Same principle as chimneys. But I also noticed this line:

> leaves which have adapted to withstand greater water stress before wilting.

That must be one of the "adjustments to water transport" mentioned. So I suggest that they do, in fact, have trouble pumping water to top branches.

Maybe it's not more trouble pumping, eh, sucking water up. But that the top branches are the last ones to get water in periods of draught, and have therefore more resilience?

Or, it’s simply a rate to variably adjust to, so the tree is neither flooding nor parching the leaf.