After some serious computations I have found the following parity algs for 4x4 solved using the reduction method. The algs should be the shortest possible using only R, Rw & U while preserving the lower layers.
The move count is higher than the well known parity alg, but this might still be...