Aircraft noise emissions are a significant restraint to aircraft operating at low-level altitudes, i.e., during take-off and landing. Residencies located near airports are affected in particular, so flight procedures have been subject to noise-
abatement policies for decades. Such strategies rely either on measured noise at ground receiver locations in airport proximity or on analytical models predicting noise locally for given aircraft and routing data. Both approaches face various limitations in accuracy. This paper evaluates the suitability of the analytic noise model of ECAC Doc. 29 for individual aircraft noise calculation using rebuilt trajectories, weather and performance data compared with noise measurements at Munich Airport. Various uncertainties regarding aircraft configuration, equipment, and thrust levels, as well as no consideration of prevailing atmospheric conditions that may influence noise propagation, are identified as reasons for related discrepancies. Consequentially, we suggest using machine-learning models relying on noise, track and weather data to improve the noise prediction quality.