Add George's original recipe for divergenece of sea current vectors

Author: George Pulickan

docs/source/recipes/plot_20_recipe.py

@@ -0,0 +1,72 @@
+import cfplot as cfp
+import cf
+
+f = cf.read("~/recipes_break/new/POLCOMS_WAM_ZUV_01_16012006.nc")
+print(f)
+
+# Get separate vector components
+u = f[0]
+v = f[1]
+print(u)
+print(v)
+
+# First get rid of the ocean sigma coord size 1 axis
+u = u.squeeze()
+v = v.squeeze()
+
+# Now we need to use some means to condense the u and v fields in the same way into
+# having 1 time point, not 720 - for example we can just pick a time value out:
+chosen_time = "2006-01-16 00:00:00"
+v_1 = v.subspace(T=cf.dt(chosen_time))
+u_1 = u.subspace(T=cf.dt(chosen_time))
+v_1 = v_1.squeeze()
+u_1 = u_1.squeeze()
+print(u_1)
+print(v_1)
+
+# Are now in a plottable form! Let's give it a go:
+### cfp.vect(u=u_1, v=v_1)
+# Need to play around with the relative length and spacing of the vectors, using these paramters:
+###cfp.vect(u=u_1, v=v_1, key_length=10, scale=50, stride=2)
+
+# Note that there appear to be some really large vectors all pointing in the
+# same direction which are spamming the plot. We need to remove these. By
+# looking at the data we can see what these are and work out how to remove them:
+print(u.data)
+print(u[:10].data.array)
+
+# ... shows more of the array
+
+# Can see there are lots of -9999 values, seemingly used as a fill/placeholder value
+# so we need to remove those so we can plot the menaingful vectors
+# Apply steps to mask the -9999 fill values, which spam the plot, to x_1
+u_2 = u_1.where(cf.lt(-9e+03), cf.masked)
+v_2 = v_1.where(cf.lt(-9e+03), cf.masked)
+print(u_2)
+print(v_2)
+
+# We can even plot the final field, effective wave height, as the
+# background contour!
+w = f[2]
+w_1 = w.subspace(T=cf.dt(chosen_time))
+# This field also needs masking for those data points.
+w_2 = w_1.where(cf.lt(-9e+03), cf.masked)
+print(w_2)
+print(w_2, w_2[:10].data.array)
+
+# Our final basic plot:
+cfp.mapset(resolution="10m")  # makes UK coastline more high-res
+cfp.gopen(file="irish-sea-currents.png")
+# BTW ignore the warnings below - they aren't relevant.
+cfp.vect(u=u_2, v=v_2, stride=2, scale=8, key_length=5)
+cfp.levs(min=-5, max=5, step=0.5)
+cfp.con(w_1, blockfill=True, lines=False)
+cfp.gclose()
+
+# Ideas for TODOs:
+# investigate difference days (do this by changing the 'T=cf.dt("2006-01-16 00:00:00")') datetime
+# values to different ones in the time coordinate data so you look at different days, or repace it
+# with a collapse over some stat e.g. mean to show the mean over all the times,
+# calculate divergence, calculate curl / relative voriticity, calculate absolute voriticity,
+# explore the other dataset as well (that covers other dates/times) - you could compare the
+# two to effectively compare the currents across different dates.