Response comments and proposed edits are satisfactory overall, and if implemented fully I expect the manuscript will be accepted for publication. There are, however, some remaining issues I would ask the authors to address.

The abstract ends with the claim that constraint from biomass observations reduced the uncertainty of the historical carbon budget, however this does not appear to be consistent with the results of the manuscript. The study reports that the mean and standard deviation on the constrained LUC emissions is almost identical to that from the TRENDY DGVMs if one assumes that they have a normal distribution. It would thus be difficult to argue that uncertainty has been reduced and such statements need to be removed from the paper. You could potentially claim to have identified one or more of the TRENDY members that are inconsistent with the observations. You can also comment, as in the Conclusions, that the wide range in the constrained estimates derives from wide uncertainty in biomass observations which acts as the constraint, and that reductions in uncertainties on observed biomass would improve the constraint that these observations could provide.

R1 asks a good question about what this study offers given the above result. The so-called "lessons learned" are well communicated in the response comment and should be reflected in the paper's discussion and/or conclusions.

Given concerns about non-normality of the distribution of ELUC estimates across the TRENDY DGVM ensemble, you might consider using quantiles for both the original _and_ the constrained estimates when you compare them.