')
pv.chunks
# -> ((1, 1, 1, 1, 1, 1, 1, 1, 1, 1), (20,), (30,))
# and this propagates to polyval

# align back against the original data
(da - pv).chunks
# -> ((1, 1, 1, 1, 1, 1, 1, 1, 1, 1), (5, 5, 5, 5), (30,))
# hides the fact that we have chunk consolidation happening upstream
```

Example: number of chunks increases
```python
nz, ny, nx = (6, 10, 4)
data = dsa.ones((nz, ny, nx), chunks=(2, 10, 2))
da = xr.DataArray(data, dims=['z', 'y', 'x'])
da.chunks
# -> ((2, 2, 2), (10,), (2, 2))

pf = da.polyfit('y', 1)
pf.polyfit_coefficients.chunks
# -> ((2,), (1, 1, 1, 1, 1, 1), (4,))

pv = xr.polyval(da.y, pf.polyfit_coefficients).transpose('z', 'y', 'x')
pv.chunks
# -> ((1, 1, 1, 1, 1, 1), (10,), (4,))

(da - pv).chunks
# -> ((1, 1, 1, 1, 1, 1), (10,), (2, 2))
```
(This discussion started in [https://github.com/xgcm/xrft/issues/116](https://github.com/xgcm/xrft/issues/116) with @rabernat and @navidcy.)

**Environment**:

Running on Pangeo Cloud

<details><summary>Output of <tt>xr.show_versions()</tt></summary>


INSTALLED VERSIONS
------------------
commit: None
python: 3.8.6 | packaged by conda-forge | (default, Oct  7 2020, 19:08:05) 
[GCC 7.5.0]
python-bits: 64
OS: Linux
OS-release: 4.19.112+
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: C.UTF-8
LANG: C.UTF-8
LOCALE: en_US.UTF-8
libhdf5: 1.10.6
libnetcdf: 4.7.4

xarray: 0.16.1
pandas: 1.1.3
numpy: 1.19.2
scipy: 1.5.2
netCDF4: 1.5.4
pydap: installed
h5netcdf: 0.8.1
h5py: 2.10.0
Nio: None
zarr: 2.4.0
cftime: 1.2.1
nc_time_axis: 1.2.0
PseudoNetCDF: None
rasterio: 1.1.7
cfgrib: 0.9.8.4
iris: None
bottleneck: 1.3.2
dask: 2.30.0
distributed: 2.30.0
matplotlib: 3.3.2
cartopy: 0.18.0
seaborn: None
numbagg: None
pint: 0.16.1
setuptools: 49.6.0.post20201009
pip: 20.2.3
conda: None
pytest: 6.1.1
IPython: 7.18.1
sphinx: 3.2.1


</details>","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/4554/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,,13221727,issue