5th iLEAPS Science Conference Abstracts - D1

Abstracts – Session D1


D101ORAL-0146: Role of global wetlands in renewed atmospheric growth of methane

Ben Poulter1

1NASA-GSFC, Greenbelt, The United States of America

The renewed growth of atmospheric methane concentrations since 2007 is associated with a depletion of carbon-13, indicating that an increase in biogenic sources are responsible. Biogenic sources include wetlands and agricultural emissions, which together account for almost two thirds of total methane emissions to the atmosphere. Here, as part of the Global Carbon Project CH4 budget, we evaluate the role of global wetlands in the renewed growth period 2007-2012, where atmospheric concentrations of methane rose by ~6 ppb per year, or ~17 Tg CH4 per year. Wetland methane emissions are modeled by integrated the area and dynamics of saturated soils with knowledge of methanogenesis and methanotrophy represented by empirical or process-based relationships. The area and monthly temporal dynamics of wetlands were derived from passive microwave observations of surface inundation from the NASA SWAMPS database integrated with a global-gridded inventory of wetlands based on Global Lakes and Wetlands Database (GLWD). Eleven biogeochemical models were used to provide an ensemble of methane emissions using the same diagnostic wetlands database and the CRU-NCEP meteorological information for temperature, precipitation and radiation. Global methane emissions were estimated at 184 Tg CH4 per year (+/- 22 Tg), and between 2000-2012, no globally significant trend was simulated. Regionally, however, an increase in high-latitude emissions by 1.2 Tg CH4 per year was balanced by a decrease in tropical emissions of 0.9 Tg CH4 per year. High interannual variability in emissions was observed and corresponded with the El Nino Southern Oscillation and its effects on tropical moisture and temperature. Uncertainties in wetland area and dynamics, as well as the reliance of models on estimating 'net' emissions, rather than gross component fluxes of production and consumption, remain to be addressed with use of tower, airborne, and remote sensing platforms.

D102 ORAL-0116: Evaluating year-to-year anomalies in tropical wetland methane emissions using satellite CH4 observations

Robert Parker1, Hartmut Boesch1, Joe McNorton2, Edward Comyn-Platt3, Manuel Gloor4, Chris Wilson2, Martyn Chipperfield2, Garry Hayman3, Anthony Bloom5

1Earth Observation Science, Department of Physics and Astronomy, University of Leicester, Leicester, United Kingdom 2School of Earth and Environment, University of Leeds, Leeds, United Kingdom 3Centre for Ecology & Hydrology, Wallingford, United Kingdom 4School of Geography, University of Leeds, Leeds, United Kingdom 5Jet Propulsion Laboratory, California Institute of Technology, Pasadena, The United States of America

In this work we use atmospheric observations of total column methane from the GOSAT satellite to evaluate current state-of-the-art methane wetland emission estimates, including two versions of the JULES land surface model. To assess how well the models reproduce the observed methane inter-annual variability, we evaluate the detrended model seasonal cycle against observations. Globally, we find that the latitudinal means of the detrended seasonal cycle agree well between simulations and observations but maximum differences in the tropics of 28.1-34.8 ppb suggest that all simulations fail to capture the full extent of the tropical wetland seasonal cycle.
We perform a more detailed analysis for the major natural wetland regions in South America: the major seasonally flooded savannah of the Pantanal (Brazil) and Llanos de Moxos (Bolivia) regions; and the riverine wetlands formed by the Parana River (Argentina). We find that for certain time periods there are large discrepancies between simulation and observation. In particular, strong enhancements in the methane seasonal cycle are observed over the Pantanal and Llanos de Moxos region in 2010, 2011 and 2014 and over the Parana River region in 2010 and 2014. We find highly consistent behaviour between the time and location of these methane anomalies and the change in wetland extent, itself driven by precipitation related to El Nino Southern Oscillation (ENSO) activity.
Specific mechanisms within land surface models, such as the inability to increase wetland extent through overbank inundation, are highlighted as reasons for observed discrepancies and can lead to under-estimation by as much as 50% of the observed emissions. As the hydrology of these regions is heavily linked to ENSO variability, in both the El Nino and La Nina phases, being able to reproduce such changes in wetland behaviour is important for successfully predicting methane emissions from these regions.

D103 ORAL-0356: Large methane emissions from inundated trees of the Amazon basin

Sunitha Pangala1, David Bastviken2, Alex Prast3, 2, Vincent Gauci1

1The Open University, Milton Keynes, United Kingdom 2Linköping University, Linköping, Sweden 3University Federal of Rio de Janeiro, Rio de Janeiro, Brazil

Methane emissions from wetland trees are an overlooked source of methane. Although, recent studies suggest trees to emit up-to 90% of methane in forested wetlands, tree-mediated methane emissions have not been investigated to date in the seasonally flooded, dense forests of the Amazon floodplains. This region has also been the centre of competing hypotheses to address the discrepancy that currently exists between the top-down and bottom-up measurements. Given that tree emissions have not been included in the bottom-up measurements to date, we reappraised bottom-up measurements by measuring methane emissions from water surfaces, soil surfaces, herbaceous vegetation and 2500 trees across the Amazon region. Our data demonstrate that regionally, tree stems are the dominant means of methane emissions to the atmosphere and tree fluxes are approximately 150 times larger than fluxes reported for southeast Asia. Our models suggest that regionally, inclusion of tree methane emissions in bottom-up measurements has the potential to close the gap between the bottom-up and top-down estimates.

D104ORAL-0380: Implications of precipitation underestimation and ice-wedge degradation on Arctic tundra soil moisture

Anna Liljedahl1, Donatella Zona2, Larry Hinzman1, Craig Tweedie3, Walter Oechel2, Douglas Kane1

1University of Alaska Fairbanks, Fairbanks, The United States of America 2San Diego State University, San Diego, The United States of America 3University of Texas at El Paso, El Paso, The United States of America

The impact of modeled Arctic wetland permafrost thaw on methane fluxes has primarily been forced by measured precipitation and described via increasing active layer depths, i.e. the soil column that experiences seasonal freeze and thaw above permafrost. Field measurements aimed to track increasing active layer depths are challenging as the ground subsides when ice-rich soil thaws. Similarly, difficulties in obtaining accurate precipitation measurements have limited meaningful hydrologic assessment for over a century due to performance challenges of conventional snowfall and rainfall gauges in windy Arctic environments. In agreement with other studies, and not accounting for sublimation, conventional snowfall gauges captured 23 to 56% of end-of-winter snow accumulation. Applying conventional precipitation observations to a water balance analysis produced consistent water storage deficits that are larger than what was observed in an unusually low rainfall summer. Field observations throughout the Arctic present abundant examples of long-term warming of cold permafrost and differential ground subsidence, where the latter is due to melting of the top of ice-wedges. Ice-wedge degradation leads to surface moisture shifting from the wide ice-wedge polygon center to the narrow surrounding troughs. The troughs themselves may also drain if the trough-channels become connected. Incorrect representation of precipitation in field and modeling studies and/or not accounting for ice-wedge degradation in models can therefore have major implications for Arctic water balance studies due to cascading effects on storage and runoff. There is a potential to dramatically refine the representation of methane fluxes in larger scale modeling efforts by including differential thaw and addressing precipitation underestimation.

D105 ORAL-0366: Evaluating Spatiotemporal Differences in Methane Fluxes on the North Slope of Alaska via Eddy Covariance Footprint Modelling

Kassandra Reuss-Schmidt1, Peter Levy2, Cathy Wilson3, Donatella Z Zona1, 4

1University of Sheffield, Sheffield, United Kingdom 2Center for Ecology and Hydrology, Edinburgh, United Kingdom 3Los Alamos National Laboratory, Los Alamos, The United States of America 4San Diego State University, San Diego, The United States of America

Arctic permafrost soils store 1300-1370 Pg of organic carbon, twice the current atmospheric stock. This region is warming at approximately 1°C per decade, and permafrost soils could lose 381-616 Pg C by 2300, with a large portion potentially being released as the potent greenhouse gas methane (CH4). Despite intensive investigation, uncertainty estimates of CH4 emissions have changed little since the first estimates in 1974.  Two main difficulties in creating a baseline flux estimate is the region’s remote nature and the high spatiotemporal variability in methane fluxes. This project examines fluxes from three eddy covariance sites in Barrow, Alaska by applying the Kormann and Meixner (2001) footprint model to investigate the spatio-temporal variability. A LiDAR digital elevation model collected by NGEE Artic at a very fine resolution (0.25m) and WorldView2 data have been used to give quantitative metrics for vegetation and microtopographic differences over these three sites. Preliminary results show significant differences (p-value < 0.05) in CH4 emission patterns in the footprints that could bias flux estimates by 20%.  Furthermore, the pattern of footprint variability shows divergent spatial patterns between summer and winter fluxes. The largest mean summer fluxes were observed in a low lying sedge-dominated drained lake basin (7.74 mg CH4 m-2 day-1) with the less degraded, more polygonal area having an average flux of (5.82 mg CH4 m-2 day-1). In the winter “zero curtain” period, the pattern reversed with higher fluxes coming from the polygonal area (3.58 mg CH4 m-2 day-1) and slightly lower fluxes (3.35 mg CH4 m-2 day-1) observed from the lake basin. This highlights that flux drivers differ by season and that these dynamics should be considered for estimating annual and regional fluxes.

D106 ORAL-0208: Comparison of regional d13CH4 from a land-surface model with atmospheric measurements

Anita Ganesan1, Nic Gedney2, Edward Comyn-Platt3, Garry Hayman3, Rebecca Fisher4

1University of Bristol, Bristol, United Kingdom 2Met Office Hadley Centre, Exeter, United Kingdom 3Centre for Ecology & Hydrology, Wallingford, United Kingdom 4Royal Holloway University of London, Egham, United Kingdom

Natural wetlands are the single largest source of methane (CH4) emissions to the atmosphere but there is a large discrepancy in the magnitude predicted by top-down inverse studies and bottom-up process models (Saunois et al., 2016). Measurements of the main isotopologue of CH4, 13CH4, can provide an additional fingerprint of specific sources and sinks and could help to resolve this mismatch. At present, most top-down inverse studies that assimilate atmospheric d13CH4 measurements use bulk d13CH4 signatures for all wetlands, typically -60‰ (permil). Any errors in this signature will have serious implications for inferred wetland fluxes due to the small variations in atmospheric d13CH4. It has been shown that different wetland ecosystems have different pathways of emissions (e.g. CO2 reduction or acetoclastic fermentation) and this directly impacts both the magnitude of CH4 flux as well as the d13CH4 source signature. Bogs, which occur in acidic environments are more highly depleted in 13CH4 (-110 to -60 ‰) than fens, which are found in more alkaline conditions (-65 to -50 ‰) (E. Hornibrook, 2009). Here, we present the combined effect of (1) CH4 fluxes that have been tuned for different wetland ecosystems and (2) ecosystem-specific d13CH4 signatures on ‘upscaled’ regional d13CH4 values. We investigate the ability of the model to represent different wetland ecosystems by comparing the modelled regional d13CH4­with atmospheric measurements from towers and aircraft. Finally, we investigate the impact of variations in wetland source signature on global atmospheric d13CH4.

D107 ORAL-0175: Role of wetlands and permafrost thaw in modulating emission profiles to stabilise climate at 1.5° or 2.0°C.

Edward Comyn-Platt1, Garry Hayman1, Chris Huntingford1, Sarah Chadburn2, Eleanor Burke3, Anna Harper2, Chris Webber4, Peter Cox2, Bill Collins4, Nic Gedney3

1Centre for Ecology & Hydrology, Wallingford, United Kingdom 2University of Exeter, Exeter, United Kingdom 3Met Office, Exeter, United Kingdom 4University of Reading, Reading, United Kingdom

Methane (CH4) is a greenhouse gas, and changes to its atmospheric concentration will influence global warming and efforts by society to stabilise global warming at 1.5°C or 2.0°C. However under global warming, additional factors influencing CH4 levels are climate change-induced alterations of land-atmosphere methane fluxes, and particularly from wetlands and permafrost. We demonstrate how climate models can be used to quantify permissible CO2 emissions and future carbon capture requirements, given these additional CH4 considerations.
The JULES-IMOGEN framework (Huntingford et al., 2010) provides an intermediate complexity climate modelling system to examine the climate feedbacks of the land surface within the climate sensitivity bounds described by the 34 global circulation models (GCMs) from CMIP5 (climate model inter-comparison, phase 5; Taylor et al., 2012). Traditionally, JULES-IMOGEN simulations have used prescribed emission scenarios to drive future projections and examine the resulting changes in climate. This work uses an inversion of the JULES-IMOGEN framework, i.e. driven with a prescribed future climate trajectory, such that the resultant atmospheric composition, and hence the permitted/required human contribution to it, is the focus of analysis. Results presented here demonstrate emission reductions and uptake pathways required given the JULES vn4.8 release, and the additional reductions and uptake required when natural wetland methane and permafrost thaw feedbacks are also considered.

D108 - ORAL-0316: Tree stem CH4 fluxes suggest that uptake dominates over emission in tropical, temperate and boreal upland forests.

Vincent Gauci1, Bertie Welch1, Sunitha Pangala1, Emma Sayer2, Alex Enrich-Prast3, David Bastviken3

1The Open University, Milton Keynes, United Kingdom 2Lancaster University, Lancaster, United Kingdom 3Linkoping University, Linkoping, Sweden

Forests play an important role in the exchange of radiatively important gases with the atmosphere. Previous studies have shown that in both temperate and tropical wetland forests tree stems are significant sources of methane (CH4), yet little is known about trace greenhouse gas dynamics in ‘upland’ free-draining soils that dominate global forested areas. We examined methane fluxes from both soils and tree stems in lowland tropical forest on free-draining soils in Panama, Central America (Barro Colorado Nature Monument), in the Amazon (Cunia), from a deciduous woodland in the United Kingdom (Wytham, Oxfordshire) and boreal forest in Sweden. Tree stem samples were collected via syringe or via in situ laser analyser measurement using temporary chambers strapped to the trees. Soil fluxes were sampled from installed collars. We found that trees behaved as both sources (near the tree base) and sinks (higher up the tree stem) of methane across Swedish, Panamanian and UK sites, however, this pattern was only apparent in a subset of trees in the Amazon where the dominant process was stem CH4 uptake for the majority of trees. We synthesise these results and report the consequences for ecosystem budgets of methane which show that trees are a consistent net methane sink in tropical and boreal ecosystems whereas temperate forest trees alternate between sources and sinks depending on climatic variables.

D109 -ORAL-0209: Methane flux in the Amazon forest: First steep to understand  its seasonal and spatial variation in forested and deforested areas, as well as upland and wetland areas

Jose Mauro S. Moura1, Joost van Harren2, Jorge Rodrigues3, Scott R Saleska2, Adelaine Michela S Figueira1, Rodrigo da Silva4, Daniel A Jati5, Katrine S Escher4, Tsai S Mui6, Julia B Gontijo6

1Federal University of Western Pará, Santarém, Brazil 2Ecology and Evolutionary Biology Department, University of Arizona, Tucson, The United States of America 3University of California - Davis, Davis, The United States of America 4Federal University of Western Para, Santarém, Brazil 5Federal University of Western Para, Santarem, Brazil 6Centro de Enegia Nuclear na Agricultura, Piracicaba, Brazil

Methane (CH4) is an important but understudied component of the C-cycle, a greenhouse gas 30 times more powerful by mass than CO2. However little is known regarding the impact of land use change and seasonal inundation of wetlands on microbial biodiversity and methane flux, especially in the tropics. It still remains unknown how changes in microbial biodiversity affect ecosystem functions. Our proposed research addresses the intersection of these questions in the context of biodiversity conservation by asking: “how does the interaction between soil microbial and forest tree biodiversity control cycling of methane along gradients of land use and seasonal water inundation in Amazon forests?” To predict the future of methane as a driver of climate change in this system, we will combine novel gas flux measurement instrumentation with cutting edge molecular microbial ecology. We propose to address biodiversity and environmental controls on methane (CH4) production from tropical regions, by measuring CH4 fluxes from a variety of potential sources – surfaces of tree stems and leaves, soil, water in forested and deforested areas, as well as upland and wetland areas. We propose to combine microbial-scale genome-enabled techniques (high throughput functional gene biodiversity measurements and metagenomics) with ecosystem-scale biogeochemical techniques (field and controlled core based flux and isotopic measurements) to quantify the effects of (a) land use change and (b) seasonal variation in inundation levels in tropical wetlands on microbial-mediated CH4 cycling. The site is centred around the City of Santarém, State of Pará, Brazil. Our project goals are to advance biodiversity conservation science in Amazônia by (1) quantifying methane-cycling microbial diversity as a function of land use and seasonal inundation, (2) quantifying interactions between methane-cycling microbes and methane cycling, and (3) incorporating knowledge of interactions between methane-cycling microbes and plants into conservation and management plans for mitigating the climate impact of methane emissions.

D110 ORAL-0350: Methane emissions along a tropical peat dome: Insights on spatial variability and transport mechanisms

Sunitha Pangala1, Alison Hoyt2, Alex Cobb3, Charles Harvey2, Vincent Gauci1

1The Open University, Milton Keynes, United Kingdom 2Massachusetts Institute of Technology, Cambridge, MA, The United States of America 3 Singapore-MIT Alliance for Research and Technology, Singapore, Singapore

Methane emissions from tropical peat swamp forests of southeast Asia are thought to be relatively modest, owing to low soil pH, high methane oxidation rates and recalcitrant carbon impeding rates of methanogenesis. However, there is growing evidence to suggest that inundated trees are an efficient means of methane transport and can emit large quantities of methane in peat swamp forests. Here we demonstrate by measuring methane emissions from soils, tree stems and leaves that tree-mediated methane emissions are the dominant means of methane emissions in one of the last remaining pristine tropical peatlands in Southeast Asia. Soil emissions only equated to less than 30% of the total ecosystem methane flux. Methane emissions, measured within the three plots along a transect (from the edge to the centre of the peat dome), revealed that both tree and soil emissions vary between and within the three plots. Soil emissions decreased from the edge to the centre of the peat dome with increasing peat depth and decreasing water table depths and tree emissions following an opposite trend. Within each plot, tree-mediated methane emissions displayed large variability with fluxes ranging between 0.2 – 14.4 mg m-2 hr-1. Relationships between tree-mediated emissions and pore-water methane concentrations point towards the possibility of some of these trees transporting methane produced in the deeper layers of the peat profile to the atmosphere. Using the models we developed, the efficiency of diffusion and transpiration driven mechanisms to drive methane emissions from tree stems will be discussed. 

D111 - ORAL-0333: Methane emission from the Three Gorges Reservoir and their influencing factors

Xiaoke Wang1, Le Yang1, Fei Lu1

1Research Center for Eco-Environmental Sciences, CAS, Beijing, China

The methane emission from the Three Gorges Reservoir (TGR), currently the largest hydroelectric reservoir in the world, is very interested by scientists and politicians. In this study, methane (CH4) were measured in the water surface and the drawdown area using the static chamber and gas chromatograph method. The results showed that the average annual CH4 flux was 0.62 mg CH4 m-2 h-1, including 0.27± 0.15 mg CH4 m-2 h-1 of diffusion and 0.35 ± 0.57 mg CH4 m-2 h-1 of ebullition, which was similar to CH4 emission from temperate and boreal reservoirs but significantly lower than those from tropical reservoirs. Seasonal variation in CH4 flux showed that CH4 flux reached the maximum in the summer and related with the seasonal variations in water velocity, temperature, and dissolved oxygen. Moreover, the yearly average CH4 flux decreased from the upstream to the downstream before the Three Gorges Dam (TGD), but CH4 emission from the river surface down TGD was higher than that from the water surface before TGD.
In the drawdown areas of TGR, cropland, fallow land and deforested land were sources in the inundated season (0.22 ± 0.26 mg CH4 m-2 h-1) and a sink in the drained season (-0.008 ± 0.035 mg CH4 m-2 h-1). The average CH4 flux across the entire sampled season was 2.61 ± 3.76 mg CH4 m-2 h-1 in the rice paddy, which was the highest among the four types of land that were examined.
The total GHGs emission ranged from 0.7399–1.331 Tg C yr-1 (equivalent CO2) in 2010, which was only 3.0–5.4% of the GHGs that was released by thermo-power plants if the same amount of electricity (84.7 billion kWh) was generated.

D112 -POSTER-0090: Environmental control of methane exchange in a poorly drained black spruce forest over permafrost

Yasunori Igarashi1, Hiroki Iwata2, Masahito Ueyama3, Hirohiko Nagano4, Yoshinobu Harazono3, Tetsuya Hiyama1

1Institute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan 2Department of Environmental Sciences, Faculty of Science, Shinshu University, Matsumoto, Japan 3Graduate School of Life and Environmental Sciences, Osaka Prefecture University, Sakai, Japan 4International Arctic Research Center, University of Alaska Fairbanks, Fairbanks, The United States of America

Methane exchanges between atmospheric and surface vegetation in a poorly drained black spruce (Picea mariana) forest over permafrost was observed over a 6-year period. During summer (from May to October), the thawing depth and soil moisture content increased with seasonal progression. The methane flux also corresponded to the seasonality of thawing depth and soil moisture content, and its maximum value was observed in late summer. On the contrary, the seasonality of air and soil temperature did not correspond to that of methane flux. We revealed that the magnitude of methane flux corresponded to the water level above the thawing depth (thickness of the anaerobic layer), regardless of the season. Our site was flooded for 12 weeks during the observation period. The methane flux did not correspond to temperature and ground temperature but did correspond to the thawing depth (in this case, the thickness of the anaerobic layer), even in flood conditions. We concluded that the temperature dependence of methane flux at this site was small and that the thickness of the anaerobic layer dramatically determined the magnitude of the flux. The results of this study suggest that it is important to consider the influence of soil moisture and thawing depth when modeling methane flux in northern regions, especially over permafrost.

D113 - POSTER-0097: Measurement of methane flux over a larch forest in eastern Siberia

Taro Nakai1, Ayumi Kotani2, Trofim C. Maximov3, Takeshi Ohta2, Tetsuya Hiyama1

1Institute for Space-Earth Environmental Research, Nagoya University, Nagoya, Japan 2Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Japan 3Institute for Biological Problems of Cryolithozone, Siberian Branch of Russian Academy of Sciences, Yakutsk, The Russian Federation

Methane flux measurement by the eddy covariance method has been conducted over a larch forest in eastern Siberia during snow-free seasons since 2015. According to the previous study of methane flux measurements using static chambers in this forest site, the dry forest floor showed uptake of methane, even at sites where the water table was situated a few centimeters below the soil surface. However, the flux data by the eddy covariance method in this study were highly variable, and the average values showed a slightly positive value. This means that methane emission occurred as a whole forest ecosystem. In addition, a clear diurnal variation in methane flux was observed in June 2016, showing maximum methane emission in the daytime. This forest ecosystem includes wetlands (fens) and small lakes, at which a large methane emission was observed in previous studies. Therefore, the methane flux observed by the eddy covariance method might cover the effect of various types of surfaces including wet surface. We should clarify the cause of diurnal variation in methane flux.

D114 - POSTER-0125: Methane emissions dominate warming potential of Alaskan Arctic tundra

Kyle Arndt1, Walter Oechel, Donatella Zona

1San Diego State University, San Diego, The United States of America

The Arctic is warming at twice the global rate. Arctic soils contain twice as much carbon as the global atmosphere giving the Arctic potential for positive feedbacks impacting global change. Fall emissions of carbon dioxide (CO2) and methane (CH4) are increasingly contributing to the yearly carbon budget especially during a the zero curtain, where soils freeze from the top down and bottom up, forming an insulated layer that remains active into the cold season.
Using the eddy-flux covariance technique at five sites on the North Slope of Alaska, we measured CO2 and CH4 flux as well as a suite of meteorological data. Three of the sites are located in Barrow, AK (BEO [71.2810016, -156.6123454], BES [71.280881, -156.596467], & CMDL [71.3225269, -156.6091798]), one 70 km south in Atqasuk (70.4696228,-157.4089471), and one in Ivotuk (68.48649,-155.75022) in the foothills of the Brooks Range. The sites represent a variety of ecotypes and vegetation regimes as well as a latitudinal gradient.
We calculated the yearly budget of CO2 and CH4 across the sites and found that our sites that sequestered the most CO2 (Ivotuk & BES; mean 27.7 and 31.1 g C CO2 m-2 y-1 respectively) were also the larger emitters of CH4 (mean 5.6 and 3.4 g C CH4 m-2 y-1 respectively). During the last 15 years, warming has been delaying the soil freezing and increasing the duration of the zero curtain, which is resulting in an increase in CH4 emissions by 2.136 – 17.280 mg C CH4 day-1. The largest estimated increase was again in the most southernmost site Ivotuk. These results imply that areas across the Arctic could be contributing to positive feedbacks at unexpected rates due to their high methane emissions despite stronger CO2 sequestration.

D115  - POSTER-0415: Evaluation of pan-Arctic ecosystem scale greenhouse gas measurements

Martijn Pallandt1, Jitendra Kumar2, Mathias Gockede3

1Max Planck Institute for Biogeochemistry, Jena, Germany 2Oak Ridge National Laboratory, Oak Ridge, The United States of America 3Biogeochemistry Department, Max Planck Institute for Chemistry, Mainz, Germany

Due to logistics difficulties and winter extreme weather there are practical limitations on where and when measurements can be taken in the Arctic. To address these limitations this study aims to assess the current state of the arctic greenhouse gas monitoring networks, with the aim of identifying possible gaps in data acquisition.
We made an inventory of all arctic (for this purpose north of 60ON) greenhouse gas monitoring sites, with a focus on CO2 and or CH4 and eddy covariance or atmospheric towers. This inventory consists of sites listed on mayor eddy covariance networks like, Fluxnet, AmeriFlux, Neon, ICOS, Asiaflux, etc .and those reported in personal communication with PI’s.  More than 100 sites where identified. Sites activity differs drastically; some have been active for a season while others have been operational for decades.  CO2 was monitored at all these sites whereas CH4 measurements took only place at approximately one third of the sites. 
From these 100 sites a selection of 31 prime sites was made, these sites where chosen based on a combination of data coverage, auxiliary measurements and location.  Based on Multivariate spatiotemporal clustering a map of pan-Arctic ecoregions was produced.  This served as the basis for a similarity metric to assess the representativeness of these prime sites.  Combining the representativeness with data availability yielded maps of spatiotemporal variability in CO2 and CH4. These maps clearly show the gaps in data acquisition both spatially and temporally, with wintertime fluxes and CH4 measurements being clear examples of areas with coverage gaps.
Further research will aim to add more sites and identify realistic locations for new sites to optimize the networks representativeness.