Introduction

replying to Y. Zhao et al. Nature Communications https://doi.org/10.1038/s41467-025-65470-z (2025)

We read with interest the commentary by Zhao et al.1 on our study on the acclimation of ecosystem photosynthesis as measured by gross primary productivity (GPP) to soil moisture2. The additional analysis provided by Zhao et al. contributes to the understanding of the soil moisture effect on GPP by considering field capacity (θFC). However, three key issues in the commentary by Zhao et al. warrant attention: (i) confusion of the concepts of θFC and apparent optimum soil moisture (\({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), beyond which the effect of soil moisture on GPP shifts from positive to negative); (ii) lack of direct evidence that \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and θFC are equivalent; (iii) unrobust analysis with significant uncertainties. In the following sections, we provide details on each of these issues.

Conceptual confusion between \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and θ FC

Zhao et al.1 proposed a conceptual framework for GPP as a function of soil moisture. In the framework, GPP is expected to increase, plateau at a broad optimum, and then decline in response to increasing soil moisture. This generally aligns with our study, which shows that both excessively low and high soil moisture can inhibit GPP. Unfortunately, the authors directly defined \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) as θFC, conflating these two different concepts. Specifically, \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) is derived from the GPP-soil moisture response curve and is considered to be an ecosystem property, indicating the soil water requirement to maximize GPP. In contrast, θFC reflects soil hydraulic property that indicates the soil’s capacity to retain water3. It is measured as the soil water content remaining after excess water has been drained away and the rate of downward movement has been substantially decreased3. Given this conceptual distinction, it is not appropriate to treat \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) as θFC.

No direct evidence that \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and θ FC are equivalent

Zhao et al.1 analyzed the relationship between \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and water content at different soil water potentials to justify that \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and θFC are equivalent, but the authors relied on simulated rather than field measured data. This may introduce biases in the subsequent analysis, which were not discussed by Zhao et al. In particular, we note that shifting the soil water potential from −60 to −330 hPa only altered water content, but had no effect on \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). This would be theoretically unexpected, as a shift in soil water potential typically changes the ability of plants to absorb and utilize water4,5, which in turn would potentially alter the response of GPP to soil moisture.

Furthermore, it is important to note that the linear regression reflects the change in the dependent variable with respect to the change in the independent variable6, and is therefore not applicable to testing whether \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) is equivalent to θFC, as Zhao et al.1 did. In fact, when we extracted the data from Zhao et al. for reanalysis, we found that the ratio of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) to θFC at soil water potential of −330 hPa varied largely, ranging from 0.81 to 1.41 (Fig. 1). This contradicts Zhao et al.’s claim that \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) equals θFC at soil water potential of −330 hPa. Overall, the evidence provided by the authors to support the equivalence of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and θFC is not compelling.

Fig. 1: Distribution of ratio of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) to θFC at soil water potential of −330 hPa (\({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\):θFC).
figure 1

The data is from Zhao et al.1. The red vertical dotted line indicates that \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) equals θFC (i.e., \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\):θFC = 1).

Full size image

Uncertainties in the analysis by Zhao et al.

Zhao et al.1 analyzed the \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) distribution across different soil texture classifications at different soil water potentials. Based on this analysis, they conclude that the variation in \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) is driven by inherent soil water retention properties rather than the acclimation process. We acknowledge the additional efforts made by Zhao et al. to stress the importance of soil texture in influencing \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). Nevertheless, there are significant uncertainties in their analysis that potentially weaken their conclusion. In particular, although Zhao et al. found an association between soil texture and \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), this association was based on correlation and did not indicate a causal effect of soil texture on \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). In addition, when estimating the relationship between soil texture and \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), the authors failed to account for potential co-varying factors, especially local soil water availability (SMgrowth), which is a key factor influencing \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) identified in our study.

In fact, our study also attempted to test the effect of soil texture (i.e., soil sand fraction) on \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), which was likely overlooked by Zhao et al.1. Specifically, we found a negative correlation between soil sand fraction and \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) as shown in the bivariate plot (see Supplementary Fig. 7 in Peng et al.2), but this correlation became negligible when controlling for other climatic, soil, and vegetation factors (see Supplementary Fig. 6 in Peng et al.2). In contrast, even after controlling for other climatic, soil, and vegetation factors, SMgrowth was always closely correlated with \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), possibly reflecting water acclimation of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). Notably, we further used a field experiment in which we manipulated only the amount of water and kept all other factors constant to verify the causal effect of SMgrowth on \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), providing direct experimental evidence for the water acclimation of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). This field experiment also shed light on the mechanisms involved by revealing the potential role of plant traits in driving the water acclimation of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). Additionally, we would like to emphasize the ecological significance of quantifying the water acclimation of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\). Whether \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) shows a proportional adjustment to SMgrowth (i.e., SMgrowth-\({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) slope) has important implications for the way in which ecosystem responds to soil moisture (Fig. 2)2. For example, if an increase in SMgrowth does not result in a proportional increase in \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) (i.e., SMgrowth-\({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) slope < 1), the increased SMgrowth would be more likely to suppress GPP once \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) falls below SMgrowth. Conversely, if an increase in SMgrowth results in a proportional or greater increase in \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) (i.e., SMgrowth-\({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) slope ≥ 1), the increased SMgrowth would be more likely to stimulate GPP.

Fig. 2: Conceptual framework for the change in gross primary productivity (GPP)-soil moisture response curve due to water acclimation of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\).
figure 2

The black solid line represents the response curve under low SMgrowth, whereas the green solid line represents the response curve under high SMgrowth.

Full size image

In conclusion, our study provides compelling evidence for the water acclimation of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\), and Zhao et al.’s analysis is not sufficiently robust to support their point that \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) is equivalent to θFC and that the variation in \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) is solely due to inherent soil water retention properties. We appreciate Zhao et al.’s interest in our study and their efforts in developing an empirical function that accounts for the inhibitory effects of both excessively low and high soil moisture on GPP, while also including a potential plateau of optimal soil moisture levels. The performance of this empirical function deserves to be tested in future studies. Further research into the dynamics of \({{{{\rm{SM}}}}}_{{{{\rm{opt}}}}}^{{{{\rm{GPP}}}}}\) and its control mechanisms globally would also be valuable, as it is expected to improve our understanding of carbon-climate feedbacks. A global network of control experimental studies, such as International Drought Experiment7, could provide insights into this issue.