Modeling the transport of passive tracers in sea ice

Abstract

A model for the vertical transport of passive tracers in sea ice is presented. Two Reynolds flux closure approximations are proposed for the tracer velocity dispersion term in the new ice tracer model (IceT). The schemes, "enhanced molecular diffusion" (EMD) and "mixing length diffusion" (MLD), are suggested parameterizations of gravity drainage, the dominant mechanism of desalination in growing sea ice. The performance of IceT using each parameterization is tested against two experimental studies: simulations I compare profiles of brine salinity with passive tracer concentrations during sea ice growth and melt, and simulations II compare bottom boundary salt and brine volume fluxes with passive tracer fluxes and bottom diffusivities. IceT behavior is evaluated under three salinity forcing prescriptions. Both schemes successfully reproduce expected tracer concentrations in simulations I when the salinity forcing evolves logarithmically. However in simulations II, the MLD scheme outperforms EMD in two significant ways: 1) growth rate trends in the MLD diffusivity parallel measurements of brine volume flux, while EMD bottom diffusivities remain almost constant; and 2) MLD bottom tracer concentrations decrease with ice growth rate similar to excluded brine salinity, while EMD trends are contrary. Hence, enhanced molecular diffusion is highly questionable for implementation in models of reactive passive tracers. The IceT mixing length diffusivity is a viable choice in sea ice biogeochemical modeling. It remains a challenge to assess whether MLD can provide quantitative estimates of desalination at each time step for use in forcing IceT or for implementation in a model of sea ice salinity. Copyright 2011 by the American Geophysical Union.