A Marine Autonomous Deep SEA Storm Monitor

(MAD SEA)

Specialty Devices, Inc., working with the Gas Hydrate Research Consortium directed by the Mississippi Minerals Resource Institute, has developed and installed a Severe Storm Monitoring System. The MAD SEA Storm Monitor was installed in approximately 1,000 meters of water depth at Mississippi Canyon 118 in the Gulf of Mexico on June 12, 2006. The Storm Monitor is to be recovered in November of 2006 at the conclusion of the hurricane season in the Gulf of Mexico. This web-page discusses the MAD SEA Storm Monitor and its capabilities. The Storm Monitor can also be used for marine mammal research and for monitoring man made acoustic signatures.

Goal

Develop a long term acoustic storm monitoring system and have it in place for the 2006 hurricane season in the Gulf of Mexico.

Theory

Monitoring of storm induced sound requires estimation of the frequency domain of interest, the dynamic range of amplitudes versus frequency and an understanding of natural and man made background noise which may interfere. Ideally, we would be able to measure calm sea natural back ground levels, be unaffected by man made interference and still be capable of recording sound levels resulting from a maximum storm event directly over the monitor location. Our frequency spectrum of interest is dictated by the sound generation spectrum from storms, by the acoustic transmission characteristics of the sea and by the frequency spectrum that provides the best indicator of storm intensity.

There is some historical information on sound generation. Sound that will not travel long distances will not be useful in monitoring storms over long distances and more recent studies have predicted the frequency spectrum that may provide storm intensity indicators. For sound source information we look at storm driven environmental noise contributions from wind, rain and waves. Wind energy has dominant frequencies in the 1 kHz to 25 kHz range. This energy does not transit long distances through the oceans and in there for is out of the frequency range of interest. Rain energy is broad spectrum, or white, and will contribute to the environmental noises which vary in storms. Rain Sound levels can vary from 0 db to 80 db in heavy rain. (McCauley 1994). Wave energy including wind contribution to wave energy varies from 74 db in calm seas to >100 db re 1 uPa in rough seas and has dominate frequencies between 70 and 140 Hz. (re; Vella et al 2001) Knudsen suggest approximately 50 db variation in sound level from calm seas to SS6 over 5 Hz to 2000 Hz. Little data exists for near field acoustic sound pressure levels for category 4 or 5 hurricanes, however there are several data sets available for Hurricanes category 3 and below. Makris and Wilson 2004 propose the best frequency indicators are below 100 Hertz as these frequencies are not as subject to blanking by bubble curtains that form in the storm generation areas and the resulting energy peaking as increasing storm intensity creates these bubble curtains.

Normally transmission characteristics allow us to estimate sound pressure levels to compare calm seas and local rough seas to distant rough seas. Transmission losses for sound in the sea are affected by the absorption loss which is frequency dependant, the spreading loss characteristics, bottom and surface absorption loss and channeling of the sound. A hurricane or other large storm is an interesting sound source as, in addition to bubbles entrained in the upper layers, they are not point radiators even at the very long wave lengths of the frequency spectrum of interest. The storm source is essentially a large circular planar radiator with areas significantly larger than the water depth. As such the spreading loss is not the 6 db per doubling of the distance due to spherical spreading but instead is more directly proportional to distance and has a 3 db spreading loss per doubling of the distance from the source. There is also a channeling effect for the lower frequencies with the sound channeled between the water surface and the sea floor. This channeling further reduces transmission losses. Absorption losses are frequency dependant with lower absorption at lower frequencies. By way of example, at 12 kHz the absorption loss for one way travel is approximately 1 db per 1,000 m. The absorption loss at 3 kHz is less than 0.1 db per 1,000 m and losses due to absorption for frequencies in the 250 Hz range are nearly negligible at 0.0003 db/ 1,000 m. (USN NAVORD 1945)

Requirements

A monitoring system should sample and record well above the frequency of interest and have sufficient dynamic range to accommodate the weakest and strongest sound pressure levels anticipated from background to nearby extreme storm conditions.

Installation Depth

The MC118 site for installation of the MAD Sea Storm Monitor is 980 m. The operational depth should be at least 1,500 meters.

Installation Duration

The GOM Hurricane season is generally June through November creating a requirement for a 180 day instrument recording and power supply life.

Frequency Spectrum of Interest

A frequency spectrum from sub-Hertz to 250 Hertz should allow detection of local and distant storms with sufficient bandwidth to detect and remove low frequency harmonic man made noises from rigs and ships. To achieve this band width, a sampling rate of 2,000 Hertz with a low pass anti-alias filter set at 500 Hertz will provide undisturbed signal at 250 Hertz and sufficient over sampling to allow removal of local rig induced harmonics. The hydrophones have a response to within 1 db from 1 to well above 2,000 Hertz. Below 1 Hertz the signal rolls off to 6 db down at ½ Hertz.

Sampling Plan

A 20 minute record should provide sufficient record length to process the lower end of the spectrum. Four records of this length per day will provide the ability to detect changes in intensity and with the addition of later systems, changes in location multiple times per day.

Dynamic Range

Sound generation levels have been shown to vary by 30 to 40 db and transmission loss even in ideal frequency spectrum to vary an additional 30 to 40 db between a local storm and one at 300 to 500 km distant. Additional dynamic range may be required to prevent signal clipping be local man made noises such as rigs and shipping. A 16 bit, 96 db dynamic range should suffice if the sensitivity and amplification of the receiver and signal processing is selected with care.

Data Storage

Sufficient data storage should exist for a 180 day hurricane season. The data storage should accommodate a 2,000 Hz sampling rate, 16 bit, two byte word and 20 minute samples occurring 4 times daily.

Power Supply

A low power uProcessor and duty cycle powering up and down of the sampling circuitry can limit the power requirement to allow use of alkaline batteries within the pressure housing of the instrument.

Design

The MAD SEA Storm Monitor has been designed as an autonomous self contained sensor, recorder and power system designed for deployment at 1,000 meter depths in the Gulf of Mexico.

The hydrophone is mounted directly to the instrument pressure housing. The pressure housing includes a wide dynamic range (> 118 db) instrumentation preamplifier, a low pass anti-alias filter at 1000 Hz , and an additional wide dynamic range amplifier. These components are supplied with +/- 12 volt operational range. The output is later divide by 10 limited to the +/- 1.25 input voltage range of the A/D. Other than during data acquisition, this input circuitry is powered down and the uProcessor is in low power mode. MAD SEA software includes watch dog timer and failsafe restart features to assure data collection throughout the design deployment.

The pressure housing is a two layer system designed to provide 10 to 20 year life in the sea with an inner pressure housing which is enclosed in an oil filled and pressure compensated outer plastic housing. Prevention of the sea water contact with the inner metal housing prevents corrosion without the expense of aTitanium housing. The MAD SEA Storm Monitor was mounted in a 316 stainless steel cage equipped with isolation from the rest of the mooring and anode protection to extend the life of the cage. While this instrument is only being installed for a 6 month installation, the MAD SEA Storm Monitor concept is under consideration as a permanent part of the Sea Floor Observatory and as such should be capable of much longer installations.

Figure 1 shows the MAD SEA Storm Monitor inner and outer housings. Figure 2 shows the MAD SEA Strom Monitor in its cage ready for deployment.

Figure 1 MAD SEA Inner and outer housings

Figure 2 MAD SEA Storm Monitor ready for Deployment

MAD SEA Specification Sheet

Main Processor	CF2
Sampling Rate	2 kHz
Resolution	16 bits
Main Power Supply	6 volt and 12 vdc Alkaline batteries with 15 A-H
Power Supply Operational Life	291 days with standard battery pack
Anti-Alias Filter	1000 Hz low pass 2 pole
Hydrophone Freq. Range	1 Hz to 10,000 Hz
Hydrophone Sensitivity	-101 db re 1 uBar
Total preamplifier/Amplifier gain	72.9 db
Data storage Capacity	180 Days can be increased to 720 days
Data interface	serial high speed
Data Transfer	CF flash card
Operational Depth	1,500 meters
In Water Weight ( w/o cage)	6 lbs.

Home Products Survey Services SDI News Free Downloads

Send mail to Paul D. Higley with questions about our products.

Last modified February 22, 2008