Search

We present the first application of Stochastic Heterogeneity Mapping based on the band-limited von Kármán function to a seismic reflection stack of a Mediterranean water eddy (meddy), a large salt lens of Mediterranean water. This process extracts two stochastic parameters directly from the reflectivity field of the seismic data: the Hurst number, which ranges from 0 to 1, and the correlation length (scale length). Lower Hurst numbers represent a richer range of high wavenumbers and correspond to a broader range of heterogeneity in reflection events. The Hurst number estimate for the top of the meddy (0.39) compares well with recent theoretical work, which required values between 0.25 and 0.5 to model internal wave surfaces in open ocean conditions based on simulating a Garrett-Munk spectrum (GM76) slope of −2. The scale lengths obtained do not fit as well to seismic reflection events as those used in other studies to model internal waves. We suggest two explanations for this discrepancy: (1) due to the fact that the stochastic parameters are derived from the reflectivity field rather than the impedance field the estimated scale lengths may be underestimated, as has been reported; and (2) because the meddy seismic image is a two-dimensional slice of a complex and dynamic three-dimensional object, the derived scale lengths are biased to the direction of flow. Nonetheless, varying stochastic parameters, which correspond to different spectral slopes in the Garrett-Munk spectrum (horizontal wavenumber spectrum), can provide an estimate of different internal wave scales from seismic data alone. We hence introduce Stochastic Heterogeneity Mapping as a novel tool in physical oceanography.

The observation of spatial and temporal dynamics of the ocean is fundamental to understand global and regional aspects of water mixing. Physical oceanography has traditionally observed ocean structures with in situ measurements, often limited in temporal and/or spatial resolution. In exploration seismology a set of techniques has been developed over the last decades to image and characterize the physical properties of sub-seafloor structures by inversion methods at high horizontal resolution. The two different fields have made contact in seismic oceanography where the well developed methods of marine reflection seismology have been applied to the dynamic ocean. However, one aspect, so far ignored in seismic oceanography, is the dynamical, temporally varying nature of water structures. Here we show that it is possible to estimate temporal variations of reflectors in water structures as an inversion parameter. The new dynamic property reflector movement velocity gives an additional parameter to characterize ocean water dynamics.

Multi-Channel Seismic method (MCS), with its ability to image events down to a lateral resolution of 10 m has been successfully applied to address questions in physical oceanography. However, to date, these analyses have overlooked an important detail; the imaged boundaries are dynamic and move on a timescale that can be resolved by the MCS method. An important step in understanding the effect of the movement is calibration against constrained models. We demonstrate in this paper that it is possible using careful interpolation to take high resolution models of dynamic water (160 m x 2 m spatial resolution and 15 min temporal resolution) and generate models for synthetic seismic simulations (20 m x 4 m spatial resolution and 20 sec temporal resolution). We show that moving water, when ignored, will distort analyses of wavenumber spectra estimated from seismic data since the relative movement of water masses and the seismic acquisition vessel will change the apparent slope of spectra. Citation: Vsemirnova, E., R. Hobbs, N. Serra, D. Klaeschen, and E. Quentel (2009), Estimating internal wave spectra using constrained models of the dynamic ocean, Geophys. Res. Lett., 36, L00D07, doi: 10.1029/2009GL039598.

The seismic reflection profile is the convolution of the seismic acquisition impulse response and the target impedance contrasts. In the ocean, these contrasts are mainly determined by the widths and gradients of the temperature transitions between the different water masses. Hence seismic oceanography profiles are sensitive to the frequency bandwidth of the seismic acquisition system. We tested a novel seismic source that allowed us to simultaneously profile the ocean with differing impulse responses. We show that frequencies ∼20 Hz are best to delineate large impedance contrasts that occur over a vertical scale of several tens of meters whereas frequencies ∼80 Hz image the boundaries of layers of around 10 m. We demonstrate a towed acquisition system that can operate from a research vessel to give a bandwidth from 10 to 120 Hz that could, if required, be modified to provide frequencies up to 200 Hz.