ABSTRACT

TitleAbstract
Underwater Object Size/shape Analysis for Fast Object Recognition Using Sonar Images
   In this paper we address the problem of fast recognition of man-made underwater objects. It is very important for an autonomous underwater vehicle (AUV) to complete its missions such as relocation of lost objects, removal of toxic wastes and mine counter measures etc. Here our work is focused mainly on the estimation of underwater object size and/or shape from a sequence of 2-D acoustic images acquired with a forward looking sonar (SEABAT 6012).
    Generally the first step for a recognizing/classification task is to extract some useful features of the objects. Traditionally the size and/or shape features extracted from sonar images (e.g., the side scan sonar images) are largely dependent upon the acoustic shadows, or only upon the geometric clues from the sonar images. But in many cases an underwater object may not have a detectable shadow associated with it because of reverberation, multi-path, and side-lobe effects. In addition, the imaging sonar may pass the object at a vertical angle of approximately 90ª°causing little or no shadow at all. The forward looking sonar used in this paper is steered along the motion direction and provides for real-time 2-D (distance and orientation) images of the region in front of the AUV. So it can not get shadows in its images. On the other hand, the images obtained from a forward looking sonar are quite different from the optical images. For example the geometric shape of a steel ball in sonar images is not round or just one point. In fact they are a series of light points in the sonar images (the real images provided in this paper). For an object with cylindrical shape, it is very difficult to estimate the object size using just one sonar image if the image quality is low and/or the insonified angle is inadequate (e.g., the insonified wave is perpendicular to the axis of the cylinder). For these reasons in this paper we make use of the geometric clues embodied in the sequence of sonar images, together with the near field acoustic scattering characteristics of underwater objects with regular shapes (sphere, cylinder etc.), to dynamically estimate and
   analyze the object's shape/size. We get the sonar image sequence by making the AUV (with the forward looking sonar system) move around the object or up and down. The raw images is pre-processed by morphological approach and Hough transform or voting approach to get more qualified images and its more precise geometric features, and to make the methods more robust to noise. The highlight structure characteristics in sonar images are also analyzed to get the relations between the object size and the highlights time-space distribution. A blackboard system (or a dynamic database) is used here to match a known object on the basis of a priori knowledge and to learn the new ones. Experiments are performed to
    how the feasibility of the methods.

Underwater Laser Viewing System and Its Application
   Viewing through turbid water, which is obstructed by the immixture of scattered rays prematurely reflected backward from intervening suspended particles and reflected forward from such particles after impinging on the viewed object, is improved by adopting the range gate camera, combined with pulsed illumination of the viewed object by laser rays. The attenuation of laser intensity through a body of water contained within a given boundary proved upon experiment to depend on suspended particle size for equal water turbidity.
   A trial model underwater laser viewing unit incorporating the foregoing system was tested in actual turbid river water environment, and proved it to be capable of appreciably enhancing the discernibility of viewed image. The system presents possibilities of future applications in remotely monitoring various underwater submarine cables and pipelines.

Computer Biochemistry and Molecular Technology of the Underwater Biosensor Creation
    There is a great diversity of different chemical substances found in the World Ocean. Revealing of their physicochemical features and location using special sensors is a typical example of the application of computer biochemistry and molecular technology [1-2]. The whole population of marine chemical substances contain natural oligopeptides and their analogies. They are produced by many marine living organisms and can be a result of human activity. These molecules possessing unique structures can efficiently bind to specialized receptor structures. This work is devoted to the computer study of common chemical features of oligopeptides and related substances, in particular specific radicals, protruding out of the molecule backbone, and to the interaction of these ligands with sensor structures under underwater conditions.
    The data bank EROP-Moscow (Endogenous Regulatory OligoPeptides) [3] on the structure, functions, and location of about 2000 different natural oligopeptides ranging from 2 to 50 amino acid residues was used to analyze the common chemical features. To do this, the following problems had to be solved: extraction of the primary structures of neuropeptides, oligopeptide hormones, toxins, and antimicrobial peptides from a data bank, determination of the range of existence of these molecules on the scale of the number of amino acid residues, comparison of oligopeptide primary structures to reveal their structural homologues and the possible sites of binding to receptors [4], and analysis of chemical radical composition [5].
    The analysis of chemical radicals showed that all these oligopeptides contained a positively charged group and a cyclic or hydrophobic radical of particular amino acids. The results of the present study indicate that most chemical radicals of these molecules are accessible for the interaction with the potential receptors of a target cell or sensor. The mechanism for the binding of oligopeptide ligand to the receptors and the possible role of different specific chemical radicals in the realization of this process are discussed to elucidate the main principles of the creation of the most precise biosensors. These results have been extended to the detection of many marine substances of other chemical classes including simple structures, e.g. metal ions, methane, etc.
   
   [1] Zamyatnin, A.A. Ann. Rev. Biophys. Bioeng. (1984) 13, 145-165.
   [2] Zamyatnin, A.A. J. Biochem. Organization (1992) 1, 2641-2644.
   [3] Zamyatnin, A.A. Prot. Seq. & Data Anal. (1991) 4, 49-52.
   [4] Zamyatnin, A.A. Prot. Seq. & Data Anal. (1991) 4, 57-60.
   [5] Zamyatnin, A.A. Prot. Seq. & Data Anal. (1991) 4, 53.
   

Acoustic and Magnetic Probe for Forecast Atmospheric Storms and Underwater Earthquakes
   In this paper the using of synchronized registration of changing magnetic and acoustic fields for forecast appearing underwater earthquakes or atmospheric storms is offered. It's known that changing of magnetic and electric fields and also propagation of mechanical waves along Earth surface could be properties of appearing earthquake. Synchronized measurements of these fields and calculation of correlation between different physic fields let increase probability of true forecast of appearing atmospheric storm. Ocean bottom waves generate acoustic waves in water volume. And changing of magnetic field of Earth is induced through sea medium in registration point (locality of acoustic probe). This report is devoted to consideration of acoustic magnetic probe. The probe consists of an oscillation velocity three-component receiver, register of acoustic pressure and three-component receiver of vector of magnetic field. Mathematical programs of such a complex probe allow us to measure and calculate correlation characteristics between magnetic and acoustic fields synchronously. The probe measures acoustic field by scanning of beam patterns in frequency range of 15-200 Hz. Magnetic field is measured in range of 5000-6000 nT. It is assumed that this method can be used more widely in oceanic
   exploration.
   

High Accuracy Pressure Instrumentation for Underwater Applications
   For many underwater observation and construction activities the accurate measurement of pressure is a key requirement. The pressure measurement may be used as primary observation data such as in Tsunami detection, wave and tide gauges, and platform leveling applications, or the pressure measurements may be used as associated observation data such as in depth sensors for ROV's, profiling instruments, and towed arrays.
   In all these applications pressure transducers employing quartz resonator technology have been successfully used in underwater systems that required the highest resolution, accuracy, and stability. This paper describes the construction, operation, and performance of the quartz resonator technology with specific examples of underwater applications. In addition, advances in materials and electronics promise to extend the usefulness of these devices into wider use within virtual instrumentation arrays aimed at synoptic observations.
   

Optically-Linked Laser-Interferometric Tiltmeter for Ocean Borehole Observations
    Observation of crustal deformation has significance in understanding background of occurrence of earthquakes. Tilt measurement is one of such observations, and has been performed mainly on the ground using tunnels or boreholes. Although the ground observation is relatively easy to approach, signals related to earthquake are reduced mainly due to disturbances of surface noise and large distance from the region where earthquakes often occur.
    One challenging approach, to get rid of above situations, is observations in the deep borehole drilled on the seafloor near the trench. To make best use of the merits of the seafloor borehole, such as low background noise and stable environment, tilt sensors should be highly sensitive to observe small tilt change, and should have low drift performance which can detect secular tilt changes due to plate subduction.
    In this point of view, we have developed a borehole tiltmeter based on laser interferometry. The tiltmeter, in principle, is a simple pendulum whose horizontal 2-D motion is detected laser interferometers with high sensitivity, and wavelength-stabilized laser light will make it possible to reduce drift of the detected signal (less than 0.1 nano-radians). By virtue of principle of interferometry, output signals are easily calibrated itself by wavelength of light which is very accurately determined. Moreover, we adopted quadrature-coding detection method which can detect both sub-fringe displacement and periodical interference fringe, resulting in large dynamic range (more than 140dB in our case) with even poor-resolution (12bit, or 72dB) A/D converters.
    The entire design of the tilt measurement system is as follows. A laser source with a wavelength stabilizer is housed in a pressure-resistant box located on the seafloor, and output light is introduced through an optical fiber to the tilt sensor in the borehole. The detected tilt signals, which are interference light, are also led out through optical fibers to the box on the seafloor and are electrically detected and converted as digital signals. In this configuration, the sensor in the borehole is optically linked with a light source and detetion units, resulting in being free from any electrical trouble and any harmful electric heating of the sensor, and it can work in a relatively high temperature environment (more than 100C) such as in a deep borehole.
    At the symposium, we will introduce the above-mentioned borehole tiltmeter using optical measurement techniques, and show the observation data obtained with a preliminary model in the 80m-deep ground borehole at Nokogiriyama observatory (Chiba Pref.). Although there were several problems related to the installation stability, we found the observed data were very well agreed with those obtained by water-tube tiltmeters which are located at the same observatory. From the result obtained so far, we could confirm expected performances and find several advantages of such kind of sensors to the future borehole observations.

Comprehensive Deep Seafloor Monitoring System in JAMSTEC
   More than 80 % of earthquakes around Japan in every year occured on the seafloor. However, seafloor seismic network is quite few and insufficient compared with the land based seismic network which has been established extensively in these 30 years. After the giant disaster wrought by the earthquake in Kobe in January 1995, the government reorganized new earthquake survey and research system under the policies laid by the Headquarters of Earthquake Research Promotion. Several national research institutions and universities in Japan are involved in this system to carry out basic studies on seismology rather than to predict the earthquake. As only six seafloor seismic networks have been in operation, the importance to expand seafloor seismic network around the islands of Japan was strongly recognized in the new research system. As a step to increase seafloor seismic network, Comprehensive Deep Seafloor Monitoring System was deployed in Nankai Trough off Cape Muroto in March 1997. The system is a combination of observatories with a cable and without cable. The former system comprises of two seismometers, two Tsunami pressure gauges, seafloor observatory with multiple sensors and 125 km long optical cable. The data are sent in realtime to the land station at Muroto and they are also transmitted to JAMSTEC in Yokosuka and Meteorological Agency of Japan. The system without cable, which could be deployed at any place, comprises of seafloor observatory with multiple sensors and four long-term digital ocean bottom seismometers. The data could be recovered once every month by releasing pop-up buoys to the surface through the satellite. The system with a cable was deployed on the land ward slope of Nankai Trough off Muroto at water depths between 1286 m and 3572 m. The system without cable will be deployed 200 km off Muroto in Shikoku Basin at a depth of 4300 m in 1998. Five similar systems will be deployed until the year of 2002.
   

The VENUS Project - Instrumentation and Underwater Work System
   The VENUS Project is a five-year plan between 1995 and 1999 to utilize the retired TPC-2 submarine telephone cable between Okinawa and Guam Islands for the multi-disciplinary geo-scientific studies. Deployment of the first seafloor observatory off Okinawa Island was completed in March 1998 in the land ward slope of Nansei Shoto (Ryukyu) Trench at a depth of 4,180 m. Extensive survey was carried out how to put sophisticated sensors and instruments into the cable. Electro-magnetic or inductive data coupling is the simplest method as it is not necessary to cut and recover the cable. However, it is impossible to get power from the cable for the instruments, which is a serious problem for long-term observations. Another simple method is to connect the instruments with the cable electrically in parallel. As very little power could be obtained from the cable, it is difficult to supply sufficient power for the instruments. Among several ideas to connect instruments into the SF coaxial cable, direct data coupling system was adopted in the VENUS phase I to maintain continuous electric power supply for the instruments. Integrated seafloor observatory is comprised of broadband and velocity-type seismometers, Tsunami pressure gauge, magnetometer, hydrophone array, heat-flow temperature probe, potentiometer for telluric current, transponders for precise baseline measurement, digital camera, CTD, current meter, etc.. These sensors were deployed by the deep tow and connected directly with the main cable by manned submersible Shinkai 6500 and ROV Kaiko. In order to carry out underwater works in situ, ROV matable coaxial, electrical and optical connectors were developed. Also, cable cutter and gripper were developed for the underwater operations by Shinkai 6500. This will lead revolutionary changes in the deep sea technology and deepsea research. The VENUS project will contribute greatly to expand global seismic network, and to reuse economically retired submarine coaxial and optical telephone cables over the world ocean floor.

Acoustical/optical Technology Integration with a Manned Submersible and a ROV for the Investigation of a Radiation Dumpsite and Sewage Diffuser Outfall
   Between September 22 and October1, 1997 several federal and state agencies, with the active support of industry and institutional partners, undertook investigations of several underwater sites located within the geographical limits of Massachusetts Bay. Surveyed sites included the former Lightship Dumping Ground and an inactive industrial East Site where radiation waste containers were permitted for disposal, and the completed but non-operating diffuser structures that are part of the new Boston sewage disposal system. The survey objectives included documenting the location of waste containers, identifying natural resources which are potentially at risk, and evaluating the integrity of underwater man-made structures.
   Specific mobile assets used during this program included the two-man submarine Clelia that is owned and operated by the Harbor Branch Oceanographic Institution (HBOI) and a Phantom S2 Remotely Operated Vehicle (ROV) that is owned and operated by the University of Conn. The Clelia was fitted with an EDO 258 CTFM forward looking sonar for target detection and a multi-spectral laser line scan system manufactured by Faytheon Electronic Systems that was used for target identification. The Phantom S2 was fitted with a Tritech SeaKing forward looking sonar, a color television camera, a 35mm photographic camera, and a sodium iodide gamma ray detector that was provided by Bechtel, Las Vegas. Both the Clelia and the ROV were deployed from the support ship R/V Sea Diver operated by HBOI.
   Investigation of the areas of interest within the waste disposal sites was prosecuted on several spatial scales. Broad scale survey with a low frequency 135kHz side scan sonar identified 43 potential target sites at depths of Ca. 50-70 meters within the 16 sqnm Lightship area. Potential targets were localized using a vehicle mounted high frequency forward looking sonar. The localized targets were then identified using the laser line scan system. And finally, targets that were identified as waste containers were prosecuted with the ROV which was used to place a gamma ray detector and a photographic camera in close proximity to the container to make and record detailed measurements.
   The Clelia was then deployed to survey the site in accordance with the track plan. The position of the Clelia relative to the SeaDiver was monitored using a Track Point II ultra-short baseline acoustic positioning system and the absolute position of the SeaDiver was monitored using a Differential GPS. The Track Point and GPS positioning data were fed into a PC based integrated navigation system (INS) which continuously displayed and recorded the position of both the Clelia and the SeaDiver.
   After the planned dive was completed images were downloaded, processed, displayed, printed, and evaluated within minutes which allowed the scientific party to efficiently direct the laser imaging operations planned for subsequent dives.
   A total of 12 Clelia dives were logged and approximately 20 laser images were recorded during each dive. Imagery was acquired at attitudes as low as 1m (1.4 m swath width) in areas where an intense nepheloid layer was encountered and at altitudes as high as 10 meters where more moderate optical conditions prevailed. These images determined that many acoustic signals that were thought to be generated by waste barrels were actually generated by rock formations or lobster traps. In the area of the sewer structures imagery was obtained at altitudes of 7-8 meters which was high enough to provide wide scale images that included the entire 6 meter diameter diffuser and the surrounding rip-rap. Fluorescence imagery the fluorescent signatures of some benthic biota attached to the diffuser structures and to rocks.
   The value of using a multiplicity of sensors with carefully selected range/resolution characteristics was conclusively demonstrated. For example, the long range, low frequency side scan sonar was useful in localizing specific areas of interest. The shorter range, higher resolution sector-scan sonars proved invaluable in verifying the presence or absence of previously detected targets and in nullifying navigation uncertainties. The laser line scan system demonstrated its ability to unambiguously establish the identity of detected targets. And the ROV operations complimented the larger scale laser imagery by providing close-up video that could be used to inspect the identified waste barrels and evaluate their condition.

The Application of Side-Scan Sonar System in Monitoring Benthic Artificial Reefs
    The physical status of two types of cubic artificial reefs was investigated by means of a side-scan sonar system on a testing site off Yung-an LNG port, south-western part of Taiwan. The first type of the reefs, which was named "traditional type", has dimensions of 2x2x2m, and the second type with dimensions of 2x2x2.3m was named "modified type". A total number of 100 units of these reefs (with 50 units for each type) were deployed to the testing site in November, 1996. The purpose of this investigation was to monitoring and inspecting the variations of position and height of these reefs with time. The ultimate goal of this investigation was intended to deduce the mechanisms which cause the excess subsidence of these reefs, and thereafter, to propose pertinent procedures in preventing this excess subsidence.
    After deployment, these reefs were monitored regularly at a frequency of about one per month. The positions and heights of these reefs as well as the morphology of the seabed around this site could be properly quantified. Based on their positions, these reefs could be recognized by three sets, i.e., set A, B and C. Very limited relative displacement of these reefs could be detected. The heights of the reefs were fairly constant for the first 5 months. However, a sudden height decrease of about 0.3m was detected for most of the reefs thereafter. This sudden height decrease event was closely correlated with the strong sea state caused by a storm. According to limited information obtained during the past 10 months, the major factors influence the subsidence of the artificial reefs in this testing site were scouring effects. This assumption could be further proved by the comparison of information obtained before and after the passage of a typhoon in the future.
    The side-scan sonar system illustrated extremely effectiveness in searching, positioning, and height measurement of benthic artificial reefs.

Tools and Techniques for Deep Water Sediment Sampling around a Ship Wreck
   In August 1997, a series of near-bottom operations were conducted in an acoustic transponder net deployed around the wreck of the ex USS-Agerholm,about 130 nm offshore Southern California in 800 m of water depth. The Marine Physical Laboratory's Deep Tow control vehicle equipped with two orthogonal horizontal thrusters, a sector scanning sonar, cameras,lights, a navigation sonar, and an altimeter sonar was used to first locate the wreck in a debris field spanning about 40,000 m2.
   Unlike conventional neutrally buoyant ROVs the control vehicle can handle payloads of a few thousand kilograms, depending on the water depth. The 0.680" electomechanical tow cable from which the vehicle and payload are suspended carries the full load and provides electrical power. The vehicle provides lateral position control.
   We used this capability to suspend a 0.25m2 box core below the vehicle and take sediment samples at locations determined via transponder navigation. In spite of moderately rough weather conditions with 20 knot winds and 1.5-3 m seas on average, and in spite of operations from a ship without dynamic positioning, eight usable core samples were taken at ranges of 10 m to 1 km from the wreck. The vehicle was also used to deploy four sediment traps and an ADCP at precisely located bottom positions on a 50 m radius around the wreck.
   The tools and techniques used during these operations will be presented and their relative merits or drawbacks will be discussed.

VENUS Project Overview: Multi-Disciplinary Geophysical Measurements at Ocean Bottom Using Demission Submarine Telecommunication Cables
   Real time scientific measurements at deep-sea are getting very important for geophysical studies, specially for the seismology. Submarine cables are on of the most suitable devices for real-time measurements. The construction costs for a submarine cable system, however, are expensive in the scientific world. Recently, coax submarine cables are being replaced by high capacity fiber optics and then the utilizing of demission telecommunication submarine cables greatly helps to reduce construction costs of real-time ocean bottom stations.
    On January, 1997, the GeO-TOC project deployed a seismic station at the Izu-Bonin Trench slope using the TPC-1, one of demission submarine coax-cables. The VENUS (Versatile Eco-monitoring Network by Under-cable System) project is going to reuse TPC-2, which is an another demission submarine coax-cable system, for the multi-purpose geophysical measurements. In the VENUS project, there are twelve instrument set: a three components of broadband seismometers, a hydrophone array, precise distance measurements,
   tsunamis, geomagnetic-geoelectric instruments, an acoustic data link, a CTD, a digital camera, a current meter, a nephrometer, a ground thermometer and short period seismometers. Digital telemetry using analog telephone cables is going to be used. Several new technologies in instrumental deployment and data communication are adopted. Each instrument and the data transmitting unit will be connected through each deep-sea matable connector by use of a submersible at 4200m deep. The branching unit will be connected to the Okinawa and the Guam cable ends by splicing and by deep-sea coax connectors, respectively. GPS time signals will be transmitted from shore to each instrument through cables. A 24V DC for instrument powers is generated by a DC/DC converter by withdraw of 1000V DC from the submarine cable voltage. The DC/DC converter is designed to keep a constant current due to the constrain for the cable power supply. Rush current for each instrument is designed to be less than 1A.
    A part of system is going to be deployed at 4200m deep of the Ryukyu Trench slope on November, 1997 and the rest of instruments will be deployed in March, 1998. The data obtained by the system are going to be transmitted to the main office of JAMSTEC through the shore station.
   

Design and Testing of an Underwater Microscopy for the Study of Zooplankton Distribution
    In this paper, design consideration and testing of an underwater imaging system equipped with a shipboard computer vision system for the study of zooplankton distribution is presented.
    A method that can monitor zooplankton density at adequate spatio/temporal resolution is desired in the field of oceanic ecosystem studies.
    We have developed a submergible microscopy equipped with a non-interlace CCD camera. As target plankton, ploima, cladocera, calanoida, cyclopoida, ciliata are seclected those are dominant species in coastal sea in Japan. Requirements of an system for an underwater imaging of zooplankton are discussed. Selections of lens, camera and illumination are key issues, as they may influence on performance of the system.
    Second order auto-correlational masks are used to extract features from images. The combination of local auto-correlational masks and multivariate analysis, which is a two-step feature extraction, is a powerful tool as a general information extracting -method from images. In our procedures, a set of these masks is 33 dimensional vector. To identify and count zooplankton, multiple regression analysis is performed. Transformed feature have properties similar to a vector. The direction of a vector represents a species, and the length of it matches the number of the species. When more than two plankton is in one image, transformed features are the linear combination of vectors of each plankton. Therefore, we can count and classify zooplankton into its taxonomic unit. Through another multiple regression, sizing of plankton can also be performed. As far as we tested, accuracies of identifications were more than 90% when discrimination analysis was performed between any two taxa of preserved plankton. Proof of principle experiment was also performed with images of living copepoda in real time.
    In conclusion, combination of a submergible microscope and local auto-correlation masks are proved to be an effective method to identify, count and size zooplankton.

Space Continuous Measurements on Ocean Current and Chemical Properties with an Intelligent Towed Vehicle "Flying Fish"
    A research project on " Development of the Observation System for Heat, Momentum and Material Circulation in the Ocean and Atmosphere " was carried out in the Research Institute for Applied Mechanics, Kyushu University from 1992 to 1997. In this project a pitch, roll and depth controllable towed vehicle called " Flying Fish " has been developed. Flying Fish can houses an acoustic Doppler current profiler (ADCP), CO2 analyzer and sensors for measuring temperature, salinity, dissolved oxygen, pH, turbidity and chlorophyll. Flying Fish enables us to obtain the space continuous data of physical and chemical properties efficiently in the upper mixed layer of the ocean from the surface to a depth of 200m. Length of the vehicle is 3.84m, breadth 2.26m height 1.4m, weight in air 1400kg and weight in water is about 0kg. From 1994 to 1997, ocean observations were carried out in the northern, southern and central part of Japan Sea with Flying Fish as a collaboration study between Japan, Korea and Russia. A few examples of data obtained from these observations will be shown in the meeting and the results of the current velocities will be compared with those of numerical simulations.

Estimation of Vertical Profile of Chlorophyll Concentration around the Antarctic Peninsula Obtained from Nimbus-7/CZCS Imagery by the Principal Component Analysis
    Due to the importance of monitoring a vast amount of resources in the Antarctic Ocean such as the Antarctic Krill (Euphausia Superba), the biological process and production assessment in this region has become a matter of world wide concern. However, field oceanographic investigation in the Antarctic Ocean involves many difficulties (e.g. low temperature, icebergs and strong wind). In recent years, the visible remote sensing (Nimbus-7/CZCS : Coastal Zone Color Scanner) has proved to be a promising method for the estimation of phytoplankton pigment concentration and hence of primary production by virtue of its spatial and temporal capability. However, CZCS imagery can observe only surface chlorophyll concentration. The purpose of this presentation is to develop an estimated model of vertical profile from satellite-derived surface chlorophyll concentration using a statistical method (Principal Component Analysis : PCA). This presentation is composed of three elements as follows:
   (1) Propose the model by means of PCA. This part proposes the estimated model, analyzing more than 200 ship observations from the surface to 150m depth.
   (2) Validate the model. This part compares ship-observed data with estimated the vertical profile of chlorophyll concentration using ship-observed surface chlorophyll concentration.
   (3) Apply of the estimation model to satellite imagery. This part compares CZCS-derived the vertical profile with ship-observed profile data at three points. This comparison in term of vertical profile demonstrates good agreement within the error of 40%.
   The prediction by the model only need one input, surface chlorophyll concentration, which can be easily derived from the satellite remote sensing data.
   

Experimental Observations of Vertical and Horizontal Distribution of Total Suspended Sediment Using High Frequency Acoustic Backscattering Technique
   Accumulation of heavy metals such as Hg, Pb, and Cd is increasingly becomes a critical environmental concern, especially when they enter into a higher food chains and consumed by human. It has been identified that these heavy metals are attached to a particle or aggregate of particles in the water column. This fact prompted a need for a better understanding of the dynamics of these particles in order to predict their spread or the potential accumulation. Recent experimental observation at Jakarta Bay, Indonesia, employing high frequency acoustic backscattering technique is used to describe this dynamic of suspension and resuspension of particles in the water column. Vertical and horizontal distribution and concentration of the total suspended sediment will be presented and discussed.

GEOSTAR: the First European Observatory for Geophysical and Oceanographic Research at Abyssal Depths
   GEOSTAR is an European project with Italian coordination, funded by EU in the framework of Marine Science and Technology Programme (MAST-III). Aim of the project is the development of the first prototype of an innovative deep sea benthic observatory capable to carry out long-term geophysical and oceanographic observations at abyssal depths (4000 m). The observatory is conceived to act as a node of existing and future environmental monitoring networks (e.g. the Italian seismic network) and will make possible their extension to marine environment. The project, started in November 1995, will end in mid 1998 with the availability of a prototype that will undertake a demonstration phase in Adriatic Sea. The prototype will be equipped with a triaxial seismometer, two magnetometers, CTD+transmissometer, an electrochemical package for the measurement of pH, Eh, H2, H2S, a short range ADCP. Its configuration includes some innovative technological solutions such as the use of an original deployment/recovery vehicle designed for heavy loads, the modularity and operational subsystems of the observatory, a communication system based on expendable capsules released from the station. The first long-term (6-12 months) scientific mission is foreseen within the end of the century in the abyssal plain of the Southern Tyrrhenian Sea, offshore Ustica Island, Italy (3400 m.w.d.) where for the first time key information about the geodynamics and oceanography of the whole Mediterranean basin can be acquired simultaneously. In the present paper the main objectives of the project and the results achieved in the design, development and laboratory test phases are reported and discussed.