Thermal and photo-thermal stresses are major considerations in managing coral reef ecosystems. Physical insights into sea temperature variability offer a way to identify reefs or larger areas that may be more prone to temperature extremes under varying light and weather conditions. Networks of autonomous reef monitoring stations, including both ICON/CREWS stations in the Caribbean and ICON/SEAKEYS stations in the Florida Keys, measure hourly wind speed and direction, barometric pressure, air and sea temperature, and in some cases dew-point temperature, tide height, and incident light, generally on the shallow reef crest or promontories near the edge of the deeper ocean. These quality-controlled, in situ data have been combined with high-resolution atmospheric reanalyses from the NOAA National Centers for Environmental Prediction (NCEP), together with models and satellite data, to estimate surface radiative and turbulent heat fluxes and heat advection for monitored sites. This ocean heat budget is used to evaluate the effectiveness of bulk formulae and reanalysis in explaining observed sea temperature variability at coral reef sites. Based on the heat budget, ICON researchers find that another dynamic process, different from either direct air-sea heat flux or larger-scale advection, accounts for a significant portion of sea temperature variability at intraday and longer periods. This dynamic process is horizontal convection – thermally induced exchange currents between the crest and the deeper waters of the reef slope, also known as the thermal siphon (Monismith et al., 2006).

A manuscript currently in preparation compares total heat budget predictions to observed sea temperature variability at several sites in the Florida Keys. A total ocean heat budget for shallow reef sites may be modeled as T_t = –u·∇_hT + (Q₀/C_ph) - u_HC·∇_hT_HC. Here T_t is Eulerian time-rate of change in sea temperature; u the horizontal ocean current; ∇_hT horizontal sea temperature gradient; Q₀ net heat flux; u_HC·∇_hT_HC is an additional heat advection term for horizontal convection; h varying water depth; and ρ and C_p density and heat capacity of sea water, resp. Net heat flux has five constituents: Q₀=γ·Q_SW+Q_LW+Q_SH+Q_LH+Q_RH. Short- and longwave radiative fluxes, Q_SW and Q_LW, are derived from NCEP reanalysis, with comparisons to available in situ data;factor γ models radiation absorption by the water column and reef floor. Tropical Ocean Global Atmosphere – Coupled Ocean-Atmosphere Response Experiment (COARE 3.0a) algorithms are used to estimate sensible (Q_SH), latent (Q_LH), and rain (Q_RH) heat fluxes directly from in situ data, using reanalysis for insolation, long-wave flux, and where in situ data are not available, for precipitation. Larger-scale heat advection is estimated using either the 1/25-degree Gulf of Mexico HYCOM experiment, or the operational Naval Research Lab Global HYCOM. Finally, the horizontal convection terms u_HC and ∇_hT_HC are calculated assuming unsteady thermal and momentum balances, as lagged functions of air-sea heat flux, Q₀, and sea-floor depths based on 3-arcsec NOAA NGDC gridded topographic products.

The figure below shows a schematic of the horizontal convection process (from Chubarenko, 2010) over a steep topographic slope, like that of the fore-reef in the Florida Keys.

Figure 1 from Chubarenko IP (2010) Horizontal convective water exchange above a sloping bottom: The mechanism of its formation and an analysis of its development. Oceanology 50:166-174. ISSN 0001-4370. ©Pleiades Publishing, Inc. Reprinted with permission of the author.

Below is a table from the 2006 paper of Monismith et al. cited above, using momentum / thermodynamic balances to estimate various scaling relationships between vertical shear ΔV from horizontal convective currents and several physical attributes. These attributes are bottom depth (D), bottom slope (β), characteristic turbulent velocity scale (q), mean buoyancy flux (note characteristic convective velocity term u_f is a function of net sea surface flux), and time scale of variability in the flux (T).

Table 1 from Monismith SG, Genin A, Reidenbach MA, Yahel G, Koseff JR (2006) Thermally driven exchanges between a coral reef and the adjoining ocean. Journal of Physical Oceanography 36:1332-1347. ISSN 0022-3670. ©American Meteorological Society. Reprinted with permission.