Diagnosing the Temperature Sensitivity of Ecosystem Respiration in Northern High Latitude Regions

The accurate estimation of the temperature sensitivity of ecosystem respiration (Q₁₀) is important to understanding the terrestrial ecosystem carbon cycle, especially in northern high latitude regions (NHLs). Q₁₀ estimated by the conventional approach at the annual scale is influenced by seasonal confounding effects. Based on singular spectrum analysis, scale dependent parameter estimation (SCAPE) is considered an effective approach to eliminate confounding effects. Nevertheless, the performance of the decomposition and reconstruction schemes in the SCAPE approach in Q₁₀ estimation remains limited, which hampers its further application in larger scale systems. In this study, we utilized an improved scale dependent parameter estimation (iSCAPE) approach to analyze trends of the unconfounded Q₁₀ and its environmental controls in NHLs. The results showed that in NHLs, the confounding effects in forest ecosystems were smaller than those in cropland and grassland ecosystems. The apparent Q₁₀ estimated by the conventional approach varied among 32 sites with a mean value of 2.82 (95% confidence interval (CI): 2.72-2.91), while the mean intrinsic Q₁₀ estimated by the iSCAPE approach across the 32 sites was 1.53 (95% CI: 1.48-1.57). The apparent Q₁₀ increased with the annual mean temperature. The intrinsic Q₁₀ decreased with the increasing of spatial temperature gradient. The current study indicates that ecosystem respiration in NHLs is less sensitive to climate warming than previously reported. The seasonality of ecosystem respiration should be eliminated when estimating Q₁₀ to avoid overestimating climate carbon cycle feedback.