# Note that when we extract ``u`` and ``v`` from the file,
# we only read every third latitude and longitude.
# This choice was made because ``geocat.viz`` doesn't offer an
# equivalent function to ncl's ``vcMinDistanceF`` yet.
file_in = xr.open_dataset('../../data/netcdf_files/uv300.nc')
# Our dataset is a subset of the data from the file
ds = file_in.isel(time=1, lon=slice(0, -1, 3), lat=slice(1, -1, 3))
###############################################################################
# Make the plot
# Set up figure and axes
fig, ax = plt.subplots(figsize=(10, 5.25))
ax = plt.axes(projection=ccrs.PlateCarree())
nclize_axis(ax)
add_lat_lon_ticklabels(ax)
# Set major and minor ticks
plt.xticks(range(-180, 181, 30))
plt.yticks(range(-90, 91, 30))
# Draw vector plot
# Notes
# 1. We plot every third vector in each direction, which is not as nice as vcMinDistanceF in NCL
# 2. There is no matplotlib equivalent to "CurlyVector"
Q = plt.quiver(ds['lon'],
ds['lat'],
ds['U'].data,
ds['V'].data,
color='black',
lat_sst = sst['lat']
lon_sst = sst['lon']
lat_uv = u['lat']
lon_uv = u['lon']
###############################################################################
# Make the plot
# Define levels for contour map
# Levels are [24, 24.1, ..., 28.8, 28.9]
levels = np.linspace(24, 28.9, 50)
# Set up figure
fig, ax = plt.subplots(figsize=(10, 7))
ax = plt.axes(projection=ccrs.PlateCarree())
nclize_axis(ax, minor_per_major=5)
add_lat_lon_ticklabels(ax)
# Set major and minor ticks
plt.xlim([65, 95])
plt.ylim([5, 25])
plt.xticks(range(70, 95, 10))
plt.yticks(range(5, 27, 5))
# Draw vector plot
Q = plt.quiver(lon_uv,
lat_uv,
u,
v,
color='white',
pivot='middle',
import numpy as np
import xarray as xr
import geocat.datafiles
from geocat.viz.util import nclize_axis
################################################################################
# Open data file and extract a slice of the data
dset = xr.open_dataset(geocat.datafiles.get("netcdf_files/TestData.xy3.nc"))
ds = dset.isel(case=0, time=slice(0, 36))
################################################################################
# Create XY plot with two different Y axes
fig, ax1 = plt.subplots(figsize=(12, 8))
nclize_axis(ax1)
ax1.set_xlabel(ds.time.long_name, fontsize=24)
ax1.set_ylabel(f"{ds.T.long_name} [solid]", fontsize=24)
ax1.plot(ds.time, ds.T, color="blue", linestyle="-", linewidth=2.0)
ax1.tick_params(axis="both", labelsize=20)
ax1.set_xlim(xmin=1970, xmax=1973)
ax1.set_ylim(ymin=0.0, ymax=16.0)
ax1.set_yticks(np.arange(0, 17, 3))
ax2 = ax1.twinx()
nclize_axis(ax2)
# we already handled the x-label with ax1
ax2.set_ylabel(f"{ds.P.long_name} [dash]", fontsize=24)
ax2.plot(ds.time,
# now draw arrows
# xarray doesn't have a quiver method (yet)
# the NCL code plots every 4th value in lat, lon
# this is the equivalent of u(::4, ::4)
subset = ds.isel(lat=slice(None, None, 4), lon=slice(None, None, 4))
ax[2].quiver(
subset.lon,
subset.lat,
subset.U,
subset.V,
width=0.0015, # thinner than default
transform=ccrs.PlateCarree(),
zorder=2, # hack to make sure quiver plots on top of continents
scale=1100, # adjust till it looks right
)
ax[2].set_title("Vector Wind", loc="left", y=1.05)
# cartopy axes require this to be manual
ax[2].set_xticks(kwargs["xticks"])
ax[2].set_yticks(kwargs["yticks"])
# make axes look nice and add coastlines
[nclize_axis(axes) for axes in ax.flat]
[add_lat_lon_ticklabels(axes) for axes in ax.flat]
# nice figure size in inches
f.set_size_inches((5, 8))
plt.show()