Divergent NEE balances from manual‐chamber CO₂ fluxes linked to different measurement and gap‐filling strategies: A source for uncertainty of estimated terrestrial C sources and sinks?

Abstract

Manual closed‐chamber measurements are commonly used to quantify annual net CO2 ecosystem exchange (NEE) in a wide range of terrestrial ecosystems. However, differences in both the acquisition and gap filling of manual closed‐chamber data are large in the existing literature, complicating inter‐study comparisons and meta analyses. The aim of this study was to compare common approaches for quantifying CO2 exchange at three methodological levels. (1) The first level included two different CO2 flux measurement methods: one via measurements during mid‐day applying net coverages (mid‐day approach) and one via measurements from sunrise to noon (sunrise approach) to capture a span of light conditions for measurements of NEE with transparent chambers. (2) The second level included three different methods of pooling measured ecosystem respiration (RECO) fluxes for empirical modeling of RECO: campaign‐wise (19 single‐measurement‐day RECO models), season‐wise (one RECO model for the entire study period), and cluster‐wise (two RECO models representing a low and a high vegetation status). (3) The third level included two different methods of deriving fluxes of gross primary production (GPP): by subtracting either proximately measured RECO fluxes (direct GPP modeling) or empirically modeled RECO fluxes from measured NEE fluxes (indirect GPP modeling). Measurements were made during 2013–2014 in a lucerne‐clover‐grass field in NE Germany. Across the different combinations of measurement and gap‐filling options, the NEE balances of the agricultural field diverged strongly (–200 to 425 g CO2‐C m−2). NEE balances were most similar to previous studies when derived from sunrise measurements and indirect GPP modeling. Overall, the large variation in NEE balances resulting from different data‐acquisition or gap‐filling strategies indicates that these methodological decisions should be made very carefully and that they likely add to the overall uncertainty of greenhouse gas emission factors. Preferably, a standard approach should be developed to reduce the uncertainty of upscaled estimates.

Assessing the impact of peat erosion on growing season CO2 fluxes by comparing erosional peat pans and surrounding vegetated haggs

“…This uncertainty consists of both natural variability which is known to be significant when using multiple plots (Laine et al 2009) and uncertainty associated with modelling. It has been shown that different treatment of closed chamber data can result in variation in estimated NEE of 0.25 gCO 2 m -2 d -1 over annual estimates (Huth et al 2017) sufficient to change the estimate of ecosystem exchange from a net CO 2 source to a net CO 2 sink. Accepting this uncertainty, the models suggest both landscape components are losing carbon with greater loss from the haggs even with a greater vegetation cover.…”

Upland grasslands in Northern England were atmospheric carbon sinks regardless of management regimes

“…To model GPP, we parameterised and compared the performance of two non-linear regression equations: a rectangular hyperbolic saturation curve (Thornley and Johnson, 1990; Equation 4) which has been widely used for modelling grasslands' GPP (e.g. Campbell et al, 2015;Du et al, 2014;Dyukarev, 2017;Elsgaard et al, 2012;Huth et al, 2017), and a multiplicative variant (Equation 5) which helps to account for subtle changes in moisture and temperature that may affect photosynthesis.…”

“…Most models of ER are based on exponential dependence on temperature (Moffat et al, 2007) but the most commonly used is the Arrhenius model (Equation 6; e.g. Du et al, 2014;Elsgaard et al, 2012;Huth et al, 2017;Jacobs et al, 2007).…”

“…is temperature when ER is zero, usually constrained to 227.13 K to avoid over-parameterisation (Elsgaard et al, 2012;Huth et al, 2017); E0 is temperature sensitivity coefficient.…”

Negative effects of climate change on upland grassland productivity and carbon fluxes are not attenuated by nitrogen status

“…A rectangular hyperbolic saturation curve (Equation 3) is widely used for modelling GPP in grasslands (e.g. Dyukarev, 2017;Elsgaard et al, 2012;Huth et al, 2017).…”