diff --git a/pyproject.toml b/pyproject.toml
index 60e02cc..1e77f12 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -21,8 +21,8 @@ classifiers = [
 ]
 requires-python = ">=3.10"
 dependencies = [
-    "numpy>=1.22.4, <2.0",
-    "pandas>=2.0.0, <3.0",
+    "numpy>=1.22.4",
+    "pandas>=2.0.0, <4.0",
     "scipy>=1.9.1, <2.0",
     "bigtree>=0.21.3",
     "scikit-learn>=1.2.2, <2.0",
diff --git a/tests/test_processing.py b/tests/test_processing.py
index 5dd07c3..88b4ba5 100644
--- a/tests/test_processing.py
+++ b/tests/test_processing.py
@@ -511,8 +511,8 @@ def test_pd_add_time_lag(self):
                 "1:00:00_col1": [0.0],
             },
             index=pd.DatetimeIndex(
-                ["2009-01-01 01:00:00"], dtype="datetime64[ns, UTC]", freq="h", tz="UTC"
-            ),
+                ["2009-01-01 01:00:00"], freq="h", tz="UTC"
+            ).as_unit(df.index.unit),
         )
 
         lager = AddTimeLag(time_lag=dt.timedelta(hours=1), drop_resulting_nan=True)
@@ -541,8 +541,8 @@ def test_pd_add_time_lag_string_time_lag(self):
                 "1h_col1": [0.0],
             },
             index=pd.DatetimeIndex(
-                ["2009-01-01 01:00:00"], dtype="datetime64[ns, UTC]", freq="h", tz="UTC"
-            ),
+                ["2009-01-01 01:00:00"], freq="h", tz="UTC"
+            ).as_unit(df.index.unit),
         )
 
         lager = AddTimeLag(time_lag="1h", drop_resulting_nan=True)
@@ -1399,7 +1399,7 @@ def test_drop_quantile(self):
             filtered_noise["Noise_1"][filtered_noise["Noise_1"].isna()].index,
             pd.DatetimeIndex(
                 ["2009-01-01 10:00:00+00:00", "2009-01-01 15:30:00+00:00"],
-                dtype="datetime64[ns, UTC]",
+                dtype=f"datetime64[{toy_df.index.unit}, UTC]",
                 freq=None,
             ),
         )
@@ -1407,7 +1407,7 @@ def test_drop_quantile(self):
         pd.testing.assert_index_equal(
             filtered_noise["Noise_2"][filtered_noise["Noise_2"].isna()].index,
             pd.DatetimeIndex(
-                ["2009-01-01 03:30:00+00:00"], dtype="datetime64[ns, UTC]", freq=None
+                ["2009-01-01 03:30:00+00:00"], dtype=f"datetime64[{toy_df.index.unit}, UTC]", freq=None
             ),
         )
 
@@ -1440,7 +1440,7 @@ def test_drop_quantile(self):
         pd.testing.assert_index_equal(
             filtered["Temp_2"][filtered["Temp_2"].isna()].index,
             pd.DatetimeIndex(
-                ["2009-01-01 11:30:00+00:00"], dtype="datetime64[ns, UTC]", freq=None
+                ["2009-01-01 11:30:00+00:00"], dtype=f"datetime64[{toy_df.index.unit}, UTC]", freq=None
             ),
         )
 
diff --git a/tide/math.py b/tide/math.py
index 369e5b7..697c731 100644
--- a/tide/math.py
+++ b/tide/math.py
@@ -95,7 +95,7 @@ def time_integrate(data: pd.DataFrame | pd.Series) -> pd.Series:
     if data.empty:
         return pd.Series(index=data.columns, dtype=float)
 
-    t = (data.index.view("int64") - data.index[0].value) * 1e-9  # seconds
+    t = ((data.index - data.index[0]) / pd.Timedelta(seconds=1)).to_numpy()  # seconds
     y = data.to_numpy()
     result = trapezoid(y, t, axis=0)
 
diff --git a/tide/processing.py b/tide/processing.py
index ab70d17..5cdc688 100644
--- a/tide/processing.py
+++ b/tide/processing.py
@@ -1050,7 +1050,7 @@ def _transform_implementation(self, X: pd.Series | pd.DataFrame):
         if len(X) == 0:
             return X
 
-        time_diffs = np.diff(X.index.view("int64")) * 1e-9
+        time_diffs = np.diff(X.index) / np.timedelta64(1, "s")
 
         # Create DataFrame for integrated values
         integrated = pd.DataFrame(index=X.index, columns=X.columns, dtype=float)
@@ -3846,7 +3846,7 @@ def _transform_implementation(self, X: pd.Series | pd.DataFrame):
 
         new_index = X.index.to_frame().diff().squeeze()
         sec_dt = [element.total_seconds() for element in new_index]
-        increasing_seconds = pd.Series(sec_dt).cumsum().to_numpy()
+        increasing_seconds = pd.Series(sec_dt).cumsum().to_numpy(copy=True)
         increasing_seconds[0] = 0
 
         omega = 2 * np.pi * frequency