The authors present an interesting multi-year study of the sensitivity of primary production of different functional groups to experimental water manipulation along an elevation gradient, in the context of a rare multi-year drought. The authors’ experimental water manipulation maintained over a steep elevation gradient for 5 years makes this a particularly impressive study. This study is valuable for its ability to reveal the context-dependency of responses and to highlight cumulative effects through time, which are poorly understood. The authors’ monitoring of soil moisture and incoming vs redistributed precipitation is thorough and impressive, and the paper is well-written and likely of very broad interest.
The presentation of figures and results is generally clear, however I have made some recommendations to improve clarity of how the statistical methods directly address the objectives and pertain to which results, and to clarify when the experimental watering treatment is being referred to versus ambient precipitation.
Major comments:
• At times it is not clear whether the precipitation manipulation treatment is being referred to, or the ambient precipitation amount. To improve clarity, I recommend keeping terminology consistent when referring to the precipitation manipulation treatments (drought and water addition) vs ambient precipitation (whether that be drought or wet, for example). For example, in L291, I would replace ‘to extreme drought or heavy rainfall’ with ‘to extreme drought or water addition’. Similarly, L30 states ‘We experimentally imposed extreme drought and heavy rainfall from 2016 – 2020, during which ambient warm-season precipitation declined to reach history lows…’ Since precipitation fell to such low extremes in later years (L255-260), it does not seem accurate to say that heavy rainfall was imposed up until 2020. I would rephrase this to clarify that water was experimentally reduced and added relative to ambient amounts.
• It would be beneficial to clarify how the relationship between cover and biomass was determined. Please provide more detail of how biomass was sampled outside of the plots (e.g. over what area, two 1m2 quadrats? L197), and can the authors reassure the reader of the quality of the regression and amount of uncertainty associated with biomass?
• When reading the methods for the statistical analyses, it is not clear which method is being used to address which objective. One solution to clarify this for the reader would be to structure the analytical methods around the objectives they are each addressing.
• In Figure 5, the readers are shown results for the sensitivity of C3 and C4 grasses to warm-season precipitation. In the results text, responses for other life forms (annuals L313 and perennial forbs and shrubs L319-323) to warm-season precipitation, climate and soil moisture are stated but there is nowhere for the reader to look at these results. I recommend including a supplementary table showing these results, which would also help the reader to understand what was modelled. When referring to these analyses, it would be helpful to refer the reader to this table.
Minor comments:
• L85 spell out GCM acronym
• L227 typo, replace ‘non-liner’ with ‘non-linear’
• L270 typo, space needed in Fig. 3d
• L273 typo, replace ‘Figs. 3j’ with ‘Fig. 3j’
• L275 add ‘through time’ at the end of the sentence. i.e. ‘increased at the desert scrub (Fig. 3f) and ponderosa pine meadow (Fig. 3i) through time.’
• L287 clarify sensitivity to warm-season precipitation, i.e. ‘C4 perennial grass ANPP had higher sensitivity to warm-season precipitation’
• L302-308 remove ‘(Fig. 5b)’ from second sentence
• L312 add a sentence on how water addition influenced ANPP of annuals (no response)
• L345 why would the plants already be water-limited, given that the control plots at those two sites were close to the 95th percentile of highest precipitation in 2016? I disagree with this assumption
• L360-363 split this sentence into two – 1) results of Hoover et al. (2015), 2) suggestion of acceleration mortality etc. Please include a reference to support the forecasted aridification in southwestern US
• L393-396 surprising finding that water addition led to a decrease in production. The suggestion that competition between functional types due to increased C3 ANPP is reasonable, but please insert references to support this
• L416 typo, replace ‘on our study’ with ‘in our study’
• Figure 1 legend typo – replace ‘d) ponderosa pine meadow, and d) mixed conifer meadow’ with ‘d) ponderosa pine meadow, and e) mixed conifer meadow’
• L463-464 this is an interesting explanation for this result. However, if annuals increased opportunistically in response to pulses of moisture that were available because there was less perennial grass biomass, why wouldn’t the annuals have positive responses to water addition as well?
• L489 add hyphen, warm-season
• Figure 1 and Figure 5 legends – replace ‘dash’ with ‘dashed’
• Table 1 adjust spacing so headings aren’t split between two lines
• Table 2 add hyphen for ‘three way’
• Figure 3 legend – replace ‘and the ratio of C4 perennial grass ANPP in treatments’ with ‘and the ratio of all plant functional types (total) ANPP in treatments’
• Figures 3, 4 and 6 legends – clarify where solid lines indicate significant relationships, grey lines are for drought and black lines are for water addition
• Figure 5 legend – suggestion to clarify warm-season precipitation, e.g. ‘in relation to May – September (warm-season) precipitation at each site’. Also, add slope, r squared and p-value info for ponderosa pine meadow (water)
• Make sure you upload high resolution figures
• Supplemental Figure Legends – ‘Figure S1’ is duplicated