Collecting raw data from a wireless sensor network for environmental monitoring applications can be a difficult task due to the high energy consumption involved. This is especially difficult when the application requires specialized sensors that have very high energy consumption, e.g. hydrological sensors for monitoring marine environments. This paper introduces a technique for reducing energy consumption by minimizing sensor sampling operations. In addition, we illustrate how a randomized algorithm can be used to improve temporal coverage such that the time between the occurrence of an event and its detection can be minimized. We evaluate our approach using real data collected from a sensor network deployment on the Great Barrier Reef. [ABSTRACT FROM AUTHOR]
Wireless Sensor Networks promised to do for observation systems what consumer electronics have done for areas like photography-drive down the price per observation (photograph), introduce new functionality and capabilities, and make, what had been a relatively exclusive set of technologies and capabilities, ubiquitous. While this may have been true for some terrestrial sensor networks there are issues in the marine environment that have limited the realization of ubiquitous cheap sensing. This paper reports on the lessons learned from two years of operation of wireless sensor networks deployed at seven coral reefs along the Great Barrier Reef in north-eastern Australia. [ABSTRACT FROM AUTHOR]