Deliberately changing terrain undulation and ground characteristics (“acoustical landscaping”) is an potential noise abatement solution near roads. However, there is hardly any research regarding the validity of sound propagation models to predict its effectiveness. Long-term continuous sound pressure level measurements near a complex road traffic and sound propagation case were performed. Three types of modeling approaches were validated, covering the full spectrum of available techniques. A two-dimensional full-wave technique (the finite-difference time-domain method, FDTD), but also an advanced engineering model (the Harmonoise point-to-point model), provide accurate transmission loss predictions, both in 1/3 octave bands and for total A-weighted sound pressure levels. Two common and widely used semi-empirical engineering methods (ISO9613-2 and CNOSSOS) yield rather inaccurate results, notwithstanding the short propagation distance. The sensitivity to input data was assessed by modeling various scenarios with the FDTD method. Detailed ground effect modeling was shown to be of main importance.

故意改变地形起伏和地面特征（“声学美化”）是消除道路附近噪音的潜在解决方案。但是，几乎没有任何关于声音传播模型的有效性来预测其有效性的研究。在复杂的道路交通和声音传播的情况下，进行了长期连续的声压级测量。验证了三种类型的建模方法，涵盖了可用技术的全部范围。二维全波技术（时域有限差分时域方法，FDTD）以及先进的工程模型（Harmonoise点对点模型）均可以以1/3倍频程提供准确的传输损耗预测频段和总A加权声压级。尽管传播距离短，但两种普遍且广泛使用的半经验工程方法（ISO9613-2和CNOSSOS）产生的结果相当不准确。通过使用FDTD方法对各种情况进行建模，可以评估对输入数据的敏感性。详细的地面效应建模被证明是最重要的。