diff --git a/RELEASE_NOTES.md b/RELEASE_NOTES.md
index 808b70eeb..1e416404a 100644
--- a/RELEASE_NOTES.md
+++ b/RELEASE_NOTES.md
@@ -2,6 +2,7 @@
 
 The next release must be bumped to v3.0.0.
 
+- [#899](https://github.com/IAMconsortium/pyam/pull/899) Add `to_netcdf()` method
 - [#896](https://github.com/IAMconsortium/pyam/pull/896) Add `sort_data()` method
 - [#896](https://github.com/IAMconsortium/pyam/pull/896) Sort columns of `timeseries()` with mixed time domain
 - [#893](https://github.com/IAMconsortium/pyam/pull/893) No sorting of timeseries data on initialization or append
diff --git a/pyam/core.py b/pyam/core.py
index ab8e21fe0..7aeeecf0f 100755
--- a/pyam/core.py
+++ b/pyam/core.py
@@ -11,6 +11,8 @@
 import pandas as pd
 from pandas.api.types import is_integer
 
+from pyam.netcdf import to_xarray
+
 try:
     from datapackage import Package
 
@@ -2513,7 +2515,7 @@ def to_datapackage(self, path):
         Parameters
         ----------
         path : string or path object
-            any valid string path or :class:`pathlib.Path`
+            Any valid string path or :class:`pathlib.Path`.
         """
         if not HAS_DATAPACKAGE:
             raise ImportError("Required package `datapackage` not found!")
@@ -2533,6 +2535,30 @@ def to_datapackage(self, path):
         # return the package (needs to reloaded because `tmp` was deleted)
         return Package(path)
 
+    def to_netcdf(self, path):
+        """Write object to a NetCDF file
+
+        Parameters
+        ----------
+        path : string or path object
+            Any valid string path or :class:`pathlib.Path`.
+
+        See Also
+        --------
+        pyam.read_netcdf
+        """
+        self.to_xarray().to_netcdf(path)
+
+    def to_xarray(self):
+        """Convert object to an :class:`xarray.Dataset`
+
+        Returns
+        -------
+        :class:`xarray.Dataset`
+        """
+        df = swap_year_for_time(self) if self.time_domain == "year" else self
+        return to_xarray(df._data, df.meta)
+
     def load_meta(self, path, sheet_name="meta", ignore_conflict=False, **kwargs):
         """Load 'meta' indicators from file
 
diff --git a/pyam/netcdf.py b/pyam/netcdf.py
index a69f963ad..cf736ab19 100644
--- a/pyam/netcdf.py
+++ b/pyam/netcdf.py
@@ -3,6 +3,8 @@
 import numpy as np
 import pandas as pd
 
+from pyam.index import get_index_levels
+
 try:
     import xarray as xr
 
@@ -10,9 +12,10 @@
 except ModuleNotFoundError:
     xr = None
     HAS_XARRAY = False
-from pyam.core import IamDataFrame
 from pyam.utils import IAMC_IDX, META_IDX
 
+NETCDF_IDX = ["time", "model", "scenario", "region"]
+
 
 def read_netcdf(path):
     """Read timeseries data and meta indicators from a netCDF file
@@ -26,11 +29,15 @@ def read_netcdf(path):
     ----------
     :class:`IamDataFrame`
 
+    See Also
+    --------
+    pyam.IamDataFrame.to_netcdf
     """
+    from pyam import IamDataFrame
+
     if not HAS_XARRAY:
-        raise ModuleNotFoundError("Reading netcdf files requires 'xarray'")
+        raise ModuleNotFoundError("Reading netcdf files requires 'xarray'.")
     _ds = xr.open_dataset(path)
-    NETCDF_IDX = ["time", "model", "scenario", "region"]
     _list_variables = [i for i in _ds.to_dict()["data_vars"].keys()]
 
     # Check if the time coordinate is years (integers) or date time-format
@@ -86,3 +93,45 @@ def read_netcdf(path):
         data,
         meta=_ds[_meta].to_dataframe().replace("nan", np.nan) if _meta else None,
     )
+
+
+def to_xarray(data_series: pd.Series, meta: pd.DataFrame):
+    """Convert timeseries data and meta indicators to an xarray Dataset
+
+    Returns
+    -------
+    :class:`xarray.Dataset`
+
+    """
+    if not HAS_XARRAY:
+        raise ModuleNotFoundError("Converting to xarray requires 'xarray'.")
+
+    dataset = xr.Dataset()
+
+    # add timeseries data-variables
+    for variable, _variable_data in data_series.groupby("variable"):
+        unit = get_index_levels(_variable_data, "unit")
+
+        if len(unit) > 1:
+            raise ValueError(
+                "Cannot write to xarray for non-unique units in '" + variable + "'."
+            )
+
+        dataset[variable] = xr.DataArray(
+            _variable_data.droplevel(["variable", "unit"]).to_xarray(),
+        )
+        dataset[variable].attrs = {
+            "unit": unit[0],
+            "long_name": variable,
+        }
+
+    # add meta indicators as data-variables
+    for meta_indicator, meta_data in meta.items():
+        meta_data = meta_data.replace(np.nan, "nan")
+        dataset[meta_indicator] = xr.DataArray(
+            meta_data.to_xarray(),
+            dims=META_IDX,
+            name=meta_indicator,
+        )
+
+    return dataset
diff --git a/pyam/time.py b/pyam/time.py
index 34077d9bc..356b7d146 100644
--- a/pyam/time.py
+++ b/pyam/time.py
@@ -5,7 +5,7 @@
 from pyam.logging import raise_data_error
 
 
-def swap_time_for_year(df, inplace, subannual=False):
+def swap_time_for_year(df, inplace=False, subannual=False):
     """Internal implementation to swap 'time' domain to 'year' (as int)"""
     if not df.time_col == "time":
         raise ValueError("Time domain must be datetime to use this method")
@@ -49,7 +49,7 @@ def swap_time_for_year(df, inplace, subannual=False):
         return ret
 
 
-def swap_year_for_time(df, inplace):
+def swap_year_for_time(df, inplace=False):
     """Internal implementation to swap 'year' domain to 'time' (as datetime)"""
 
     if not df.time_col == "year":
diff --git a/tests/test_io.py b/tests/test_io.py
index 4c97d4615..8f117297d 100644
--- a/tests/test_io.py
+++ b/tests/test_io.py
@@ -307,6 +307,7 @@ def test_io_datapackage(test_df, tmpdir):
     assert_iamframe_equal(test_df, import_df)
 
 
-def test_io_netcdf(test_df_year):
-    obs = read_netcdf(TEST_DATA_DIR / "test_df.nc")
-    assert_iamframe_equal(obs, test_df_year)
+def test_io_netcdf(test_df, tmpdir):
+    file = Path(tmpdir) / "foo.nc"
+    test_df.to_netcdf(file)
+    assert_iamframe_equal(read_netcdf(file), test_df)