论文部分内容阅读
通过长期监测声波在海洋声道中远距离传播的时间，以对全球海水平均温度的变化进行观测时，对于潮汐的影响必须加以修正。本文讨论ＡＴＯＣ项目中潮汐信号预报的计算机模拟。基于潮汐信号的线性模型，根据在一段预定时间内每隔４小时测量一次所得到的数据，利用伪逆矩阵法可对潮汐所引起的传播时间变化进行预报。传播时间的这种变化是正压潮在整个声波传播路径上所产生的总效果。为了得到所要求的预报精度，必须能估计出包括长周期分潮在内的所有分潮，并分辨一些频率非常接近的谐波分潮。为此采样期不应短于１８个月。在建立潮汐模型时，分潮频率是唯一需要用到的参数。由于所有天文分潮的频率都能十分精确地知道，而非天文成分对声波传播的影响很小，因此在线性模型中包括尽可能多的分潮即可有效地减小预报误差。量化是预报误差的主要来源之一。通过对多径信号的平均能有效地减小量化误差。
The impact on tides must be corrected by long-term monitoring of the time it takes a sound wave to travel long distances through ocean channels to observe changes in the mean global sea temperature. This article discusses the computer simulation of tidal forecasting in the ATOC project. Based on the linear model of the tidal signal, the tidal-induced changes in propagation time can be predicted using the pseudo-inverse matrix method based on data measured every 4 hours over a predetermined period of time. This change in propagation time is the total effect of a positive pressure wave on the entire path of sound wave propagation. In order to obtain the required prediction accuracy, it is necessary to be able to estimate all the tidal currents, including the long period tides, and to distinguish between some of the harmonic tides that are very close in frequency. The sampling period should not be shorter than 18 months. The tidal frequency is the only parameter that needs to be used when establishing a tidal model. Since the frequencies of all astronomical tidal waves can be known with great accuracy, and the non-astronomical components have little impact on the sound wave propagation, including as many tidal scales as possible in the linear model can effectively reduce the prediction error. Quantification is one of the main sources of forecast error. Through the average of multipath signals can effectively reduce the quantization error.