Researching automatic methods to interpolate between sound speed profiles at different locations

PDF

Authors
  1. McCammon, D.
Corporate Authors
Defence R&D Canada - Atlantic, Dartmouth NS (CAN);McCammon Acoustical Consulting, Waterville NS (CAN)
Abstract
When predicting the propagation of sound in the ocean, knowledge of the sound speed profile (SSP) is a critical factor in determining transmission loss because it controls the refraction of sound as it travels throughout the water column. Naturally, to obtain the best possible predictions from propagation loss models, all available profiles along a track should be input, which means that it’s the model’s job to interpolate between them. The purpose of this report is to examine what range-interpolation schemes are being used (or could be used) by transmission loss models and to assess the relative merits of each scheme. Examples of the differences between linear interpolation and feature-following techniques are shown. The study finds that from a physics standpoint, a feature-following technique is best suited to track the rising and falling of the sound channel axis with range while in the mixed layer, a linear technique is a better choice. In reality, however, the Bellhop ray tracing mathematics are not suited for feature-following methods because the changes in depth gradient cause rays to be refracted out of channels and into the bottom. In addition, there is a speed penalty of roughly 55% when using the feature-following methods compared to no interpolation, but only from 2.3% – 18% when using linear interpolation. Therefore, the recommendation is to use Linear Range Interpolation in BellhopDRDC.

Il y a un résumé en français ici.

Keywords
range dependent sound speed interpolation;triangular interpolation;trapezoidal interpolation
Report Number
DRDC-ATLANTIC-CR-2008-292 — Contractor Report
Date of publication
01 Mar 2009
Number of Pages
82
DSTKIM No
CA032472
CANDIS No
531638
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: