How can I get rid of diagonal tilting of shifting plt.pcolormesh() when shifting row positions?

Question

I am plotting a 2D heat map as part of my research in python using plt.colormesh() and I am able to get nice plots this way: raw data plot

I want to get rid of the bowing (artifact of experimental set up) by shifting each row to make that hot curve a vertical hot line.

I have written code to locate and shift each x row in the Xmesh to align the hot line. When I plot the data with this shifted Xmesh, the centers of each 'pixel' are aligned nicely, but each 'pixel' effectively gets tilted: data plotted with shifted Xmesh

Is there a way to shift the rows and not have this tilting effect?

Below is a simple example code demonstrating this phenomena:

import numpy as np
import matplotlib.pyplot as plt

xs = [1,3,4,8]
ys = [5,6,7,8]

zGrid = np.random.rand(len(xs),len(ys))

X,Y = np.meshgrid(xs,ys)

print(X)
print(Y)

Xshifted = np.zeros([len(X),len(Y)])
shifts = [0,1,2,1]
for i in range(len(Y[0])):
    for j in range(len(X[0])):
        Xshifted[i,j] = X[i,j] - shifts[i]

fig, axs = plt.subplots(2,1)

axs[0].pcolormesh(X,Y,zGrid,shading='nearest',cmap=plt.cm.jet,)
axs[1].pcolormesh(Xshifted,Y,zGrid,shading='nearest',cmap=plt.cm.jet,)

plt.show()

which gives these two plots showing un-shifted (top) and shifted (bottom) plots:

simple case plots

I looked around in the documentation and found an "offsets" key word argument, but the documentation around it is nearly non-existent and I could not find any examples.