| 27 Nov 2025
Carbon–Water Flux Coupling Characteristics and Driving Factors at Multi-temporal Scales in an Alpine Meadow Ecosystem on the Tibetan Plateau
Abstract. Alpine meadow ecosystems play a crucial role in the global carbon and water cycles, with water use efficiency (WUE) serving as a key indicator of carbon-water coupling. Investigating the characteristics of carbon and water fluxes and WUE in alpine meadows on the Tibetan Plateau (TP) is essential for accurately assessing carbon budget, water cycling, and carbon–water interactions under climate change. This study utilized eddy covariance observations from 2012 to 2017 in an alpine meadow of the eastern TP to analyze the temporal dynamics of carbon fluxes (net ecosystem carbon exchange, NEE; ecosystem respiration, Re; gross primary productivity, GPP), water flux (evapotranspiration, ET), and WUE across daily, monthly to seasonal, and inter-annual timescales. Ridge regression was applied to identify the main drivers of carbon and water fluxes and WUE at different time-scale. The results indicate that: (1) the alpine meadow acted as a carbon sink, with a multi-year average NEE of 109.7 gC m-2y-1, and carbon and water fluxes as well as WUE exhibited pronounced temporal variations across daily, monthly to seasonal, and inter-annual timescales; (2) daily and monthly to seasonal variations of carbon fluxes were primarily driven by soil temperature (Ts), while ET was mainly controlled by radiation. At the inter-annual timescale, precipitation (PRE) and leaf area index (LAI) were the dominant factors influencing carbon and water fluxes; (3) Ts regulated WUE at daily, monthly to seasonal scales, whereas PRE was the key factor controlling carbon–water coupling at the inter-annual timescale. These findings enhance our understanding of the coupling characteristics and driving mechanisms of carbon and water fluxes in alpine meadows, providing a scientific basis for predicting the responses of grassland ecosystems on the TP to future climate change.
1. Summary and General Assessment
The manuscript presents an analysis of carbon and water fluxes (GPP, Re, NEE, and ET) and Water Use Efficiency (WUE) in an alpine meadow ecosystem on the Tibetan Plateau, utilizing a 6-year dataset (2012–2017) derived from Eddy Covariance observations. The authors aim to disentangle the driving factors of these fluxes across multiple temporal scales—daily, seasonal, and inter-annual—employing Ridge Regression to address multicollinearity among environmental variables.
I would like to commend the authors on several strong aspects of this work:
Regional Importance: The study focuses on the Tibetan Plateau, a region often described as the "Asian Water Tower." Given the sensitivity of this ecosystem to global climate change and the logistical challenges of maintaining long-term monitoring sites in such high-altitude environments, in situ data from this region are highly valuable to the scientific community.
Multi-scale Approach: The attempt to explicitly differentiate driving mechanisms across different time scales (daily vs. seasonal vs. inter-annual) is a robust conceptual framework. It moves beyond simple averages and acknowledges that the biological response of the meadow varies depending on the temporal window.
Methodological Choice: The use of Ridge Regression is an appropriate statistical choice for this type of environmental data, demonstrating the authors' awareness of the issues related to multicollinearity between temperature, radiation, and vapor pressure deficit (VPD).
However, despite these clear merits, I have significant concerns regarding the statistical reliability of the conclusions drawn at the inter-annual scale. The primary issue is the sample size (N = 6 years) used to determine climatic drivers via regression analysis. While the daily and seasonal analyses appear sound, the inter-annual results carry a high degree of uncertainty that needs to be addressed or re-framed before publication.
2. Specific Comments
Major Concerns:
Statistical Power at the Inter-annual Scale:
The authors perform regression analyses to identify the dominant drivers of inter-annual variability (e.g., identifying Precipitation and LAI as key factors). With a dataset spanning only 6 years (2012–2017), there are only 6 data points available for these regressions. From a statistical standpoint, this sample size is insufficient to reliably disentangle the effects of multiple predictors (Precipitation, Ta, Ts, SWC, etc.). There is a high risk that the identified correlations are spurious or driven by a single anomalous year.
Suggestion: The authors should explicitly acknowledge this limitation. I strongly suggest reducing the emphasis on the quantitative regression results for the inter-annual scale. Instead, the discussion should focus on qualitative comparisons with other long-term studies in the region or focus more on the process-based response to specific extreme events within the 6-year period (e.g., a specific drought or heatwave year vs. a normal year).
Ecological Interpretation (Precipitation vs. Temperature):
The site is characterized as a "humid" alpine meadow. The results suggest that soil temperature ($T_s$) drives fluxes at the daily scale, which is consistent with energy-limited environments. However, the conclusion that Precipitation drives inter-annual variability is somewhat counter-intuitive for a humid ecosystem where water is typically not the limiting factor.
Question: Could this dominance of precipitation be an artifact of the short time series (e.g., one significantly dry year skewing the regression)? The authors need to provide a deeper biological explanation for why a humid ecosystem would be so sensitive to annual precipitation amounts.
Minor Comments / Technical Corrections:
Ridge Regression Details:
While the choice of Ridge Regression is praised, the manuscript should provide more details on how the regularization parameter ($\lambda$ or $k$) was optimized. Did the authors use cross-validation? Providing a trace plot of the coefficients would strengthen the methodological description.
Energy Balance Closure:
For any study based on Eddy Covariance, reporting the Energy Balance Closure (EBC) is standard practice to assess data quality. I could not find a clear evaluation of the EBC (slope of $LE+H$ vs. $Rn-G$) for the observation period. Please add this information to the Materials and Methods or Results section.
Gap-Filling Uncertainty:
Please specify the percentage of missing data for the carbon and water fluxes and describe how the uncertainty introduced by gap-filling was handled or evaluated, especially given the harsh environmental conditions that often lead to instrument failure.