The spatial structure of the semidiurnal internal tide in the vicinity of Tasmania is characterized using temperature and salinity data from Seaglider and Slocum glider deployments. Wavelet analysis of isopycnal displacements measured by the gliders was used to isolate the semidiurnal internal tide, with a solid signal observed both to the east and to the south of Tasmania. The signal south of Tasmania was attributed to local forcing, while that to the east of Tasmania was found to have propagated from the south east to the north west—a result which supports previous studies indicating the presence of an internal tidal beam originating over the Macquarie Ridge, south of New Zealand. Displacement amplitudes were observed to be amplified in the vicinity of the continental slope, with the incoming tidal beam shown to be both reflected and scattered on the continental slope and shelf, and energy transferred to higher modes.

A dynamical understanding of the physical process of surface gravity wave breaking remains an unresolved problem in fluid dynamics. Conceptually, breaking can be described by inception and onset, where breaking inception is the initiation of unknown irreversible processes within a wave crest that precede the visible manifestation of breaking onset. In the search for an energetic indicator of breaking inception, we use an ensemble of non-breaking and breaking crests evolving within unsteady wave packets simulated in a numerical wave tank to investigate the evolution of each term in the kinetic energy balance equation. We observe that breaking onset is preceded by around one quarter of a wave period by a rapid increase in the rate of convergence of kinetic energy that triggers an irreversible acceleration of the kinetic energy growth rate. This energetic signature, which is present only for crests that subsequently break, arises when the kinetic energy growth rate exceeds a critical threshold. At this point the additional kinetic energy convergence cannot be offset by converting excess kinetic energy to potential energy or by dissipation through friction. Our results suggest that the ratio of the leading terms of the kinetic energy balance equation at the time of this energetic signature is proportional to the strength of the breaking crest. Hence this energetic inception point both predicts the occurrence of breaking onset and indicates the strength of the breaking event.

Abstract. The ocean mixed layer depth is an important parameter describing the exchange of fluxes between the atmosphere and ocean. In ocean modelling a key factor in the accurate representation of the mixed layer is the parameterization of vertical mixing. An ideal opportunity to investigate the impact of different mixing schemes was provided when the Australian Bureau of Meteorology upgraded its operational ocean forecasting model, OceanMAPS to version 3.0. In terms of the mixed layer, the main difference between the old and new model versions was a change of vertical mixing scheme from that of Chen et al. (1994) to the General Ocean Turbulence Model. The model estimates of the mixed layer depth were compared with those derived from Argo observations. Both versions of the model exhibited a deep bias in tropical latitudes and a shallow bias in the Southern Ocean, consistent with previous studies. The bias, however, was greatly reduced in version 3.0, and variance between model runs decreased. Additionally, model skill against climatology also improved significantly. Further analysis discounted changes to model resolution outside of the Australian region having a significant impact on these results, leaving the change in vertical mixing scheme as the main factor in the assessed improvements to mixed layer depth representation.

Scientific commentsOn a spatially dependent MLD definition. The reviewer questions our suggestion that the suitability of the MLD criteria thresholds may vary spatially. We adopted de Boyer Montegut et al., (2004) as a well performing and widely cited scheme. However, there are a number of clear points that would not support this as an optimal definition. As we discuss in section 3, the calculation of the MLD is complicated by the wide variety of temperature and density profiles observed in the ocean, making any MLD criteria susceptible to errors in certain circumstances. The temperature criterion is not set to be optimal in every global location or season but a single criterion that is "fairly successful" C1

Abstract. The ocean mixed layer depth is an important parameter describing the exchange of fluxes between the atmosphere and ocean. In ocean modelling a key factor in the accurate representation of the mixed layer is the parameterisation of vertical mixing. An ideal opportunity to investigate the impact of different mixing schemes was provided when the Australian Bureau of Meteorology upgraded its operational ocean forecasting model, OceanMAPS to version 3.0. In terms of the mixed layer, the main difference between the old and new model versions was a change of vertical mixing scheme from that of Chen et al to the General Ocean Turbulence Model. The model estimates of the mixed layer depth were compared with those derived from Argo observations. Both versions of the model exhibited a deep bias in tropical latitudes and a shallow bias in the Southern Ocean, consistent with previous studies. The bias however, was greatly reduced in version 3.0, and variance between model runs decreased. Additionally, model skill against climatology also improved significantly. Further analysis discounted changes to model resolution outside of the Australian region having a significant impact on these results, leaving the change in vertical mixing scheme as the main factor in the assessed improvements to mixed layer depth representation.

No abstract

No abstract

1 General comments A1 .On the temporal extent of the dataset. While the relatively short duration of the dataset had been mentioned in the manuscript, we have expanded upon the implications of this for our results and modified the title to include the season.A2. On the impact of model changes on MLD estimation. The reviewer questioned whether the impact on MLD performance of other changes between model versions can be assumed to have a minor impact. We agree that it is not possible to completely isolate these other impacts, but our analysis was designed to minimise the impact of changes other than the vertical mixing scheme. We also note that the configuration of C1