Towards determining spatial methane distribution on Arctic permafrost bluffs with an unmanned aerial system
Arctic permafrost stores vast amounts of methane (CH4) in subsurface reservoirs. Thawing permafrost creates areas for this potent greenhouse gas to be released to the atmosphere. Identifying ‘hot spots’ of methane flux on a local scale has been limited by the spatial scales of traditional ground-based or satellite-based methane-sampling methods. Here we present a reliable and an easily replicable design using only off-the-shelf, cost-effective methane sensor components and an Unmanned Aerial System (UAS). Our results demonstrate the high efficiency of the design and the advantages of this methodology for environmental methane studies that are subjected to the high spatial variability of methane levels. On Barter Island, NE Alaska, we noted spikes in CH4 concentrations coincident with topographic features or anomalies. Such spikes may be attributed to enhanced land/air transfer and may reveal zones of high methane production and/or minimal oxidation in areas of thermoerosional gullies along thawing coastal zones. Thermoerosional gullies represent hotspots that release significantly higher levels of methane than the surrounding areas, thus suggesting that point sampling is inadequate in characterizing methane releases and that increasing rates of permafrost thaw may result in increasing point sources of high CH4 emissions.
KeywordsMethane Permafrost UAS Thermoerosion North Slope Alaska Absorption spectroscopy
The permafrost landscape of the Arctic shoreline is one of the most dramatically changing environments in the world. Despite the fact that Arctic permafrost coastlines represent 34% of Earth’s coast and are affected by permafrost , the physical processes and mechanisms that drive change along these coastlines remain poorly understood [2, 3]. Along the Arctic shoreline, coastal erosion is expected to dramatically increase over the next decades due to the combined effect of longer and warmer thawing seasons, declining sea-ice cover that allows larger waves and storms to collide with the coastline at both longer durations and at higher coastal elevations. With this heightened erosion one can expect that constituent fluxes both into the sea and air will also increase over time.
Erosion rates appear to be increasing along some sections of the Arctic coast [4, 5] reaching up to 25 m/year . This has resulted in an annual efflux of 14 Tg of particulate organic carbon into the coastal ocean . This carbon flux is of equal magnitude as the annual delivery from all Arctic rivers, or the vertical net methane (CH4) emissions from terrestrial permafrost [8, 9].
Spatial identification and quantification of CH4 efflux hotspots may serve a twofold purpose along coastal bluffs. First, quantifying CH4 efflux from hotspots is essential for accurate methane budgets. Secondly, these hotspots could also indicate areas of lower soil stability, thereby identifying coastal sites highly vulnerable to erosion. Identification of methane pathways may therefore help gain better local insights on permafrost integrity and help with erosion prediction models.
Currently, spatial methane distribution in the Arctic is being recorded mostly with top-down methods by several different means—all of which come with major drawbacks: (1) tower station measurements have a sparse spatial coverage and cannot identify individual methane sources; (2) satellite measurements can reveal large scale atmospheric background signals but give little indication on the ground-based sources of methane sources; (3) airplane-based measurements can only be carried out at a minimum height of ~ 100 m and are not as accurate as in situ data. All of the above methodologies are associated with significant costs unattainable for most institutions and stakeholders (e.g. ).
Until recently, methane measurements with systems sensitive to parts-per-billion levels were also confined to large, heavy and expensive laboratory equipment; now, a new class of significantly more cost-effective and lighter tunable diode-laser using absorption spectroscopy  allows for the measurement of methane using small Unmanned Aerial Systems (UAS). While few of these methane-detection systems have been developed (i.e. [19, 20]) none have aimed to create an off-the-shelf, cost effective UAS design. An overview of the challenges and limitations of using small UAS systems to monitor methane and other atmospheric gases has been compiled by Villa et al. . While this manuscript focuses on the application of an unmanned methane sensor in a very remote, largely undisturbed location north of the Arctic circle, where permafrost is thawing, the proposed system could also be utilized for more urban environments and for pipeline leak detection. Emran et al.  for example, have recently demonstrated the feasibility and advantages of UAS-based methane monitoring for grass-topped landfill degassing, and Bretschneider et al.  showed how such a system could detect leaks in pipelines. The methane laser sensor used herein was first developed in 1992  and became commercially available in 2013 . The sensor has been shown to function in a natural environmental setting to demonstrate time variation of low CH4 concentrations over rice paddies  or small ponds and ditches . We here present the design and components of a CH4—UAS system.
In this paper, we demonstrate our efforts to detect methane being emitted from thawing coastal permafrost and map these concentrations. Our study was carried out along the coastal bluff environment of Barter Island, AK, which is part of a larger Arctic coastal erosion investigation study by the U.S. Geological Survey (USGS; i.e. [28, 29, 30]).
2.1 The UVA-platform
The vibration dampening methane sensor mount was 3D-printed (Fig. 2) and attached to the central Accessory Bay, the area behind the gimbal under the 3DR Solo. It is intended for secondary accessories, including additional communications hardware and other high power devices . Total cost for UAS and 1 month sensor rental was < $2500.
2.2 The methane sensor
Methane sensor technical information
ppm × m
Wavelength 1653 nm
− 17 °C to 50 °C
Methane concentration measurements were updated at a rate of 10 Hz. In addition to the internal calibration that the instrument undergoes during each reboot, a calibration experiment was carried out using quartz glass calibration gas cells at varying distances and concentrations. While the manufacturer lists the CH4 levels with a ± 10% accuracy our calibration experiment suggested an accuracy of ± 30–40%. Calibration curves are shown in the “Appendix”. All data were GPS referenced and stored via a standard Bluetooth link on a UAV mounted Android OS device (Fig. 2).
2.3 Flight plans and post processing
Automatic flight paths were created using Mission Planner  and executed to fly at a continuous speed of 3 m/s. A high precision Lidar-based Digital Elevation Model (DEM)  was used to guide the methane sensor at a constant distance of 5 m above ground. The normalization of the data to 1 m distance occurred by simply dividing the set flight elevation by five. The flight elevation was chosen in part because the sensor is most precise at distances of 5–10 m (see “Appendix”) and because it was the minimum height at which the drone could be safely operated above potential human activity. However, there was no recorded human activity nor engineering construction that could have affected the measurements within the study site. The flight paths were created to be above the flat vegetation tundra landscape, slightly inland (~ 5 m) of the coastal bluff (Fig. 5). All flights were carried out during zero to low (< 3 m/s) wind conditions and flown against the prevailing western wind direction. All human infrastructure that could potentially act as an anthropogenic methane source was downwind from our study site. This strategy was chosen in order to identify individual hotspots along the flight path rather than following a plume. Air temperatures during all flights were at a constant 2 °C and between 35 and 45% humidity as determined by the nearby (located 650 m away) airport weather station. The methane sensor automatically marks data that was recorded with a below minimum reflected energy as containing an error such as over bodies of water. This data was filtered from the utilized dataset following a previously published MATLAB (R2007a) methodology described by Emran et al. . Visualization of the methane data was carried out using Google Earth software but can be performed using any GIS software program.
3 Results and discussion
Our first results from Barter Island, AK show that spatial methane distribution along the 2 km coastal path is mostly at atmospheric background levels of 2 ppm ± 35%. There are however several large methane peaks within the flown path. These peaks range between 20 and 210 ppm and are spread out along the path. After closer analysis of the surface topography it became clear that the peaks were linked with areas of high permafrost thaw and thermoerosional gullies (Fig. 5).
High methane fluxes from thawing permafrost are often associated with saturated systems such as thermokarst bogs, fens, and lakes or low centered polygonal tundra . Much less research has been focused on thermoerosional gullies because they may not typically be saturated and they may represent a small fraction of the landscape. However, recent studies  have shown high CH4 fluxes from gullies such as ours, even if soil moisture is not changed relative to the undegraded landscape. The mechanisms driving these fluxes and the contribution of these fluxes to the overall coastal methane budget, are further issues to be resolved.
The main uncertainties associated with this study are caused by potential variations in the ground to sensor distance. These variations are most likely associated with non-linear movements of the UAS such as roll, tilt and yaw that are produced by flight turns and variable wind speeds and directions. While the data on off-nadir angle could not be recovered due to an overwrite, common deviations from Nadir for this UAS-class are < 15° [39, 40]. Assuming a 15° deviation from Nadir, this would lead to a maximum 3.5% increase in measured methane levels, which is significantly below the published accuracy (± 10%) or our measured accuracy (± 35%) of the instrument (Table 1, “Appendix”). Nonetheless, in order to account for this effect we included many turns in our test grid at constant flight speed (Fig. 4a) forcing yaw, tilt and roll and thus establishing a conservative measure of instrument background noise. Conversely, in the permafrost bluff survey we attempt to minimize this effect by avoiding all turns in our flight path (Fig. 5). Unfortunately, wind, fog and snow prevented additional flights during the time-limited field research campaign, and locations north of the Arctic Circle such as Barter Island are not easily accessed to allow for low-cost follow up studies. Nonetheless, with respect to future deployments, we hope to employ traditional ground-based point-sampling techniques to better quantify the relative limitation of each methodology in more detail. Additional uncertainties in our measurements may be the result of the under or over represented gullies in the DEM used in this study. Gullies ranged in depth by 1–5 m and an additional conservative ± 20% measurement error can be estimated from comparisons between field elevation observations of gullies and DEM results. While this leads to a total uncertainty of ± 55%, which is mirrored by our field observed uncertainty of ± 56%, the reported peaks found in gullies are more than 2000% above background levels (Fig. 5) and thus remain significant with these uncertainties applied.
Advantages and disadvantages of major techniques measuring methane emissions
Satellite-based absorption spectroscopy using reflected sunlight or thermal emissions
Global, complete spatial coverage; long or short timescales
Coarse resolution; limited to cloud and snow free scenarios; entire atmospheric column is processed; very high cost
Aircraft-based absorption spectroscopy or infrared spectrometry
Targets specific emission source regions and obtain vertical profiles of methane concentrations; high-time-resolution
Limited to a temporal snapshot; challenging to account for changing ground-winds; high cost
Tower-based infrared spectrometry analyzed by eddy covariance or by inverse modeling
Very precise across the towers sensitivity footprint; repeatable and consistent measurements; long time series
Individual methane sources can be difficult to distinguish; permanent singular position; high cost
Chamber-based techniques analyze changes in CH4 concentrations in chamber volume or changes in inlet/outlet concentrations
Can accurately quantify diffusive and point-source emission rates from a small source area; can quantify “negative” methane soil oxidation rates; independent of wind conditions
Limited to a very small area; labor intensive to measure the temporal variability of emissions; usually requires large numbers of chambers
UAV-based absorption spectroscopy
High temporal and spatial resolution. Can target specific emission sources. Very low cost
Limited to a temporal snapshot. Limited to very low wind conditions. Daylight operation only. Battery limited spatial range
The herein presented technology could also be used to look for areas of geological methane release and CH4 release from methane hydrates. However, in those cases, isotopic methane composition will help in answering questions of geologic methane origin. Additionally, planned future studies focusing on the onset of the thawing season are likely to answer the question if the herein presented UAS methane-detection methodology is capable at predicting erosion hotspots through early identification of methane hotspots.
Spatial methane distribution data that can be used to identify source pathways is currently still very rare and traditionally its monitoring has been reserved to large institutions capable of absorbing the associated high costs. Compared to previous efforts that collected spatial methane data in the Arctic by helicopter [41, 42], by small aircraft , by ground based laser scanning , or with a ground based solid-state trace gas sensor  the methodology presented herein exemplifies a significantly more affordable (< $2500) and more flexible method towards acquiring spatial methane distributions.
With the rapid commercial spread of UAVs the presented UAS system has the potential to accelerate the availability of high resolution spatial methane data through university and citizens science and thus gain significantly better insights into methane pathways and its related processes.
Funding for this research was provided by the U.S. Geological Survey’s Coastal and Marine Geology Program and the U.S. Geological Survey’s Mendenhall Program. We thank the U.S. Fish and Wildlife Services for their local support. We also thank C. Johnson and T. Lorenson for assistance with instrument deployment and data collection. The IAEA is grateful for the support provided to its Environment Laboratories by the Government of the Principality of Monaco. A special thanks goes to the community of Kaktovik for their continued support for scientific research. We do not intend to claim that the methane sensor’s manufacturer accuracy statement or any other statements by the manufacturer are incorrect as differences in laboratory setups and methodologies can lead to different results. Data on which this paper is based are availabe for download at https://doi.pangaea.de/10.1594/PANGAEA.898636. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
- 1.Lantuit H, Overduin PP, Couture N, Wetterich S, Aré F, Atkinson D, Brown J, Cherkashov G, Drozdov D, Forbes DL, Graves-Gaylord A, Grigoriev M, Hubberten H-W, Jordan J, Jorgenson T, Ødegård RS, Ogorodov S, Pollard WH, Rachold V, Sedenko S, Solomon S, Steenhuisen F, Streletskaya I, Vasiliev A (2012) The Arctic coastal dynamics database: a new classification scheme and statistics on Arctic permafrost coastlines. Estuaries Coasts 35:383–400. https://doi.org/10.1007/s12237-010-9362-6 CrossRefGoogle Scholar
- 7.Wegner C, Bennett KE, de Vernal A, Forwick M, Fritz M, Heikkilä M, Łącka M, Lantuit H, Laska M, Moskalik M, O’Regan M, Pawłowska J, Promińska A, Rachold V, Vonk JE, Werner K (2015) Variability in transport of terrigenous material on the shelves and the deep Arctic Ocean during the Holocene. Polar Res 34:24964. https://doi.org/10.3402/polar.v34.24964 CrossRefGoogle Scholar
- 8.Schuur EAG, McGuire AD, Schädel C, Grosse G, Harden JW, Hayes DJ, Hugelius G, Koven CD, Kuhry P, Lawrence DM, Natali SM, Olefeldt D, Romanovsky VE, Schaefer K, Turetsky MR, Treat CC, Vonk JE (2015) Climate change and the permafrost carbon feedback. Nature 520:171–179. https://doi.org/10.1038/nature14338 CrossRefGoogle Scholar
- 12.Turetsky MR, Kotowska A, Bubier J, Dise NB, Crill P, Hornibrook ERC, Minkkinen K, Moore TR, Myers-Smith IH, Nykänen H, Olefeldt D, Rinne J, Saarnio S, Shurpali N, Tuittila E-S, Waddington JM, White JR, Wickland KP, Wilmking M (2014) A synthesis of methane emissions from 71 northern, temperate, and subtropical wetlands. Glob Change Biol 20:2183–2197. https://doi.org/10.1111/gcb.12580 CrossRefGoogle Scholar
- 14.Walter Anthony K, Schneider von Deimling T, Nitze I, Frolking S, Emond A, Daanen R, Anthony P, Lindgren P, Jones B, Grosse G (2018) 21st-century modeled permafrost carbon emissions accelerated by abrupt thaw beneath lakes. Nat Commun 9:3262. https://doi.org/10.1038/s41467-018-05738-9 CrossRefGoogle Scholar
- 17.National Academies of Sciences Engineering and Medicine (2018) Improving characterization of anthropogenic methane emissions in the United States. The National Academies Press, Washington, D.C.; ISBN 978-0-309-47050-6Google Scholar
- 19.Smith BJ, John G, Christensen LE, Chen Y (2017) Fugitive methane leak detection using sUAS and miniature laser spectrometer payload: system, application and groundtruthing tests. In: 2017 International conference on unmanned aircraft systems (ICUAS); IEEE, pp 369–374Google Scholar
- 23.Bretschneider TR, Shetti K (2014) UAV-based gas pipeline leak detection. In: Asian conference on remote sensing at Nay Pyi Taw, Myanmar, pp 1–6Google Scholar
- 25.Pergam and Tokyo Gas Engineering Co. Operation manual for laser methane mini-G. https://www.crowcon.com/product/download/640/SA3C50A+Manual.pdf
- 28.Gibbs AE, Nolan M, Richmond BM, Kinsman N, Richmond BM (2015) Evaluating changes to arctic coastal bluffs using repeat aerial photography and Structure-from-motion elevation models. In: The proceedings of the coastal sediments 2015; WORLD SCIENTIFIC, pp 1–14Google Scholar
- 30.Erikson LH, Storlazzi CD, Jensen RE (2011) Wave climate and trends along the Eastern Chukchi Arctic Alaska Coast. In: Solutions to coastal disasters 2011; American Society of Civil Engineers: Reston, VA, pp 273–285Google Scholar
- 31.MRobotics Technical Information for GPS chipset mRo u-Blox Neo-M8 N. https://store.mrobotics.io/product-p/gps001-mr.htm. Accessed 6 Jan 2018
- 32.3D Robotics 4.2. The Accessory Bay. https://dev.3dr.com/hardware-accessorybay.html. Accessed 26 Aug 2018
- 34.Fukuchi T (2012) Gas sensing using laser absorption spectroscopy. In: Fukuchi T, Shiina T (eds) Industrial applications of laser remote sensing. Bentham Science Publishers: Chiba University, Japan, 2012; pp 37–59. ISBN: 9781608053407Google Scholar
- 35.Liebetrau J, Reinelt T, Agostini A, Linke B (2017) Methane emissions from biogas plants greenhouse gas balance of electricity produced. In: Murphy JD (ed) IEA Bioenergy; ISBN: 9781910154359Google Scholar
- 36.ArduPilot Mission Planner HomeGoogle Scholar
- 37.Dlugokencky EJ, Lang PM, Masarie KA, Crotwell AM, Crotwell MJ. Database for atmospheric methane dry air mole fractions from quasi-continuous measurements at Barrow, Alaska and Mauna Loa, Hawaii. ftp://aftp.cmdl.noaa.gov/data/trace_gases/ch4/insitu/surface/. Accessed 5 Aug 2018
- 38.Gibbs AE, Richmond BM, Erikson L, Jones BM (2018) Long-term retreat of coastal permafrost bluffs, Barter Island, Alaska. In: 5th European Conference on permafrost; EUCOP5: Chamonix, France, vol 1, pp 1–2Google Scholar
- 41.Dzikowski M, Klyashitsky A, Jaeger W, Tulip J (2009) Open path spectroscopy of methane using a battery operated vertical cavity surface-emitting laser system. In: Vallée R (ed), p 73861HGoogle Scholar
- 42.Fix A, Ehret G, Hoffstädt A, Klingenberg HH, Lemmerz C, Mahnke P, Ulbricht M, Wirth M, Wittig R, Zirnig W (2004) Charm—a helicopter-borne lidar system for pipeline monitoring. Eur Sp Agency Spec Publ ESA SP 1:45–48Google Scholar
- 45.Rothman LS, Gordon IE, Babikov Y, Barbe A, Chris Benner D, Bernath PF, Birk M, Bizzocchi L, Boudon V, Brown LR, Campargue A, Chance K, Cohen EA, Coudert LH, Devi VM, Drouin BJ, Fayt A, Flaud J-M, Gamache RR, Harrison JJ, Hartmann J-M, Hill C, Hodges JT, Jacquemart D, Jolly A, Lamouroux J, Le Roy RJ, Li G, Long DA, Lyulin OM, Mackie CJ, Massie ST, Mikhailenko S, Müller HSP, Naumenko OV, Nikitin AV, Orphal J, Perevalov VI, Perrin A, Polovtseva ER, Richard C, Smith MAH, Starikova E, Sung K, Tashkun SA, Tennyson J, Toon GC, Tyuterev VG, Wagner G (2013) The HITRAN2012 molecular spectroscopic database. J Quant Spectrosc Radiat Transf 130:4–50. https://doi.org/10.1016/j.jqsrt.2013.07.002 CrossRefGoogle Scholar
- 46.Knights Optical Material specifications for fused quartz JGS3 with 185–3500 nm transmission. https://www.knightoptical.com/_public/documents/1382532741_irmaterialirfusedquartzjgs3opmi-jgs3.pdf. Accessed 6 Jan 2018