On the inadvisability of using single parameter impedance models for representing the acoustical properties of ground surfaces

Although semi-empirical one parameter models are used widely for representing outdoor ground impedance, they are not physically admissible. Even when corrected to satisfy a passivity condition in respect of surface impedance they do not satisfy the condition that the real part of complex density must be greater than zero. Comparison of predictions with frequency-domain data for short range propagation have indicated that physically admissible models provide superior overall agreement. A two parameter variable porosity model yields better agreement for many grassland surfaces and a two parameter version of the slit pore microstructural impedance model yields better agreement with data obtained over low flow resistivity surfaces such as forest floors and gravel. Impedance models and conditions for physical admissibility are summarised. In addition to those examined previously, the slit pore model is shown to be physically admissible. After providing further examples of the better agreement with short range data that can be achieved using two parameter models, it is shown that differences between frequency domain predictions at longer ranges using physically admissible models rather than one parameter models are significantly greater than those resulting from short range spatial variability and comparable with seasonal variability over grassland.