The Use of Microprocessor Technology for the Conditional Sampling of Intermittent Ocean Processes

View More View Less
  • 1 Ocean Process Analysis Laboratory and Department of Earth Sciences, University of New Hampshire, Durham, NH 03824
© Get Permissions
Full access

Abstract

Geophysical signals are often intermittent, having statistics which vary with time. Optimal sampling of these signals requires a so-called “conditional sampling” scheme, a technique which changes the sampling program to match the time scales of the processes of interest. To optimize the limited tape storage capacity of remote oceanographic instruments, a conditional sampling scheme has been implemented using the computational power of microprocessor-controlled instruments, and several deployments have been made with various configurations of the conditional sampling algorithm. This algorithm monitors short-term changes in the energy of an incoming signal within a designated high-frequency band (by digital filtering techniques) and compares the resulting intensity with the longer term statistics of the signal. If the energy exceeds an intensity defined as critical according to some criteria, then an “event” is declared and the data are recorded at a higher than normal rate for the duration of the event. When the statistics of the expected signals are not well known, and criteria cannot be predetermined with confidence, an “adaptive” technique is required whereby the instrument makes an in situ determination of the critical intensity level for each signal based on the statistics of that signal.

Several deployments of the conditional sampling instruments have been made which demonstrate the operation of the technique. In Massachusetts Bay, a burst of high-frequency internal wave energy was identified and recorded by the adaptive critical algorithm applied to a moored temperature sensor array. On the northern California shelf, salinity was calculated in situ from moored temperature and conductivity sensors, and the resulting salinity time series conditionally sampled to identify salinity events as separate from temperature or pressure events.

Conditional sampling techniques may not be optimum for exploratory work. However, where the processes and expected signals are intermittent and have a specific signature, then the use of a conditional sampling technique can make more efficient use of the limited storage capacity of remote instrumentation.

Abstract

Geophysical signals are often intermittent, having statistics which vary with time. Optimal sampling of these signals requires a so-called “conditional sampling” scheme, a technique which changes the sampling program to match the time scales of the processes of interest. To optimize the limited tape storage capacity of remote oceanographic instruments, a conditional sampling scheme has been implemented using the computational power of microprocessor-controlled instruments, and several deployments have been made with various configurations of the conditional sampling algorithm. This algorithm monitors short-term changes in the energy of an incoming signal within a designated high-frequency band (by digital filtering techniques) and compares the resulting intensity with the longer term statistics of the signal. If the energy exceeds an intensity defined as critical according to some criteria, then an “event” is declared and the data are recorded at a higher than normal rate for the duration of the event. When the statistics of the expected signals are not well known, and criteria cannot be predetermined with confidence, an “adaptive” technique is required whereby the instrument makes an in situ determination of the critical intensity level for each signal based on the statistics of that signal.

Several deployments of the conditional sampling instruments have been made which demonstrate the operation of the technique. In Massachusetts Bay, a burst of high-frequency internal wave energy was identified and recorded by the adaptive critical algorithm applied to a moored temperature sensor array. On the northern California shelf, salinity was calculated in situ from moored temperature and conductivity sensors, and the resulting salinity time series conditionally sampled to identify salinity events as separate from temperature or pressure events.

Conditional sampling techniques may not be optimum for exploratory work. However, where the processes and expected signals are intermittent and have a specific signature, then the use of a conditional sampling technique can make more efficient use of the limited storage capacity of remote instrumentation.

Save