The K (potassium (or potash)) to h (distance from surface to the subducting slab) relationship is an older concept based on a common observation of increasing potassium concentrations with distance into the rear arc of subduction zones. Here is a plot from a Geology paper by William R. Dickinson in 1975 showing the relationship.
And here is a graph of my new data from southwest Bolivia. (click here for the location map (red circles are the new data)).
wt. % K2O vs. distance from arc front (equivalent to 125 km above the slab in the west (left side of graph) to 180 km in the east (right side)).
As is common in the central Andes (but rarely seen with such awesome data), there isn’t much (if at all) of a K-h relationship.
The relationship (when apparent) is often attributed to smaller degrees of partial melting with distance behind the arc front.