Spacing is determined by the energy-weighted wind rose.  But it takes anywhere from 12 to >20 diameters before upwind turbulence has decayed and boundary layer mixing has replenished the energy taken out by the upwind row.

No project developer uses such spacing today, which puts extra load cycles on the downwind turbines in any direction.. There should be a happy medium, with well understood tradeoffs between energy capture and excessive load avoidance. But proper spacing greatly increases cable costs as well, so... it's often not under major consideration.

We'll have to wait for more operational data from the low rpm greater diameter WTs to see how great the problem is before there's a chance of establishing a rule of thumb. This can also be an underestimated problem between projects, when they are clustered with a narrow shipping lane between.

The science is obtained incrementally.

"Life shrinks or expands in proportion to one's courage." - Anaïs Nin