Dear Author,
Please revise your manuscript as described in your reply to thereferees's comments. The revised manuscript will be sent to referee #2.
Regards,
JeanPierre Gattuso
BG editor 
Suggestions for revision or reasons for rejection  Spilling and collaborators have overall made a pretty good job revising this manuscript. I would recommend its publication following minor revisions.
My main “minor” concern is related to the way fluxes and associated errors have been estimated.
Here is the answer provided by the authors to my initial comments and the added paragraph on data treatment:
. Answer on the way pools and rates were calculated: This is a good point. It was calculated based on the difference between the start of each period, and using the average of the first two sampling days as the initial value for each period. So they are not slopes per se. There is no statistical testing of the differences in this paper, but we have explained that this was done in the paper where the original data is presented (and here linear regressions were used e.g. Paul et al 2015).
. Answer on the way errors were calculated: We agree that the error estimates needs to be better explained. In the case of NPPe the SE was calculated from the square root of the sum of variance of the three parameters used to calculate the NPPe: DOC TPC and Exported TPC. We will include the sample size as suggested.
. Added paragraph in the revised version: Based on the primary variables the experiment where divided into three distinct phases: Phase I: t0t16; Phase II: t17t30 and Phase III: t31t43, where e.g. Chlorophyll a (Chl a) concentration was relatively high during Phase I, decreased during Phase II and remained low
during Phase III (Paul et al. 2015). Measurements of pools and rates were average for the two first sampling points of each experimental phase (n = 2) and where normalized to m2 knowing the total depth and volume of the mesocosms. The three different phases of the experiments were of different length (16, 14 and 13 days respectively). For fluxes and
biological rates we used the average for the whole periods normalized to days (day1) All error estimates were calculated as standard error (SE), with n = 16, n = 14 and n = 13 for Phases I – III respectively. SE for estimated rates were calculated from the square root of the sum of variance for all the variables (Eq 611 below) The primary papers present detailed statistical analyses and we only refer to those here.
I am afraid I still don’t understand correctly as it seems that answers provided and the new paragraph that was added do not match. If I understand correctly, rates of change (i.e. Delta DOC) were calculated as the difference between the start (2 first days) of each period. My first question would be: how did you do for the last period? In the added paragraph, it does not seem to be explained correctly. The same is true for the legends of Table 13, as it is said that: “… net community production estimated based on organic carbon pools (NCPo) are all average for Phase I in mmol C m2 d1 ± SE (n = 16).” How can you have n = 16 if based on a difference between 2 time points? Moreover, regarding measured rates (i.e. BP), I doubt that they were measured on a daily basis (n is not 16), but I might be wrong.
If I take the example of Delta DOC in M1 for Phase 1, that would be 16.4 mmol C/m2/d (considering a 16 day period, 7435 for the start of phase 1  7172 for the start of phase 2), a value of 15.5 is reported. I end up with a NCPo estimate of 18.2 mmol C/m2/d. More importantly, I don’t understand that, if indeed you propagated errors the way you explained (i.e. square root of the sum of variance), how you end up with a propagated error that is lower than one of the terms. For instance, still for M1 in Phase 1, delta DOC should have an error of SQRT(87^2+38^2) = ca. 95. Doing the same for delta TPC (SE ~40), you would end up with a propagated error on NCPo of SQRT(95^2+40^2+0.1^2) = 103, far from the value of 33 that is reported.
I might have misunderstood, however if this is the case, first I apologize, but also I would suggest the authors to better clarify their methodology.
Very minor corrections:
L1821: Affiliations have been swapped
L50: Please remove “fixed”
L180: A parenthesis is missing in the equation
L650: since you calculate GPP from your budgets, you should change to: (i.e. NCP + TR)
Regards
FG  

Dear Author,
Thank you for submitting a revised version of your manuscript which is acceptable pending minor revision. Referee #2 still has a hard time to understand how fluxes and their error were calculated. I concur with this concern; please address them in the next, hopefully the last, revision. Note that there are 4 additional comments at the end of this message.
Regards,
JeanPierre Gattuso
BG editor

Spilling and collaborators have overall made a pretty good job revising this manuscript. I would recommend its publication following minor revisions.
My main “minor” concern is related to the way fluxes and associated errors have been estimated.
Here is the answer provided by the authors to my initial comments and the added paragraph on data treatment:
 Answer on the way pools and rates were calculated: This is a good point. It was calculated based on the difference between the start of each period, and using the average of the first two sampling days as the initial value for each period. So they are not slopes per se. There is no statistical testing of the differences in this paper, but we have explained that this was done in the paper where the original data is presented (and here linear regressions were used e.g. Paul et al 2015).
 Answer on the way errors were calculated: We agree that the error estimates needs to be better explained. In the case of NPPe the SE was calculated from the square root of the sum of variance of the three parameters used to calculate the NPPe: DOC TPC and Exported TPC. We will include the sample size as suggested.
 Added paragraph in the revised version: Based on the primary variables the experiment where divided into three distinct phases: Phase I: t0t16; Phase II: t17t30 and Phase III: t31t43, where e.g. Chlorophyll a (Chl a) concentration was relatively high during Phase I, decreased during Phase II and remained low
during Phase III (Paul et al. 2015). Measurements of pools and rates were average for the two first sampling points of each experimental phase (n = 2) and where normalized to m2 knowing the total depth and volume of the mesocosms. The three different phases of the experiments were of different length (16, 14 and 13 days respectively). For fluxes and biological rates we used the average for the whole periods normalized to days (day1) All error estimates were calculated as standard error (SE), with n = 16, n = 14 and n = 13 for Phases I – III respectively. SE for estimated rates were calculated from the square root of the sum of variance for all the variables (Eq 611 below) The primary papers present detailed statistical analyses and we only refer to those here.
I am afraid I still don’t understand correctly as it seems that answers provided and the new paragraph that was added do not match. If I understand correctly, rates of change (i.e. Delta DOC) were calculated as the difference between the start (2 first days) of each period. My first question would be: how did you do for the last period? In the added paragraph, it does not seem to be explained correctly. The same is true for the legends of Table 13, as it is said that: “… net community production estimated based on organic carbon pools (NCPo) are all average for Phase I in mmol C m2 d1 ± SE (n = 16).” How can you have n = 16 if based on a difference between 2 time points? Moreover, regarding measured rates (i.e. BP), I doubt that they were measured on a daily basis (n is not 16), but I might be wrong.
If I take the example of Delta DOC in M1 for Phase 1, that would be 16.4 mmol C/m2/d (considering a 16 day period, 7435 for the start of phase 1  7172 for the start of phase 2), a value of 15.5 is reported. I end up with a NCPo estimate of 18.2 mmol C/m2/d. More importantly, I don’t understand that, if indeed you propagated errors the way you explained (i.e. square root of the sum of variance), how you end up with a propagated error that is lower than one of the terms. For instance, still for M1 in Phase 1, delta DOC should have an error of SQRT(87^2+38^2) = ca. 95. Doing the same for delta TPC (SE ~40), you would end up with a propagated error on NCPo of SQRT(95^2+40^2+0.1^2) = 103, far from the value of 33 that is reported.
I might have misunderstood, however if this is the case, first I apologize, but also I would suggest the authors to better clarify their methodology.
Very minor corrections:
L1821: Affiliations have been swapped
L50: Please remove “fixed”
L180: A parenthesis is missing in the equation
L650: since you calculate GPP from your budgets, you should change to: (i.e. NCP + TR) 
Dear Author,
Thank you for submitting a revised version of your manuscript submitted to Biogeosciences, which I am glad to accept for publication.
I look forward to seeing this paper published and thank you for considering Biogeosciences to publish these interesting results.
Best regards,
JeanPierre Gattuso
BG Editor 