Resolving herbivore influences under climate variability

In this impressive and well-designed work, the authors explored the role of grazing intensity (cattle) and drought on above and below biomass production in a Lolium + Trifolium dominated grassland over ten years. Root production was assessed monthly by cores, and leaf biomass by cutting and weighing. Other roots and leaf traits such as root length, density, and area, as well as community-weighted mean, dry matter content, and specific lead area and reproductive plant height, were also recorded and considered. The authors tested the following hypothesis: Whether grazing intensity increases above-ground biomass and other related traits (e.g., length, density, area) at expenses of root production (I), whether inter-annual weather conditions influences the effects of grazing (II), and finally whether root traits and thus below net primary production respond to treatments. On the other hand, the authors also describe weather conditions, soil temperature, and water content dynamics throughout the experiment. These additional objectives, however, have not properly been introduced in the aims of the study. The authors conclude that root production did not differ between grazing intensities. Below net primary production, however, decreased in ungrazed plots. Grass species composition varied across grazing treatments (e.g., Lolium grasses are more abundant in highly grazed areas), and specific root length and area were slightly lower in cattle-free areas. I think this work provides an exhaustive and detailed description of the temporal dynamics of the above and below biomass, as well as other related rood and leaf traits, in a very well managed grassland. Cattle pressure is low and rotational and thus is expected to have little impact on the traits considered. On the other hand, grasslands have been fertilized increasing the resilience of grasslands on the inter-annual variations in weather conditions. In summary, the authors have shown the outcomes of an excellent grassland management practice. To my understanding, grazing intensity plays a secondary role in this work. I have included an attached file with other minor comments and suggestions in the main body of the manuscript as popup notes. Check the use of acronyms through the manuscript, they are sometimes confusing and repetitive. I strongly recommend deep proofreading and editing by a native speaker. I hope my comments will be useful for your investigation.

This is an interesting study, which describes the productive response of grass at different levels of grazing intensity comparing a wet year with a dry year. The subject of study is interesting, despite the fact that already there are some related works, more information is still needed to understand the effect of the interaction between grazing and environmental conditions on pasture functioning.
The introduction is very extensive, greater concretion would be desirable. Some expressions are difficult to understand, changing long sentences for several short sentences will make understanding easier, for example, L53-56 or L62-66. Furthermore, in general, the use of language sometimes makes it difficult to understand the content. Perhaps the review of the work by someone specialized can help make the speech more fluid.
Regarding the statistical analysis, I have a doubt. The use of the plot nested in the block as a radon factor in the linear mixed-effects model is considered for accounting for temporal pseudo-replication, however, in my opinion at the same time it can control the spatial pseudo-replication and therefore it would not be necessary to calculate the mean values for each plot, being able to use the data of each ingrowth-cores which would increase the degrees of freedom and with it the statistical power of the test.
I am not sure I understand what is the intention of carrying out a PCA, since, in the results, I miss that each of the plots is represented in figure 3. Its representation would allow us to observe if the set of measures of the variables separate the different plots depending on the treatment to which they belong. Otherwise, if the objective is only to observe the correlation between variables, a matrix of correlations between parameters would be more suitable than a PCA, since it would allow us to observe the correlation values and their significance for each pair of variables. In the discussion, the authors state "in our field conditions and after 10 years of treatment application, soil moisture was not affected by the rotational grazing, probably because the temporal scale used (monthly-based) buffer shorter-term response", however, there are other possible explanations for it. Is it possible that the livestock loads used do not increase the apparent density of the soil? Or does the change in plant species between treatments compensate for the lower LAI in the grazed areas? Not all plant species have the same efficiency in the use of water. Authors affirmed “these treatments seem to be better adapted to buffering the negative effect of drought on grassland production than for abandoned grasslands. This is consistent with previous work showing that moderate grazing could be more beneficial than no grazing for drought resistance and recovery of ANPP and BNPP” I'm not sure if the data showed reflects this. Figure 2 shows a marked decrease in the dry year with respect to the wet year in the ANPP for grazing treatments, while the abandoned treatment maintains a similar value in both years. Therefore, the interpretation that I give to these results is that in wet years grazing increases productivity with respect to abandonment, but this increase in productivity disappears in the event of a drought, therefore, the grazing treatment is more sensitive to drought than the abandoned ones. In this context, it would be interesting to know how the years prior to the study have been, at least 2012 and 2013, to know if there is any type of accumulated effect from one year to the next.

On some occasions the authors do not find the expected effect regarding the grazing treatment, giving different explanations for it. However, a possible explanation not considered is the low number of replicates used in this study. When the expected differences between treatments are small, high statistical power is necessary to detect them, which means a sufficient number of replications. Even more so when the effects are heterogeneous in space, as the authors well indicate since the impact of livestock is highly variable in space. Perhaps reanalyze the data without calculating the means per plot can give new results. But regardless of this, I would not disregard the need for a broader sampling to be able to detect the impacts of grazing on the functional characteristics of the roots.
It seems that there is some confusion among tables S1; S2 and S3 since in the text name S2 to refer to the information shown in S3 and names S1 for the information shown in S2.
References:
There is a lack of some important studies as:
Li, W., Li, X., Zhao, Y., Zheng, S., & Bai, Y. (2018). Ecosystem structure, functioning, and stability under climate change and grazing in grasslands: current status and future prospects. Current Opinion in Environmental Sustainability, 33, 124-135.
Aldezabal, A., Odriozola, I., & García-Baquero, G. (2019). Grazing abandonment delays the effect of temperature on aboveground net primary production in Atlantic grasslands. Rangeland Ecology & Management, 72(5), 822-831.
In the manuscript authors indicate that “Data are available online: https://zenodo.org/deposit/4034903#” however, the web page indicates Permission required, so in reality, data are not available.
In general, I consider it to be a very interesting article that provides interesting data on certain aspects of pasture operations that are not yet clear, and that will certainly be cited after its publication.

General The paper "Intra and inter-annual climatic conditions have stronger effect than grazing intensity on root growth of permanent grasslands" addresses the important question of how climate and management (herbivory) influences the main drivers of soil C sequestration or soil organic carbon (SOC), i.e. above and belowground plant growth. The manuscriopt is well written and clear in its objectives to address the importance of root growth, soil fertility, and species composition as drivers of SOC.

The approach using an existing long-term trial is appropriate to capture short-term changes. While we know that SOC changes occur over decadal and longer time periods, and thus a longer trial would have been ideal, the scarcity of such trials must be acknowledged. Neverthless, their findings that climate has a larger influence than management (in this case a gradient of grazing regimes) is in agreement with other studies, and thus supports the overall idea that SOC is driven primarily by climate, and how this affects ANPP (and BNPP). The observations regarding grazing, relatively increased root growth and drought tolerance are not novel (see Klumpp et al 2011 in manuscript) but an important confirmation. An important contribution of the work is that they had access to different grazing intensities, as it is the variety of grazing intensities in various studies that confounds generalizations on its effect on SOC or other components of the biosphere.

As a suggestion, I would have liked to see more than 2 years of data, at least for the ANPP and possibly this is available for an existing 10-year trial? If they have 2 years of BNPP and ANPP as well as longer-term ANPP data available, this would allow some estimation of BNPP over the 10 years. Also once they have data over the longer time period it becomes viable to model time periods applicable to climate models. Without this longer term data I would say the paper still provides valuable data, especially considering the paucity of belowground studies of roots and SOC.

Specific comments:
The abstract could be improved with some result details