Infrasonic Monitoring of Eruptions at Tungurahua Volcano, Ecuador using a Wireless Sensor Network

Wireless sensor networks, consisting of small, low-power devices integrating a modest amount of CPU, memory, and wireless communication, could play an important role in volcanic monitoring applications. Wireless sensor nodes have lower power requirements, are easier to deploy, can can support a larger number of sensors distributed over a wider area than wired arrays currently used in many campaign studies. Using long-distance wireless links, data could be monitored in real time, avoiding the need for manual data collection from remote stations. We developed and deployed a wireless infrasonic sensor array at Volcán Tungurahua, Ecuador, in July 2004. This network consisted of three wireless sensor nodes that digitized infrasonic signals, transmitting data to a remote base station. The sensors are based on the Mica2 mote platform, which integrates a 7.3 MHz Atmel Atmega128L embedded controller with 4 KB of RAM and 128 KB of ROM. The Mica2 uses a low-power, single-chip radio, the Chipcon CC1000, capable of transmitting data at 22.5 kbps with an outdoor range of approximately 100 m. The node measures 5.7 cm x 3.2 cm x 2.2 cm and is operated on 2 AA batteries, with a lifetime of about 157 hours without duty-cycling the radio or CPU. These nodes run a specialized operating system called TinyOS that is specifically designed for wireless embedded devices. Each sensor node sampled infrasonic signals continuously at 102 Hz, transmitting data over a short-range radio link to a local aggregator node. The aggregator relayed the data over a 9 km wireless link to a laptop station at the volcano observatory, using a pair of spread-spectrum FreeWave modems and 9 dBi Yagi antennas. Nodes were time-synchronized using a separate GPS receiver that transmitted periodic timestamp messages, allowing our data to be later correlated with signals acquired at a nearby wired seismoacoustic sensor array. During the deployment, we collected over 54 hours of continuous data which included at least 9 verified explosions. In addition to continuous sampling, we have developed a distributed event detector that automatically triggers data transmission when a well-correlated signal is received by multiple nodes. This approach greatly reduces radio bandwidth and energy consumption, and we plan to deploy this new system as part of a larger wireless sensor array in the near future.