I suspected that multipath reflections were causing a problem and decided to add a reflector element to the VHF folded dipole. A cardboard box is used as a temporary support.
The reflector is 13" behind the VHF dipole and consists of a 24" length of 1" aluminum tubing with a 12" length of 7/8" tubing in each end and clamps. The length has been adjusted to 27.7". I also tried a strip of aluminum foil on a yardstick with the same results.
The reflector improves the reception:
and the scan looks a lot better: