Battan noted that ducting of radar energy by anomalous atmospheric refractive index profiles and resulting abnormally strong ground clutter can occur during three types of meteorological circumstance: (i) large scale boundary layer temperature inversions and associated sharp decrease in moisture with height -- these are often created by nocturnal radiative cooling; (ii) warm, dry air moving over cooler bodies of water, resulting in cooling and moistening of air in the lowest levels; (iii) cool, moist outflows from thunderclouds. In contrast to the first two types of anaomalous propagation (AP), radar ducting associated with thunderstorm outflows is quite dynamic and may mimic echoes from precipitating clouds in terms of spatial scale and temporal evolution. While non-coherent weather radars (e.g. WSR-57) are obviously susceptible to false storm indications from this phenomemenon, Doppler radars that select the level of ground clutter suppression based on "clear day maps" may also fail to suppress the AP-induced ground clutter echoes. Operational Doppler radar systems known to be susceptible to this phenomena are the National Weather Service's WSR-88D and the Federal Aviation Administration's Airport Surveillance Radar (ASR-9) six-level weather channel. In this paper, characteristics of thunderstorm outflow-generated AP are documented using data from a testbed ASR-9 operated at Orlando, Florida. The testbed radar's rapid temporal update (4.8 seconds per PPI scan) and accurate scan-to-scan registration of radar resolution cells enabled characterization of the spatial and temporal evolution of the AP-induced clutter echoes. We discuss implications of these phenomenological characteristics on operational systems, specifically the ASR-9. Algorithms for discrimination between true precipitation echoes and AP-induced ground clutter are discussed.