Acoustical Society of America
ASA/EAA/DAGA '99 Meeting
Lay Language Papers
Ola M. Johannessen - ola.johannesen@nersc.no, S. Sandven , H. Sagen, T. Hamre, H. Hobaek
Nansen Environmental and Remote Sensing Center (NERSC)
University of Bergen, Norway
K. Hasselmann - klaus.hasselmann@dkrz.de, E. Maier-Reimer, U. Mikolajewicz - Max Planck Institute for Meteorology, Hamburg, Germany
P. Wadhams - pw11@cam.ac.uk. and A. Kaletzky, Scott Polar Research Institute, University of Cambridge, England
L. Bobylev, E. Evert, V. Troyan, Nansen International Environmental and remote Sensing Center, St. Petersburg, Russia - nansen@online.ru
K.A.Naugolnykh knaugolnykh@etl.noaa.gov and I. Esipov, Nonlinear Acoustics Laboratory, N.Andreyev
Acoustics Institute, State Research Center, Moscow
Popular version of paper 3aAOb4
Presented Wednesday Morning, March 17, 1999
ASA/EAA/DAGA '99 Meeting, Berlin, Germany
The overall objective of AMOC is to develop and design an acoustic system that will monitor ocean temperature and ice thickness in the Arctic Ocean (including the Fram Strait), study long-term climate variability, and thus detect global warming. The acoustic monitoring system will be designed (using existing acoustic source/receiver technology) to fulfill the following requirements:
The source of energy which drives the climate on earth is radiation from the Sun. The global climate determines the energy budgets of the atmosphere, cryosphere, biosphere and other parts of the earth system. Usually when a climate change is discussed, one hears about increasing atmospheric temperatures - the Greenhouse effect. The importance of the ocean is often underestimated. Today the oceans cover 70% of the Earth's surface and their ability to store solar energy is high, which delays and reduces temperature changes on a daily, seasonal and regional scale. The oceans are also very important in the absorption of CO2, which is one of the main anthropogenic greenhouse gases.
Recent results from global climate models indicate that global warming in the next decades will be most pronounced in the Arctic region (Manabe et al., 1991, Cattle et al.,1995). Furthermore, time-dependent greenhouse warming simulations indicate that within the next decade the global warming signal should become detectable over the natural variability. This may cause a net melting of the sea ice in the Arctic due to advection of warmer Atlantic Water through the Fram Strait (Semtner, 1987, Johannessen et al., 1995 a, Johannessen et al., 1995 b) and St. Anna Trough.
Sound waves, however, propagate in water with very little absorption and have been used extensively in underwater defense systems for several decades. In oceanography, acoustical methods such as acoustic tomography and thermometry (Munk et al., 1995), have become increasingly important in studying physical processes such as surface waves, currents and temperature distribution. These methods employ configurations of sound sources and receivers in arrays.
Figure: Suggested acoustic tracks in a Arctic monitoring system.
In AMOC a system of well known acoustic source and receiver technology will be established and mounted underwater. The source will send acoustic signals with a given frequency content. The signals travel through the ocean and parts of it will be received by a receiver positioned on, for example, the other side of the Fram Strait (See Figure above). Usually the travel time between transmitter and receiver is measured together with the transmission loss, showing how much the signal is reduced compared to the source signal. The signal travels at a sound speed modified by ocean temperature, sea water pressure and salinity. As a rule of thumb, the speed of sound increases with around 4 m/s per degree Celsius; 1.5 m/s per 100 m depth increase; and 1 m/s for a salinity increase of 1 %. The sound speed is most sensitive to ocean tempereature. If the travel time for sound waves across the Arctic Basin changes over a period of 10-30 years the change in mean ocean temperature can be estimated.
During a pilot Transarctic Acoustic Propagation experiment (TAP), conducted by Mikhalevsky in the spring of 1994, long range propagation (2600 km) in the Arctic Basin was performed. This experiment showed that transmission across the Arctic basin is possible at frequencies around 20 Hz. Furthermore, the results of this experiment showed that travel time and phase measurements are sufficiently accurate to pick up changes caused by climate change in the Arctic basin (Mikhalevsky et al. 1999). This demonstrates the feasibility of using acoustic travel time measurements to monitor changes in average ocean temperature. A new low frequency source has been deployed North-East of Svalbard, transmitting sound at frequency close to 20 Hz every fourth day for 2.5 years (Mikhalevsky et al. 1999).
Estimates have been made which combine hydrologic data from different times and locations, resulting in a large spread in values. Previous estimates of current flux of the West Spitzbergen Current (WSC) across 79 N vary from 1.9 Sv to 8 Sv (Simonsen and Haugan, 1996). This large variation in water mass exchange introduces errors in ocean circulation models. A promising method of monitoring temperature and current velocity through the Fram Strait is acoustic signal travel time measurement at the cross-section of the strait (which has a typical width of 300 km and depth of 2700 m - see Figure). This method is based on the fact that the propagation time of acoustic pulses along the paths connecting a source and a receiver (eigen rays) is determined primarily by the distributions of temperature and longitudinal stream velocity components. This allows the acquisition of appropriate data from acoustic measurements within an acoustic tomography framework (K.A. Naugolnykh, et. al., 1998a). In the sensitivity study of acoustic propagation to current velocity changes the methods of reciprocal transmission, scintillation, and Horizontal - Refraction Modal Tomography (HRTM) are used. It is shown that HRMT potentially can be used in the Fram Strait for transverse current monitoring.
It was also found that the multi-year ice area has decreased more rapidly, about 9% per decade (i.e., 18%) over the observation period, which represents a reduction of nearly 900,000 km2, of which about half is replaced by first-year ice. This is the first time that quantitative changes in the character of the sea ice cover have been derived from passive microwave satellite data. These findings show an Arctic ice cover in apparent transformation, and suggest a thinner winter ice cover, consistent with very recent analyses of two decades of submarine sonar transect data (Wadhams, 1998). However, it must be noted that even two decades of observations, whether satellite or in situ, are inadequate to establish whether these represent long-term trends. Furthermore, the most fundamental parameter in studies of the responses of the ice cover to global warming - the ice thickness - cannot be measured by existing satellite techniques.
In the Arctic Ocean a cold surface layer, 80 - 200 m deep, is always present in ice-covered areas. Depending on the thickness of the surface duct, sound above a given frequency will be trapped within the duct and repeatedly interact with the ice cover. After several bounces exclusively with the sea ice cover, the sound has been exposed to several reflection losses and scattering losses due to the sea ice. The reflection loss is caused by conversion of acoustic energy into elastic waves within the sea ice cover. The conversion of energy depends strongly on the age of the sea ice (given by the elastic properties) and its thickness. By considering the reflection loss a strong sensitivity to changes in ice thickness and elastic properties of the ice cover is revealed at frequencies above 100 Hz. Based on our simulations a thinning of the ice cover will make the received signal from a broad band source less attenuated for a broader range of frequencies. Below 100 Hz the reflection loss is insignificant and the sound starts to leak out of the surface duct, causing it to be less and less sensitive to the internal properties of the sea ice, including sea ice thickness. So, in order to obtain information about the internal properties and ice thickness, acoustic measurements must be made at frequencies or in frequency bands which are sensible to the sea ice, not at low frequencies, which would cause total specular reflection.
If the frequency is below 100 Hz the sound will not sense the presence of the cold duct, and is attenuated and scattered from reflections off the rough underside of the sea ice. At these frequencies reflection loss caused by the rough sea floor is introduced at shallow and intermediate water depths. Loss measurements at low frequencies will provide information about the roughness of the sea ice and the reflectivity of the sea floor. Thus the low frequency - 19.6 Hz - used in the TAP experiment can only provide limited information about sea ice roughness. As a consequence, a system for sea ice has to use a time-limited pulse or a coded pulse to obtain information about the internal properties of the sea ice. Since it is the transmission loss caused by the sea ice cover that will be used, the ranges have to be much shorter, in order to have a good enough signal to noise ratio.
Ambient noise is defined as sound which is generated by many natural processes within or beneath the sea-ice cover. The main sources in the Arctic Ocean are related to building of ridges, thermal cracking and breakup of sea ice. These noise generating processes reflect the dynamic processes of the sea ice. On the other hand, long term changes in the Arctic Ocean climate will cause changes in overall sea ice properties, such as the observed reduction of more than one-year-old ice during winter time (Johannessen et. al., 1999). This again will change the response of the sea ice to dynamic processes caused by wind stress and current. Finally, this will cause a change in the ambient noise sources.
Similar to the acoustic energy generated by specific man-made sources, ambient noise is strongly affected by acoustic propagation conditions, which in the Arctic Ocean are characterized by strong surface ducts and sea ice cover. The ambient noise recorded at a location therefore contains combined information of the environment in which it was generated and has propagated through, including the sea ice cover. In 1974 Diachoke and Winikur concluded after a brief argumentation that "Reflection and scattering losses at the interfaces affect transmission loss and, consequently, the shapes of the ambient noise versus frequency curves". After this study very little attention has been paid to the effect of propagation loss on ambient noise characteristics.
The main effort in the present work was to analyze broad band ambient noise data obtained during the Marginal Ice Zone Experiment (MIZEX - 1985, 1987) and the Seasonal Ice Zone Experiment (SIZEX - 1989, 1992) and relate it to the complex oceanographic conditions found in the complicated region at the border of the sea ice-covered Arctic and the open ocean. Ambient noise data from 82 different locations in the Barents Sea and 72 locations in the Greenland Sea were analyzed. Environmental conditions were observed both by in situ measurements and remote sensing data from satellites. Briefly, our results show a correspondence between averaged ambient noise observations in the frequency band from 20 Hz up to 5 kHz and ocean stratification and sea ice properties. This correspondence is generally explained by sound propagation conditions that obtain under ocean stratification and sea ice reflrvtivity.
This makes the monitoring of changes in broad band ambient noise characteristics very interesting as a component in future Acoustic Monitoring System concepts in the Arctic Ocean. Ambient noise measurements have several benefits. First of all, they produce no additional man-made noise in the Arctic ocean. Secondly, receiver arrays are much cheaper than configurations using large, low frequency acoustic sources. Thirdly, ambient noise recording systems need much less energy supply and can also easily be mounted at the sea bottom or under the ice. Finally, the ambient noise can be used in a variety of ways, both to identify dynamic processes causing break up, swell and ridging, and to retrieve changes in averaged ice parameters and ocean stratification.
Remote sensing from space measures the ice extent efficiently, whereas no synoptic means for measuring the average ocean temperature or ice thickness distribution exists. Up to now average temperature of the Arctic ocean and sea ice thickness distribution estimates have been based on point measurements, and are thus very rough. Acoustic intensity tomography, with its potential for obtaining sea ice thickness, will compliment satellite data by correlating coverage and thickness changes. Acoustic thermometry will monitor changes in average ocean temperature. The unique combination of underwater acoustic remote sensing (AMOC) with satellite remote sensing of the ice cover (including modelling and data assimilation) in this sensitive Arctic region is key to monitoring and understanding global climate changes and early detection of global warming.
This research was supported by the European "Environment and climate programme" under contract: ENV4-CT97-0463, the Norwegian Research Council project number 121290/720, 101798/410 and by NATO linkage grant: (E.NVIR.L G 960352)528 (96) LVdC.
For more information, go to the AMOC Home Page.