The Boreal-Arctic Wetland and Lake Dataset (BAWLD)

Abstract

Methane emissions from boreal and arctic wetlands, lakes, and rivers are expected to increase in response to warming and associated permafrost thaw. However, the lack of appropriate land cover datasets for scaling field-measured methane emissions to circumpolar scales has contributed to a large uncertainty for our understanding of present-day and future methane emissions. Here we present the Boreal-Arctic Wetland and Lake Dataset (BAWLD), a land cover dataset based on an expert assessment, extrapolated using random forest modelling from available spatial datasets of climate, topography, soils, permafrost conditions, vegetation, wetlands, and surface water extents and dynamics. In BAWLD, we estimate the fractional coverage of five wetland, seven lake, and three river classes within 0.5 × 0.5° grid cells that cover the northern boreal and tundra biomes (17 % of the global land surface). Land cover classes were defined using criteria that ensured distinct methane emissions among classes, as indicated by a co-developed comprehensive dataset of methane flux observations. In BAWLD, wetlands occupied 3.2 × 106 km2 (14 % of domain) with a 95 % confidence interval between 2.8 and 3.8 × 106 km2. Bog, fen, and permafrost bog were the most abundant wetland classes, covering ~28 % each of the total wetland area, while the highest methane emitting marsh and tundra wetland classes occupied 5 and 12 %, respectively. Lakes, defined to include all lentic open-water ecosystems regardless of size, covered 1.4 × 106 km2 (6 % of domain). Low methane-emitting large lakes (> 10 km2) and glacial lakes jointly represented 78 % of the total lake area, while high-emitting peatland and yedoma lakes covered 18 and 4 %, respectively. Small (

Full text

1
The Boreal-Arctic Wetland and Lake Dataset (BAWLD)
David Olefeldt1, Mikael Hovemyr2, McKenzie A. Kuhn1, David Bastviken3, Theodore J. Bohn4, John
Connolly5, Patrick Crill6, Eugénie S. Euskirchen7,8, Sarah A. Finkelstein9, Hélène Genet8, Guido
Grosse10,11, Lorna I. Harris1, Liam Heffernan12, Manuel Helbig13, Gustaf Hugelius2,14, Ryan
Hutchins15, Sari Juutinen16, Mark J. Lara17,18, Avni Malhotra19, Kristen Manies20, A. David McGuire8,
5
Susan M. Natali21, Jonathan A. O’Donnell22, Frans-Jan W. Parmentier23,24, Aleksi Räsänen25, Christina
Schädel26, Oliver Sonnentag27, Maria Strack28, Suzanne E. Tank29, Claire Treat10, Ruth K. Varner2,30,
Tarmo Virtanen25, Rebecca K. Warren31, Jennifer D. Watts21
1 Department of Renewable Resources, University of Alberta, Edmonton, AB, T6G 2G7, Canada
2 Department of Physical Geography, Stockholm University, 10691 Stockholm, Sweden
10
3 Department of Thematic Studies – Environmental Change, Linköping University, 58183 Linköping, Sweden
4 WattIQ, 400 Oyster Point Blvd. Suite 414, South San Francisco, CA, 94080, USA
5 Department of Geography, School of Natural Sciences, Trinity College Dublin, Dublin 2, Ireland
6 Department of Geological Sciences, Stockholm University, 10691 Stockholm, Sweden
7 Department of Biology and Wildlife, University of Alaska Fairbanks, Fairbanks, AK 99775, USA
15
8 Institute of Arctic Biology, University of Alaska Fairbanks, Fairbanks, AK 99775, USA
9 Department of Earth Sciences, University of Toronto, Toronto, ON, M5S 3B1, Canada
10 Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, Permafrost Research Section, 14473 Potsdam,
Germany
11 Institute of Geosciences, University of Potsdam, 14476 Potsdam, Germany
20
12 Department of Ecology and Genetics, Uppsala University, 752 36 Uppsala, Sweden
13 Department of Physics and Atmospheric Science, Dalhousie University, Halifax, NS, B3H 4R2, Canada
14 Bolin Centre for Climate Research, Stockholm University, 10691 Stockholm, Sweden
15 Department of Earth and Environmental Sciences, University of Waterloo, Waterloo, ON, N2L 3G1, Canada
16 Ecosystems and Environment Research Program, University of Helsinki, FI-00014 Helsinki, Finland
25
17 Department of Plant Biology, University of Illinois, Urbana, IL 61801, USA
18 Department of Geography, University of Illinois, Urbana, IL 61801, USA
19 Department of Earth System Science, Stanford University, Stanford, CA 94305, USA
20 U.S. Geological Survey, Menlo Park, CA, USA
21 Woodwell Climate Research Center, Falmouth, MA 02540, USA
30
22 Arctic Network, National Park Service, Anchorage, AK 99501 USA
23 Centre for Biogeochemistry in the Anthropocene, Department of Geosciences, University of Oslo, 0315 Oslo, Norway
24 Department of Physical Geography and Ecosystem Science, Lund University, 223 62 Lund, Sweden
25 Ecosystems and Environment Research Programme, Faculty of Biological and Environmental Sciences, 00014 University
of Helsinki, Finland
35
26 Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ 86011, USA
27 Département de Géographie, Université de Montréal, Montréal, QC, Canada
28 Department of Geography and Environmental Management, University of Waterloo, Waterloo, ON, N2L 3G1, Canada
29 Department of Biological Sciences, University of Alberta, Edmonton, AB, T6G 2E9, Canada
30 Department of Earth Sciences and Institute for the Study of Earth, Oceans and Space, University of New Hampshire,
40
Durhan, NH 03824, USA
31 National Boreal Program, Ducks Unlimited Canada, Edmonton, AB, T5S 0A2, Canada
Correspondence to: David Olefeldt (olefeldt@ualberta.ca)
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

2
Abstract.
Methane emissions from boreal and arctic wetlands, lakes, and rivers are expected to increase in response to warming and
45
associated permafrost thaw. However, the lack of appropriate land cover datasets for scaling field-measured methane
emissions to circumpolar scales has contributed to a large uncertainty for our understanding of present-day and future
methane emissions. Here we present the Boreal-Arctic Wetland and Lake Dataset (BAWLD), a land cover dataset based on
an expert assessment, extrapolated using random forest modelling from available spatial datasets of climate, topography,
soils, permafrost conditions, vegetation, wetlands, and surface water extents and dynamics. In BAWLD, we estimate the
50
fractional coverage of five wetland, seven lake, and three river classes within 0.5×0.5˚ grid cells that cover the northern
boreal and tundra biomes (17% of the global land surface). Land cover classes were defined using criteria that ensured
distinct methane emissions among classes, as indicated by a co-developed comprehensive dataset of methane flux
observations. In BAWLD, wetlands occupied 3.2×106 km2 (14% of domain) with a 95% confidence interval between 2.8 and
3.8×106 km2. Bog, fen, and permafrost bog were the most abundant wetland classes, covering ~28% each of the total wetland
55
area, while the highest methane emitting marsh and tundra wetland classes occupied 5 and 12%, respectively. Lakes, defined
to include all lentic open-water ecosystems regardless of size, covered 1.4×106 km2 (6% of domain). Low methane-emitting
large lakes (>10 km2) and glacial lakes jointly represented 78% of the total lake area, while high-emitting peatland and
yedoma lakes covered 18 and 4%, respectively. Small (<0.1 km2) glacial, peatland, and yedoma lakes combined covered
17% of the total lake area, but contributed disproportionally to the overall spatial uncertainty of lake area with a 95%
60
confidence interval between 0.15 and 0.38×106 km2. Rivers and streams were estimated to cover 0.12 ×106 km2 (0.5% of
domain) of which 8% was associated with high-methane emitting headwaters that drain organic-rich landscapes. Distinct
combinations of spatially co-occurring wetland and lake classes were identified across the BAWLD domain, allowing for the
mapping of “wetscapes” that will have characteristic methane emission magnitudes and sensitivities to climate change at
regional scales. With BAWLD, we provide a dataset which avoids double-accounting of wetland, lake and river extents, and
65
which includes confidence intervals for each land cover class. As such, BAWLD will be suitable for many hydrological and
biogeochemical modelling and upscaling efforts for the northern Boreal and Arctic region, in particular those aimed at
improving assessments of current and future methane emissions. Data is freely available at
https://doi.org/10.18739/A2C824F9X (Olefeldt et al., 2021).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

3
1 Introduction
70
Emissions of methane (CH4) from abundant wetlands, lakes, and rivers located in boreal and arctic regions are expected to
substantially increase this century due to rapid climate warming and associated permafrost thaw (Walter Anthony et al.,
2018; Ito, 2019; Hugelius et al., 2020; Schneider von Deimling et al., 2015; Zhang et al., 2017). However, predicting future
CH4 emissions is highly uncertain, as estimates of present-day CH4 emissions from boreal and arctic regions are poorly
constrained, ranging between 21 and 77 Tg CH4 yr-1 (Saunois et al., 2020; Peltola et al., 2019; Wik et al., 2016; Treat et al.,
75
2018; McGuire et al., 2012; Watts et al., 2014; Thompson et al., 2018; Zhu et al., 2015; Tan et al., 2016; Walter Anthony et
al., 2016). Estimates of high-latitude CH4 emissions vary between approaches, with generally lower estimates from
atmospheric inversions (top-down estimates), than from field-measured CH4 emissions data paired with land cover data
(bottom-up estimates) (Saunois et al., 2020; McGuire et al., 2012). Low accuracy of high-latitude land cover datasets for
wetland and lake distributions, and their classification, represent key sources of uncertainty for estimates of high-latitude
80
CH4 emissions and may contribute to the discrepancies between bottom-up and top-down estimates. A limitation of many
currently available land cover datasets is an insufficient differentiation between wetland, lake, and river classes that are
known to have distinct CH4 emissions (Bruhwiler et al., 2021; Bohn et al., 2015; Marushchak et al., 2016; Melton et al.,
2013).
85
There are several challenges when using remote sensing approaches to map distinct wetland, lake, and river classes at the
circumpolar scale. Many small or narrow wetland ecosystems with high methane CH4 emissions are located along lake
shorelines, along stream networks, or in polygonal tundra terrain, and are thus difficult to map as image resolution can be
inadequate (Wickland et al., 2020; Cooley et al., 2017; Virtanen and Ek, 2014; Liljedahl et al., 2016). Wetland detection can
further be complicated by the presence of tree species in wetlands, e.g. Scots pine (Pinus sylvestris), black spruce (Picea
90
mariana), and tamarack (Larix laricina), that are also found in non-wetland boreal forests, making differentiation of treed
wetlands from non-wetland forests difficult. Using spectral signatures to differentiate and map distinct wetland classes can
further be difficult due to seasonal variation in inundation or phenology, poor differentiation between ecosystems (e.g.
similarities between different peatland classes), or high spectral diversity within classes due to shifts in vegetation along
subtle environmental gradients (Räsänen and Virtanen, 2019; Vitt and Chee, 1990; Chasmer et al., 2020). Vegetation
95
composition and spectral signatures of wetland classes can also vary between different high-latitude regions, e.g. with shifts
in dominant tree and shrub species between North America and Eurasia (Raynolds et al., 2019), and be influenced for
decades by wildfires (Chen et al., 2021; Helbig et al., 2016). Active microwave remote sensing can help detect inundated
wetlands and saturated soils, but has limitations due to its computational requirements, coarse resolution, and issues with
detecting rarely inundated peatlands (Beck et al., 2021; Duncan et al., 2020). Accurate mapping of wetlands that include
100
differentiation among distinct wetland classes requires substantial ground truthing, something which has only been done
consistently at local and regional scales (Terentieva et al., 2016; Chasmer et al., 2020; Bryn et al., 2018; Lara et al., 2018;
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

4
Canadian Wetland Inventory Technical Committee, 2016). Similar issues arise for lakes, rivers, and streams. While larger
lakes and rivers have been mapped with high precision (Messager et al., 2016; Linke et al., 2019), the highest CH4 emissions
are generally from ponds, pools, and low-order streams that are too small to be accurately detected by anything other than
105
very high-resolution imagery (Muster et al., 2017). Statistical approaches are often used to model the distribution and
abundance of small open-water ecosystems, yielding large uncertainties (Holgerson and Raymond, 2016; Cael and Seekell,
2016; Muster et al., 2019). Remote sensing approaches are also inadequate in assessing other key variables known to
influence lake CH4 emissions, including lake genesis, depth, and sediment characteristics (Messager et al., 2016; Brosius et
al., 2021; Smith et al., 2007; Lara et al., 2021). Another key issue is that wetlands and lakes often are mapped separately,
110
allowing for potential double-counting of ecosystems in both wetland and lake inventories (Thornton et al., 2016; Saunois et
al., 2020).
Emissions of CH4 from boreal and arctic ecosystems range from uptake to some of the highest emissions observed globally
(Turetsky et al., 2014; Knox et al., 2019; Glagolev et al., 2011; St Pierre et al., 2019). Net ecosystem CH4 emissions are a
115
balance between microbial CH4 production (methanogenesis) and oxidation (methanotrophy), a balance further influenced
by the dominant transport pathway; diffusion, ebullition, and plant-mediated transport (Bridgham et al., 2013; Bastviken et
al., 2004). For wetlands, defined as ecosystems with temporally or permanently saturated soils and biota adapted to anoxic
conditions, CH4 emissions in boreal and arctic regions are primarily influenced by water table position, soil temperatures,
and vegetation composition and productivity (Olefeldt et al., 2013; Treat et al., 2018). Marshes and tundra wetlands are
120
characterized by frequent or permanent inundation and dominant graminoid vegetation that enhance methanogenesis and
facilitates plant-mediated transport, and thus generally have high CH4 emissions (Knoblauch et al., 2015; Juutinen et al.,
2003). Conversely, peat-forming bogs and fens generally have a water table at or below the soil surface, less graminoid
vegetation and instead vegetation dominated by mosses, lichens, and shrubs, resulting in typically low to moderate CH4
emissions (Bubier et al., 1995; Pelletier et al., 2007). Permafrost conditions in peatlands can cause the surface to be elevated
125
and dry, with cold soil conditions where methanogenesis is inhibited, leading to low CH4 emissions or even uptake
(Bäckstrand et al., 2008; Glagolev et al., 2011). Non-wetland boreal forests and tundra ecosystems generally have net CH4
uptake, as methanotrophy outweighs any methanogenesis (Lau et al., 2015; Juncher Jørgensen et al., 2015; Whalen et al.,
1992). The transition from terrestrial to aquatic ecosystems is not always well defined, and several wetland classification
systems consider shallow, open-water ecosystems as a distinct wetland class (Rubec, 2018). The transition from vegetated to
130
open water ecosystems is however associated with shifts in apparent primary controls of CH4 emissions, including a shift
towards increased importance of ebullition (Bastviken et al., 2004). For lakes, when defined to include all lentic open-water
ecosystems regardless of size (e.g. including peatland ponds), spatial variability in CH4 emissions is primarily linked to
water depth and the quantity and origin of the organic matter of the sediment (Heslop et al., 2020; Li et al., 2020). As such,
lake CH4 emissions are generally higher for smaller lakes and for lakes with organic-rich sediments (Wik et al., 2016;
135
Holgerson and Raymond, 2016), which are extremely abundant in many high-latitude regions (Muster et al., 2017). The CH4
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

5
emitted from streams and rivers is largely derived from the soils that are drained, and as such emissions generally are higher
in smaller streams draining wetland-rich watersheds (Wallin et al., 2018; Stanley et al., 2016). It is overall likely that studies
of CH4 emissions from boreal and arctic ecosystems have focused disproportionally on sites with higher CH4 emissions
(Olefeldt et al. 2013). A focus on high-emitting sites is warranted for understanding site-level controls on CH4 emissions, but
140
may potentially cause bias of bottom-up CH4 scaling approaches if they lack appropriate differentiation between various
wetland and lakes classes in the used land cover datasets.
There is currently no spatial dataset available that has information on the distribution and abundance of wetland, lake, and
river classes defined specifically for the purpose of estimating boreal and arctic CH4 emissions. However, a large number of
145
spatial datasets have partial, but relevant, information. This includes circumpolar spatial data of soil types (Hugelius et al.,
2013; Strauss et al., 2017), vegetation (Olson et al., 2001; Walker et al., 2005), surface water extent and dynamics (Pekel et
al., 2016), lake sizes and numbers (Messager et al., 2016), topography (Gruber, 2012), climate (Fick and Hijmans, 2017),
permafrost conditions (Gruber, 2012; Brown et al., 2002), river networks (Linke et al., 2019), and previous estimates of total
wetland cover (Matthews and Fung, 1987; Bartholomé and Belward, 2005). By integrating quantitative spatial data with
150
expert knowledge it is possible to model new spatial data for specific purposes (Olefeldt et al., 2016). Researchers with
interests in the boreal and arctic have considerable knowledge of the presence and relative abundance of typical wetland and
lake classes in various high-latitude regions, along with the ability for satellite image interpretation and the judgement to
define parsimonious land cover classes suitable for CH4 scaling.
155
Here we present the Boreal-Arctic Wetland and Lake Dataset (BAWLD), an expert knowledge-based land cover dataset.
Developed in concert with a comprehensive dataset of observed CH4 fluxes from high-latitude aquatic ecosystems (Kuhn et
al., 2021), BAWLD includes five wetland, seven lake, and three river classes with distinct CH4 emissions. Coverage of each
wetland, lake, and river class within 0.5° grid cells was modelled through random forest regressions based on expert
assessment data and available relevant spatial data. The approach aims to reduce issues with bias in representativeness of
160
empirical data, to reduce issues of overlaps in wetland and lake extents, and to allow for the partitioning of uncertainty of
CH4 emissions estimates to CH4 emission magnitudes or areal extents of different land cover classes. As such, BAWLD will
facilitate improved bottom-up estimates of high-latitude CH4 emissions, and will be suitable for use in process-based models
and as an a-priori input to inverse modelling approaches. The land cover dataset will be suitable for further uses, especially
for questions related to high-latitude hydrology and biogeochemistry. Lastly, BAWLD allows for the definition of
165
“wetscapes”; regions with distinct co-occurrences of specific wetland and lake classes, and which thus can be used to
understand regional responses to climate change and as a way to visualize the landscape diversity of the boreal and arctic
domain.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

6
Table 1. Description of data sources and layers extracted into the BAWLD 0.5° grid cell network.
170
Dataset and extracted layers
Dataset and extracted layers
WorldClim V2 (Fick and Hijmans, 2017)
- WC2-MAAT: Mean annual average air temperature 1970-2000. (˚C)
- WC2-MAAP: Mean annual average precipitation 1970-2000 (mm)
- WC2-CMI: Climate moisture index 1970-2000 (mm)
Reference information
- LAT: Latitude (˚)
- LONG: Longitude (˚)
- SHORE: Coastal shoreline presence in cell (yes/no)
Circum-Arctic Map of Permafrost and Ground-Ice (Brown et al., 2002)
- CAPG-CON: Continuous permafrost. (%)
- CAPG-DIS: Discontinuous permafrost. (%)
- CAPG-SPO: Sporadic permafrost. (%)
- CAPG-ISO: Isolated permafrost. (%)
- CAPG-XHF: Land with thick overburden and >20% ground-ice. (%)
- CAPG-XMF: Land with thick overburden and 10-20% ground-ice. (%)
- CAPG-XLF: Land with thick overburden and <10% ground-ice. (%)
- CAPG-XHR: Land with thin overburden and >10% ground-ice. (%).
- CAPG-XLR: Land with thin overburden and <10% ground-ice. (%)
- CAPG-REL: Land with relict permafrost. (%)
Northern Circumpolar Soil Carbon Dataset (Hugelius et al., 2014)
- NCS-HSO: Histosol soils; non-permafrost organic soils. (%)
- NCS-HSE Histel soils: permafrost organic soils. (%)
- NCS-AQU: Aqueous soils: non-organic wetland soils. (%)
- NCS-ROC: Rocklands. (%)
- NCS-GLA: Glaciers. (%)
- NCS-H2O: Open water. (%)
BasinATLAS (Linke et al., 2019)
- BAS-RIV: River area. (%)
Global Lakes and Wetland Dataset (Lehner and Döll, 2004)
- GLWD-RIV: Rivers, 6th order rivers or greater. (%)
Circumpolar Arctic Vegetation Map (CAVM Team 2003)
- CAVM-BAR: Barren Tundra. (%)
- CAVM-GRA: Graminoid Tundra. (%)
- CAVM-SHR: Shrubby Tundra. (%)
- CAVM-WET: Wet Tundra. (%)
Terrestrial Ecoregions of the World (Olson et al., 2001)
- TEW-BOR: Fractional cover of boreal ecoregion/ (%)
- TEW-TUN: Fractional cover of tundra ecoregion. (%)
- TEW-GLA: Fractional cover of glaciers. (%)
HydroLakes (Messager et al., 2016)
- HL-LAR: Lakes >10 km2. (%)
- HL-MID: Lakes between 10 km2 and 0.1 km2. (%)
- HL-SHO: Shoreline density (length/area) of lakes >0.1 km2. (m/m2)
Global Land Cover Database 2000 (Bartholomé and Belward, 2005)
- GLC2-H2O: Water Bodies, natural and artificial. (%)
- GLC2-RFSM: Regularly Flooded Shrub and/or Herbaceous Cover. (%)
- GLC2-FOR: Forest cover. (%)
Global Inundation Map (Fluet-Chouinard et al., 2015)
- GIM-MAMI: Mean annual minimum inundation. (%)
- GIM-MAMA: Mean annual maximum inundation. (%)
Dataset of Ice-Rich Yedoma Permafrost (Strauss et al., 2017)
- IRYP-YED: Yedoma ground. (%)
GlobLand30 (Chen et al., 2015)
- GL30-H2O: Water bodies: including lakes, rivers, reservoirs. (%)
- GL30-WET: Wetlands: marshes, floodplains, shrub wetland, peatlands. (%)
- GL30-TUN: Tundra: shrub, herbaceous, wet, and barren tundra. (%)
- GL30-ART: Artificial Surfaces: cities, industry, transport. (%)
- GL30-ICE: Permanent snow and ice. (%)
Permafrost zonation and Terrain Ruggedness Index (Gruber, 2012)
- PZI-PERM: Permafrost ground. (%)
- PZI-FLAT: Flat topography. (%)
- PZI-UND: Undulating topography. (%)
- PZI-HILL: Hilly topography. (%)
- PZI-MTN: Mountainous topography. (%)
- PZI-RUG: Rugged topography. (%)
Global Surface Water (Pekel et al., 2016)
- GSW-RAR: Rarely inundated; open water in 0 to 5% of occasions. (%)
- GSW-OCC: Occasionally inundated; open water 5 to 50%. (%)
- GSW-REG: Regularly inundated; open water 50 to 95%. (%)
- GSW-PER: Permanent open water; open water 95 to 100%. (%)
Global Wetlands (Matthews and Fung, 1987)
-GWET-IN: Inundation and presence of wetlands. (%).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

7
2 Development of the Boreal-Arctic Wetland and Lake Dataset
2.1 Study domain and harmonization of available spatial data
The BAWLD domain includes all of the northern boreal and tundra ecoregions, and also areas of rock and ice at latitudes
>50°N (Olson et al., 2001). The BAWLD domain thus covers 25.5×106 km2, or 17% of the global land surface. Although
175
northern peat-forming wetlands can also be found in temperate ecoregions, our decision to define the southern limit of
BAWLD by the transition from boreal to temperate ecoregions was based on the greater human footprint and the increased
biogeographic diversity of temperate ecoregions, which would require additional land cover classes (Venter et al., 2016). A
network of 0.5° grid cells, cropped along coasts and at the transition from boreal to temperate ecoregions, was created for the
BAWLD domain.
180
Grid cells in BAWLD were populated with data from 15 publicly available spatial datasets, yielding 53 variables with spatial
information (Table 1). Most datasets that were included have data at higher resolution than the 0.5° BAWLD grid cells,
hence information was averaged for each grid cell. For datasets where the spatial resolution was coarser or where spatial data
were not aligned with the 0.5° grid cells, data was first apportioned into BAWLD grid cells before area-weighted averages
185
were calculated. Climate data from the WorldClim2 (WC2) dataset (Fick and Hijmans, 2017) were averaged for each grid
cell, including “mean annual air temperature”, “mean annual precipitation”, and “climate moisture index”. Information on
soils and permafrost conditions were summarized as fractional coverage within each grid cell, and included “permafrost
extent” from the Permafrost Zonation and Terrain Ruggedness Index (PZI) dataset (Gruber, 2012), permafrost zonation,
ground ice content, and overburden thickness from the Circum-Arctic Map of Permafrost and Ground-Ice (CAPG) dataset
190
(Brown et al., 2002), “yedoma ground” from the Ice-Rich Yedoma Permafrost (IRYP) dataset (Strauss et al., 2017), and non-
permafrost peat “histosol”, permafrost peat “histel”, and “aqueous” wetland soils from the Northern Circumpolar Soil
Carbon Database (NCSCD; hereafter NCS) (Hugelius et al., 2013). Four independent datasets provided information on
wetland coverage, although without further differentiation between distinct wetland classes; the “regularly flooded shrub
and/or herbaceous cover” area from the Global Land Cover Database 2000 (GLC2) (Bartholomé and Belward, 2005), the
195
“wetlands” area in the GlobLand30 (GL30) dataset (Chen et al., 2015), and the “inundation and presence of wetlands” area
from the Global Wetlands (GWET) dataset (Matthews and Fung, 1987), and the Circumpolar Arctic Vegetation Map
(CAVM) dataset (Walker et al., 2005). Two datasets provided information of the extent of forested regions; the GLC2 and
the Terrestrial Ecoregions of the World (TEW) dataset (Olson et al., 2001), while three datasets provided information on the
extents of tundra vegetation; the CAVM, the GL30, and the TEW. Three datasets provided information on extent of glaciers
200
and permanent snow; the NCS, the GL30, and the TEW. The NCS dataset also provided information about the extents of
“rocklands”, while the PZI dataset had extents of topographic ruggedness (“flat”, “undulating”, “hilly”, “mountainous”, and
“rugged”). Information on river extents was found in two datasets; the “river area” in the BasinATLAS (BAS) dataset (Linke
et al., 2019), and “rivers” in the Global Lakes and Wetland (GLW) dataset, which includes 6th order rivers and greater
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

8
(Lehner and Döll, 2004). Inundation dynamics was provided by two datasets, with “mean annual minimum” and “mean
205
annual maximum” inundation in the Global Inundation Map (GIM) dataset (Fluet-Chouinard et al., 2015), and an analysis of
temporal inundation from the Global Surface Water (GSW) dataset (Pekel et al., 2016) where we defined inundation of
individual 30 m pixels as being inundated “rarely” (>0 to 5% of all available Landsat images), “occasionally” (5 to 50%),
“regularly” (50 to 95%), or “permanently” (95 to 100%). Four datasets included information about static extents of open
water, including “open water” in NCS, “water bodies” in GL30, “water bodies” in GLC2, and information about lakes in the
210
Hydrolakes (HL) dataset (Messager et al., 2016), where we differentiated between the area of “large lakes” (lakes >10 km2),
and “midsize lakes” (lakes between 0.1 and 10 km2).
2.2 Land cover classes in BAWLD
The land cover classification in BAWLD was constructed with the goal to enable upscaling of CH4 fluxes for large spatial
extents. As such, we aimed to include as few classes as possible to facilitate for large-scale mapping, while still including
215
classes that allow for separation among ecosystems with distinct hydrology, ecology, biogeochemistry and thus net CH4
fluxes. The BAWLD land cover classification is hierarchical; with five wetland classes, seven lake classes, and three river
classes, along with four other classes; glaciers, dry tundra, boreal forest, and rocklands. The class descriptions below were
provided to all experts for their land cover assessments, and thus effectively serve as the BAWLD class definitions.
2.2.1 Wetland Classes
220
Wetlands are defined by having a water table near or above the land surface for sufficient time to cause the development of
wetland soils (either mineral soils with redoximorphic features, or organic soils with > 40 cm peat), and the presence of plant
species with adaptations to wet environments (Hugelius et al., 2020; Canada Committee on Ecological (Biophysical) Land
Classification et al., 1997; Jorgenson et al., 2001). Wetland classifications for boreal and arctic biomes can focus either on
small-scale wetland classes that have distinct hydrological regimes, vegetation composition, and biogeochemistry, or on
225
larger-scale wetland complexes that are comprised of distinct patterns of smaller wetland and open-water classes
(Gunnarsson et al., 2014; Terentieva et al., 2016; Masing et al., 2010; Glaser et al., 2004). While larger-scale wetland
complexes are easier to identify through remote sensing techniques (e.g. patterned fens comprised of higher elevation ridges
and inundated hollows), our classification focuses on wetland classes due to greater homogeneity of hydrological, ecological,
and biogeochemical characteristics that regulate CH4 fluxes (Heiskanen et al., 2021).
230
Several boreal countries identify four main wetland classes, differentiated primarily based on hydrodynamic
characterization; bogs, fens, marshes, and swamps (Gunnarsson et al., 2014; Canada Committee on Ecological (Biophysical)
Land Classification et al., 1997; Masing et al., 2010). The BAWLD classification follows this general framework, but further
uses the presence or absence of permafrost as a primary characteristic for classification and excludes a distinct swamp class,
235
yielding five classes; Bogs, Fens, Marshes, Permafrost Bogs, and Tundra Wetlands (Figure 1). The swamp class was omitted
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

9
due to the wide range of moisture and nutrient conditions of swamps, as well as the limited number of studies of swamp CH4
fluxes (Kuhn et al., 2021). We instead included swamp ecosystems in expanded descriptions of Bogs, Fens, and Marshes.
The presence or absence of near-surface permafrost was used as a primary characteristic to distinguish between Permafrost
Bogs and Bogs, and to distinguish Tundra Wetlands from Marshes and Fens. The presence or absence of near-surface
240
permafrost is considered key for controlling CH4 emissions given its influence on hydrology, and for the potential of
permafrost thaw and thermokarst collapse to cause rapid non-linear shifts to CH4 emissions (Bubier et al., 1995; Turetsky et
al., 2002; Malhotra and Roulet, 2015). Finally, while some classifications include shallow (e.g. 2 m depth), open-water
ecosystems within the definition of wetlands (Gunnarsson et al., 2014; Canada Committee on Ecological (Biophysical) Land
Classification et al., 1997), we have included all open-water ecosystems without emergent vegetation within the lake classes
245
(see below) due to the strong influence of emergent vegetation in controlling CH4 emissions (Juutinen et al., 2003).
Figure 1. Descriptions of wetland classes in BAWLD as distinguished based on the moisture regime, the nutrient/pH regime,
hydrodynamics, and the presence/absence of permafrost.
250
Bogs are described as ombrotrophic peatland ecosystems, i.e. only dependent on precipitation, and snowmelt for water
inputs. Peat thickness is at least 40 cm, with maximum thickness > 10 m. The peat profile is not affected by permafrost,
although in some climatically colder settings there may be permafrost below the peat profile. Bogs are wet to saturated
ecosystems, often with small-scale (<10 m) microtopographic variability, with stagnant water and a water table that rarely is
255
above the surface or more than 50 cm below the surface (Figure 1). Bogs have low pH (<5), low concentrations of dissolved
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

10
ions, and low nutrient availability resulting from a lack of hydrological connectivity to surrounding mineral soils. Vegetation
is commonly dominated by Sphagnum mosses, lichens, and woody shrubs, and can be either treed or treeless (Beaulne et al.,
2021). Our description of Bogs also includes what is commonly classified as treed swamps, which generally represent
ecotonal transitions between peatlands and upland forests (Canada Committee on Ecological (Biophysical) Land
260
Classification et al., 1997). Emissions of CH4 from Bogs are low to moderate in comparison with other classes (Kuhn et al.,
2021).
Fens are described as minerotrophic peatland ecosystems, i.e. hydrologically connected to surrounding mineral soils through
surface water or groundwater inputs. A Fen peat profile is at least 40 cm thick (Gorham et al. 1991), although maximum
265
thickness is generally less than for bogs. The peat profile is not affected by permafrost. Fens are wet to saturated ecosystems,
with generally slow-moving water (Figure 1). Fens have widely ranging nutrient regimes and levels of dissolved ions
depending on the degree and type of hydrological connectivity to their surroundings, ranging from poor fens to rich fens.
Vegetation largely depends on wetness and nutrient availability, where more nutrient poor fens can have Sphagnum mosses,
shrubs, and trees, while rich fens are dominated by brown mosses, graminoids (sedges, rushes), herbaceous plants, and
270
sometimes coniferous or deciduous trees (e.g. willows, birch, larch). Our description of Fens also includes what is
commonly classified as shrubby swamps, which often are associated with riparian ecotones and lake shorelines. Fen CH4
emissions are moderate to high in comparison to other classes (Kuhn et al., 2021).
Marshes are minerotrophic wetlands with dynamic hydrology, and often high nutrient availability (Figure 1). Vegetation is
275
dominated by emergent macrophytes, including tall graminoids such as rushes, reeds, grasses and sedges – some of which
can persist in settings with >1.5 m of standing water. Marshes are saturated to inundated wetlands, often with highly
fluctuating water levels as they generally are located along shorelines of lake or coasts, along streams and rivers, or on
floodplains and deltas. It is common for marshes to exhibit both flooded and dry periods. Dry periods facilitate
decomposition of organic matter, and can prevent the build-up of peat. As such Marshes generally have mineral soils,
280
although some settings allow for the accumulation of highly humified organic layers – sometimes indicating ongoing
succession towards a peatland ecosystem. Salinity can vary depending on water sources, with brackish to saline conditions in
some areas of groundwater discharge, or in coastal settings. While highly saline marshes can have low CH4 emissions due to
inhibition of methanogenesis, Marshes generally have high to very high CH4 emissions compared to other classes (Kuhn et
al., 2021).
285
Permafrost Bogs are peatland ecosystems, although the peat thickness in cold climates is often relatively shallow.
Permafrost Bogs have a seasonally thawed active layer that is 30 to 70 cm thick, with the remainder of the peat profile
perennially frozen (i.e. permafrost). Excess ground-ice and ice expansion often elevate Permafrost Bogs up to a few meters
above their surroundings, and as such they are ombrotrophic and generally the wetland class with the driest soils (Figure 1).
290
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

11
Permafrost Bogs have moist to wet soil conditions, often with a water table that follows the base of the seasonally
developing thawed soil layer. Ombrotrophic conditions cause nutrient-poor conditions, and the vegetation is dominated by
lichens, Sphagnum mosses, woody shrubs, and sometimes stunted coniferous trees. Permafrost Bogs are often interspersed in
a fine-scale mosaic (10 to 100 m) with other wetland classes, e.g. Bogs and Fens. Common Permafrost Bog landforms
include palsas, peat plateaus, and the elevated portions of high- and low-center polygonal peatlands. Permafrost Bogs have
295
very low CH4 emissions, and sometimes even CH4 uptake (Kuhn et al., 2021).
Tundra Wetlands are treeless ecosystems with saturated to inundated conditions, most commonly with near surface
permafrost (Figure 1). Tundra Wetlands can have either mineral soils or shallow organic soils, and generally receive surface
or near-surface waters from their surroundings, as permafrost conditions preclude connectivity to deeper groundwater
300
sources. Vegetation is dominated by short emergent vegetation, including sedges and grasses, with mosses and shrubs in
slightly drier sites. Tundra Wetlands have lower maximum depth of standing water than Marshes, due to the shorter
vegetation. Tundra Wetlands can be found in basin depressions, in low-center polygonal wetlands, and along rivers, deltas,
lake shorelines, and on floodplains in regions of continuous permafrost. Despite the name, limited wetlands with these
characteristics (hydrology, permafrost conditions, and vegetation) can also be found within the continuous permafrost zone
305
in boreal and sub-arctic regions (Virtanen et al., 2016). Tundra Wetlands have moderate to high CH4 emissions compared to
other classes (Kuhn et al., 2021).
2.2.2 Lake Classes
Lakes in BAWLD are considered to include all lentic open-water ecosystems, regardless of surface area and depth of
standing water. It is common in ice-rich permafrost lowlands and peatlands for open-water bodies to have shallow depths,
310
often less than two meters, even when surface areas are up to hundreds of km2 in size (Grosse et al., 2013). While small,
shallow open-water bodies often are included in definitions of wetlands (Canada Committee on Ecological (Biophysical)
Land Classification et al., 1997; Gunnarsson et al., 2014; Treat et al., 2018), we include them here within the lake classes as
controls on net CH4 emissions depend strongly on the presence or absence of emergent macrophytes (Juutinen et al., 2003).
Further classification of lakes in BAWLD is based lake size and lake genesis where lake genesis influences lake bathymetry
315
and sediment characteristics. Previous global spatial inventories of lakes include detailed information on size and location of
individual larger lakes (Messager et al., 2016; Downing et al., 2012), but do not include open-water ecosystems <0.1 km2 in
size, and do not differentiate between lakes of different genesis (e.g. tectonic, glacial, organic, and yedoma lakes). Small
water bodies are disproportionately abundant in some high latitude environments (Muster et al., 2019), have high emissions
of CH4 (Holgerson and Raymond, 2016), and therefore require explicit classification apart from larger water-bodies.
320
Furthermore, lake genesis and sediment type haven been shown to influence net CH4 flux from lakes (Wik et al., 2016). In
BAWLD we thus differentiate between large (>10 km2), midsize (0.1 to 10 km2) and small (<0.1 km2) lake classes, and
further differentiate between three lake types for midsize and small lakes; peatland, yedoma, and glacial lakes.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

12
Small and Midsize Peatland Lakes are described as lakes with thick organic sediments that are mainly found adjacent to or
325
surrounded by peatlands, or in lowland tundra regions with organic-rich soils. Small Peatland Lakes includes the numerous
small pools often found in extensive peatlands and lowland tundra regions, e.g. including the open-water parts of string fens
and polygonal peatlands. Peatland Lakes generally form as a result of interactions between local hydrology and the
accumulation of peat which can create open water pools and lakes (Garneau et al., 2018; Harris et al., 2020), but can also
form in peatlands as a result of permafrost dynamics (Liljedahl et al., 2016; Sannel and Kuhry, 2011). As such, these lakes
330
with thick organic sediments are often shallow and have a relatively low shoreline development index. Peatland lakes have
dark waters with high concentrations of dissolved organic carbon. Small Peatland Lakes generally have higher CH4
emissions than Midsize Peatland Lakes, but observations from both classes range from moderate to very high CH4
emissions, with roughly half of emissions attributed to ebullition (Kuhn et al., 2021).
335
Small and Midsize Yedoma Lakes are exclusive to non-glaciated regions of eastern Siberia, Alaska, and the Yukon where
yedoma deposits accumulated during the Pleistocene (Strauss et al., 2017). Yedoma permafrost soils are ice-rich and contain
fine-grained, organic-rich loess which was deposited by wind and accumulated upwards in parallel with permafrost
aggradation, thus limiting decomposition and facilitating organic matter burial (Schirrmeister et al., 2013). Notable
thermokarst features, including lakes, often develop when yedoma permafrost thaws, causing labile organic matter to
340
become available for microbial mineralization (Walter Anthony et al., 2016). Small Yedoma lakes typically represent
younger thermokarst features, whereas Midsized Yedoma Lakes represent later stages of thermokarst lake development.
Small Yedoma Lakes are thus more likely to have actively thawing and expanding lake edges where CH4 emissions can be
extremely high, largely driven by high ebullition emissions (Walter Anthony et al., 2016). Century-scale development of
yedoma lakes can shift the main source of CH4 production from yedoma deposits to new organic-rich sediment that
345
accumulate from allochtonous and autochthonous sources – resulting in such lakes here being considered as Peatland Lakes.
Midsize Yedoma Lakes have lower CH4 emissions than Midsize Peatland Lakes, while Small Yedoma Lakes have similar
CH4 emissions as Small Peatland Lakes – albeit with a greater proportion attributed to ebullition (Kuhn et al., 2021).
Small and Midsize Glacial Lakes include all lakes with organic-poor sediments – predominately those formed through glacial
350
or post-glacial processes, e.g. kettle lakes and bedrock depressions. However, due to similarities in CH4 emissions and
controls thereof, we also include all other lakes with organic-poor sediments within these classes. Glacial Lakes typically
have rocky bottoms or mineral sediments with limited organic content. Lakes in this class are abundant on the Canadian
Shield and in Fennoscandia, but can be found throughout the boreal and tundra biomes. Many Glacial Lakes have high
shoreline development index, with irregular, elongated shapes. Generally, Glacial Lakes are deeper than lakes in the other
355
classes, when comparing lakes with similar lake area. Glacial Lakes have very low to moderate CH4 emissions, with slightly
greater emissions from Small than Midsize Glacial Lakes, driven by greater ebullitive emissions (Kuhn et al., 2021).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

13
Large Lakes are greater than 10 km2 in surface area. Most Large Lakes are glacial or structural/tectonic in origin. Lake
genesis is not considered for further differentiation within this land cover class. Emissions of CH4 from Large Lakes are very
360
low to low (Kuhn et al., 2021).
2.2.3 River Classes
We include three river classes in BAWLD, Large Rivers, Small Organic-Rich Rivers, and Small Organic-Poor Rivers. Large
Rivers are described as 6th Strahler order rivers or greater, and generally have river widths >~75 m (Downing et al., 2012;
Lehner and Döll, 2004). Small Organic-Rich Rivers include all 1st to 5th order streams and rivers that drain peatlands or other
365
wetland soils, thus being associated with high concentrations of dissolved organic carbon and high supersaturation of CH4.
Conversely, Small Organic-Poor Rivers drain regions with less wetlands and organic-rich soils, and generally have lower
concentrations of dissolved organic carbon and dissolved CH4.
2.2.4 Other Classes
Four additional classes are included in BAWLD; Glaciers, Rocklands, Dry Tundra, and Boreal Forests. Glaciers include
370
both glaciers and other permanent snow and ice on land. Rocklands include areas with very poor soil formation and where
vegetation is largely absent. Rocky outcrops in shield landscapes, slopes of mountains, and high Arctic barren landscapes are
included in the class. The Rocklands class also includes artificial surfaces such as roads and towns. Glaciers and Rocklands
are largely considered to be neutral with respect to CH4 emissions. The Dry Tundra class includes both lowland arctic tundra
and alpine tundra; both treeless ecosystems dominated by graminoid or shrub vegetation. Dry Tundra ecosystems generally
375
have near-surface permafrost, with seasonally thawed active layers between 20 and 150 cm depending on climate, soil
texture, and landscape position (van der Molen et al., 2007; Heikkinen et al., 2004). Near-surface permafrost in Dry Tundra
prevents vertical drainage, but lateral drainage ensures predominately oxic soil conditions. A water table is either absent or
close to the base of the seasonally thawing active layer. Dry Tundra is differentiated from Permafrost Bogs by having
thinner organic soil (<40 cm), and from Tundra Wetlands by their drained soils (average water table position >5 cm below
380
soil surface). Dry Tundra generally have net CH4 uptake, but very low CH4 emissions are also common (Kuhn et al., 2021).
Boreal Forests are treed ecosystems with non-wetland soils. Coniferous trees are dominant, but the class also includes
deciduous trees in warmer climates and landscape positions. Boreal Forests may have permafrost or non-permafrost ground,
where absence of permafrost often allow for better drainage. Overall, it is rare for anoxic conditions to occur in Boreal
Forest soils, and CH4 uptake is prevalent, although low CH4 emissions have been observed during brief periods during
385
snowmelt or following summer storms (Matson et al., 2009), or conveyed through tree stems and shoots (Machacova et al.,
2016). The Boreal Forest class also includes the few agricultural/pasture ecosystems within the boreal biome.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

14
2.3 Expert assessment
Expert assessments can be used to inform various environmental assessments, and are particularly useful to assess levels of
uncertainty and to provide data that cannot be obtained through other means (Olefeldt et al., 2016; Loisel et al., 2021; Abbott
390
et al., 2016; Sayedi et al., 2020). We solicited an expert assessment to aid in the modelling of fractional coverage of the 19
land cover classes within each BAWLD grid cell. Researchers associated with the Permafrost Carbon Network
(www.permafrostcarbon.org) with expertise from wetland, lake, and/or river ecosystems within the BAWLD domain were
invited to participate. We also included a few additional referrals to suitable experts outside the Permafrost Carbon Network.
A total of 29 researchers completed the expert assessment, and are included as co-authors of the BAWLD dataset. Each
395
expert was asked to identify a region within the BAWLD domain for which they considered themselves familiar. Experts
were then assigned 10 random cells from their region of familiarity and 10 cells distributed across the BAWLD domain that
allowed for an overall balanced distribution of training cells (Figure S1). No cell was assessed more than once, and in total
~3% of the area of the BAWLD domain was included in the expert assessment. Each expert was asked to assess the percent
coverage of each of the 19 land cover classes within their 20 training cells. To guide their assessment, each expert was
400
provided step-by-step instructions, plus information on the definitions of each land cover class, and a KML file with the data
extracted from available spatial datasets for each grid cell (Table 1). Experts were asked to use their knowledge of typical
wetland and lake classes within specific high-latitude regions, their ability to interpret satellite imagery as provided by
Google Earth, and their judgement of the quality and relevance of available spatial datasets to make their assessments of
fractional cover. The information provided to experts to carry out the assessment is provided in the Supplementary
405
Information.
2.4 Random forest model and uncertainty analysis
Random forest regression models were created to predict the percent coverage of all 19 individual BAWLD land cover
classes, along with three additional models for total wetland, lake, and river coverage. Each land cover class was at first
modelled separately, which was followed by minor adjustments, described below, that ensured that the total land cover
410
within each cell added up to 100%. All statistical analysis and modelling were done using R 4.0.2 (R Core Team, 2020), and
the packages Boruta (v7.0.0; Miron et al., 2010), caret (v6.0-86; Kuhn, 2020), randomForest (v4.6-14; Liaw and Wiener,
2002), and factoextra (v1.0.7; Kassambara and Mundt, 2020).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

15
Table 2. Summary of random forest models for each land cover class in BAWLD.
415
Land cover classes
RMSE (%)
%Var
mtrya
Var.b
Relative Variable Importancec
Glaciers
2.32
95.9
13
24
GL30-ICE (100) NCS-GLA (6)
Rocklands
9.79
67.2
21
41
NCS-ROC (100) PZI-MTN (43) CAVM-BAR (39)
PZI-RUG (24) CAPG-XLR (16) WC2-MAAP (14) WC2-MAAT (14)
WC2-CMI (11) TEW-TUN (10) GLC2-FOR (10)
Tundra
14.7
75.2
22
43
TEW-TUN (100) TEW-BOR (41) PZI-PERM (26) GL30-TUN (16)
WC2-MAAP (8) CAPG-CON (7) LAT (7) PZI-PERM (7) NCS-ROC(5)
Boreal Forest
15.5
79.8
20
39
GLC2-FOR (100) TEW-BOR (61) GL30-WET (23) TEW-TUN (21)
GIEMS_MAMA (12) GIM_MAMI (11) GL30_TUN (10)
Wetland Classes
8.5
85.8
25
48
GL30-WET (100) NCS-HSE (46) NCS-HSO (44) PZI_FLAT (28)
GWET-IN (10) WC2-MAAT (6) GLC2-WET (5)
Bog
4.7
75.0
22
42
NCS-HSO (100) GL30-WET (47) WC2-MAAT (23) PZI-PERM (17)
PZI_FLAT (12) GLC2-WET (12) WC2-MAAP (6)
Fen
4.5
76.3
21
40
NCS-HSO (100) GL30-WET (56) WC2-MAAT (16) PZI-PERM (7)
Marsh
1.3
54.1
18
34
GL30-WET (100) GSW-OCC (80) GLC2-WET (56) BAS-RIV (31)
NCS-HSO (17) WC2-MAAT (14) GLWD-RIV (13) PZI-PERM (12)
IRYP-YED (10) GSW-RAR (10)
Permafrost Bog
4.1
84.0
22
42
NCS-HSE (100) CAPG_DIS (6) CAPG-XMF (5)
Tundra Wetland
4.1
47.2
2
36
CAVM-WET (100) GSW-OCC (95) NCS-HSE (80) CAPG-XHF (77)
PZI-FLAT (63) GL30-TUN (61) WC2-CMI (57) HL-MID (57)
IRYP-YED (56) LAT (53)
Lentic Classes
2.03
97.8
32
32
GL30-H2O (100)
Large Lake
0.75
99.5
30
30
HL-LAR (100) GL30-H2O (18) NCS-H2O (10)
Midsize Glacial Lake
1.49
75.3
18
35
HL-SHO (100) HL-MID (56) GSW-REG (18) GL30-H2O (8)
NCS-H2O (8) GSW-PER (6) NCS-ROC (5) GLC2-WET (5) \
PZI-PERM (5)
Midsize Peatland Lake
1.44
68.5
17
32
HL-MID (100) NCS-HSO (24) GL30-WET (18) GWET-IN (18)
HL-SHO (17) NCS-HSE (14) GSW-OCC (13) GLC2-WET (11)
GIM-MAMI (9) GSW-PER (8)
Midsize Yedoma Lake
0.86
68.4
16
31
IRYP-YED (100) HL-MID (42) HL-SHO (23) CAVM-WET (14)
CAPG-XHF (12) GSW-OCC (12) PZI-PERM (9) WC2-CMI (8)
GL30-H2O (7) GSW-REG (6)
Small Glacial Lake
0.89
15.6
2
29
GSW-OCC (100) GSW-REG (81) HL-SHO (49) HL-MID (39)
GIM-MAMA (37) GIM-MAMI (29) GSW-RAR (27) WC2-MAAT (26)
PZI-PERM (25) GL30-H2O (23)
Small Yedoma Lake
0.47
39.2
17
32
IRYP-YED (100) WC2-MAAP (25) WC2-CMI (21) GLC2-H2O (15)
GSW-OCC (14) CAVM-WET (14) PZI-FLAT (13) GSW_REG (11)
CAPG-XHF (10) HL-MID (10)
Small Peatland Lake
1.22
65.9
20
39
GLC2-WET (100) GL30-WET (95) WC2-MAAT (34) CAPG-REL (32)
NCS-HSE (26) PZI-PERM (24) NCS-HSO (23) GSW_OCC (23)
LAT (17) WC2-CMI (16)
Lotic Classes
0.49
90.3
16
31
GLWD-RIV (100) BAS-RIV (39) GSW-OCC (11)
Large River
0.48
90.4
17
32
GLWD-RIV (100) BAS-RIV (39) GSW-OCC (10)
Small Org.-Poor Rivers
0.09
18.7
2
41
GSW-OCC (100) GLC2-WET (62) BAS-RIV(48) GSW-PER (42)
GLWD-RIV (39) PZI-FLAT (38) GL30-H2O (33) GLC2-H2O (32)
GIM-MAMA (32) WC2-MAAP (31)
Small Org.-Rich Rivers
0.04
59.3
23
45
GLC3-WET (100) PZI-FLAT (41) NCS-HSE (37) NCS-HSO (18)
GLC2-WET (17) GWET-IN (13) GSW-OCC (12) CAPG-XLR (12)
GLWD-RIV (11) BAS-RIV (11)
amtry – is a fitted variable which decides how many variables that were randomly chosen at each split in the random forest analysis.
bVar. – indicates the number of variables that were included (out of 53) in the random forest analysis after the Boruta automatic feature selection.
cRelative variable importance – the most influential variable in the random forest analysis is assigned a 100% rating, and the importance of other variables
are relative to this. See Table 1 for full descriptions of the variables. Here we list either all variable with >5% influence, or the top ten variables.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

16
Prior to running the random forest analyses, we performed an automatic feature selection using a Boruta algorithm, (Miron et
al., 2010). The Boruta algorithm completed 150 runs for each land cover class, after which subsets of the 53 possible data
variables (Table 1) were deemed important and selected for inclusion in subsequent random forest models (Table 2). The
random forest models (Kuhn 2020, Liaw and Wiener, 2002) then used boot-strapped samples (i.e. the expert assessments of
420
land cover fractional grid cell coverages) to grow 500 decision-trees (ntree), with a subset of randomized data variables as
predictors at each tree node (mtry). We used a 10-fold cross-validation with five repetitions providing mtry as a tuneable
parameter for model training. The random forest model output included the root mean squared error (RMSE), the percent of
the expert assessment variability that was explained (%Var), and relative variable importance (Table 2). Relative variable
importance assigns a 100% importance to the variable with the most influence on the model, and then ranks all other
425
variables relative to the influence of that variable. A bias correction (Song, 2015) was applied to the predicted data of land
cover class coverages, as the models were found to overestimate low coverages and underestimate high coverages. After the
bias correction, all bias-adjusted predictions <0% were set to 0%, while those >100% were set to 100% (for examples of the
bias correction, see Figure S2). Next, we ensured that the combined coverage of all 19 land cover classes within each grid
cell added up to 100% by applying a proportional adjustment. In order to estimate the 5th and 95th percentile confidence
430
bounds of the land cover predictions, we repeated the random forest analysis, as outlined above, an additional 20 times for
each class. Each new run completely excluded 20% of the expert assessments, and the data were reshuffled four times. Each
grid cell thus had 21 predictions of coverage for each of the 19 land cover classes, and for the cumulative wetland, lake, and
river coverages, and the variability of these predictions were used to define the 5th and 95th percentile confidence bounds.
435
While each cell in BAWLD has a distinct land cover combination, we were also interested in identifying cells with
similarities in their land cover compositions to distinguish between regions of the boreal and arctic domain that represent
characteristic landscapes. We carried out a k-means clustering (Kassambara and Mundt, 2020) to group grid cells with
similarities in their predicted land cover compositions. The k-means clustering was based on within-cluster sum of squares,
and we evaluated resulting maps with between 10 and 20 distinct classes. Using 15 clusters was deemed to balance within-
440
cluster sum of squares and interpretability of the resulting map. As each cluster was defined largely by the relative
dominance (or absence) of different wetland, lake, and river classes, we henceforth refer to these clusters as “wetscapes”.
2.5 Evaluation against regional wetland datasets
We evaluated the predictions of wetland coverage in BAWLD against four independent, high-resolution regional land cover
datasets. These four datasets were chosen as they included more than one wetland class, thus enabling both evaluation
445
against total wetland coverage and subsets of wetland classes. Two of these datasets were specifically aimed at mapping of
wetlands, including Ducks Unlimited Canada’s wetland inventories for western Canada as part of the Canadian Wetland
Inventory (CWI; Canadian Wetland Inventory Technical Committee, 2016), and wetland mapping of the West Siberian
Lowlands (WSL) (Terentieva et al., 2016). The other two datasets, the 2016 National Land Cover Database (NLCD) of
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

17
Alaska (Homer et al., 2020), and the 2018 CORINE Land Cover (Büttner, 2014) of northern Europe, represent more general
450
land cover datasets. Data from these four datasets were summarized for each BAWLD grid cell where there was complete
coverage. Data filtration was done for the CWI to remove cells if >10% of the cell was classified as burned, cloud, or
shadow. There were few cases where there were equivalent wetland classes in BAWLD and these four regional datasets, and
as such comparisons were generally made between groups of wetland classes that were considered generally comparable.
Similar evaluations were not possible for the lake classes, as there are no regional or circumpolar spatial datasets with
455
information on lake genesis or sediment type.
3 Results and Discussion
The fractional land cover estimates of the Boreal-Arctic Wetland and Lake Dataset (BAWLD) is freely available online at
https://doi.org/10.18739/A2C824F9X (Olefeldt et al., 2021), and includes both the central estimates and the 95% high and
low estimates of each land cover class in each grid cell.
460
3.1 Wetlands
Wetlands were predicted to cover a total of 3.2 × 106 km2, or 12.5% of the BAWLD domain. The wetland area was
dominated by Fens (29% of total wetland area), Bogs (28%), and Permafrost Bogs (27%), while Marshes and Tundra
Wetlands, which have relatively higher CH4 emissions, covered 5 and 12% of the wetland area, respectively (Table 3). This
estimate of total wetland area was greater than previously mapped within the BAWLD domain in GLC2 at 0.9 × 106 km2
465
(Bartholomé and Belward, 2005), GL30 at 1.4 ×106 km2 (Chen et al., 2015), and GWET at 2.3 × 106 km2 (Matthews and
Fung, 1987), but similar to the area of wetland soils in NCS (sum of “histosols”, “histels”, and “aqueous” soil coverage) at
3.0 × 106 km2 (Hugelius et al., 2014). Differences between BAWLD and other estimates of wetland area likely stem partially
from differences in wetland definitions, where e.g. definitions of wetlands in GLC2 and GL30 likely do not include wooded
bogs, fens, and permafrost bogs. While estimates of total wetland area GWET and in the NCS were closer to BAWLD, there
470
were differences in the spatial distribution. Wetland cover in BAWLD was generally greater than in GWET and NCS in
regions with low wetland cover. This likely reflects the ability of experts to infer the presence of small, or transitional
wetlands that may otherwise be underestimated when mapped using other methodologies. Conversely, wetland cover in
BAWLD was generally lower than in GWET and NCS in regions with high wetland cover. This was likely due to
differences in definitions, especially the exclusion of all open water ecosystems from wetlands in BAWLD. For example, it
475
was common in the West Siberian Lowlands for the summed coverage of wetland soils in NCS and the “open water”
coverage in GL30 to be substantially greater than 100%, suggesting that NCS included peatland pools and small ponds
within its wetland soil coverage. Overall, the predictive random forest model of total wetland coverage was able to explain
86% of the variability in the expert assessments, and it was primarily influenced by the area of “wetlands” in GLC30 and the
wetland soil categories in NCS, followed by the coverage of “flat topography” in PZI (Gruber, 2012) (Table 2).
480
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

18
Table 3. Summary of central estimates, 95% low and high confidence bounds, and the range of the 95% confidence interval
expressed as a percent of the central estimate, for each of the land cover classes within the BAWLD domain.
Land cover classes
Central
estimate
(106 km2)
Low
confidence
bound
(106 km2)
High
confidence
bound
(106 km2)
95%
confidence
interval
(106 km2)
95% CI
(% of central
estimate)
Glaciers
2.09
1.99
2.21
0.22
11
Rocklands
2.74
2.21
3.40
1.19
44
Tundra
5.28
4.56
6.37
1.82
34
Boreal Forest
10.66
9.77
11.39
1.61
15
Wetlands
3.18
2.79
3.79
1.00
31
Bog
0.88
0.71
1.24
0.53
60
Fen
0.91
0.76
1.14
0.38
42
Marsh
0.16
0.12
0.23
0.11
71
Permafrost Bog
0.86
0.67
1.17
0.50
58
Tundra Wetland
0.38
0.31
0.53
0.22
59
Lakes
1.44
1.34
1.59
0.24
17
Large Lake
0.64
0.61
0.72
0.11
18
Midsize Peatland Lake
0.14
0.11
0.21
0.10
69
Midsize Yedoma Lake
0.034
0.023
0.071
0.05
140
Midsize Glacial Lake
0.38
0.33
0.43
0.10
26
Small Peatland Lake
0.12
0.085
0.17
0.08
71
Small Yedoma Lake
0.028
0.015
0.046
0.03
114
Small Glacial Lake
0.094
0.051
0.16
0.11
119
Rivers
0.12
0.094
0.19
0.10
81
Large River
0.080
0.072
0.11
0.04
50
Small Organic-Rich Rivers
0.010
0.005
0.054
0.05
502
Small Organic-Poor Rivers
0.033
0.020
0.067
0.05
143
485
The predictive random forest models for individual wetland classes differed both in terms of how much of the variability of
the expert assessment data was explained, and in terms of which spatial data were most influential (Table 2). The model for
Permafrost Bog coverage explained 84% of the variability in the expert assessments and was very strongly influenced by
“histel” distribution in the NCS (Hugelius et al., 2014). Predictive models explained ~75% of the variability in the expert
assessments for Bogs and Fens separately (Table 2), but 87% when considered jointly. This shows that the available
490
predictor variables were less suitable for modelling Bogs and Fens separately rather than jointly, which could partly be due
to lower agreement among experts in assessments of Bog and Fen coverages compared to their sum. This would not be
surprising as bogs and fens (and swamps) occur along hydrological and nutrient gradients, and can have vegetation
characteristics that make them difficult to distinguish. Models for Bogs and Fens were both strongly influenced by the
“histosol” distribution in NCS, with secondary influences from the area of “wetlands” in GL30, “permafrost extent” in PZI,
495
and “mean annual air temperature” in WC2. Predictive models for Marsh and Tundra Wetlands explained less of the
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

19
variability in expert assessments, at 54 and 47%, respectively. The predictive models for Marsh and Tundra Wetlands were
influenced by variables that indicate a transition between terrestrial and aquatic ecosystems, e.g. area of “occasional
inundation” in GSW, “rivers” in BAS, and “midsize lakes” in HL, but then differed in the influence of climate and
permafrost conditions.
500
Figure 2. Predicted distribution of wetland classes across the BAWLD domain; a) Bog, b) Fen, c) Marsh, d) Permafrost Bog, and
e) Tundra Wetland.
505
Each wetland class had a distinct spatial distribution (Figure 2). Bogs and Fens were the dominant wetland classes in
relatively warmer climates, with high densities in the West Siberian Lowlands, Hudson Bay Lowlands, and the Mackenzie
River Basin. While Bogs and Fens had similarities in their spatial distributions, there was also a relative shift in dominance
from Bogs to Fens in relatively colder and drier climates (Figure S3). These trends are supported by bog to fen transitions
observed both within and between regions (Packalen et al., 2016; Vitt et al., 2000a; Väliranta et al., 2017), but may not be
510
universal (Kremenetski et al., 2003). Marshes were also found in warmer climates and largely associated with Bogs and
Fens, but with a more evenly spread distribution. The highest abundance of Marsh coverage was predicted for the Ob River
floodplains, a region with very few studies of CH4 emissions (Terentieva et al., 2019; Glagolev et al., 2011). Bogs, Fens, and
Marshes all decreased in abundance in colder climates, with Permafrost Bogs becoming more abundant than Bogs when
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

20
mean annual temperatures were below -2.5°C, corresponding to findings from western Canada, Fennoscandia, and the West
515
Siberian Lowlands (Vitt et al., 2000b; Seppälä, 2011; Terentieva et al., 2016). Tundra Wetlands became dominant over Fens
and Marshes when mean annual air temperatures were below -5.5°C (Figure 3). Tundra Wetlands were predicted to be most
abundant in the lowland regions across the Arctic Ocean coast, with especially high abundance in northern Alaska, eastern
Siberia, and on the Yamal and Gydan Peninsulas in western Siberia.
520
Figure 3. Relative abundance of the five wetland classes across a gradient of mean annual temperatures.
We found good agreement between the distribution of wetlands in BAWLD and that of four independent regional spatial
datasets (Figure 4, Figure S4). The best agreements for total wetland cover were between BAWLD and the two datasets
525
dedicated specifically to wetland mapping; with R2 of 0.76 with the WSL dataset and 0.72 with the CWI dataset. There were
also strong relationships between BAWLD and the WSL dataset for the distribution of specific wetland classes, for both
drier wetland classes (“ridge”+”ryam”+”palsa” vs. Permafrost Bog + Bog) and wetter classes (“fen” + “hollow” vs. Fen).
When comparing “wet hollow” of the WSL dataset and Marsh in BAWLD there were discrepancies, but they were primarily
attributed to the explicit exclusion of the Ob river floodplains in the WSL dataset (Figure S4). For the wettest classes, we had
530
only a weak relationship (R2 = 0.19) between the CWI “Marsh” class and the sum of the BAWLD Marsh and Tundra
Wetland classes, but the overall average abundance for comparable grid cells was similar at 1.4 and 2.2%, respectively.
Agreements between BAWLD and the NLCD and CLC datasets were lower, especially for the relatively drier wetland
classes (Figure S4). Lower agreement between BAWLD and some classes of regional wetland datasets should not be
interpreted to demonstrate poor accuracy of BAWLD, as differences can be due to class definitions, large mapping units, and
535
relatively low accuracy of the non-wetland specific regional datasets.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

21
Figure 4. Comparison of total wetland extent between BAWLD and four regional independent wetland inventories; the National
Land Cover Database (NLCD), the Canadian Wetland Inventory (CWI), the wetland mapping of the West Siberian Lowlands
(WSL), and the CORINE Land Cover dataset (CLC). a) Spatial extents of the regional datasets, b) Correlations between grid cell
540
wetland coverages in BAWLD and the regional datasets, c) Spatial distribution of total wetland coverages in the four regional
datasets, d) Spatial distribution of total wetland coverage in BAWLD for grid cells corresponding with the regional datasets.
The 95% confidence intervals for predictions of abundance varied both between wetland classes and among regions (Table
545
3, Figure S5, S6). The confidence interval for total wetland area ranged between 2.8 and 3.8 × 106 km2, i.e. a range that
represented 31% of the central estimate. The range of the confidence interval depends both on how much consensus there is
among experts in their assessments, and how well the available spatial datasets used in the random forest modelling can
explain the expert assessments. The considerable range of the confidence interval for wetlands likely stems from a
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

22
combination of these two components. The confidence interval for total area of individual wetland classes varied between
550
representing 42% (Fens) and 71% (Marshes) of respective central estimates. The absolute range of confidence intervals for
individual cells generally increased with higher central estimates of abundances, but the range of confidence intervals
decreased if expressed as a percent of the central estimate (Figure S7).
3.2 Lakes
Lakes were predicted to cover a total of 1.44 ×106 km2, or 5.6% of the BAWLD domain. Large Lakes had the greatest lake
555
area (44% of total lake area), followed by Midsize Glacial Lakes (26%) and Midsize Peatland Lakes (10%) (Table 3). The
lake classes with the highest CH4 emissions, Small Yedoma Lakes and Small Peatland Lakes, jointly covered 10% of the
total lake area. The total predicted lake area in BAWLD was higher than the area of lakes in HL (1.20 ×106 km2) which only
includes lakes >0.1 km2, and was similar to the area of “open water” in GL30 (1.43 ×106 km2). The “open water” class in the
GL30 dataset is, however, based on Landsat 30 m resolution data and thus excludes very small open water areas, while it
560
includes both lentic and lotic open water. The 95% confidence interval for the total lake area in BAWLD was 0.24 ×106 km2,
or 17% of the central estimate.
The predictive models for the three midsize lake classes each explained between 69 and 75% of the variability in expert
assessments, while a model for the sum of the three midsize lake classes explained 99.1%. The predictive model for the sum
565
of the three midsize lake classes was almost exclusively influenced by the area of “midsize lakes” in HL, while the three
midsize lake classes were differentiated through further influences by the area of “yedoma ground” (Midsize Yedoma Lakes),
by the area of “histosols” and “histels” in NCS, and “wetlands” in GL30 (Midsize Peatland Lakes), and by “shoreline
length” in HL (Midsize Glacial Lakes). The influence of “shoreline length” for Midsize Glacial Lakes shows that experts
associated glacial lakes with high shoreline development, and low shoreline development with peatland and yedoma lakes.
570
Despite similarities in how much of the expert assessments were explained by the predictive models (69-75%), the
extrapolation to the BAWLD domain led to large differences in the 95% confidence interval, which represented only 26% of
the central estimate for Midsize Glacial Lakes, while representing 69 and 140% for Midsize Peatland and Midsize Yedoma
Lakes, respectively (Table 3, Figure S8). Midsize Glacial Lakes were predominately predicted to have high abundances on
the Canadian Shield and in Fennoscandia, while Midsize Yedoma Lakes were associated with the lowland, coastal tundra
575
regions of Northeast Siberia and Alaska, and Midsize Peatland Lakes were especially common in the West Siberian
Lowlands, but also common in the peatland regions of the Hudson Bay Lowlands and the Mackenzie River Basin and in
coastal lowland regions (Figure 5).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

23
Figure 5. Predicted distributions of lake and river classes within the BAWLD domain; a) Large Lakes, b) Midsize Glacial Lakes, c)
580
Midsize Peatland Lakes, d) Midsize Yedoma Lakes, e) Small Glacial Lakes, f) Small Peatland Lakes, g) Small Yedoma Lakes. h)
Large Rivers, i) Small Organic-Rich Rivers, j) Small Organic-Poor Rivers.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

24
Small Glacial, Yedoma, and Peatland Lakes were jointly estimated to cover 0.9% of the BAWLD domain. The predictive
models explained 16, 39, and 66% of the variability, respectively (Table 2). The relatively lower predictive power for small
lakes was not unexpected, given lack of information on the smallest open water systems in the available spatial data, the
585
variable abundance of very small open water systems among landscapes (Muster et al., 2019), and a lower relative consensus
among experts when assessing classes with generally small fractional coverages. Models for all three small lake classes were
influenced by the area of “occasional inundation” in GSW but were then differentiated by variables largely similar to those
that were characteristic of the corresponding midsize lake classes (Table 2). The predicted distributions of the small lake
classes were also largely similar to that of the corresponding midsize lake type classes (Figure 5). The overall predicted area
590
of small lakes was 0.24 ×106 km2, representing 17% of the total lake area. The combined 95% uncertainty for the three
classes ranged between 0.15 and 0.38 ×106 km2 (Table 3, Figure S8), suggesting that small lakes represent between 11 and
26% of the total lake area. Previous assessments have estimated that open water ecosystems <0.1 km2 represent between 21
and 31% of global lake area (Holgerson and Raymond, 2016), but relied on assumptions in the statistical modelling which
may lead to bias for boreal and arctic regions (Cael and Seekell, 2016; Muster et al., 2019).
595
3.3 Rivers
Rivers were predicted to cover a total of 0.12 ×106 km2, or 0.47% of the BAWLD domain. Large Rivers accounted for 65%
of the total river area in BAWLD. These estimates were similar to global assessments, where streams and rivers have been
estimated to cover between 0.30 and 0.56% of the land area, with 65% of the river area consisting of large rivers of 6th or
600
greater stream order (Downing et al., 2012). The predictive model for Large Rivers was strongly influenced by the area of
“large rivers” in GLWD, but experts consistently made lower assessments which led to an overall 15% lower area of Large
Rivers compared to the area of rivers in GLWD within the BAWLD domain.
Small Organic-Poor and Small Organic-Rich Rivers were estimated to represent 27% and 8%, respectively, of the total river
605
area. The predictive models for the Small Organic-Poor and Organic-Rich Rivers explained 19 and 59% of the expert
assessments, and were distinctly influenced by the area of “occasional inundation” in GSW and “wetlands” in GLC30,
respectively. The estimated area of small rivers varied among experts, reflecting difficulties in consistent assessments among
experts for land cover classes with low extents (<1% in most grid cells). The distributions of expert assessments for small
river areas were non-normal, leading to a long upper tail for the 95% confidence interval (Figure S9). For example, the low,
610
central, and high estimates for the area of Small Organic Rich Rivers were 0.005, 0.10 and 0.54 ×106 km2, respectively. The
predicted distributions showed that Small Organic-Rich Rivers was closely associated with the distribution of the BAWLD
wetland classes, while Small Organic-Poor Rivers dominated elsewhere, with especially high abundances in regions with
higher mean annual precipitation (Figure 5).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

25
3.4 Other Classes
615
Boreal Forest, Dry Tundra, Rocklands, and Glaciers were predicted to cover 10.7, 5.3, 2.7 and 2.1 ×106 km2, respectively,
within the BAWLD domain (Figure S10). The predictive models explained between 96% (Glaciers) and 67% (Rocklands) of
the variability in expert assessments. While the predictive models for Glaciers was almost exclusively influenced by the area
of “permanent snow and ice” in GL30, several variables influenced predictions of Rocklands – including area of “rocklands”
in NCS, “mountainous” and “rugged” terrain in PZI, and “barrens” in CAVM. The predictive models for Boreal Forest and
620
Tundra suggested that the transition between these classes was strongly influenced by the area “forest” in GLC2, and by the
distinction between “tundra” and “boreal” terrestrial ecoregions in TEW.
3.5 Wetscapes
We defined “wetscapes” as regions with characteristic composition of specific wetland, lake, and river classes. Our
clustering analysis distinguished 15 typical wetscapes within the BAWLD domain (Figure 6), each defined by the relative
625
presence or absence of the 19 BAWLD classes (Table S1). Visualising the distribution of wetscapes provides information on
regions that are likely to have similarities in the magnitude, seasonality, and climatic controls over CH4 emissions.
Three wetscapes common in boreal regions were differentiated based on the abundance of non-permafrost wetlands. The
Sparse, Common, and Dominant Boreal Wetlands wetscapes all had limited lake coverage (<6% on average), but had 15, 35,
630
and 60% combined coverages of Bogs, Fens, and Marshes, respectively. The Dominant Boreal Wetlands wetscape was
almost exclusive to the non-permafrost regions of the Hudson Bay Lowlands and the West Siberian Lowlands. The Common
Boreal Wetlands wetscape was more widespread, found adjacent to the core areas of the Hudson Bay Lowlands and the West
Siberian Lowlands, but also in the Mackenzie River Basin, northern Finland, European Russia, and in the Kamchatka
Lowlands. The Sparse Boreal Wetlands wetscape was widespread in Sweden, Finland, European Russia, and the southern
635
boreal regions of Canada outside of Yukon. Emissions of CH4 from these regions are likely dominated by wetlands rather
than lakes, with main sensitivity to climate change being altered water balance (Tarnocai, 2006; Olefeldt et al., 2017; Olson
et al., 2013).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

26
Figure 6. Wetscapes of the Boreal-Arctic Wetland and Lake Dataset. Wetscapes are defined by their characteristic composition of
640
the BAWLD land cover classes, and thus groups regions with similar abundances (or absences) of specific wetland, lake, and river
classes. The 15 wetscapes have their average land cover composition indicated by pie charts, with the legend shown in the bottom
left. For clarity, the small and mid-sized lakes classes were combined for glacial, peatland, and yedoma lakes, and the river classes
were omitted from the pie charts. No land cover pie charts are shown for the Large Lakes, Rivers, and Glaciers wetscapes.
645
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

27
The Lake-rich Peatlands and the Permafrost Peatlands wetscapes were both found in lowland regions with discontinuous
permafrost, near the boreal to tundra transition. The Lake-rich Peatlands wetscape was almost exclusively found in the West
Siberian Lowlands, north of the Ob River. This wetscape was characterized by roughly equal abundances of Bogs, Fens and
Permafrost Bogs (each 14-16%), along with 8% Marshes, 9% Small Peatland Lakes and 5% Midsize Peatland Lakes. It is
650
notable that this wetscape, with the highest coverages of high-CH4 emitting marshes and peatland lakes, has no presence in
North America. The Permafrost Peatlands wetscape was conversely primarily found in the Hudson Bay Lowlands and the
Mackenzie River Basin, with additional coverage along the Arctic Ocean coast in European Russia, in interior Alaska, and in
the Anadyr Lowlands of far eastern Russia. This wetscape had the greatest abundance of Permafrost Bogs (27%), with less
contribution from other wetland classes (16%), and relatively low abundance of lakes (7%). The Lake-rich Peatlands
655
wetscape likely has the highest regional CH4 emissions, while the Permafrost Peatlands wetscape likely has low to moderate
emissions. However, CH4 emissions from both these wetscapes are likely highly sensitive to climate change due to the rapid
ongoing and future permafrost thaw that causes expansion of thermokarst lakes and non-permafrost wetlands at the expense
of Permafrost Bogs (Bäckstrand et al., 2008; Turetsky et al., 2002).
660
Three wetscapes were found in lowland tundra regions, and varied in relative dominance of different wetland and lake
classes. Wetland-rich Tundra had 23% wetlands but only 7% lakes, and was found on the Gydan and Taymyr peninsulas in
North Siberia, with minor extents in far eastern Siberia and in Alaska. Wetland and Lake-rich Tundra had similar wetland
cover (24%) but twice the coverage of lakes (15%), split equally between glacial and peatland lakes. It was found on the
Alaska North Slope along with minor extents on the Yamal Peninsula, the Mackenzie River Delta, and on sections of Baffin
665
Island. Lastly, the Wetland and Lake-rich Yedoma Tundra was characterized by the highest abundance of yedoma lakes
(8%), a total wetland and lake coverage of 46%, and was primarily found in the Kolyma Lowlands, with minor extents in the
Yukon-Kuskokwim Delta and on the Alaska North Slope. These regions may have sensitive CH4 emissions, particularly
associated with thermokarst lake expansion where highly labile yedoma sediments fuel high CH4 production (Walter
Anthony et al., 2016).
670
The remaining seven wetscapes are likely to have overall low CH4 emissions, or even net uptake, as a result either from the
dominance of low-CH4 emitting classes or due to the relative absence of wetland and lake classes. The Dry Tundra wetscape
was common in regions of undulating topography of northernmost Siberia, the Alaska North Slope, and the western
Canadian Arctic, and was characterized by relatively low abundances of wetlands (9%) and lakes (3%). The Lake-rich Shield
675
wetscape was exclusive to the Canadian Shield, and although it had a high abundance of lakes (18%), these were almost
completely dominated by low-CH4 emitting large lakes and glacial lakes. The Upland Boreal wetscape dominates boreal
regions of Siberia but is also found in the Yukon, Alaska, and Quebec, and was defined by having <5% wetlands and 0.5%
lakes. The Alpine and Tundra Barrens wetscape had <2% wetlands and ~1.5% lakes, and dominates the Greenland coast, the
high-latitude polar deserts of the Canadian Arctic Archipelago, and the mountain ranges in Fennoscandia, Alaska, Yukon,
680
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

28
and eastern Siberia. Lastly, the Glaciers, Large Lakes, and Large Rivers wetscapes were defined by the dominance of the
namesake BAWLD classes.
5 Data Availability
The fractional land cover estimates from the Boreal-Arctic Wetland and Lake Dataset (BAWLD) is freely available at the
Arctic Data Center (Olefeldt et al., 2021): https://doi.org/10.18739/A2C824F9X. The dataset is provided as an ESRI
685
shapefile (.shp) and as a Keyhole Markup Language (.kml) file.
6 Conclusions
The Boreal-Arctic Wetland and Lake Dataset (BAWLD) was developed to provide improved estimates of areal extents of
five wetland classes, seven lentic ecosystem classes, and three lotic ecosystem classes by leveraging expert knowledge along
with available spatial data. By differentiating between wetland, lake and river classes with distinct characteristics, BAWLD
690
will be suitable to support large-scale modelling of high-latitude hydrological and biogeochemical impacts of climate
change. In particular, BAWLD has been developed with the aim to facilitate improved modelling of current and future CH4
emissions. For example, a dataset of empirical CH4 data was co-developed with BAWLD (Kuhn et al., 2021), ensuring that
the land cover classification was meaningful for the separation of classes based on distinct magnitudes and controls of CH4
emissions. By being based on an expert assessment and existing spatial dataset rather than a remote sensing approach,
695
BAWLD was able to provide predictions for abundance of high-CH4 emitting wetland and lake classes that have limited
extents but disproportionate influences on regional and overall CH4 emission. Using BAWLD for upscaling of CH4
emissions will reduce issues of representativeness of empirical data for upscaling, reduce the risk of overlap between
wetland and lake classes, and allow for more rigorous uncertainty analysis.
Author contributions
700
This study was conceived by DO. The GIS work was done by MH. The information sent to experts to complete the expert
assessment was compiled by DO, MH, and MAK. All co-authors completed the expert assessment. The random forest
modelling was led by DO, with input from TB, AR, and MJL. Data analysis and visualizations was led by DO with input
from all co-authors. The manuscript was written by DO with contributions from all co-authors.
Competing Interests
705
The authors declare no competing interests.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

29
Acknowledgements
Financial support to DO was provided the National Science and Engineering Research Council of Canada (NSERC)
Discovery grant (RGPIN-2016-04688) and the Campus Alberta Innovates Program. CT was supported by ERC (#851181)
and the Helmholtz Impulse and Networking Fund. AM was supported by the Gordon and Betty Moore Foundation (Grant
710
GBMF5439, 839; Stanford University). DB was supported by ERC (#725546), Swedish Research Council VR (#2016-
04829), and FORMAS (#2018-01794). FJWP was supported by the Norwegian Research Council under grant agreement
274711, and the Swedish Research Council under registration no. 2017-05268. GG was supported through the BMBF KoPf
Synthesis project (03F0834B). JDW was supported by NASA Earth Science (NNH17ZDA001N). MJL was supported by
NSF-EnvE (#1928048). MS was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC)
715
through the Canada Research Chairs program. RKV was supported by the National Aeronautics and Space Administration
IDS program (NASA grant NNX17AK10G). SAF was supported by the Natural Sciences and Engineering Research Council
of Canada. SET was supported by funding from the Campus Alberta Innovates Program. Ducks Unlimited Canada’s
Wetland Inventories were funded by various partnering organizations: Environment and Climate Change Canada, Canadian
Space Agency, Government of Alberta, Government of Saskatchewan, U.S. Forest Service, U.S. Fish and Wildlife Service,
720
PEW Charitable Trusts, Canadian Boreal Initiative, Alberta-Pacific Forest Industries Inc., Mistik Management Ltd.,
Louisiana-Pacific, Forest Products Association of Canada, Weyerhaeuser, Lakeland Industry and Community, Encana,
Imperial Oil, Devon Energy Corporation, Shell Canada Energy, Suncor Foundation, Treaty 8 Tribal Corporation
(“Akaitcho”), and Dehcho First Nations. The Permafrost Carbon Network provided coordination support, and is funded by
the NSF PLR Arctic System Science Research Networking Activities (RNA) Permafrost Carbon Network: Synthesizing Flux
725
Observations for Benchmarking Model Projections of Permafrost Carbon Exchange, Grant # 1931333 (2019-2023).
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

30
References
Abbott, B. W., Jones, J. B., Schuur, E. A. G., III, F. S. C., Bowden, W. B., Bret-Harte, M. S., Epstein, H. E., Flannigan, M.
D., Harms, T. K., Hollingsworth, T. N., Mack, M. C., McGuire, A. D., Natali, S. M., Rocha, A. V., Tank, S. E., Turetsky, M.
730
R., Vonk, J. E., Wickland, K. P., Aiken, G. R., Alexander, H. D., Amon, R. M. W., Benscoter, B. W., Yves Bergeron,
Bishop, K., Blarquez, O., Bond-Lamberty, B., Breen, A. L., Buffam, I., Yihua Cai, Carcaillet, C., Carey, S. K., Chen, J. M.,
Chen, H. Y. H., Christensen, T. R., Cooper, L. W., Cornelissen, J. H. C., Groot, W. J. de, DeLuca, T. H., Dorrepaal, E.,
Fetcher, N., Finlay, J. C., Forbes, B. C., French, N. H. F., Gauthier, S., Girardin, M. P., Goetz, S. J., Goldammer, J. G.,
Gough, L., Grogan, P., Guo, L., Higuera, P. E., Hinzman, L., Hu, F. S., Gustaf Hugelius, Jafarov, E. E., Jandt, R., Johnstone,
735
J. F., Karlsson, J., Kasischke, E. S., Gerhard Kattner, Kelly, R., Keuper, F., Kling, G. W., Kortelainen, P., Kouki, J., Kuhry,
P., Hjalmar Laudon, Laurion, I., Macdonald, R. W., Mann, P. J., Martikainen, P. J., McClelland, J. W., Ulf Molau,
Oberbauer, S. F., Olefeldt, D., Paré, D., Parisien, M.-A., Payette, S., Changhui Peng, Pokrovsky, O. S., Rastetter, E. B.,
Raymond, P. A., Raynolds, M. K., Rein, G., Reynolds, J. F., Robards, M., Rogers, B. M., Schädel, C., Schaefer, K., Schmidt,
I. K., Anatoly Shvidenko, Sky, J., Spencer, R. G. M., Starr, G., Striegl, R. G., Teisserenc, R., Tranvik, L. J., Virtanen, T.,
740
Welker, J. M., et al.: Biomass offsets little or none of permafrost carbon release from soils, streams, and wildfire: an expert
assessment, Environ. Res. Lett., 11, 034014, https://doi.org/10.1088/1748-9326/11/3/034014, 2016.
Bäckstrand, K., Crill, P. M., Mastepanov, M., Christensen, T. R., and Bastviken, D.: Total hydrocarbon flux dynamics at a
subarctic mire in northern Sweden, J. Geophys. Res. – Biogeo., 113, https://doi.org/10.1029/2008JG000703, 2008.
Bartholomé and Belward, A. S.: GLC2000: a new approach to global land cover mapping from Earth observation data, Int. J.
745
Remot. Sens., 26, 1959–1977, https://doi.org/10.1080/01431160412331291297, 2005.
Bastviken, D., Cole, J., Pace, M., and Tranvik, L.: Methane emissions from lakes: Dependence of lake characteristics, two
regional assessments, and a global estimate, 18, https://doi.org/10.1029/2004GB002238, 2004.
Beaulne, J., Garneau, M., Magnan, G., and Boucher, É.: Peat deposits store more carbon than trees in forested peatlands of
the boreal biome, 11, 2657, https://doi.org/10.1038/s41598-021-82004-x, 2021.
750
Beck, H. E., Pan, M., Miralles, D. G., Reichle, R. H., Dorigo, W. A., Hahn, S., Sheffield, J., Karthikeyan, L., Balsamo, G.,
Parinussa, R. M., van Dijk, A. I. J. M., Du, J., Kimball, J. S., Vergopolan, N., and Wood, E. F.: Evaluation of 18 satellite-
and model-based soil moisture products using in situ measurements from 826 sensors, 25, 17–40,
https://doi.org/10.5194/hess-25-17-2021, 2021.
Bohn, T. J., Melton, J. R., Ito, A., Kleinen, T., Spahni, R., Stocker, B. D., Zhang, B., Zhu, X., Schroeder, R., Glagolev, M.
755
V., Maksyutov, S., Brovkin, V., Chen, G., Denisov, S. N., Eliseev, A. V., Gallego-Sala, A., McDonald, K. C., Rawlins, M.
A., Riley, W. J., Subin, Z. M., Tian, H., Zhuang, Q., and Kaplan, J. O.: WETCHIMP-WSL: intercomparison of wetland
methane emissions models over West Siberia, 12, 3321–3349, https://doi.org/10.5194/bg-12-3321-2015, 2015.
Bridgham, S. D., Cadillo‐Quiroz, H., Keller, J. K., and Zhuang, Q.: Methane emissions from wetlands: biogeochemical,
microbial, and modeling perspectives from local to global scales, 19, 1325–1346, https://doi.org/10.1111/gcb.12131, 2013.
760
Brosius, L. S., Anthony, K. M. W., Treat, C. C., Lenz, J., Jones, M. C., Bret-Harte, M. S., and Grosse, G.: Spatiotemporal
patterns of northern lake formation since the Last Glacial Maximum, Quaternary Science Reviews, 253, 106773,
https://doi.org/10.1016/j.quascirev.2020.106773, 2021.
Brown, J., Ferrians, O., Heginbottom, J.A., and Melnikov, E. Circum-Arctic Map of Permafrost and Ground-Ice Conditions,
Version 2. Boulder, Colorado USA. NSIDC: National Snow and Ice Data Center. doi: https://doi.org/10.7265/skbg-kf16.
765
2002.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

31
Bruhwiler, L., Parmentier, F.-J. W., Crill, P., Leonard, M., and Palmer, P. I.: The Arctic Carbon Cycle and Its Response to
Changing Climate, Curr Clim Change Rep, 7, 14–34, https://doi.org/10.1007/s40641-020-00169-5, 2021.
Bryn, A., Strand, G.-H., Angeloff, M., and Rekdal, Y.: Land cover in Norway based on an area frame survey of vegetation
types, 72, 131–145, https://doi.org/10.1080/00291951.2018.1468356, 2018.
770
Bubier, J. L., Moore, T. R., Bellisario, L., Comer, N. T., and Crill, P. M.: Ecological controls on methane emissions from a
Northern Peatland Complex in the zone of discontinuous permafrost, Manitoba, Canada, 9, 455–470,
https://doi.org/10.1029/95GB02379, 1995.
Büttner, G.: CORINE Land Cover and Land Cover Change Products, in: Land Use and Land Cover Mapping in Europe:
Practices & Trends, edited by: Manakos, I. and Braun, M., Springer Netherlands, Dordrecht, 55–74,
775
https://doi.org/10.1007/978-94-007-7969-3_5, 2014.
Cael, B. B. and Seekell, D. A.: The size-distribution of Earth’s lakes, 6, 29633, https://doi.org/10.1038/srep29633, 2016.
Canada Committee on Ecological (Biophysical) Land Classification, National Wetlands Working Group, Warner, B. G., and
Rubec, C. D. A.: The Canadian wetland classification system, Wetlands Research Branch, University of Waterloo, Waterloo,
Ont., 1997.
780
Canadian Wetland Inventory Technical Committee, 2016. Canadian Wetland Inventory (Data Model), version 7.0. Prepared
by the Canadian Wetland Inventory Technical Committee. URL: http://www.ducks.ca/initiatives/canadian-wetland-
inventory/
Chasmer, L., Mahoney, C., Millard, K., Nelson, K., Peters, D., Merchant, M., Hopkinson, C., Brisco, B., Niemann, O.,
Montgomery, J., Devito, K., and Cobbaert, D.: Remote Sensing of Boreal Wetlands 2: Methods for Evaluating Boreal
785
Wetland Ecosystem State and Drivers of Change, 12, 1321, https://doi.org/10.3390/rs12081321, 2020.
Chen, J., Chen, J., Liao, A., Cao, X., Chen, L., Chen, X., He, C., Han, G., Peng, S., Lu, M., Zhang, W., Tong, X., and Mills,
J.: Global land cover mapping at 30m resolution: A POK-based operational approach, ISPRS Journal of Photogrammetry
and Remote Sensing, 103, 7–27, https://doi.org/10.1016/j.isprsjprs.2014.09.002, 2015.
Chen, Y., Hu, F. S., and Lara, M. J.: Divergent shrub-cover responses driven by climate, wildfire, and permafrost
790
interactions in Arctic tundra ecosystems, 27, 652–663, https://doi.org/10.1111/gcb.15451, 2021.
Cooley, S. W., Smith, L. C., Stepan, L., and Mascaro, J.: Tracking Dynamic Northern Surface Water Changes with High-
Frequency Planet CubeSat Imagery, 9, 1306, https://doi.org/10.3390/rs9121306, 2017.
Downing, J. A., Cole, J. J., Duarte, C. M., Middelburg, J. J., Melack, J. M., Prairie, Y. T., Kortelainen, P., Striegl, R. G.,
McDowell, W. H., and Tranvik, L. J.: Global abundance and size distribution of streams and rivers, 2, 229–236,
795
https://doi.org/10.5268/IW-2.4.502, 2012.
Duncan, B. N., Ott, L. E., Abshire, J. B., Brucker, L., Carroll, M. L., Carton, J., Comiso, J. C., Dinnat, E. P., Forbes, B. C.,
Gonsamo, A., Gregg, W. W., Hall, D. K., Ialongo, I., Jandt, R., Kahn, R. A., Karpechko, A., Kawa, S. R., Kato, S.,
Kumpula, T., Kyrölä, E., Loboda, T. V., McDonald, K. C., Montesano, P. M., Nassar, R., Neigh, C. S. R., Parkinson, C. L.,
Poulter, B., Pulliainen, J., Rautiainen, K., Rogers, B. M., Rousseaux, C. S., Soja, A. J., Steiner, N., Tamminen, J., Taylor, P.
800
C., Tzortziou, M. A., Virta, H., Wang, J. S., Watts, J. D., Winker, D. M., and Wu, D. L.: Space-Based Observations for
Understanding Changes in the Arctic-Boreal Zone, 58, e2019RG000652, https://doi.org/10.1029/2019RG000652, 2020.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

32
Fick, S. E. and Hijmans, R. J.: WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas, 37, 4302–
4315, https://doi.org/10.1002/joc.5086, 2017.
Fluet-Chouinard, E., Lehner, B., Rebelo, L.-M., Papa, F., and Hamilton, S. K.: Development of a global inundation map at
805
high spatial resolution from topographic downscaling of coarse-scale remote sensing data, Remote Sensing of Environment,
158, 348–361, https://doi.org/10.1016/j.rse.2014.10.015, 2015.
Garneau, M., Tremblay, L., and Magnan, G.: Holocene pool formation in oligotrophic fens from boreal Québec in
northeastern Canada, The Holocene, 28, 396–407, https://doi.org/10.1177/0959683617729439, 2018.
Glagolev, M., Kleptsova, I., Filippov, I., Maksyutov, S., and Machida, T.: Regional methane emission from West Siberia
810
mire landscapes, Environ. Res. Lett., 6, 045214, https://doi.org/10.1088/1748-9326/6/4/045214, 2011.
Glaser, P. H., Siegel, D. I., Reeve, A. S., Janssens, J. A., and Janecky, D. R.: Tectonic drivers for vegetation patterning and
landscape evolution in the Albany River region of the Hudson Bay Lowlands, 92, 1054–1070,
https://doi.org/10.1111/j.0022-0477.2004.00930.x, 2004.
Gorham, E. Northern Peatlands: Role in the Carbon Cycle and Probable Responses to Climatic Warming, Ecological
815
Applications, 1, 182-195, https://doi.org/10.2307/1941811, 1991.
Grosse, G., Jones, B., and Arp, C.: 8.21 Thermokarst Lakes, Drainage, and Drained Basins, in: Treatise on Geomorphology,
edited by: Shroder, J. F., Academic Press, San Diego, 325–353, https://doi.org/10.1016/B978-0-12-374739-6.00216-5, 2013.
Gruber, S.: Derivation and analysis of a high-resolution estimate of global permafrost zonation, 6, 221–233,
https://doi.org/10.5194/tc-6-221-2012, 2012.
820
Gunnarsson, U., Löfroth, M., and Sandring, S.: The Swedish wetland survey: compiled excerpts from the national final
report, Swedish Environmental Protection Agency, Stockholm, 37 pp., 2014.
Harris, L. I., Roulet, N. T., and Moore, T. R.: Mechanisms for the Development of Microform Patterns in Peatlands of the
Hudson Bay Lowland, Ecosystems, 23, 741–767, https://doi.org/10.1007/s10021-019-00436-z, 2020.
Heikkinen, J. E. P., Virtanen, T., Huttunen, J. T., Elsakov, V., and Martikainen, P. J.: Carbon balance in East European
825
tundra, 18, https://doi.org/10.1029/2003GB002054, 2004.
Heiskanen, L., Tuovinen, J.-P., Räsänen, A., Virtanen, T., Juutinen, S., Lohila, A., Penttilä, T., Linkosalmi, M., Mikola, J.,
Laurila, T., and Aurela, M.: Carbon dioxide and methane exchange of a patterned subarctic fen during two contrasting
growing seasons, 18, 873–896, https://doi.org/10.5194/bg-18-873-2021, 2021.
Helbig, M., Pappas, C., and Sonnentag, O.: Permafrost thaw and wildfire: Equally important drivers of boreal tree cover
830
changes in the Taiga Plains, Canada, 43, 1598–1606, https://doi.org/10.1002/2015GL067193, 2016.
Heslop, J. K., Walter Anthony, K. M., Winkel, M., Sepulveda-Jauregui, A., Martinez-Cruz, K., Bondurant, A., Grosse, G.,
and Liebner, S.: A synthesis of methane dynamics in thermokarst lake environments, Earth-Science Reviews, 210, 103365,
https://doi.org/10.1016/j.earscirev.2020.103365, 2020.
Holgerson, M. A. and Raymond, P. A.: Large contribution to inland water CO 2 and CH 4 emissions from very small ponds,
835
9, 222–226, https://doi.org/10.1038/ngeo2654, 2016.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

33
Homer, C., Dewitz, J., Jin, S., Xian, G., Costello, C., Danielson, P., Gass, L., Funk, M., Wickham, J., Stehman, S., Auch, R.,
and Riitters, K.: Conterminous United States land cover change patterns 2001–2016 from the 2016 National Land Cover
Database, ISPRS Journal of Photogrammetry and Remote Sensing, 162, 184–199,
https://doi.org/10.1016/j.isprsjprs.2020.02.019, 2020.
840
Hugelius, G., Tarnocai, C., Broll, G., Canadell, J. G., Kuhry, P., and Swanson, D. K.: The Northern Circumpolar Soil Carbon
Database: spatially distributed datasets of soil coverage and soil carbon storage in the northern permafrost regions, 5, 3–13,
https://doi.org/10.5194/essd-5-3-2013, 2013.
Hugelius, G., Strauss, J., Zubrzycki, S., Harden, J. W., Schuur, E. a. G., Ping, C.-L., Schirrmeister, L., Grosse, G.,
Michaelson, G. J., Koven, C. D., O’Donnell, J. A., Elberling, B., Mishra, U., Camill, P., Yu, Z., Palmtag, J., and Kuhry, P.:
845
Estimated stocks of circumpolar permafrost carbon with quantified uncertainty ranges and identified data gaps, 11, 6573–
6593, https://doi.org/10.5194/bg-11-6573-2014, 2014.
Hugelius, G., Loisel, J., Chadburn, S., Jackson, R. B., Jones, M., MacDonald, G., Marushchak, M., Olefeldt, D., Packalen,
M., Siewert, M. B., Treat, C., Turetsky, M., Voigt, C., and Yu, Z.: Large stocks of peatland carbon and nitrogen are
vulnerable to permafrost thaw, PNAS, 117, 20438–20446, https://doi.org/10.1073/pnas.1916387117, 2020.
850
Ito, A.: Methane emission from pan-Arctic natural wetlands estimated using a process-based model, 1901–2016, Polar
Science, 21, 26–36, https://doi.org/10.1016/j.polar.2018.12.001, 2019.
Jorgenson, M. T., Racine, C. H., Walters, J. C., and Osterkamp, T. E.: Permafrost Degradation and Ecological Changes
Associated with a WarmingClimate in Central Alaska, Climatic Change, 48, 551–579,
https://doi.org/10.1023/A:1005667424292, 2001.
855
Juncher Jørgensen, C., Lund Johansen, K. M., Westergaard-Nielsen, A., and Elberling, B.: Net regional methane sink in
High Arctic soils of northeast Greenland, 8, 20–23, https://doi.org/10.1038/ngeo2305, 2015.
Juutinen, S., Alm, J., Larmola, T., Huttunen, J. T., Morero, M., Martikainen, P. J., and Silvola, J.: Major implication of the
littoral zone for methane release from boreal lakes, 17, https://doi.org/10.1029/2003GB002105, 2003.
Kassambara, A., and Mundt, F. factoextra: Extract and Visualize the Results of Multivariate Data Analyses. R package
860
version 1.0.7. https://CRAN.R-project.org/package=factoextra, 2020.
Knoblauch, C., Spott, O., Evgrafova, S., Kutzbach, L., and Pfeiffer, E.-M.: Regulation of methane production, oxidation, and
emission by vascular plants and bryophytes in ponds of the northeast Siberian polygonal tundra, 120, 2525–2541,
https://doi.org/10.1002/2015JG003053, 2015.
Knox, S. H., Jackson, R. B., Poulter, B., McNicol, G., Fluet-Chouinard, E., Zhang, Z., Hugelius, G., Bousquet, P., Canadell,
865
J. G., Saunois, M., Papale, D., Chu, H., Keenan, T. F., Baldocchi, D., Torn, M. S., Mammarella, I., Trotta, C., Aurela, M.,
Bohrer, G., Campbell, D. I., Cescatti, A., Chamberlain, S., Chen, J., Chen, W., Dengel, S., Desai, A. R., Euskirchen, E.,
Friborg, T., Gasbarra, D., Goded, I., Goeckede, M., Heimann, M., Helbig, M., Hirano, T., Hollinger, D. Y., Iwata, H., Kang,
M., Klatt, J., Krauss, K. W., Kutzbach, L., Lohila, A., Mitra, B., Morin, T. H., Nilsson, M. B., Niu, S., Noormets, A., Oechel,
W. C., Peichl, M., Peltola, O., Reba, M. L., Richardson, A. D., Runkle, B. R. K., Ryu, Y., Sachs, T., Schäfer, K. V. R.,
870
Schmid, H. P., Shurpali, N., Sonnentag, O., Tang, A. C. I., Ueyama, M., Vargas, R., Vesala, T., Ward, E. J., Windham-
Myers, L., Wohlfahrt, G., and Zona, D.: FLUXNET-CH4 Synthesis Activity: Objectives, Observations, and Future
Directions, 100, 2607–2632, https://doi.org/10.1175/BAMS-D-18-0268.1, 2019.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

34
Kremenetski, K. V., Velichko, A. A., Borisova, O. K., MacDonald, G. M., Smith, L. C., Frey, K. E., and Orlova, L. A.:
Peatlands of the Western Siberian lowlands: current knowledge on zonation, carbon content and Late Quaternary history,
875
Quaternary Science Reviews, 22, 703–723, https://doi.org/10.1016/S0277-3791(02)00196-8, 2003.
Kuhn, M. caret: Classification and Regression Training. R package version 6.0-86. https://CRAN.R-
project.org/package=caret, 2020
Kuhn, M., Varner, R., Bastviken, D., Crill, P., MacIntyre, S., Turetsky, M.R., Walter Anthony, K., McGuire, A.D., and
Olefeldt, D.: BAWLD-CH4: Methane Fluxes from Boreal and Arctic Ecosystems, urn:node:ARCTIC.
880
doi:10.18739/A27H1DN5S, 2021
Lara, M. J., and Chipman, M. L.: Periglacial Lake Origin Influences the Likelihood of Lake Drainage in Northern Alaska.
Remote Sensing, 13(5), 853, https://doi.org/10.3390/rs13050852, 2021
Lara, M. J., Nitze, I., Grosse, G., and McGuire, A. D.: Tundra landform and vegetation productivity trend maps for the
Arctic Coastal Plain of northern Alaska, 5, 180058, https://doi.org/10.1038/sdata.2018.58, 2018.
885
Lau, M. C. Y., Stackhouse, B. T., Layton, A. C., Chauhan, A., Vishnivetskaya, T. A., Chourey, K., Ronholm, J., Mykytczuk,
N. C. S., Bennett, P. C., Lamarche-Gagnon, G., Burton, N., Pollard, W. H., Omelon, C. R., Medvigy, D. M., Hettich, R. L.,
Pfiffner, S. M., Whyte, L. G., and Onstott, T. C.: An active atmospheric methane sink in high Arctic mineral cryosols, 9,
1880–1891, https://doi.org/10.1038/ismej.2015.13, 2015.
Lehner, B. and Döll, P.: Development and validation of a global database of lakes, reservoirs and wetlands, Journal of
890
Hydrology, 296, 1–22, https://doi.org/10.1016/j.jhydrol.2004.03.028, 2004.
Li, M., Peng, C., Zhu, Q., Zhou, X., Yang, G., Song, X., and Zhang, K.: The significant contribution of lake depth in
regulating global lake diffusive methane emissions, Water Research, 172, 115465,
https://doi.org/10.1016/j.watres.2020.115465, 2020.
Liaw, A., and Wiener, M. Classification and Regression by randomForest. R News 2(3), 18—22, 2002
895
Liljedahl, A. K., Boike, J., Daanen, R. P., Fedorov, A. N., Frost, G. V., Grosse, G., Hinzman, L. D., Iijma, Y., Jorgenson, J.
C., Matveyeva, N., Necsoiu, M., Raynolds, M. K., Romanovsky, V. E., Schulla, J., Tape, K. D., Walker, D. A., Wilson, C. J.,
Yabuki, H., and Zona, D.: Pan-Arctic ice-wedge degradation in warming permafrost and its influence on tundra hydrology,
9, 312–318, https://doi.org/10.1038/ngeo2674, 2016.
Linke, S., Lehner, B., Ouellet Dallaire, C., Ariwi, J., Grill, G., Anand, M., Beames, P., Burchard-Levine, V., Maxwell, S.,
900
Moidu, H., Tan, F., and Thieme, M.: Global hydro-environmental sub-basin and river reach characteristics at high spatial
resolution, 6, 283, https://doi.org/10.1038/s41597-019-0300-6, 2019.
Loisel, J., Gallego-Sala, A. V., Amesbury, M. J., Magnan, G., Anshari, G., Beilman, D. W., Benavides, J. C., Blewett, J.,
Camill, P., Charman, D. J., Chawchai, S., Hedgpeth, A., Kleinen, T., Korhola, A., Large, D., Mansilla, C. A., Müller, J., van
Bellen, S., West, J. B., Yu, Z., Bubier, J. L., Garneau, M., Moore, T., Sannel, A. B. K., Page, S., Väliranta, M., Bechtold, M.,
905
Brovkin, V., Cole, L. E. S., Chanton, J. P., Christensen, T. R., Davies, M. A., De Vleeschouwer, F., Finkelstein, S. A.,
Frolking, S., Gałka, M., Gandois, L., Girkin, N., Harris, L. I., Heinemeyer, A., Hoyt, A. M., Jones, M. C., Joos, F., Juutinen,
S., Kaiser, K., Lacourse, T., Lamentowicz, M., Larmola, T., Leifeld, J., Lohila, A., Milner, A. M., Minkkinen, K., Moss, P.,
Naafs, B. D. A., Nichols, J., O’Donnell, J., Payne, R., Philben, M., Piilo, S., Quillet, A., Ratnayake, A. S., Roland, T. P.,
Sjögersten, S., Sonnentag, O., Swindles, G. T., Swinnen, W., Talbot, J., Treat, C., Valach, A. C., and Wu, J.: Expert
910
assessment of future vulnerability of the global peatland carbon sink, 11, 70–77, https://doi.org/10.1038/s41558-020-00944-
0, 2021.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

35
Machacova, K., Bäck, J., Vanhatalo, A., Halmeenmäki, E., Kolari, P., Mammarella, I., Pumpanen, J., Acosta, M., Urban, O.,
and Pihlatie, M.: Pinus sylvestris as a missing source of nitrous oxide and methane in boreal forest, 6, 23410,
https://doi.org/10.1038/srep23410, 2016.
915
Malhotra, A. and Roulet, N. T.: Environmental correlates of peatland carbon fluxes in a thawing landscape: do transitional
thaw stages matter?, 12, 3119–3130, https://doi.org/10.5194/bg-12-3119-2015, 2015.
Marushchak, M. E., Friborg, T., Biasi, C., Herbst, M., Johansson, T., Kiepe, I., Liimatainen, M., Lind, S. E., Martikainen, P.
J., Virtanen, T., Soegaard, H., and Shurpali, N. J.: Methane dynamics in the subarctic tundra: combining stable isotope
analyses, plot- and ecosystem-scale flux measurements, 13, 597–608, https://doi.org/10.5194/bg-13-597-2016, 2016.
920
Masing, V., Botch, M., and Läänelaid, A.: Mires of the former Soviet Union, Wetlands Ecol Manage, 18, 397–433,
https://doi.org/10.1007/s11273-008-9130-6, 2010.
Matson, A., Pennock, D., and Bedard-Haughn, A.: Methane and nitrous oxide emissions from mature forest stands in the
boreal forest, Saskatchewan, Canada, Forest Ecology and Management, 258, 1073–1083,
https://doi.org/10.1016/j.foreco.2009.05.034, 2009.
925
Matthews, E. and Fung, I.: Methane emission from natural wetlands: Global distribution, area, and environmental
characteristics of sources, 1, 61–86, https://doi.org/10.1029/GB001i001p00061, 1987.
McGuire, A. D., Christensen, T. R., Hayes, D., Heroult, A., Euskirchen, E., Kimball, J. S., Koven, C., Lafleur, P., Miller, P.
A., Oechel, W., Peylin, P., Williams, M., and Yi, Y.: An assessment of the carbon balance of Arctic tundra: comparisons
among observations, process models, and atmospheric inversions, 9, 3185–3204, https://doi.org/10.5194/bg-9-3185-2012,
930
2012.
Melton, J. R., Wania, R., Hodson, E. L., Poulter, B., Ringeval, B., Spahni, R., Bohn, T., Avis, C. A., Beerling, D. J., Chen,
G., Eliseev, A. V., Denisov, S. N., Hopcroft, P. O., Lettenmaier, D. P., Riley, W. J., Singarayer, J. S., Subin, Z. M., Tian, H.,
Zürcher, S., Brovkin, V., Bodegom, P. M. van, Kleinen, T., Yu, Z. C., and Kaplan, J. O.: Present state of global wetland
extent and wetland methane modelling: conclusions from a model inter-comparison project (WETCHIMP), 10, 753–788,
935
https://doi.org/10.5194/bg-10-753-2013, 2013.
Messager, M. L., Lehner, B., Grill, G., Nedeva, I., and Schmitt, O.: Estimating the volume and age of water stored in global
lakes using a geo-statistical approach, 7, 13603, https://doi.org/10.1038/ncomms13603, 2016.
Miron B. Kursa, Witold R. Rudnicki. Feature Selection with the Boruta Package. Journal of Statistical Software, 36(11), 1-
13. URL http://www.jstatsoft.org/v36/i11/, 2010
940
van der Molen, M. K., van Huissteden, J., Parmentier, F. J. W., Petrescu, A. M. R., Dolman, A. J., Maximov, T. C.,
Kononov, A. V., Karsanaev, S. V., and Suzdalov, D. A.: The growing season greenhouse gas balance of a continental tundra
site in the Indigirka lowlands, NE Siberia, 4, 985–1003, https://doi.org/10.5194/bg-4-985-2007, 2007.
Muster, S., Roth, K., Langer, M., Lange, S., Cresto Aleina, F., Bartsch, A., Morgenstern, A., Grosse, G., Jones, B., Sannel,
A. B. K., Sjöberg, Y., Günther, F., Andresen, C., Veremeeva, A., Lindgren, P. R., Bouchard, F., Lara, M. J., Fortier, D.,
945
Charbonneau, S., Virtanen, T. A., Hugelius, G., Palmtag, J., Siewert, M. B., Riley, W. J., Koven, C. D., and Boike, J.: PeRL:
a circum-Arctic Permafrost Region Pond and Lake database, 9, 317–348, https://doi.org/10.5194/essd-9-317-2017, 2017.
Muster, S., Riley, W. J., Roth, K., Langer, M., Cresto Aleina, F., Koven, C. D., Lange, S., Bartsch, A., Grosse, G., Wilson,
C. J., Jones, B. M., and Boike, J.: Size Distributions of Arctic Waterbodies Reveal Consistent Relations in Their Statistical
Moments in Space and Time, Front. Earth Sci., 7, https://doi.org/10.3389/feart.2019.00005, 2019.
950
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

36
Olefeldt, D., Turetsky, M. R., Crill, P. M., and McGuire, A. D.: Environmental and physical controls on northern terrestrial
methane emissions across permafrost zones, 19, 589–603, https://doi.org/10.1111/gcb.12071, 2013.
Olefeldt, D., Goswami, S., Grosse, G., Hayes, D., Hugelius, G., Kuhry, P., McGuire, A. D., Romanovsky, V. E., Sannel, A.
B. K., Schuur, E. a. G., and Turetsky, M. R.: Circumpolar distribution and carbon storage of thermokarst landscapes, 7,
13043, https://doi.org/10.1038/ncomms13043, 2016.
955
Olefeldt, D., Euskirchen, E. S., Harden, J., Kane, E., McGuire, A. D., Waldrop, M. P., and Turetsky, M. R.: A decade of
boreal rich fen greenhouse gas fluxes in response to natural and experimental water table variability, 23, 2428–2440,
https://doi.org/10.1111/gcb.13612, 2017.
Olefeldt, D., Hovemyr, M., Kuhn, MA, Bastviken, D., Bohn, T.J., Connolly, J., Crill, P., Euskirchen, E.S., Finkelstein, S.A.,
Genet, H., Grosse, G., Harris, L.I., Heffernan, L., Helbig, M., Hugelius, G., Hutchins, R., Juutinen, S., Lara, M.J., Malhotra,
960
A., Manies, K., McGuire, A.D., Natali, S.M., O’Donnell, J.A., Parmentier, F.-J.W., Räsänen, A., Schädel, C., Sonnentag, O.,
Strack, M., Tank, S.E., Treat, C., Varner, R.K., Virtanen, T., Warren, R.K., and Watts, J.D.: The fractional land cover
estimates from the Boreal-Arctic Wetland and Lake Dataset (BAWLD), 2021. Arctic Data Center.
doi:10.18739/A2C824F9X, 2021.
Olson, D. M., Dinerstein, E., Wikramanayake, E. D., Burgess, N. D., Powell, G. V. N., Underwood, E. C., D’amico, J. A.,
965
Itoua, I., Strand, H. E., Morrison, J. C., Loucks, C. J., Allnutt, T. F., Ricketts, T. H., Kura, Y., Lamoreux, J. F., Wettengel,
W. W., Hedao, P., and Kassem, K. R.: Terrestrial Ecoregions of the World: A New Map of Life on Earth: A new global map
of terrestrial ecoregions provides an innovative tool for conserving biodiversity, BioScience, 51, 933–938,
https://doi.org/10.1641/0006-3568(2001)051[0933:TEOTWA]2.0.CO;2, 2001.
Olson, D. M., Griffis, T. J., Noormets, A., Kolka, R., and Chen, J.: Interannual, seasonal, and retrospective analysis of the
970
methane and carbon dioxide budgets of a temperate peatland, 118, 226–238, https://doi.org/10.1002/jgrg.20031, 2013.
Packalen, M. S., Finkelstein, S. A., and McLaughlin, J. W.: Climate and peat type in relation to spatial variation of the
peatland carbon mass in the Hudson Bay Lowlands, Canada, 121, 1104–1117, https://doi.org/10.1002/2015JG002938, 2016.
Pekel, J.-F., Cottam, A., Gorelick, N., and Belward, A. S.: High-resolution mapping of global surface water and its long-term
changes, 540, 418–422, https://doi.org/10.1038/nature20584, 2016.
975
Pelletier, L., Moore, T. R., Roulet, N. T., Garneau, M., and Beaulieu‐Audy, V.: Methane fluxes from three peatlands in the
La Grande Rivière watershed, James Bay lowland, Canada, 112, https://doi.org/10.1029/2006JG000216, 2007.
Peltola, O., Vesala, T., Gao, Y., Räty, O., Alekseychik, P., Aurela, M., Chojnicki, B., Desai, A. R., Dolman, A. J.,
Euskirchen, E. S., Friborg, T., Göckede, M., Helbig, M., Humphreys, E., Jackson, R. B., Jocher, G., Joos, F., Klatt, J., Knox,
S. H., Kowalska, N., Kutzbach, L., Lienert, S., Lohila, A., Mammarella, I., Nadeau, D. F., Nilsson, M. B., Oechel, W. C.,
980
Peichl, M., Pypker, T., Quinton, W., Rinne, J., Sachs, T., Samson, M., Schmid, H. P., Sonnentag, O., Wille, C., Zona, D.,
and Aalto, T.: Monthly gridded data product of northern wetland methane emissions based on upscaling eddy covariance
observations, 11, 1263–1289, https://doi.org/10.5194/essd-11-1263-2019, 2019.
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna,
Austria. URL https://www.R-project.org/, 2020
985
Räsänen, A. and Virtanen, T.: Data and resolution requirements in mapping vegetation in spatially heterogeneous
landscapes, Remote Sensing of Environment, 230, 111207, https://doi.org/10.1016/j.rse.2019.05.026, 2019.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

37
Raynolds, M. K., Walker, D. A., Balser, A., Bay, C., Campbell, M., Cherosov, M. M., Daniëls, F. J. A., Eidesen, P. B.,
Ermokhina, K. A., Frost, G. V., Jedrzejek, B., Jorgenson, M. T., Kennedy, B. E., Kholod, S. S., Lavrinenko, I. A.,
Lavrinenko, O. V., Magnússon, B., Matveyeva, N. V., Metúsalemsson, S., Nilsen, L., Olthof, I., Pospelov, I. N., Pospelova,
990
E. B., Pouliot, D., Razzhivin, V., Schaepman-Strub, G., Šibík, J., Telyatnikov, M. Yu., and Troeva, E.: A raster version of
the Circumpolar Arctic Vegetation Map (CAVM), Remote Sensing of Environment, 232, 111297,
https://doi.org/10.1016/j.rse.2019.111297, 2019.
Rubec, C.: The Canadian Wetland Classification System, in: The Wetland Book: I: Structure and Function, Management,
and Methods, edited by: Finlayson, C. M., Everard, M., Irvine, K., McInnes, R. J., Middleton, B. A., van Dam, A. A., and
995
Davidson, N. C., Springer Netherlands, Dordrecht, 1577–1581, https://doi.org/10.1007/978-90-481-9659-3_340, 2018.
Sannel, A. B. K. and Kuhry, P.: Warming-induced destabilization of peat plateau/thermokarst lake complexes, 116,
https://doi.org/10.1029/2010JG001635, 2011.
Saunois, M., Stavert, A. R., Poulter, B., Bousquet, P., Canadell, J. G., Jackson, R. B., Raymond, P. A., Dlugokencky, E. J.,
Houweling, S., Patra, P. K., Ciais, P., Arora, V. K., Bastviken, D., Bergamaschi, P., Blake, D. R., Brailsford, G., Bruhwiler,
1000
L., Carlson, K. M., Carrol, M., Castaldi, S., Chandra, N., Crevoisier, C., Crill, P. M., Covey, K., Curry, C. L., Etiope, G.,
Frankenberg, C., Gedney, N., Hegglin, M. I., Höglund-Isaksson, L., Hugelius, G., Ishizawa, M., Ito, A., Janssens-Maenhout,
G., Jensen, K. M., Joos, F., Kleinen, T., Krummel, P. B., Langenfelds, R. L., Laruelle, G. G., Liu, L., Machida, T.,
Maksyutov, S., McDonald, K. C., McNorton, J., Miller, P. A., Melton, J. R., Morino, I., Müller, J., Murguia-Flores, F., Naik,
V., Niwa, Y., Noce, S., O’Doherty, S., Parker, R. J., Peng, C., Peng, S., Peters, G. P., Prigent, C., Prinn, R., Ramonet, M.,
1005
Regnier, P., Riley, W. J., Rosentreter, J. A., Segers, A., Simpson, I. J., Shi, H., Smith, S. J., Steele, L. P., Thornton, B. F.,
Tian, H., Tohjima, Y., Tubiello, F. N., Tsuruta, A., Viovy, N., Voulgarakis, A., Weber, T. S., van Weele, M., van der Werf,
G. R., Weiss, R. F., Worthy, D., Wunch, D., Yin, Y., Yoshida, Y., Zhang, W., Zhang, Z., Zhao, Y., Zheng, B., Zhu, Q., Zhu,
Q., and Zhuang, Q.: The Global Methane Budget 2000–2017, 12, 1561–1623, https://doi.org/10.5194/essd-12-1561-2020,
2020.
1010
Sayedi, S. S., Abbott, B. W., Thornton, B. F., Frederick, J. M., Vonk, J. E., Overduin, P., Schädel, C., Schuur, E. A. G.,
Bourbonnais, A., Demidov, N., Gavrilov, A., He, S., Hugelius, G., Jakobsson, M., Jones, M. C., Joung, D., Kraev, G.,
Macdonald, R. W., McGuire, A. D., Mu, C., O’Regan, M., Schreiner, K. M., Stranne, C., Pizhankova, E., Vasiliev, A.,
Westermann, S., Zarnetske, J. P., Zhang, T., Ghandehari, M., Baeumler, S., Brown, B. C., and Frei, R. J.: Subsea permafrost
carbon stocks and climate change sensitivity estimated by expert assessment, Environ. Res. Lett., 15, 124075,
1015
https://doi.org/10.1088/1748-9326/abcc29, 2020.
Schirrmeister, L., Froese, D., Tumskoy, V., Grosse, G., and Wetterich, S.: PERMAFROST AND PERIGLACIAL
FEATURES | Yedoma: Late Pleistocene Ice-Rich Syngenetic Permafrost of Beringia, in: Encyclopedia of Quaternary
Science (Second Edition), edited by: Elias, S. A. and Mock, C. J., Elsevier, Amsterdam, 542–552,
https://doi.org/10.1016/B978-0-444-53643-3.00106-0, 2013.
1020
Schneider von Deimling, T., Grosse, G., Strauss, J., Schirrmeister, L., Morgenstern, A., Schaphoff, S., Meinshausen, M., and
Boike, J.: Observation-based modelling of permafrost carbon fluxes with accounting for deep carbon deposits and
thermokarst activity, 12, 3469–3488, https://doi.org/10.5194/bg-12-3469-2015, 2015.
Seppälä, M.: Synthesis of studies of palsa formation underlining the importance of local environmental and physical
characteristics, 75, 366–370, https://doi.org/10.1016/j.yqres.2010.09.007, 2011.
1025
Smith, L. C., Sheng, Y., and MacDonald, G. M.: A first pan-Arctic assessment of the influence of glaciation, permafrost,
topography and peatlands on northern hemisphere lake distribution, 18, 201–208, https://doi.org/10.1002/ppp.581, 2007.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

38
Song, J.: Bias corrections for Random Forest in regression using residual rotation, 2, 321–326,
https://doi.org/10.1016/j.jkss.2015.01.003, 2015.
St Pierre, K. A., Danielsen, B. K., Hermesdorf, L., D’Imperio, L., Iversen, L. L., and Elberling, B.: Drivers of net methane
1030
uptake across Greenlandic dry heath tundra landscapes, Soil Biology and Biochemistry, 138, 107605,
https://doi.org/10.1016/j.soilbio.2019.107605, 2019.
Stanley, E. H., Casson, N. J., Christel, S. T., Crawford, J. T., Loken, L. C., and Oliver, S. K.: The ecology of methane in
streams and rivers: patterns, controls, and global significance, 86, 146–171, https://doi.org/10.1890/15-1027, 2016.
Strauss, J., Schirrmeister, L., Grosse, G., Fortier, D., Hugelius, G., Knoblauch, C., Romanovsky, V., Schädel, C., Schneider
1035
von Deimling, T., Schuur, E. A. G., Shmelev, D., Ulrich, M., and Veremeeva, A.: Deep Yedoma permafrost: A synthesis of
depositional characteristics and carbon vulnerability, Earth-Science Reviews, 172, 75–86,
https://doi.org/10.1016/j.earscirev.2017.07.007, 2017.
Tan, Z., Zhuang, Q., Henze, D. K., Frankenberg, C., Dlugokencky, E., Sweeney, C., Turner, A. J., Sasakawa, M., and
Machida, T.: Inverse modeling of pan-Arctic methane emissions at high spatial resolution: what can we learn from
1040
assimilating satellite retrievals and using different process-based wetland and lake biogeochemical models?, 16, 12649–
12666, https://doi.org/10.5194/acp-16-12649-2016, 2016.
Tarnocai, C.: The effect of climate change on carbon in Canadian peatlands, Global and Planetary Change, 53, 222–232,
https://doi.org/10.1016/j.gloplacha.2006.03.012, 2006.
Terentieva, I. E., Glagolev, M. V., Lapshina, E. D., Sabrekov, A. F., and Maksyutov, S.: Mapping of West Siberian taiga
1045
wetland complexes using Landsat imagery: implications for methane emissions, 13, 4615–4626, https://doi.org/10.5194/bg-
13-4615-2016, 2016.
Terentieva, I. E., Sabrekov, A. F., Ilyasov, D., Ebrahimi, A., Glagolev, M. V., and Maksyutov, S.: Highly Dynamic Methane
Emission from the West Siberian Boreal Floodplains, Wetlands, 39, 217–226, https://doi.org/10.1007/s13157-018-1088-4,
2019.
1050
Thompson, R. L., Nisbet, E. G., Pisso, I., Stohl, A., Blake, D., Dlugokencky, E. J., Helmig, D., and White, J. W. C.:
Variability in Atmospheric Methane From Fossil Fuel and Microbial Sources Over the Last Three Decades, 45, 11,499-
11,508, https://doi.org/10.1029/2018GL078127, 2018.
Thornton, B. F., Wik, M., and Crill, P. M.: Double-counting challenges the accuracy of high-latitude methane inventories,
43, 12,569-12,577, https://doi.org/10.1002/2016GL071772, 2016.
1055
Treat, C. C., Bloom, A. A., and Marushchak, M. E.: Nongrowing season methane emissions–a significant component of
annual emissions across northern ecosystems, 24, 3331–3343, https://doi.org/10.1111/gcb.14137, 2018.
Turetsky, M. R., Wieder, R. K., and Vitt, D. H.: Boreal peatland C fluxes under varying permafrost regimes, Soil Biology
and Biochemistry, 34, 907–912, https://doi.org/10.1016/S0038-0717(02)00022-6, 2002.
Turetsky, M. R., Kotowska, A., Bubier, J., Dise, N. B., Crill, P., Hornibrook, E. R. C., Minkkinen, K., Moore, T. R., Myers‐
1060
Smith, I. H., Nykänen, H., Olefeldt, D., Rinne, J., Saarnio, S., Shurpali, N., Tuittila, E.-S., Waddington, J. M., White, J. R.,
Wickland, K. P., and Wilmking, M.: A synthesis of methane emissions from 71 northern, temperate, and subtropical
wetlands, 20, 2183–2197, https://doi.org/10.1111/gcb.12580, 2014.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

39
Väliranta, M., Salojärvi, N., Vuorsalo, A., Juutinen, S., Korhola, A., Luoto, M., and Tuittila, E.-S.: Holocene fen–bog
transitions, current status in Finland and future perspectives, The Holocene, 27, 752–764,
1065
https://doi.org/10.1177/0959683616670471, 2017.
Venter, O., Sanderson, E. W., Magrach, A., Allan, J. R., Beher, J., Jones, K. R., Possingham, H. P., Laurance, W. F., Wood,
P., Fekete, B. M., Levy, M. A., and Watson, J. E. M.: Sixteen years of change in the global terrestrial human footprint and
implications for biodiversity conservation, 7, 12558, https://doi.org/10.1038/ncomms12558, 2016.
Virtanen, R., Oksanen, L., Oksanen, T., Cohen, J., Forbes, B. C., Johansen, B., Käyhkö, J., Olofsson, J., Pulliainen, J., and
1070
Tømmervik, H.: Where do the treeless tundra areas of northern highlands fit in the global biome system: toward an
ecologically natural subdivision of the tundra biome, 6, 143–158, https://doi.org/10.1002/ece3.1837, 2016.
Virtanen, T. and Ek, M.: The fragmented nature of tundra landscape, International Journal of Applied Earth Observation and
Geoinformation, 27, 4–12, https://doi.org/10.1016/j.jag.2013.05.010, 2014.
Vitt, D. H. and Chee, W.-L.: The relationships of vegetation to surface water chemistry and peat chemistry in fens of
1075
Alberta, Canada, Vegetatio, 89, 87–106, https://doi.org/10.1007/BF00032163, 1990.
Vitt, D. H., Halsey, L. A., Bauer, I. E., and Campbell, C.: Spatial and temporal trends in carbon storage of peatlands of
continental western Canada through the Holocene, 37, 12, 2000a.
Vitt, D. H., Halsey, L. A., and Zoltai, S. C.: The changing landscape of Canada’s western boreal forest: the current dynamics
of permafrost, Can. J. For. Res., 30, 283–287, https://doi.org/10.1139/x99-214, 2000b.
1080
Walker, D. A., Raynolds, M. K., Daniëls, F. J. A., Einarsson, E., Elvebakk, A., Gould, W. A., Katenin, A. E., Kholod, S. S.,
Markon, C. J., Melnikov, E. S., Moskalenko, N. G., Talbot, S. S., Yurtsev, B. A. (†), and Team, T. other members of the C.:
The Circumpolar Arctic vegetation map, 16, 267–282, https://doi.org/10.1111/j.1654-1103.2005.tb02365.x, 2005.
Wallin, M. B., Campeau, A., Audet, J., Bastviken, D., Bishop, K., Kokic, J., Laudon, H., Lundin, E., Löfgren, S.,
Natchimuthu, S., Sobek, S., Teutschbein, C., Weyhenmeyer, G. A., and Grabs, T.: Carbon dioxide and methane emissions of
1085
Swedish low-order streams—a national estimate and lessons learnt from more than a decade of observations, 3, 156–167,
https://doi.org/10.1002/lol2.10061, 2018.
Walter Anthony, K., Daanen, R., Anthony, P., Schneider von Deimling, T., Ping, C.-L., Chanton, J. P., and Grosse, G.:
Methane emissions proportional to permafrost carbon thawed in Arctic lakes since the 1950s, 9, 679–682,
https://doi.org/10.1038/ngeo2795, 2016.
1090
Walter Anthony, K., Schneider von Deimling, T., Nitze, I., Frolking, S., Emond, A., Daanen, R., Anthony, P., Lindgren, P.,
Jones, B., and Grosse, G.: 21st-century modeled permafrost carbon emissions accelerated by abrupt thaw beneath lakes, 9,
3262, https://doi.org/10.1038/s41467-018-05738-9, 2018.
Watts, J. D., Kimball, J. S., Bartsch, A., and McDonald, K. C.: Surface water inundation in the boreal-Arctic: potential
impacts on regional methane emissions, Environ. Res. Lett., 9, 075001, https://doi.org/10.1088/1748-9326/9/7/075001, 2014.
1095
Whalen, S. C., Reeburgh, W. S., and Barber, V. A.: Oxidation of methane in boreal forest soils: a comparison of seven
measures, Biogeochemistry, 16, 181–211, https://doi.org/10.1007/BF00002818, 1992.
Wickland, K. P., Jorgenson, M. T., Koch, J. C., Kanevskiy, M., and Striegl, R. G.: Carbon Dioxide and Methane Flux in a
Dynamic Arctic Tundra Landscape: Decadal-Scale Impacts of Ice Wedge Degradation and Stabilization, 47,
e2020GL089894, https://doi.org/10.1029/2020GL089894, 2020.
1100
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.

40
Wik, M., Varner, R. K., Anthony, K. W., MacIntyre, S., and Bastviken, D.: Climate-sensitive northern lakes and ponds are
critical components of methane release, 9, 99–105, https://doi.org/10.1038/ngeo2578, 2016.
Zhang, Z., Zimmermann, N. E., Stenke, A., Li, X., Hodson, E. L., Zhu, G., Huang, C., and Poulter, B.: Emerging role of
wetland methane emissions in driving 21st century climate change, PNAS, 114, 9647–9652,
https://doi.org/10.1073/pnas.1618765114, 2017.
1105
Zhu, Q., Peng, C., Chen, H., Fang, X., Liu, J., Jiang, H., Yang, Y., and Yang, G.: Estimating global natural wetland methane
emissions using process modelling: spatio-temporal patterns and contributions to atmospheric methane fluctuations, 24, 959–
972, https://doi.org/10.1111/geb.12307, 2015.
https://doi.org/10.5194/essd-2021-140
Open Access
Earth System
Science
Data
Discussions
Preprint. Discussion started: 7 May 2021
c
Author(s) 2021. CC BY 4.0 License.