Submarine Estimates of Arctic Turbulence Spectra (SEATS)

The arctic ocean and it's peripheral seas comprise a seasonally variable mix of ice covered and ice-free open ocean and shelf waters that are considered to be one of the most sensitive regions to global change processes. Current estimated rates of multi-annual sea-ice loss appear to be exceeding model predictions. Despite the insulation from direct atmospheric forcing, mesoscale baroclinically unstable eddies (the oceanic's internal 'storm scale', 10-100 km) have been clearly observed through repeated salinity and temperature against depth profiles deployed through ice holes and more recently by the deployment of ice-tethered profiling instruments. Measurements of a turbulent well mixed boundary layer under sea-ice have been examined, establishing a mixed layer thickness closely coincident with an ice speed related frictional drag. For the same reason that different processes lead to the establishment of a turbulent mixed layer under ice compared to open waters, we expect that different mechanisms will lead to mixed layer re-stratification and therefore that the sub-mesoscale (1-10's km) interaction with mesoscale eddies and frontal boundaries will vary between ice-free and ice-covered regions. The aim of this study is an improved understanding of the characteristics of mixed layer variability in marginal or seasonally variable ice-covered regions, through the analysis of a large pre-existing data resource.

Royal Navy (RN) submarines have collected temperature and salinity data in the Arctic, and elsewhere for a long time. For over 15 years RN submarines have been equipped with a sensor suite that also includes a number of biogeochemical instruments particularly for pigments such as chlorophyll(a) and 'yellow substance'. These data are normally highly classified and unavailable for research purposes outside and even largely within the MOD itself. This is because traditional research requirements include acquisition time/date and position. However, only vessel velocity, elapsed time (instrument sampling interval) and depth, within well known published ranges, are required to accompany the data-streams for us to examine the spatial cross correlation coefficients in the physical and biological parameters. This is a mechanism for looking at data collected spatially and temporally in terms of the time and space scales of the inherent variability contained within it. Specifically, discussions with our MOD collaborators indicate that significant amounts of data will be available both in the mixed layer and deeper. The wavenumber spectrum is simply proportional to the Fourier transform of these cross-correlation coefficients, and quantifies the contribution to variability from features at a range of length scales, including eddies and filaments. For general underway vessel velocities, the 'speed' of the observer is such that the time taken to traverse a structure is very much less than the time taken for its evolution for all but the smallest structures. The disturbance scale of the vessel O(100 m) will limit our study to sub-mesoscale filaments and above (~0.2-100 km). A key issue we can address is how closely related the spectra of other properties are to the physical spectra.

The SEATS project will make the first direct estimates of turbulence at the sub-mesoscale to mesoscale across the arctic basin. The 10-15 year history and wide spatial coverage of a very large, currently untapped, RN submarine dataset, makes this a valuable study of both temporal change and regional variation in regards to ice-coverage of this rapidly changing ocean basin. The climate modeling community, both at NOC and within the UK as a whole will be the first direct beneficiary as the derived co-spectra of buoyancy, and buoyancy against potential temperature and salinity, will support the implementation of new mixed layer parameterisations for coupled models such as Fox-Kemper et al. (2010). Specifically these spectra will indicate if, and under what conditions, the spectral slope is significantly different from k**-2. The spectra will also act as a valuable test for model comparisons at different spatial resolutions and both mixed layer and sea-ice parameterisations.

The earth system modeling community is another direct beneficiary. Even if the spectral slope for buoyancy and its components scales as k**-2, the processes that make up this turbulent spectrum are thought to have a non-linear impact on biogeochemical cycling and ecosystem progression partly because the inputs to the systems are very different and partly because they depend on biological growth terms as well as fluid mechanics. Within the NERC arctic programme ecosystem models will be coupled into the various NEMO basin and global scale model runs provided both by NOC national capability and other discrete targeted projects. Our observed co-spectra of biological parameters against buoyancy and temperature will provide a simple and objective comparator for model assessment.

The existing close links between the Ocean Biogeochenistry and Ecosystems group and the Ocean Modelling and Forecasting group at the NOC will be used to invite key members of the UK's modeling community, including the Meteorological Office, to the workshop on the use of upper ocean turbulence spectra to be held by SEATS in July 2013. The PI's and CoI's close involvement in the preparation of a consortium bid to the NERC's Ocean Surface Boundary Layer programme, will be used to expand this invite list to colleagues involved in large eddy simulation modeling and the observations of small scale wind induced turbulent processes such as Langmuir circulations, right at the highest wavenumber end of the SEATS project.
Indirectly our study will benefit the MOD and DSTL in providing a route to export for a valuable data source otherwise hampered by security issues; and it is expected that similar opportunities may exist elsewhere in military data gathering. The benefit from this to the MOD and DSTL is both socio-political, demonstrating their continued enthusiasm for playing a role in UK environmental monitoring as a whole, and military, improving operational forecast modeling through a better objective understanding of the impact of non-linear processes on regional scale forecast skill. Our routes to MOD and to the UK marine instrument manufacturing industry are already in place through the OBS (Observational Biogeochemistry for Submariners) Knowledge Exchange project. Discussions between NOC, DSTL, the Royal Navy and UK industry have highlighted where UK advantage can be gained from further real-time exploitation of routine underway data through expert scientific guidance.

SEATS will be readily available to both direct and indirect beneficiaries as a result of it's short time frame. Having a dataset already present, but as yet un-tapped, the first results from SEATS will be available for use by the beneficiaries within 6 months of the beginning of the project (i.e. April 2012). The final analysis and report, including initial reports on comparisons with models, will be available by October 2013, with nearly two years of the arctic programme as a whole left to run.