Hi Paul,
The precip RTMA is much less accurate than the precip URMA. The RTMA is made at 33min past the top of the hour with whatever RFC QPEs that are available at that time and the rest is filled with "early Stage II" (from radar DPAs and METAR gauges, which are available at ~30min past the hour), while the URMA is from RFC QPEs and re-made 1/2/3/5/7 days later to catch any corrections/human intervention at the RFCs. There are no hourly QPEs from NWRFC and CNRFC and the radar-based Stage II analysis, even with the gauge correction from the rather sparse METAR) can be quite poor in these areas. For example, in the hour ending at 02Z 13 Mar, which falls into one the 6h periods you looked into, the spurious precip at the south end of CA came from the radar KSOX (Santa Ana Mountains), with upwards of 40mm/hr of precip. The much closer KNKX radar did not have any data available within +/-10min from the top of the hour, so the KSOX coverage remained. With the terrain and time constraint, the precip RTMA can be quite problematic out west.
Ying