> 3. Differences between DNS and linear theory: The WKB approximation might
> be applicable to the system because k_f H_rho ~ 30 is assumed in the
> authors’ DNS, at least in the linear stage. Then, according to the linear
> theory given in Losada+12, the max growth rate of the NEMPI is expected to
> decrease with the colatitude, i.e., higher in the pole and lower in the
> equator because the frequency of the inertia wave becomes higher in the
> larger colatitude. Nevertheless, it seems that the growth rate of the NEMPI
> measured in the DNS does not show such a monotonic behavior (e.g., Figs.6
> and 13). The authors should clarify why it is so different from the
> prediction of the linear theory. I guess the fastest growing mode of the
> NEMPI may not be included in the range k < kf/6 (cut-off wavenumber) for
> some models , OR, maybe not fully-resolved by the DNS because the
> horizontal wavenumber of the fastest growing mode should be varied with the
> latitude. It might be better to do the convergence check at the lower
> latitudinal region.

The growth rate of NEMPI derived in Losada+12 in the presence of
rotation was obtained for the horizontal imposed magnetic field in
the absence of coronal envelope. In the presence of coronal
envelope the situation is more complicated. In particular, in the
presence of a coronal envelope, there is a change in the direction of
the large-scale magnetic field from the horizontal field at t=0 to
nearly vertical field. The growth rate of NEMPI for horizontal and
vertical fields are very different [see review by Brandenburg et
al. (2016)]. Also NEMPI for horizontal field is saturated very fast
due to "potato-sack" effect (A local increase of the magnetic field
causes a decrease of the negative effective magnetic pressure,
which is compensated for by enhanced gas pressure, leading to
enhanced gas density, so the gas is heavier than its surroundings
and sinks.). On the other hand, for a vertical field, the
"potato-sack" effect is absent and the operation of NEMPI results
in formation of strong concentrations [see Brandenburg et al.
(2013, 2016)]. An additional complication arises with increasing
rotation rate, when a kinetic helicity is produced by a combined
effect of uniform rotation and density stratified turbulence. It
results in the excitation of an $\alpha \Omega$ or $\alpha^2\Omega$
dynamo instabilities. All this causes a non-monotonic behavior of
the growth rate of magnetic field as the function of the Coriolis
number for different colatitudes.

A discussion on these aspects has been added to the paper; see the
last two paragraphs in Section 4.2.

> 4. Interpretation on the rotational dependency: Based on the linear theory
> of Losada+12, the increase of the rotation rate enhances the frequency of
> the inertia wave, resulting in greater suppression of the NEMPI. The result
> of the DNS is however not quite so simple. It shows non-monotonic
> dependence of the NEMPI on the Coriolis number. It may be one of
> interesting findings in the authors’ results that the relatively-moderate
> rotation rate, Co ~ 0.002-0.003, rather strengthens the NEMPI-induced BRs.
> Is this a consequence of the coronal layer or of the rotational shrinking
> of the turbulent eddies ? Is the rotational dependence a natural result of
> the competition between the enhancement of the NEMPI due to the rotational
> shrinking of eddies and the rotational suppression ? The authors should
> give some physical interpretation on the non-monotonic rotational
> dependency at least qualitatively.

%IR:
 As was pointed out in answer to previous comment,
 with increasing of rotation rate, a kinetic helicity is
 produced by a combined effect of uniform rotation and density
 stratified turbulence. It results in the excitation of an $\alpha
 \Omega$ or $\alpha^2\Omega$ dynamo instabilities.
 This implies that two different instabilities are excited
 in the system, i.e., NEMPI at low Coriolis numbers
 and mean-field dynamo instability at larger values of the
 Coriolis numbers. This causes a non-monotonic behavior of the
 growth rate of magnetic field as the function of the Coriolis
 number for different colatitudes.

%JW:
 The scale of the turbulence is given by
 the forcing wave number and it is not influenced by rotation unlike
 in convection, where the scale of convection decreases for larger
 rotation.

 Again, a discussion on these aspects has can be found in Section 4.2,
 see the last two paragraphs.