Wrong wind direction of vertcross

Background

Hi, all,

I'm trying to plot the wind quiver on the crosssection. But, I find the wind direction is wrong. Here're the details.

Read data and vertcross

import numpy as np
from wrf import getvar, interplevel, ALL_TIMES, CoordPair, vertcross, to_np, latlon_coords
from netCDF4 import Dataset
import matplotlib.pyplot as plt

data_dir = '../data/wrfchem/wrfout/20200901/lnox_500/'
file = data_dir+'wrfout_d04_2020-09-01_05:00:00_subset'
ncfile = Dataset(file)

# set cross
start_lon = 118.6
end_lon = 119
start_lat = 32.1
end_lat = 31.7

start_point = CoordPair(lat=start_lat, lon=start_lon)
end_point = CoordPair(lat=end_lat, lon=end_lon)

# get vars
u = getvar(ncfile, 'ua')
v = getvar(ncfile, 'va')
w = getvar(ncfile, 'wa')
wspd = getvar(ncfile, 'uvmet_wspd_wdir')
dbz = getvar(ncfile, 'dbz')
z = getvar(ncfile, 'z', units='km')

# vertcross and subset to 0-17 km
cross_u = vertcross(u, z, wrfin=ncfile, start_point=start_point, end_point=end_point, latlon=True, meta=True).sel(vertical=slice(0, 17))
cross_v = vertcross(v, z, wrfin=ncfile, start_point=start_point, end_point=end_point, latlon=True, meta=True).sel(vertical=slice(0, 17))
cross_w = vertcross(w, z, wrfin=ncfile, start_point=start_point, end_point=end_point, latlon=True, meta=True).sel(vertical=slice(0, 17))
cross_dbz = vertcross(dbz, z, wrfin=ncfile, start_point=start_point, end_point=end_point, latlon=True, meta=True).sel(vertical=slice(0, 17))
cross_wspd = vertcross(wspd, z, wrfin=ncfile, start_point=start_point, end_point=end_point, latlon=True, meta=True).sel(vertical=slice(0, 17))

# get xlabels
coord_pairs = to_np(cross_dbz.coords["xy_loc"])
xticks = np.arange(coord_pairs.shape[0])
xlabels = [pair.latlon_str() for pair in to_np(coord_pairs)]

fig, ax = plt.subplots(figsize=(10, 6))
ax.set_xticks(xticks[::20])
ax.set_xticklabels(['(' +s.replace(',', '\n') + ')' for s in xlabels[::20]], fontsize=4) # rotation

# plot wind
xs = np.arange(0, cross_w.shape[-1], 1)
ys = to_np(cross_w.coords['vertical'])
step = 2

# plot dbz
m = ax.contourf(xs[::step],
                ys[::step],
                to_np(cross_dbz[::step, ::step]),
                cmap='Reds'
               )


q = ax.quiver(xs[::step],
              ys[::step],
              to_np(cross_u[::step, ::step]+cross_v[::step, ::step]),
              to_np(cross_w[::step, ::step])
             )

plt.colorbar(m, label='dBZ')
plt.savefig('./cross.jpg', dpi=300)

Result:

Wind interpolated to 15 km

# Get the lat/lon coordinates
lats, lons = latlon_coords(u)

zlevel = 15  # km
u_zlevel = interplevel(u, z, zlevel)
v_zlevel = interplevel(v, z, zlevel)

step = 10

fig, ax = plt.subplots(figsize=(10, 6))
q = plt.quiver(to_np(lons[::step,::step]), to_np(lats[::step,::step]),
              to_np(u_zlevel[::step, ::step]), to_np(v_zlevel[::step, ::step])
             )

qk = plt.quiverkey(q, 1, 1.05, 10, r'$10 m/s}$', labelpos='E')
plt.plot([start_lon, end_lon], [start_lat, end_lat], '-r')
plt.title('Wind at 15 km level')
plt.savefig('./zlev.jpg', dpi=300)

Problem

As you can see, the wind at the 15 km level should blow from the start point to the endpoint.
But, the 15 km wind in the crosssection is the opposite.

Is there anything wrong with my method?
Thanks in advance!

Reproduce the example using vertcross, interplevel, and matplotlib.quiver, then compare the cross-section components with the 15 km horizontal wind plot. Start by checking the vertcross documentation and any related tests or implementation; done means explaining or correcting the direction discrepancy and documenting a verified plotting method.