Abstract: ‍ ‍With the expansion of civil aviation transportation and the improvement of ecological environments， bird strikes occur frequently， posing a threat to aviation safety. Birds are typical “low， slow， and small” targets， characterized by low-altitude flight， slow speed， and small size. The core challenge in recognizing and classifying such targets lies in extracting meaningful features from data through deep feature learning and combining the physical characteristics of the target with a multidimensional feature model for comprehensive analysis. Meteorological radar is suitable for large-scale monitoring of migratory bird flocks. It can be effectively networked and is capable of acquiring various target characteristic parameters. Taking bird flocks and meteorological targets in the context of aviation safety as examples， we analyzed measured meteorological radar data. By parametrically representing radar multidimensional features， the characteristics of bird flocks， precipitation， and cloud targets were examined， with a focus on features such as target reflectivity， spectral width， and differential phase characteristics， among others. Through histogram visualization analysis， principal component analysis， and the Fisher linear discriminant method， significant statistical differences were identified in key parameters—such as reflectivity and correlation coefficient—demonstrating their effectiveness in distinguishing bird flocks from meteorological targets. The experimental results show that the correlation coefficient of bird flock targets presents a single-peak normal distribution， the spectrum width distribution is relatively dispersed， the reflectivity is highly concentrated in the 10~20 dBZ interval， and the differential phase is concentrated in the high value interval of 75°~125°， all of which differ significantly from meteorological targets. By clustering different target types and calculating the two-norm distance between targets， further evidence of significant separability between bird flocks and meteorological targets is provided. These findings offer both data-driven and theoretical support for developing a bird-strike warning system.