Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Integrating local environmental observations and remote sensing to better understand the life cycle of a thermokarst lake in Arctic Alaska

Integrating local environmental observations and remote sensing to better understand the life... ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 2023, VOL. 55, NO. 1, 2195518 https://doi.org/10.1080/15230430.2023.2195518 Integrating local environmental observations and remote sensing to better understand the life cycle of a thermokarst lake in Arctic Alaska a b b c c c Benjamin M. Jones , Susan Schaeffer Tessier , Tim Tessier , Michael Brubaker , Mike Brook , Jackie Schaeffer , a d d d e,f Melissa K. Ward Jones , Guido Grosse , Ingmar Nitze , Tabea Rettelbach , Sebastian Zavoico , e e Jason A. Clark , and Ken D. Tape a b Institute of Northern Engineering, University of Alaska Fairbanks, Fairbanks, Alaska, USA; Local Environmental Observer, Kotzebue, Alaska, c d USA; Community Environment and Health, Alaska Native Tribal Health Consortium, Anchorage, Alaska, USA; Helmholtz Centre for Polar and Marine Research, Alfred Wegener Institute, Potsdam, Germany; Geophysical Institute, University of Alaska Fairbanks, Fairbanks, Alaska, USA; Alaska Cooperative Fish and Wildlife Research Unit, University of Alaska Fairbanks, Fairbanks, Alaska, USA ABSTRACT ARTICLE HISTORY Received 16 November 2022 On 29 June 2022, local observers reported the drainage of a 0.5 ha lake near Qikiqtaġruk (Kotzebue), Revised 26 February 2023 Alaska, that prompted this collaborative study on the life cycle of a thermokarst lake in the Arctic. Prior Accepted 22 March 2023 to its drainage, the lake expanded from 0.13 ha in 1951 to 0.54 ha in 2021 at lateral rates that ranged from 0.25 to 0.35 m/year. During the drainage event, we estimate that 18,500 m of water drained KEYWORDS from the lake into Kotzebue Sound, forming a 125-m-long thermo-erosional gully that incised 2 to Arctic; lakes; lake drainage; 14 m in ice-rich permafrost. Between 29 June and 18 August 2022, the drainage gully expanded from local observations; 1 m to >10 m wide, mobilizing ~8,500 m of material through erosion and thaw. By reconstructing permafrost; thermokarst a pre-lake disturbance terrain model, we show that thaw subsidence occurs rapidly (0.78 m/year) upon transition from tundra to lake but that over a seventy-year period it slows to 0.12 m/year. The combination of multiple remote sensing tools and local environmental observations provided a rich data set that allowed us to assess rates of lake expansion relative to rates of sub-lake permafrost thaw subsidence as well as hypothesizing about the potential role of beavers in arctic lake drainage. Introduction (Turetsky et al. 2020; B. M. Jones et al. 2022). Better The formation and drainage of lakes in the Arctic repre- understanding lake drainage processes and document- sent dominant landscape change processes in perma- ing lake drainage events is of importance to climate change research, wildlife and habitat studies, access to frost regions (Grosse, Jones, and Arp 2013; Brosius et al. 2021; B. M. Jones et al. 2022). Their dynamics subsistence resources, and the well-being of northern during the Holocene and Anthropocene have shaped socioecological systems (B. M. Jones et al. 2022). L-DLB systems in the northwestern Alaska Arctic the current state of lowland arctic landscapes and lake and drained lake basin (L-DLB) systems (Farquharson region are continuing to shift to a landscape that is et al. 2016; Bouchard et al. 2020; S. Wolfe et al. 2020; increasingly dominated by DLBs. B. M. Jones et al. B. M. Jones et al. 2022). Taken together, L-DLB systems (2011) found that between 1951 and 2007 on the Cape Espenberg Lowlands of the northern Seward Peninsula, may cover up to 80 percent of the landscape in lowland permafrost regions in the Arctic (Hinkel et al. 2005; the land area gained through lake drainage was nearly 4 M. C. Jones et al. 2012; Grosse, Jones, and Arp 2013; times greater than land area lost through lake formation and expansion. Swanson (2019) conducted lake change B. M. Jones and Arp 2015; Bergstedt et al. 2021). Several recent studies have documented an increase in lake studies for the U.S. National Park Service lands in the evolution and, in particular, lake drainage across the region showing widespread lake losses (mean regional rate loss of 16 ha/year) between 2000 and 2017. Nitze permafrost region (Nitze et al. 2018, 2020; Lara, Chen, and Jones 2021; Webb et al. 2022) with implica- et al. (2020) further documented widespread lake area tions for local-, regional-, and global-scale feedbacks loss for the northern Seward Peninsula and Baldwin CONTACT Benjamin M. Jones bmjones3@alaska.edu Institute of Northern Engineering, University of Alaska Fairbanks, 1764 Tanana Loop Road, Fairbanks, AK 99775 © 2023 The Author(s). Published with license by Taylor & Francis Group, LLC. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. 2 B. M. JONES ET AL. Peninsula in 2017 and 2018, where 192 lakes were iden- had caused the drainage of some thermokarst lakes. Our tified as draining, which exceeded the long-term average recent observations, both with remote sensing and dur- drainage rate by a factor of approximately ten. Lara, ing field studies, in part prompted this study to investi- Chen, and Jones (2021) assessed lake drainage dynamics gate the complex processes triggering lake drainage. across the entirety of the Brooks Range and northwes- On 29 June 2022, between 08:00 and 20:00, a 0.5 ha tern Arctic Alaska between 1975 and 2019, showing that thermokarst lake, with bank heights that ranged from 2 100 to 250 lakes per year had drained between 2015 and to 4 m high, drained on the Schaeffer’s Native allotment 2019 and that by 2050 lake area will likely decline by an on the Baldwin Peninsula, near Qikiqtaġruk (Kotzebue), additional 15 to 21 percent. The factors responsible for Alaska (Figure 1). The event was reported on the Local the rapid nature of lake drainage in the region have been Environmental Observer (LEO) Network by Susan and attributed to increases in mean annual air temperature Tim Tessier, who have their home on the allotment near and precipitation, lengthening of the thaw season, and the lake (Tessier et al. 2022). Over the course of the next increasing near-surface permafrost ground temperature few months, a cross-disciplinary group of scientists and that destabilize ice-rich permafrost soils (Swanson 2019; local knowledge experts from a variety of countries and Nitze et al. 2020; Lara, Chen, and Jones 2021). However, institutions explored the lake basin using aerial and an increase in the presence of beaver engineering in the satellite imagery as well as on-site surveys and discus- region has also been well documented through the for- sions with local observers. The presence of a beaver in mation of more than 12,000 new beaver ponds since the the lake just days before the drainage event and beaver- early 1950s (Tape et al. 2018, 2022). The recent and chewed shrubs on the shore indicate that it may have rapid expansion of beavers in the region adds a level of played a role in the lake drainage event. To better under- complexity when trying to infer the factors responsible stand the factors leading up to this lake drainage event, for lake dynamics. For example, B. M. Jones, Tape, et al. we investigated various forms of imagery to reconstruct (2020) found that beavers were the primary driver of lake water volume and topographic change and estimate surface water area increases on the Baldwin Peninsula permafrost thaw subsidence over time. More specifi - between 2002 and 2019 but their engineering activities cally, we used historical aerial photography and Figure 1. Photos from the LEO post on the sudden lake drainage event that happened on 29 June 2022 (Tessier et al. 2022). (a) Oblique aerial photo of the drained lake basin on 10 August 2022 (photo credit: Sebastian Zavoico). (b) Ground photo of the recently exposed drained lake basin floor and (c) the thermo-erosional drainage gully that formed during the drainage event (photo credits: Susan Tessier, 29 June 2022). (d) The thermo-erosional drainage gully fifty-four days after the lake drained showing continued permafrost thaw and ground ice melt since the drainage event (photo credit: Benjamin Jones). ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 3 contemporary very high-resolution airborne and satel- are signals of environmental change. In 2012, the LEO lite imagery to reconstruct lake evolution prior to drai- Network was launched as an online tool to help the tribal nage as well as corroborating the timing of the lake health system and local observers to share information drainage event. Structure-from-motion (SfM) techni- about climate and other drivers of environmental change. ques were used to process digital photography acquired The event was reported on the LEO Network by Tim and from a plane mounted camera (2016 and 2021) and Susan Tessier, who have their home on a Native allotment a drone system camera (2022) to provide an estimate near the lake, which is located about 5 km northeast of of lake water volume loss and topographic change asso- Kotzebue, Alaska (Figure 2; Tessier et al. 2022). Several ciated with the lake drainage. Simulating a pre-lake researchers were then approached by the LEO Network to formation terrain surface allowed for the estimate of verify and comment on this reported event, which led to permafrost thaw subsidence associated with thermo- the scientific analysis reported here in close collaboration karst lake evolution prior to its drainage in 2022. between local observers and researchers. Reconstructing lake area changes over time Methods We reconstructed the environmental history and drai- Local environmental observations nage of Schaeffer Lake by analyzing several remote sen- The observation of the sudden lake drainage event was sing datasets in a geographic information system (Jones provided through the Alaska Native Tribal Health 2023). Historical aerial photography from 14 July 1951 Consortium’s LEO Network. The LEO Network is (1:20,000 scale), 6 July 1974 (1:6,000 scale), and a community of local and topic experts who share knowl- 31 August 1993 (1:48,000 scale) provided a historical edge about the time and location of specific events that data set to map changes in the lake area over time due Figure 2. The northwestern Arctic region of Alaska where the thermokarst lake drained on 29 June 2022. (a) An inset map showing the location of the northern part of the Baldwin Peninsula (red box). (b) A Landsat satellite image showing the northern Baldwin Peninsula and the location of the drained thermokarst lake (yellow box). (c) A Worldview-2 satellite image from 20 July 2017 showing a series of small thermokarst lakes that have formed in ice-rich permafrost. Schaeffer Lake is the lake that drained on 29 June 2022. ©2023 Maxar. 4 B. M. JONES ET AL. Table 1. Historic aerial photograph and contemporary satellite imagery used in our analysis of Schaeffer Lake evolution between 1951 and 2021. Image Ground control points Spatial resolution Georegistration root mean square error Lake area Image date Image source scale (n) (m) (m) (ha) 14 July 51 Aerial photography 1:20,000 10 1 0.95 0.13 6 July 74 Aerial photography 1:6,000 10 1 0.80 0.24 31 August 93 Aerial photography 1:48,000 10 1 0.47 0.36 4 July 07 Quickbird-2 N/A 10 1 0.71 0.44 4 August 21 Worldview-2 N/A N/A 1 N/A 0.54 Figure 3. Very high-resolution satellite image time series bracketing the drainage of Schaeffer Lake to the time period observed by local environmental observers. The lake drained between 24 June 2022 (note smoke is present) and 30 June 2022 as observed in the 0.5 m resolution satellite images. The observation from 4 August 2021 was the last non-smoke and snow/ice-affected image available for the site prior to the lake drainage. The image acquired on 1 August 2022 demonstrates that the lake drainage was complete. ©2023 Maxar. to lake expansion into ice-rich permafrost. The historical 2007 versus 2021) were measured using Digital Shoreline aerial photograph time series was complemented by very Analysis System (DSAS) Version 5.1 (Himmelstoss et al. high-resolution satellite imagery from 4 July 2007 2021) extension installed in ArcGIS Desktop 10.8.1 (ESRI (Quickbird-2) and 4 August 2021 (Worldview-2) to 2022). This method had previously been used to success- map lake area in the two decades prior to its drainage fully quantify thermokarst lake expansion rates (Table 1). The Worldview-2 image from 2021 was used as (B. M. Jones et al. 2011). Transects were cast 3 m apart the base image to georegister the other four images in the using the baseline polyline drawn on the outside of the time series. The georegistration error was less than 1.0 m 2021 shoreline for a total of 110 transects. Transects were based on using ten ground control points in each image initially cast using automated features in DSAS and edited (Table 1). Visual analysis of very high-resolution satellite to ensure that all transects crossed each lake shoreline in images from 24 June 2022 (Worldview-3), 30 June 2022 a perpendicular direction. The same transects were used for (Worldview-2), and 1 August 2022 (Worldview-2) each DSAS run for each time period. The net expansion bracketed the drainage event in 2022 and demonstrated (m) and mean expansion rate (m/year) were calculated for its complete drainage (Figure 3). each time period using the 110 perpendicular transect lines. Quantifying thermokarst lake margin expansion Developing high-resolution digital surface and rates terrain models Lake expansion rates between each time period (1951 Predrainage digital images were acquired on versus 1974, 1974 versus 1993, 1993 versus 2007, and 1 September 2016 from a small single-engine airplane ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 5 with a Nikon D800E DSLR camera with a 24 mm lens varied from 7 to 8 m/s. The along-track overlap and mounted vertically and connected to a Trimble 5700 across-track overlap of the mission were set at 80 per- survey-grade Global Positioning System (Fairbanks cent and 70 percent, respectively. All images were Fodar 2019). Photo exposure was continuously moni- processed in the software Pix4D Mapper (v4.7.5) to tored and adjusted in flight as needed by the pilot. The produce an orthophoto mosaic and a DSM at spatial bulk of the missions were planned with a 17-cm ground resolutions of 2 and 5 cm, respectively and a point sample distance using a two-pass method along the cloud with ~400 points per square meter. A Leica coast. The coast was followed by eye and then subse- Viva differential Global Positioning System provided quent lines offset the appropriate distance to get ground control for the mission, and the data were a 60 percent across-track overlap. Along-track overlap postprocessed to WGS84 UTM Zone 3 North in was generally 80 percent or more. The photos were ellipsoid heights. processed using the software Agisoft Photoscan We developed predrainage (1 September 2016 and (Nolan, Larsen, and Sturm 2015) to create an ortho- 3 July 2021) and postdrainage (22 August 2022) digital photo mosaic, a digital surface model (DSM), and terrain models (DTMs) using the AGL Analyst function a point cloud with twelve points per square meter in the software Quick Terrain Modeler (QTM v8.31; using North American Datum (NAD)83 UTM Zone 3 QTM 2022). We reprocessed all point cloud data sets to North in orthometric heights (North American Vertical a spatial resolution of 30 cm by calculating a bare earth Datum of 1988 (NAVD)88). DTM using the LidarSense algorithm in QTM. The 2016 Predrainage MACS (Modular Aerial Camera System) and 2021 data sets were reprojected to WGS84 UTM aerial images with red–green–blue and near-infrared sen- Zone 3 North in ellipsoid heights to match the reference frame of the 2022 data prior to differencing the DTM sors were acquired on 3 July 2021 during the Perma-X products to estimate the volume of water lost through 2021 flight campaign. Images were captured on a single lake drainage as well as thermal erosion caused by the lake flight line with an 80 percent along-track overlap and drainage event. Prior to differencing the DTMs, Schaeffer a flight altitude of 1,000 m, which translates to an approx- Lake was hydro-flattened to an ellipsoid height of 31 m imate ground resolution of 10 cm. We processed these in QTM. data to an orthophoto mosaic, DSM, and colorized point cloud in Pix4D Mapper (v4.7.5; PIX4D 2022). The single- track acquisition pattern of the MACS data required the Simulating a predisturbance land surface manual selection of ground control points using the post- drainage drone system orthophoto mosaic and DSM to We simulated a pre-lake formation land surface by ensure spatial accuracy across data sets. extracting topographic information from a 0.5 ha area Postdrainage digital images were acquired from adjacent to Schaeffer Lake in the 2016 DTM. The area a DJI P4RTK quadcopter with a DJI D-RTK where the elevation values were extracted represented 2 Mobile Base Station at the Schaeffer Lake drainage the same inferred topography as the site where study site on 22 August 2022. The drone system was Schaeffer Lake had developed. These inferences were flown at 50 m above ground level, and flight speeds based on information in historical U.S. Geological Figure 4. Assessing the accuracy of differencing the three digital terrain models (DTMs) used in our assessments of elevation change. The difference in elevation values was extracted from a 30 m x 30 m area that appeared to have remained stable and undisturbed between 2016 and 2022. Left, 2022 DTM versus 2016 DTM; middle, simulated DTM versus 2016 DTM; right, simulated DTM versus 2022 DTM. 6 B. M. JONES ET AL. Survey topographic maps as well as features visible in allotment near the lake, which is located 5 km northeast of Kotzebue, Alaska (Tessier et al. 2022): the historical aerial photography. The hydro-flattened surface of Schaeffer Lake in the 2016 DTM was replaced The event occurred on June 29th, on our native allot- with the extracted elevation values from the adjacent ment (Illivak) near Kotzebue. We left home in the upland land surface. We downsampled the merged morning and when we came back around 8:00 PM in the evening the whole lake had drained! It looked like it product using an aggregation factor of seventy-five to was blown up with dynamite. (Susan Schaeffer Tessier, smooth out the terrain surface. We then resampled the 29 June 2022) simulated DTM to a spatial resolution of 30 cm for differencing with the 2022 DTM to estimate the total The Tessiers observed that the drainage gully was 1 m thaw subsidence associated with the evolution of the wide on the north end of the lake and that it eroded lake prior to its drainage. We performed an accuracy downwards at least 6 to 7 m into ice-rich permafrost assessment of all three DTMs (2016, 2022, and simu- during the drainage event. Very high-resolution satellite lated) by differencing each product inside what images acquired on 24 June 2022 and 30 June 2022 also appeared to be an undisturbed 30 m × 30 m area detected lake drainage in this same time span (Figure 3). The Tessiers also noted the pattern of degraded ice between 2016 and 2022 (Figure 4). wedges and intervening ice wedge polygon centers evi- dent in the drained thermokarst lake basin bottom as Results well as accumulations of dead aquatic invertebrates that were providing forage for ducks. The week before the Local environmental observations of the sudden sudden lake drainage, Susan saw a beaver in the lake for thermokarst lake drainage the first time (Figure 5a). Following the sudden lake On 29 June 2022, a sudden lake drainage event was drainage, Tim also observed signs of beaver-chewed reported on the LEO Network by Susan Schaeffer wood in the willow shrubs adjacent to where the lake Tessier and Tim Tessier, who have their home on the drainage occurred (Figure 5c,d). Figure 5. Evidence for beaver burrowing as a thermokarst lake drainage mechanism (red circles and ovals). (a) A beaver showed up in the lake a week before the lake drained for the first time in at least four years (photo credit: Susan Schaeffer Tessier). (b) An image acquired by the drone system during a flight down the drainage gully showing beaver-chewed wood (photo credit: Benjamin Jones). (c) and (d) Beaver chew on the bush near the location where the drainage gully formed (photo credit: Tim Tessier). ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 7 Decadal-scale thermokarst lake evolution prior to Differencing digital terrain models—Lake drainage drainage and gully erosion Historic aerial photographs and contemporary very We differenced the pre- and post-drainage DTM pro- high-resolution satellite image time series show ducts from 2016 and 2022 to assess changes in elevation a gradually expanding thermokarst lake between associated with the loss of water in the lake basin as well 1951 and 2021 (Figure 6). In 1951, Schaeffer Lake as formation of a 125-m-long drainage gully (Figure 7). consisted of two small waterbodies totaling an area of The mean difference in elevations between the two DTMs 0.13 ha (Table 1). Between 1951 and 1974, the two in the undisturbed accuracy assessment area was 0.05 m, waterbodies coalesced, and the lake expanded at an with a range from −0.34 m to 0.26 m (Figure 4). We average rate of 0.28 m/year, nearly doubling in sur- estimate that the lake was between 3 and 4 m deep on face water area (Table 2). From 1974 to 1993, lake average prior to drainage, and microtopography asso- expansion rates increased to 0.35 m/year, which ciated with degrading sub-lake ice wedges was on the added an additional 0.12 ha of surface water area to order of 0.6 to 1.0 m, indicating active sub-lake thermo- the lake. Between 1993 and 2007, lake expansion karst processes prior to its drainage. Downcutting of ice- rates decreased slightly to 0.25 m/year, and the sur- rich permafrost during the drainage event resulted in face water area increased by 0.08 ha, for a total area a thermo-erosional gully extending for 125 m and exceed- of 0.44 ha. Lake expansion rates remained just below ing 13 m in depth adjacent to the old lake shoreline average (0.27 m/year) between 2007 and 2021, (Figure 8). Further downstream, near the base of the whereas the lake area increased to 0.54 ha. Over the ocean bluff, the depth of incision was on the order of 2 most recent observation period, the lake expanded to 4 m. The estimated lake water volume lost through the most rapidly toward the direction it eventually drainage was 18,500 m and the estimated amount of drained, northwards into Kotzebue Sound. material (sediment and ice) transported through erosion Figure 6. Remote sensing time series showing the evolution of Schaeffer Lake between 1951 and 2021. Historic photography is from 1951, 1974, and 1993. High-resolution commercial satellite imagery is from 2007 and 2021. The lake area increased by >300 percent between 1951 and 2021, prior to its drainage in 2022. The grid in the lower row measures 30 m × 30 m. ©2023 Maxar. Table 2. Reconstructing the evolution of Schaeffer Lake between 1951 and 2021. Time period Area change (ha) Area change (%/year) Mean expansion rate (m/year) Mean subsidence rate (m/year) 1951 to 1974 0.11 3.7 0.28 0.12 1974 to 1993 0.12 2.6 0.35 0.17 1993 to 2007 0.08 1.6 0.25 0.27 2007 to 2021 0.10 1.6 0.27 0.78 Note. Area change was determined by manually digitizing the lake perimeter in georeferenced historic aerial photography and contemporary very high- resolution satellite imagery. Mean expansion rates were determined using the DSAS tool (Himmelstoss et al. 2021). Mean subsidence rates were determined by overlaying the past lake perimeters on the estimated permafrost thaw subsidence raster created from the simulated DTM and the 2022 DTM. The mean subsidence rate is relative to the midpoint in the time period relative to the number of years prior to 2021. 8 B. M. JONES ET AL. Figure 7. Elevation changes using DTMs acquired in 2016 and 2022. (a) A DTM derived from a point cloud based on structure-from- motion analysis of airborne digital photography acquired on 1 September 2016. (b) A DTM derived from a point cloud based on structure-from-motion analysis of drone system digital photography acquired on 22 August 2022. (c) Elevation changes exceeding 0.5 m for the thermokarst lake and drainage gully between the 2016 and 2022 DTMs. Elevation change values indicate subsidence, lake level lowering, and thermo-erosion. and thawing was 8,555 m . The 2021 DTM was partially 10.6 m, thus exceeding the estimated uncertainty in affected by lingering snow cover and limited spatial reso- the DTM products. Drained lake basin microtopo- lution relative to the other two DTMs, but it did provide graphy associated with melting ice wedges and an assessment of the perimeter near where the lake degrading polygon centers is on the order of 0.6 to drained. It showed that the lake bank was more than 1.0 m, with troughs being between 8 and 10 m wide 3 m tall and that the distance from the lake edge to an in some cases. elevation equal to or lower than that of the lake surface We further assessed subsidence rates over time by exceeded 25 m, which allowed us to rule out bank over- quantifying subsidence magnitudes using the time ser- topping as a likely drainage mechanism for Schaeffer ies of lake perimeters digitized from the aerial photo- Lake. graphy and very high-resolution satellite imagery (Figure 10). For example, the mean estimated subsi- dence based on the perimeters of the two small ponds Differencing digital terrain models—Subsidence due that already existed in 1951 was 7.4 m. Next, by iso- to thermokarst lake development lating out the area of lake expansion that occurred We estimated the total subsidence due to thermo- between 1951 and 1974, we estimate that the mean karst lake formation and drainage processes by dif- subsidence of the land surface was 7.2 m. Isolating out ferencing the simulated DTM that was developed to the area of lake expansion and estimating the respec- represent the ice-rich permafrost terrain prior to the tive subsidence for each of the remaining three time formation of any thermokarst lake at the site with the periods showed that between 1974 and 1993 it was postdrainage DTM from 2022 (Figure 9). The mean 6.3 m, between 1993 and 2007 it was 5.7 m, and difference in elevations in the undisturbed accuracy between 2007 and 2021 it was 5.5 m. These data assessment area between the simulated DTM and the show that the mean rate of subsidence was initially 2016 and 2022 DTMs was −0.06 m and 0.11 m, quite rapid (0.78 m/year for the area that transitioned respectively, and the range was −0.47 m to 0.27 m from tundra to lake between 2007 and 2021) but that and −0.49 m to 0.22 m, respectively (Figure 4). The it slowed over time (0.12 m/year for the area that estimated mean permafrost thaw subsidence magni- transitioned from tundra to lake between 1951 and tude for the basin was 6.5 m, with a range of 1.9 m to 1974). ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 9 Figure 8. (top) Colorized point clouds developed using structure-from-motion techniques with imagery acquired from a plane (1 September 2016; Fairbanks Fodar 2019) and a drone system (22 August 2022). The thermo-erosional drainage gully is evident in the 22 August 2022, fifty-four days after the thermokarst lake drained. Note the patterned ground associated with melting ice wedges and degrading permafrost that was still present on the lake bottom. The dashed transect lines refer to the topographic plots shown below. DTM profiles showing the elevation differences associated with the lake drainage event. Transect a is a profile across the middle of the lake. It shows thermokarst lake expansion between 2016 versus 2022 as well as the drainage of the lake. Transect b is a profile across the upper portion of the gully that was formed during the drainage event. The maximum downcutting is estimated at 13 to 14 m. Transect c is a profile across the lower portion of the gully, ~100 m downslope of where the lake failure occurred. Note that the x-axis and y-axis are the same scale for each of the transects. 10 B. M. JONES ET AL. Figure 9. Estimating the magnitude of permafrost thaw subsidence using a simulated DTM and the drone system DTM acquired on 18 August 2022. (a) We simulated an “original surface” DTM by digitizing the predrainage lake perimeter, masking it from the 2016 DTM, replacing the values from an adjacent representative upland of the same slope, and then aggregating and resampling the gridded elevation data. (b) The drone system DTM clipped to the extent of Schaeffer Lake in 2021, prior to drainage. (c) Thaw subsidence estimated by differencing the simulated DTM and the drone system DTM. The white polygons in the image represent the lake perimeter at various times in the past. This was used to estimate thaw subsidence magnitude and rates over time. Figure 10. Estimated mean terrain subsidence magnitude and rates based on the difference in elevation between the simulated DTM and the drone system DTM relative to the time periods of thermokarst lake expansion determined with the remote sensing time series imagery. As shown here, thermokarst processes led to rapid rates of subsidence in the first decade following disturbance but diminished over time as the lake deepened. Discussion to assess rates of lake expansion relative to rates of sub- lake permafrost thaw subsidence. There are many similar The life cycle of a thermokarst lake on the Baldwin small thermokarst lakes actively evolving in the ice-rich Peninsula Yedoma uplands of the Baldwin Peninsula. In the case of Major knowledge gaps still exist concerning the environ- Schaeffer Lake, we were lucky to have a historical image mental factors that drive the initiation and long-term from 1951 of the site showing a very young thermokarst growth of permafrost-region lakes such that reliable pro- lake system that matured relatively quickly. The most jections of future lake formation remain limited despite rapid increase in lake surface area occurred in the first several decades of research focused on these aspects of four decades of observation. The lake area more than L-DLB systems (Jorgenson and Shur 2007; S. Wolfe et al. tripled, increasing by 4.2 percent/year between 1951 and 2020; B. M. Jones et al. 2022). Our reconstruction of the 1993, whereas lake area increases slowed to 1.8 percent/ evolution and drainage of the small thermokarst lake on year from 1993 to 2021. Lake expansion rates were also the Schaeffer’s Native allotment provides a unique data set higher in the first two time periods (0.28 and 0.35 m/year) ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 11 relative to the latter time periods of observation (0.25 and that formed more than seventy years ago. However, the 0.27 m/year). In addition, we were able to assess the rate rates that we measured most recently over the approxi- mately fifteen-year period from 2007 to 2021 are nearly of permafrost thaw subsidence between 2007 and 2021, four times higher than thermokarst lake subsidence rates estimated to be 0.78 m/year, which contrasts with the for the middle part of the Lena River basin in eastern lateral rate of expansion of 0.27 m/year over that same Siberia (Fedorov et al. 2014). This difference in the rate time period. of subsidence can likely be explained by differences in The ability to visit the site during the same summer as the near-surface ground ice content and the distribution the lake drainage event provided an opportunity to of ice wedges in the upper permafrost between the two acquire data to create a high-resolution orthophoto regions (Nitzbon et al. 2021). mosaic, DSM, and derived DTM of the freshly drained lake basin. These data allowed us to use a simulated pre- thermokarst lake-affected land surface to estimate per- Reassessing potential lake drainage mechanisms mafrost thaw subsidence rates below a thermokarst lake using our remote sensing time series, which is informa- Common mechanisms that may lead to lake drainage in tion that is not readily available in the literature. Our the continuous permafrost zone include ice wedge degra- results showed the rapid nature of permafrost thaw dation (flow through ice wedge troughs), headward subsidence in the first decade following the transition stream erosion, snow dam accumulation, bank overtop- from tundra to lake through thermokarst but that the ping, river channel migration, coastal erosion, under- rate of permafrost thaw subsidence declined over time. ground piping or tunnel flow (drainage through open These findings are similar to permafrost thaw subsi- frost cracks or layers of permeable material), human dis- dence rates as observed by Ulrich et al. (2017) for turbance, and expansion of a lake toward a drainage Yedoma lakes in Central Yakutia where subsidence gradient (Mackay 1988; Hinkel et al. 2007; Grosse, rates were nearly double the rate for younger lakes that Jones, and Arp, 2013; B. M. Jones, Arp, et al. 2020). In had formed in the last thirty years relative to older lakes “first-order” lakes like Schaeffer Lake, Mackay (1988) Figure 11. (a) Figure 3d from Mackay (1988) depicting the process of lateral, subsurficial drainage via underground tunnel flow, through a process known as ice wedge tunneling, which happens through contraction cracks developed during the previous winter. This lake drainage process likely caused the drainage of Schaeffer Lake. (b) Topographic profiles extracted from the DTM created using the MACS overflights in June 2021 (black line), the year prior to the lake drainage, and in August 2022 (blue dashed line) after drainage. The lake bank was 2 to 3 m tall where the lake drainage occurred, and a drainage gradient extended away from the lake edge more than 25 m. 12 B. M. JONES ET AL. Figure 12. A photo taken by the drone system in August 2022 of the Schaeffer Lake drainage gully. Annotations have been added to the photo showing the lake level prior to drainage, the nearshore permafrost table and thawed bank sediments, a potential remnant contraction crack, wedge ice, and ice-rich permafrost exposed in the drainage gully. The lake likely drained because of tunnel flow through thermal contraction cracks in ice wedges. Thawed bank and sub-lake sediments could have permitted beaver burrowing near where the lake drained. previously identified lateral, subsurficial drainage via drainage occurred, which could have provided thawed underground tunnel flow, through a process known as sediments that enabled burrowing (Figure 12). Thus, we ice wedge tunneling, as a frequent cause of thermokarst hypothesize that beaver burrowing in the banks of ther- lake drainage in the northwestern Canadian Arctic. Lakes mokarst lakes could promote underground tunnel flow draining via this mechanism typically drain during the and erosion of ice wedges, contributing to further snowmelt period via tunnel flow through interconnected increases in Arctic lake drainage. This valuable local ice wedge cracks. However, in the case of Schaeffer Lake, it observation made by the Tessiers provides a potentially drained several weeks following the snowmelt period, so it new thermokarst lake drainage mechanism to be con- is unlikely that snowmelt water triggered the under- sidered in regions where beavers also occur. ground erosion that caused the lake to drain (Figure 11). Prior to drainage, Schaeffer Lake also lacked a drainage L-DLB systems in a warming climate outlet and the banks around the lake perimeter were between 2 and 4 m tall (Figure 11), so it is unlikely that L-DLB systems are rapidly shifting toward one that is bank overtopping triggered the lake drainage. more heavily dominated by drained lake basins where Though we can only speculate here, the direct obser- the trajectory of the evolution of drained lake basins is vation by Susan Tessier of a beaver in the lake for the also likely to be disrupted by a changing climate first time in the week prior to its drainage could point at (B. M. Jones et al. 2022; Webb et al. 2022). Under a cold a key factor contributing to the sudden lake drainage. climate, following lake drainage, permafrost would typi- Beavers are known to dig tunnels and burrows adjacent cally form through top-down freezing (epigenetically) in to waterbodies with bluffs and banks (Zurowski 1992; the freshly exposed basin bottom as well as through quasi- Rozhkova-Timina et al. 2018). The burrows are gener- syngenetic processes associated with peat accumulation in ally short (1–4 m long) but can have many underground the basin (Kanevskiy et al. 2014). The thickness of ice-rich branches and levels, with the entrance of the burrow permafrost below DLBs tends to be restricted to the upper generally being located underwater for protection several meters of the land surface because epigenetic ice against predators (Rozhkova-Timina et al. 2018). wedges tend to grow wider rather than deeper compared Observations at the recently drained lake basin on to their syngenetic counterparts (Mackay 1997). However, 18 August 2022 showed that the permafrost table was given a warmer climate and possibly one with more winter more than 3 m below the lake bed near where the snowfall, there is likely to be a regime shift in drained lake ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 13 basins in continuous permafrost regions such that near- water source has drained, they will either need to melt snow surface permafrost no longer aggrades and new ice wedges for drinking, which is not as desirable, or travel further to do not form in the basin following drainage (B. M. Jones access and harvest ice for their drinking water needs. et al. 2022; Lantz, Zhang, and Kokelj 2022). The unique opportunity provided by the observations documented The importance of local observations in a changing through the LEO Network demonstrate the value of the Arctic observation network in connecting local observers to researchers, which could potentially result in identifying Despite widespread observations of lake drainage in Arctic long-term observation sites for future study. It also docu- Alaska using remote sensing data (Hinkel et al. 2007; ments the mutual benefit from this type of rapid commu- B. M. Jones et al. 2011; Nitze et al. 2018, 2020; Swanson nity-based observation and reporting by jointly helping 2019; Lara, Chen, and Jones 2021), there are very few field- with the interpretation of local events associated with measured or eyewitness records of lake drainage events rapid environmental change in the Arctic. (Mackay 1997; Hinkel et al. 2007; B. M. Jones and Arp The local observations also add insights into the likely 2015; Burn 2020; Rozell 2022; Turner, Wolfe, and effects of a warming Arctic on the cascading effects on McDonald 2022). If the Tessiers had not documented L-DLB landscapes and ecosystems. Range expansion of the lake drainage event that occurred on their Native beaver has been reported as soil conditions change and allotment on 29 June 2022 (Tessier et al. 2022), it is likely warming allows for the growth of shrubs providing habitat that the drainage of the small lake would have gone unno- for moose, muskrat, beaver, and other wildlife (Tape, ticed and/or it would have taken time for it to be detected Christie, et al. 2016; Tape, Gustine, et al. 2016). The expand- in remote sensing data. The rapid reporting associated ing presence of beavers in the region, with more than 12,000 with the LEO Network post along with the back-and- new ponds mapped in imagery since the early 1950s (Tape forth dialogue between the local observers and national et al. 2022), offers unique opportunities to study the inter- and international scientists provided an opportunity to actions between beavers, permafrost, and thermokarst lake visit the site soon after the drainage event so that timely dynamics (B. M. Jones, Tape, et al. 2020). In addition, from observations and data collection could take place in a public health standpoint, beaver range expansion in the August 2022. The presence of historical remote sensing Arctic has been associated with increased risk of water- observations available for the site combined with the borne illness, which may further compromise subsistence detailed local observations has provided new insights drinking water sources in the Arctic (DeMarban 2011). into thermokarst lake dynamics in the Arctic. For exam- ple, the presence of a beaver in the lake for the first time in the week prior to its drainage and the presence of beaver- The impact of lake drainage on drinking water chewed wood near the drainage outlet would have been sources overlooked as a possible drainage mechanism in the Lakes are a key subsistence resource in the Arctic (Berkes absence of local observations and a mechanism to connect 1990; Brinkman et al. 2016; Arp et al. 2019). They provide communities in the Arctic through the LEO Network. important habitat for the subsistence harvest of fish, birds, Since 2009, the Alaska Native Tribal Health Consortium and other wildlife resources (R. J. Wolfe and Walker 1987; has been developing a conversation with Arctic residents to Fauchald et al. 2017), are important transportation routes facilitate sharing their experiences, better understand these (Brinkman et al. 2016; Wrona et al. 2016), and serve as event-based symptoms, and describe relationships between a source of year-round drinking water (Martin et al. 2007; climate change, environmental impacts, and health effects White et al. 2007; Medeiros et al. 2017). Lakes and streams through their Center for Climate and Health. In 2012, the are often the only viable source of drinking water in perma- LEO Network was launched as a tool to help the tribal frost regions due to the lack of access to groundwater health system and local observers to share information resources (Martin et al. 2007; Alessa et al. 2008). In the about climate and other drivers of environmental change. winter, lakes are particularly important drinking water LEO serves as an almanac of first-person accounts of sources because freshwater ice may get harvested to provide unusual environmental events, which are often the first water during the eight- to nine-month-long cold season signals of significant and systemic changes occurring in (Crate 2012; Medeiros et al. 2017). The drainage of their environment. The LEO Network itself is part obser- Schaeffer Lake has resulted in the loss of the Tessiers’ vation system, part social and consultation network, and preferred wintertime source of drinking water. To them, part publishing platform. This approach works because the clear water and once reliable source provided an unlim- people who have intimate knowledge about their local ited supply of freshwater ice that sustained their drinking environment share their observations on the platform water needs during the entire winter. Now that their local and, by doing so, share information with people in other 14 B. M. JONES ET AL. communities and countries. The LEO Network now has Funding approximately 3,600 members from 700 different commu- This research was supported by the U.S. National Science nities worldwide. The drainage of Schaeffer Lake has Foundation under awards OPP-1806213, OPP-1850578, undoubtedly sparked the most in terms of observations OPP-2114051, OIA-1929170, RISE-1927872, and RISE- and contributions from local observers being connected 2052107. GG and IN were additionally supported by the EU H2020 project Arctic PASSION [Grant No. 101003472], and to national and international scientists. The example of the TR was supported by a HEIBRiDS/Geo.X grant. Aerial image sudden lake drainage event documented by the Tessiers’ is acquisitions with the DLR Modular Aerial Camera System evidence that the LEO Network is effective at connecting (MACS) in 2021 was supported through the AWI-funded with others in the community, sharing observations, rais- Perma-X airborne campaign with Polar-6. Geospatial support ing awareness, and discussing the significance of environ- for this work was provided by the Polar Geospatial Center mental events. It also provides that often missing link under NSF-OPP awards 1043681 and 1559691. between local environmental observations and engaging with topic experts in different organizations and from ORCID different disciplinary backgrounds that serves to strengthen the broader Arctic community. Benjamin M. Jones http://orcid.org/0000-0002-1517-4711 Michael Brubaker http://orcid.org/0000-0003-4983-5829 Melissa K. Ward Jones http://orcid.org/0000-0002-3401- Conclusions Guido Grosse http://orcid.org/0000-0001-5895-2141 In this study, we combine local environmental observations Ingmar Nitze http://orcid.org/0000-0002-1165-6852 and remote sensing data to reconstruct the evolution and Tabea Rettelbach http://orcid.org/0000-0002-4187-4113 drainage of a small thermokarst lake formed in ice-rich Sebastian Zavoico http://orcid.org/0000-0002-1570-9122 permafrost on the Baldwin Peninsula in northwestern Ken D. Tape http://orcid.org/0000-0002-1039-6868 Arctic Alaska. The lake drained on 29 June 2022 as observed and documented by the LEO Network. The LEO Network provided a valuable platform for local obser- References vers and national and international scientists to exchange Alessa, L., A. Kliskey, R. Lammers, C. Arp, D. White, observations and information. The online discussion L. Hinzman, and R. Busey. 2008. The Arctic water resource prompted a site visit invitation to the local observer’s vulnerability index: An integrated assessment tool for com- Native allotment that resulted in the collection of valuable munity resilience and vulnerability with respect to freshwater. Environmental Management 42, no. 3: 523–41. scientific data that has provided new insights into the doi:10.1007/s00267-008-9152-0. evolution of thermokarst lakes in the Arctic and this colla- Arp, C. D., M. S. Whitman, B. M. Jones, D. A. Nigro, borative study. In particular, new information on sub-lake V. A. Alexeev, A. Gädeke, S. Fritz, et al. 2019. Ice roads permafrost thaw subsidence rates was made possible as well through lake-rich Arctic watersheds: Integrating climate as observations that led to the hypothesis that beaver bur- uncertainty and freshwater habitat responses into adaptive rowing may facilitate tunnel erosion of ice wedges and lead management. Arctic, Antarctic, and Alpine Research 51, no. 1: 9–23. doi:10.1080/15230430.2018.1560839. to the sudden drainage of Arctic lakes. This study highlights Bergstedt, H., B. M. Jones, K. Hinkel, L. Farquharson, the importance of local environmental observations for B. V. Gaglioti, A. D. Parsekian, M. Kanevskiy, et al. 2021. better understanding and addressing changes occurring in Remote sensing-based statistical approach for defining the Arctic in a more timely and collaborative manner. drained lake basins in a continuous permafrost region, North Slope of Alaska. Remote Sensing 13, no. 13: 2539. doi:10.3390/rs13132539. Acknowledgments Berkes, F. 1990. Native subsistence fisheries: A synthesis of harvest studies in Canada. Arctic 43, no. 1: 35–42. doi:10. We thank Associate Editor Duane Froese (University of Alberta) 14430/arctic1588. and two reviewers, Kevin Turner (Brock University) and Trevor Bouchard, F., D. Fortier, M. Paquette, V. Boucher, R. Pienitz, Lantz (University of Victoria), for providing a thorough review and I. Laurion. 2020. Thermokarst lake inception and and assessment of our article that has improved this published development in syngenetic ice-wedge polygon terrain dur- version. We would also like to thank the National Science ing a cooling climatic trend, Bylot Island (Nunavut), eastern Foundation Research Support and Logistics Program for field Canadian Arctic. The Cryosphere 14, no. 8: 2607–27. doi:10. support provided through contracts with Battelle Arctic Research 5194/tc-14-2607-2020. Operations and Polar Field Services. Brinkman, T. J., W. D. Hansen, F. S. Chapin, G. Kofinas, S. BurnSilver, and T. S. Rupp. 2016. Arctic communities perceive climate impacts on access as a critical challenge to Disclosure statement availability of subsistence resources. Climatic Change 139, No potential conflict of interest was reported by the author(s). no. 3–4: 413–27. doi:10.1007/s10584-016-1819-6. ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 15 Brosius, L. S., K. M. W. Anthony, C. C. Treat, J. Lenz, Identifying historical and future potential lake drainage M. C. Jones, M. S. Bret-Harte, and G. Grosse. 2021. events on the western Arctic coastal plain of Alaska. Spatiotemporal patterns of northern lake formation since Permafrost and Periglacial Processes 31, no. 1: 110–27. the Last Glacial Maximum. Quaternary Science Reviews doi:10.1002/ppp.2038. 253:106773. doi:10.1016/j.quascirev.2020.106773. Jones, B. M., G. Grosse, C. D. Arp, M. C. Jones, K. M. Walter Burn, C. 2020. The Illisarvik drained-lake field experiment: Anthony, and V. E. Romanovsky. 2011. Modern thermo- A legacy of J. Ross Mackay. In Curious about nature: karst lake dynamics in the continuous permafrost zone, A passion for fieldwork , ed. D. Thompson and T. Burt, northern Seward Peninsula, Alaska. Journal of Geophysical 156–60. Cambridge: Cambridge University Press. Research: Biogeosciences 116:G2. doi:10.1029/2011JG Crate, S. A. 2012. Climate change and ice dependent commu- 001666. nities: Perspectives from Siberia and Labrador. The Polar Jones, B. M., G. Grosse, L. M. Farquharson, P. Roy-Léveillée, Journal 2, no. 1: 61–75. doi:10.1080/2154896X.2012.679560. A. Veremeeva, M. Z. Kanevskiy, B. V. Gaglioti, et al. 2022. DeMarban, A. 2011. In Arctic, climate-change threats include Lake and drained lake basin systems in lowland permafrost giardia, food poisoning [online]. http://www.thearcticsoun regions. Nature Reviews Earth & Environment 3, no. 1: der.com/article/1109in_arctic_climate-change_threats_ 85–98. doi:10.1038/s43017-021-00238-9. include (accessed November 15, 2022). Jones, M. C., G. Grosse, B. M. Jones, and K. Walter Anthony. ESRI. 2022. ArcGIS Desktop. Redlands, CA: Environmental 2012. Peat accumulation in drained thermokarst lake basins Systems Research Institute. in continuous, ice-rich permafrost, northern Seward Fairbanks Fodar. 2019. Digital mapping of coast between Cape Peninsula, Alaska. Journal of Geophysical Research: Thompson and Cape Prince of Wales [online]. https://eleva Biogeosciences 117, no. G2: G2. doi:10.1029/2011JG001766. tion.alaska.gov/ (accessed August 21, 2022). Jones, B. M., K. D. Tape, J. A. Clark, I. Nitze, G. Grosse, and Farquharson, L. M., D. H. Mann, G. Grosse, B. M. Jones, and J. Disbrow. 2020. Increase in beaver dams controls surface V. E. Romanovsky. 2016. Spatial distribution of thermo- water and thermokarst dynamics in an Arctic tundra karst terrain in Arctic Alaska. Geomorphology 273:116–33. region. Baldwin Peninsula, Northwestern Alaska 15, no. 7: doi:10.1016/j.geomorph.2016.08.007. 75005. Fauchald, P., V. Hausner, J. Schmidt, and D. Clark. 2017. Jorgenson, M. T., and Y. Shur. 2007. Evolution of lakes and Transitions of social-ecological subsistence systems in the basins in northern Alaska and discussion of the thaw lake Arctic. International Journal of the Commons 11, no. 1: cycle. Journal of Geophysical Research: Earth Surface 112, 275–329. doi:10.18352/ijc.698. no. F2: F2. doi:10.1029/2006JF000531. Fedorov, A. N., P. P. Gavriliev, P. Y. Konstantinov, T. Hiyama, Kanevskiy, M., T. Jorgenson, Y. Shur, J. A. O’Donnell, Y. Iijima, and G. Iwahana. 2014. Estimating the water J. W. Harden, Q. Zhuang, and D. Fortier. 2014. balance of a thermokarst lake in the middle of the Lena Cryostratigraphy and permafrost evolution in the lacustrine River basin, eastern Siberia. Ecohydrology 7, no. 2: 188–96. lowlands of West-Central Alaska. Permafrost and doi:10.1002/eco.1378. Periglacial Processes 25, no. 1: 14–34. doi:10.1002/ppp.1800. Grosse, G., B. Jones, C. Arp. 2013. 8.21 Thermokarst lakes, Lantz, T. C., Y. Zhang, and S. V. Kokelj. 2022. Impacts of drainage, and drained basins. In Treatise on geomorphology, ecological succession and climate warming on permafrost and J. F. Shroder, 325–53. San Diego: Academic Press. aggradation in drained lake basins of the Tuktoyaktuk Himmelstoss, E. A., R. E. Henderson, M. G. Kratzmann, and Coastlands, Northwest Territories, Canada. Permafrost A. S. Farris. 2021. Digital Shoreline Analysis System (DSAS) and Periglacial Processes 33, no. 2: 176–92. doi:10.1002/ version 5.1 user guide. Open-File Report. U.S. Geological ppp.2143. Survey, No. 2021–1091. Lara, M. J., Y. Chen, and B. M. Jones. 2021. Recent warming Hinkel, K. M., R. C. Frohn, F. E. Nelson, W. R. Eisner, and reverses forty-year decline in catastrophic lake drainage and R. A. Beck. 2005. Morphometric and spatial analysis of thaw hastens gradual lake drainage across northern Alaska. lakes and drained thaw lake basins in the western Arctic Environmental Research Letters 16, no. 12: 124019. doi:10. Coastal Plain, Alaska. Permafrost and Periglacial Processes 1088/1748-9326/ac3602. 16, no. 4: 327–41. doi:10.1002/ppp.532. Mackay, J. R. 1988. Catastrophic lake drainage, Tuktoyaktuk Hinkel, K. M., B. M. Jones, W. R. Eisner, C. J. Cuomo, peninsula area, District of Mackenzie. Current Research, R. A. Beck, and R. Frohn. 2007. Methods to assess natural Part D, Geological Survey of Canada, Paper 88:83–90. and anthropogenic thaw lake drainage on the western Mackay, J. R. 1997. A full-scale field experiment (1978–1995) Arctic coastal plain of northern Alaska. Journal of on the growth of permafrost by means of lake drainage, Geophysical Research: Earth Surface 112 (F2). doi:10.1029/ western Arctic coast: A discussion of the method and some 2006JF000584. results. Canadian Journal of Earth Sciences 34, no. 1: 17–33. Jones, B. M. 2023. Geospatial Data Documenting the doi:10.1139/e17-002. Expansion and Drainage of Schaeffer Lake, Northwestern Martin, D., D. Bélanger, P. Gosselin, J. Brazeau, C. Furgal, and Arctic Alaska, 1951-2022. Arctic Data Center. doi:10.18739/ S. Déry. 2007. Drinking water and potential threats to A2930NW7W. human health in Nunavik: Adaptation strategies under Jones, B. M., and C. D. Arp. 2015. Observing a catastrophic climate change conditions. Arctic 60, no. 2: 195–202. thermokarst lake drainage in Northern Alaska. Permafrost Medeiros, A. S., P. Wood, S. D. Wesche, M. Bakaic, and and Periglacial Processes 26, no. 2: 119–28. doi:10.1002/ppp. J. F. Peters. 2017. Water security for northern peoples: 1842. Review of threats to Arctic freshwater systems in Jones, B. M., C. D. Arp, G. Grosse, I. Nitze, M. J. Lara, Nunavut, Canada. Regional Environmental Change 17, no. M. S. Whitman, L. M. Farquharson, et al. 2020. 3: 635–47. doi:10.1007/s10113-016-1084-2. 16 B. M. JONES ET AL. Nitzbon, J., M. Langer, L. C. P. Martin, S. Westermann, Alaska linked to warming and increased shrub habitat. T. Schneider von Deimling, and J. Boike. 2021. Effects of PloS one 11, no. 4: e0152636. doi:10.1371/journal.pone. multi-scale heterogeneity on the simulated evolution of 0152636. ice-rich permafrost lowlands under a warming climate. Tape, K. D., B. M. Jones, C. D. Arp, I. Nitze, and G. Grosse. 2018. The Cryosphere 15, no. 3: 1399–422. doi:10.5194/tc-15- Tundra be dammed: Beaver colonization of the Arctic. Global 1399-2021. Change Biology 24, no. 10: 4478–88. doi:10.1111/gcb.14332. Nitze, I., S. W. Cooley, C. R. Duguay, B. M. Jones, and Tessier, S., B. Forbes, G. Grosse, B. M. Jones, I. Nitze, and G. Grosse. 2020. The catastrophic thermokarst lake drai- J. Schaeffer. 2022. Sudden lake draining event - Local nage events of 2018 in northwestern Alaska: Fast-forward Environmental Observer (LEO) Network [online]. https:// into the future. The Cryosphere 14, no. 12: 4279–97. doi:10. www.leonetwork.org/en/posts/preview/DC4753BF-6270- 5194/tc-14-4279-2020. 46EE-816B-6B97C36B30D3 (accessed August 24, 2022). Nitze, I., G. Grosse, B. M. Jones, V. E. Romanovsky, and Turetsky, M. R., B. W. Abbott, M. C. Jones, K. W. Anthony, J. Boike. 2018. Remote sensing quantifies widespread abun- D. Olefeldt, E. A. G. Schuur, G. Grosse, et al. 2020. Carbon dance of permafrost region disturbances across the Arctic release through abrupt permafrost thaw. Nature Geoscience and Subarctic. Nature Communications 9, no. 1: 5423. 13, no. 2: 138–43. doi:10.1038/s41561-019-0526-0. doi:10.1038/s41467-018-07663-3. Turner, K. W., B. B. Wolfe, and I. McDonald. 2022. Nolan, M., C. Larsen, and M. Sturm. 2015. Mapping snow Monitoring 13 years of drastic catchment change and the depth from manned aircraft on landscape scales at centi- hydroecological responses of a drained thermokarst lake. meter resolution using structure-from-motion photogram- Arctic Science 8, no. 4: 1094–115. metry. The Cryosphere 9, no. 4: 1445–63. Ulrich, M., H. Matthes, L. Schirrmeister, J. Schütze, H. Park, PIX4D. 2022. Pix4dmapper: Release 4.7.5. Prilly, Switzerland: Y. Iijima, and A. N. Fedorov. 2017. Differences in behavior PIX4D S.A. and distribution of permafrost-related lakes in Central Yakutia QTM. 2022. Quick Terrain Modeler: Release 8.3.1. Baltimore, and their response to climatic drivers. Water Resources MD: Applied Imagery, John Hopkins University Applied Research 53, no. 2: 1167–88. doi:10.1002/2016WR019267. Physics Laboratory. Webb, E. E., A. K. Liljedahl, J. A. Cordeiro, M. M. Loranty, Rozell, N. 2022. If a lake drains in northern Alaska . . . [online]. C. Witharana, and J. W. Lichstein. 2022. Permafrost thaw https://uaf.edu/news/if-a-lake-drains-in-northern-alaska. drives surface water decline across lake-rich regions of the php (accessed October 12, 2022). Arctic. Nature Climate Change 12, no. 9: 841–6. doi:10. Rozhkova-Timina, I. O., V. K. Popkov, P. J. Mitchell, and 1038/s41558-022-01455-w. S. N. Kirpotin. 2018. Beavers as ecosystem engineers – White, D. M., S. C. Gerlach, P. Loring, A. C. Tidwell, and A review of their positive and negative effects. IOP M. C. Chambers. 2007 Food and water security in achan- Conference Series: Earth and Environmental Science. ging Arctic climate. Environmental Research Letters. 2, no. 201:012015. 4: 045018. Swanson, D. K. 2019. Thermokarst and precipitation drive Wolfe, S., J. Murton, M. Bateman, and J. Barlow. 2020. changes in the area of lakes and ponds in the National Oriented-lake development in the context of late Parks of northwestern Alaska, 1984–2018. Arctic, Quaternary landscape evolution, McKinley Bay Coastal Antarctic, and Alpine Research 51, no. 1: 265–79. doi:10. Plain, western Arctic Canada. Quaternary Science Reviews 1080/15230430.2019.1629222. 242:106414. doi:10.1016/j.quascirev.2020.106414. Tape, K. D., K. Christie, G. Carroll, and J. A. O’Donnell. 2016. Wolfe, R. J., and R. J. Walker. 1987. Subsistence economies in Novel wildlife in the Arctic: The influence of changing Alaska: Productivity, geography, and development impacts. riparian ecosystems and shrub habitat expansion on snow- Arctic Anthropology 24, no. 2: 56–81. shoe hares. Global Change Biology 22, no. 1: 208–19. doi:10. Wrona, F. J., M. Johansson, J. M. Culp, A. Jenkins, J. Mård, 1111/gcb.13058. I. H. Myers-Smith, T. D. Prowse, W. F. Vincent, and Tape, K. D., J. A. Clark, B. M. Jones, S. Kantner, B. V. Gaglioti, P. A. Wookey. 2016. Transitions in Arctic ecosystems: G. Grosse, and I. Nitze. 2022. Expanding beaver engineering Ecological implications of a changing hydrological regime. creates widespread disturbances in the Arctic. Scientific Journal of Geophysical Research: Biogeosciences 121, no. 3: reports. 650–74. Tape, K. D., D. D. Gustine, R. W. Ruess, L. G. Adams, and Zurowski, W. 1992. Building activity of beavers. Acta J. A. Clark. 2016. Range expansion of moose in Arctic Theriologica 37, no. 4: 403–11. doi:10.4098/AT.arch.92-41. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Arctic Antarctic and Alpine Research Taylor & Francis

Integrating local environmental observations and remote sensing to better understand the life cycle of a thermokarst lake in Arctic Alaska

Download PDF
16 pages

Loading next page...
 
/lp/taylor-francis/integrating-local-environmental-observations-and-remote-sensing-to-zVbmz2dh1i

References (31)

Publisher
Taylor & Francis
Copyright
© 2023 The Author(s). Published with license by Taylor & Francis Group, LLC.
ISSN
1523-0430
eISSN
1938-4246
DOI
10.1080/15230430.2023.2195518
Publisher site
See Article on Publisher Site

Abstract

ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 2023, VOL. 55, NO. 1, 2195518 https://doi.org/10.1080/15230430.2023.2195518 Integrating local environmental observations and remote sensing to better understand the life cycle of a thermokarst lake in Arctic Alaska a b b c c c Benjamin M. Jones , Susan Schaeffer Tessier , Tim Tessier , Michael Brubaker , Mike Brook , Jackie Schaeffer , a d d d e,f Melissa K. Ward Jones , Guido Grosse , Ingmar Nitze , Tabea Rettelbach , Sebastian Zavoico , e e Jason A. Clark , and Ken D. Tape a b Institute of Northern Engineering, University of Alaska Fairbanks, Fairbanks, Alaska, USA; Local Environmental Observer, Kotzebue, Alaska, c d USA; Community Environment and Health, Alaska Native Tribal Health Consortium, Anchorage, Alaska, USA; Helmholtz Centre for Polar and Marine Research, Alfred Wegener Institute, Potsdam, Germany; Geophysical Institute, University of Alaska Fairbanks, Fairbanks, Alaska, USA; Alaska Cooperative Fish and Wildlife Research Unit, University of Alaska Fairbanks, Fairbanks, Alaska, USA ABSTRACT ARTICLE HISTORY Received 16 November 2022 On 29 June 2022, local observers reported the drainage of a 0.5 ha lake near Qikiqtaġruk (Kotzebue), Revised 26 February 2023 Alaska, that prompted this collaborative study on the life cycle of a thermokarst lake in the Arctic. Prior Accepted 22 March 2023 to its drainage, the lake expanded from 0.13 ha in 1951 to 0.54 ha in 2021 at lateral rates that ranged from 0.25 to 0.35 m/year. During the drainage event, we estimate that 18,500 m of water drained KEYWORDS from the lake into Kotzebue Sound, forming a 125-m-long thermo-erosional gully that incised 2 to Arctic; lakes; lake drainage; 14 m in ice-rich permafrost. Between 29 June and 18 August 2022, the drainage gully expanded from local observations; 1 m to >10 m wide, mobilizing ~8,500 m of material through erosion and thaw. By reconstructing permafrost; thermokarst a pre-lake disturbance terrain model, we show that thaw subsidence occurs rapidly (0.78 m/year) upon transition from tundra to lake but that over a seventy-year period it slows to 0.12 m/year. The combination of multiple remote sensing tools and local environmental observations provided a rich data set that allowed us to assess rates of lake expansion relative to rates of sub-lake permafrost thaw subsidence as well as hypothesizing about the potential role of beavers in arctic lake drainage. Introduction (Turetsky et al. 2020; B. M. Jones et al. 2022). Better The formation and drainage of lakes in the Arctic repre- understanding lake drainage processes and document- sent dominant landscape change processes in perma- ing lake drainage events is of importance to climate change research, wildlife and habitat studies, access to frost regions (Grosse, Jones, and Arp 2013; Brosius et al. 2021; B. M. Jones et al. 2022). Their dynamics subsistence resources, and the well-being of northern during the Holocene and Anthropocene have shaped socioecological systems (B. M. Jones et al. 2022). L-DLB systems in the northwestern Alaska Arctic the current state of lowland arctic landscapes and lake and drained lake basin (L-DLB) systems (Farquharson region are continuing to shift to a landscape that is et al. 2016; Bouchard et al. 2020; S. Wolfe et al. 2020; increasingly dominated by DLBs. B. M. Jones et al. B. M. Jones et al. 2022). Taken together, L-DLB systems (2011) found that between 1951 and 2007 on the Cape Espenberg Lowlands of the northern Seward Peninsula, may cover up to 80 percent of the landscape in lowland permafrost regions in the Arctic (Hinkel et al. 2005; the land area gained through lake drainage was nearly 4 M. C. Jones et al. 2012; Grosse, Jones, and Arp 2013; times greater than land area lost through lake formation and expansion. Swanson (2019) conducted lake change B. M. Jones and Arp 2015; Bergstedt et al. 2021). Several recent studies have documented an increase in lake studies for the U.S. National Park Service lands in the evolution and, in particular, lake drainage across the region showing widespread lake losses (mean regional rate loss of 16 ha/year) between 2000 and 2017. Nitze permafrost region (Nitze et al. 2018, 2020; Lara, Chen, and Jones 2021; Webb et al. 2022) with implica- et al. (2020) further documented widespread lake area tions for local-, regional-, and global-scale feedbacks loss for the northern Seward Peninsula and Baldwin CONTACT Benjamin M. Jones bmjones3@alaska.edu Institute of Northern Engineering, University of Alaska Fairbanks, 1764 Tanana Loop Road, Fairbanks, AK 99775 © 2023 The Author(s). Published with license by Taylor & Francis Group, LLC. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. 2 B. M. JONES ET AL. Peninsula in 2017 and 2018, where 192 lakes were iden- had caused the drainage of some thermokarst lakes. Our tified as draining, which exceeded the long-term average recent observations, both with remote sensing and dur- drainage rate by a factor of approximately ten. Lara, ing field studies, in part prompted this study to investi- Chen, and Jones (2021) assessed lake drainage dynamics gate the complex processes triggering lake drainage. across the entirety of the Brooks Range and northwes- On 29 June 2022, between 08:00 and 20:00, a 0.5 ha tern Arctic Alaska between 1975 and 2019, showing that thermokarst lake, with bank heights that ranged from 2 100 to 250 lakes per year had drained between 2015 and to 4 m high, drained on the Schaeffer’s Native allotment 2019 and that by 2050 lake area will likely decline by an on the Baldwin Peninsula, near Qikiqtaġruk (Kotzebue), additional 15 to 21 percent. The factors responsible for Alaska (Figure 1). The event was reported on the Local the rapid nature of lake drainage in the region have been Environmental Observer (LEO) Network by Susan and attributed to increases in mean annual air temperature Tim Tessier, who have their home on the allotment near and precipitation, lengthening of the thaw season, and the lake (Tessier et al. 2022). Over the course of the next increasing near-surface permafrost ground temperature few months, a cross-disciplinary group of scientists and that destabilize ice-rich permafrost soils (Swanson 2019; local knowledge experts from a variety of countries and Nitze et al. 2020; Lara, Chen, and Jones 2021). However, institutions explored the lake basin using aerial and an increase in the presence of beaver engineering in the satellite imagery as well as on-site surveys and discus- region has also been well documented through the for- sions with local observers. The presence of a beaver in mation of more than 12,000 new beaver ponds since the the lake just days before the drainage event and beaver- early 1950s (Tape et al. 2018, 2022). The recent and chewed shrubs on the shore indicate that it may have rapid expansion of beavers in the region adds a level of played a role in the lake drainage event. To better under- complexity when trying to infer the factors responsible stand the factors leading up to this lake drainage event, for lake dynamics. For example, B. M. Jones, Tape, et al. we investigated various forms of imagery to reconstruct (2020) found that beavers were the primary driver of lake water volume and topographic change and estimate surface water area increases on the Baldwin Peninsula permafrost thaw subsidence over time. More specifi - between 2002 and 2019 but their engineering activities cally, we used historical aerial photography and Figure 1. Photos from the LEO post on the sudden lake drainage event that happened on 29 June 2022 (Tessier et al. 2022). (a) Oblique aerial photo of the drained lake basin on 10 August 2022 (photo credit: Sebastian Zavoico). (b) Ground photo of the recently exposed drained lake basin floor and (c) the thermo-erosional drainage gully that formed during the drainage event (photo credits: Susan Tessier, 29 June 2022). (d) The thermo-erosional drainage gully fifty-four days after the lake drained showing continued permafrost thaw and ground ice melt since the drainage event (photo credit: Benjamin Jones). ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 3 contemporary very high-resolution airborne and satel- are signals of environmental change. In 2012, the LEO lite imagery to reconstruct lake evolution prior to drai- Network was launched as an online tool to help the tribal nage as well as corroborating the timing of the lake health system and local observers to share information drainage event. Structure-from-motion (SfM) techni- about climate and other drivers of environmental change. ques were used to process digital photography acquired The event was reported on the LEO Network by Tim and from a plane mounted camera (2016 and 2021) and Susan Tessier, who have their home on a Native allotment a drone system camera (2022) to provide an estimate near the lake, which is located about 5 km northeast of of lake water volume loss and topographic change asso- Kotzebue, Alaska (Figure 2; Tessier et al. 2022). Several ciated with the lake drainage. Simulating a pre-lake researchers were then approached by the LEO Network to formation terrain surface allowed for the estimate of verify and comment on this reported event, which led to permafrost thaw subsidence associated with thermo- the scientific analysis reported here in close collaboration karst lake evolution prior to its drainage in 2022. between local observers and researchers. Reconstructing lake area changes over time Methods We reconstructed the environmental history and drai- Local environmental observations nage of Schaeffer Lake by analyzing several remote sen- The observation of the sudden lake drainage event was sing datasets in a geographic information system (Jones provided through the Alaska Native Tribal Health 2023). Historical aerial photography from 14 July 1951 Consortium’s LEO Network. The LEO Network is (1:20,000 scale), 6 July 1974 (1:6,000 scale), and a community of local and topic experts who share knowl- 31 August 1993 (1:48,000 scale) provided a historical edge about the time and location of specific events that data set to map changes in the lake area over time due Figure 2. The northwestern Arctic region of Alaska where the thermokarst lake drained on 29 June 2022. (a) An inset map showing the location of the northern part of the Baldwin Peninsula (red box). (b) A Landsat satellite image showing the northern Baldwin Peninsula and the location of the drained thermokarst lake (yellow box). (c) A Worldview-2 satellite image from 20 July 2017 showing a series of small thermokarst lakes that have formed in ice-rich permafrost. Schaeffer Lake is the lake that drained on 29 June 2022. ©2023 Maxar. 4 B. M. JONES ET AL. Table 1. Historic aerial photograph and contemporary satellite imagery used in our analysis of Schaeffer Lake evolution between 1951 and 2021. Image Ground control points Spatial resolution Georegistration root mean square error Lake area Image date Image source scale (n) (m) (m) (ha) 14 July 51 Aerial photography 1:20,000 10 1 0.95 0.13 6 July 74 Aerial photography 1:6,000 10 1 0.80 0.24 31 August 93 Aerial photography 1:48,000 10 1 0.47 0.36 4 July 07 Quickbird-2 N/A 10 1 0.71 0.44 4 August 21 Worldview-2 N/A N/A 1 N/A 0.54 Figure 3. Very high-resolution satellite image time series bracketing the drainage of Schaeffer Lake to the time period observed by local environmental observers. The lake drained between 24 June 2022 (note smoke is present) and 30 June 2022 as observed in the 0.5 m resolution satellite images. The observation from 4 August 2021 was the last non-smoke and snow/ice-affected image available for the site prior to the lake drainage. The image acquired on 1 August 2022 demonstrates that the lake drainage was complete. ©2023 Maxar. to lake expansion into ice-rich permafrost. The historical 2007 versus 2021) were measured using Digital Shoreline aerial photograph time series was complemented by very Analysis System (DSAS) Version 5.1 (Himmelstoss et al. high-resolution satellite imagery from 4 July 2007 2021) extension installed in ArcGIS Desktop 10.8.1 (ESRI (Quickbird-2) and 4 August 2021 (Worldview-2) to 2022). This method had previously been used to success- map lake area in the two decades prior to its drainage fully quantify thermokarst lake expansion rates (Table 1). The Worldview-2 image from 2021 was used as (B. M. Jones et al. 2011). Transects were cast 3 m apart the base image to georegister the other four images in the using the baseline polyline drawn on the outside of the time series. The georegistration error was less than 1.0 m 2021 shoreline for a total of 110 transects. Transects were based on using ten ground control points in each image initially cast using automated features in DSAS and edited (Table 1). Visual analysis of very high-resolution satellite to ensure that all transects crossed each lake shoreline in images from 24 June 2022 (Worldview-3), 30 June 2022 a perpendicular direction. The same transects were used for (Worldview-2), and 1 August 2022 (Worldview-2) each DSAS run for each time period. The net expansion bracketed the drainage event in 2022 and demonstrated (m) and mean expansion rate (m/year) were calculated for its complete drainage (Figure 3). each time period using the 110 perpendicular transect lines. Quantifying thermokarst lake margin expansion Developing high-resolution digital surface and rates terrain models Lake expansion rates between each time period (1951 Predrainage digital images were acquired on versus 1974, 1974 versus 1993, 1993 versus 2007, and 1 September 2016 from a small single-engine airplane ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 5 with a Nikon D800E DSLR camera with a 24 mm lens varied from 7 to 8 m/s. The along-track overlap and mounted vertically and connected to a Trimble 5700 across-track overlap of the mission were set at 80 per- survey-grade Global Positioning System (Fairbanks cent and 70 percent, respectively. All images were Fodar 2019). Photo exposure was continuously moni- processed in the software Pix4D Mapper (v4.7.5) to tored and adjusted in flight as needed by the pilot. The produce an orthophoto mosaic and a DSM at spatial bulk of the missions were planned with a 17-cm ground resolutions of 2 and 5 cm, respectively and a point sample distance using a two-pass method along the cloud with ~400 points per square meter. A Leica coast. The coast was followed by eye and then subse- Viva differential Global Positioning System provided quent lines offset the appropriate distance to get ground control for the mission, and the data were a 60 percent across-track overlap. Along-track overlap postprocessed to WGS84 UTM Zone 3 North in was generally 80 percent or more. The photos were ellipsoid heights. processed using the software Agisoft Photoscan We developed predrainage (1 September 2016 and (Nolan, Larsen, and Sturm 2015) to create an ortho- 3 July 2021) and postdrainage (22 August 2022) digital photo mosaic, a digital surface model (DSM), and terrain models (DTMs) using the AGL Analyst function a point cloud with twelve points per square meter in the software Quick Terrain Modeler (QTM v8.31; using North American Datum (NAD)83 UTM Zone 3 QTM 2022). We reprocessed all point cloud data sets to North in orthometric heights (North American Vertical a spatial resolution of 30 cm by calculating a bare earth Datum of 1988 (NAVD)88). DTM using the LidarSense algorithm in QTM. The 2016 Predrainage MACS (Modular Aerial Camera System) and 2021 data sets were reprojected to WGS84 UTM aerial images with red–green–blue and near-infrared sen- Zone 3 North in ellipsoid heights to match the reference frame of the 2022 data prior to differencing the DTM sors were acquired on 3 July 2021 during the Perma-X products to estimate the volume of water lost through 2021 flight campaign. Images were captured on a single lake drainage as well as thermal erosion caused by the lake flight line with an 80 percent along-track overlap and drainage event. Prior to differencing the DTMs, Schaeffer a flight altitude of 1,000 m, which translates to an approx- Lake was hydro-flattened to an ellipsoid height of 31 m imate ground resolution of 10 cm. We processed these in QTM. data to an orthophoto mosaic, DSM, and colorized point cloud in Pix4D Mapper (v4.7.5; PIX4D 2022). The single- track acquisition pattern of the MACS data required the Simulating a predisturbance land surface manual selection of ground control points using the post- drainage drone system orthophoto mosaic and DSM to We simulated a pre-lake formation land surface by ensure spatial accuracy across data sets. extracting topographic information from a 0.5 ha area Postdrainage digital images were acquired from adjacent to Schaeffer Lake in the 2016 DTM. The area a DJI P4RTK quadcopter with a DJI D-RTK where the elevation values were extracted represented 2 Mobile Base Station at the Schaeffer Lake drainage the same inferred topography as the site where study site on 22 August 2022. The drone system was Schaeffer Lake had developed. These inferences were flown at 50 m above ground level, and flight speeds based on information in historical U.S. Geological Figure 4. Assessing the accuracy of differencing the three digital terrain models (DTMs) used in our assessments of elevation change. The difference in elevation values was extracted from a 30 m x 30 m area that appeared to have remained stable and undisturbed between 2016 and 2022. Left, 2022 DTM versus 2016 DTM; middle, simulated DTM versus 2016 DTM; right, simulated DTM versus 2022 DTM. 6 B. M. JONES ET AL. Survey topographic maps as well as features visible in allotment near the lake, which is located 5 km northeast of Kotzebue, Alaska (Tessier et al. 2022): the historical aerial photography. The hydro-flattened surface of Schaeffer Lake in the 2016 DTM was replaced The event occurred on June 29th, on our native allot- with the extracted elevation values from the adjacent ment (Illivak) near Kotzebue. We left home in the upland land surface. We downsampled the merged morning and when we came back around 8:00 PM in the evening the whole lake had drained! It looked like it product using an aggregation factor of seventy-five to was blown up with dynamite. (Susan Schaeffer Tessier, smooth out the terrain surface. We then resampled the 29 June 2022) simulated DTM to a spatial resolution of 30 cm for differencing with the 2022 DTM to estimate the total The Tessiers observed that the drainage gully was 1 m thaw subsidence associated with the evolution of the wide on the north end of the lake and that it eroded lake prior to its drainage. We performed an accuracy downwards at least 6 to 7 m into ice-rich permafrost assessment of all three DTMs (2016, 2022, and simu- during the drainage event. Very high-resolution satellite lated) by differencing each product inside what images acquired on 24 June 2022 and 30 June 2022 also appeared to be an undisturbed 30 m × 30 m area detected lake drainage in this same time span (Figure 3). The Tessiers also noted the pattern of degraded ice between 2016 and 2022 (Figure 4). wedges and intervening ice wedge polygon centers evi- dent in the drained thermokarst lake basin bottom as Results well as accumulations of dead aquatic invertebrates that were providing forage for ducks. The week before the Local environmental observations of the sudden sudden lake drainage, Susan saw a beaver in the lake for thermokarst lake drainage the first time (Figure 5a). Following the sudden lake On 29 June 2022, a sudden lake drainage event was drainage, Tim also observed signs of beaver-chewed reported on the LEO Network by Susan Schaeffer wood in the willow shrubs adjacent to where the lake Tessier and Tim Tessier, who have their home on the drainage occurred (Figure 5c,d). Figure 5. Evidence for beaver burrowing as a thermokarst lake drainage mechanism (red circles and ovals). (a) A beaver showed up in the lake a week before the lake drained for the first time in at least four years (photo credit: Susan Schaeffer Tessier). (b) An image acquired by the drone system during a flight down the drainage gully showing beaver-chewed wood (photo credit: Benjamin Jones). (c) and (d) Beaver chew on the bush near the location where the drainage gully formed (photo credit: Tim Tessier). ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 7 Decadal-scale thermokarst lake evolution prior to Differencing digital terrain models—Lake drainage drainage and gully erosion Historic aerial photographs and contemporary very We differenced the pre- and post-drainage DTM pro- high-resolution satellite image time series show ducts from 2016 and 2022 to assess changes in elevation a gradually expanding thermokarst lake between associated with the loss of water in the lake basin as well 1951 and 2021 (Figure 6). In 1951, Schaeffer Lake as formation of a 125-m-long drainage gully (Figure 7). consisted of two small waterbodies totaling an area of The mean difference in elevations between the two DTMs 0.13 ha (Table 1). Between 1951 and 1974, the two in the undisturbed accuracy assessment area was 0.05 m, waterbodies coalesced, and the lake expanded at an with a range from −0.34 m to 0.26 m (Figure 4). We average rate of 0.28 m/year, nearly doubling in sur- estimate that the lake was between 3 and 4 m deep on face water area (Table 2). From 1974 to 1993, lake average prior to drainage, and microtopography asso- expansion rates increased to 0.35 m/year, which ciated with degrading sub-lake ice wedges was on the added an additional 0.12 ha of surface water area to order of 0.6 to 1.0 m, indicating active sub-lake thermo- the lake. Between 1993 and 2007, lake expansion karst processes prior to its drainage. Downcutting of ice- rates decreased slightly to 0.25 m/year, and the sur- rich permafrost during the drainage event resulted in face water area increased by 0.08 ha, for a total area a thermo-erosional gully extending for 125 m and exceed- of 0.44 ha. Lake expansion rates remained just below ing 13 m in depth adjacent to the old lake shoreline average (0.27 m/year) between 2007 and 2021, (Figure 8). Further downstream, near the base of the whereas the lake area increased to 0.54 ha. Over the ocean bluff, the depth of incision was on the order of 2 most recent observation period, the lake expanded to 4 m. The estimated lake water volume lost through the most rapidly toward the direction it eventually drainage was 18,500 m and the estimated amount of drained, northwards into Kotzebue Sound. material (sediment and ice) transported through erosion Figure 6. Remote sensing time series showing the evolution of Schaeffer Lake between 1951 and 2021. Historic photography is from 1951, 1974, and 1993. High-resolution commercial satellite imagery is from 2007 and 2021. The lake area increased by >300 percent between 1951 and 2021, prior to its drainage in 2022. The grid in the lower row measures 30 m × 30 m. ©2023 Maxar. Table 2. Reconstructing the evolution of Schaeffer Lake between 1951 and 2021. Time period Area change (ha) Area change (%/year) Mean expansion rate (m/year) Mean subsidence rate (m/year) 1951 to 1974 0.11 3.7 0.28 0.12 1974 to 1993 0.12 2.6 0.35 0.17 1993 to 2007 0.08 1.6 0.25 0.27 2007 to 2021 0.10 1.6 0.27 0.78 Note. Area change was determined by manually digitizing the lake perimeter in georeferenced historic aerial photography and contemporary very high- resolution satellite imagery. Mean expansion rates were determined using the DSAS tool (Himmelstoss et al. 2021). Mean subsidence rates were determined by overlaying the past lake perimeters on the estimated permafrost thaw subsidence raster created from the simulated DTM and the 2022 DTM. The mean subsidence rate is relative to the midpoint in the time period relative to the number of years prior to 2021. 8 B. M. JONES ET AL. Figure 7. Elevation changes using DTMs acquired in 2016 and 2022. (a) A DTM derived from a point cloud based on structure-from- motion analysis of airborne digital photography acquired on 1 September 2016. (b) A DTM derived from a point cloud based on structure-from-motion analysis of drone system digital photography acquired on 22 August 2022. (c) Elevation changes exceeding 0.5 m for the thermokarst lake and drainage gully between the 2016 and 2022 DTMs. Elevation change values indicate subsidence, lake level lowering, and thermo-erosion. and thawing was 8,555 m . The 2021 DTM was partially 10.6 m, thus exceeding the estimated uncertainty in affected by lingering snow cover and limited spatial reso- the DTM products. Drained lake basin microtopo- lution relative to the other two DTMs, but it did provide graphy associated with melting ice wedges and an assessment of the perimeter near where the lake degrading polygon centers is on the order of 0.6 to drained. It showed that the lake bank was more than 1.0 m, with troughs being between 8 and 10 m wide 3 m tall and that the distance from the lake edge to an in some cases. elevation equal to or lower than that of the lake surface We further assessed subsidence rates over time by exceeded 25 m, which allowed us to rule out bank over- quantifying subsidence magnitudes using the time ser- topping as a likely drainage mechanism for Schaeffer ies of lake perimeters digitized from the aerial photo- Lake. graphy and very high-resolution satellite imagery (Figure 10). For example, the mean estimated subsi- dence based on the perimeters of the two small ponds Differencing digital terrain models—Subsidence due that already existed in 1951 was 7.4 m. Next, by iso- to thermokarst lake development lating out the area of lake expansion that occurred We estimated the total subsidence due to thermo- between 1951 and 1974, we estimate that the mean karst lake formation and drainage processes by dif- subsidence of the land surface was 7.2 m. Isolating out ferencing the simulated DTM that was developed to the area of lake expansion and estimating the respec- represent the ice-rich permafrost terrain prior to the tive subsidence for each of the remaining three time formation of any thermokarst lake at the site with the periods showed that between 1974 and 1993 it was postdrainage DTM from 2022 (Figure 9). The mean 6.3 m, between 1993 and 2007 it was 5.7 m, and difference in elevations in the undisturbed accuracy between 2007 and 2021 it was 5.5 m. These data assessment area between the simulated DTM and the show that the mean rate of subsidence was initially 2016 and 2022 DTMs was −0.06 m and 0.11 m, quite rapid (0.78 m/year for the area that transitioned respectively, and the range was −0.47 m to 0.27 m from tundra to lake between 2007 and 2021) but that and −0.49 m to 0.22 m, respectively (Figure 4). The it slowed over time (0.12 m/year for the area that estimated mean permafrost thaw subsidence magni- transitioned from tundra to lake between 1951 and tude for the basin was 6.5 m, with a range of 1.9 m to 1974). ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 9 Figure 8. (top) Colorized point clouds developed using structure-from-motion techniques with imagery acquired from a plane (1 September 2016; Fairbanks Fodar 2019) and a drone system (22 August 2022). The thermo-erosional drainage gully is evident in the 22 August 2022, fifty-four days after the thermokarst lake drained. Note the patterned ground associated with melting ice wedges and degrading permafrost that was still present on the lake bottom. The dashed transect lines refer to the topographic plots shown below. DTM profiles showing the elevation differences associated with the lake drainage event. Transect a is a profile across the middle of the lake. It shows thermokarst lake expansion between 2016 versus 2022 as well as the drainage of the lake. Transect b is a profile across the upper portion of the gully that was formed during the drainage event. The maximum downcutting is estimated at 13 to 14 m. Transect c is a profile across the lower portion of the gully, ~100 m downslope of where the lake failure occurred. Note that the x-axis and y-axis are the same scale for each of the transects. 10 B. M. JONES ET AL. Figure 9. Estimating the magnitude of permafrost thaw subsidence using a simulated DTM and the drone system DTM acquired on 18 August 2022. (a) We simulated an “original surface” DTM by digitizing the predrainage lake perimeter, masking it from the 2016 DTM, replacing the values from an adjacent representative upland of the same slope, and then aggregating and resampling the gridded elevation data. (b) The drone system DTM clipped to the extent of Schaeffer Lake in 2021, prior to drainage. (c) Thaw subsidence estimated by differencing the simulated DTM and the drone system DTM. The white polygons in the image represent the lake perimeter at various times in the past. This was used to estimate thaw subsidence magnitude and rates over time. Figure 10. Estimated mean terrain subsidence magnitude and rates based on the difference in elevation between the simulated DTM and the drone system DTM relative to the time periods of thermokarst lake expansion determined with the remote sensing time series imagery. As shown here, thermokarst processes led to rapid rates of subsidence in the first decade following disturbance but diminished over time as the lake deepened. Discussion to assess rates of lake expansion relative to rates of sub- lake permafrost thaw subsidence. There are many similar The life cycle of a thermokarst lake on the Baldwin small thermokarst lakes actively evolving in the ice-rich Peninsula Yedoma uplands of the Baldwin Peninsula. In the case of Major knowledge gaps still exist concerning the environ- Schaeffer Lake, we were lucky to have a historical image mental factors that drive the initiation and long-term from 1951 of the site showing a very young thermokarst growth of permafrost-region lakes such that reliable pro- lake system that matured relatively quickly. The most jections of future lake formation remain limited despite rapid increase in lake surface area occurred in the first several decades of research focused on these aspects of four decades of observation. The lake area more than L-DLB systems (Jorgenson and Shur 2007; S. Wolfe et al. tripled, increasing by 4.2 percent/year between 1951 and 2020; B. M. Jones et al. 2022). Our reconstruction of the 1993, whereas lake area increases slowed to 1.8 percent/ evolution and drainage of the small thermokarst lake on year from 1993 to 2021. Lake expansion rates were also the Schaeffer’s Native allotment provides a unique data set higher in the first two time periods (0.28 and 0.35 m/year) ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 11 relative to the latter time periods of observation (0.25 and that formed more than seventy years ago. However, the 0.27 m/year). In addition, we were able to assess the rate rates that we measured most recently over the approxi- mately fifteen-year period from 2007 to 2021 are nearly of permafrost thaw subsidence between 2007 and 2021, four times higher than thermokarst lake subsidence rates estimated to be 0.78 m/year, which contrasts with the for the middle part of the Lena River basin in eastern lateral rate of expansion of 0.27 m/year over that same Siberia (Fedorov et al. 2014). This difference in the rate time period. of subsidence can likely be explained by differences in The ability to visit the site during the same summer as the near-surface ground ice content and the distribution the lake drainage event provided an opportunity to of ice wedges in the upper permafrost between the two acquire data to create a high-resolution orthophoto regions (Nitzbon et al. 2021). mosaic, DSM, and derived DTM of the freshly drained lake basin. These data allowed us to use a simulated pre- thermokarst lake-affected land surface to estimate per- Reassessing potential lake drainage mechanisms mafrost thaw subsidence rates below a thermokarst lake using our remote sensing time series, which is informa- Common mechanisms that may lead to lake drainage in tion that is not readily available in the literature. Our the continuous permafrost zone include ice wedge degra- results showed the rapid nature of permafrost thaw dation (flow through ice wedge troughs), headward subsidence in the first decade following the transition stream erosion, snow dam accumulation, bank overtop- from tundra to lake through thermokarst but that the ping, river channel migration, coastal erosion, under- rate of permafrost thaw subsidence declined over time. ground piping or tunnel flow (drainage through open These findings are similar to permafrost thaw subsi- frost cracks or layers of permeable material), human dis- dence rates as observed by Ulrich et al. (2017) for turbance, and expansion of a lake toward a drainage Yedoma lakes in Central Yakutia where subsidence gradient (Mackay 1988; Hinkel et al. 2007; Grosse, rates were nearly double the rate for younger lakes that Jones, and Arp, 2013; B. M. Jones, Arp, et al. 2020). In had formed in the last thirty years relative to older lakes “first-order” lakes like Schaeffer Lake, Mackay (1988) Figure 11. (a) Figure 3d from Mackay (1988) depicting the process of lateral, subsurficial drainage via underground tunnel flow, through a process known as ice wedge tunneling, which happens through contraction cracks developed during the previous winter. This lake drainage process likely caused the drainage of Schaeffer Lake. (b) Topographic profiles extracted from the DTM created using the MACS overflights in June 2021 (black line), the year prior to the lake drainage, and in August 2022 (blue dashed line) after drainage. The lake bank was 2 to 3 m tall where the lake drainage occurred, and a drainage gradient extended away from the lake edge more than 25 m. 12 B. M. JONES ET AL. Figure 12. A photo taken by the drone system in August 2022 of the Schaeffer Lake drainage gully. Annotations have been added to the photo showing the lake level prior to drainage, the nearshore permafrost table and thawed bank sediments, a potential remnant contraction crack, wedge ice, and ice-rich permafrost exposed in the drainage gully. The lake likely drained because of tunnel flow through thermal contraction cracks in ice wedges. Thawed bank and sub-lake sediments could have permitted beaver burrowing near where the lake drained. previously identified lateral, subsurficial drainage via drainage occurred, which could have provided thawed underground tunnel flow, through a process known as sediments that enabled burrowing (Figure 12). Thus, we ice wedge tunneling, as a frequent cause of thermokarst hypothesize that beaver burrowing in the banks of ther- lake drainage in the northwestern Canadian Arctic. Lakes mokarst lakes could promote underground tunnel flow draining via this mechanism typically drain during the and erosion of ice wedges, contributing to further snowmelt period via tunnel flow through interconnected increases in Arctic lake drainage. This valuable local ice wedge cracks. However, in the case of Schaeffer Lake, it observation made by the Tessiers provides a potentially drained several weeks following the snowmelt period, so it new thermokarst lake drainage mechanism to be con- is unlikely that snowmelt water triggered the under- sidered in regions where beavers also occur. ground erosion that caused the lake to drain (Figure 11). Prior to drainage, Schaeffer Lake also lacked a drainage L-DLB systems in a warming climate outlet and the banks around the lake perimeter were between 2 and 4 m tall (Figure 11), so it is unlikely that L-DLB systems are rapidly shifting toward one that is bank overtopping triggered the lake drainage. more heavily dominated by drained lake basins where Though we can only speculate here, the direct obser- the trajectory of the evolution of drained lake basins is vation by Susan Tessier of a beaver in the lake for the also likely to be disrupted by a changing climate first time in the week prior to its drainage could point at (B. M. Jones et al. 2022; Webb et al. 2022). Under a cold a key factor contributing to the sudden lake drainage. climate, following lake drainage, permafrost would typi- Beavers are known to dig tunnels and burrows adjacent cally form through top-down freezing (epigenetically) in to waterbodies with bluffs and banks (Zurowski 1992; the freshly exposed basin bottom as well as through quasi- Rozhkova-Timina et al. 2018). The burrows are gener- syngenetic processes associated with peat accumulation in ally short (1–4 m long) but can have many underground the basin (Kanevskiy et al. 2014). The thickness of ice-rich branches and levels, with the entrance of the burrow permafrost below DLBs tends to be restricted to the upper generally being located underwater for protection several meters of the land surface because epigenetic ice against predators (Rozhkova-Timina et al. 2018). wedges tend to grow wider rather than deeper compared Observations at the recently drained lake basin on to their syngenetic counterparts (Mackay 1997). However, 18 August 2022 showed that the permafrost table was given a warmer climate and possibly one with more winter more than 3 m below the lake bed near where the snowfall, there is likely to be a regime shift in drained lake ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 13 basins in continuous permafrost regions such that near- water source has drained, they will either need to melt snow surface permafrost no longer aggrades and new ice wedges for drinking, which is not as desirable, or travel further to do not form in the basin following drainage (B. M. Jones access and harvest ice for their drinking water needs. et al. 2022; Lantz, Zhang, and Kokelj 2022). The unique opportunity provided by the observations documented The importance of local observations in a changing through the LEO Network demonstrate the value of the Arctic observation network in connecting local observers to researchers, which could potentially result in identifying Despite widespread observations of lake drainage in Arctic long-term observation sites for future study. It also docu- Alaska using remote sensing data (Hinkel et al. 2007; ments the mutual benefit from this type of rapid commu- B. M. Jones et al. 2011; Nitze et al. 2018, 2020; Swanson nity-based observation and reporting by jointly helping 2019; Lara, Chen, and Jones 2021), there are very few field- with the interpretation of local events associated with measured or eyewitness records of lake drainage events rapid environmental change in the Arctic. (Mackay 1997; Hinkel et al. 2007; B. M. Jones and Arp The local observations also add insights into the likely 2015; Burn 2020; Rozell 2022; Turner, Wolfe, and effects of a warming Arctic on the cascading effects on McDonald 2022). If the Tessiers had not documented L-DLB landscapes and ecosystems. Range expansion of the lake drainage event that occurred on their Native beaver has been reported as soil conditions change and allotment on 29 June 2022 (Tessier et al. 2022), it is likely warming allows for the growth of shrubs providing habitat that the drainage of the small lake would have gone unno- for moose, muskrat, beaver, and other wildlife (Tape, ticed and/or it would have taken time for it to be detected Christie, et al. 2016; Tape, Gustine, et al. 2016). The expand- in remote sensing data. The rapid reporting associated ing presence of beavers in the region, with more than 12,000 with the LEO Network post along with the back-and- new ponds mapped in imagery since the early 1950s (Tape forth dialogue between the local observers and national et al. 2022), offers unique opportunities to study the inter- and international scientists provided an opportunity to actions between beavers, permafrost, and thermokarst lake visit the site soon after the drainage event so that timely dynamics (B. M. Jones, Tape, et al. 2020). In addition, from observations and data collection could take place in a public health standpoint, beaver range expansion in the August 2022. The presence of historical remote sensing Arctic has been associated with increased risk of water- observations available for the site combined with the borne illness, which may further compromise subsistence detailed local observations has provided new insights drinking water sources in the Arctic (DeMarban 2011). into thermokarst lake dynamics in the Arctic. For exam- ple, the presence of a beaver in the lake for the first time in the week prior to its drainage and the presence of beaver- The impact of lake drainage on drinking water chewed wood near the drainage outlet would have been sources overlooked as a possible drainage mechanism in the Lakes are a key subsistence resource in the Arctic (Berkes absence of local observations and a mechanism to connect 1990; Brinkman et al. 2016; Arp et al. 2019). They provide communities in the Arctic through the LEO Network. important habitat for the subsistence harvest of fish, birds, Since 2009, the Alaska Native Tribal Health Consortium and other wildlife resources (R. J. Wolfe and Walker 1987; has been developing a conversation with Arctic residents to Fauchald et al. 2017), are important transportation routes facilitate sharing their experiences, better understand these (Brinkman et al. 2016; Wrona et al. 2016), and serve as event-based symptoms, and describe relationships between a source of year-round drinking water (Martin et al. 2007; climate change, environmental impacts, and health effects White et al. 2007; Medeiros et al. 2017). Lakes and streams through their Center for Climate and Health. In 2012, the are often the only viable source of drinking water in perma- LEO Network was launched as a tool to help the tribal frost regions due to the lack of access to groundwater health system and local observers to share information resources (Martin et al. 2007; Alessa et al. 2008). In the about climate and other drivers of environmental change. winter, lakes are particularly important drinking water LEO serves as an almanac of first-person accounts of sources because freshwater ice may get harvested to provide unusual environmental events, which are often the first water during the eight- to nine-month-long cold season signals of significant and systemic changes occurring in (Crate 2012; Medeiros et al. 2017). The drainage of their environment. The LEO Network itself is part obser- Schaeffer Lake has resulted in the loss of the Tessiers’ vation system, part social and consultation network, and preferred wintertime source of drinking water. To them, part publishing platform. This approach works because the clear water and once reliable source provided an unlim- people who have intimate knowledge about their local ited supply of freshwater ice that sustained their drinking environment share their observations on the platform water needs during the entire winter. Now that their local and, by doing so, share information with people in other 14 B. M. JONES ET AL. communities and countries. The LEO Network now has Funding approximately 3,600 members from 700 different commu- This research was supported by the U.S. National Science nities worldwide. The drainage of Schaeffer Lake has Foundation under awards OPP-1806213, OPP-1850578, undoubtedly sparked the most in terms of observations OPP-2114051, OIA-1929170, RISE-1927872, and RISE- and contributions from local observers being connected 2052107. GG and IN were additionally supported by the EU H2020 project Arctic PASSION [Grant No. 101003472], and to national and international scientists. The example of the TR was supported by a HEIBRiDS/Geo.X grant. Aerial image sudden lake drainage event documented by the Tessiers’ is acquisitions with the DLR Modular Aerial Camera System evidence that the LEO Network is effective at connecting (MACS) in 2021 was supported through the AWI-funded with others in the community, sharing observations, rais- Perma-X airborne campaign with Polar-6. Geospatial support ing awareness, and discussing the significance of environ- for this work was provided by the Polar Geospatial Center mental events. It also provides that often missing link under NSF-OPP awards 1043681 and 1559691. between local environmental observations and engaging with topic experts in different organizations and from ORCID different disciplinary backgrounds that serves to strengthen the broader Arctic community. Benjamin M. Jones http://orcid.org/0000-0002-1517-4711 Michael Brubaker http://orcid.org/0000-0003-4983-5829 Melissa K. Ward Jones http://orcid.org/0000-0002-3401- Conclusions Guido Grosse http://orcid.org/0000-0001-5895-2141 In this study, we combine local environmental observations Ingmar Nitze http://orcid.org/0000-0002-1165-6852 and remote sensing data to reconstruct the evolution and Tabea Rettelbach http://orcid.org/0000-0002-4187-4113 drainage of a small thermokarst lake formed in ice-rich Sebastian Zavoico http://orcid.org/0000-0002-1570-9122 permafrost on the Baldwin Peninsula in northwestern Ken D. Tape http://orcid.org/0000-0002-1039-6868 Arctic Alaska. The lake drained on 29 June 2022 as observed and documented by the LEO Network. The LEO Network provided a valuable platform for local obser- References vers and national and international scientists to exchange Alessa, L., A. Kliskey, R. Lammers, C. Arp, D. White, observations and information. The online discussion L. Hinzman, and R. Busey. 2008. The Arctic water resource prompted a site visit invitation to the local observer’s vulnerability index: An integrated assessment tool for com- Native allotment that resulted in the collection of valuable munity resilience and vulnerability with respect to freshwater. Environmental Management 42, no. 3: 523–41. scientific data that has provided new insights into the doi:10.1007/s00267-008-9152-0. evolution of thermokarst lakes in the Arctic and this colla- Arp, C. D., M. S. Whitman, B. M. Jones, D. A. Nigro, borative study. In particular, new information on sub-lake V. A. Alexeev, A. Gädeke, S. Fritz, et al. 2019. Ice roads permafrost thaw subsidence rates was made possible as well through lake-rich Arctic watersheds: Integrating climate as observations that led to the hypothesis that beaver bur- uncertainty and freshwater habitat responses into adaptive rowing may facilitate tunnel erosion of ice wedges and lead management. Arctic, Antarctic, and Alpine Research 51, no. 1: 9–23. doi:10.1080/15230430.2018.1560839. to the sudden drainage of Arctic lakes. This study highlights Bergstedt, H., B. M. Jones, K. Hinkel, L. Farquharson, the importance of local environmental observations for B. V. Gaglioti, A. D. Parsekian, M. Kanevskiy, et al. 2021. better understanding and addressing changes occurring in Remote sensing-based statistical approach for defining the Arctic in a more timely and collaborative manner. drained lake basins in a continuous permafrost region, North Slope of Alaska. Remote Sensing 13, no. 13: 2539. doi:10.3390/rs13132539. Acknowledgments Berkes, F. 1990. Native subsistence fisheries: A synthesis of harvest studies in Canada. Arctic 43, no. 1: 35–42. doi:10. We thank Associate Editor Duane Froese (University of Alberta) 14430/arctic1588. and two reviewers, Kevin Turner (Brock University) and Trevor Bouchard, F., D. Fortier, M. Paquette, V. Boucher, R. Pienitz, Lantz (University of Victoria), for providing a thorough review and I. Laurion. 2020. Thermokarst lake inception and and assessment of our article that has improved this published development in syngenetic ice-wedge polygon terrain dur- version. We would also like to thank the National Science ing a cooling climatic trend, Bylot Island (Nunavut), eastern Foundation Research Support and Logistics Program for field Canadian Arctic. The Cryosphere 14, no. 8: 2607–27. doi:10. support provided through contracts with Battelle Arctic Research 5194/tc-14-2607-2020. Operations and Polar Field Services. Brinkman, T. J., W. D. Hansen, F. S. Chapin, G. Kofinas, S. BurnSilver, and T. S. Rupp. 2016. Arctic communities perceive climate impacts on access as a critical challenge to Disclosure statement availability of subsistence resources. Climatic Change 139, No potential conflict of interest was reported by the author(s). no. 3–4: 413–27. doi:10.1007/s10584-016-1819-6. ARCTIC, ANTARCTIC, AND ALPINE RESEARCH 15 Brosius, L. S., K. M. W. Anthony, C. C. Treat, J. Lenz, Identifying historical and future potential lake drainage M. C. Jones, M. S. Bret-Harte, and G. Grosse. 2021. events on the western Arctic coastal plain of Alaska. Spatiotemporal patterns of northern lake formation since Permafrost and Periglacial Processes 31, no. 1: 110–27. the Last Glacial Maximum. Quaternary Science Reviews doi:10.1002/ppp.2038. 253:106773. doi:10.1016/j.quascirev.2020.106773. Jones, B. M., G. Grosse, C. D. Arp, M. C. Jones, K. M. Walter Burn, C. 2020. The Illisarvik drained-lake field experiment: Anthony, and V. E. Romanovsky. 2011. Modern thermo- A legacy of J. Ross Mackay. In Curious about nature: karst lake dynamics in the continuous permafrost zone, A passion for fieldwork , ed. D. Thompson and T. Burt, northern Seward Peninsula, Alaska. Journal of Geophysical 156–60. Cambridge: Cambridge University Press. Research: Biogeosciences 116:G2. doi:10.1029/2011JG Crate, S. A. 2012. Climate change and ice dependent commu- 001666. nities: Perspectives from Siberia and Labrador. The Polar Jones, B. M., G. Grosse, L. M. Farquharson, P. Roy-Léveillée, Journal 2, no. 1: 61–75. doi:10.1080/2154896X.2012.679560. A. Veremeeva, M. Z. Kanevskiy, B. V. Gaglioti, et al. 2022. DeMarban, A. 2011. In Arctic, climate-change threats include Lake and drained lake basin systems in lowland permafrost giardia, food poisoning [online]. http://www.thearcticsoun regions. Nature Reviews Earth & Environment 3, no. 1: der.com/article/1109in_arctic_climate-change_threats_ 85–98. doi:10.1038/s43017-021-00238-9. include (accessed November 15, 2022). Jones, M. C., G. Grosse, B. M. Jones, and K. Walter Anthony. ESRI. 2022. ArcGIS Desktop. Redlands, CA: Environmental 2012. Peat accumulation in drained thermokarst lake basins Systems Research Institute. in continuous, ice-rich permafrost, northern Seward Fairbanks Fodar. 2019. Digital mapping of coast between Cape Peninsula, Alaska. Journal of Geophysical Research: Thompson and Cape Prince of Wales [online]. https://eleva Biogeosciences 117, no. G2: G2. doi:10.1029/2011JG001766. tion.alaska.gov/ (accessed August 21, 2022). Jones, B. M., K. D. Tape, J. A. Clark, I. Nitze, G. Grosse, and Farquharson, L. M., D. H. Mann, G. Grosse, B. M. Jones, and J. Disbrow. 2020. Increase in beaver dams controls surface V. E. Romanovsky. 2016. Spatial distribution of thermo- water and thermokarst dynamics in an Arctic tundra karst terrain in Arctic Alaska. Geomorphology 273:116–33. region. Baldwin Peninsula, Northwestern Alaska 15, no. 7: doi:10.1016/j.geomorph.2016.08.007. 75005. Fauchald, P., V. Hausner, J. Schmidt, and D. Clark. 2017. Jorgenson, M. T., and Y. Shur. 2007. Evolution of lakes and Transitions of social-ecological subsistence systems in the basins in northern Alaska and discussion of the thaw lake Arctic. International Journal of the Commons 11, no. 1: cycle. Journal of Geophysical Research: Earth Surface 112, 275–329. doi:10.18352/ijc.698. no. F2: F2. doi:10.1029/2006JF000531. Fedorov, A. N., P. P. Gavriliev, P. Y. Konstantinov, T. Hiyama, Kanevskiy, M., T. Jorgenson, Y. Shur, J. A. O’Donnell, Y. Iijima, and G. Iwahana. 2014. Estimating the water J. W. Harden, Q. Zhuang, and D. Fortier. 2014. balance of a thermokarst lake in the middle of the Lena Cryostratigraphy and permafrost evolution in the lacustrine River basin, eastern Siberia. Ecohydrology 7, no. 2: 188–96. lowlands of West-Central Alaska. Permafrost and doi:10.1002/eco.1378. Periglacial Processes 25, no. 1: 14–34. doi:10.1002/ppp.1800. Grosse, G., B. Jones, C. Arp. 2013. 8.21 Thermokarst lakes, Lantz, T. C., Y. Zhang, and S. V. Kokelj. 2022. Impacts of drainage, and drained basins. In Treatise on geomorphology, ecological succession and climate warming on permafrost and J. F. Shroder, 325–53. San Diego: Academic Press. aggradation in drained lake basins of the Tuktoyaktuk Himmelstoss, E. A., R. E. Henderson, M. G. Kratzmann, and Coastlands, Northwest Territories, Canada. Permafrost A. S. Farris. 2021. Digital Shoreline Analysis System (DSAS) and Periglacial Processes 33, no. 2: 176–92. doi:10.1002/ version 5.1 user guide. Open-File Report. U.S. Geological ppp.2143. Survey, No. 2021–1091. Lara, M. J., Y. Chen, and B. M. Jones. 2021. Recent warming Hinkel, K. M., R. C. Frohn, F. E. Nelson, W. R. Eisner, and reverses forty-year decline in catastrophic lake drainage and R. A. Beck. 2005. Morphometric and spatial analysis of thaw hastens gradual lake drainage across northern Alaska. lakes and drained thaw lake basins in the western Arctic Environmental Research Letters 16, no. 12: 124019. doi:10. Coastal Plain, Alaska. Permafrost and Periglacial Processes 1088/1748-9326/ac3602. 16, no. 4: 327–41. doi:10.1002/ppp.532. Mackay, J. R. 1988. Catastrophic lake drainage, Tuktoyaktuk Hinkel, K. M., B. M. Jones, W. R. Eisner, C. J. Cuomo, peninsula area, District of Mackenzie. Current Research, R. A. Beck, and R. Frohn. 2007. Methods to assess natural Part D, Geological Survey of Canada, Paper 88:83–90. and anthropogenic thaw lake drainage on the western Mackay, J. R. 1997. A full-scale field experiment (1978–1995) Arctic coastal plain of northern Alaska. Journal of on the growth of permafrost by means of lake drainage, Geophysical Research: Earth Surface 112 (F2). doi:10.1029/ western Arctic coast: A discussion of the method and some 2006JF000584. results. Canadian Journal of Earth Sciences 34, no. 1: 17–33. Jones, B. M. 2023. Geospatial Data Documenting the doi:10.1139/e17-002. Expansion and Drainage of Schaeffer Lake, Northwestern Martin, D., D. Bélanger, P. Gosselin, J. Brazeau, C. Furgal, and Arctic Alaska, 1951-2022. Arctic Data Center. doi:10.18739/ S. Déry. 2007. Drinking water and potential threats to A2930NW7W. human health in Nunavik: Adaptation strategies under Jones, B. M., and C. D. Arp. 2015. Observing a catastrophic climate change conditions. Arctic 60, no. 2: 195–202. thermokarst lake drainage in Northern Alaska. Permafrost Medeiros, A. S., P. Wood, S. D. Wesche, M. Bakaic, and and Periglacial Processes 26, no. 2: 119–28. doi:10.1002/ppp. J. F. Peters. 2017. Water security for northern peoples: 1842. Review of threats to Arctic freshwater systems in Jones, B. M., C. D. Arp, G. Grosse, I. Nitze, M. J. Lara, Nunavut, Canada. Regional Environmental Change 17, no. M. S. Whitman, L. M. Farquharson, et al. 2020. 3: 635–47. doi:10.1007/s10113-016-1084-2. 16 B. M. JONES ET AL. Nitzbon, J., M. Langer, L. C. P. Martin, S. Westermann, Alaska linked to warming and increased shrub habitat. T. Schneider von Deimling, and J. Boike. 2021. Effects of PloS one 11, no. 4: e0152636. doi:10.1371/journal.pone. multi-scale heterogeneity on the simulated evolution of 0152636. ice-rich permafrost lowlands under a warming climate. Tape, K. D., B. M. Jones, C. D. Arp, I. Nitze, and G. Grosse. 2018. The Cryosphere 15, no. 3: 1399–422. doi:10.5194/tc-15- Tundra be dammed: Beaver colonization of the Arctic. Global 1399-2021. Change Biology 24, no. 10: 4478–88. doi:10.1111/gcb.14332. Nitze, I., S. W. Cooley, C. R. Duguay, B. M. Jones, and Tessier, S., B. Forbes, G. Grosse, B. M. Jones, I. Nitze, and G. Grosse. 2020. The catastrophic thermokarst lake drai- J. Schaeffer. 2022. Sudden lake draining event - Local nage events of 2018 in northwestern Alaska: Fast-forward Environmental Observer (LEO) Network [online]. https:// into the future. The Cryosphere 14, no. 12: 4279–97. doi:10. www.leonetwork.org/en/posts/preview/DC4753BF-6270- 5194/tc-14-4279-2020. 46EE-816B-6B97C36B30D3 (accessed August 24, 2022). Nitze, I., G. Grosse, B. M. Jones, V. E. Romanovsky, and Turetsky, M. R., B. W. Abbott, M. C. Jones, K. W. Anthony, J. Boike. 2018. Remote sensing quantifies widespread abun- D. Olefeldt, E. A. G. Schuur, G. Grosse, et al. 2020. Carbon dance of permafrost region disturbances across the Arctic release through abrupt permafrost thaw. Nature Geoscience and Subarctic. Nature Communications 9, no. 1: 5423. 13, no. 2: 138–43. doi:10.1038/s41561-019-0526-0. doi:10.1038/s41467-018-07663-3. Turner, K. W., B. B. Wolfe, and I. McDonald. 2022. Nolan, M., C. Larsen, and M. Sturm. 2015. Mapping snow Monitoring 13 years of drastic catchment change and the depth from manned aircraft on landscape scales at centi- hydroecological responses of a drained thermokarst lake. meter resolution using structure-from-motion photogram- Arctic Science 8, no. 4: 1094–115. metry. The Cryosphere 9, no. 4: 1445–63. Ulrich, M., H. Matthes, L. Schirrmeister, J. Schütze, H. Park, PIX4D. 2022. Pix4dmapper: Release 4.7.5. Prilly, Switzerland: Y. Iijima, and A. N. Fedorov. 2017. Differences in behavior PIX4D S.A. and distribution of permafrost-related lakes in Central Yakutia QTM. 2022. Quick Terrain Modeler: Release 8.3.1. Baltimore, and their response to climatic drivers. Water Resources MD: Applied Imagery, John Hopkins University Applied Research 53, no. 2: 1167–88. doi:10.1002/2016WR019267. Physics Laboratory. Webb, E. E., A. K. Liljedahl, J. A. Cordeiro, M. M. Loranty, Rozell, N. 2022. If a lake drains in northern Alaska . . . [online]. C. Witharana, and J. W. Lichstein. 2022. Permafrost thaw https://uaf.edu/news/if-a-lake-drains-in-northern-alaska. drives surface water decline across lake-rich regions of the php (accessed October 12, 2022). Arctic. Nature Climate Change 12, no. 9: 841–6. doi:10. Rozhkova-Timina, I. O., V. K. Popkov, P. J. Mitchell, and 1038/s41558-022-01455-w. S. N. Kirpotin. 2018. Beavers as ecosystem engineers – White, D. M., S. C. Gerlach, P. Loring, A. C. Tidwell, and A review of their positive and negative effects. IOP M. C. Chambers. 2007 Food and water security in achan- Conference Series: Earth and Environmental Science. ging Arctic climate. Environmental Research Letters. 2, no. 201:012015. 4: 045018. Swanson, D. K. 2019. Thermokarst and precipitation drive Wolfe, S., J. Murton, M. Bateman, and J. Barlow. 2020. changes in the area of lakes and ponds in the National Oriented-lake development in the context of late Parks of northwestern Alaska, 1984–2018. Arctic, Quaternary landscape evolution, McKinley Bay Coastal Antarctic, and Alpine Research 51, no. 1: 265–79. doi:10. Plain, western Arctic Canada. Quaternary Science Reviews 1080/15230430.2019.1629222. 242:106414. doi:10.1016/j.quascirev.2020.106414. Tape, K. D., K. Christie, G. Carroll, and J. A. O’Donnell. 2016. Wolfe, R. J., and R. J. Walker. 1987. Subsistence economies in Novel wildlife in the Arctic: The influence of changing Alaska: Productivity, geography, and development impacts. riparian ecosystems and shrub habitat expansion on snow- Arctic Anthropology 24, no. 2: 56–81. shoe hares. Global Change Biology 22, no. 1: 208–19. doi:10. Wrona, F. J., M. Johansson, J. M. Culp, A. Jenkins, J. Mård, 1111/gcb.13058. I. H. Myers-Smith, T. D. Prowse, W. F. Vincent, and Tape, K. D., J. A. Clark, B. M. Jones, S. Kantner, B. V. Gaglioti, P. A. Wookey. 2016. Transitions in Arctic ecosystems: G. Grosse, and I. Nitze. 2022. Expanding beaver engineering Ecological implications of a changing hydrological regime. creates widespread disturbances in the Arctic. Scientific Journal of Geophysical Research: Biogeosciences 121, no. 3: reports. 650–74. Tape, K. D., D. D. Gustine, R. W. Ruess, L. G. Adams, and Zurowski, W. 1992. Building activity of beavers. Acta J. A. Clark. 2016. Range expansion of moose in Arctic Theriologica 37, no. 4: 403–11. doi:10.4098/AT.arch.92-41.

Journal

Arctic Antarctic and Alpine Research – Taylor & Francis

Published: Dec 31, 2023

Keywords: Arctic; lakes; lake drainage; local observations; permafrost; thermokarst

There are no references for this article.