The path to COVIS: A review of acoustic imaging of hydrothermal flow regimes

Abstract

Acoustic imaging of hydrothermal flow regimes started with the incidental recognition of a plume on a routine sonar scan for obstacles in the path of the human-occupied submersible ALVIN. Developments in sonar engineering, acoustic data processing and scientific visualization have been combined to develop technology which can effectively capture the behavior of focused and diffuse hydrothermal discharge. This paper traces the development of these acoustic imaging techniques for hydrothermal flow regimes from their conception through to the development of the Cabled Observatory Vent Imaging Sonar (COVIS). COVIS has monitored such flow eight times a day for several years. Successful acoustic techniques for estimating plume entrainment, bending, vertical rise, volume flux, and heat flux are presented as is the state-of-the-art in diffuse flow detection.

Introduction

While scanning the seafloor terrain for obstacles, a 37 kHz sonar on ALVIN intersected hydrothermal plumes at the East Pacific Rise (EPR) (Palmer et al., 1986). In this 1984 survey of the Feather Duster vent field (10°56′N EPR), photographs of sonar range-finding showed a variety of targets, including pillow lava forms, the rim of a lava lake and hydrothermal plumes (Palmer et al., 1986). In observations reported by Palmer et al. (1986), the hydrothermal plume targets are differentiated from the underlying sulfide mound based on the estimated dimensions of the target (7–10 m wide), the known location and dimensions (4–5 m wide) of the sulfide mounds relative to the submersible, the apparent radial stretching due to the vertical extent of the target (an implied 16 m vertical extent where chimneys are generally less than 4 m), and the inferred rising and expanding behavior of plume(s). Thus began a series of studies leading to the development of specialized acoustic imaging techniques for hydrothermal plumes.

Seafloor hydrothermal systems form a fundamental part of magmatic systems, wherein seawater percolates downward into the deep crust, absorbs volcanic heat, rises buoyantly, and discharges at the seafloor (German and Von Damm, 2006). Discharge occurs both as high temperature fluids exiting in a focused manner through chimneys and flanges and as lower temperature fluids exiting diffusely through mounds, faults, and cracks in the seafloor.

On a global scale, such hydrothermal discharge is important for several reasons. First, the convective flow of the mid-oceanic ridge hydrothermal system is responsible for ~33% of the transfer of heat from the Earth's crust to the ocean (German and Von Damm, 2006). Second, the impact of hydrothermal discharge on ocean chemistry is on the scale of riverine impacts for many chemical processes (Bickle and Elderfield, 2004). Third, the heat and chemical output of hydrothermal discharge supports a local ecosystem through chemosynthesis based on reduced gases (e.g., H₂S and CH₄) carried by the hydrothermal fluids (Govenar, 2012).

Measuring the fluxes of heat and chemicals transported by hydrothermal systems from the mantle to the ocean provides significant logistical and measurement challenges (Rona et al., 2015). Approaches to measuring fluxes have either focused on detailed accounting for discharge sites (e.g., black smokers, diffuse discharge) on or near the seafloor or attempted to capture the integrated flux (e.g., buoyant plume and neutrally-buoyant plume) high in the water column (Baker, 2007). The need for a large number of measurements across a significant area (or volume) limits the accuracy of both approaches. This motivates the development of remote sensing techniques that can capture areas and volumes of flows, flow rates and possibly temperatures rapidly (if not instantaneously) over large (or at least larger) distances. The acoustic imaging techniques discussed herein are one such technique.

In general, heat flux can be estimated by measuring the amount of heat (temperature) flowing through a given area. If accounting for a collection of discharge sites, the obvious approach relies on direct measurement of temperature, discharge or orifice area, and exit velocity. The integrated (water column) approaches measure the spatial variations of flow velocity and/or heat content to estimate heat flux. Two measurement challenges are common to both approaches: (1) measuring flow rates whether at a black smoker, above a diffuse discharge site, or in the plume rising above such sites and (2) identifying and delimiting accurately the areas through which such flow occurs.

The attempt to quantify flow rates began with the first discoveries of hydrothermal vents discharging hot water into the ocean on mid-oceanic ridges (Corliss et al., 1979; Macdonald et al., 1980; Converse et al., 1984). These early efforts used vane-type or turbine flowmeters and video-captured particle motions to estimate vertical rise rates (Corliss et al., 1979; Macdonald et al., 1980; Converse et al., 1984). Later studies combined video with graded meter sticks or backgrounds, used increasingly sophisticated software for video analysis, or modified flowmeters to reduce clogging to increase reliable accuracy (e.g., Rona and Trivett, 1992; Crone et al., 2008; Ramondenc et al., 2006). Temperature arrays in a variety of configurations have also been introduced to estimate flux or flow rates based on spatial variations in plumes (vertically) and diffuse discharge (horizontally) (e.g., Little et al., 1987; Bemis et al., 1993; Bates et al., 2010). Nevertheless, achieving considerable spatial coverage with these spot measurements takes significant time, and high degrees of variability prevent accurate use of extrapolation. Alternative water column approaches, such as towing CTDs through the plumes, also lack synchronicity and cannot unambiguously attribute fluxes to a single vent or vent cluster due to the complexities of ocean currents (e.g., Baker and Massoth, 1987; Viers et al., 2006).

The measurement of areas for heat flux estimation presents similar challenges. Accurate measurement of the area of a single black smoker orifice depends on good video of the exit point (tricky with the high velocity and high temperature flow in the way). Furthermore, determining the total number of black smokers traditionally requires a systematic (and time consuming) survey of the relevant region. Similarly, diffuse discharge areas are traditionally located by video survey. Recent developments of photo mosaic and photogrammetric techniques are improving the accuracy of such surveys (Kwasnitschka et al., 2013, Barreyre et al., 2012); however, significant underwater vehicle time is still required to collect the video data.

Beyond the quantitative estimation of heat flux, efficient and accurate visualization of the behavior of underwater hydrothermal systems has the potential to support qualitative assessment of the spatial and temporal variability of hydrothermal discharge. Data acquisition techniques, which are suited to imaging and potentially relevant to studying hydrothermal discharge, include both optical and acoustic imaging. As discussed above, non-imaging data acquisition methods generally do not supply sufficient spatial coverage to efficiently describe or image any significant portion of a hydrothermal system (even a single vent can have a 10–30 m diameter footprint on the seafloor and the discharging fluids can rise as plumes for hundreds of meters). Optical imaging is challenged by the rapid attenuation and high scattering rates of light in seawater (Jaffe, 1990, Kocak et al., 2008). Optical imaging of hydrothermal plumes is impractical: video and photographs rarely show features more than 10–20 m from the camera. Close-up detection of diffuse flow by optical imaging is feasible (Smart et al., 2013, Mittelstaedt et al., 2012); broad imaging of the seafloor distribution of discharge by photogrammatic or photomosaic methods is also feasible (Kwasnitschka et al., 2013, Barreyre et al., 2012). However, both require considerable underwater vehicle time. In contrast, acoustic imaging offers the ability to detect discharge over areas from 10s to 100s m² with only minutes of imaging from a single location.

Palmer et al. (1986) recognized that acoustic imaging had the potential for acquiring measurements over a broader area (>10 m² rather than <0.1–1 m² for spot and small volume measurements) in a relatively short time interval (seconds to minutes rather than the minutes to hours of deep towing or vehicle surveys). Furthermore, acoustic techniques for determining flow rates, temperature fluctuations, and particle concentrations already existed or were being developed (Weston, 1959, Orr and Hess, 1978, Jackson and Dworski, 1992). In short, the anticipated advantages of acoustic imaging included the definition of plume boundaries to estimate entrainment and injection into the ocean (useful as the dilute plume is difficult to see at 10's of meters above its source with cameras), the estimation of particulate motion along sonar imaging direction with potential to estimate internal plume motion (that is flow rates and fluxes), and the estimation of particle concentration to investigate suspended particle distribution (Palmer et al., 1986).

In this paper, we will establish that acoustic imaging has three advantages: (1) estimation of the vertical velocity distribution within the lower buoyant plume provides volume flux and heat flux estimates for an integrated portion of the hydrothermal site; (2) a modification of the acoustic imaging approach can detect diffuse discharge over a significant (hundreds of m²) area; and (3) the remote sensing nature of acoustic imaging data combined with visualization techniques effectively captures the quantitative distribution of both buoyant plumes and diffuse discharge over a significant (hundreds of m²) area.

While the development of acoustic imaging was mainly motivated by its ability to quantify properties of hydrothermal discharge, acoustic imaging also has the potential to communicate aspects of the spatial distribution (and temporal variation) in a 3D visual mode. Visualization of the acoustic data facilitates the perception and quantification of spatial and dynamic variations in hydrothermal discharge. As a result, the development and application of 3D visualization techniques accompanied the development of the acoustic techniques. The visualization process starts with data acquisition, proceeds to data processing and ends with visual rendering techniques; the development of acoustic data processing and illustration methods becomes intertwined with the development of visualization methods. This paper will thus trace the contemporaneous developments in scientific visualization that have enabled both effective presentation and three-dimensional quantitative analysis of the acoustic imaging data.

This paper traces the following: (1) the development of acoustic imaging as a tool for exploring hydrothermal plumes from the initial adaptive use of navigation sonar to the use of a sonar specifically designed for attachment to a cabled underwater observatory; and (2) the related development of visualization applications for illustrating and quantifying hydrothermal fluxes. Table 1 presents an overview of the developments discussed in the following sections. Section 2 covers the initial sonar detection of hydrothermal plumes and the mechanisms of scattering enabling detection. Section 3 discusses the acoustical engineering developments. Section 4 details the visualization and quantification developments, including changes in data processing leading to additional scientific information. Section 5 highlights the key scientific results concerning hydrothermal plumes. Section 6 gives a brief overview of COVIS and the scientific information it is capable of acquiring. Section 7 provides some concluding thoughts.