An approximate mirror symmetry is seen when we plot the partial sums of $$\sum_{1}^N{n^{-(1/2+it)}}$$ where $N\approx{t/2 \pi}$. The point $N_0=\lfloor \sqrt{t/2\pi}\rfloor$ divides the plot into two parts that are pretty much mirror images of each other. This has been discussed in phenomenological terms by, for example, Nickel [1].
Has this mirror symmetry been explained or proven?
[1] Nickel, G., Geometry of the Riemann Zeta Function, https://arxiv.org/abs/1310.6396