At first this position is unlikely to stir thoughts of the Arabian mate; White’s knight isn’t on f6 and the king isn’t diagonal to it. The key point to see, though, is that it quickly could become an Arabian mating pattern if Black's king could be forced onto f7 or h7 by a heavy White piece (a) dropped onto the back rank to force the king to one of those squares, and then (b) moved to f8 or h8 to complete the mate.
Hopefully you would see this much for yourself by now and finish the mate easily if Black's queen were off the board. But instead the queen currently guards the back rank and meets Re8 with QxR, apparently spoiling the pattern. Everything depends on how you think about this obstacle. The temptation is to conclude that there's no usable mating idea here because Black's queen is so well-placed. The winning train of thought turns that point around: because Black's queen is separating you from mate, it is stuck on the back rank and you can attack it with an unusual sense of comfort. How? Hit it with a fork: 1. Qb3+. If Black plays QxQ, his queen has abandoned its defensive responsibilities and White has 2. Re8+, Kf7; 3. Rf8# (it seems surprising that Black’s king has no flight squares, but there it is). If Black instead replies to Qb3+ by moving his king, White has QxQ—and if Black then thinks about Rh2+ in an effort to create trouble with his rooks, White has QxR.
Really this all is an application of principles we considered in the section on removing the guard: if an enemy piece prevents you from mating, think of it as a target that may be especially vulnerable. This idea naturally becomes more potent as you learn more different ways to threaten mate. In this case White wins material if he recognizes that one of the checks he can give, Re8+, is a mating threat. The key to realizing this is a knowledge of the Arabian mate.