| Phone: | +1 (541) 753-6695 |
| Fax: | +1 (541) 753-1999 |
| Phone: | +1 (206) 726-0501 |
| Fax: | +1 (206) 726-0524 |
|
||||
| Laurence Padman | ||||
| ||||
| Robin D. Muench | ||||
|
This page summarizes the specific contributions of ESR personnel to AnSlope. More general information on the multi-institution AnSlope program, including goals, objectives and an overview of activities, can be found here.
We attempt a quantitative description of vertical mixing through analyses of a variety of data types: CTD data (Thorpe scale analyses); CTD-based microstructure; and microstructure measurements from a tethered, free-fall profiler.
In principle, a CTD can measure turbulence-related hydrographic (T and S) structure, provided the turbulence length scales greatly exceed the resolution of the CTD sensors. In practice, preparing CTD data for this type of use is a delicate art, and we are not yet convinced that this is a useful approach to making turbulence observations. However, AnSlope provides an excellent test environment: the CTD descent rate is relatively constant because there is no wave-induced line heave on the CTD package when the ship is in high-concentration sea ice (as in most of AnSlope Cruise 1); and the turbulence levels in the benthic boundary layer are known from even a casual look at the data to be extremely high, with large vertical scales of turbulence.
Thorpe-reordering analysis of density data is one approach to finding the length scales of turbulence (see Thorpe, 1977; Dillon, 1982 ). Once the length scale L for an identified turbulent patch is determined, the vertical diffusivity Kz can be estimated as a function of L and local buoyancy frequency N. We will apply these methods to CTD data from AnSlope Cruise 1. Our primary goals is to estimate Kz for AnSlope, however the study will also serve as a test of the limitations of this approach for application to other CTD data sets.
CMiPS was designed and built by RGL Consulting (Rolf Lueck, Principal) in Victoria BC, Canada, for use in Southern Ocean GLOBEC. The package mounts on the CTD rosette (see photo) and collects pressure (P), temperature (T) and conductivity (C) data at small scales. The sampling rate is 512 Hz, so that at a typical CTD descent rate of ~0.5-1 m s-1, CMiPS samples every ~1-2 mm. Actual resolution is limited by the sensor characteristics, but is roughly 1-5 cm. (Oversampling allows some reduction in electronic noise.) These scales resolve the inertial subrange of the scalar gradient spectra where ocean mixing is significant.
CMiPS was deployed on nearly every CTD cast during AnSlope Cruise 1 (Figure 1; also see the Cruise Report). In combination with the LDEO lowered ADCP ("LADCP"), CMiPS provided detailed information on the turbulence of the benthic "plume" of outflowing High-Salinity Shelf Water (HSSW).
ESR personnel will use our VMP on AnSlope Cruise 3, in late 2004. The VMP offers a different view of turbulence. Compared with CMiPS, the VMP provides superior-quality data of microscale T and C, since the instrument falls independently and thus without the CTD vibration and line motion (primarily vessel heave when waves are significant) contaminating the signals. The VMP can also measure microscale velocity shear fluctuations, which allow a direct calculation of the turbulent kinetic energy dissipation rate (ε ). However, the VMP cannot reach great depths because of limitations on the tether length caused by both winch capacity and lateral drag on the tether cable from ship-relative currents.
The decision to attempt deployment of the VMP on Cruise 3, rather than CMiPS on Cruise 2 or 3, is based on the high quality and exciting science already captured by CMiPS and the CTD/LADCP package during Cruise 1. We judge that there is only a modest potential scientific value in obtaining more data similar to Cruise 1 CMiPS, but that VMP data might provide a different and valuable perspective on mixing along the shelf break.
The need to consider tides in the AnSlope study area was demonstrated by our Circum-Antarctic Tidal Simulation version 02.01 ("CATS02.01") model, which predicted spring tidal currents exceeding 1 m s-1 near the shelf break in the NW Ross Sea (see Padman et al. (2003). Measurements made by vessel-mounted ADCP (VM-ADCP), the Lowered ADCP (LADCP) on the CTD rosette frame, and short-deployment moorings all support the model's qualitative view of the time-averaged intensity of the shelf-break currents. It is not so simple, however, to model the time-dependence of these currents at sufficient accuracy to allow the tide to be extracted from VM-ADCP and LADCP records. Most tidal kinetic energy along the Ross Sea shelf break is in the form of diurnal topographic vorticity waves ("DTVWs"); see, e.g., Middleton et al. (1987), Padman et al. (1992), and Padman and Kottmeier (2000). DTVWs have very small across-slope scales of variability (of order the slope width (~10-50 km)) and similar along-slope wavelengths. Their dispersion characteristics are sensitive to mean flows, stratification (even if it is too weak to create significant baroclinicity in the wave), and bathymetric structures on small scales. Barotropic tidal models that ignore mean flows and stratification, and have poorly known bathymetry, therefore struggle to accurately model DTVWs.
Nevertheless, to achieve the AnSlope objectives, we must be able to model tides accurately over a relatively large domain, i.e., the entire survey area rather than simply the small region surrounding the current meter moorings. Once the TAMU/OSU moorings are recovered and analyzed, we will have very good tidal information in the region covered by the moorings. However, modeling the tidal currents over a larger region requires improvements in our tides model. We will achieve this primarily by data assimilation, whereby available current meter data (ADCP and moorings) can be used to nudge a "prior" dynamical model into compliance with the data. At the end of the AnSlope field program we will have many moored current meter records (and one tidal height gauge) for this study, as well as 3 cruises of VM-ADCP data over a broader modeling domain. Limited data available as of late 2003 (mostly AnSlope Cruise 1 VM-ADCP and LADCP) suggests that tidal currents are indeed primarily barotropic, so that an assimilation model based on the depth-integrated shallow water equations is probably a sufficient first step.
We have already made one attempt to improve the quality of our tide model for the AnSlope study region. A paper was submitted in November 2003 to the Journal of Atmospheric and Oceanic Research (Erofeeva et al., 2003), detailing a full Ross Sea tidal inverse model in which 3 cruises of VM-ADCP data have been assimilated. Contact Laurie Padman to request a copy. The model has been compared with a short-duration mooring that was deployed and recovered during AnSlope Cruise 1 ("Central-E" on the central slope, near the 1400 m isobath), and appears to be performing quite well (Figure 2), although further improvements are necessary.
From an oceanographic perspective, the most important feature of the tide is the magnitude of tidal currents . As a guide to tidal kinetic energy, we calculate the "mean tidal current speed" |U|mean (Figure 3). |U|mean exceeds 50 cm s-1 along much of the northwest Ross Sea shelf break, implying maximum tidal currents exceeding 1 m s-1 at spring tides. (For a diurnal-dominated region, spring tides occur when the K1 and O1 tidal harmonics are in phase, roughly every 2 weeks.)
Preliminary analyses of AnSlope Cruise 1 data suggest that tidal currents play 3 distinct roles in the AnSlope region:
We postulate that if mixing were the dominant contribution from tides, the diluted HSSW may be insufficiently dense to contribute to deep ocean hydrographic properties and circulation. On the other hand, tidal advection may act to improve deep-ocean ventilation by HSSW. ESR investigators will continue to study the influences of tides on the cross-slope exchange and mixing of water masses in the AnSlope region. We note that the AnSlope-area tides are qualitatively similar to tides in the southern and western Weddell Sea (see Padman and Kottmeier, 2000), which is another significant source for dense water entering the global ocean deep circulation. Thus, the AnSlope studies will also contribute to our understanding of environments that are colelctively responsible for well over half of the input of dense water from Antarctic shelf seas into the World Ocean.
The current Ross Sea ADCP tidal inverse model ("RossTIM") is available with a Matlab Graphical User Interface (GUI) driver, as well as simple scripts for batch processing of tidal predictions and extracting tidal amplitude, phase, and current ellipse properties.
To obtain RossTIM, please follow these Download Instructions.
ESR's contribution to AnSlope is funded by grant OPP-0125602 from the National Science Foundation, Office of Polar Programs. The purchase of the VMP, proposed for use in AnSlope Cruise 3, is funded by a Major Research Instrumentation (MRI) grant, number 0320622, from NSF.
The CMiPS instrument is owned by Raytheon Polar Services (RPSC), who provided all the logistic support necessary to run it during AnSlope. The excellent performance of CMiPS during AnSlope Cruise 1 depended heavily on Chris MacKay's patient and methodical handling of the instrument, and the able assistance of the RPSC Marine Techs, in particular, Steve Tarrant. The large volume of data from CMiPS was managed on the ship's network and backed up by the RPSC Electronics Techs. Jay Simpkins and Kathryn Brooksforce (OSU) ably carried out mooring operations in extremely difficult ice conditions. Alex Orsi (TAMU) kindly provided the current meter data from the short-term moorings during Cruise 1, used in Figure 2 for tide model validation. Grateful thanks also to Karl Newyear (Raytheon MPC) for facilitating the logistics associated with CMiPS use in AnSlope, and to Captain Joe, the officers and crew of N. B. Palmer, for getting us through the ice to where we wanted to go, and providing a great ship to spend time in.
Tide modeling at ESR is a collaborative venture with Oregon State University. We thank Gary Egbert and Lana Erofeeva for their contributions to the Ross Sea tidal studies. Additional contributions to tides modeling and web maintenance are provided by Susan Howard at ESR.
Photograph of the nose of CMiPS mounted on the CTD Rosette.
Figure 1: Cruise track and CTD locations for AnSlope Cruise 1. Selected CTD stations are numbered. Intensive CTD/LADCP/CMiPS work was focused around the moorings (yellow dots) near 72oS, 172oE. The red line is the approximate location of the Ross Ice Shelf front. Click on the figure for a clearer (*.png) version.
Figure 2: Comparison of north-south velocity component (red) from a mid-depth current meter (nominally at 407 m below the surface in ~1400 m of water) on the AnSlope Cruise 1 short-term Central-E mooring, with predictions from a dynamics-only tide model CATS02.01 (blue) and the Ross Sea tidal inverse model for which AnSlope Cruise 1 VM-ADCP data have been assimilated (green). Current meter data are provided courtesy of Alex Orsi (TAMU) and Dale Pillsbury (OSU). The inverse model performs significantly better than the dynamics-only model during spring tides (~days 70-77). During neap tides (~days 63-68), neither model accurately captures the diurnal tidal variability, suggesting that at least one component is being poorly modeled. Additional assimilation data (VM-ADCP from AnSlope Cruises 2 and 3, and from the moorings themselves) should improve the inverse solution. Click on the figure for a clearer (*.png) version.
Figure 3: Map of mean tidal speed (|U|mean)for the AnSlope region, based on the CATS02.01 tidal model. The color scale ranges from 0 to 50 cm s-1. Maximum ("spring") tidal currents are about 2|U|mean. The 500, 1000, and 2000 m isobaths are contoured in white. The red line is the approximate location of the Ross Ice Shelf front. Click on the figure for a clearer (*.png) version.