Dear authors

Thank you for the responses to the reviewers’ comments that indicate that major revisions are necessary in order to publish you MS in BG. I found appropriate the scientific and technical argumentation in your answers. However, on several occasions, your responses remained relatively vague concerning the way you will alter your MS in order to satisfy the comments. In the following step, please revise intensively your manuscript and provide a new response to the reviewers’ comments that details all the alterations made in the MS. Looking forward reading this soon.

With best regards

Gwenaël Abril

Dear authors

Your revised paper has now been reviewed a second time by Dr Ribas-Ribas and she thinks you have satisfactorily improved you MS to allow its publication. From my own reading, I also found that you paper can be published in BG. Please find below two minor marginal but significant comment concerning the description of the state of the art and the citation of literature in your introduction (particularly concerning inland waters). I think you could consider some small changes in your MS in order to make it consistent with this important literature. This would strengthen you paper.

thank you for submitting your interesting work to Biogeosciences

best regards
G Abril

L26 The sentence “Regions such as lakes, rivers and reservoirs, are only more recently becoming recognized as significant elements of the global carbon budget (e.g. Regnier et al. 2013; Borges et al., 2015)” is not fully exact. In fact, CO2 emissions from lakes and rivers have been first recognized as significant more than 20 years ago. In 1995 by Cole et al. 1995 in Science (lakes ) and Cole and Caraco 2001 Mar. Freshwater Res., 2001, 52, 101–10 (rivers) and then updated with new (pH/TA) data compilation by Tranvik et al. (2009 L&O) and Raymond et al 2013. Nothing really new on the quantification of lakes and rivers CO2 emissions in Regnier’s 2013 paper, which does not provide any new data, but rather uses the previously published estimates.

L32 &75. Sometimes high-resolution measurements as those you are proposing here allow some necessary changes in paradigms on aquatic C cycle, compared to those derived from discrete "blind" sampling (see eg Richey et al. 2002 and Abril et al. 2014 both in Nature). This might be another powerful argument to promote the development of in situ, high resolution and real-time instrumentations.