Merge pull request #1403 from knutfrode/dev

Added example to illustrate parameterisation of surface Stokes drift …

examples/example_parameterized_stokesdrift.py

@@ -0,0 +1,31 @@
+#!/usr/bin/env python
+"""
+Parameterised Stokesdrift
+=========================
+"""
+
+#%%
+#
+# If surface Stokes drift is not awailable from a wave model, there are two alternatives:
+#    - one can increase the wind_drift_factor by e.g. 1.5%, as the Stokes Drift is typically 1.5% of the wind speed
+#    - or the surface Stokes drift can be parameterized from wind speed and fetch distance
+
+#%%
+# The latter option is activated in OpenDrift with the config setting
+# `o.set_config('drift:use_tabularised_stokes_drift', True)`
+# This activates a paramterisation of Stokes drift with the following method, as implemented by Petter Nygren from SMHI:
+# https://opendrift.github.io/_modules/opendrift/models/physics_methods.html#wave_stokes_drift_parameterised
+# The code and corresponding plot below shows how the Stokes drift factor (fraction of wind speed) varies with wind speed and fetch (3 different tabulated fetch distances).
+
+import numpy as np
+import matplotlib.pyplot as plt
+from opendrift.models.physics_methods import wave_stokes_drift_parameterised
+
+for fetch in ['5000', '25000', '50000']:
+    wind = ([np.arange(1, 35), np.array([0])])
+    sx, sy = wave_stokes_drift_parameterised(wind=wind, fetch=fetch)
+    plt.plot(wind[0], sx/wind[0], label=f'Fetch: {fetch[:-3]} km')
+plt.xlabel('Wind speed  [m/s]')
+plt.ylabel('Stokes drift / wind speed  [ratio]')
+plt.legend()
+plt.show()