Application of an open-path eddy covariance methane flux measurement system to a larch forest in eastern Siberia

https://doi.org/10.1016/j.agrformet.2019.107860Get rights and content

Highlights

  • A canopy-scale daily-mean CH4 flux over a Cajander larch forest was net emission.

  • CH4 flux was nearly the same as the upper boundary of the limit of flux detection.

  • Daytime CH4 flux showed dependencies on both temperature and soil moisture.

Abstract

Canopy-Scale methane (CH4) flux measurement over a larch forest in eastern Siberia was conducted by eddy covariance method using an open-path CH4 gas analyzer. Though the uncorrected flux showed strong CH4 uptake in the daytime, this changed to CH4 emission after density and spectroscopic effects were corrected. Random flux errors calculated from cross-covariance functions suggested that CH4 flux was nearly the same as the upper boundary of the limit of flux detection at the 95th percentile, being barely resolved by the measurement system; and that most of the daytime CH4 flux remained positive even after uncertainties due to random flux errors were taken into consideration. CH4 flux showed clear diurnal variation, representing emission in the daytime and near-zero in the nighttime, irrespective of wind direction. The daytime CH4 flux was dependent on both air temperature and volumetric soil water content. The CH4 flux from May 29, to June 12, was calculated as net emissions of 4.9–13.8 nmol m2 s1 in daily average, ranging between the forest floor and a mesotrophic fen near this site measured by static chambers in a previous study.

Introduction

Monitoring of methane (CH4) flux dynamics at various terrestrial ecosystems is becoming important, as CH4 is the third most important greenhouse gas after water (H2O) and carbon dioxide (CO2). However, the role of forests in CH4 cycling has not yet been well understood, as the monitoring of CH4 fluxes at the canopy- or ecosystem-scale is still difficult (Shoemaker et al., 2014). Though previous chamber measurements at the forest floor show that many upland forest soils are a net CH4 sink (Megonigal and Guenther, 2008), the canopy- or ecosystem-scale CH4 flux over forests can be positive (i.e., net CH4 source), due to the presence of CH4 emissions from “hotspots” such as small wet areas within the micrometeorological flux footprint (e.g., Ueyama et al., 2018). Focusing on boreal forests, canopy-scale CH4 flux measurements have been attempted in Saskatchewan, Canada by a gradient method (Simpson et al., 1997), in Interior Alaska by a closed-path eddy covariance method (Iwata et al., 2015), and in central Sweden by Bowen ratio and combined eddy covariance and gradient methods (Sundqvist et al., 2015). All these studies have shown CH4 emission (positive flux) over forests at the canopy-scale, while Simpson et al. (1997) and Sundqvist et al. (2015) reported CH4 uptake (negative flux) at the forest floor by chamber measurements. Besides, even though approximately 44% of all boreal forests are located east of the Urals in northern Russia (Jarvis et al., 2001), canopy-scale CH4 flux over forests in such regions (e.g., eastern Siberia) has not been reported. Therefore, knowledge regarding canopy-scale CH4 flux over boreal forests in eastern Siberia can contribute to the further understanding of global CH4 cycling.

In the last decade, an open-path CH4 analyzer LI-7700 (LI-COR Biogeosciences, Lincoln, NE, USA) has been developed (LI-COR Inc., 2011, McDermitt, Burba, Xu, Anderson, Komissarov, Riensche, Schedlbauer, Starr, Zona, Oechel, Oberbauer, Hastings, 2011) which enables us to measure the variation in CH4 concentration, without the need for a power-consuming vacuum pump. As emphasized by McDermitt et al. (2011), the low-power characteristics of the LI-7700 will be a strong advantage when CH4 flux measurements are explored by solar-powered operation in remote areas without access to electricity, including in a vast portion of boreal forests. On the other hand, the open-path CH4 analyzer inevitably presents two errors: one is induced by the fluctuation of air density due to variations in temperature and humidity (Webb et al., 1980); the other is the spectroscopic effects induced by line broadening due to pressure, humidity, and temperature of the air (McDermitt et al., 2011). Though the methods for correcting these errors are theoretically grounded and robust, the extent of these corrections can be significantly large relative to the density-corrected (or ‘true’) CH4 flux, which sometimes changes the sign of fluxes from negative measurement (or raw, uncorrected) values to positive corrected ones, especially in the case of ecosystems where the ‘true’ CH4 flux is very small. For example, Chamberlain et al. (2017) evaluated the effect of these corrections on CH4 flux over sites with different flux strengths, and showed that corrections accounted for a greater than 100% increase in daily CH4 flux at “negligible-flux” sites (alfalfa and pasture). Chamberlain et al. (2017) also observed positive CH4 flux after correction in the daytime over a pavement airfield (zero-flux site), which was interpreted as an overcorrection. Since the absolute value of canopy-scale CH4 flux over forests is reported to be generally small compared to wetlands, we must pay especially close attention to instrument uncertainties when the LI-7700 open-path CH4 analyzer is applied to forest CH4 flux measurements.

When measuring eddy covariance fluxes, sampling (random) errors due to instrumental noise and natural variability (i.e., stochastic nature of turbulence) are unavoidable, resulting in uncertainties in the obtained data and finally calculated fluxes. Several methods for assessing such uncertainties have been proposed (e.g. Finkelstein, Sims, 2001, Langford, Acton, Ammann, Valach, Nemitz, 2015, Wienhold, Welling, Harris, 1995) and used for evaluation of random errors in fluxes (e.g. Iwata, Harazono, Ueyama, Sakabe, Nagano, Kosugi, Takahashi, Kim, 2015., Mauder, Cuntz, Druee, Graf, Rebmann, Schmid, Schmidt, Steinbrecher, 2013), and incorporated in a real-time flux monitoring and surveillance system (Kim et al., 2015). Random flux error due to instrumental noise has also been investigated (e.g. Billesbach, 2011, Mauder, Cuntz, Druee, Graf, Rebmann, Schmid, Schmidt, Steinbrecher, 2013); Rannik et al. (2016) reviewed these studies and suggested the method by Finkelstein and Sims (2001) is robust and accurate for estimating random flux error.

In the case of small CH4 flux measurements over forests, the signal-to noise ratio (SNR) of the open-path CH4 analyzer can be low, which is similar to cases of flux measurements of trace gases and volatile organic compounds (Langford et al., 2015). Langford et al. (2015) focused on eddy covariance data with low SNR, and showed procedures for quantifying random white noise and SNR for the measured data, based on the practical method for determining the components of variance attributed to a genuine (structured) signal and unstructured white noise using the auto-covariance function (Lenschow et al., 2000). They also assessed the random error and limit to the detection of fluxes using the root-mean-square error of the cross-covariance function within a time window defined well away from the point of zero time lag, a modification of the method used by Wienhold et al. (1995) and Spirig et al. (2005). Using these procedures, Langford et al. (2015) clearly showed large random errors and a large fraction of errors associated with instrument noise in the case of isoprene and acetone fluxes, and assessed the quality ensemble mean of diurnal variation in acetone and benzene fluxes, comparing different methods for determining time lag between vertical wind velocity and measured scalar concentrations. The procedure of Langford et al. (2015) is considered useful for the assessment of the applicability of the LI-7700 open-path CH4 analyzer to small CH4 flux sites.

The objectives of this study are to evaluate the applicability of the open-path CH4 flux measurement system to a larch forest in eastern Siberia using random flux error analysis according to Langford et al. (2015), and to analyze the characteristics of the CH4 flux measured in this forest.

Section snippets

Study site and instruments

Eddy covariance flux measurements were conducted in a larch forest in Spasskaya Pad Scientific Forest Station (6215′18″N, 12914′29″E, 220 m a.s.l.), which is located on the left bank of the Lena River, about 20 km northwest of Yakutsk, in the Republic of Sakha, Russia (Fig. 1), where long-term flux observation has been conducted since 1998 (e.g. Kotani, Saito, Kononov, Petrov, Maximov, Iijima, Ohta, 2019, Ohta, Hiyama, Tanaka, Kuwada, Maximov, Ohata, Fukushima, 2001, Ohta, Kotani, Iijima,

Auto-covariance and cross-covariance functions

Given the time series of variable x, the auto-covariance function fx,x(t) is defined as follows.fx,x(t)=fx,x(t)=1Nti=0Nt(xix¯)(xi+tx¯),where t is the number of data points associated with the time lag, and N is the number of data points in the time series. Note that fx,x(0)=σx2, where σx is the standard deviation of x. If the time series of x is affected by white noise, the contribution of this noise to fx,x(t) appears only at t=0 (Langford, Acton, Ammann, Valach, Nemitz, 2015, Lenschow,

Results

During the study period, the daytime latent heat flux λE gradually increased from about 50 W m2 to 150 W m2, where the energy balance closure was 79.9% (Fig. 4a). This might be due to the extremely dry atmospheric conditions with its water vapor pressure deficit (VPD) up to about 30 hPa (Fig. 4b). On the other hand, there were only three precipitation events, totaling 8 mm. These conditions resulted in a gradual decrease in volumetric soil water content at 10-cm depth (Fig. 4d). Air

Validity of positive daytime methane flux after corrections

The daytime positive Fm in this study was a consequence of the cumulative effects of high frequency, WPL, and SS corrections, whereas the uncorrected daytime Fm showed strong negative values. Since whether the daytime Fm is positive or negative is dependent on the magnitude of the cumulative effects of corrections, the sign of outcomes of Fm after corrections should be carefully validated. Such changes in the sign of Fm after corrections using the LI-7700 open-path CH4 analyzer have been

Conclusions

A canopy-scale CH4 flux over a larch forest in eastern Siberia in 2016 measured by an open-path eddy covariance flux system was shown here, together with random flux error and limit of flux detection (LoD). CH4 flux after corrections showed clear diurnal variation, showing emission in the daytime and near-zero in the nighttime, irrespective of wind direction. Though the calculated methane flux was nearly the same as the upper boundary of the LoD at the 95th percentile, this flux was considered

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We thank Nate Bauer of International Arctic Research Center, University of Alaska Fairbanks for English proofreading of the manuscript. We are also grateful to anonymous reviewers for valuable comments. This research was supported by the Arctic Challenge for Sustainability (ArCS) Project of the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan. Methane flux data was provided by North-Eastern Federal University (NEFU), Russia, under the agreement between NEFU and

References (54)

  • T. Nakai et al.

    Ultrasonic anemometer angle of attack errors under turbulent conditions

    Agric. For. Meteorol.

    (2012)
  • T. Nakai et al.

    Parameterisation of aerodynamic roughness over boreal, cool- and warm-temperate forests

    Agric. For. Meteorol.

    (2008)
  • T. Nakai et al.

    A comparison between various definitions of forest stand height and aerodynamic canopy height

    Agric. For. Meteorol.

    (2010)
  • T. Ohta et al.

    Seasonal variation in the energy and water exchanges above and below a larch forest in eastern Siberia

    Hydrol. Process.

    (2001)
  • T. Ohta et al.

    Interannual variation of water balance and summer evapotranspiration in an eastern Siberian larch forest over a 7-year period (1998–2006)

    Agric. For. Meteorol.

    (2008)
  • O. Peltola et al.

    Field intercomparison of four methane gas analyzers suitable for eddy covariance flux measurements

    Biogeosci.

    (2013)
  • D. Vickers et al.

    Quality control and flux sampling problems for tower and aircraft data

    J. Atmos. Oceanic Tech.

    (1997)
  • G. Burba et al.

    Accounting for spectroscopic effects in laser-based open-path eddy covariance flux measurements

    Glob. Change Biol.

    (2019)
  • G. Burba et al.

    Calculating CO2 and H2O eddy covariance fluxes from an enclosed gas analyzer using an instantaneous mixing ratio

    Glob. Change Biol.

    (2012)
  • S.D. Chamberlain et al.

    Evaluation of density corrections to methane fluxes measured by open-path eddy covariance over contrasting landscapes

    Bound. Layer Meteorol.

    (2017)
  • K.R. Covey et al.

    Methane production and emissions in trees and forests

    New Phytol.

    (2019)
  • M. Detto et al.

    Simplified expressions for adjusting higher-order turbulent statistics obtained from open path gas analyzers

    Bound. Layer Meteorol.

    (2007)
  • P.L. Finkelstein et al.

    Sampling error in eddy correlation flux measurements

    J. Geophys. Res.

    (2001)
  • J.H.C. Gash et al.

    Applying a linear detrend to eddy correlation data in real time

    Bound. Layer Meteorol.

    (1996)
  • D.P. Gatley et al.

    A twenty-first century molar mass for dry air

    HVAC&R Res.

    (2008)
  • Gill Instruments

    Software bug affecting ‘w’ wind component of the WindMaster family

    Technical key note series number kn11509v6

    (2016)
  • J. van Huissteden et al.

    Summer soil CH4 emission and uptake in taiga forest near Yakutsk, Eastern Siberia

    Agric. For. Meteorol.

    (2008)
  • Cited by (9)

    • Long-term dynamics of soil, tree stem and ecosystem methane fluxes in a riparian forest

      2022, Science of the Total Environment
      Citation Excerpt :

      Ecosystem-level studies using the eddy covariance (EC) technique suggest that tree canopies contribute to the ecosystem CH4 budget (do Carmo et al., 2006). Eddy covariance measurements of CH4 fluxes have been conducted in boreal forests (Iwata et al., 2015; Nakai et al., 2020), temperate forests (Wang et al., 2013; Hommeltenberg et al., 2014; Shoemaker et al., 2014; Flanagan et al., 2020), and tropical/sub-tropical forests (do Carmo et al., 2006; Dalmagro et al., 2019; Griffis et al., 2020; Liu et al., 2020). In addition, data are available from FLUXNET network sites in Finland, Switzerland, Russia and Indonesia.

    • Constraining models for methane oxidation based on long-term continuous chamber measurements in a temperate forest soil

      2021, Agricultural and Forest Meteorology
      Citation Excerpt :

      CH4 fluxes in upland soils decrease with increasing water-filled porosity, which determines the gas diffusivity, and has a low temperature sensitivity (Del Grosso et al., 2000; Hashimoto et al., 2011b; Smith et al., 2000). Canopy-scale tower flux measurements showed that upland forests are a net CH4 source with microspots for CH4 emissions, such as banks of streams and small ponds having a strong emission potential, within a forest (Sakabe et al., 2012; Sundqvist et al., 2015a; Ueyama et al., 2018; Nakai et al., 2020). Compared with studies on spatial variations, long-term observations are limited for CH4 fluxes in upland soils, except for a few long-term studies (Wu et al., 2011).

    • Lessons learned from more than a decade of greenhouse gas flux measurements at boreal forests in eastern Siberia and interior Alaska

      2021, Polar Science
      Citation Excerpt :

      The clear diurnal variation of the CH4 flux, emission in the daytime and near-zero in the nighttime, was mostly explained by the atmospheric stability. The range of the daily mean CH4 flux from May 29 to June 12 was in net emissions of 4.9–13.8 nmol m−2 s−1 (Fig. 6; see also Figur 16 in Nakai et al., 2020). This was within the range of the CH4 emissions measured by static chambers on the forest floor and at a mesotrophic fen near the site, reported by a previous study (van Huissteden et al., 2008).

    View all citing articles on Scopus
    View full text