Effects of soil water content on carbon sink strength in an alpine swamp meadow of the northeastern Qinghai-Tibet Plateau

Abstract. Predicted intensified climate warming will likely alter the ecosystem net carbon (C) uptake of the Qinghai-Tibet Plateau (QTP). Variations in C sink/source responses to climate warming have been linked to water availability; however, the mechanisms by which net C uptake responds to soil water content in water-saturated swamp meadow ecosystems remain unclear. To explore how soil moisture and other environmental drivers modulate net C uptake in the QTP, field measurements were conducted using the eddy covariance technique in 2014, 2015, 2017, and 2018. The alpine swamp meadow presented in this study was a consistent and strong C sink of CO2 (−168.0 ± −62.5 gC m−2 y−1, average ± standard deviation) across the entire 4-year study period. A random forest machine-learning analysis suggests that the diurnal, seasonal, and annual variations of net ecosystem exchange (NEE) and gross primary productivity (GPP) were controlled by temperature and solar radiation. Ecosystem respiration (Re), however, was found mainly regulated by the variability of soil water content (SWC) at different temporal aggregations followed by temperature, the second contributing driver. We further explored how Re is controlled by nearly saturated soil moisture and temperature comparing two different periods featuring identical temperatures and significantly differences on SWC and vice versa. Our data suggest that, despite the relatively abundant water supply, periods with a substantial decrease of SWC or increase of temperature produced higher Re lowering the C sink strength. Our results reveal that nearly saturated soil conditions during the warm seasons can help to maintain lower ecosystem respiration rates and thus enhance the overall C sequestration capacity in this alpine swamp meadow. We argue that changes in soil hydrological conditions induced by a warming climate near permafrost (or seasonal frozen layers) may affect the C sink magnitude of wet and cold ecosystems through changes in soil hydrology and the subsequent effect on respiration losses.


reservations about the innovation of scientific questions and the reliability of some results that need to be addressed before the publication of this manuscript.
We thank Reviewer #1 for taking the time to assess our manuscript and for providing general positive comments and main concerns. We believe the comments have helped to improve the manuscript and we carefully considered them. Here specifically we clarify the innovation of our scientific questions addressed in this study. As REF#1 pointed out, alpine swamps in Tibetan Plateau are less focused at regional scale than alpine steppes and alpine meadows. This study highlights among other things the importance of soil water availability regulating carbon sink strength of an alpine swamp, which is characterized by saturated water condition and high soil water content (SWC). The role of soil water has often been neglected or assumed to be less important relative to other factors for carbon (C) cycling. This study provides a four-year field observation dataset to characterize and quantify the importance of soil water controlling the C sink strength of an alpine swamp -one key finding is that a 15% decrease in soil water can induce 25% higher respiration and therefore weaken C sink strength by 20%, and an additional 44% increase of temperature at annual scale can also weaken the C sink strength by about 50% (see answer to comment number 4). These new insights will help us to better understand, model and predict the complex C cycle dynamics in the Tibetan Plateau driven by the almost certain future intensified climate warming.
(1) The experimental site is located at Haibei in the northeastern Tibetan Plateau. We appreciate REF#1 for sharing this very useful information about other existing eddy covariance sites and related references. We have now cited these previous studies in the revised manuscript see L76 and L77. According to Wei et. al (2021), there are six observational studies about C fluxes at Haibei. However, only three of the them focused on alpine swamp meadows (or wetland in Wet et. al (2021)). Among them, one study only focused on a 1-year dataset (Zhang et al., 2008), and the other two characterized the same location . Moreover, these alpine swamp meadows were reported as a net C source while we our site showed a consistent C sink. The different directions of C exchange suggest that there are still uncertainties in our understanding of C exchange in this alpine swamp meadows, and further insights are obtained from studying multiple years of observations. Therefore, further efforts are still needed to improve our projection of C balance change of this ecosystem under changing climate.
Additionally, as mentioned before, previous studies focusing on C fluxes in alpine swamp meadows did not give enough consideration to the effects of soil water content on C fluxes given their nearly saturated nature. A number of previous studies have shown that temperature is an important driver of ecosystem respiration in similar alpine swamp meadows. For example, in the papers from Zhao et al. (2010) and Zhao et al. (2005), the authors showed that ecosystem respiration follows the exponential variation of soil temperature without considering soil water content. Zhu et al. (2020) also suggested that soil temperature plays the most important role in the change of monthly ecosystem respiration in the alpine wetland at Luanhaizi, northeastern Qinghai-Tibet Plateau. Therefore, in this study, we wanted to characterize and estimate the terrestrial C exchange while considering the potential effects of soil moisture. Meanwhile, it should be noted that Zhao et al. (2010) also noticed that the CO2 emission rates decrease notably after rain events, and Zhu et al. (2020) confirmed that annual precipitation exhibits significant impact on variation of annual net C uptake. All these existing studies have suggested that C fluxes are related to water availability condition, but few studies have found that soil moisture explicitly affects respiration, and its decrease further reduces net C uptake in alpine swamp meadows, this finding could be an important factor for carbon modelling in the future.
Finally, the addition and further analysis of multiple years of data from new sites is always very important also in a more regional/global context -there is a generalized sparsity of in situ observations where their temporal and spatial coverage is very limited. We believe that any effort and addition to this generalized lack of in situ data will be useful for both flux communities such as FLUXNET and ASIAFLUX and modelling community in general. Thank you for this insightful comment. Our first draft didn't present the findings clearly and it was somewhat misleading. First of all, we should acknowledge that the original title did not fully reflect our conclusions, therefore we revised it as follows: "Radiation, soil water content, and temperature interactions with carbon cycling in an alpine swamp meadow of the northeastern Qinghai-Tibet Plateau".
In the S4.3, we intend to put our results into a broader context by comparing with other surrounding alpine swamp meadow sites to highlight the effects of the complex interactions between temperature and soil water content on carbon fluxes. In this section, we did not intend to explicitly disentangle the most important drivers for the NEE at annual scales among these different sites, given that detailed observations for net radiation are lacking for other sites. Such comparison highlights the importance of SWC (precipitation as a proxy for SWC as explicit SWC is not present in these either) in controlling NEE. This further validates our Random Forest (RF) findings regarding the drivers for the Re in our site ( Figure 5). We should keep in mind that NEE is the difference between Re and GPP, environmental variables affecting Re and GPP could affect NEE indirectly (Song et al. 2011).We revealed that the main drivers for the GPP, Re and NEE are remarkably different. Net radiation is a key driver of seasonal and annual NEE and GPP, while soil water content is most important for Re at diurnal, seasonal and annual scales. Therefore, our findings from the RF analyses are not contradictory to the discussions regarding the difference in NEE and the potential influence factors among different sites.
However, your constructive comments point out an interesting scientific issue regarding the divergent drivers for the NEE dynamics at different time scales across different alpine swamp meadow sites. This will be an excellent point to be addressed in our future study. Many thanks for this comment -we have not clearly described the comparisons in S4.1 between years in the text. Since C fluxes are affected by plant phenology and climate factors including temperature, soil moisture, and radiation simultaneously ( Figure 5), in order to analyze the effects of single factor, ideally, other factors need to be identical or at least close (no significant differences). Based on this theory, we made our comparisons of specific time periods other than all the observation time. We now explicitly implemented the following text in S4.1, L288-293: "Since C fluxes are affected by plant phenology and climate factors including temperature, soil moisture, and radiation simultaneously (Fig. 5), to analyze the effects of single factor, ideally, other factors need to be identical or at least close (no significant differences).
Based on this theory and to better understand the underlying mechanisms around how SWC interacts with the C fluxes in the studied alpine swamp meadow ecosystem, we selected a specific group of data for further evaluation other than the entire observation time. " As described in S2.4, L177-184, we chose explicitly 2014 and 2015 for comparison in S4.1 because there was not a significant change in temperature (<1%) between these two periods, while soil water content decreased significantly more (15.4%) in 2015 (see Table S2). Therefore, within this set of conditions we can compare the influence of soil water content reduction on Re, and its influence further over NEE. Similarly, in S4.2, we also chose 2014 and 2018 for comparison, because the temperature difference between these two periods was greatest (25%) while there was no significant difference in soil water content (0.1%) (see again Table S2), so we could isolate and study the impact of temperature increase over the C fluxes.
Although the soil water content and ecosystem respiration in 2017 were at their highest, the temperature was also higher than in 2015, so we cannot compare it with 2015 to study separately the influence of soil water content change on ecosystem respiration and further net C uptake. To clarify this, we included a more detailed explanation about this comparison in the revision, see  SWC and Rn (i.e. 2017vs 2014, and 2018vs 2014. We made the comparison in each group to exclude the influence of plant phenology, which can influence C fluxes significantly. The magnitude of the differences between C fluxes in the same group were analysed by the independent-sample T-test method." Please note also that we included a new column with net radiation to complement the comparison and expand the discussion in order to improve the manuscript's clarity and align better with RF findings, this is including also radiation as suggested by REF#1 indirectly in the earlier point but also explicitly suggested by REF#2 later on.   (4) Another key conclusion of this study is that warming leads to higher C losses rather than enhanced C uptake. This is This is a great point, thanks for bringing this up. In the revised manuscript, we added a new Table   S4 to characterize the effect of temperature increase on net C uptake at annual basis and added more discussion accordingly.   (Table S4)." We thank REF#1 for pointing us towards this nice paper by Wei et al., 2021 "Plant uptake of CO2 outpaces losses from permafrost and plant respiration on the Tibetan Plateau". We benefited a lot from this paper. The conclusion from Wei et al. (2021) are based on data of 32 eddy covariance sites during 2002 to 2020, while our study only covers 4-years of year-round observations. Differences in time and space scales may help to explain the differences we found.
In fact, in the figure 4 from Wei et al. (2021) there are specific sites from Haibei where RRNEP (Response ratio of NEP to the warming rate) is smaller than 1.0, indicating a negative effect of warming on the NEP. Potential site-specific differences together with the particularity of water condition in of alpine swamp ecosystem could be the possible reason. We have added more text in the discussion section S4.2, L347-353.
Typo, thanks for picking it up. Changed accordingly.