Marine Seafloor Sediment Profiling and Classification in the Arctic Ocean

Abstract
Ocean Data Equipment Corporation redesigned their FM Chirp sub-bottom profiler, the Bathy 2000P, for operation aboard a US Navy Sturgeon-class nuclear powered submarine. Normally operated from a surface vessel, the Bathy 2000P was customized to meet specifications provided by scientists from Lamont-Doherty Earth Observatory of Columbia University for use as an Arctic research tool. An overview of the system design criteria is given that details the difficulties posed by operation aboard a submarine surveying in the Arctic and how the system was tailored to meet the scientific objectives for operation in a complex environment.

Successful operation of the Bathy 2000P during two annual Submarine Science Ice Expeditions (SCICEX) has provided seafloor and sediment sonar data previously unavailable to the scientific community. High-resolution acoustic images of the Arctic marine sub-bottom and topographic contours are presented and corroborate the use of a submarine as a platform for acoustic acquisition of geophysical data. Additional processed data, generated using a unique suite of algorithms to derive spectral and absorption loss characteristics are presented. These techniques illustrate the possibility of extending our geophysical knowledge through sediment characterization of the Arctic basin without the need for extended physical sampling.

I. INTRODUCTION

Technological advances have greatly improved man's knowledge of the world's oceans. Despite our scientific innovations, however, knowledge of geophysical features of the Arctic seafloor has remained obscure. Extreme climate and the inaccessibility of the Arctic region are mainly responsible for the lack of physical data. In order to assist scientists, the US Navy has made a Sturgeon-class nuclear powered attack submarine available for science cruises to the Arctic Ocean since 1993. Use of submarines as mobile scientific platforms have proven invaluable and vast amounts of new data has been realized from annual cruises in the Arctic depths. Submarine cruises offer many possibilities for scientific exploration in the Arctic region but also pose new difficulties for scientists and engineers responsible for designing the data collection systems. Limited space, extreme cold and pressure for external sensors, non-stationary vertical reference, longevity of the expedition and isolation of the study area are factors that contribute to the complexity of system and instrumentation design.

In 1998, the US Navy submarine USS Hawkbill sailed for the Arctic Ocean and carried, for the first time, the Seafloor Characterization and Mapping Pods (SCAMP). SCAMP was designed to examine and document geophysical properties of the Arctic seafloor. A main component of SCAMP was a high-resolution sub-bottom profiler that was an adapted Bathy 2000P sub-bottom profiler delivered by Ocean Data Equipment Corporation. The Bathy 2000P installed aboard the USS Hawkbill incorporated special features to survive the Arctic and submarine environments. A major redesign effort was undertaken that saw changes to transducers, connectors and cables, data acquisition hardware and software, processing and communication algorithms, and the addition of a new serial messaging interface as well as a high bandwidth interface to provide data to an advanced sonar data collection platform.

The successful missions of the Submarine Science Ice Expeditions (SCICEX) are in large part due to the reliability of the various instruments designed for use aboard the submarine. Images of sub-bottom records taken from the 1998 cruise, given in the next section, illustrate the successful customization of the Bathy 2000P sub-bottom profiling system. The high quality of the data, combined with the fact that the system operated with minimal operator intervention, is testimony to the dedication of the scientists and engineers responsible for system re-engineering and operation.

A major breakthrough of the SCAMP project was proving the feasibility of acoustic geophysical mapping in ice-covered seas using a submarine as a survey vehicle. The quiet platform was a primary factor contributing to data quality. Many data sets from the initial 1998 cruise, evaluated to verify system performance by Ocean Data Equipment Corporation, were of exemplary resolution and detail. Most data records detail sub-bottom penetration past 100 meters with several extending beyond 200 meters. The scientific value of these data sets to aid geophysical research in Polar Regions is easily recognized. Further demonstration of the data's importance has been shown by Caulfield Engineering Services Group of Ocean Data Equipment Corporation. Using analytical calibration techniques, Caulfield Engineering Services, has generated spectral and absorption loss characteristics as well as estimates of sediment content for a few select data records. The techniques, previously applied in regions of known sediment content with positive results, provide much insight to the physical properties of the Arctic sediment strata. A brief overview of the analytical techniques and preliminary results are given in the next section.

II. DETAILS OF THE SCAMP SUB-BOTTOM PROFILER PROJECT

A. SCICEX/SCAMP Background
The Submarine Science Ice Expeditions (SCICEX) program was initiated in 1993 with surveys planned for a five year period (1995-1999). The US Navy has made available a Sturgeon-class, nuclear powered, attack submarine for unclassified science cruises to the Arctic Ocean with the main focus being oceanographic and geophysical data collection in the Arctic Ocean. In 1998, a geophysical component, consisting of a sub-bottom profiler and swath bathymetric sonar systems, were added to the existing instrument set. Funding support for the purchase, installation and operation of the SCAMP Bathy 2000P system was provided by Palisades Geophysical Inc., the National Science Foundation and the Canadian Government.

B. Bathy 2000P Design Criteria
A Bathy 2000P sub-bottom profiler was selected to perform surveys of the Arctic sediment strata. Many modifications to the Bathy 2000P were required including the development of transducers capable of withstanding extreme pressure and drag forces. New transducers designed by Ocean Data Equipment's sister company, International Transducer Corporation, were engineered to withstand the pressure exerted at depths greater than 250 meters. Newly designed, nine ITC-5465 3.5kHz transducers were mounted within a pod attached to the keel of the submarine. Fig. 1 details the transducer array pod mounting location.

Limited space aboard the submarine for additional instrumentation made necessary a requirement that the sub-bottom profiler be located in the torpedo room. This constraint posed a serious challenge, as the Bathy 2000P electronics, designed as a rack mount system, is too deep in its original mechanical configuration. Scientists from Lamont-Doherty Earth Observatory, with technical advise from the Engineering Department of Ocean Data Equipment Corporation, were able to split the system into two shallow cases. Hardware components were connected between the two cases with an assortment of ribbon and custom cables. Extensive testing permitted removal of electrical noise sources created during this modification. Fig. 2 details the arrangement of instruments aboard the Hawkbill.

Fig. 1: Illustration of submarine showing approximate location of the ITC-5465 transducer array pod and sub-bottom profiling system.

¹Courtesy of the University of Hawaii's Mapping Research Group, http://imina.soest.hawaii.edu
² Map courtesy of the National Geographic Society MapMachine web site, http://www.nationalgeographic.com.

Designed for operation from a surface vessel, the non-stationary vertical reference inherent to submarine operations created new challenges for the design engineers. Under normal operation from a surface vessel, the Bathy 2000P system makes adjustments to the recorded data for transducer draft and vertical heave motion. Requirements for the Bathy 2000P system included automatic adjustment of the recorded data to account for the constantly varying submarine depth. This was required to correctly reference the recorded depth to the ocean surface. Navigational systems aboard the submarine provided the Bathy 2000P with a serial data stream that included the depth of the submarine. The submarine depth could vary between the sea surface and the maximum operational depth. This data was incorporated by the Bathy 2000P and necessary adjustments made to correct the digital display of acoustic sub-bottom data as well as the hardcopy output. Depth corrections were applied to archived data allowing proper display of data during post processing and insuring that the true depth of the sea floor could be recovered after the surveys were completed. Furthermore, submarine depth data allowed the vertical track of the submarine to be displayed on digital and hardcopy output. This entire process appears to be a simple matter of having the draft setting utilized in a real time manner. The range of possible submarine depths, however, is an order of magnitude (or two) larger than possible draft settings for surface vessels and thus requires very large adjustments. As a result many internal changes were required to the data acquisition and processing software of the Bathy 2000P to accommodate the varying submarine depth.

In order to provide a high bandwidth stream of acoustic data to the Data Acquisition and Quality Control System (DAQCS) aboard the submarine an Ethernet-based component was added to the Bathy 2000P architecture. Due to the high data rates achieved in shallow depths where the acoustic transmit rate is high, the new Ethernet connection utilized the latest technology. Network interface cards internal to the Bathy 2000P were rated for 100 Mbs and Ethernet cable was category 5 (CAT 5) rated. The Ethernet interface was utilized to archive acoustic data on the DAQCS. In addition, the Bathy 2000P system was modified for the SCICEX project to allow the user to select the maximum allowable file size of archived data files. Taking advantage of the full duplex capabilities of the Ethernet interface, the Bathy 2000P performed the file management operations automatically to further minimize operator intervention.

An additional messaging system was also designed for the SCAMP Bathy 2000P system. The messaging system provided a direct link to the DAQCS via a serial interface that allowed scientists to monitor the Bathy 2000P system status directly from DAQCS rather than requiring direct monitoring of the system from the overcrowded torpedo room. Transmitted via the serial link, the various messages provided acquisition status information and relayed operator and system level warnings as well as the status of acoustic data archiving processes.

C. Cruise Summaries
SCICEX cruises during 1998 and 1999 each lasted about four months in duration. During the portion of the 1998 cruise allotted for geophysical data acquisition, the Bathy 2000P recorded data continuously for 31 days (August 1 - September 1, 1998). Over 17,000 kilometers of survey track was covered with concentrated surveys focused on the Gakkel Mid-Ocean Ridge, a suspected buried fault scarp, the Alpha-Mendeleyev Ridge, and the Chukchi Cap. In addition a near-shelf transect and cross-basin transect were performed. Over a six day stretch geophysical data acquisition occurred over the Gakkel Mid-Ocean Ridge, a deep axial valley, characterized by high relief flanks. A total of 3,500 kilometers of survey track was covered along a 282 kilometer stretch of the ridge axis. The deepest depth observed in the Gakkel Ridge region was 5,173 meters and the shallowest was approximately 600 meters. Fig. 3 is a map of the Arctic basin and shows the approximate track lines of the 1998 survey. The 1999 cruise took place from March to May and extended the coverage of the Gakkel Ridge, Alaska Margin, and Chukchi Cap. New surveys of the Lomonosov Ridge and Yermak Plateau were also conducted in 1999.

Fig. 3: Map of Arctic basin detailing the submarine track lines of the 1998 SCICEX geophysical survey2.

Fig. 4: Sample SCAMP sub-bottom data collected with a Bathy 2000P in the Arctic basin on August 30, 1998.

D. Data Analysis, Processing and Classification
The Bathy 2000P sub-bottom Profiler provides a real time animated display of acoustic sub-bottom data. This display aids the scientists monitoring data collection during the surveys. In addition, data may be post-processed with Windows-based processing software. Figures 4 and 5, generated using the Windows-based processing and visualization software, are two examples of data collected during the 1998 survey as part of the SCICEX project. Fig. 4 shows the lower portion of a steep hillside leading into a sharp intersection with a flat valley floor. The high quality of the data is evidenced by the fine scale resolution of the acoustically reflective materials. Sub-bottom penetration on the steep slope is on the order of fifty meters and in excess of 100 meters on the valley floor. Fig. 5 is a high-resolution display of sub-bottom data collected on August 29, 1999 and shows, in better detail, the numerous vertical variations of acoustical energy over a fifty meter scale.

Ocean Data Equipment, in order to derive spectral and absorption loss characteristics from the Arctic data, processed an arbitrary section of the data shown in Fig. 5. The data processing was performed as a scientific exercise designed to compare the spectral and absorption characteristics of the SCAMP data with physical data previously collected in the Arctic. Caulfield Engineering Services Group of Ocean Data Equipment Corporation performed this analysis and has been developing an extensive database of acoustic properties (impedance, absorption, and velocity) versus sediment physical properties (density and material type) designed to assist acoustic quantification and classification of sub-bottom sediments. To create the acoustic properties database, an extensive array of Quality Assurance (QA) and analytical analysis software have been generated to both calibrate sonar hardware and to compute absolute bottom loss (reflectivity) to obtain estimates of sub-bottom material type. Recent publications (Caulfield [1], Caulfield [2, 3, 4], McGee, Ballard, Caulfield [8]) outline both the theoretical applications of the sonar equation to the system calibration and the use of impedance for estimating sub-bottom material type.

Fig. 5: High-resolution image of SCAMP sub-bottom data collected august 29, 1998.

The Bathy 2000P data was acquired without the advantage of the above QA programs for source, gain, and receiver characteristic determination. However, a new absorption database and analysis program allows the generation of rough estimates of bottom densities and material types. This is because the Bathy 2000P does not apply Time-Varying Gain (TVG) or smoothing functions to its archived data and, therefore, the frequency deviations due to absorption in the bottom are not lost.

The first step in the absorption-modeling step is to select a trace and examine the spectral content as a function of travel time in the Digital Spectral Analysis software (DSA50). Fig. 6 contains selected output from DSA50 with, from left to right, a cross-section of the data samples to be analyzed, an amplitude trace, the spectral content, and the absorption estimates by selected layer. The spectral analysis program slides a window down the trace and computes the spectral content of the waveform at each time increment of one-half the window size. The spectral array at each time increment is shown graphically as blocks with each block being 7 dB increments of energy. Increasing energy is represented by increasingly darker gray shades.

Bathy 2000P data is full wave rectified, hence, the base frequency is zero. The spectral deviations in this plot can be considered as side bands about the center frequency of 4750 Hz, the center frequency of the chirp signal. The spectral data contains both reflectivity and absorption information. DSA50 has the ability to eliminate the reflection effects and to estimate the absorption for each layer. The section on the far right of Fig. 6 illustrates the estimated absorption as a function of frequency for the layers selected. In this example the two top layers have virtually no absorption and then the absorption increases as a function of depth in the bottom two layers. For this example major event horizons were chosen and not individual micro-layers. Table 1 provides the color code for absorption in decibels/meter. Note that DSA50 displays results in color despite the gray scale depicted in the figures of this paper.