Oceanic winds estimated from underwater ambient noise observations in SWADE
Oceanic winds estimated from underwater ambient noise observations in SWADE
复制标题
根据 SWADE 的水下环境噪声观测估计的海洋风
DOI：
--
复制
发表时间：
1996
期刊：
影响因子：
--
通讯作者：
W. Large
中科院分区：
文献类型：
--
作者：
Ram Vakkayil;H. Graber;W. Large
利用表面波动力学实验（SWADE）的测量结果，评估了在FASINEX实验中开发的从水下环境声估计海洋风的WOTAN技术。WOTAN是从Minimet浮标MET-3部署的，MET-3是SWADE阵列的一部分，并携带传感器，用于以足够的精度测量表面气象。WOTAN传感器位于513米的深度，并在6个多月内每7.5分钟记录3 - 25 kHz范围内11个频率的环境噪声水平。将这些噪声声谱级与观测到的浮标风速和其他计算出的风变量进行比较，以确定频率和风速的相关性。首先筛选环境声音时间序列以消除来自运输和降水（雨）噪声的污染。然后，使用现场和频率相关的校正因子，将这一安静的数据集标准化到1米深。进一步研究了1kHz处的声谱级和谱斜率随风速的变化。推导出噪声级与风变量之间的线性和二次关系。这些关系，然后反演估计风变量使用的噪声谱和压力水平。WOTAN推导的估计和浮标观测之间的统计比较。FASINEX和SWADE算法之间的差异也进行了研究。
The WOTAN technique to estimate oceanic winds from underwater ambient sound developed in the FASINEX experiment is evaluated using measurements from the Surface Wave Dynamics Experiment (SWADE). The WOTAN was deployed from the Minimet buoy, MET-3, which was part of the SWADE array, and carried sensors for measuring the surface meteorology with sufficient accuracy. The WOTAN sensor was located at a depth of 513 m and recorded ambient noise levels at 11 frequencies in the range 3-25 kHz every 7.5 minutes for over six months. These noise sound spectrum levels are compared with the observed buoy wind speeds and other computed wind variables to establish frequency and wind speed dependencies. The ambient sound timeseries are first screened to eliminate contamination from shipping and precipitation (rain) noise. This quiet data set is then standardized to 1 m depth using site and frequency-dependent correction factors. The variation with wind speed of the sound spectrum level at 1 kHz and the spectral slope is further examined. Linear and quadratic relationships between noise level and wind variables are derived. These relationships are then inverted for estimating the wind variables using the noise spectrum and pressure levels. Statistical comparisons are made between the WOTAN derived estimates and the buoy observations. Differences between the FASINEX and SWADE algorithm are also examined.