|This manuscript by L. Wu and X. Chen examines the motion of a tropical cyclone (TC) in a numerical experiment using a potential vorticity (PV) tendency framework. The authors set out to investigate two allegedly open fundamental questions about storm motion: 1) Why steering by environmental flow plays the dominant role in TC motion and 2) When TC motion deviates considerably from the environmental steering. It is my impression that the authors basically revisit their own study from (Wu and Wang, 2000), which was restricted by a horizontal resolution of the underlying numerical experiment inadequate to represent TC inner-core processes sufficiently well.|
The manuscript contains interesting aspects but over large parts it remains elusive to me what the new insight gained by the authors’ analysis actually is. Partly, this is because the manuscript lacks clarity in a) describing the current state of understanding of TC motion and b) introducing their PV tendency methodology. Mainly, I miss a rationale for dividing the horizontal advection term into different contributions and linking these individual contributions to processes involved in the current understanding of TC motion.
I believe that the TC community’s answer to question 1) is that a TC can be approximated by a point vortex that is advected by the environmental flow (e.g. Emanuel, 2005, “Divine Wind”). This approximation implies that a) PV tendencies are dominated by horizontal advection and b) that the TC’s (horizontal) PV structure and changes thereof do not play a leading-order role in TC motion. Why would this well-known picture be in question? Deviations of this leading-order picture have been demonstrated in the context of track deflection near Taiwan, tilted TC vortices, and trochoidal motion (see references in the manuscript), i.e. these studies give answers to question 2) above. For this manuscript to be suitable for publication, the authors need to clarify to the reader how their PV tendency analysis improves the current state of knowledge.
Before suitable for publication, the presentation needs to be improved considerably, i.e. several further, partly major revisions are needed.
* A secondary eyewall cycle is described in some detail early in the manuscript but is not referred to later in relation to TC motion.
* The terminology of steering, conventional steering, secondary steering, etc. is confusing.
* A rationale for dividing and approximating the horizontal advection into HA1 and HA2 – it looks akin to a linearization – needs to be given and the calculation of the TC motion C in Eq. 1 needs clarification.
* Importantly, the authors consider finite differences over 2 hours to determine TC motion, which sets the timescale of the resolved deviations from steering. Why 2 h?
* I cannot follow the authors’ distinction between layer-wise steering and the attempt to find a vertical average that best represents the steering flow for the TC as different processes. Velden and Leslie (1991) and Galarneau and Davis (2013), e.g., consider a PV(or pressure)-weighted vertical average as the best way to define a steering flow, making explicit the idea that steering is governed by horizontal PV advection on individual levels.
* Fig. 5 seems incorrect as the individual motion components seem not to add up to the motion speed.
* What is angular RMS in the caption of Fig. 6?
* Compensation between HA1 and HA2 (pg 13/14): Is this basically saying that the wavenumber 0 and 1 PV structure of the TC is stationary (in the storm-relative frame of reference)? If yes, this observation is inconsistent with reference to the argument of vorticity stretching/ compression (Bender, 1997) on pg 15. Recently (Riemer, 2016), a similar compensation between advection terms has been described, consistent with your observation. Riemer (2016) argues that Bender’s mechanism is not at play when compensation occurs between symmetric and asymmetric vorticity advection.
* Fig. 8: It is unclear to me what is shown in the individual panels. There are 4 panels but only labels a) and b). In addition, terms HA1 and HA2 need to have the same units.
* The correlations found between individual terms of the PV tendency equation are rather low. Are they statistically significant? If not, I would argue against a physical interpretation of the potentially spurious correlations.
* Section 5 describes the occurrence of trochoidal motion in the authors’ experiments and analysis the motion using the PV tendency framework. It is unclear to me, however, what the novel insight is that would be gained in this section.
* line 348: Do we miss a “sink”?
* line 356: Unclear what sort of spectral analysis was conducted.
* Fig. 12 misses units at the axes.
Galarneau Jr, Thomas J., and Christopher A. Davis. "Diagnosing forecast errors in tropical cyclone motion." Monthly Weather Review 141.2 (2013): 405-430.
Riemer, Michael. "Meso‐β‐scale environment for the stationary band complex of vertically‐sheared tropical cyclones." Quarterly Journal of the Royal Meteorological Society (2016).
Velden, Christopher S., and Lance M. Leslie. "The basic relationship between tropical cyclone intensity and the depth of the environmental steering layer in the Australian region." Weather and Forecasting 6.2 (1991): 244-253.