|I appreciate the authors' effort in substantially revising this manuscript. In my opinion, the discussions have improved. Nevertheless, there are still some issues that the authors should address before the paper can be accepted.|
1) My main remaining issue is the authors' contention that the air-sea fluxes act to decrease baroclinicity (lines 189-200). I pointed out in my previous review that the fact that baroclinicity decreases while the FT increases does not imply the flux causes decrease in baroclinicity, as the synoptic eddy is growing when FT increases, and the synoptic eddy growth could be the cause of the decrease in baroclinicity. In the response and in the revised manuscript, the authors acknowledged that they can't separate the effect of the heat flux from the effect of the eddy growth in reducing baroclinicity. Nevertheless, they still contended that since the air-sea flux damps the temperature variance, this provide strong evidence for the eddy flux damping baroclinicity. I don't really follow this point how the air-sea flux damping eddy amplitude would imply it would erode the MEAN temperature gradient. This needs to be better explained. This also relates to the discussions in the paragraph on lines 210-213. I can see how the air-sea flux is important in the budget of eddy APE, but I can't really see any "strong" evidence (apart from Fig. 5, which I have argued that this does not show causality) that these fluxes deplete the mean baroclinicity.
2) Lines 32-33: Note that as I pointed out in my previous review, Chang et al (2002) showed that total diabatic generation of transient EPE is largely negative except over parts of the Pacific, with the damping effect of sensible heating largely cancelling the positive generating due to latent heating. In fact, Chang and Zurita-Gotor (2007, JAS 64, 2309, see their table 1) have shown that over the northern hemisphere mid latitudes, and especially in the Atlantic, diabatic heating damps transient eddies. These results do not contradict those cited in this study (Li et al, 2007; Marques et al., 2009), because in Lorenz's definition of EAPE, eddies are defined as deviations from the zonal mean, and includes stationary eddies. In winter, net diabatic heating is positive over the warm oceans, and negative over the cold continents (e.g. Held et al., 2002, J. Climate, 15, 2125, Fig. 8), leading to generation of EAPE but mainly for stationary eddies and not necessarily for transient eddies that form the storms. My point is that there exists previous studies that showed that diabatic heating acts to damp baroclinic waves instead of being a source of energy in storm development (lines 33, 337-338).
3) Equation 1. Here the authors should discuss why they used 850 hPa T' rather than 2m T'. This is discussed below (starting line 148) but in my opinion should be discussed here instead.
4) Lines 112-114: I still don't understand how the effects of ocean eddies on F at the resolved scales would be capture by the reanalysis system. As far as I know, F is model generated, hence if the model does not resolve the ocean eddies in the SST, how can F contain the effects of that? You could argue that the effects of the ocean eddies on the resolved scale temperature structure may be captured since temperature is an observed input that is assimilated, but I am not convinced that the effects of ocean eddies on F can be captured by the model. Either remove this discussion or explain more clearly.
5) Line 127: The heat flux anomalies seems to be skewed towards positive values to me?
6) Lines 155-156: Perhaps the near surface temperature anomalies are more strongly damped by the underlying SST may also contribute?
7) Line 162: "derives from a gain/loss of signal due to averaging" - perhaps add something like "partially"? Not clear to me that all of the signal is due to that as the authors also argued below.
8) Lines 174-175: I don't quite understand the phrase "it is reasonable to interpret any positive instances or moderately negative values as indicative of a relatively weak heat exchange" means.
9) Fig. 3 shows that the covariance between F and T is positive along the Canadian coast. Can the authors explain that?
10) Line 182: phases should be phase?
11) Lines 189-190: The Eady growth rate maximum should probably be defined?
12) Line 196: "to reduce" should be changed to "to decrease"
13) Line 254: "consistently" should be "consistent"?
14) Line 317: As discussed in item 1), I'm not convinced that this is the case.