在以前的博文中我讲过，我的第一篇理论文章从开始投稿(1995年）到在美国大气科学<<J.Atmos. Sci.>>上发表花了10年（2005年发表）。

说实话，作理论研究是一项苦差事，国际上已没有多少人愿意干这事了，主要原因：一是这方面的课题拿到项目不容易；二是理论研究太难，而且即使文章发出来了，人们关注的也少，引用也不会太多；三是理论研究要使用假设，所以审稿人能找到的漏洞比较多，要说服审稿人不那么容易。所以现在作理论研究完全是凭兴趣。

2007年我与朋友写的关于海气相互作用的理论文章，到现在也还没发出来，快7年了。先是投JPO被拒，接着投JFM和QJ也被拒。直到两年前投DAO，差点被拒。编辑前后找了6个人审，我们同审稿人论战很多次。现在的两位审稿人已基本上同意我们的文章可以发表了，但还存在一些问题。下面是两位审稿人的审稿意见：

Reviewer #5: In my previous review, my concern was primarily on the nonlinear Ekman effect, because I was asked for it by an editor.  In the present review, my comments are not limited to the nonlinear Ekman effect.  I am sorry for this, but I believe additional comments will help improving your study further.(1) It is nice to include the nonlinear Ekman effect into your analytical model.  In the revised manuscript, the nonlinear Ekman effect and the SST-Wind coupling effect are discussed separately in 5.2.3 and 5.2.4.  This separation would be helpful for better understanding of each effect (though physical (not mathematical) aspects of each effect should be explained in more detail).  However, in real oceans, both effects work together.  There should be some conditions where stabilization due to the nonlinear effect is larger than the destabilization due to SST-wind coupling.  When such a condition is realized will be important if one wants to understand actual frontal phenomena in oceans.  Discussions on the stability of currents with both effects are necessary. (2) Growth rates are larger for larger wave number (Fig.1), showing that horizontally smaller perturbations are more preferred.  Because aspect ratios (D/H) of such perturbations are quite small, I think perturbations should be analyzed in non-hydrostatic system with some dissipation terms.  I am afraid that SG approximation is not appropriate for the instability described in this study.(3) Can we consider the most unstable perturbations as modified unstable waves of baroclinic instability?  Because of quite different nature of dependence of growth rates on wave number, it seems originated from another instability (such as modified symmetric instability, for example).  Without more detailed information than growth rates (e.g., structure of perturbation, source of perturbation energy, ...), readers could not understand "what will happen in frontal region under the SST-Wind coupling".  Validity of model configuration (hydrostatic approximation for example) can also be checked by these information.(4) The nonlinear Ekman effects in your study will be different from those investigated by Thomas and Lee (2005), because the former are induced by across-front wind while the later are caused by along-front wind.  More descriptions of perturbations of your instability are necessary.(4) Suggestion.  If you will start your explanation from perturbations with m=0 (no meridional variations) in section 5, then your explanation will become more clearer, because (a) nonlinear Ekman effect is absent and (b) difference between QG and SG will be minimized for m=0.  After describing m=0 mode, then you can extend your explanation to m=0 modes, which might be affected by other instabilities than baroclinic instability (I wonder).  The nonlinear Ekman effects in your analysis will also be made clearer.  

Reviewer #6: In this revised version of the manuscript, the authors have followed the recommendation of the reviewers. They now consider the non-linear Ekman transport in the SG case and discuss its impact on the stability of the SG model. They also detail more the impacts of the SST-wind stress coupling on both the SG and the QG models. I think this improves the manuscript considerably and that it is ready for publication pending few minor modifications. The writing is also better in this version and makes the paper easier to read, however there are still a lot of typos. Here are the typos I noted:Abstract: - Please specify which article by Spall you are referring to.- Ekamn should be Ekman. - There's a double parenthesis towards the end. Section 2: - The Rossby number (the article is missing before equation 2.1).- What do you mean by "As in the S07a"? Does this refer to Spall (2007)? In this case, you should specify before that Spall (2007) will be referred to as S07a in the rest of the manuscript and be consistent with the notation, it will make it easier for the reader to follow.Section 3: - Just after the geostrophic velocities definition, "<PHI> is geopotential", "the" is missing.- Before equation 3.14 "...boundary conditions are are reflected...".  Section 5: after equation 5.10, please modify the sentence "where k and m is wavenumber".