Search

Methane (CH4) emissions from natural landscapes constitute roughly half of global CH4 contributions to the atmosphere, yet large uncertainties remain in the absolute magnitude and the seasonality of emission quantities and drivers. Eddy covariance (EC) measurements of CH4 flux are ideal for constraining ecosystem-scale CH4 emissions due to quasi-continuous and high-temporal-resolution CH4 flux measurements, coincident carbon dioxide, water, and energy flux measurements, lack of ecosystem disturbance, and increased availability of datasets over the last decade. Here, we (1) describe the newly published dataset, FLUXNET-CH4 Version 1.0, the first open-source global dataset of CH4 EC measurements (available at https://fluxnet.org/data/fluxnet-ch4-community-product/, last access: 7 April 2021). FLUXNET-CH4 includes half-hourly and daily gap-filled and non-gap-filled aggregated CH4 fluxes and meteorological data from 79 sites globally: 42 freshwater wetlands, 6 brackish and saline wetlands, 7 formerly drained ecosystems, 7 rice paddy sites, 2 lakes, and 15 uplands. Then, we (2) evaluate FLUXNET-CH4 representativeness for freshwater wetland coverage globally because the majority of sites in FLUXNET-CH4 Version 1.0 are freshwater wetlands which are a substantial source of total atmospheric CH4 emissions; and (3) we provide the first global estimates of the seasonal variability and seasonality predictors of freshwater wetland CH4 fluxes. Our representativeness analysis suggests that the freshwater wetland sites in the dataset cover global wetland bioclimatic attributes (encompassing energy, moisture, and vegetation-related parameters) in arctic, boreal, and temperate regions but only sparsely cover humid tropical regions. Seasonality metrics of wetland CH4 emissions vary considerably across latitudinal bands. In freshwater wetlands (except those between 20∘ S to 20∘ N) the spring onset of elevated CH4 emissions starts 3 d earlier, and the CH4 emission season lasts 4 d longer, for each degree Celsius increase in mean annual air temperature. On average, the spring onset of increasing CH4 emissions lags behind soil warming by 1 month, with very few sites experiencing increased CH4 emissions prior to the onset of soil warming. In contrast, roughly half of these sites experience the spring onset of rising CH4 emissions prior to the spring increase in gross primary productivity (GPP). The timing of peak summer CH4 emissions does not correlate with the timing for either peak summer temperature or peak GPP. Our results provide seasonality parameters for CH4 modeling and highlight seasonality metrics that cannot be predicted by temperature or GPP (i.e., seasonality of CH4 peak). FLUXNET-CH4 is a powerful new resource for diagnosing and understanding the role of terrestrial ecosystems and climate drivers in the global CH4 cycle, and future additions of sites in tropical ecosystems and site years of data collection will provide added value to this database. All seasonality parameters are available at https://doi.org/10.5281/zenodo.4672601 (Delwiche et al., 2021). Additionally, raw FLUXNET-CH4 data used to extract seasonality parameters can be downloaded from https://fluxnet.org/data/fluxnet-ch4-community-product/ (last access: 7 April 2021), and a complete list of the 79 individual site data DOIs is provided in Table 2 of this paper.

While wetlands are the largest natural source of methane (CH4 ) to the atmosphere, they represent a large source of uncertainty in the global CH4 budget due to the complex biogeochemical controls on CH4 dynamics. Here we present, to our knowledge, the first multi-site synthesis of how predictors of CH4 fluxes (FCH4) in freshwater wetlands vary across wetland types at diel, multiday (synoptic), and seasonal time scales. We used several statistical approaches (correlation analysis, generalized additive modeling, mutual information, and random forests) in a wavelet-based multi-resolution framework to assess the importance of environmental predictors, nonlinearities and lags on FCH4 across 23 eddy covariance sites. Seasonally, soil and air temperature were dominant predictors of FCH4 at sites with smaller seasonal variation in water table depth (WTD). In contrast, WTD was the dominant predictor for wetlands with smaller variations in temperature (e.g., seasonal tropical/subtropical wetlands). Changes in seasonal FCH4 lagged fluctuations in WTD by ~17±11days, and lagged air and soil temperature by median values of 8±16 and 5±15days, respectively. Temperature and WTD were also dominant predictors at the multiday scale. Atmospheric pressure (PA) was another important multiday scale predictor for peat-dominated sites, with drops in PA coinciding with synchronous releases of CH4 . At the diel scale, synchronous relationships with latent heat flux and vapor pressure deficit suggest that physical processes controlling evaporation and boundary layer mixing exert similar controls on CH4 volatilization, and suggest the influence of pressurized ventilation in aerenchymatous vegetation. In addition, 1- to 4-h lagged relationships with ecosystem photosynthesis indicate recent carbon substrates, such as root exudates, may also control FCH4. By addressing issues of scale, asynchrony, and nonlinearity, this work improves understanding of the predictors and timing of wetland FCH4 that can inform future studies and models, and help constrain wetland CH4 emissions.

The National Ecological Observatory Network (NEON) is a multidecadal and continental-scale observatory with sites across the United States. Having entered its operational phase in 2018, NEON data products, software, and services become available to facilitate research on the impacts of climate change, land-use change, and invasive species. An essential component of NEON are its 47 tower sites, where eddy-covariance (EC) sensors are operated to determine the surface–atmosphere exchange of momentum, heat, water, and CO2. EC tower networks such as AmeriFlux, the Integrated Carbon Observation System (ICOS), and NEON are vital for providing the distributed observations to address interactions at the soil–vegetation–atmosphere interface. NEON represents the largest single-provider EC network globally, with standardized observations and data processing explicitly designed for intersite comparability and analysis of feedbacks across multiple spatial and temporal scales. Furthermore, EC is tightly integrated with soil, meteorology, atmospheric chemistry, isotope, phenology, and rich contextual observations such as airborne remote sensing and in situ sampling bouts. Here, we present an overview of NEON’s observational design, field operation, and data processing that yield community resources for the study of surface–atmosphere interactions. Near-real-time data products become available from the NEON Data Portal, and EC and meteorological data are ingested into AmeriFlux and FLUXNET globally harmonized data releases. Open-source software for reproducible, extensible, and portable data analysis includes the eddy4R family of R packages underlying the EC data product generation. These resources strive to integrate with existing infrastructures and networks, to suggest novel systemic solutions, and to synergize ongoing research efforts across science communities.

Time series of wetland methane fluxes measured by eddy covariance require gap-filling to estimate daily, seasonal, and annual emissions. Gap-filling methane fluxes is challenging because of high variability and complex responses to multiple drivers. To date, there is no widely established gap-filling standard for wetland methane fluxes, with regards both to the best model algorithms and predictors. This study synthesizes results of different gap-filling methods systematically applied at 17 wetland sites spanning boreal to tropical regions and including all major wetland classes and two rice paddies. Procedures are proposed for: 1) creating realistic artificial gap scenarios, 2) training and evaluating gap-filling models without overstating performance, and 3) predicting half-hourly methane fluxes and annual emissions with realistic uncertainty estimates. Performance is compared between a conventional method (marginal distribution sampling) and four machine learning algorithms. The conventional method achieved similar median performance as the machine learning models but was worse than the best machine learning models and relatively insensitive to predictor choices. Of the machine learning models, decision tree algorithms performed the best in cross-validation experiments, even with a baseline predictor set, and artificial neural networks showed comparable performance when using all predictors. Soil temperature was frequently the most important predictor whilst water table depth was important at sites with substantial water table fluctuations, highlighting the value of data on wetland soil conditions. Raw gap-filling uncertainties from the machine learning models were underestimated and we propose a method to calibrate uncertainties to observations. The python code for model development, evaluation, and uncertainty estimation is publicly available. This study outlines a modular and robust machine learning workflow and makes recommendations for, and evaluates an improved baseline of, methane gap-filling models that can be implemented in multi-site syntheses or standardized products from regional and global flux networks (e.g., FLUXNET). ISSN:0168-1923 ISSN:1873-2240

The FLUXNET2015 dataset provides ecosystem-scale data on CO2, water, and energy exchange between the biosphere and the atmosphere, and other meteorological and biological measurements, from 212 sites around the globe (over 1500 site-years, up to and including year 2014). These sites, independently managed and operated, voluntarily contributed their data to create global datasets. Data were quality controlled and processed using uniform methods, to improve consistency and intercomparability across sites. The dataset is already being used in a number of applications, including ecophysiology studies, remote sensing studies, and development of ecosystem and Earth system models. FLUXNET2015 includes derived-data products, such as gap-flled time series, ecosystem respiration and photosynthetic uptake estimates, estimation of uncertainties, and metadata about the measurements, presented for the frst time in this paper. In addition, 206 of these sites are for the frst time distributed under a Creative Commons (CC-BY 4.0) license. This paper details this enhanced dataset and the processing methods, now made available as open-source codes, making the dataset more accessible, transparent, and reproducible., European Union (EU), United States Department of Energy (DOE)

Carbon dioxide (CO2) exchange of tidal brackish wetlands and how they may be affected by restoration methods are largely unknown. The New Jersey Meadowlands, a tidal brackish estuary system, have had a long history of pollution, hydrological alterations, multiple restoration and mitigation treatments since the early 1970ies. To understand the effects of restoration, net ecosystem exchange (NEE) was measured with the eddy covariance technique in three urban tidal brackish wetland sites over three years: an organic matter amended restored site, a non-amended restored site and a natural (non-restored) wetland site. Our results showed all three sites to be a CO2 source in the wintertime reverting to a weak sink of CO2 in the summer and annually to be sources of CO2 for some of the sites for some of the time. However, the amended restored wetland was found to have the highest release of CO2. Annual destructive harvest of the major species at these wetlands (Spartina ssp and Phragmites australis) did not show significant differences in the aboveground net primary productivity and leaf area over the course of this study. Overall, aboveground biomass production of the three wetlands sites was within the expected range of tidal brackish marshes, but the sites differed in their NEE, ecosystem respiration, and gross ecosystem production. This investigation suggests that CO2 uptake and potential subsequent carbon sequestration may be strongly affected by the management and restoration history of the wetlands, and in our case these wetlands are often acting as a source of CO2, and not as a sink, as may have been expected.

Methane (CH4) emissions from natural landscapes constitute roughly half of global CH4 contributions to the atmosphere, yet large uncertainties remain in the absolute magnitude and the seasonality of emission quantities and drivers. Eddy covariance (EC) measurements of CH4 flux are ideal for constraining ecosystem-scale CH4 emissions, including their seasonality, due to quasi-continuous and high temporal resolution of flux measurements, coincident measurements of carbon, water, and energy fluxes, lack of ecosystem disturbance, and increased availability of datasets over the last decade. Here, we 1) describe the newly published dataset, FLUXNET-CH4 Version 1.0, the first global dataset of CH4 EC measurements (available at https://fluxnet.org/data/fluxnet-ch4- community-product/). FLUXNET-CH4 includes half-hourly and daily gap-filled and non gap-filled aggregated CH4 fluxes and meteorological data from 79 sites globally: 42 freshwater wetlands, 6 brackish and saline wetlands, 7 formerly drained ecosystems, 7 rice paddy sites, 2 lakes, and 15 uplands. Then, we 2) evaluate FLUXNET-CH4 representativeness for freshwater wetland coverage globally, because the majority of sites in FLUXNET-CH4 Version 1.0 are freshwater wetlands and because freshwater wetlands are a substantial source of total atmospheric CH4 emissions; and 3) provide the first global estimates of the seasonal variability and seasonality predictors of freshwater wetland CH4 fluxes. Our representativeness analysis suggests that the freshwater wetland sites in the dataset cover global wetland bioclimatic attributes (encompassing energy, moisture, and vegetation-related parameters) in arctic, boreal, and temperate regions, but only sparsely cover humid tropical regions. Seasonality metrics of wetland CH4 emissions vary considerably across latitudinal bands. In freshwater wetlands (except those between 20° S to 20° N) the spring onset of elevated CH4 emissions starts three days earlier, and the CH4 emission season lasts 4 days longer, for each degree C increase in mean annual air temperature. On average, the onset of increasing CH4 emissions lags soil warming by one month, with very few sites experiencing increased CH4 emissions prior to the onset of soil warming. In contrast, roughly half of these sites experience the spring onset of rising CH4 emissions prior to the spring increase in gross primary productivity (GPP). The timing of peak summer CH4 emissions does not correlate with the timing for either peak summer temperature or peak GPP. Our results provide seasonality parameters for CH4 modeling, and highlight seasonality metrics that cannot be predicted by temperature or GPP (i.e., seasonality of CH4 peak). The FLUXNET-CH4 dataset provides an open-access resource for CH4 flux synthesis, has a range of applications, and is unique in that it includes coupled measurements of important CH4 drivers such as GPP and temperature. Although FLUXNET-CH4 could certainly be improved by adding more sites in tropical ecosystems and by increasing the number of site-years at existing sites, it is a powerful new resource for diagnosing and understanding the role of terrestrial ecosystems and climate drivers in the global CH4 cycle. All seasonality parameters are available at https://doi.org/10.5281/zenodo.4408468. Additionally, raw FLUXNET-CH4 data used to extract seasonality parameters can be downloaded from https://fluxnet.org/data/fluxnet-ch4-community-product/, and a complete list of the 79 individual site data DOIs is provided in Table 2 in the Data Availability section of this document., Earth System Science Data Discussions, ISSN:1866-3591

Quality assurance and control is fundamental to ensuring the scientific usefulness of atmospheric measurements. Quality is relevant to all stages of data generation, from site selection and system design to all physical components of a measurement system (including their calibration, operation, and maintenance) as well as data handling and processing. This chapter describes useful practices and techniques for a comprehensive quality management program that is designed to minimize problems and quantify quality along the entire data generation chain. Widely applicable methods of post-field data quality control for instrumented (in-situ), visual, and remotely sensed observations are presented. The chapter concludes with example applications of QA/QC in measurement networks, and a discussion of common data problems. Finally, future developments in quality assurance and quality control are presented.

Water is a limited and valuable resource in California. A large proportion of the fresh water for southern California is supplied by the Sacramento and San Joaquin rivers. With recent efforts to restore large areas of land in the Sacramento–San Joaquin Delta region from farmland to managed wetlands, it is important to investigate the effect of wetland restoration on the local water cycle. We measured evapotranspiration using the eddy covariance method over six different sites in the Sacramento–San Joaquin Delta from 2013 to 2016. The three restored wetlands sites differed in time since restoration and wetland structure, i.e. the area of open water compared to closed vegetation. The three agricultural sites were a flooded rice field, an alfalfa field, and a grazed cattle pasture. In most years, annual evapotranspiration was significantly lower at the drained agricultural sites, with 652 ± 131 mm (2 year average ± standard deviation, SD) for the pasture and 901 ± 24 mm (3 year average ± SD) for the alfalfa field, compared to the two open water wetlands, with 1091 ± 144 mm (4 year average ± SD) for the Mayberry wetland and 1140 ± 67 mm (3 year average ± SD) at the East End wetland. Annual evapotranspiration at the flooded rice site (975 ± 50 mm, 4 year average ± SD) or the closed vegetation wetland (West Pond wetland, 996 ± 63 mm, 4 year average ± SD) was not significantly different from the drained alfalfa site across individual years. Our analysis showed that the structural difference between the wetlands, specifically the fraction of open water compared to closed vegetation, has a large impact on evapotranspiration dynamics. Through analysis of normalized equilibrium evaporation and nighttime evapotranspiration measurements we deduced that evapotranspiration at the wetland with a low fraction of open water surfaces was almost entirely dominated by plant transpiration with very little contribution from evaporation, despite the fact that the site was flooded and water was readily available. Both evaporation and transpiration contributed substantially to evapotranspiration at the two wetlands with larger fraction of open water surfaces. At the closed canopy site evaporation from subcanopy water seemed to be inhibited through two mechanisms: first, the closed canopy prevented heating of the water column and led to significantly cooler water temperatures, which reduced surface vapor pressure. Second, the closed canopy decoupled the water surface from the atmosphere and inhibited turbulent transport of water vapor away from the water surface. Our results provide valuable insights into the water use in California wetlands and can inform decisions on how to maximize water conservation during wetland restoration.

The following authors were omitted from the original version of this Data Descriptor: Markus Reichstein and Nicolas Vuichard. Both contributed to the code development and N. Vuichard contributed to the processing of the ERA-Interim data downscaling. Furthermore, the contribution of the co-author Frank Tiedemann was re-evaluated relative to the colleague Corinna Rebmann, both working at the same sites, and based on this re-evaluation a substitution in the co-author list is implemented (with Rebmann replacing Tiedemann). Finally, two affiliations were listed incorrectly and are corrected here (entries 190 and 193). The author list and affiliations have been amended to address these omissions in both the HTML and PDF versions. © 2021, This is a U.S. government work and not under copyright protection in the U.S.; foreign copyright protection may apply.

Large differences in instrumentation, site setup, data format, and operating system stymie the adoption of a universal computational environment for processing and analyzing eddy-covariance (EC) data. This results in limited software applicability and extensibility in addition to often substantial inconsistencies in flux estimates. Addressing these concerns, this paper presents the systematic development of portable, reproducible, and extensible EC software achieved by adopting a development and systems operation (DevOps) approach. This software development model is used for the creation of the eddy4R family of EC code packages in the open-source R language for statistical computing. These packages are community developed, iterated via the Git distributed version control system, and wrapped into a portable and reproducible Docker filesystem that is independent of the underlying host operating system. The HDF5 hierarchical data format then provides a streamlined mechanism for highly compressed and fully self-documented data ingest and output. The usefulness of the DevOps approach was evaluated for three test applications. First, the resultant EC processing software was used to analyze standard flux tower data from the first EC instruments installed at a National Ecological Observatory (NEON) field site. Second, through an aircraft test application, we demonstrate the modular extensibility of eddy4R to analyze EC data from other platforms. Third, an intercomparison with commercial-grade software showed excellent agreement (R2 = 1.0 for CO2 flux). In conjunction with this study, a Docker image containing the first two eddy4R packages and an executable example workflow, as well as first NEON EC data products are released publicly. We conclude by describing the work remaining to arrive at the automated generation of science-grade EC fluxes and benefits to the science community at large. This software development model is applicable beyond EC and more generally builds the capacity to deploy complex algorithms developed by scientists in an efficient and scalable manner. In addition, modularity permits meeting project milestones while retaining extensibility with time.

Wetlands have the ability to accumulate large amounts of carbon (C), and therefore wetland restoration has been proposed as a means of sequestering atmospheric carbon dioxide (CO2) to help mitigate climate change. There is a growing interest in using the C services of wetlands to help reduce habitat loss and finance restoration projects. However, including wetlands in C markets worldwide requires a better mechanistic understanding of CO2 and methane exchange and instruments and models that can accurately and inexpensively monitor and predict these fluxes across global wetlands. Remote sensing technology, including near-surface and satellite instruments/approaches, is an effective tool for modeling C fluxes including gross primary productivity (GPP) from the site to global scale. In this study, we evaluate the potential of using digital cameras as a simple, cost-effective means of estimating GPP in restored wetlands, and assess the suitability of using Landsat data to model GPP in these environments for regional upscaling. Our research focused on restored temperate freshwater marshes due to their high C sequestration potential. As observed in other ecosystems, daily GPP was strongly correlated with site greenness derived from camera imagery (GCCcam). Based on this, we show the potential of using GCCcam and eddy covariance data to adapt and parameterize a light use efficiency (LUE) model to predict daily GPP. The LUE model combining GCCcam and meteorological data was able to explain up to 91% of the variation in daily GPP at the restored marshes, and predict annual GPP budgets within 0% to 20% of observed budgets. However, model performance decreased with increasing site complexity, highlighting the need to explicitly consider spatial heterogeneity in LUE models. We also tested a similar model using Landsat-derived indices, and found that although model performance was high at a homogeneous wetland dominated by emergent vegetation, data-model agreement decreased at a site comprised of a mixture of open water and vegetation, reflecting limitations of Landsat data. Nonetheless, we show that digital camera and Landsat imagery can be used to model photosynthesis in restored wetlands, providing low-cost methods for monitoring C capture that can be used in C market-funded wetland conservation and restoration.

The partitioning of net ecosystem exchange of CO2 (NEE) into photosynthesis and respiration can be challenging and is often associated with assumptions that yield unknown amounts of uncertainty, thereby hindering model development. This occurs because we are inferring two pieces of information from one equation and measurement, NEE. While there are multiple methods for partitioning NEE, each has unique limitations that are difficult to evaluate as these techniques are often not implemented simultaneously. Here we present an analysis of multiple partitioning methods (a non-linear regression model known as the Reichstein method, artificial neural networks (ANN), stable carbon isotopes, and soil respiration (Rsoil) measurements) under ideal field conditions. We measured ecosystem-scale fluxes of stable C isotopes via a new quantum cascade laser (QCL) spectrometer and paired those measurements with a biophysical model CANVEG in order to close a system of equations and solve for gross primary productivity (GPP) and ecosystem respiration (Reco). Isotope-partitioned GPP and Reco were on average 10–13% lower than Reichstein- and ANN-partitioned GPP and Reco. These results suggest that the Reichstein and ANN approaches, which use nighttime NEE to infer daytime Reco, may be overestimating Reco and GPP during the day. This may be attributed to higher plant respiration rates at night compared to the day, otherwise known as the Kok effect. As isotope measurements and theory become more accessible to diverse ecosystems it will grow to be more than a benchmarking technique and into a valuable means of improving understanding of carbon cycling. Future studies are encouraged to evaluate these techniques in increasingly complex ecosystems to determine when significant differences appear between partitioning methods and how those differences influence modeling of terrestrial carbon budgets.

Wetlands and flooded peatlands can sequester large amounts of carbon (C) and have high greenhouse gas mitigation potential. There is growing interest in financing wetland restoration using C markets; however, this requires careful accounting of both CO2 and CH4 exchange at the ecosystem scale. Here we present a new model, the PEPRMT model (Peatland Ecosystem Photosynthesis Respiration and Methane Transport), which consists of a hierarchy of biogeochemical models designed to estimate CO2 and CH4 exchange in restored managed wetlands. Empirical models using temperature and/or photosynthesis to predict respiration and CH4 production were contrasted with a more process-based model that simulated substrate-limited respiration and CH4 production using multiple carbon pools. Models were parameterized by using a model-data fusion approach with multiple years of eddy covariance data collected in a recently restored wetland and a mature restored wetland. A third recently restored wetland site was used for model validation. During model validation, the process-based model explained 70% of the variance in net ecosystem exchange of CO2 (NEE) and 50% of the variance in CH4 exchange. Not accounting for high respiration following restoration led to empirical models overestimating annual NEE by 33–51%. By employing a model-data fusion approach we provide rigorous estimates of uncertainty in model predictions, accounting for uncertainty in data, model parameters, and model structure. The PEPRMT model is a valuable tool for understanding carbon cycling in restored wetlands and for application in carbon market-funded wetland restoration, thereby advancing opportunity to counteract the vast degradation of wetlands and flooded peatlands.

Temperate freshwater wetlands are among the most productive terrestrial ecosystems, stimulating interest in using restored wetlands as biological carbon sequestration projects for greenhouse gas reduction programs. In this study, we used the eddy covariance technique to measure surface energy carbon fluxes from a constructed, impounded freshwater wetland during two annual periods that were 8 years apart: 2002–2003 and 2010–2011. During 2010–2011, we measured methane (CH4) fluxes to quantify the annual atmospheric carbon mass balance and its concomitant influence on global warming potential (GWP). Peak growing season fluxes of latent heat and carbon dioxide (CO2) were greater in 2002–2003 compared to 2010–2011. In 2002, the daily net ecosystem exchange reached as low as −10.6 g C m−2 d−1, which was greater than 3 times the magnitude observed in 2010 (−2.9 g C m−2 d−1). CH4 fluxes during 2010–2011 were positive throughout the year and followed a strong seasonal pattern, ranging from 38.1 mg C m−2 d−1 in the winter to 375.9 mg C m−2 d−1 during the summer. The results of this study suggest that the wetland had reduced gross ecosystem productivity in 2010–2011, likely due to the increase in dead plant biomass (standing litter) that inhibited the generation of new vegetation growth. In 2010–2011, there was a net positive GWP (675.3 g C m−2 yr−1), and when these values are evaluated as a sustained flux, the wetland will not reach radiative balance even after 500 years.

We present 6.5 years of eddy covariance measurements of fluxes of methane (FCH4) and carbon dioxide (FCO2) from a flooded rice paddy in Northern California, USA. A pronounced warming trend throughout the study associated with drought and record high temperatures strongly influenced carbon (C) budgets and provided insights into biophysical controls of FCO2 and FCH4. Wavelet analysis indicated that photosynthesis (gross ecosystem production, GEP) induced the diel pattern in FCH4, but soil temperature (Ts) modulated its amplitude. Forward stepwise linear models and neural networking modeling were used to assess the variables regulating seasonal FCH4. As expected due to their competence in modeling nonlinear relationships, neural network models explained considerably more of the variance in daily average FCH4 than linear models. During the growing season, GEP and water levels typically explained most of the variance in daily average FCH4. However, Ts explained much of the interannual variability in annual and growing season CH4 sums. Higher Ts also increased the annual and growing season ratio of FCH4 to GEP. The observation that the FCH4 to GEP ratio scales predictably with Ts may help improve global estimates of FCH4 from rice agriculture. Additionally, Ts strongly influenced ecosystem respiration, resulting in large interannual variability in the net C budget at the paddy, emphasizing the need for long-term measurements particularly under changing climatic conditions.

Methane (CH4) exchange in wetlands is complex, involving nonlinear asynchronous processes across diverse time scales. These processes and time scales are poorly characterized at the whole-ecosystem level, yet are crucial for accurate representation of CH4 exchange in process models. We used a combination of wavelet analysis and information theory to analyze interactions between whole-ecosystem CH4 flux and biophysical drivers in two restored wetlands of Northern California from hourly to seasonal time scales, explicitly questioning assumptions of linear, synchronous, single-scale analysis. Although seasonal variability in CH4 exchange was dominantly and synchronously controlled by soil temperature, water table fluctuations, and plant activity were important synchronous and asynchronous controls at shorter time scales that propagated to the seasonal scale. Intermittent, subsurface water table decline promoted short-term pulses of methane emission but ultimately decreased seasonal CH4 emission through subsequent inhibition after rewetting. Methane efflux also shared information with evapotranspiration from hourly to multiday scales and the strength and timing of hourly and diel interactions suggested the strong importance of internal gas transport in regulating short-term emission. Traditional linear correlation analysis was generally capable of capturing the major diel and seasonal relationships, but mesoscale, asynchronous interactions and nonlinear, cross-scale effects were unresolved yet important for a deeper understanding of methane flux dynamics. We encourage wider use of these methods to aid interpretation and modeling of long-term continuous measurements of trace gas and energy exchange.

© 2018 John Wiley & Sons Ltd Wetlands are the largest source of methane (CH4) globally, yet our understanding of how process-level controls scale to ecosystem fluxes remains limited. It is particularly uncertain how variable soil properties influence ecosystem CH4emissions on annual time scales. We measured ecosystem carbon dioxide (CO2) and CH4fluxes by eddy covariance from two wetlands recently restored on peat and alluvium soils within the Sacramento–San Joaquin Delta of California. Annual CH4fluxes from the alluvium wetland were significantly lower than the peat site for multiple years following restoration, but these differences were not explained by variation in dominant climate drivers or productivity across wetlands. Soil iron (Fe) concentrations were significantly higher in alluvium soils, and alluvium CH4fluxes were decoupled from plant processes compared with the peat site, as expected when Fe reduction inhibits CH4production in the rhizosphere. Soil carbon content and CO2uptake rates did not vary across wetlands and, thus, could also be ruled out as drivers of initial CH4flux differences. Differences in wetland CH4fluxes across soil types were transient; alluvium wetland fluxes were similar to peat wetland fluxes 3 years after restoration. Changing alluvium CH4emissions with time could not be explained by an empirical model based on dominant CH4flux biophysical drivers, suggesting that other factors, not measured by our eddy covariance towers, were responsible for these changes. Recently accreted alluvium soils were less acidic and contained more reduced Fe compared with the pre-restoration parent soils, suggesting that CH4emissions increased as conditions became more favorable to methanogenesis within wetland sediments. This study suggests that alluvium soil properties, likely Fe content, are capable of inhibiting ecosystem-scale wetland CH4flux, but these effects appear to be transient without continued input of alluvium to wetland sediments.

In the widely-used eddy-covariance (EC) technique, it is often assumed that the air storage term, i.e. the change of below-turbulence-sensor scalar abundance, is negligible or comprises a small part of net surface-atmosphere exchange (NSAE). Previous studies have demonstrated that this assumption is often violated where non-turbulent processes prevail, and thus it is important to measure and calculate air storage in flux measurements. However, the implementation of air storage measurement and calculation is not ubiquitous as EC standard turbulent flux. In most cases, air storage is not a standard data product or even neglected in EC flux tower measurements. In other cases, air storage term is calculated simply using only the measurements at the tower top. This gap between the ideal initiative and actual implementation motivates us to derive and release one of the first community resources to facilitate the consistent measurement and calculation of EC air storage across sites. These resources include (i) the standardized air storage term measurement setup design at National Ecological Observatory Network (NEON) sites; (ii) the development and public release of the eddy4R.stor open-source air storage R-package; (iii) the derivation and public release of storage term data products, measured and calculated consistently across 47 NEON sites; and (iv) exploration the scientific usefulness of these resources through example use cases, specifically the exploration of the bias of the air storage term when different measurement level intensity used and exploration of the air storage term pattern. We expect the consistent air storage measurement and calculation can better serve the overall purpose of the EC technique to provide more reliable measurement of NSAE for the community. This can further benefit the community accurate depiction of the sub-daily to diurnal cycle of surface fluxes in doing carbon cycle flux partitioning, land modeling, and studying ecosystem response to weather extremes.

Measurements of hyperspectral canopy reflectance provide a detailed snapshot of information regarding canopy biochemistry, structure and physiology. In this study, we collected five years of repeated canopy hyperspectral reflectance measurements for a total of over 100 site visits within the flux footprints of two eddy covariance towers at a pasture and rice paddy in Northern California. The vegetation at both sites exhibited dynamic phenology, with significant inter-annual variability in the timing of seasonal patterns that propagated into inter-annual variability in measured hyperspectral reflectance. We used partial least-squares regression (PLSR) modeling to leverage the information contained within the entire continuous canopy reflectance spectra (400–900 nm) in order to investigate questions regarding the connection between measured hyperspectral reflectance and landscape-scale fluxes of net ecosystem exchange (NEE) and gross primary productivity (GPP) across multiple timescales, from instantaneous flux to monthly-integrated flux. With the PLSR models developed from this large dataset we achieved a high level of predictability for both NEE and GPP flux in these two ecosystems, where the R2 of prediction with an independent validation dataset ranged from 0.24 to 0.69. The PLSR models achieved the highest skill at predicting the integrated GPP flux for the week prior to the hyperspectral canopy reflectance collection, whereas the NEE flux often achieved the same high predictive power at the daily- through monthly-integrated flux timescales. The high level of predictability achieved by PLSR regression in this study demonstrated the potential for using repeated hyperspectral canopy reflectance measurements to help partition NEE measurements into its component fluxes, GPP and ecosystem respiration, and for using continuous hyperspectral reflectance measurements to model regional carbon flux in future analyses.

It is necessary to partition eddy covariance measurements of carbon dioxide exchange into its offsetting gross fluxes, canopy photosynthesis, and ecosystem respiration, to understand the biophysical controls on the net fluxes. And independent estimates of canopy photosynthesis (G) and ecosystem respiration (R) are needed to validate and parametrize carbon cycle models that are coupled with climate and ecosystem dynamics models. Yet there is a concern that carbon flux partitioning methods may suffer from spurious correlation because derived values of canopy photosynthesis and ecosystem respiration both contain common information on net carbon fluxes at annual time scales. We hypothesized that spurious correlation among canopy photosynthesis and ecosystem respiration can be minimized using day–night conditional sampling of CO2 exchange; daytime fluxes are dominated by photosynthesis and nighttime fluxes are dominated by respiration. To test this hypothesis, we derived explicit equations that quantify the degree of spurious correlation between photosynthesis and respiration. Theoretically, day and night samples of net carbon exchange share a different common variable, daytime ecosystem respiration, and the degree of spurious correlation depends upon the variance of this shared variable. We then applied this theory to ideal measurements of carbon exchange of over a vigorous, irrigated, and frequently harvested alfalfa field in the sunny and windy region of California, the Sacramento-San Joaquin Delta, where soil CO2 efflux is strong. In this case, we found that the correlation coefficient between canopy photosynthesis and ecosystem respiration was −0.79. This relatively high correlation between canopy photosynthesis and respiration was mostly real as the degree of spurious correlation was only −0.32. We then expanded this analysis to the FLUXNET database that spans a spectrum of climate and plant functional types. We found, on average, that the correlation between gross photosynthesis and ecosystem respiration, using day–night sampling, was close to minus one (−0.828 ± 0.130). For perspective, a large fraction of this correlation was real, as the degree of spurious correlation (Eq. (22)) was −0.526. We conclude that the potential for spurious correlation between canopy photosynthesis and ecosystem respiration across the FLUXNET database was moderate. Looking across the database, we found that the least negative spurious correlations coefficients (>−0.3) were associated with seasonal deciduous forests. The most negative spurious correlations coefficients (

This study presents the systematic development of an open-source, flexible and modular eddy-covariance (EC) data processing framework. This is achieved through adopting a Development and Systems Operation (DevOps) philosophy, building on the eddy4R family of EC code packages in the R Language for Statistical Computing as foundation. These packages are community-developed via the GitHub distributed version control system and wrapped into a portable and reproducible Docker filesystem that is independent of the underlying host operating system. The HDF5 hierarchical data format then provides a streamlined mechanism for highly compressed and fully self-documented data ingest and output. This framework is applicable beyond EC, and more generally builds the capacity to deploy complex algorithms developed by scientists in an efficient and scalable manner. In addition, modularity permits meeting project milestones while retaining extensibility with time. The efficiency and consistency of this framework is demonstrated in the form of three application examples. These include tower EC data from first instruments installed at a National Ecological Observatory (NEON) field site, aircraft flux measurements in combination with remote sensing data, as well as a software intercomparison. In conjunction with this study, the first two eddy4R packages and simple NEON EC data products are released publicly. While this proof-of-concept represents a significant advance, substantial work remains to arrive at the automated framework needed for the streaming generation of science-grade EC fluxes.

Agricultural drainage of organic soils has resulted in vast soil subsidence and contributed to increased atmospheric carbon dioxide (CO2) concentrations. The Sacramento-San Joaquin Delta in California was drained over a century ago for agriculture and human settlement and has since experienced subsidence rates that are among the highest in the world. It is recognized that drained agriculture in the Delta is unsustainable in the long-term, and to help reverse subsidence and capture carbon (C) there is an interest in restoring drained agricultural land-use types to flooded conditions. However, flooding may increase methane (CH4) emissions. We conducted a full year of simultaneous eddy covariance measurements at two conventional drained agricultural peatlands (a pasture and a corn field) and three flooded land-use types (a rice paddy and two restored wetlands) to assess the impact of drained to flooded land-use change on CO2 and CH4 fluxes in the Delta. We found that the drained sites were net C and greenhouse gas (GHG) sources, releasing up to 341 g C m(-2) yr(-1) as CO2 and 11.4 g C m(-2) yr(-1) as CH4. Conversely, the restored wetlands were net sinks of atmospheric CO2, sequestering up to 397 g C m(-2) yr(-1). However, they were large sources of CH4, with emissions ranging from 39 to 53 g C m(-2) yr(-1). In terms of the full GHG budget, the restored wetlands could be either GHG sources or sinks. Although the rice paddy was a small atmospheric CO2 sink, when considering harvest and CH4 emissions, it acted as both a C and GHG source. Annual photosynthesis was similar between sites, but flooding at the restored sites inhibited ecosystem respiration, making them net CO2 sinks. This study suggests that converting drained agricultural peat soils to flooded land-use types can help reduce or reverse soil subsidence and reduce GHG emissions.

The importance and consequences of extreme events on the global carbon budget are inadequately understood. This includes the differential impact of extreme events on various ecosystem components, lag effects, recovery times, and compensatory processes. In the summer of 2007 in Barrow, Arctic Alaska, there were unusually high air temperatures (the fifth warmest summer over a 65-year period) and record low precipitation (the lowest over a 65-year period). These abnormal conditions were associated with substantial desiccation of the Sphagnum layer and a reduced net Sphagnum CO2 sink but did not affect net ecosystem exchange (NEE) from this wet-sedge arctic tundra ecosystem. Microbial biomass, NH4+ availability, gross primary production (GPP), and ecosystem respiration (Reco) were generally greater during this extreme summer. The cumulative ecosystem CO2 sink in 2007 was similar to the previous summers, suggesting that vascular plants were able to compensate for Sphagnum CO2 uptake, despite the impact on other functions and structure such as desiccation of the Sphagnum layer. Surprisingly, the lowest ecosystem CO2 sink over a five summer record (2005–2009) was observed during the 2008 summer (~70% lower), directly following the unusually warm and dry summer, rather than during the extreme summer. This sink reduction cannot solely be attributed to the potential damage to mosses, which typically contribute ~40% of the entire ecosystem CO2 sink. Importantly, the return to a substantial cumulative CO2 sink occurred two summers after the extreme event, which suggests a substantial resilience of this tundra ecosystem to at least an isolated extreme event. Overall, these results show a complex response of the CO2 sink and its sub-components to atypically warm and dry conditions. The impact of multiple extreme events requires further investigation.

Restored wetlands are a complex mosaic of open water and new and old emergent vegetation patches, where multiple environmental and biological drivers contribute to the measured heterogeneity in methane (CH4) flux. In this analysis, we replicated the measurements of CH4 flux using the eddy covariance technique at three tower locations within the same wetland site to parse the spatiotemporal variability in CH4 flux contributed by large-scale seasonal variations in climate and phenology and short-term variations in flux footprint movement over a mosaic of vegetation and open water. Using a hierarchical statistical model accounting for site-level environmental effects, tower-level footprint and biological effects, and temporal autocorrelation, we partitioned the key drivers of the daily CH4 flux variability among the three replicated towers. The daily mean air temperature and mean friction velocity, a measure of momentum transfer, explained a significant variability in CH4 flux across the three towers, and the abundance and spatial aggregation of vegetation in the flux footprint along with the daily gross primary productivity explained much of the tower-level variability. This statistical model captured 67% of the total variance in the daily integrated growing season CH4 fluxes at this site, which bridged an order of magnitude from 80 to 480 mg C m−2 d−1 during the measurement period from 10 May 2012 to 24 October 2012.

Wetlands can influence global climate via greenhouse gas (GHG) exchange of carbon dioxide (CO2 ), methane (CH4 ), and nitrous oxide (N2 O). Few studies have quantified the full GHG budget of wetlands due to the high spatial and temporal variability of fluxes. We report annual open-water diffusion and ebullition fluxes of CO2 , CH4 , and N2 O from a restored emergent marsh ecosystem. We combined these data with concurrent eddy-covariance measurements of whole-ecosystem CO2 and CH4 exchange to estimate GHG fluxes and associated radiative forcing effects for the whole wetland, and separately for open-water and vegetated cover types. Annual open-water CO2 , CH4 , and N2 O emissions were 915 ± 95 g C-CO2 m-2 yr-1 , 2.9 ± 0.5 g C-CH4 m-2 yr-1 , and 62 ± 17 mg N-N2 O m-2 yr-1 , respectively. Diffusion dominated open-water GHG transport, accounting for >99% of CO2 and N2 O emissions, and ~71% of CH4 emissions. Seasonality was minor for CO2 emissions, whereas CH4 and N2 O fluxes displayed strong and asynchronous seasonal dynamics. Notably, the overall radiative forcing of open-water fluxes (3.5 ± 0.3 kg CO2 -eq m-2 yr-1 ) exceeded that of vegetated zones (1.4 ± 0.4 kg CO2 -eq m-2 yr-1 ) due to high ecosystem respiration. After scaling results to the entire wetland using object-based cover classification of remote sensing imagery, net uptake of CO2 (-1.4 ± 0.6 kt CO2 -eq yr-1 ) did not offset CH4 emission (3.7 ± 0.03 kt CO2 -eq yr-1 ), producing an overall positive radiative forcing effect of 2.4 ± 0.3 kt CO2 -eq yr-1 . These results demonstrate clear effects of seasonality, spatial structure, and transport pathway on the magnitude and composition of wetland GHG emissions, and the efficacy of multiscale flux measurement to overcome challenges of wetland heterogeneity.