Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research.
Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research.
复制标题
水生生态系统研究中使用的十种声后向散射模型的比较。
DOI：
10.1121/1.4937607
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
K. Sawada
中科院分区：
文献类型：
--
作者：
J. Jech;John K. Horne;D. Chu;D. Demer;David T. I. Francis;N. Górska;Benjamin A. Jones;A. Lavery;T. Stanton;Gavin J. Macaulay;D. B. Reeder;K. Sawada
在研究和生态系统监测应用中，带有数字表示的分析和数值散射模型越来越多地用于预测鱼类和浮游动物的声学反向散射。十个这样的模型被应用于具有简单几何形状和参数化（例如尺寸和材料属性）的目标，以代表浮游动物和鱼类等生物有机体，并将它们对声学反向散射的预测与精确或近似分析模型（即基准）的预测进行比较。这些比较是针对球体、球壳、长椭球体和有限圆柱体进行的，每种都具有均匀的成分。对于每种形状，考虑了四种目标边界条件：刚性固定、压力释放、充气和弱散射。目标强度（dB re 1 m(2)）计算为声波频率（f = 12 至 400kHz）和入射角（θ = 0° 至 90°）的函数。一般来说，数值模型（即边界元和有限元）在整个模拟参数范围内与基准相匹配。虽然说明了与近似分析模型相关的固有误差，但它们的优点也得到了说明，因为它们计算效率高，并且在某些情况下，在数值模型不收敛的条件下优于数值模型。
Analytical and numerical scattering models with accompanying digital representations are used increasingly to predict acoustic backscatter by fish and zooplankton in research and ecosystem monitoring applications. Ten such models were applied to targets with simple geometric shapes and parameterized (e.g., size and material properties) to represent biological organisms such as zooplankton and fish, and their predictions of acoustic backscatter were compared to those from exact or approximate analytical models, i.e., benchmarks. These comparisons were made for a sphere, spherical shell, prolate spheroid, and finite cylinder, each with homogeneous composition. For each shape, four target boundary conditions were considered: rigid-fixed, pressure-release, gas-filled, and weakly scattering. Target strength (dB re 1 m(2)) was calculated as a function of insonifying frequency (f = 12 to 400 kHz) and angle of incidence (θ = 0° to 90°). In general, the numerical models (i.e., boundary- and finite-element) matched the benchmarks over the full range of simulation parameters. While inherent errors associated with the approximate analytical models were illustrated, so were the advantages as they are computationally efficient and in certain cases, outperformed the numerical models under conditions where the numerical models did not converge.