Sound generation and radiation properties are studied of full-scale tactical jet engine noise. This is motivated by the high sound exposure levels from jet noise, particularly for tactical engines. Acoustic source reconstruction methods are implemented computationally on existing jet noise data. A comparative study is performed using numerical simulations to understand the capabilities of more advanced beamforming methods to successfully estimate the source properties of a distributed, partially correlated source distribution. The properties and limitations of each beamforming method are described. Having validated the methods, beamforming with regularization---via the Hybrid Method---is implemented on linear array measurements near an installed tactical engine. A detailed analysis of the correlation and coherence properties associated with the phased array measurements guides the implementation of the beamforming. When the measurements are used as inputs to the beamforming, they produce partially correlated, distributed sources in a full-order model representation. A processing technique is also implemented that increases the usable bandwidth of the array measurements to almost an order of magnitude above the array design frequency. To more appropriately study the equivalent sources, a decomposition technique is designed and implemented to create a reduced-order wavepacket model of the jet noise. The wavepacket model is modular and scalable to allow for the efficient characterization of similar jet noise measurements. It is also appropriate for its physical significance, as wavepackets are attributed to the turbulent flow as well as the hydrodynamic and acoustic properties of the radiation. The reduced order model can estimate the levels and coherence properties of the acoustic radiation and represents a significant step towards a complete jet noise prediction model.