SUMMARY

Because of the role of the deep oceans in absorbing atmospheric heat trapped by the greenhouse effect, it is important to have accurate estimations of underwater temperature change at large depths at the global scale. Although active acoustic and non-acoustic direct measurement methods currently exist to monitor the ocean, these methods are plagued by environmental concerns and large measurement uncertainty errors, respectively. Cross-correlation processing of ocean ambient noise has been suggested as a potential means for developing noise-based modalities of acoustic ocean monitoring techniques such as acoustic thermometry without using any controlled active sources. This research examines the feasibility of using passive acoustic monitoring to monitor changes in sound speed in the deep ocean, which correspond to changes in temperature. This work investigates the feasibility of performing acoustic thermometry between pairs of hydrophones of the International Monitoring System (IMS) of the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) using only low-frequency ambient noise (f<40Hz). The first part of the proposed work is to investigate coherence of ambient noise over long periods of time (up to 6 years) for the purposes of monitoring long-term changes in ocean temperature. The second part is to examine the coherence of ambient noise at 5 sites throughout the world to develop recommendations for the design of a sensor network to monitor the deep oceans. The final part is to optimize the emergence rate of the coherent components of the noise using predictive state estimation techniques.