Here’s an intuitive “proof” (not mathematically rigorous but still a solid line of reasoning):
Consider a constant angle $θ$ (angle between the planes), and an angle $α$ (angle between the plane bisecting the upper and lower planes described in the problem and the horizontal plane).

Imagine positioning the notch so that the upper plane is vertical. The volume of the notch would obviously be infinite. The current value of $α$ is the maximum possible value of $α$.

Now imagine slowly changing $α$ so that the notch tilts downwards. At any instant in time, the change in volume is equal to the infinitesimal amount of $($volume gained$)$ $-$ $($volume lost$)$. These volumes can be approximated with the area of the two moving boundaries.*

As you can see, when $α$ is greater than $0$, the upper plane is always larger in area than the lower plane, as the cross-section of a cylinder at a steeper angle is always larger. This means that the volume lost is greater than the volume gained as $α$ decreases when $α > 0$, meaning the notch is decreasing in volume.
Similarly, when $α < 0$, the bottom plane is larger, meaning that the volume gained is larger. The notch increases in volume as it moves downward.
Eventually, when $α$ reaches its minimum value, the volume of the notch will again be infinite. (Imagine flipping the second diagram upside-down).
Here’s my rough sketch of a volume vs time graph for the notch based on the logic above:

It makes sense that the minimum value is reached when $α$ is $0$ and the ‘middle’ plane is horizontal, meaning that the angles between the horizontal plane (which is also the bisecting plane) and the planes above and below are the same!
It doesn't matter what the tree's radius or $θ$ are (as long as $θ$ is less than 180^\circ rad).
In summary, by showing that the derivative of a volume function $V(α)$ wrt. α is negative when $α > 0$ and positive when $α < 0$ for $|α|<90^\circ-{1\over2}θ$, you can conclude that volume is minimum when $α=0$.
*The infinitesimal changes in volume are not exactly proportional to the areas since the planes are rotating and not just translating. But since the relationship between the area and change in volume is monotonic, the reasoning holds true.