Using a Bayesian framework to account for advection in seven years of snowpack CO2 fluxes in a mortality-impacted subalpine forest
Introduction
Soil respiration, from plant roots and soil heterotrophs, is an indispensable measurement for terrestrial carbon (C) cycle research, but characterizing its variability across time and space is often a challenge (Ryan and Law, 2005, Phillips et al., 2016). Automated measurement techniques have allowed higher-resolution temporal measurements of soil respiration, advancing our understanding of diel, seasonal and interannual controls over this significant portion of the annual C budget (Carbone and Vargas, 2008, Vargas and Allen, 2008, van Asperen et al., 2017). However, in non-tropical areas, most of these advancements deal only with respiration during the growing season, leaving much to be understood about respiration during the winter. Wintertime respiration is especially important in seasonally snow-covered ecosystems where the insulative properties of snow keep soil temperature at or above freezing, maintaining biological activity (McDowell et al., 2000, McGuire et al., 2000, Brooks et al., 2004, Tucker et al., 2016). In fact, wintertime snowpack respiration may show some seasonality (Hubbard et al., 2005), which should be taken into account in sampling designs if the goal is to close annual C budgets. Improving our ability to capture spatial and temporal variability in wintertime respiration is necessary to understand interannual dynamics of the C cycle in snow-dominated forests. Eddy covariance captures high (e.g., half-hourly) temporal variability in C fluxes, and its large footprint makes it a useful tool for aggregating spatial variability.
However, it is challenging to glean detailed information about wintertime soil respiration from EC. The method tends to have signal-to-noise issues with low CO2 fluxes typical of wintertime (Webb et al., 2016). Also, EC fluxes at half-hourly timescales may not represent contemporaneous CO2 production in the soil due to temporal dynamics of transport processes through snowpack, including advection (Euskirchen et al., 2012). Consequently, there is a need for an independent method to measure wintertime soil respiration underneath snowpack at a high temporal resolution similar to that of EC.
Automated measurements of snowpack CO2 concentrations can be made at fine time scales (Seok et al., 2009), and these data can help address a major methodological challenge with measuring soil respiration that exits through the snowpack. Typically, wintertime soil respiration is calculated by applying Fick's first law of diffusion to a vertical CO2 gradient measured in the snowpack (Sommerfeld et al., 1993, Hubbard et al., 2005), assuming steady-state conditions wherein flux emanating from the snowpack surface is equivalent to respiration from the soil into the snowpack. Under conditions where snowpack CO2 concentrations are largely diffusion dominated, this method estimates soil respiration with high confidence. However, wind generates advective fluxes in which CO2 is pumped from the surface layer of snow, generating non-linear CO2 profiles that result in erroneous fluxes when only Fick's first law is used for calculation (Massman et al., 1995, Takagi et al., 2005, Bowling and Massman, 2011). Thus, we need some way to infer soil respiration into the bottom of the snowpack when advective fluxes from the snowpack surface create nonlinearity in snowpack CO2 profiles. Measuring vertical snowpack CO2 profiles at frequent (e.g., hourly) timescales can provide enough information to calculate (1) storage, advective and diffusive fluxes, (2) flux from the soil into the snowpack for ecological interpretation, and (3) flux from the top of the snowpack to the atmosphere for comparison to eddy covariance.
Despite its low magnitude, winter respiration may be an indicator of changing C cycle dynamics in forests that have experienced mass tree mortality events, but no results have been presented to date. Studies have found either no significant effect of bark beetle mortality on growing season soil respiration or only a modest, short-term change (Morehouse et al., 2008, Moore et al., 2013, Speckman et al., 2015, Borkhuu et al., 2015). Detecting change in growing season soil respiration is complicated by its sensitivity to short-term changes in soil temperature and moisture (Davidson et al., 1998) that may obscure mortality effects. However, microclimate forcings on respiration are lessened during the winter, when snowpack insulates the soil surface and maintains soil temperature slightly above freezing and holds soil moisture at a more constant level (Contosta et al., 2016). Thus, mortality impacts on respiration may be easier to detect in under-snowpack respiration compared to growing season soil respiration. We hypothesize that tree mortality effects on substrate availability for respiration will result in a noticeable increase in wintertime soil respiration underneath snowpack in the first few years following tree mortality from fine root mortality and needlefall.
One key challenge with ecological measurements that are made with high temporal frequency but low spatial replication is assigning measures of uncertainty. Calculation of soil respiration from vertical CO2 profiles requires information about variables that influence the rate at which CO2 moves through the snowpack. The precise values of these variables (e.g., snowpack temperature, porosity, and density) at the locations of CO2 measurements are usually unknown; rather, they are estimated from nearby snow pits. These sources of uncertainty, even when not precisely known, can be taken into account in a Bayesian analytical framework. Having a reliable measure of uncertainty around these fluxes will allow a comparison between our snowpack CO2 profile method and accompanying EC tower, as well as allowing the quantification of significance in observed variation of fluxes over time.
In this study, we calculate total CO2 flux from the snowpack surface at hourly timescales to compare with contemporaneous half-hourly EC measurements. In addition, we examine factors related to overall flux uncertainty and the contribution of advective versus diffusive processes to total flux from the snowpack. Lastly, we hypothesize that the biological flux of CO2 from soil into the bottom of the snowpack would increase as a result of the tree mortality event. To achieve this, we develop a Bayesian-based method to estimate both advective and diffusive components of CO2 fluxes from both the bottom of the snowpack (interpreted as biological soil respiration) and from the top of the snowpack (comparable to eddy covariance measurements) and apply it over a 7-year time period at an AmeriFlux site located within a subalpine forest in the Rocky Mountains of southern Wyoming, USA, that had recently experienced tree mortality.
Section snippets
Study site and data collection
Research was conducted at the Glacier Lakes Ecosystem Experiments Site (GLEES) in southeastern Wyoming which is part of the AmeriFlux network of eddy covariance sites (US-GLE, ameriflux.lbl.gov). GLEES is a mature subalpine forest at 3190 m elevation and receives 1250 mm of average annual precipitation, about 80% as snow. Snowpack persists an average of 8 months per year; development typically begins in October and persists until June or July. Forest species composition consists of old growth
CO2 Profile model
The shape of the CO2 profile with depth in the snowpack varied over time, sometimes fluctuating between linear and concave shapes within the course of less than a day (i.e., Fig. 2a–d). The profile shape is reflected by α in the profile model (Eq. (1)), with high α corresponding to high concavity in the profile shape. Because concave profiles had low CO2 gradients at the top of the snowpack and high CO2 gradients at the bottom of the snowpack (Fig. 3), per Fick's first law of diffusion, these
Discussion
Using 38,237 hourly measurements of snowpack CO2 profiles over the course of 7 years, we constructed a detailed time series of wintertime soil respiration, and CO2 fluxes out of the snowpack, in a subalpine forest in southern Wyoming, USA following a bark beetle mortality event during which 75% to 85% of the tree canopy was lost (Frank et al., 2014). Our work builds upon a strong history of snowpack CO2 research at this site (Sommerfeld et al., 1993, Sommerfeld et al., 1996, Massman et al., 1997
Acknowledgements
This work was funded by the U.S. Forest Service, the AmeriFlux program, and the Climate and Land Use Change Mission Area of U.S. Geological Survey. We acknowledge the assistance of Brianna Miles for collection of snowpack survey CO2 measurements and Dr. Jace Fahnestock for the use of his data. We also thank Monica Dorning and anonymous peer reviewers, whose comments and suggestions improved the quality of this paper. Any use of trade, product or firm names is for descriptive purposes only and
References (58)
- et al.
Does soil respiration decline following bark beetle induced forest mortality? Evidence from a lodgepole pine forest
Agric. For. Meteorol.
(2015) - et al.
Underestimates of sensible heat flux due to vertical velocity measurement errors in non-orthogonal sonic anemometers
Agric. For. Meteorol.
(2013) - et al.
Carbon and nitrogen cycling immediately following bark beetle outbreaks in southwestern ponderosa pine forests
For. Ecol. Manag.
(2008) - et al.
Diel variation in isotopic composition of soil respiratory CO2 fluxes: the role of non-steady state conditions
Agric. For. Meteorol.
(2017) - et al.
Seasonality of soil CO2 efflux in a temperate forest: biophysical effects of snowpack and spring freeze-thaw cycles
Agric. For. Meteorol.
(2013) - et al.
Strong resilience of soil respiration components to drought-induced die-off resulting in forest secondary succession?
Oecologia
(2016) - et al.
Total belowground carbon flux in subalpine forests is related to leaf area index, soil nitrogen, and tree height
Ecosphere
(2016) - et al.
Advective and Diffusive Snowpack Carbon Dioxide Fluxes from 2009 to 2016 at GLEES
(2017) - et al.
Winter carbon dioxide effluxes from Arctic ecosystems: an overview and comparison of methodologies
Global Biogeochem. Cycles
(2010) - et al.
Persistent wind-induced enhancement of diffusive CO2 transport in a mountain forest snowpack
J. Geophys. Res.
(2011)
Carbon and nitrogen cycling in snow-covered environments?
Geogr. Compass
Carbon limitation of soil respiration under winter snowpacks: potential feedbacks between growing season and winter carbon fluxes
Glob. Change Biol.
Addressing the influence of instrument surface heat exchange on the measurements of CO2 flux from open-path gas analyzers?
Glob. Change Biol.
Automated soil respiration measurements: new information, opportunities, and challenges
New Phytol.
Winter soil respiration in a humid temperate forest: the roles of moisture, temperature, and snowpack
J. Geophys. Res. Biogeosci.
Soil water content and temperature as independent or confounded factors controlling soil respiration in a temperate mixed hardwood forest
Glob. Change Biol.
Simulating coupled carbon and nitrogen dynamics following mountain pine beetle outbreaks in the western United States
J. Geophys. Res.
Seasonal patterns of carbon dioxide and water fluxes in three representative tundra ecosystems in northern Alaska?
Ecosphere
Ecosystem CO2/H2O fluxes are explained by hydraulically limited gas exchange during tree mortality from spruce bark beetles
J. Geophys. Res. Biogeosci.
Bayesian Data Analysis
Seasonal patterns in soil surface CO2 flux under snow cover in 50 and 300 year old subalpine forests
Biogeochemistry
Wet meadow ecosystems contribute the majority of over-winter soil respiration from snow-scoured alpine tundra
J. Geophys. Res. Biogeosci.
Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan
Bayesian estimation supersedes the t test
J. Exp. Psychol.: Gen.
Rapid rebound of soil respiration following partial stand disturbance by tree girdling in a temperate deciduous forest
Oecologia
Process-level controls on CO2 fluxes from a seasonally snow-covered subalpine meadow soil, Niwot Ridge, Colorado
Biogeochemistry
On the temperature dependence of soil respiration
Funct. Ecol.
Advective transport of CO2 in permeable media induced by atmospheric pressure fluctuations: 1. An analytical model
J. Geophys. Res.
CO2 flux through a Wyoming seasonal snowpack: diffusional and pressure pumping effects
Biogeochemistry of Seasonally Snow-Covered Catchments (Proceedings of a Boulder Symposium, July 1995)
Cited by (11)
The effect of static pressure-wind covariance on vertical carbon dioxide exchange at a windy subalpine forest site
2021, Agricultural and Forest MeteorologyBark beetle outbreaks alter biotic components of forested ecosystems
2021, Bark Beetle Management, Ecology, and Climate ChangeA new perspective on the open-path infrared gas analyzer self-heating correction
2020, Agricultural and Forest MeteorologyCitation Excerpt :Other studies have investigated on the impact of the self-heating correction on the net ecosystem exchange of carbon dioxide in cold-weather environments (Amiro, 2010; Reverter et al., 2011; Wang et al., 2017a; Wang et al., 2017b). Though this is not our focus, our revised self-heating correction has been incorporated by Berryman et al. (2018) into a comparison between winter CO2 respiration measured with eddy covariance versus snowpack CO2 gradients at GLEES; their results indicated that the average winter CO2 fluxes measured with the two methodologies were similar. In summary, we conclude that significant IRGA self-heating exists in a cold, windy, high elevation, mountainous field site.
A Bayesian inversion framework to evaluate parameter and predictive inference of a simple soil respiration model in a cool-temperate forest in western Japan
2020, Ecological ModellingCitation Excerpt :Variations in soil respiratory CO2 and the associated uncertainty therein (i.e. predictive uncertainty) can then be inferred. This technique has been thus widely used in the biogeochemical and atmosphere-surface energy exchange studies (Hashimoto et al., 2011; van Oijen et al., 2011; Kim et al., 2014; Minet et al., 2015; Lu et al., 2017; Berryman et al., 2018), while few studies have employed the GL approach with residuals varying in skewness and kurtosis (Schoups and Vrugt, 2010; Elshall et al., 2019). In the present study, we use 2 years’ worth of soil respiratory CO2 flux observations collected from two open-top chambers in a cool-temperate forest in western Japan during the period 2017–2018 (hereafter referred to as the ‘long-term simulation’), to constrain parameter and predictive uncertainty of a soil respiration model based on Bayesian statistics.
Progress in the theory, hypothesis and application of the methods measuring soil CO<inf>2</inf> flux gradient
2022, Chinese Journal of Plant EcologyModeling Gas Flow Velocities in Soils Induced by Variations in Surface Pressure, Heat, and Moisture Dynamics
2022, Journal of Advances in Modeling Earth Systems