u CHAPTER TWO ICE-OCEAN THERMODYNAMICS ‘The major forcing on ice shelf evolution that is considered here is the melting and freezing processes that occur beneath the ice shelf, Processes that are occur internally to the ice shelf, such as strain thinning and ice shelf calving, are not considered in this thesis. As such, the purpose of this chapter is to develop a set of equations to describe ice-ocean thermodynamics to be used as a basis for the rest of the thesis. The well defined ice shelf-ocean thermodynamics are briefly summarised before a longer review of the current state of frazil ice-ocean thermodynamics. ‘The formulations for the ice-ocean parametrisations that are shown in this chapter are tested in simplified cavities (Chapter 4) and for the Amery Ice Shelf cavity (Chapters 5, 6 and 7), All known published forms of ice-ocean interactions are derived from Fourier’s Law, where the conductive heat transfer rate q. for a given surface area A, is given by [Pitts and Sissom, 1977]: Ge _ _, OP Aan’ Wm Qn Here, 3© is the temperature gradient in the direction normal to the area A. The thermal conductivity i is an empirically determined constant for the medium, and may depend on other properties such as temperature and pressure. The minus sign in Fourier’s law is required by the second law of thermodynamics that is thermal energy transfer resulting from a thermal gradient ‘must be from a warmer to a colder region. Eqn 2.1 is appropriate for ice growth in freshwater and care must be given when examining ice growth rates in saltwater. 2.1. Ice Shelf-Ocean Dynamics At the ice-ocean shelf interface, a parameterisation with a viscous sub-layer model is used (see Fig. 2.1). Three equations represent the conservation of heat and salt and a linearised version of the equation of freezing point of seawater as a function of salinity and pressure [Holland and Jenkins, 1999], ‘The free variables that are found by solving the three equations simultaneously are temperature, Tp, salinity, Sp, at the ice shelf base, and the melt rate, m. The parameterization of the fluxes of heat and freshwater that occur at the ice shelf-ocean interface is effective due to the difference in time-scales between the slowly flowing ice shelf (for example, 1x 10~5 m s~!) and the faster sub-ice shelf ocean current (for example, 1x10! m s~!). This approach allows the spatial pattern of melting and freezing at an ice shelf base to be simulated. The assumption is that the ice shelf is in a steady-state balance with respect to sources and sinks of mass and heat. The conservation of heat and salt are: pi(L — GAT)m = powrr(Ts-T) (2a) pi(AS)m = pys(Se— $) 22)2 CHapTER 2 Ice shelf qT S=0 ‘Viscous sub-layer Ty=T, Ss uu Figure 2.1: _ Schematic showing the ice shelf-ocean thermodynamic processes T and S are the water temperature and salinity, with subscript b indicating the condition at the boundary and subscript i in the ice. The free variables are Ts, which is the temperature at the ice shelf base, Sp, which is the sainity atthe ice shelf base, and m , which is the melting (m < 0) or freezing (m > 0) rate (m s~1), The symbols are further described in-text. where, p; is the density of ice (assumed to be 916 kg m7), p is the density of ocean water, L is the latent heat of ice fusion (3.35 x 10°), cj and cy are the specific heats of ice (2009 J kg! K-) and water (3974 J kg“! K~), respectively. AT is the temperature difference between the ice shelf interior, 7; (-20 °C ), and the freezing temperature at the base of the ice shelf, Ty. T is the temperature and S the salinity of the water away from the base of the ice shelf base. The assumption is that no salt is present in the ice shelf, which simplifies eqn 2.2b to, piSsm = pns(Sb — S) @3) The parameters ‘yr and + are coefficients that represent the transfer of heat and salt across the boundary layer. Jenkins [1991] used a molecular sub-layer approximation to formulate expressions for yr and 7s as: ua 2A2In(ugh/r) + 12.5Pr28 —9 Ud 212In(ugh/v) + 12.5827 —9 or (2.4a) a (2.40) where, the molecular Prandt! number (Pr) is the ratio of viscosity to thermal diffusivity and the molecular Schmidt number ($c), is the ratio of viscosity to salinity diffusivity. The kinematic viscosity of seawater, v (1.95 x 10-© m? s~!), is considered constant [Holland and Jenkins, 1999] over the thickness of the boundary layer, h. The friction velocity, ug, is defined in terms of the shear stress at the ice-ocean interface: ud = equ? (2.5)FRAZIL ICE-OCEAN DYNAMICS 13 where cg is a dimensionless drag coefficient (0.0025) and w is the velocity of the ocean. Note that the ice is considered to be stationary. The linearised version of the freezing point of seawater as a fun Ty =aSy+b+cP (2.6) where, Ty is the freezing point at the ice-ocean interface, a is the slope of liq (—5.73 x 107? °C psu~!), b is the offset of liquidus for seawater (8.32 10~? °C ), cis the change in freezing temperature with pressure (—7.61 x 104°C dbar~'), and P is the pressure at the ice shelf base Eqns 2.2a, 2.3 and 2.6 can be solved simultaneously (using known mixed layer and ice properties) to calculate heat and freshwater (salt) fluxes into the ocean (Schedwikat and Olbers, 1990; Hellmer et al., 1998; Holland and Jenkins, 1999). This has been done previously for several simulations of the flow beneath ice shelves [for example, Beckmann et al., 1999; Timmermann et al., 2002; Holland et al., 2003]. The calculation of the actual heat and salt fluxes into the top model layer of the ocean includes the meltwater advection term that can be important in long simulations or with high basal melt rates [Jenkins et al., 2001]. Simplifications of these equations can be made by assuming that the melt/freeze rates is only driven by the temperature gradient; that is, by setting S$ = Sy in eqn 2.2b. This assumption reduces the equations to the “two equation” parametrisation as has been discussed in Holland and Jenkins [1999]. 2.2 Frazil Ice-Ocean Dynamics Frazil is important to sub-ice ocean dynamics and overall glacial ice mass balance for two reasons described by Smedsrud and Jenkins (2004): (1) Frazil ice growth is thought to be a more effective sink for supercooling than is the growth of columnar ice directly onto the ice shelf base. (2) The presence of suspended ice crystals makes the ISW more buoyant. The formation of frazil ice thus modifies the forcing on the overturning circulation within the cavity, which determines the location and rate of marine ice accumulation at the ice shelf base [Jenkins and Bombosch, 1995} 2.2.1 Fundamental equations Frazil laden water is considered to be a two-component mixture of ice and seawater that is treated as a homogeneous fluid with spatially-averaged properties. The total mass of frazil ice crystals within a fluid parcel is, dMe = (aM)C, Qn where dM = pdV is the total mass of the parcel of seawater (mass of water plus mass of frazil ice), and C is the mass fraction of frazil ice. For a unit volume of mixture the mass of the seawater fraction is pg/(1 —C) and the mass of the frazil ice fraction is p;/C. For a Boussinesq ocean model, the bulk density of the seawater is,