(1) It seems the authors are using CIC and carbonate-intersecting subduction zones (arcs) interchangeably in the text, which can be two different things. What is the exact definition of continental subduction zones and CIC in this paper? I am confused to see the subduction zones (purple curves) in European Alps during 50-0 Ma in Figure 5. Is this classified as “continental subduction zone” or “continental arc” in this study? Did Andean-type continental arcs exist in European Alps since 50 Ma? The magenta curves in Figure 5 are subduction zones, what are relationships between subduction zones, CIC, non-CIC, oceanic subduction zone (oceanic arc), continental subduction zone (collision belt)? Does non-CIC also include oceanic subduction zone?
In the caption of Figure 4, the solid magenta curve is the “all subduction zones that intersect with carbonate platforms.” So are you actually dealing with carbonate-intersecting subduction zones instead of carbonate-intersecting continental subduction zones? I am very confused here. Also, can you use another color to highlight the CIC in Figure 5?
Page 18, Line 23: “Our investigation encompasses the total segment length of all subduction zones.” So CIC is “all subduction zones that intersect with carbonate platforms” not limiting to continental subduction zones? If this is the case, you may want to clarify/change your terminology and definitions. And, the conclusions about the correlation between CIC and CO2 have to be changed since it is the “all subduction zones that intersect with carbonate platforms” being studied.
(2) Only Phanerozoic Carbonate platforms are used. So how about pre-Cambrian ones? The authors mentioned this issue in Section 2.2, but they did not clearly state how such implementation affect their conclusions: would this overestimate or underestimate the length of CIC? Page 8, Line 8, “it is expected that accounting for Precambrian platforms would not drastically change the analysis” Why? For example, during the Proterozoic, extensive carbon platform exists in northern margins of North China and Tarim plates. In particular, ~ 1.4 Ga Jixian Formation contains about several-km-thick carbonate (e.g., Meng et al., 2011, Stratigraphic and sedimentary records of the rift to drift evolution of the northern North China craton at the Paleo-to Mesoproterozoic transition. Gondwana Research). These carbonates are within 450 km (if use the scheme proposed in this paper) from the northern margin of North China and likely to be intruded by arc magma during Permian-Triassic when Solonker ocean between Mongolia and North China closed.
(3) Please provide more information on how to read XWT, CWT, WTC figures. For example, in Figure 2, the area in the cone of influence is less reliable? What is a “red noise background spectrum”? What does “wavelet power” mean? The Results section 4.1-4.2 used many technical terms. For example, “signals of mid- to long-wavelength component remain strong”, “waveband”, “wavelength”, “wavelength band”, “periodic component”, “peak”, what do they mean? Can you translate some of them into the geological language?
Some minor issues:
Page 3, Line 18: Each arc may display 50 Myr flare-ups, but not all of them had flare-ups at the same time (Krisch et al., 2016, American Mineralogist). So the overall global carbon release pattern from these arcs could have the longer temporal wavelength.
Page 4, Line 8: Cao et al. (2017)’s compilation of lengths and spatial distributions of continental arcs is based on regional geology and not directly depends on the paleomaps. The history of continental arcs is continuous (see Cao et al.’s Figure 2) and is not based on discrete plate reconstruction. The paleomaps in the Cao et al. (2017) only show snapshots of some specific geological periods.
Page 4, Line 9: Cao et al. (2017) used surface exposures of granitoids as one of the proxies of continental arcs. They are not arc “volcanic” products. They also (1) compensated older arcs with more surface areas of granitoids; (2) used geology to restore the lengths of arcs, which is independent of surface exposures of granitoids; (3) the results from two approaches correlated well in the first order. Given the above points, why does the Cao et al., (2017) represent the lower-limit? It is fine to argue their compilation may miss some continental arcs or underestimate lengths or durations of some arcs. If your length of CIC is carbonate-interesting subduction zones not limiting to continental arcs, you could argue Cao et al., (2017) may underestimate the length of carbonate-interesting arcs.
Page 9, Line 15: why not just using the average trench-arc distance of 287 km? Wondering if the longer distance tends to smooth the temporal pattern. Will the results be affected if 287 km is used instead of 448 km?
Page 11, Line 1: Why using 7 Myr (not 6 or 8) window to remove high-frequency signals?
Page 11, Line 4: Why do you use 5 Myr here instead of keeping the window size (7 Myr) consistent?
Page 18: Line 23: you mean “ less than half”?