A new algorithm for sound speed in seawater

Abstract

Travel times of acoustic pulses across a 3000-km section in the northeast Pacific are used to estimate an algorithm for the speed of sound in seawater. This algorithm, derived from tomographic techniques, is inconsistent both with the international standard algorithm derived by Chen and Millero [J. Acoust. Soc. Am. 62, 1129–1135 (1977)] and with the algorithm of Del Grosso [J. Acoust. Soc. Am. 56, 1084–1091 (1974)]. Both previous algorithms were derived from laboratory experiments. The additive correction, δc (m s−1), to Del Grosso’s sound speeds between 0- and 4-km depth is δc(p)=−0.104 189 813 8×10−2p+0.210 143 906 0×10−4p2−0.141 411 578 8×10−6p3+0.317 812 236 0×10−9p4−0.217 806 024 3×10−12p5 with p being pressure-gauge units in kg cm−2. The rms error of δc is about 0.05 m s−1 and 0.1 m s−1 between the intervals of 0 to 2 km and 2 to 4 km, respectively. At about 3-km depth, sound speeds predicted by Chen and Millero and Del Grosso are about 0.7 m s−1 and 0.2 m s−1 too fast, respectively. An accurate algorithm for sound speed is of fundamental importance in acoustics and in tomographic measures of ocean temperature.