The cause is that the wood cells are very much longer in the vertical [alive and well] direction as compared to the horizontal. Think of a lot of thin tubes lightly glued [bonded] together. Suppose they all shrink equally in all directions. Now think of the structure of the tree, and simplify it. One cell in the center. More around that, more around that, more around that ...ever increasing circles of cells. Now let all shrink in all directions equally, so losing the same percentage per cell. The inner ones will each shrink by the same absolute amount as those individual cells around the outside, but since there are more around the outside, the total absolute amount will be greater there.
If this is correct, then [near as dammit is to swearing], thinking in term of a sequence of circles, the percentage of crack-space in an inner circle should be the same as the percentage of space in the outer edge, and in fact throughout the material: More material, more space, same percentage.
If still confused, think of expansion and contraction of a steel rod due to heating/cooling. A longer rod will expand and contract further than a small one. The measure of wood around the outside of a log is greater than around the center, and the tree grows in definite rings.
Why does it crack in the first place? Bonding. Even steel cracks under pressure since it is uneven throughout, even though it looks uniform outside of a microscope. I'll quit now. My brain is exploding.