** We thank the referee for a truly careful reading.  We have revised the manuscript in response to the many thoughtful suggestions.

A major issue that threaded through the paper that continually tripped me up while reading it surrounds the concept of radial gradients in stellar populations. One could be talking about two things here:

1) gradients within a particular structural component (for example, changes in the stellar populations of the disk as a function of radius),

or

2) changes in the relative strength of two differing structural components (for example, the starlight being dominated by the disk in the inner regions and the halo in the outer regions).

In different spots, the paper seemed to talk about processes that shape gradients within one component, such a radial migration of disk stars, and in other spots the emphasis focused more on disk vs halo populations.

All of that is fine, and certainly from an observational perspective both effects can give rise to observed gradients, but the discussion seemed to switch back and forth between these different perspectives, leading to a real lack of clarity.

As examples, the abstract talks about radial migration affecting bulge stars, when migration is largely a disk process. Similarly, the results section suggests that migration can lead to red halos, but migrating disk stars will remain largely within the original disk plane, without enough vertical heating to form a true halo.

Similarly even talking about gradients in the star formation history of a galaxy gives the implication that we're talking about star formation *within* the galaxy, but in the case of external accretion, the star formation being probed would have taken place well before the stars were incorporated into the galaxy. Again, this doesn't mean the SED fitting approach is incorrect in what it shows, but this ambiguity in meaning makes much of the results and discussion difficult to interpret.

** We have carefully modified the text to more accurately reflect the implications of the results and how they potentially relate to radial migration and external accretion.

==================
Observational data
==================

- S3.3, the authors give the limiting point source sensitivities, but for surface photometry this is not the important number-- the important number is the limiting surface brightness, which depends largely on things like sky uncertainty, flat fielding error, and scattered light.

** We have modified the text to report the limiting optical surface brightness of ~28 mag/arcsec^2.

- It would be extremely helpful to give the optical surface brightnesses in mag/arcsec^2, to allow comparison to other optical studies, all of which use those units. (I realize that the convention for Spitzer IR data is MJy/sr, but that's not the convention for optical data...) In looking at Fig 3, it looks like the optical data is plotted down to 10^-3 MJy/sr (more or less) which is roughly equivalent to mu_g ~ 28 mags/arcsec^2, I think? So is that the limiting surface brightness? This needs to be addressed somewhere in the paper, giving limiting surface brightness in mag/arcsec^2 and perhaps giving Fig 3 an alternate y-axis that shows a mag/arcsec^2 scale. This is particularly important since the authors claim to go significantly deeper than other optical studies, all of which report in mags/arcsec^2.

** We have modified Figure 3 to include a righthand axis scaled to mags/arcsec^2.  We have modified the text to report the limiting optical surface brightness of ~28 mag/arcsec^2.

- Also, if these profiles are going down to 28 mag/arcsec^2 in the optical, at those levels, a dominant source of contamination in the optical and UV is going to be scattered optical light from galactic dust, particularly for galaxies at low galactic latitude. Have the authors checked for this?

** Fortunately all our targets are at Galactic latitudes north of +66 degrees.  There is no correlation with Galactic latitude and surface brightness sensitivity, and inspection of available far-infrared maps shows a lack of conspicous Galactic cirrus along these lines-of-sight.  This has been noted in the text.

- At those low surface brightnesses, the signal from the galaxy is going to be swamped by the myriad of background sources and faint foreground stars. The authors say they use imedit to hand-mask "conspicuous" stars and galaxies, but down to what magnitude? For example, when I look at Figure 1, the sky regions (and outer disk regions) all have lots of stars in them (and many more fainter ones that can't be seen in this postage stamp image) -- were they all masked? Or just the brighter ones?

Despite (or perhaps because of) Eq 2, I'm having trouble working out how the sky error is being calculated. Part of this is because I'm not sure if faint discrete sources (stars, background galaxies) are being masked or not. If sigma_sky is the per pixel variation in the sky regions, that would be a good estimator *if* discrete sources are being masked -- otherwise it will have a very non-gaussian skew with a bright tail that makes std-dev a bad estimator. If discrete sources *aren't* being masked, the total sky error is probably better calculated as the variation in the total flux measured in different sky regions (scaled by the relative size of the annuli and sky regions). I can't see how equation 2 captures either one of these regimes.

A few things might help sort this out. First, showing an example of a photometric mask, so the reader (or at least this referee) can see what level you are masking to. The text (where it says that sky regions must have representative numbers of stars) and Figure 1 (where many stars can be seen in the sky apertures) certainly give the impression that a lot of stars are being left unmasked, but perhaps that's not true. Second, give a more thorough/descriptive explanation of how the sky regions are being used to determine and propagate the sky noise in the target annuli.

** We agree with the referee and have revisited the masking/editing of foreground stars and background sources as well as the definition of sky regions.  Compared to what occurs at the other wavelengths studied here, the Spitzer 3.6um imaging easily contains the most contamination from foreground stars and thus the revised data processing was comparatively minimal outside the 3.6um data.  We have added to the text that our editing extends down to sources of several microJy (21-22 mag AB).  We have modified Figure 1 to include an example of a photometric mask as well as to show our new approach to sky aperture selection.  As can be seen from the new version of Figure 1, we now opt for sky regions that lack anything but the faintest sources.

- Measuring radial distances in units of a25 isn't particularly physical -- for two galaxies of fixed physical size, a high surface brightness galaxy and a low surface brightness galaxy will have very different a25's. If the point is to look at structural properties of galaxies, it's better to measure in terms of the half-light radius or the scale length of the disk.

** Using R25 as a normalization allows for ease of comparison with other radial studies of nearby galaxies (e.g., Ho+15, Hunt+15, Moustakas+10, Munoz-Mateos+09, etc).  We agree that for studies involving a set of finely-spaced annular apertures, R_eff is a particularly good physical scaling for ellipticals, and that R_d is a good scaling for disk galaxies.  We  indicate approximations for R_eff(r') in Figure 3 as a reference.

- Fig 4: the "surface brightness profile slope" is essentially a crude measure of the profile shape, something that is typically characterized by an exponential scale length (for pure disks) or Sersic profile. In the luminous inner regions (where this is being measured) this is sensitive to the relative strength of the bulge and disk components. Given all this, and given the range of Hubble types in the sample, simply measuring a profile slope and saying that massive galaxies show steeper slopes doesn't seem all that meaningful. What leads to this trend? Are the more massive systems more bulge dominated, for example? There is a vast literature available discussing galaxy surface brightness profiles and their correlation with mass, color, B:D ratio, etc -- this part of the paper seems to lack much connection with this body of work.

** We agree that this subsection is tangential to our goals and so it has been removed.

- In extracting the colors (Fig 5), I'm confused by the errorbars. They seem roughly constant at ~ +/- 0.15 mags over a radial range where the galaxy surface brightness is dropping by a factor of 100. If all that's going into the errorbars are a 0.02 mag calibration error and a sky uncertainty, this means that sky uncertainty dominates all the way into the center, which doesn't make sense.

** Excellent eye!  There was an error in the computation of the error bars for that figure.  We now utilize error bars that reflect the maximum error resulting from over- or under-subtracting the sky by +/- 1 sigma (e.g., de Jong 1996; Pohlen & Trujillo 2006).

- I was very confused by the description of the issue of PSF wings in the last para of 5.2. First, the PSF reflections can happen at any wavelength, and depend not just on the CCD but also on the reflectivity of the filter and CCD dewar window, so simply avoiding i isn't enough to demonstrate they aren't an issue. And I don't see how the second issue is going to be important unless the PSF is huge compared to the galaxy size, which isn't going to be the case in the optical for nearby galaxies. Perhaps its more important at longer wavelengths, in the mid-IR, but using the optical data to demonstrate that it's not a problem doesn't seem convincing.

** We have slightly modified the text but feel the bulk of it should remain; the red halo effect is a legitimate concern for i' data taken on thinned detectors (see, e.g., the Appendix of Wu+2005).  Our goal with this Figure is to allay any concerns a reader may have with our interpretation of the observed g'-r' colors.  Note also that we utilize this Figure to directly address the red halo effect---we don't simply ignore the effect through avoidance of i' data.  In addition, the inclination effect is not more pronounced at longer wavelengths since this analysis has been carried out at the same 6" (smoothed) resolution for all wavelengths.


==============================
SED Fitting and Interpretation
==============================

- The motivation for using two different star formation models is unclear. Is the double exponential being used to test the possibility of a late burst of star formation? If so, wouldn't a better model to be comparing the delayed burst model with an additional exponential burst? On the other hand, if it's just to test two different common models, wouldnt it be better to compare the delayed burst model to a *single* exponential model (which describes the much more commonly used "tau model")?  

** We have modified our approach and utilize both a delayed and a single exponential model (the referee's suspicion was correct that we were simply aiming to test two different common models).  

- I'm not convinced that the trends shown by the averages in Fig 6 are statistically meaningful -- the scatter in the values is very large and very nongaussian. It would help to show quartile spreads or some other non-parametric estimate of the variation, to compare with the amplitude of the trend.

** We have added lines that indicate the 25th-75th quartile spreads.

- Fig 11: While I appreciate the test of how robust the derived parameters are, the real question is how do the parameter estimates couple to one another. For example, are there systematic tradeoffs between tau and Av (or any other parameter pairs)? The answer to this is well beyond the scope of this paper, and may well be answered in the more technical papers about the CIGALE modeling technique. I'd suggest discussing these kinds of systematic uncertainties as they impact the results. This discussion can be qualitative, with a reference to such papers (that hopefully exist) for more details. 

** We have added text that provides some more background on the parameter robustness, citing previous efforts that have explored this aspect in some detail.


==============================
Minor corrections/suggestions:
==============================

S4, para 1, last sentence: I think you mean overestimated, not underestimated.

** Thanks.  This paragraph has been restructured.

Fig 3: it's really difficult to see the vertical dotted lines. maybe make them solid and thicker instead?

** Done

S4.2, para 1: can the authors explain what "attentuation curve modifier" means?

** It reflects by how much the slope of the extinction law needs to be changed.  We have added further explanation to the text.