Uranium isotopes and dissolved organic 
carbon in loess permafrost: Modeling the 
age of ancient ice
Authors: S.A. Ewing, J.B. Paces, J.A. O’Donnell, M.T. 
Jorgenson, M.Z. Kanevskiy, G.R. Aiken, Y. Shur, 
J.W. Harden, & R. Striegl
NOTICE: this is the author’s version of a work that was accepted for publication in Geochimica 
et Cosmochimica Acta. Changes resulting from the publishing process, such as peer review, 
editing, corrections, structural formatting, and other quality control mechanisms may not be 
reflected in this document. Changes may have been made to this work since it was submitted 
for publication. A definitive version was subsequently published in Geochimica et Cosmochimica 
Acta, Volume 152, March 2015, DOI# 10.1016/j.gca.2014.11.008
Ewing, S.A., J.B. Paces, J.A. O’Donnell, M.T. Jorgenson, M.Z. Kanevskiy, G.R. Aiken, Y. Shur, 
J.W. Harden, and R. Striegl. “Uranium Isotopes and Dissolved Organic Carbon in Loess 
Permafrost: Modeling the Age of Ancient Ice.” Geochimica et Cosmochimica Acta 152 (March 
2015): 143–165. doi:10.1016/j.gca.2014.11.008.
Made available through Montana State University’s ScholarWorks 
scholarworks.montana.edu 
Uranium isotopes and dissolved organic carbon in 
loess permafrost: Modeling the age of ancient ice
S.A. Ewing a,b,⇑, J.B. Paces c, J.A. O’Donnell a,d,e,1, M.T. Jorgenson f, M.Z. Kanevskiy 
d, G.R. Aiken a, Y. Shur d, J.W. Harden g, R. Striegl a
aU.S. Geological Survey, 3215 Marine St., Suite E-127, Boulder, CO 80303, U.S.
.S.
the same time, stratigraphic evidence indi-cates some past sediment redistribution and possibly shallow thaw among 
cores, with local mixing of aged thaw waters. Using measures of surface area and a leaching experiment to determine U 
distribution, a geometric model of (234U/238U) evolution suggests mean ages of up to 200 ky BP in the deepest core, 
with estimated uncertainties of up to an order of magnitude. Evidence of secondary coatings on loess grains with 
elevated (234U/238U) values and U concentrations suggests that refine-ment of the geometric model to account for 
weathering processes is needed to reduce uncertainty. We suggest that in this area of deep ice-rich loess permafrost, ice 
bodies have been preserved from the last glacial period (10–100 ky BP), despite subse-quent fluctuations in climate, fire 
disturbance and vegetation. Radiocarbon (14C) analysis of dissolved organic carbon (DOC) in thaw waters supports ages 
greater than 40 ky BP below 10 m. DOC concentrations in thaw waters increased with depth to maxima of >1000 ppm, 
despite little change in ice content or cryostructures. These relations suggest time-dependent production of old DOC that 
will be released upon permafrost thaw at a rate that is mediated by sediment transport, among other factors.bMontana State University, Dept. of Land Resources and Environmental Sciences, Bozeman, MT 59717, U
cU.S. Geological Survey, Box 25046, MS963, Denver Federal Center, Denver, CO 80225-0046, U.S.
dUniversity of Alaska, Fairbanks, AK, United States
eNational Park Service, Arctic Network, 240 W. 5th Ave., Anchorage, AK 99501, United States
fAlaska Ecoscience, 2332 Cordes Way, Fairbanks, AK 99709, United States
gU.S. Geological Survey, 345 Middlefield Rd., MS962, Menlo Park, CA 94025, United States
Abstract
The residence time of ice in permafrost is an indicator of past climate history, and of the resilience and vulnerability of 
high-latitude ecosystems to global change. Development of geochemical indicators of ground-ice residence times in 
perma-frost will advance understanding of the circumstances and evidence of permafrost formation, preservation, and 
thaw in response to climate warming and other disturbance. We used uranium isotopes to evaluate the residence time of 
segregated ground ice from ice-rich loess permafrost cores in central Alaska. Activity ratios of 234U vs. 238U ( 234U/238U) 
in water from thawed core sections ranged between 1.163 and 1.904 due to contact of ice and associated liquid water 
with mineral surfaces over time. Measured (234U/238U) values in ground ice showed an overall increase with depth in a 
series of five neighboring cores up to 21 m deep. This is consistent with increasing residence time of ice with depth as a 
result of accumulation of loess over time, as well as characteristic ice morphologies, high segregated ice content, and 
wedge ice, all of which support an inter-pretation of syngenetic permafrost formation associated with loess deposition. At 
(Schirrmeister et al., 2002; Froese et al., 2008a; Jensen
et al., 2008; Blinov et al., 2009). Given evidence for preserva-
for permafrost ice for four reasons. This variety of perma-
frost (1) should increase in age with depth (French andtion of Pleistocene ice over timescales of 105 y (Froese et al.,
2008; Reyes et al., 2010a; Vaks et al., 2013), a direct measure
of the age of ice in permafrost at timescales of 104–106 y is
needed.
Shur, 2010; O’Donnell et al., 2011a), (2) should produce
measureable signal due to small grain sizes and abundant
mineral surfaces (Fireman, 1986; DePaolo et al., 2006,
2012; Maher et al., 2006; Bourdon et al., 2009), (3) should1. INTRODUCTION
Over half the belowground terrestrial organic carbon 
(C) pool resides in permafrost that is warming and thawing 
at high latitudes in response to recent warming (Hugelius et 
al., 2014; Lachenbruch and Marshall, 1986; Osterkamp et 
al., 2009; Romanovsky et al., 2010; Tarnocai et al., 2009). 
The release and decomposition of the large amounts of 
organic carbon stored in permafrost soils may constitute a 
substantial positive feedback to the climate system (Zimov 
et al., 2006, 2009; Ping et al., 2008; Schuur et al., 2008; 
Koven et al., 2011), subject to the complex interaction 
between permafrost extent and factors such as plant com-
munity changes and decreasing sea ice extent (Euskirchen et 
al., 2006; Lawrence et al., 2008). It has been argued that 
much of the carbon in permafrost soils was sequestered 
during the Late Pleistocene, followed by extensive thaw of 
permafrost and release of carbon during the Pleistocene–
Holocene transition (Zimov et al., 2006, 2009). Under-
standing the relationship between known temperature fluc-
tuation and the vulnerability or resilience of permafrost 
over multiple glacial cycles is critical for understanding 
the global carbon balance and predicting its fate in the 
future (Jorgenson et al., 2010; Grosse et al., 2011). Develop-
ment of multiple chronometers is therefore essential for 
understanding the development and evolution of perma-
frost systems that may have existed over tens of thousands 
to millions of years.
Development of tools for dating permafrost has been lim-
ited by two factors. The first is the temporal constraints of
commonly used chronological tools: 14C has a half life of
5730 y and can be used to date materials up to about
50 ky BP (Reimer et al., 2009). Tritium can be used to
directly date ice, but its 12.32 y half life restricts its use to res-
idence times from one year to a few hundred years
(Morgenstern et al., 2010). Both 14C and tritium are compli-
cated – though not intractably – by residually elevated con-
centrations due to atmospheric testing of nuclear weapons
in the mid-twentieth century. The second limiting factor is
the potential for ice bodies to be younger than the surround-
ing sediment, due to thaw and drainage followed by re-accu-
mulation and freezing of meteoric water. This possibility
necessitates secondary geochronological evidence in support
of dates obtained through 14C analysis of solids, lumines-
cence analysis of loess, and tephrachronologyPermafrost formed in conjunction with loess deposition 
occurs throughout the Arctic and Subarctic, and represents 
a specific mode of “syngenetic” permafrost formation, also 
termed yedoma (French and Shur, 2010; Kanevskiy et al., 
2011). Gradual addition of eolian and organic material to 
frozen land surfaces – particularly in the context of produc-
tive Pleistocene steppe-tundra ecosystems – results in synge-
netic loess permafrost that stores large quantities of carbon 
(Zimov et al., 2006), slows subsequent thaw due to high ice 
content (Romanovsky et al., 2010; O’Donnell et al., 2011a), 
preserves distinct ice morphologies (French and Shur, 2010), 
and by definition must increase in age with depth. 
Moreover, the restricted particle-size distribution in eolian-
derived loess and the likelihood of equiaxial grains 
minimizes the effect of variable soil materials over time 
(Fleischer, 1983; Muhs et al., 2008). Hence this type of per-
mafrost offers a useful medium for testing indicators of per-
mafrost chronology based on rock–water contact, and 
associated carbon dynamics.
In this study, we evaluate radioactive disequilibrium
among uranium isotopes in thaw water (specifically,
234U/238U activity ratios, here denoted as (234U/238U) val-
ues) as a direct measure of the age of ice in syngenetic loess
permafrost. U-series disequilibrium occurs over timescales
of up to 106 y, and results from several natural processes,
either as a consequence of the distinct chemical behavior
of different elements in the decay chain, or as a result of
alpha-recoil effects from radioactive decay (Fig. 1)
(Bourdon et al., 2003). Recoil-driven disequilibrium occurs
as a function of rock–water contact at low temperature and
has been extensively observed in groundwaters (Osmond
et al., 1983; Paces et al., 2002; Porcelli and Swarzenski,
2003), as well as marine and riverine environments as a
function of weathering (Vigier et al., 2001, 2006; Chabaux
et al., 2003a,b, 2011; Robinson et al., 2004; Dosseto
et al., 2006a,b, 2008, 2010; Pogge von Strandmann et al.,
2011). Recoil effects have also been documented from min-
eral-ice contact (Fireman, 1986; Goldstein et al., 2004;
Aciego et al., 2011) and in weathering rinds and soils
(Pelt et al., 2008, 2013; Suresh et al., 2013). In general,
recoil disequilibrium effects occur as a function of time, U
concentrations in minerals and water or ice, and mineral
grain size and surface area. Ice-rich loess permafrost that
formed syngenetically with eolian loess deposition over
time is an ideal substrate for testing a U-series age model
in mi
d (c)
will o
et alpossess characteristic ice morphologies in support of its ori-
gin (French and Shur, 2010), and (4) should have a predict-
able mineral particle-size distribution and surface area as a
result of sorting during eolian transport (Wright, 1995,
2001; Smalley et al., 2005). If present, weathering rinds
may complicate interpretation of (234U/238U) values as a
metric of time by adding surface area as well as U adsorbed
from the surrounding fluid (Fig. 1) (Pelt et al., 2008; Suresh
et al., 2013, 2014). While uncertainties of model ages based
on (234U/238U) values in both sediment and ice can be sub-
stantial, their utility has been demonstrated (Dosseto et al.,
2010; Aciego et al., 2011; Handley et al., 2013). Ultimately,
the limited chronometric tools available for permafrost
make the investigation of this approach worthwhile.
The objective of this work was to determine the utility of
U-series isotopes for modeling ages of permafrost ice and
thaw waters. Development of this approach offers a means
of tracking permafrost thaw dynamics in the context of
Fig. 1. Schematic illustration of radioactive decay of 238U to 234U
234Th from (a) smooth grains, (b) grains with surface roughness, an
silt-sized grains in loess, recoil loss of 234Th (i.e., (234U/238U) < 1)
(30 nm) relative to the radius of the grain. Modified from DePaoloboth landscape history and the ongoing response of perma-
frost soils to global change. Recent work (Koch et al.,
2013a) has suggested that (234U/238U) values can provide
a signal of subsurface thaw in small catchments bounded
by ice-rich loess and hosting ice wedge networks. We
hypothesize that this signal will become increasingly evident
in larger river systems such as the Yukon and its tributaries
with loess dominated reaches as thaw increases in these pre-
viously resilient areas (Romanovsky et al., 2010; O’Donnell
et al., 2011a, 2014; Kraemer and Brabets, 2012). With
respect to landscape history, it has been argued that retreat
from the last glacial maximum was underway by 25 ka in
the Brooks Range (Briner and Kaufman, 2008; Barclay
et al., 2009). Large, inactive ice wedges in loess deposits
have long been thought to be of late Pleistocene age. Sub-
sequent warming and partial thaw of these features are indi-
cated by flat tops several meters below the present land
surface (Pewe, 1975). However, limitations of dating tech-
niques applied to permafrost makes it difficult to evaluate
these relationships. Preservation of Pleistocene permafrost
through periods of warmer conditions would havedepended on complex interactions among topography,
water, soil, vegetation and snow over time (Jorgenson
et al., 2010). Thus the current age and state of ground ice
in upland loess deposits is both an indicator of landscape
and ecosystem history, and a predictor of thaw susceptibil-
ity or resilience. Preservation of ground ice through major
fluctuations in external climate conditions could provide
important evidence of controls on boreal landscape evolu-
tion with climate, and has implications for landscape legacy
effects on future thaw dynamics and carbon-mediated feed-
backs with climate warming (Jorgenson et al., 2013a;
Kanevskiy et al., 2014).
2. MATERIALS AND METHODS
To develop this U-series approach to dating ground ice,
we (1) measured (234U/238U) values and U concentrations
in waters from thawed segments of five deep (7–21 m) cores
neral grains and surrounding fluid (ice), including the recoil loss of
grains with surface roughness modified by secondary coatings. For
nly be measurable at the grain rim due to the small recoil length
. (2006); Aciego et al. (2011); Handley et al. (2013).obtained in ice-rich loess permafrost, (2) measured
(234U/238U) values and U concentrations in a sequence of
leachates and residual solids following thaw of select sam-
ples within a single core, (3) quantified surface area of loess
material in a subset of samples from a single core using
direct measurement of surface area and inference from par-
ticle size distributions, (4) constrained near-surface ages
using 14C in thaw-water dissolved organic carbon (DOC)
from select samples, and (5) used our results to constrain
a geometric model of ice residence time based on
(234U/238U) values (DePaolo et al., 2006; Lee et al., 2010;
Aciego et al., 2011; Handley et al., 2013). Additionally,
DOC concentrations in thaw waters provide a metric of
carbon processing that is time dependent, and are expected
to increase with depth and duration of ice preservation.
2.1. Study site
This work focuses on samples of permafrost (Fig. 2)
formed in deep loess within the Hess Creek watershed near
Livengood, Alaska (65.568N, 148.925W, NAD83), approx-
imately 150 km north of Fairbanks, Alaska (Fig. 3). All
core locations are on forested north-facing slopes underlain
by permafrost, and are generally representative of black
spruce ecosystems in the discontinuous permafrost zone
of Alaska (Kane et al., 2005) and western Canada
(Harden et al., 1997).
Interior Alaska is characterized by a continental climate,
with temperature extremes ranging from 50 C to +35 C.
In the Hess Creek area, the average daily air temperature
ranges from 25 C in January to 15 C in July. Annual
precipitation averages 270 mm, 65% of which falls during
the summer growing season (mid-May to early September).
The cold snow period in interior Alaska is typically
>210 days long with maximum snow accumulation at these
sites (average = 44 cm; see O’Donnell et al. (2011b)) occur-
ring in late March (O’Donnell et al., 2011a).
In central Alaska, loess deposits vary widely in thick-
ness, from a few mm to >100 m (Pewe, 1975). In many
upland locations, the downslope thickening of loess depos-
its has been interpreted as indicating re-transportation after
eolian deposition, with accumulation on toeslopes and in
valley bottoms (Pewe, 1975) – a process that would facili-
tate development of ground ice in addition to its burial
and protection against thaw. Our study location captures
this potential variation by evaluating depth trends along a
hillslope using a sequence of cores that show increasing
loess thickness downslope (ranging from 7 to 21 m thick;
Fig. 4).
Most central Alaskan loess is glaciogenic, and formed
during middle to late Pleistocene time (Pewe, 1975; Jensen
et al., 2008, 2011; Muhs et al., 2008; Reyes et al., 2010b).
The glacial source of loess for the study area was likely in
the south-central Brooks Range to the north (Muhs and
Budahn, 2006), where multiple glacial advances occurred
during the Wisconsin period (12–110 ky BP) (Hamilton,
2001; Briner and Kaufman, 2008), and are associated with
loess deposition (Hamilton, 1982). Greater loess production
likely occurred in the region during full glacial climates;
however, accumulation may have been reduced under those
conditions due to changes in vegetation and surface pro-
cesses related to increased aridity and decreased tempera-
tures (Muhs et al., 2003). It is probable that loess
originally generated in the Brooks Range may have been
transported in the Yukon River and its tributaries prior
to deposition at the Hess Creek site (Froese et al., 2005).
Vegetation plays a fundamental role in the formation
and preservation of permafrost (Shur and Jorgenson,
2007). Vegetation at these sites is described by
O’Donnell et al. (2011a) according to the Alaska Vegeta-
tion Classification System (Viereck et al., 1992). The dom-
CN;
ka loFig. 2. Photo of core taken at an unburned site near Hess Creek (HC
Fig. 3. Location of study site in central Alaska. (a) Location on Alas
the Tanana and Yukon Rivers.O’Donnell et al., 2011a) showing syngenetic ice and redox banding.
ess map from Muhs and Budahn (2006); (b) location with respect to
ation
ugh
haw
ed nu
abuninant forest type on north-facing slopes of the Hess Creek
region is open black spruce [Picea mariana (Mill.) B.S.P.].
In mature black spruce stands, the forest understory is
composed of small woody shrubs, primarily Vaccinium
vitis-idaea and Ledum groenlandicum. Feather mosses
(Pleurozium schreberi and Hylocomium splendens), sphag-
Fig. 4. University of Alaska Fairbanks/Alaska University Transport
squares are referenced as cores A2 through A9 or HCDT.A2 thro
illustrates loess (silt) distribution, wedge ice extent, and interpreted t
of Transportation (DOT) boreholes are also indicated with underlin
Note 10 vertical scale exaggeration. Section 1 is distinguished bynum (Sphagnum fuscum), and reindeer lichens (Cladonia
stellaris and Cladonia arbuscula) dominate ground cover
in these mature stands. In the recently burned black
spruce stands, vegetation is dominated by standing dead
P. mariana, and living V. vitis-idaea, Vaccinium uligino-
sum, L. groenlandicum, and Equisetum spp. Burned
organic soil surfaces are colonized by Ceratodon purpureus
in the recently burned stands. In ice-rich loess permafrost
on upslope positions that might otherwise be more suscep-
tible to thaw (Swanson, 1996), limited drainage upon thaw
promotes permafrost resiliency and regrowth of organic
horizons, further protecting these systems against warm-
ing climate conditions and increasing fire frequency
(Shur and Jorgenson, 2007; O’Donnell et al., 2011a).
2.2. Sample collection
Porewaters were collected from soil pits excavated to the
depth of frozen material in summers of 2008 and 2009 near
coring locations. Solutions draining into pits were field fil-
tered into acid-rinsed HDPE bottles using a 0.45 lm cap-
sule filter (Geotech, Inc.) and a peristaltic pump. Samples
were transported on ice, acidified with 1.00 mL trace-metal
grade HNO3, and refrigerated prior to analysis.
Subsamples from eight core sequences were analyzed in
this study (Table 1). Five deeper core sequences represent a
subset of those collected as part of the geotechnical investi-gation conducted in May 2008 by the University of Alaska,
Fairbanks/Alaska University Transportation Center
(UAF/AUTC) and the Alaska Department of Transporta-
tion and Public Facilities (AKDOT), in an effort to improve
a section of the Dalton Highway near Livengood, Alaska
(Fig. 4; Shur et al., 2010; Kanevskiy et al., 2012). Hollow
Center (UAF/AUTC) core sequence (numbers 2 through 9 in green
HCDT.A9 in this paper) (Shur et al., 2010). The hillslope profile
zones based on reduced ice content. Additional Alaska Department
mbers; these boreholes were logged, but samples were not retained.
dant wedge ice compared to Section 2 (Kanevskiy et al., 2012).stem drilling was conducted by AKDOT. Eight boreholes
were cored by AKDOT and logged by AUTC scientists.
The boreholes were drilled into and sometimes through
the loess mantle overlying bedrock consisting of fractured
Precambrian schist. Depths range from 7.5 m at the most
upslope position, to 21.5 m at the most downslope position
(Fig. 4). All cores were collected in an area that burned in
2003. All deep core sections were kept frozen in a portable
freezer during collection and transportation back to the
University of Alaska, Fairbanks.
In addition, subsections of three shallower cores (2 m)
were collected in fall of 2007 at two adjacent convex ups-
lope sites, one from a site that burned in 2003, and two
from a “mature” site unburned in the last 100 y (Table 1).
These shallow cores were collected using a 7.5-cm-diameter
core barrel (SIPRE corer) equipped with a hand-held pow-
erhead, and are a subset of the burn chronosequence eval-
uated by O’Donnell et al. (2011a). A subsection of
uppermost wedge ice was included in one “mature” site
core. All subsections for U and DOC analysis were double
bagged and transported on ice to Fairbanks, where they
were stored frozen prior to shipment.
Subsections of the five deep cores (A2, A4, A5, A6, A9)
and the three shallow cores were shipped frozen on dry ice
to the USGS Denver Radiogenic Isotope Laboratory,
where they were stored frozen prior to subsequent
processing.
Table 1
Uranium concentrations, ice content, and (234U/238U) values in thaw waters.
Sample namea Sample mass, g Ice content (thaw water), g [U], ppb ±2r, ppb (234U/238U) ±2r
HCDT.A2.0885 280.03 103.93 10.480 0.043 1.224 0.007
HCDT.A2.0913 297.80 111.77 6.102 0.024 1.223 0.006
HCDT.A2.0931 272.36 111.93 4.255 0.015 1.214 0.005
HCDT.A2.0996 303.09 126.04 1.957 0.007 1.214 0.006
HCDT.A2.1016 212.05 108.60 1.661 0.007 1.184 0.009
HCDT.A4.0080 180.39 114.97 4.165 0.013 1.167 0.003
HCDT.A4.0116 146.69 89.41 7.430 0.024 1.170 0.004
HCDT.A4.0338 221.94 75.44 1.165 0.008 1.326 0.007
HCDT.A4.0445 193.01 78.86 3.400 0.019 1.264 0.011
HCDT.A4.0510 239.43 148.64 0.289 0.005 1.366 0.060
HCDT.A4.0645 224.68 120.87 0.477 0.003 1.508 0.009
HCDT.A4.0674 234.92 136.36 0.396 0.004 1.395 0.027
HCDT.A4.0750 247.10 144.54 3.143 0.013 1.360 0.007
HCDT.A4.0820b 315.97 106.98 13.616 0.073 1.555 0.005
HCDT.A4.0857 308.93 122.50 12.545 0.041 1.491 0.006
HCDT.A5.0101 256.58 115.88 20.684 0.081 1.195 0.008
HCDT.A5.0361 264.93 134.39 2.551 0.034 1.373 0.078
HCDT.A5.0384 279.02 120.86 5.279 0.016 1.271 0.004
HCDT.A5.0437 233.28 111.26 8.506 0.028 1.261 0.006
HCDT.A5.0511 280.05 126.02 13.569 0.044 1.251 0.006
HCDT.A5.0545 164.34 99.70 5.872 0.019 1.248 0.005
HCDT.A5.0585 217.96 136.88 2.758 0.009 1.260 0.005
HCDT.A5.0627 259.30 165.88 5.424 0.020 1.273 0.007
HCDT.A5.0660 256.31 128.87 6.779 0.021 1.251 0.005
HCDT.A6.0151 304.63 133.98 1.981 0.008 1.173 0.005
HCDT.A6.1322 254.65 121.89 1.258 0.007 1.495 0.021
HCDT.A6.1380 264.74 120.78 2.040 0.008 1.502 0.011
HCDT.A6.1500 250.61 121.84 1.658 0.006 1.497 0.011
HCDT.A6.1560 252.29 102.14 17.915 0.057 1.379 0.004
HCDT.A6.1594 304.43 119.07 17.720 0.055 1.551 0.005
HCDT.A6.1762 218.32 153.48 6.765 0.022 1.271 0.005
HCDT.A9.0035 154.51 71.56 0.303 0.001 1.163 0.012
HCDT.A9.0055 205.34 111.23 2.105 0.008 1.168 0.014
HCDT.A9.0102 192.28 108.11 11.533 0.037 1.182 0.006
HCDT.A9.1102 188.63 104.25 0.584 0.003 1.322 0.037
HCDT.A9.1130 176.99 121.37 3.092 0.010 1.297 0.006
HCDT.A9.1237 185.63 79.56 46.453 0.149 1.386 0.010
HCDT.A9.1315b 196.59 101.16 1.532 0.008 1.468 0.014
HCDT.A9.1399 294.11 97.01 34.294 0.105 1.366 0.004
HCDT.A9.1695 226.43 100.31 12.351 0.040 1.366 0.004
HCDT.A9.1715 258.16 104.25 10.115 0.031 1.716 0.006
HCDT.A9.1770 275.16 134.38 3.973 0.013 1.703 0.005
HCDT.A9.1849 306.28 122.65 7.097 0.023 1.529 0.007
HCDT.A9.1879 178.78 32.84 4.477 0.015 1.594 0.008
HCDT.A9.1924 237.93 128.51 1.867 0.007 1.772 0.011
HCDT.A9.1956 283.01 109.07 4.532 0.015 1.904 0.014
HCDT.A9.1977 218.94 132.12 1.790 0.006 1.768 0.007
HCDT.A9.2005 166.39 96.03 1.019 0.004 1.655 0.011
HCDT.A9.2050 197.35 121.39 1.003 0.004 1.638 0.016
HCDT.A9.2069 206.79 124.10 1.482 0.007 1.589 0.017
HCDT.A9.2114 317.79 130.63 3.559 0.011 1.541 0.018
HCCN.1.053b 382.60 190.76 3.525 0.018 1.163 0.006
HCCN.1.119 379.22 115.02 17.187 0.054 1.162 0.003
HCCN.1.150 386.88 30.63 41.964 0.133 1.166 0.004
HCCN.2.127 280.44 68.63 31.729 0.097 1.165 0.003
HCCN.2.165 323.43 66.01 35.706 0.108 1.169 0.003
HCCN.2.185 313.40 32.88 24.792 0.078 1.188 0.003
HCCN.2.205 294.51 196.68 0.286 0.001 1.180 0.006
HCCN.2.255 327.90 98.49 0.094 0.0004 1.210 0.008
HC03.5.060 345.01 144.00 5.466 0.020 1.196 0.006
HC03.5.112 386.40 26.65 12.968 0.042 1.203 0.004
2.3. Core description and interpretation
Detailed description of cryostructures in the cores fol-
lowed methods of Shur et al. (2010) and Kanevskiy et al.
(2011), and are given in Kanevskiy et al. (2012). This eval-
uation concluded that the permafrost was primarily formed
syngenetically with loess deposition during Pleistocene time
(Shur et al., 2010; Kanevskiy et al., 2012). The upslope
cores were generally more ice-poor, while the down-slope
cores were more ice-rich, with abundant wedge ice (Figs. 4
and 5). This suggests better preservation of ground ice in
the downslope positions.
Stratigraphic descriptions and ice contents of the deep
cores revealed two major stratigraphic units related to tim-
ing of ground-ice development (Kanevskiy et al., 2012). An
ice-poor unit of variable thickness near the surface provides
strongly suggesting preservation following upward propa-
gation of wedges with loess accumulation. Cores A6, A7
and A9 are in the downslope “ice-rich” zone of increasing
loess depth and ice-wedge frequency downslope (Figs. 4
and 5).
Permafrost genesis within the core sequence is inter-
preted to be both syngenetic and quasi-syngenetic (French
and Shur, 2010; Shur et al., 2010). Syngenetic permafrost
forms in response to simultaneous sedimentation that
causes the top of the permafrost table to propagate
upwards, with volumetric expansion (200% and more) due
to ice aggradation, as well as high organic matter content,
characteristic cryostructures, and frequent ice wedges. At
shallower depths above or in the absence of ice wedge
sequences (e.g., core A4), ice-rich zones may indicate for-
mation of “quasi-syngenetic” permafrost by virtue of
HC03.5.150 370.30 29.04 7.785 0.033 1.226 0.013
HC03.5.190 603.47 73.59 2.169 0.015 1.273 0.027
a Last four digits of HCDT sample numbers and last three digits of HCCN and HC03 sample numbers indicate depth in cm. Sites HCCN.1
and HCCN.2 are within one kilometer of HCDT sites (deep cores); both HC03.5 and HCDT burned in 2003.
b Analyzed in duplicate for (234U/238U); uncertainties reflect propagation for two analyses.evidence of partial thaw and re-freezing (unit labeled t in
Figs. 4 and 5). Core A2 is mainly ice-poor (gravimetric
water content 40%) at 2–9 m depth, core A4 is ice-poor
(40–50%) at 1.5–3.0 m depth, and core A6 is ice-poor
(40%) at 1–3 m depths. Cryostructures within this unit
are mainly nonvisible to visible pore ice. Below this zone,
the cores typically have higher ice contents including mm-
scale cryostructures ranging from micro-lenticular to
micro-ataxitic, as well as cm-scale “belts”. Ice contents in
this ice-rich zone are higher, with values typical for synge-
netic permafrost (50–200% gravimetric moisture content).
Ice wedges were frequently encountered in the cores (up
to 47% of the total volume) and persisted with depth,Fig. 5. Interpolated subsurface ice wedge and thaw zone distribution bas
Zones of lesser thaw (less reduction in ice content) also suggested by flat
compared to Fig. 4.post-fire moss re-growth or downslope sedimentation pro-
cesses that provide the thermal insulation needed for
upward re-propagation of an ice-rich permafrost table
(Shur and Jorgenson, 2007).
We hypothesize that the layer of frozen material overly-
ing flat-topped ice wedges along the hillslope reflects the
depth affected by more frequent thaw, and some degree of
downslope sediment transport and water movement. This
process could result in age inversion of the solid-phase car-
bon and possible homogenization of the U series signal
with depth due to an overall mixing caused by thawing
and refreezing of aqueous fractions. However, this effect
notably was not observed in 14C profiles of neighboringed on AUTC/DOT cores (Shur et al., 2010; Kanevskiy et al., 2012).
tops on ice wedge sequences. Note that the area shown is expanded
shallow cores (O’Donnell et al., 2011a). The upslope ice-
poor/thaw zones, downslope thickening of accumulated
loess and downslope increase in ice-wedge frequency could
reflect the combined processes of downslope sediment redis-
tribution, upslope thaw-water loss, and downslope water
accumulation and ice aggradation. Sediment transport is
further indicated by interlayering of gravels in downslope
cores A7 and A8, and burial of ice-rich intermediate layer
in core A7. The absence of an ice-rich intermediate layer
at the top of the permafrost table in cores A6 and A7
(whereas an ice-rich intermediate layer is present in cores
A3, A4 and A5, as well as A8 and A9) indicates a limited
timeframe of slope stability at the midslope positions. How-
ever, slope angles of 2-8 degrees suggest limited potential
for downslope movement following previous redistribution
(Wu, 1984; Carey and Woo, 2002), although ongoing soli-
fluction may occur (Harris et al., 2008).
2.4. Thaw water separation and analysis
Selected subsections of frozen UAF/AUTC cores 2, 4, 5,
6 and 9 (A2, A4, A5, A6 and A9; Fig. 4; Table 1) were
placed in pre-weighed polypropylene jars, and weighed to
ensure accurate evaluation of total mass. Core sections were
thawed in the jars at room temperature and the resulting
slurry was transferred in aliquots to 50-mL centrifuge
tubes, which were then centrifuged at 10,000 rpm for
10 min. The supernatant was transferred by pipet into a
separate tube, and filtered through a 0.45 lm glass fiber fil-
ter. The solids were oven-dried overnight at 105 C, except
for the portions used in the leaching experiment.
2.5. Solid phase analysis
Evolution of (234U/238U) disequilibrium over time
depends on (1) U distribution in the solid and surrounding
water/ice, and (2) total particle surface area subject to recoil
loss. Following initial analysis that revealed increasing
(234U/238U) values with depth, we undertook quantification
of U distribution in a subset of samples from core A4, using
simple leaching steps based on prior work (Maher et al.,
2006; Lee et al., 2010). We also evaluated total particle size
distribution and Brunauer–Emmett–Teller (BET) surface
area in a subset of samples from core A9 as a test of their
influence on (234U/238U) values (DePaolo et al., 2006;
Maher et al., 2006).
2.5.1. Leaching experiment
In order to evaluate the total distribution of U in
recoil-affected samples showing a range of (234U/238U) val-
ues, we conducted leaching experiments using a subset of
samples from 1 to 9 m depth in core A4. This core was
selected early in the study because (234U/238U) values in
its thaw waters increased dramatically over a relatively
short depth interval (8 m). The leach sequence was
intended to provide a simple evaluation of the amount
and isotopic composition of U retained in secondary coat-
ings on particles (e.g., carbonate, iron oxide, organics), as
well as the amount and isotopic composition of U in the
residual solid. By measuring (234U/238U) values in allleachates and residues in these samples, we attempted to
limit the uncertainties inherent in procedurally defined
leaching steps, although in future studies, buffered solu-
tions and particle sizing would be preferred (Lee et al.,
2010; Handley et al., 2013; Suresh et al., 2014). Our aim
was simply to capture U that might be present in highly
exchangeable sites, sorbed tightly onto mineral surfaces,
or present in organic complexes.
About 2 g of wet solids from each of six samples (core
A4) were subjected to a series of leaches including: (1) ultra-
pure water, (2) trace metal grade 1 N HCl, and (3) a second
round of 1 N HCl following ashing for 1 h at 550 C (to
combust organic carbon). This procedure was selected for
simplicity based on a trial run in which substantial U
(200 ng g1 solid) was released with 1 N HCl before and
after combustion. Leaching solutions were shaken with sol-
ids in 50-mL centrifuge tubes for 30 min, then centrifuged
at 10,000 rpm for 10 min. The supernatant was separated
and one aliquot was dried for U purification and isotope
analysis. The remaining aliquot was acidified using Optima
grade HNO3 and analyzed for trace-element concentrations
by quadrupole ICP-MS. The solids were digested in a mix-
ture of hot (200 C), concentrated HF and HNO3 and
analyzed for U isotopes.
2.5.2. Particle size distribution and surface area
In order to test the effect of surface area on thaw-water
(234U/238U) values with depth, we measured particle-size
distributions and specific surface area in a subset of sam-
ples from depths of 1 to 21 m in core A9. This core was
selected because it contained the greatest range and oscil-
lation in thaw-water (234U/238U) values. In addition to
constraining the surface areas used to model ages, we
wanted to explore the degree to which variations in the
overall (234U/238U)-depth trend might be explained by
variation in particle size distribution and hence surface
area.
Five loess samples from core A9 were evaluated for spe-
cific surface area (SSA) by whole-sample BET analysis using
a Micrometricse TriStar II 3020 Automatic Physisorption
Analyzer at the University of Minnesota. After initial SSA
analysis of untreated soil, organic matter was removed from
the same samples by heating in a muffle furnace at 350 C for
12 h (Wagai et al., 2009). Following heating, samples were
then subject to the same SSA analysis. Particle size distribu-
tion (PSD) in the same samples was determined using aMal-
vern particle-size analyzer at the USGS Geosciences and
Environmental Change Science Center soils lab in Denver,
with and without treatment for carbonate using acetic acid,
and organic matter using hydrogen peroxide. The fractional
volume of particles in each of fifty size bins ranging from 0.06
to 2000 lm diameter was determined. PSD and BET results
were used to calculate recoil loss fractions (fa) as further
described below.
2.6. U concentrations and isotopic analysis
Leach solutions, porewaters and dissolved residues were
initially analyzed for U concentration by quadrupole ICP-
MS. Precise U concentrations and isotopic compositions
were then determined by isotope dilution and thermal-ion-
ization mass spectrometry. Aliquots containing 100 ng U
were weighed into 15-mL Teflon vials, acidified with 20
drops ultrapure concentrated HNO3, spiked with a highly
purified isotope tracer solution containing about 2 ng
236U, and equilibrated overnight at sub-boiling tempera-
tures. Spike addition allowed for very precise quantification
of U concentrations, and evaluation of yields following the
purification procedure. Solutions were equilibrated, dried,
and re-dissolved in 0.5 mL 7 N HNO3.
For the core-sample thaw waters, 5–20 mL of thaw
water were weighed into a pre-cleaned quartz crucible, acid-
ified with 20 drops ultrapure, concentrated HNO3, and
spiked with about 2 ng of 236U tracer solution. The solution
was dried on a hotplate in a clean hood and then heated in a
muffle furnace at 550 C for 1 h to remove organic com-
pounds. The remaining solid was dissolved in 7 N HNO3
(trace metal grade), transferred to a 15-mL Teflon vial
(Savillex Corporatione) and dried. The portion not soluble
in 7 N HNO3 was dissolved in 10–30 drops of concentrated
HF, dried, redissolved in 7 N HNO3, and combined with
the previously dissolved portion of the sample for a final
volume of 1 mL in 7 N HNO3.
Following dissolution in 7 N HNO3, uranium salts were
purified using standard ion exchange chemistry with AG 1-
X8 resin (200–400 mesh) and 7 N HNO3 followed by 6.5 N
HCl acid prior to elution of U using 0.05 N HNO3. Purified
U salts were evaporated onto one side of a double-Re fila-
ment assembly and analyzed using aThermoFinniganTriton
mass spectrometer equipped with a secondary electron mul-
tiplier (SEM) and retarding potential quadrapole (RPQ)
electrostatic filter allowing abundance sensitivity measured
at mass 237 of approximately 20 parts per billion. Measure-
ments were made using the single SEM in dynamic peak-
jumping mode on 233U, 234U, 235U, and 236U. Atomic ratios
of 234U/238U were derived using the measured 234U/235U
ratio and an assumed 238U/235U ratio of 137.88 (Steiger
and Jager, 1977) and converted to activity ratios
((234U/238U) values) using published values for decay con-
stants for 234U (2.8262  106 y1)(Cheng et al., 2000) and
238U (1.55125  1010 yr1) (Jaffey et al., 1971). Allmeasure-
ments were corrected for spike contributions, mass fraction-
ation, and blank contribution, and were normalized to the
accepted value of 5.29 ± 0.04  105 (±2r) for 234U/238U
in NIST SRM 4321B, which was included with each barrel
of unknown samples. Two hundred sixty nine analyses of
the SRM 4321B standard determined over a period from
2007 to 2012 yielded an average 234U/238U atomic ratio of
5.2891  105 (2  standard deviation = 0.0098  105,
and 2  standard error of the mean = 0.0006  105). Sev-
enty one analyses of an in-house standard assumed to be in
secular equilibrium (solution of 69.3-Ma U ore from the
Schwartzwalder mine) (Ludwig et al., 1985) analyzed over
same period yielded an average (234U/238U) of 0.9982 with
2  standard deviation of 0.0024 and a 2  standard error
of the mean of 0.0003. Full procedural blanks typically con-
tained between 5 and 50 pg U and constituted a small frac-
tion of total U from the sample, which typically ranged
between 11 and 1740 ng U.2.7. DOC concentration and isotopic analysis
To test the agemodel for permafrost based on (234U/238U)
values, we evaluated 14C ages of DOC in six samples from
three cores. DOC concentrations in all filtered thaw waters
were determined using an OI Analytical Model 700e total
organic carbon analyzer via a platinum-catalyzed persulfate
wet oxidation method (Aiken, 1992). In the samples selected
for 14C analysis, DOC concentrations were also determined
by UV-oxidation. UV-oxidation was performed using a
UV system that was directly connected to vacuum lines
where the resulting CO2 was cryogenically purified for
14C
and d13C analysis using methods described in detail else-
where (Raymond and Bauer, 2001; Raymond et al., 2004).
Briefly, a thaw water subsample containing 0.1 mg DOC
was transferred to a clean quartz tube. The samples were
acidified with 0.2 mL of ultra-high-purity 40% phosphoric
acid. The samples were then sparged with ultra-high-purity
(UHP) N2 to remove any inorganic carbon. UHP O2 was
then sparged through the system to provide an oxidant for
the UV oxidation of DOC. The sample was then oxidized
with UV. The resulting CO2 was transferred to the vacuum
line and cryogenically purified. The purified CO2 was graph-
itized and analyzed for carbon isotopes at the National
Ocean Sciences Accelerator Mass Spectrometer at the
Woods Hole Oceanographic Institution. Full procedural
blanks were below the detection limit for DOC
(0.05 ppm). DOC concentrations determined by the two
methods differed by 1–10%.
2.8. U series age model
To estimate the age of ground ice in each sample, we
combined measured (234U/238U) values, U concentrations,
and particle-size distributions with a geometric model of
(234U/238U) evolution in ice and solids with time
(Fireman, 1986; DePaolo et al., 2006; Maher et al., 2006;
Lee et al., 2010; Aciego et al., 2011; Handley et al., 2013).
Geologic materials older than about 1 Ma that have not
interacted with migrating solutions will have (234U/238U)
values of 1.000, a state defined as secular equilibrium (see
e.g., Bourdon et al., 2003). If the solid material is perme-
able, fractured, or reduced to smaller particles, the resulting
surface area is subject to radioactive disequilibrium as a
result of recoil processes associated with alpha decay of
238U. As a result, 234U is enriched in the solution because
of either direct ejection of daughter isotopes from the min-
eral completely, or preferential leaching of 234U from radi-
ation-damaged crystal lattice sites (Gascoyne, 1992;
Andersen et al., 2009). In either case, evolution of the dis-
equilibrium signal occurs as a function of particle surface
area, fluid residence time, and the relative concentration
of U in the fluid versus the mineral solid (Osmond and
Cowert, 2000; Porcelli and Swarzenski, 2003). Central Alas-
kan loess is primarily (>65%) composed of silt-sized grains
of quartz, plagioclase, mica and chlorite (Muhs et al., 2003;
Muhs and Budahn, 2006), and is suited for U series model-
ing due to its relatively small sized mineral grains with high
surface area.
For this work, a geometric model provided a relatively
simple approach to modeling U series data as an indicator
of ice-loess interaction time (DePaolo et al., 2006; Maher
et al., 2006; Bourdon et al., 2009; Dosseto et al., 2010;
Lee et al., 2010; Aciego et al., 2011; Handley et al., 2013).
In that model, the value fa is the fraction of
238U decays
that should result in the ejection of the immediate daughter
234Th atom from the grain due to the proximity of the decay
site to the mineral surface (i.e., within the a-recoil distance
of 30 nm for most silicates) (Maher et al., 2006; Lee et al.,
2010; Handley et al., 2013). The ejected 234Th atom then
undergoes rapid beta decay (half life = 24.1 days) to 234U.
The magnitude of fa is a function of recoil length (L) and
surface area (S). For samples with surface areas measured
by BET, fa values were calculated directly using (Kigoshi,
1971; DePaolo et al., 2006; Maher et al., 2006; Lee et al.,
2010; Handley et al., 2013):
f a ¼
1
4
LSqs ð1Þ
where S is measured surface area (m2 g1) and qS is the par-
ticle density (estimated as 2.69 g m3). It has been argued
that small-scale roughness included in direct measurement
of surface area (that is, at the scale of gas molecules used
to make the measurement) require use of a fractal dimen-
sion to calculate fa values (Bourdon et al., 2009; Lee
et al., 2010; Handley et al., 2013); the resulting fa values
are typically an order of magnitude smaller and the model
ages consequently greater. Here, we use Eq. (1) as a conser-
vatively high estimate of fa for calculation of minimum ice
age.
For samples without surface area measurements, values
of fa for different particle size categories were calculated
assuming spherical particles of radius r (m) (Kigoshi, 1971):
f a;r ¼
3
4
L
r
 L
3
12r3
 
ð2Þ
where r was assigned as the modal value for each of the 50
measured particle-size bins. These fa,r values were multi-
plied by volume fractions in each size bin (Xr), as well as
a factor representing surface roughness (kr) and an aspect
ratio (br), and summed to obtain an initial value fa,S for
each sample (DePaolo et al., 2006; Maher et al., 2006;
Dosseto et al., 2010; Lee et al., 2010):
f a;s ¼
Xr¼max
r¼min
f a;rX rbrkr ð3Þ
Based on previous work, a value of brkr = 10 was used to
evaluate the combined effects of aspect ratio and surface
area (DePaolo et al., 2006, 2012; Lee et al., 2010;
Handley et al., 2013). Four of the 51 size categories consist-
ing of particles <0.54 lm in diameter were assigned pro-
gressively lower values of brkr (8-2) so that their fa,r
values would not exceed 1.0.
Using this geometric approach, the activity ratio of the
solid for a given age (As,t) is (DePaolo et al., 2006, 2012;
Maher et al., 2006; Lee et al., 2010):
As;t ¼ ð1 f aÞ þ ½As;i  ð1 f aÞek234t ð4Þwhere As,i is the initial activity ratio of the sediment or loess
(which approaches an equilibrium value of (1  fa)), k234 is
the decay constant of 234U (y1) and t is time (y), or the
“comminution age” of the solid, meaning the total time
since the sediment was generated, regardless of transport
history (DePaolo et al., 2006, 2012; Lee et al., 2010). If
As,i is considered as the activity of the loess upon deposi-
tion, this equation describes the measured activity of the
loess as a function of time since deposition. However the
activity of loess particles upon deposition is not usually
assessable. Eq. (4) can be rearranged to solve for t as a func-
tion of As,i and the solid at time t (As,t):
tsolid ¼ 
1
k234
 
ln
As;t  ð1 f aÞ
As;i  ð1 f aÞ
 
ð5Þ
where (As,i) > (As,t) > (1  fa). This age equation relies only
on evolution of the mineral phase and does not assume
closed-system evolution with the surrounding fluid (ice).
For glaciogenic loess accumulation in ice-rich syngenetic
permafrost, the activity ratio of the original loess (As,i)
may have been close to unity (DePaolo et al., 2012), and
transport times arguably were short before particles were
encapsulated in ice rich permafrost. Here, because our
focus is the time the loess has spent in permafrost, we con-
sider values of As,i as the activity of the loess upon deposi-
tion, following transport times from the Brooks Range that
are unknown, but were likely brief (Muhs and Budahn,
2006).
Alternately an age equation can be derived from the geo-
metric model that describes time evolution of the disequilib-
rium signal in the surrounding ice/fluid (Fireman, 1986;
Aciego et al., 2011), based only on activity ratios in ice
and initial water, the recoil fraction, and the distribution
of U between solids and thaw water. This age equation
for ice assumes a closed system and can be expressed in
terms of activity ratios as:
tice ¼ 
1
k234
 
ln
Af ;t  f aRs  1
Af ;i  f aRs  1
 
ð6Þ
where Rs is the ratio of (
238U) in the solid to (238U) in the ice
(Xs/Xf) and Xs = 1  Xf.
Application of Eq. (6) requires knowledge of the distri-
bution of U in the solid (Xs) compared to the fluid (Xf).
With analysis of both thaw water and solids in these cores,
an assessment of assumptions about starting values in the
mineral phase can be undertaken. Initially the mass-
weighted total would reflect the activity ratio of the starting
loess (As,i) and the porewater from which the ice developed
(Af,i):
Asþf ;i ¼ As;iX s þ Af ;iX f ð7aÞ
where Xs is the fraction of total U in the solid and Xf is the
fraction of total U in the ice (which may contain a liquid
component). Under closed system conditions (no removal
of ice or water with thaw), the evolution of the mass-
weighted total (234U/238U) in the permafrost core based
on activity in the mineral solid (As,t) and fluid (ice and
water, Af,t) would occur congruently, summing to the value
of As+f,t:
Asþf ;t ¼ As;tX s þ Af ;tX f ð7bÞ
Here a key assumption requiring exploration in future
research is that the U distribution is constant over the time
of freezing, and equal to values derived in the leaching
experiment. In addition, we assume that the permafrost
developed from active layer porewater with a (234U/238U)
value of 1.161, estimated from the average of three analyses
of active-layer porewater collected nearest the core site
(Erickson Creek samples in Table 2). Assuming a closed
system, the age equation for the combined liquid and solid
would be:
A ¼ A ek234t ð8aÞ
samples ranged from 1.162 to 1.224, consistently higher
than the most proximal porewater (1.161, Table 2) as com-
pared to porewater near shallow core HC03.4 (Tables 1 and
2). The highest thaw-water (234U/238U) value (1.904)
occurred in the lowermost section of the deepest core
(A9). Within this overall trend, oscillation of thaw-water
(234U/238U) with depth was evident in most cores, with
the largest oscillation (from 1.904 to 1.541) at 20 m depth
in core A9. As with core A2, samples from A5 showed little
variation with depth over the limited interval evaluated
(1.01–6.60 m). These limited variations in (234U/238U) are
consistent with the hypothesized thaw history of these cores
based on ice morphologies (Shur et al., 2010; Kanevskiy
3a
3
1
2
ble 1sþf ;t sþf ;i
tclosed ¼ 
1
k234
 
ln
Asþf ;t
Asþf ;i
 
ð8bÞ
Based on results from our leaching experiment, we assume
that leachable fractions with values of (234U/238U) > 1 rep-
resent components of the “fluid” fraction with a value iden-
tical to that observed in the thaw water. This assumption is
not tested here but is supported by work demonstrating
that weathering rinds reflect the composition of weathering
fluids and concentrate U (Pelt et al., 2008; Andersen et al.,
2009). To our knowledge, the character of weathering rinds
in permafrost and the effect of weathering rinds on closed
system evolution of (234U/238U) values during mineral-ice
contact has not been studied. Notably, the liquid water con-
tent of permafrost is a function of mineral surface area and
solute content, both of which are expected to be high in
yedoma (Tyutyunov, 1956; Romanovsky and Osterkamp,
2000). Thus it is possible that weathering occurs at the pore
and particle scale without permafrost thaw. Field observa-
tion of redox coloration around ice lenses in frozen core is
evidence supportive of these processes (Fig. 2).
3. RESULTS AND DISCUSSION
3.1. Thaw water (234U/238U) values
In the deep cores, the primary variation observed in
(234U/238U) values of thaw water was the overall increase
with depth in all cores except A2 (Table 1, Fig. 6a), in which
no near-surface (upper 1–2 m) sample was evaluated due to
low ice content. This observation supports the general pre-
diction that thaw-water (234U/238U) values should increase
with depth in well-preserved syngenetic permafrost devel-
oped with loess deposition. Values in near-surface core
Table 2
Soil porewater samples, Hess Creek area, 2008–2009.
Date Tributary Sample ID
Sep-08 Richardson HC03.5.TK
Sep-09 Richardson HC03.5.TK
Sep-08 Erickson HC03.1b
Sep-08 Erickson HC03.1.TK
Sep-08 Erickson HC03.1.TK
Richardson Creek average (n = 2)
Erickson Creek average (n = 3)2
a HC03.TK3 is 500 m downslope of core location HC03.5 in Ta
b Closest to deep coring location.et al., 2012), although the more elevated values in core
A5 are consistent with the overall depth trend observed in
other cores (Fig. 6a).
In upslope cores A2 and A5, the lack of change in thaw-
water (234U/238U) values with depth is consistent with the
interpretation of Shur et al. (2010), who suggested that
the shallow (1–4 m) zones of reduced ice content in cores
A2 through A5 are the result of past thaw and removal
of the thaw water. However, if true, this thaw may have
occurred long enough ago to allow for re-development of
elevated thaw-water (234U/238U) values in core A5 (1.25–
1.37, Table 1); alternatively the values in core A5 may
reflect homogenization resulting from mixing of higher
and lower values. In core A4 the increase with depth is sim-
ilar to that observed in the deeper cores downslope. This
supports the mixing scenario for core A5, which may reflect
localized processes associated with cryoturbation, down-
slope sediment redistribution from the top of core A4,
and/or small scale movement of thaw waters due to ele-
vated pore pressures (Wu, 1984; Carey and Woo, 2002;
Ping et al., 2008). Overall, however, the increasing thaw-
water (234U/238U) values with increasing depth in this set
of cores is consistent with the concept that that the synge-
netic permafrost in place today has behaved predominantly
as a closed system and has largely preserved the expected
signal of increasing age with depth.
In the shallow cores (Table 1, Fig. 6b), which were sam-
pled at shorter depth intervals of 0.5–3 m, (234U/238U) val-
ues showed small increases with depth (1.162–1.273) that
were consistent with values in the upper sections of the dee-
per cores (Fig. 6). At depths shallower than 150 cm, uni-
form (234U/238U) values could indicate removal of thaw
water with periodic active layer deepening (current active
layer depths are 50 to 80 cm; fire-induced deepening of
Mass, g [U] (ppb) ±2r (ppb) (234U/238U) ±2r
435.22 0.1138 0.0004 1.184 0.009
817.78 0.2363 0.0008 1.199 0.005
438.70 3.9733 0.0057 1.157 0.003
422.83 0.0297 0.0002 1.161 0.024
433.04 0.1136 0.0011 1.166 0.023
1.191
1.161
.
34U/2
1.0 1.2 1.4
1.0 1.2 1.4
0
500
1000
1500
2000
2500
(2
de
pt
h,
 c
m20 cm was observed by O’Donnell et al. (2011a)), and/or
mixing of active layer water and sediment with downslope
movement or cryoturbation (Carey and Woo, 2002;
French and Shur, 2010). O’Donnell et al. (2011a) observed
depth trends of 14C in these soils indicating increasing ages
of solid-phase organic carbon on the order of 10 ky in the
upper 100 cm. The consistent thaw-water (234U/238U)
values are equal to or greater than those observed in pore-
water at these shallowest depths, and imply mixing of active
layer water (and perhaps sediment) with seasonal and
longer thaw cycles during the Holocene (Jorgenson et al.,
2013b).
3.2. Age model application
3.2.1. Leaching experiment: Uranium distribution and
isotopic mass balance
Fractional U abundances and isotopic compositions
(solid vs. fluid) were evaluated using the results of the
1.0 1.2 1.4
1.0 1.2 1.4
0
100
200
300
400
500
(234U/238U)tw
de
pt
h,
 c
m
(b)
Fig. 6. Measured (234U/238U) values in thaw waters from (a) five deep cor
burned and unburned sites (O’Donnell et al., 2011a), with deep core valu1.6 1.8 2.0
1.6 1.8 2.0
38U)tw
(a)sequential leach procedure for six samples from core A4
(Table 3, Fig. 7). In all samples, (234U/238U) values were
greater than 1.0 in thaw water and the first HCl leach step
(Fig. 7a), which yielded considerably more U (12–20% of
the total, Table 3) than was present in thaw waters for
these samples (0.03–0.47%; Table 4; Fig. 7b). These data
could indicate that U is concentrated in readily dissolved
secondary precipitates or loosely sorbed onto particle
surfaces due to the liquid water component in permafrost
(Osmond and Cowart, 2000), which is consistent with
observed U series isotope trends in studies of weathering
rinds and soils (Pelt et al., 2008; Suresh et al., 2014). Sec-
ondary precipitates might include carbonates, which are
known to accumulate U and would be released with
application of acid. Consistent with this, inorganic carbon
concentrations of 1% were observed in solid phase anal-
ysis (data not shown). To our knowledge, the accumula-
tion of secondary precipitates in ice-rich loess permafrost
has not been investigated, and merits further study.
A2 deep core
A4 deep core
A5 deep core
A6 deep core
A9 deep core
HCCN.1 (mature)
HCCN.2 (mature)
HC03.5 (2003 burn)
es in the sequence shown in Fig. 2, and (b) nearby shallow cores on
es for comparison. Grey box indicates ice wedge samples.
T
ab
le
3
R
es
ul
ts
of
le
ac
hi
ng
ex
pe
ri
m
en
t
(s
ix
sa
m
pl
es
fr
om
co
re
A
4)
.
U
ra
ni
um
ab
un
da
nc
e
(n
g
U
),
ac
ti
vi
ti
es
(A
),
an
d
di
st
ri
bu
ti
on
(R
s,
E
q.
(9
b)
in
th
aw
w
at
er
s
(t
w
),
le
ac
ha
te
s
an
d
so
lid
s
(s
)
at
th
e
ti
m
e
of
sa
m
pl
in
g
(t
).
D
ep
th
,
cm
T
ha
w
w
at
er
ng
U
(±
2r
)
1s
tl
ea
ch
ng
U
(±
2r
)
2n
d
le
ac
h
ng
U
(±
2r
)
R
es
id
ue
ng
U
(±
2r
)
Ic
e
A
f,
t
(2
34
U
/2
38
U
)
(±
2r
)
1s
t
le
ac
h
(2
34
U
/2
38
U
)
(±
2r
)
2n
d
le
ac
h
(2
34
U
/2
38
U
)
(±
2r
)
R
es
id
ue
(2
34
U
/2
38
U
)
(±
2r
)
T
ot
al
ng
U
g
1
(±
2r
)
A
s+
f,
t
(2
34
U
/2
38
U
)a
A
s,
t
(2
34
U
/2
38
U
)b
R
s
80
7.
53
(0
.0
2)
43
3.
2
(1
.1
)
14
6.
1
(0
.4
)
18
58
.2
(5
.9
)
1.
16
7
(0
.0
03
)
1.
10
4
(0
.0
04
)
0.
93
4
(0
.0
11
)
0.
92
1
(0
.0
03
)
24
45
(6
)
0.
95
5
0.
92
2
6.
42
11
6
11
.9
4
(0
.0
4)
50
0.
5
(0
.8
)
16
4.
4
(0
.3
)
18
56
.7
(5
.8
)
1.
17
0
(0
.0
04
)
1.
11
0
(0
.0
03
)
0.
92
1
(0
.0
06
)
0.
92
4
(0
.0
02
)
25
34
(6
)
0.
96
2
0.
92
4
5.
49
45
5
2.
39
(0
.0
1)
41
6.
3
(1
.3
)
15
9.
1
(0
.5
)
16
79
.3
(2
.4
)
1.
26
4
(0
.0
11
)
1.
22
2
(0
.0
03
)
1.
06
8
(0
.0
06
)
0.
89
2
(0
.0
02
)
22
57
(3
)
0.
96
6
0.
90
7
5.
11
64
5
0.
58
(0
.0
0)
31
4.
8
(0
.4
)
15
6.
0
(0
.2
)
17
65
.4
(5
.5
)
1.
50
8
(0
.0
09
)
1.
24
5
(0
.0
03
)
1.
18
0
(0
.0
08
)
0.
93
4
(0
.0
03
)
22
37
(6
)
0.
99
5
0.
95
4
12
.4
8
82
0
7.
32
(0
.0
3)
32
1.
2
(1
.0
)
23
8.
4
(0
.8
)
17
10
.2
(5
.3
)
1.
56
2
(0
.0
10
)
1.
24
0
(0
.0
03
)
0.
99
0
(0
.0
03
)
0.
92
7
(0
.0
03
)
22
77
(5
)
0.
98
0
0.
93
5
12
.9
2
85
7
8.
31
(0
.0
3)
27
8.
7
(0
.9
)
17
8.
3
(0
.6
)
17
84
.7
(2
.5
)
1.
49
1
(0
.0
06
)
1.
14
3
(0
.0
04
)
0.
87
9
(0
.0
04
)
0.
91
6
(0
.0
02
)
22
50
(3
)
0.
94
3
0.
91
3
17
.8
6
a
M
as
s
w
ei
gh
te
d
su
m
of
al
l
fr
ac
ti
on
s
in
le
ac
hi
ng
ex
pe
ri
m
en
t.
b
M
as
s
w
ei
gh
te
d
su
m
of
re
si
du
e
an
d
se
co
nd
le
ac
h
in
le
ac
hi
ng
ex
pe
ri
m
en
t.In the second HCl leach (following ashing), four of six
depths yielded (234U/238U) values that were less than
1.000, suggesting that this step was yielding U in the min-
eral grains subject to recoil depletion. At 857 cm depth, a
minimum (234U/238U) value of 0.879 (less than 0.916
observed in the residue) was obtained in the second leach
(Table 3).
The mass-weighted summation of (234U/238U) values
from sequential leaches and residues resulted in
(234U/238U) values consistently less than 1.000
(As+f,t = 0.9430.995; Table 3 and Fig. 7), although uncer-
tainties in the mass balance of core fractions approached
5%. This could indicate that particle transport times reduced
the (234U/238U) values of loess upon deposition and burial,
or that porewaters with (234U/238U) values greater than
1.000 were removed by drainage. However, if thaw of higher
(234U/238U) ice occurred, it likely did not travel far, given
limited hydraulic conductivity observed in the loess within
this area (Jorgenson et al., 2013b; Koch et al., 2013a,b). This
would likely affect the depth distribution of elevated
(234U/238U) values locally within the core rather than being
effectively removed from any depth other than the near sur-
face, where surficial sediment movement might occur.
For each sample in the leach experiment, the initial thaw
water values of (234U/238U) were interpreted as Af,t because
sequential leaches had (234U/238U) that consistently fell
below these values (Fig. 7, Table 3). We assume that thaw-
ing under unbuffered conditions would adequately capture
the combined ice and liquid components in permafrost; this
assumption should be tested in future with a more targeted
leaching procedure developed specifically for permafrost.
Ultimately, however, because our procedure sought to cap-
ture all U present in each sample, our age estimates are con-
strained by this mass balance. Residue (234U/238U) values
were combined with values from the second leach in a
weighted average value for As,t that ranged from 0.907 to
0.954 (Table 3). These values, along with the mass-weighted
total (234U/238U) values (As+f,t; Table 3), were used to cal-
culate the mass fraction of U in the ice (Xf) and the solid-
ice ratio (Rs) assuming a two-component mixture of solid
and fluid/ice (rearranging Eq. (7b) and using the substitu-
tion Xs = 1Xf):
X f ¼
Asþf ;t  As;t
Af ;t  As;t
ð9aÞ
Rs ¼
X s
X f
ð9bÞ
The resulting Xf values ranged from 5.3% to 16.4%
(Rs = 5.1–17.9) and did not vary systematically with depth
(Table 3). These results are consistent with previous obser-
vations and provided bounds for evaluation of minimum
and maximum model ages. The mean Xf value of 10.9%
(Rs = 10.1) was used to calculate mean model ages for all
samples using Eq. (6), and the range (Rs = 5.1–17.9) used
to calculate maximum and minimum ages (Fig. 8a,
Table 4).
In individual samples where Rs was evaluated, a fitted fa
was determined based on the linear fit of mean model age
with depth (Fig. 8a, Table 4). These fitted values
(0.0640.189) were consistent with values observed in
samples where surface area was measured (0.1110.233;
Table 5). The assumption of a relatively high (3–12%)
and constant mass fraction of U in ice is supported by
(234U/238U) values at shallow depths in cores (Table 1) that
approach those in porewaters (Table 2, Fig. 6b), even as U
concentrations remain high at shallow depths in cores
(Tables 1 and 2).
Notably the inferred values for As,t (0.907 to 0.954)
imply older ages if derived from calculated values for Af+s,i
based on Eq. (8b) and using observed Af+s,t and ice ages
Fig. 7. Leach sequence in six samples from core A4: (a) (234U/238U) valu
indicate analytical uncertainties shown in Table 3, except in summed frac
total masses of water/ice and solid.
Table 4
Model ages based on measured U distribution from leaching experiment
Depth,
cm
Thaw water Af,t
(234U/238U)a
Rs
b Mean model ice
age, ky eq 6c
Minimu
age, kyd
80 1.167 6.42 1 0
116 1.170 5.49 2 0
455 1.264 5.11 24 6
645 1.508 12.48 89 29
820 1.562 12.92 103 34
857 1.491 17.86 84 27
a Uncertainties shown in Table 3.
b Derived using Eq. (9b) and Xs = 1  Xf.
c Using mean value of faRs = 1.715.
d Minimum and maximum ages derived using Eq. (6) and mean paramet
surface area measurements (Rs = 5.117.9, fa = 0.111–0.233).
e Using linear fit for the relationship between mean model age and dep
f Rearranging Eq. (6) to solve for faRs based on the linear fit of mean
g Based on linear fit and measured Rs.based on Eq. (6). This could result if (1) As,i values were ini-
tially less than 1.0, perhaps due to transport or thaw and
refreezing; (2) U distributions were not constant through
time; and/or (3) porewater values observed today differ
from those present at permafrost development. The first
possibility is effectively assumed here, such that ages may
be underestimates. The second and third possibilities are
partially addressed here through calculation of minimum
and maximum ages based on observed ranges in values of
As,i, fa, and Rs. The second possibility could be more
es, and (b) U concentration distribution (note log scale). Error bars
tions (234U/238U) where error bars indicate uncertainty of 5% in the
(core A4).
m Maximum
age, kyd
Model ice age, ky
linear fite
Linear fit
faRs
f
Fitted
fa
g
33 5 0.417 0.065
36 8 0.433 0.079
132 30 1.020 0.200
221 43 2.261 0.181
502 54 2.115 0.164
627 57 1.707 0.096
er values over the observed range from the leaching experiment and
th (fitted age (ky) = 0.09888  depth (cm); Fig. 8).
model age with depth.
specifically examined through more targeted observations
of U distribution with depth in syngenetic permafrost cores.
All three could be better evaluated through combination of
U-series age assessment with other age constraint in ice-rich
loess permafrost.
3.2.2. Particle size distribution, surface area and recoil loss
fractions
Particle surface area is a primary control on the evolu-
tion of (234U/238U) in adjacent fluids over time (DePaolo
et al., 2006; Maher et al., 2006; Bourdon et al., 2009; Lee
et al., 2010; Handley et al., 2013). Bulk surface area in five
samples from core A9 (Table 5) reveal smoothing associ-
ated with apparent organic coatings (BET surface area
increased by a factor of 2–5 by heating to 350 C). Recoil
Fig. 8. Model ice ages with depth for all samples from cores A4 (a) and
observed values for recoil fraction based on BET surface area (fa = 0.
(Rs = 10.05). Solid grey lines indicate maximum model ages based on th
maximum faRs = 4.167 in Eq. (6). Solid black lines indicate linear fits
determined using the lm function in R (R Core Team, 2012), indicating t
zero. Circles indicate points where fa (A9, Table 5) or Rs (A4, Table 5) wa
age uncertainty, based on variation in model parameters (Table 4, Tablefractions (fa) were calculated from the BET surface area
of the treated samples using Eq. (1), which consistently
exceeded those from the particle size distributions using
Eq. (2) (Table 5). Treated BET surface area and PSD
derived spherical surface areas (bk = 1) were positively cor-
related (R2 = 0.61, n = 5) with a slope of 28, a value sugges-
tive of likely aspect ratios (particle dimensions). We note
that hydrogen peroxide treatment rather than ashing may
be preferred for preservation of mineral surface properties;
however the consistency of BET and PSD results is encour-
aging. Considered individually, BET measurements implied
roughness factors of 16–37 compared to smooth spheres,
and generally increased with depth, suggesting weathering
in cores (Table 5). Lower recoil fractions based on spherical
particles or using a fractal model to adjust BET derived
A9 (b). Squares are mean model ice ages using Eq. (6) with mean
171) and U distribution in the mineral solid compared to the ice
e minimum faRs = 0.565 in Eq. (6) and minimum values based on
of within-core mean model ages to depth, with R2 and p values
he age increases with depth are significant and likely different from
s measured. Dark grey lines indicate the upper and lower bounds of
5).
recoil fractions (D = 2.5 resulting in fa = 0.022 to 0.047)
(Bourdon et al., 2009; Lee et al., 2010; Handley et al.,
2013) were not consistent with calculated solid activity val-
ues (As,i; As,t < (1  fa) violating Eq. (5)). Recoil fractions
were based on treated BET surface area using Eq. (1) as
the more direct measure of that parameter (Table 5). In
all samples, the mean observed recoil fraction from BET
2004; Aciego et al., 2011). In individual samples where
BET was measured, a fitted Rs value was determined based
on the linear fit between mean model age and depth
(Fig. 8b, Table 5). These values (Rs = 5.27–16.25) were con-
sistent with values derived from the leaching experiment
(Table 3; Rs = 5.11–17.86).
Table 5
Model ages based on measured surface areas, DOC radiocarbon results, and particle size distributions (Core A9).
Depth,
cm
Thaw
water Af,t
(234U/238U)
Surface
area,
m2 g1
spheres
Surface
area,
m2 g1
BET
untreated
Surface
area,
m2 g1
BET -OM
Recoil
fraction, fa
Spheres
bk = 10 eq 2
Recoil
fraction, fa
BET-OM
eq 1
Mean
model ice
age, ky eq
6a,b
Minimum
age, kyc
Maximum
age, kyc
Model
ice age,
ky linear
fitd
Linear
fit
faRs
e
Fitted
Rs
f
102 1.182 0.326 1.843 9.069 0.069 0.183 7.7 0 47 7 1.252 6.84
1315 1.560 0.248 1.294 5.487 0.054 0.111 99 25 169 87 1.985 17.93
1399 1.366 0.195 3.294 7.147 0.042 0.144 105 16 173 93 1.048 7.27
1924 1.904 0.574 3.802 9.050 0.061 0.183 144 56 214 128 2.613 14.31
2069 1.541 0.551 3.887 11.566 0.059 0.233 155 37 245 137 1.342 5.75
a At 102 cm, the 14C age was 8110 y. This is the only sample in which both 14C-DOC and surface area were evaluated.
b Using mean value of faRs = 1.715.
c Minimum and maximum ages derived using Eq. (6) and mean parameter values over the observed range from the leaching experiment and
surface area measurements (Rs = 5.1–17.9, fa = 0.111–0.233; faRs = 0.565–4.1674).
d Using linear fit for the relationship between mean model age and depth (fitted age (ky) = 0.06628  depth (cm); Fig. 7).
e Rearranging Eq. (6) to solve for faRs based on the linear fit of mean model age with depth.
f Based on linear fit and measured fa.measurements (fa = 0.171) was used to calculate mean
model ages, and the range (0.111–0.233) was used to evalu-
ate maximum and minimum ages (Fig. 8b, Table 5). The
observed range of recoil fractions is consistent with previ-
ous observations in ice (Fireman, 1986; Goldstein et al.,Fig. 9. Model ice ages with depth for all samples from cores A4, A5, A
symbols are mean model ice ages using Eq. (6) with mean observed value
distribution in the mineral solid compared to the ice (Rs = 10.05). Ligh
faRs = 0.565 in Eq. (6) (crosses) and minimum values based on maximum
Core Team, 2012) are indicated, with the slope of the mean fit (dark solid
linear fit is significantly different from zero and is used to assign fitted mo
were directly measured, the mean observed faRs value (1.72) is used to cal
3000% are implied by the maximum and minimum model ages.3.2.3. Age model applied to all cores
With the assumption that the distribution of U between
the ice and mineral phases has not changed over the lifetime
of the permafrost, the age of the ice was then calculated
using Eq. (6). To derive model ice ages using Eq. (6), the6, A7 and A9 (linear fit improved slightly by omitting A2). Dark
s for recoil fraction based on BET surface area (fa = 0.171) and U
t symbols indicate maximum model ages based on the minimum
faRs = 4.167 in Eq. (6) (bars). Linear fits of model age to depth (R
line) implying a permafrost accumulation rate of 0.16 mm y1. This
del ages in Tables 4–6 based on depth. For samples where fa or Rs
culate the second parameter in Tables 4 and 5. Uncertainties of 30–
Table 6
(234U/238U), DOC concentrations, and model ages (mean and range) in thaw waters.
Sample namea Thaw water
Af,t, (
234U/238U)
±2r DOC,
ppm
DOC 14C
age, kyb
Mean
model
age, kyc
Min
age, kyd
Max
age, kyd,e
Linear with
depth model
age, kyf
faRs from
linear fit
HCDT.A2.0885 1.224 0.007 15 2 63 56 0.588
HCDT.A2.0913 1.223 0.006 14 2 62 58 0.570
HCDT.A2.0931 1.214 0.005 1081.0 12 1 53 59 0.504
HCDT.A2.0996 1.214 0.006 1066.0 12 1 53 63 0.483
HCDT.A2.1016 1.184 0.009 1077.0 5 0 24 64 0.297
HCDT.A4.0080 1.167 0.003 3.9 1 0 9 5 0.559
HCDT.A4.0116 1.170 0.004 3.7 2 0 11 7 0.582
HCDT.A4.0338 1.326 0.007 40 11 76 21 2.961
HCDT.A4.0445 1.264 0.011 1348.0 24 6 107 28 1.500
HCDT.A4.0510 1.366 0.060 50 15 136 32 2.503
HCDT.A4.0645 1.508 0.009 864.0 89 29 196 41 3.335
HCDT.A4.0674 1.395 0.027 58 18 209 43 2.213
HCDT.A4.0750 1.360 0.007 48 15 243 48 1.740
HCDT.A4.0820g 1.555 0.005 1338.0 103 33 478 52 3.040
HCDT.A4.0857 1.491 0.006 567.0 84 27 603 54 2.475
HCDT.A5.0101 1.195 0.008 75.4 9.88 8 0 35 6 2.037
HCDT.A5.0361 1.373 0.078 585.4 52 16 67 23 3.537
HCDT.A5.0384 1.271 0.004 782.2 26 6 69 24 1.809
HCDT.A5.0437 1.261 0.006 975.5 23 6 76 28 1.483
HCDT.A5.0511 1.251 0.006 1151.6 21 5 85 32 1.185
HCDT.A5.0545 1.248 0.005 599.2 20 4 89 35 1.092
HCDT.A5.0585 1.260 0.005 398.7 23 5 90 37 1.152
HCDT.A5.0627 1.273 0.007 353.7 >41.0 26 7 92 40 1.211
HCDT.A5.0660 1.251 0.005 392.6 21 5 92 42 0.964
HCDT.A6.0151 1.173 0.005 3 0 14 10 0.598
HCDT.A6.1322 1.495 0.021 85 27 622 84 1.742
HCDT.A6.1380 1.502 0.011 1499.3 87 28 659 87 1.715
HCDT.A6.1500 1.497 0.011 1550.5 86 28 632 95 1.584
HCDT.A6.1560 1.379 0.004 53 16 277 99 1.053
HCDT.A6.1594 1.551 0.005 >40.3 102 33 1187 101 1.728
HCDT.A6.1762 1.271 0.005 869.7 26 6 116 112 0.566
HCDT.A9.0035 1.163 0.012 97.9 0 0 5 2 0.428
HCDT.A9.0055 1.168 0.014 59.6 2 0 10 3 0.840
HCDT.A9.0102 1.182 0.006 49.0 8.11 5 0 22 6 1.301
HCDT.A9.1102 1.322 0.037 775.2 >36.4 39 11 145 70 1.057
HCDT.A9.1130 1.297 0.006 275.4 32 9 148 72 0.901
HCDT.A9.1237 1.386 0.010 544.8 55 17 190 78 1.290
HCDT.A9.1315g 1.468 0.014 498.7 78 25 221 83 1.621
HCDT.A9.1399 1.366 0.004 364.3 50 15 254 89 1.084
HCDT.A9.1695 1.366 0.004 711.5 50 15 254 107 0.942
HCDT.A9.1715 1.716 0.006 743.5 156 49 332 109 2.256
HCDT.A9.1770 1.703 0.005 790.4 >39.0 151 48 547 112 2.153
HCDT.A9.1849 1.529 0.007 969.7 95 31 856 117 1.464
HCDT.A9.1879 1.594 0.008 930.7 115 37 872 119 1.674
HCDT.A9.1924 1.772 0.011 682.2 176 55 896 122 2.254
HCDT.A9.1956 1.904 0.014 1274.5 230 69 914 124 2.671
HCDT.A9.1977 1.768 0.007 618.2 175 55 925 125 2.193
HCDT.A9.2005 1.655 0.011 598.3 135 43 940 127 1.796
HCDT.A9.2050 1.638 0.016 709.5 129 41 964 130 1.711
HCDT.A9.2069 1.589 0.017 742.8 114 37 974 131 1.541
HCDT.A9.2114 1.541 0.018 966.0 99 32 999 134 1.364
HCCN.1.053g 1.163 0.006 0 0 5 3 0.337
HCCN.1.119 1.162 0.003 0 0 4 8 0.193
HCCN.1.150 1.166 0.004 1 0 8 10 0.337
HCCN.2.127 1.165 0.003 1 0 7 8 0.324
HCCN.2.165 1.169 0.003 2 0 11 10 0.424
ed by
,000
s from
ramet
0.565
h dep
d dep
agatemean recoil loss fraction from the treated BET measure-
ments (fa = 0.171) was combined with the mean mass frac-
tion of U in the mineral compared to the ice based on the
leaching experiment (Rs = 10.1) and the average porewater
(234U/238U) value (Af,i = 1.161), resulting in mean model
ages ranging from 1 ky at 1 m depth to 200 ky at
21 m depth (Fig. 9, Table 6). Use of the alternate pore-
water value (Af,i = 1.191; Richardson location, Table 2),
resulted in undefined ages in the uppermost depths where
thaw water values were lower than 1.191 (Table 1), and dif-
ferences of 10% in mean ages at greatest depth. Therefore
the value of 1.161 was used, but this is a potentially sub-
stantial source of uncertainty that merits further investiga-
tion in future permafrost studies.
These mean (234U/238U) model ages are consistent with
measured 14C-DOC ages in cores where radiocarbon data
were collected (A5, A6 and A9; Table 6). Radiocarbon ages
of 9880 ± 45 14C y BP and 8110 ± 45 14C y BP were
obtained for DOC in thaw water samples at 101-cm depth
(A5.0101; Table 6) and 102-cm depth (A9.0102; Table 6),
respectively. These ages are consistent with depth sequences
of 14C values in adjacent soils and upper permafrost
(O’Donnell et al., 2011a). DOC at greater depths (6–18 m)
HCCN.2.185 1.188 0.003
HCCN.2.205 1.180 0.006
HCCN.2.255 1.210 0.008
HC03.5.060 1.196 0.006
HC03.5.112 1.203 0.004
HC03.5.150 1.226 0.013
HC03.5.190 1.273 0.027
a Sample names indicate location, core, and depth in cm, separat
b Uncertainties of 45 y; blanks of FM = 0.008 limited ages to <39
c Estimated model ages using Eq. (6) and mean parameter value
fa = 0.171).
d Minimum and maximum ages derived using Eq. (6) and mean pa
surface area measurements (Rs = 5.1–17.9, fa = 0.111–0.233; faRs =
e Maximum age in cores A4, A5 and A9 restricted to increase wit
depth.
f using linear fit for the relationship between mean model age an
g Analyzed in duplicate for (234U/238U); uncertainties reflect propin all three cores (A5, A6, A9) was 14C dead
(>50,000 y BP) within the uncertainty of the blank contribu-
tion from sample handling (fraction modern = 0.008;
Table 6).
The observed range of fa and Rs derived from the leach-
ing experiment and surface area measurements (fa-
Rs = 0.57–4.17) along with the range of observed
porewater (Af,i) values (1.157–1.199; Table 2) were used
to derive minimum and maximum ages for all core subsam-
ples (Fig. 9, Table 6). Based on the overall increase in mean
model age with depth, a linear fit between depth and the
mean model ages in cores A4 through A9 (Fig. 9) was then
used to generate a fitted relationship of model age to depth,
such that deeper samples were always modeled as older
than shallower samples (Fig. 9). The fitted mean model ages
were consistent with radiocarbon results (Table 6) and
implied values of faRs (0.19–3.54) that were close to values
based on the surface area and leaching experiment results(0.57–4.17). The model-age uncertainties of 30-3000%
reflect a degree of uncertainty consistent with other studies
(Aciego et al., 2011), and speak to the need for combining
U series dating approaches in permafrost with other age
constraints or geochronologic tools.
Generally, the observed (234U/238U) values indicate min-
imum model ages up to 70 ka and mean model ages up to
200 ka at the base of the core sequence (Table 6, Fig. 9).
This encompasses an interval of multiple glacial advances in
the Brooks Range (Briner and Kaufman, 2008), and sug-
gests that some regions of the carbon- and ice-rich loess
permafrost of interior Alaska have survived more than
one full glacial cycle. This is consistent with observations
interpreted as long-term persistence of ground ice at the
Palisades Bluff near Galena, Alaska (USA) and in the
Yukon Territory (Canada) through the last large intergla-
cial (Froese et al., 2008) and highlights the potential resil-
ience of yedoma (ice- and organic-rich loess permafrost)
to climate change (Jorgenson et al., 2013b). In addition,
recent work on speleothems suggests preservation of per-
mafrost for >400 ka at similar latitudes in Siberia (Vaks
et al., 2013). However, another 30-m-thick ice-rich loess
permafrost deposit in northern Alaska dates only about
6 0 28 12 0.978
4 0 20 13 0.678
11 1 49 16 1.250
8 0 35 4 3.400
10 0 42 7 2.256
15 2 65 10 2.597
26 7 118 12 3.494
decimal points.
y.
leaching experiment and surface area measurements (Rs = 10.1,
er values over the observed range from the leaching experiment and
–4.1674).
th where values were undefined or showed apparent decrease with
th in all cores but A2 (fitted age = 0.06338  depth).
d error for two analyses.50 ka based on 14C and thermo-luminescence dating
(Kanevskiy et al., 2011). In a separate set of samples from
this Hess Creek site, 14C results suggest younger ages
(22 to 43 ka) (Kanevskiy et al., 2012), although blank sig-
nal was uncertain in that analysis. Differences between
results from the Kanevskiy et al. study and those presented
here illustrate the need for further efforts to combine multi-
ple dating techniques in evaluating the age of yedoma
deposits. In particular, the geomorphic context of loess
sequences must be considered (Jorgenson et al., 2013b;
Koch et al., 2013a). At Hess Creek, a context of hillslope
evolution with stream incision has likely resulted in both
downslope erosion resulting in exposure of ground ice to
thaw; and protection of older ice bodies now overtopped
with material from upslope.
The relationship between sample age and depth reflects
loess net accumulation rate. Loess deposition rates of
0.002 to 1 mm y1 are reported in the region (Muhs
et al., 2003). Accounting for syngenetic ice content of 50%
by volume of the total core, total accumulation rates of
0.004 to 2 mm y1 would suggest ages of 10–500 ky in
the deepest (21 m) core here based on these rates. The linear
fits in Figs. 8 and 9 suggest accumulation rates of
0.1 mm y1, broadly consistent with regional rates. There-
fore, sediment movement by erosion and deposition would
appear to be limited at this site, or may have mainly
occurred early on following loess deposition.
3.3. DOC in thaw waters
DOC concentrations in the deep core thaw waters ana-
lyzed in this study increase with depth to maxima
>1000 ppm (Fig. 10, Table 6), suggesting long-term, in-situ
generation of DOC in the unfrozen water fraction of this
carbon-rich and presumably oxygen-poor subsurface envi-
ronment (Romanovsky and Osterkamp, 2000; Weintraub
and Schimel, 2003, 2005). Radiocarbon-dead DOC and
modeled ice ages of 3 to 160 ka imply long-term, slow
production of DOC from the included solid phase organic
matter. This closed-system production of DOC under
anoxic conditions implies that thaw of this permafrost will
release a pulse of concentrated, ancient DOC that would
likely be highly labile to decomposition upon exposure to
oxygen. This DOC character would be consistent with
observations of old, labile DOC in meltwater from retreat-
ing glaciers in coastal Alaska (Hood et al., 2009), but would
contrast with observations of young, labile DOC in arctic
rivers and streams during spring runoff (Neff et al., 2006).
face waters draining permafrost dominated areas. AlthoughFig. 10. Dissolved organic carbon (DOC) concentrations with
depth in water separated from thawed deep permafrost cores.uncertainties in the modeled ages presented here are sub-
stantial (30–3000%), other more suitable sites may permit
the combination of this U-series approach with precise indi-
cators of solid phase age, such as tephra deposits, allowing
greater refinement of model parameters to yield more pre-
cise results (Froese et al., 2008; Jensen et al., 2008, 2011,
2013; Reyes et al., 2010a, b, 2011). In addition, BET surface
area should be routinely assessed as argued in other studies,
and while the leaching procedure used here provided a rea-
sonable first step, it could be refined. In particular, use of
buffered solutions and particle sizing should be considered
(Goldstein et al., 2004; Aciego et al., 2011; Suresh et al.,
2013, 2014). Nonetheless, this work provides evidence that
preservation of ground ice through periods of climate
warming is a function of landscape processes such as insu-
lation by downslope sediment redistribution, peat accumu-
lation, and plant community succession, as well as the
intrinsic properties of syngenetic loess permafrost (high
ice content, low hydraulic conductivity upon thaw, location
on water-shedding slopes). Our results support the assertion
that thaw of the 1.5 m of material overlying ice wedges
occurred during the Holocene, as indicated by 14C ages of
<10 ka at depths of 1 m. Whether the shallow thawing
developed in response to a warmer climate during the early
Holocene or to repeated fire disturbance is uncertain.
Nonetheless, our observed (234U/238U) values in deep yed-
oma cores suggest mean model ages up to 200 ka, consis-
tent with other evidence of permafrost preservation for
timescales on the order of 100 ka (Froese et al., 2008;
Vaks et al., 2013). Moreover, increasing DOC with depth
in cores implies ongoing production of DOC from the sur-
rounding organic matter over these timescales. Upon thaw
of loess permafrost, this ancient DOC may be rapidly min-
eralized, even as the signal of ancient ice as reflected in
(234U/238U) values should be released into surface water
networks.
ACKNOWLEDGEMENTS
We thank Joshua Koch for ongoing discussion and review of
the manuscript. We also thank Leonid Neymark for useful discus-
sion and insight, Kenna Butler for fast and excellent DOC analysis,
David Butman and Peter Raymond for DOC-14C analysis, Paul
Schuster and Kathy Kelsey for IC analysis, Tom Oliver for ICP
analysis, Merith Reheis and Harland Goldstein for particle size
analysis, and Kyungsoo Yoo for surface area analysis. Funding
for this work was provided by the USGS climate program, the
NSF Carbon-Water program (NSF EAR No. 0630319), and the
Montana State University Vice President of Research. We thankThis pulsed release of old, labile DOC in permafrost thaw
waters could be substantial: DOC in core A9 accounts for
15 kg C m2, or 3.1% of the total C inventory of
480 kg m2.
4. CONCLUSIONS
This use of U isotopes to directly evaluate the residence
time of ice in permafrost demonstrates the potential utility
of this technique for understanding past chronologies of cli-
mate, and potentially for detecting a signal of thaw in sur-
six anonymous reviewers for detailed and insightful comments on
earlier versions of this manuscript. Any use of trade, firm, or prod-
uct names is for descriptive purposes only and does not imply
endorsement by the U.S. Government.
REFERENCES
Aciego S., Bourdon B., Schwander J., Baur H. and Forieri A.
(2011) Toward a radiometric ice clock: uranium ages of the
Dome C ice core. Quat. Sci. Rev. 30, 2389–2397.
Aiken G. R. (1992) Chloride interference in the analysis of
dissolved organic carbon by the wet oxidation method. Environ.
Sci. Technol. 26, 2435–2439.
Andersen M. B., Erel Y. and Bourdon B. (2009) Experimental
evidence for U-234–U-238 fractionation during granite weath-
ering with implications for U-234/U-238 in natural waters.
Geochim. Cosmochim. Acta 73, 4124–4141.
Barclay D. J., Wiles G. C. and Calkin P. E. (2009) Holocene glacier
fluctuations in Alaska. Quat. Sci. Rev. 28, 2034–2048.
Blinov A., Alfimov V., Beer J., Gilichinsky D., Schirrmeister L.,
Kholodov A., Nikolskiy P., Opel T., Tikhomirov D. and
Wetterich S. (2009) Ratio of Cl-36/Cl in ground ice of east
Siberia and its application for chronometry. Geochem. Geophys.
Geosyst., 10
Bourdon B., Turner S., Henderson G. M. and Lundstrom C. C.
(2003) Introduction to U-series geochemistry. Uranium Series
Geochem. 52, 1–21.
Bourdon B., Bureau S., Andersen M. B., Pili E. and Hubert A.
(2009) Weathering rates from top to bottom in a carbonate
environment. Chem. Geol. 258, 275–287.
Briner J. P. and Kaufman D. S. (2008) Late Pleistocene mountain
glaciation in Alaska: key chronologies. J. Quat. Sci. 23, 659–
670.
Carey S. K. and Woo M. K. (2002) Hydrogeomorphic relations
among soil pipes, flow pathways, and soil detachments within a
permafrost hillslope. Phys. Geogr. 23, 95–114.
Chabaux F., Dequincey O., Le´ve`que J.-J., Leprun J.-C., Clauer N.,
Riotte J. and Paquet H. (2003a) Tracing and dating recent
chemical transfers in weathering profiles by trace-element
geochemistry and 238U–234U–230Th disequilibria: the example
of the Kaya lateritic toposequence (Burkina-Faso). C.R. Geosci.
335, 1219–1231.
Chabaux F., Riotte J. and Dequincey O. (2003b) U–Th–Ra
fractionation during weathering and river transport. Uranium
Series Geochem. 52, 533–576.
Chabaux F., Ma L., Stille P., Pelt E., Granet M., Lemarchand D.,
Roupert R. d. C. and Brantley S. L. (2011) Determination of
chemical weathering rates from U series nuclides in soils and
weathering profiles: principles, applications and limitations.
Appl. Geochem. 26(Supplement), S20–S23.
Cheng H., Edwards R. L., Hoff J., Gallup C. D., Richards D. A.
and Asmerom Y. (2000) The half-lives of uranium-234 and
thorium-230. Chem. Geol. 169, 17–33.
DePaolo D. J., Maher K., Christensen J. N. and McManus J.
(2006) Sediment transport time measured with U-series iso-
topes: results from ODP North Atlantic drift site 984. Earth
Planet. Sci. Lett. 248, 394–410.
DePaolo D. J., Lee V. E., Christensen J. N. and Maher K. (2012)
Uranium comminution ages: sediment transport and deposition
time scales. C.R. Geosci. 344, 678–687.
Dosseto A., Bourdon B., Gaillardet J., Allegre C. J. and Filizola N.
(2006a) Time scale and conditions of weathering under tropical
climate: study of the Amazon basin with U-series. Geochim.
Cosmochim. Acta 70, 71–89.Dosseto A., Bourdon B., Gaillardet J., Maurice-Bourgoin L. and
Allegre C. J. (2006b) Weathering and transport of sediments in
the Bolivian Andes: time constraints from uranium-series
isotopes. Earth Planet. Sci. Lett. 248, 759–771.
Dosseto A., Bourdon B. and Turner S. P. (2008) Uranium-series
isotopes in river materials: insights into the timescales of
erosion and sediment transport. Earth Planet. Sci. Lett. 265, 1–
17.
Dosseto A., Hesse P. P., Maher K., Fryirs K. and Turner S. (2010)
Climatic and vegetation control on sediment dynamics during
the last glacial cycle. Geology 38, 395–398.
Euskirchen E. S., McGuire A. D., Kicklighter D. W., Zhuang Q.,
Clein J. S., Dargaville R. J., Dye D. G., Kimball J. S.,
McDonald K. C., Melillo J. M., Romanovsky V. E. and Smith
N. V. (2006) Importance of recent shifts in soil thermal
dynamics on growing season length, productivity, and carbon
sequestration in terrestrial high-latitude ecosystems. Global
Change Biol. 12, 731–750.
Fireman E. L. (1986) Uranium series dating of Allan Hills ice. J.
Geophys. Res. Solid Earth Planet. 91, D539–D544.
Fleischer R. L. (1983) Theory of alpha-recoil effects of radon
release and isotopic disequilibrium. Geochim. Cosmochim. Acta
47, 779–784.
French H. and Shur Y. (2010) The principles of cryostratigraphy.
Earth Sci. Rev. 101, 190–206.
Froese D. G., Smith D. G. and Clement D. T. (2005) Character-
izing large river history with shallow geophysics: middle Yukon
River, Yukon Territory and Alaska. Geomorphology 67, 391–
406.
Froese D. G., Westgate J. A., Reyes A. V., Enkin R. J. and Preece
S. J. (2008) Ancient permafrost and a future, warmer arctic.
Science 321, 1648-1648.
Gascoyne M. (1992) Paleoclimate determination from cave calcite
deposits. Quat. Sci. Rev. 11, 609–632. Available at: http://WOS:
A1992JZ34900001.
Goldstein S. J., Murrell M. I., Nishiizumi K. and Nunn A. J. (2004)
Uranium-series chronology and cosmogenic Be-10–Cl-36
record of Antarctic ice. Chem. Geol. 204, 125–143.
Grosse G., Harden J., Turetsky M., McGuire A. D., Camill P.,
Tarnocai C., Frolking S., Schuur E. A. G., Jorgenson T.,
Marchenko S., Romanovsky V., Wickland K. P., French N.,
Waldrop M., Bourgeau-Chavez L. and Striegl R. G. (2011)
Vulnerability of high-latitude soil organic carbon in North
America to disturbance. J. Geophys. Res. Biogeosci., 116
Hamilton T. D. (1982) A late Pleistocene glacial chronology for the
southern Brooks Range – stratigraphic record and regional
stratigraphy. Geol. Soc. Am. Bull. 93, 700–716.
Hamilton T. D. (2001) Quaternary glacial, lacustrine, and fluvial
interactions in the western Noatak basin, Northwest Alaska.
Quat. Sci. Rev. 20, 371–391.
Handley H. K., Turner S., Afonso J. C., Dosseto A. and Cohen T.
(2013) Sediment residence times constrained by uranium-series
isotopes: a critical appraisal of the comminution approach.
Geochim. Cosmochim. Acta 103, 245–262.
Harden J. W., O’Neill K. P., Trumbore S. E., Veldhuis H. and
Stocks B. J. (1997) Moss and soil contributions to the annual
net carbon flux of a maturing boreal forest. J. Geophys. Res.
Atmos. 102, 28805–28816.
Harris C., Kern-Luetschg M., Murton J., Font M., Davies M. and
Smith F. (2008) Solifluction processes on permafrost and non-
permafrost slopes: Results of a Large-scale Laboratory Simu-
lation. Permafr. Periglac. Process. 19, 359–378. Available at:
http://WOS:000261663000003.
Hood E., Fellman J., Spencer R. G. M., Hernes P. J., Edwards R.,
D’Amore D. and Scott D. (2009) Glaciers as a source of ancient
S.A. Ewing et al. /Geochimica et Cosmochimica Acta 152 (2015) 143–165 163and labile organic matter to the marine environment. Nature
462, U1044–U1100.
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. (2014) Estimated stocks of
circumpolar permafrost carbon with quantified uncertainty
ranges and identified data gaps. Biogeosciences 11, 6573–6593.
Available at: http://www.biogeosciences.net/11/6573/2014/
[Accessed December 4, 2014].
Jaffey A. H., Flynn K. F., Glendeni L, Bentley W. C. and Essling
A. M. (1971) Precision measurements of half-lives and specific
activities of 235U and 238U. Phys. Rev. C 4, 1889–1906.
Jensen B. J. L., Froese D. G., Preece S. J., Westgate J. A. and
Stachel T. (2008) An extensive middle to late Pleistocene
tephrochronologic record from east-central Alaska. Quat. Sci.
Rev. 27, 411–427.
Jensen B. J. L., Preece S. J., Lamothe M., Pearce N. J. G., Froese
D. G., Westgate J. A., Schaefer J. and Beget J. (2011) The
variegated (VT) tephra: a new regional marker for middle to
late marine isotope stage 5 across Yukon and Alaska. Quat. Int.
246, 312–323.
Jensen B. J. L., Reyes A. V., Froese D. G. and Stone D. B. (2013)
The Palisades is a key reference site for the middle Pleistocene
of eastern Beringia: new evidence from paleomagnetics and
regional tephrostratigraphy. Quat. Sci. Rev. 63, 91–108.
Jorgenson M. T., Romanovsky V., Harden J., Shur Y., O’Donnell
J., Schuur E. A. G., Kanevskiy M. and Marchenko S. (2010)
Resilience and vulnerability of permafrost to climate change.
Can. J. For. Res. 40, 1219–1236.
Jorgenson M. T., Harden J., Kanevskiy M., Donnell J. O.,
Wickland K., Ewing S., Manies K., Zhuang Q., Shur Y., Striegl
R. and Koch J. (2013a) Reorganization of vegetation, hydrol-
ogy and soil carbon after permafrost degradation across
heterogeneous boreal landscapes. Environ. Res. Lett. 8, 1–13.
Jorgenson M. T., Harden J., Kanevskiy M., O’Donnell J. A.,
Wickland K. P., Ewing S. A., Manies K., Zhuang Q., Shur Y.,
Striegl R. G. and Koch J. (2013b) Reorganization of vegetation,
hydrology and soil carbon after permafrost degradation across
heterogeneous boreal landscapes. Environ. Res. Lett., 8
Kane E. S., Valentine D. W., Schuur E. A. G. and Dutta K. (2005)
Soil carbon stabilization along climate and stand productivity
gradients in black spruce forests of interior Alaska. Can. J. For.
Res. 35, 2118–2129.
Kanevskiy M., Shur Y., Fortier D., Jorgenson M. T. and Stephani
E. (2011) Cryostratigraphy of late Pleistocene syngenetic
permafrost (yedoma) in northern Alaska, Itkillik River expo-
sure. Quat. Res. 75, 584–596.
Kanevskiy M., Shur Y., Connor B., Dillon M. R., Stephani E. and
O’Donnell J. A. (2012) Study of the ice-rich syngenetic
permafrost for road design (Interior Alaska). In Tenth Inter-
national Conference on Permafrost (ed. K. M. Hinkel). The
Northern Publisher, Salekhard, Russia, pp. 191–196.
Kanevskiy M., Jorgenson M. T., Shur Y., O’Donnell J. A., Harden
J. W., Zhuang Q. and Fortier D. (2014) Cryostratigraphy and
permafrost evolution in lacustrine lowlands of West-Central
Alaska. Permafrost Periglac. Process.
Kigoshi K. (1971) Alpha recoil thorium-234 – dissolution into
water and uranium-234/uranium-238 disequilibrium in nature.
Quat. Int. 173, 47, 47-&.
Koch J. C., Ewing S. A., Striegl R. and McKnight D. M. (2013a)
Rapid runoff via shallow throughflow and deeper preferential
flow in a boreal catchment underlain by frozen silt (Alaska,
USA). Hydrogeol. J. 21, 93–106.
Koch J. C., Runkel R. L., Striegl R. and McKnight D. M. (2013b)
Hydrologic controls on the transport and cycling of carbon andnitrogen in a boreal catchment underlain by continuous
permafrost. J. Geophys. Res. Biogeosci. 118, 698–712.
Koven C. D., Ringeval B., Friedlingstein P., Ciais P., Cadule P.,
Khvorostyanov D., Krinner G. and Tarnocai C. (2011)
Permafrost carbon-climate feedbacks accelerate global warm-
ing. Proc. Natl. Acad. Sci. USA 108, 14769–14774.
Kraemer T. F. and Brabets T. P. (2012) Uranium isotopes (U-234/
U-238) in rivers of the Yukon Basin (Alaska and Canada) as an
aid in identifying water sources, with implications for moni-
toring hydrologic change in arctic regions. Hydrogeol. J. 20,
469–481.
Lachenbruch A. H. and Marshall B. V. (1986) Changing climate –
geothermal evidence from permafrost in the Alaskan arctic.
Science 234, 689–696.
Lawrence D. M., Slater A. G., Tomas R. A., Holland M. M. and
Deser C. (2008) Accelerated Arctic land warming and perma-
frost degradation during rapid sea ice loss. Geophys. Res. Lett.,
35
Lee V. E., DePaolo D. J. and Christensen J. N. (2010) Uranium-
series comminution ages of continental sediments: Case study
of a Pleistocene alluvial fan. Earth Planet. Sci. Lett. 296, 244–
254.
Ludwig K. R., Wallace A. R. and Simmons K. R. (1985) The
Schwartzwalder uranium deposit, II: age of uranium mineral-
ization and Pb-isotope constraints on genesis. Econ. Geol. 80,
1858–1871.
Maher K., DePaolo D. J. and Christensen J. N. (2006) U–Sr
isotopic speedometer: fluid flow and chemical weathering rates
in aquifers. Geochim. Cosmochim. Acta 70, 4417–4435.
Morgenstern U., Stewart M. K. and Stenger R. (2010) Dating of
streamwater using tritium in a post nuclear bomb pulse world:
continuous variation of mean transit time with streamflow.
Hydrol. Earth Syst. Sci. 14, 2289–2301.
Muhs D. R. and Budahn J. R. (2006) Geochemical evidence for the
origin of late Quaternary loess in central Alaska. Can. J. Earth
Sci. 43, 323–337.
Muhs D. R., Ager T. A., Bettis E. A., McGeehin J., Been J. M.,
Beget J. E., Pavich M. J., Stafford T. W. and Stevens D. S. P.
(2003) Stratigraphy and palaeoclimatic significance of Late
Quaternary loess-palaeosol sequences of the Last Interglacial-
Glacial cycle in central Alaska. Quat. Sci. Rev. 22,
1947–1986.
Muhs D. R., Bettis, III, E. A., Aleinikoff J. N., McGeehin J. P.,
Beann J., Skipp G., Marshall B. D., Roberts H. M., Johnson
W. C. and Benton R. (2008) Origin and paleoclimatic signif-
icance of late Quaternary loess in Nebraska: evidence from
stratigraphy, chronology, sedimentology, and geochemistry.
Geol. Soc. Am. Bull. 120, 1378–1407.
Neff J. C., Finlay J. C., Zimov S. A., Davydov S. P., Carrasco J. J.,
Schuur E. A. G. and Davydova A. I. (2006) Seasonal changes in
the age and structure of dissolved organic carbon in Siberian
rivers and streams. Geophys. Res. Lett., 33
O’Donnell J. A., Harden J. W., McGuire A. D., Kanevskiy M. Z.,
Jorgenson M. T. and Xu X. (2011a) The effect of fire and
permafrost interactions on soil carbon accumulation in an
upland black spruce ecosystem of interior Alaska: implications
for post-thaw carbon loss. Global Change Biol. 17, 1461–1474.
O’Donnell J. A., Harden J. W., McGuire A. D. and Romanovsky
V. E. (2011b) Exploring the sensitivity of soil carbon dynamics
to climate change, fire disturbance and permafrost thaw in a
black spruce ecosystem. Biogeosciences 8, 1367–1382.
O’Donnell J., Aiken G., Walvoord M., Raymond P., Butler K.,
Dornblaser M. and Heckman K. (2014) Using dissolved
organic matter age and composition to detect permafrost thaw
in boreal watersheds of interior Alaska. J. Geophys. Res.
Biogeosciences, 2155–2170. Available at: http://onlinelibrary.
164 S.A. Ewing et al. /Geochimica et Cosmochimica Acta 152 (2015) 143–165wiley.com/doi/10.1002/2014JG002695/full [Accessed December
30, 2014].
Osmond J. K. and Cowart J. B. (2000) U-series nuclides as tracers
in groundwater hydrology. In Environmental Tracers in Catch-
ment Hydrology (eds. P. Cook and A. L. Herczeg). Kluwer
Academic Publishers, The Netherlands, pp. 145–173.
Osmond J. K., Cowart J. B. and Ivanovich M. (1983) Uranium
isotopic disequilibrium in groundwater as an indicator of
anomalies. Int. J. Appl. Radiat. Isot. 34, 283–308.
Osterkamp T. E., Jorgenson M. T., Schuur E. A. G., Shur Y. L.,
Kanevskiy M. Z., Vogel J. G. and Tumskoy V. E. (2009)
Physical and ecological changes associated with warming
permafrost and thermokarst in interior Alaska. Permafrost
Periglac. Process. 20, 235–256.
Paces J. B., Ludwig K. R., Peterman Z. E. and Neymark L. A.
(2002) U-234/U-238 evidence for local recharge and patterns of
ground-water flow in the vicinity of Yucca Mountain, Nevada,
USA. Appl. Geochem. 17, 751–779.
Pelt E., Chabaux F., Innocent C., Navarre-Sitchler A. K., Sak P. B.
and Brantley S. L. (2008) Uranium–thorium chronometry of
weathering rinds: rock alteration rate and paleo-isotopic record
of weathering fluids. Earth Planet. Sci. Lett. 276, 98–105.
Pelt E., Chabaux F., Stille P., Innocent C., Ghaleb B., Ge´rard M.
and Guntzer F. (2013) Atmospheric dust contribution to the
budget of U-series nuclides in soils from the Mount Cameroon
volcano. Chem. Geol. 341, 147–157.
Pewe T. L. (1975) Quaternary geology of Alaska. In: USGS
Professional Paper (ed. Survey, U.S.G.). Washington, DC. p.
145.
Ping C.-L., Michaelson G. J., Jorgenson M. T., Kimble J. M.,
Epstein H., Romanovsky V. E. and Walker D. A. (2008) High
stocks of soil organic carbon in the North American Arctic
region. Nat. Geosci. 1, 615–619.
Pogge von Strandmann P. A. E., Burton K. W., Porcelli D., James
R. H., van Calsteren P. and Gislason S. R. (2011) Transport
and exchange of U-series nuclides between suspended material,
dissolved load and colloids in rivers draining basaltic terrains.
Earth Planet. Sci. Lett. 301, 125–136.
Porcelli D. and Swarzenski P. W. (2003) The behavior of U- and
Th-series nuclides in groundwater. Uranium Series Geochem. 52,
317–361.
R Core Team (2012). R: A language and environment for statistical
computing. R Foundation for Stastical Computing, Vienna,
Austrailia.
Raymond P. A. and Bauer J. E. (2001) DOC cycling in a temperate
estuary: a mass balance approach using natural C-14 and C-13
isotopes. Limnol. Oceanogr. 46, 655–667.
Raymond P. A., Bauer J. E., Caraco N. F., Cole J. J., Longworth
B. and Petsch S. T. (2004) Controls on the variability of organic
matter and dissolved inorganic carbon ages in northeast US
rivers. Mar. Chem. 92, 353–366.
Reimer P. J., Baillie M. G. L., Bard E., Bayliss A., Beck J. W.,
Blackwell P. G., Ramsey C. B., Buck C. E., Burr G. S.,
Edwards R. L., Friedrich M., Grootes P. M., Guilderson T. P.,
Hajdas I., Heaton T. J., Hogg A. G., Hughen K. A., Kaiser K.
F., Kromer B., McCormac F. G., Manning S. W., Reimer R.
W., Richards D. A., Southon J. R., Talamo S., Turney C. S.
M., van der Plicht J. and Weyhenmeye C. E. (2009) INTCAL09
and MARINE09 radiocarbon age calibration curves, 0–
50,000 years cal bp. Radiocarbon 51, 1111–1150.
Reyes A. V., Froese D. G. and Jensen B. J. L. (2010a) Permafrost
response to last interglacial warming field evidence from non-
glaciated Yukon and Alaska. Quat. Sci. Rev. 29, 3256–3274.
Reyes A. V., Jensen B. J. L., Zazula G. D., Ager T. A., Kuzmina S.,
La Farge C. and Froese D. G. (2010b) A late-Middle
Pleistocene (Marine Isotope Stage 6) vegetated surface buriedby Old Crow tephra at the Palisades, interior Alaska. Quat. Sci.
Rev. 29, 801–811.
Reyes A. V., Zazula G. D., Kuzmina S., Ager T. A. and Froese D.
G. (2011) Identification of last interglacial deposits in eastern
Beringia: a cautionary note from the Palisades, interior Alaska.
J. Quat. Sci. 26, 345–352.
Robinson L. F., Henderson G. M., Hall L. and Matthews I. (2004)
Climatic control of riverine and Seawater uranium-isotope
ratios. Science 305, 851–854.
Romanovsky V. E. and Osterkamp T. E. (2000) Effects of unfrozen
water on heat and mass transport processes in the active layer
and permafrost. Permafrost Periglac. Process. 11, 219–239.
Romanovsky V. E., Smith S. L. and Christiansen H. H. (2010)
Permafrost thermal state in the polar northern hemisphere
during the international polar year 2007–2009: a synthesis.
Permafrost Periglac. Process. 21, 106–116.
Schirrmeister L., Siegert C., Kuznetsova T., Kuzmina S., Andreev
A., Kienast F., Meyer H. and Bobrov A. (2002) Paleoenviron-
mental and paleoclimatic records from permafrost deposits in
the Arctic region of Northern Siberia. Quat. Int. 89,
97–118.
Schuur E. A. G., Bockheim J., Canadell J. G., Euskirchen E., Field
C. B., Goryachkin S. V., Hagemann S., Kuhry P., Lafleur P.
M., Lee H., Mazhitova G., Nelson F. E., Rinke A., Roma-
novsky V. E., Shiklomanov N., Tarnocai C., Venevsky S.,
Vogel J. G. and Zimov S. A. (2008) Vulnerability of permafrost
carbon to climate change: implications for the global carbon
cycle. Bioscience 58, 701–714.
Shur Y. L. and Jorgenson M. T. (2007) Patterns of permafrost
formation and degradation in relation to climate and ecosys-
tems. Permafrost Periglac. Process. 18, 7–19.
Shur Y., Kanevskiy M., White D. M. and Connor B. (2010)
Geotechnical investigations for the Dalton Highway innovation
project as a case study of the ice-rich syngenetic permafrost.
Smalley I. J., Kumar R., Dhand K. O., Jefferson I. F. and Evans R.
D. (2005) The formation of silt material for terrestrial
sediments: particularly loess and dust. Sed. Geol. 179, 321–328.
Steiger R. H. and Jager E. (1977) Subcommission on geochronol-
ogy: convention on the use of decay constants in geo- and
cosmochronology. Earth Planet. Sci. Lett. 36, 359–362.
Suresh P. O., Dosseto a., Hesse P. P. and Handley H. K. (2013) Soil
formation rates determined from Uranium-series isotope dis-
equilibria in soil profiles from the southeastern Australian
highlands. Earth Planet. Sci. Lett. 379, 26–37.
Suresh P. O., Dosseto A., Handley H. K. and Hesse P. P. (2014)
Assessment of a sequential phase extraction procedure for
uranium-series isotope analysis of soils and sediments. Appl.
Radiat. Isot. 83(Pt A), 47–55.
Swanson D. K. (1996) Susceptibility of permafrost soils to deep
thaw after forest fires in interior Alaska, USA, and some
ecologic implications. Arct. Alp. Res. 28, 217–227.
Tarnocai C., Canadell J. G., Schuur E. A. G., Kuhry P., Mazhitova
G. and Zimov S. (2009) Soil organic carbon pools in the
northern circumpolar permafrost region. Global Biogeochem.
Cycles, 23.
Tyutyunov I. A. (1956) Physical and chemical processes in frozen
soils. In Fundamentals of Geocryology (Permafrost Science)
Part 1 General Geocryology (eds. P. F. Shvetsov and B. A.
Dostovalov). USSR Academy of Sciences, Moscow, pp. 115–
121.
Vaks A., Gutareva O. S., Breitenbach S. F. M., Avirmed E., Mason
A. J., Thomas A. L., Osinzev A. V., Kononov A. M. and
Henderson G. M. (2013) Speleothems reveal 500,000-year
history of Siberian permafrost. Science 340, 183–186. Available
at: http://www.ncbi.nlm.nih.gov/pubmed/23429705 [Accessed
December 17, 2014].
Viereck L. A., Dyrness C. T., Batten A. R. and Wenzlick K.J.
(1992) The Alaska vegetation classification.
Vigier N., Bourdon B., Turner S. and Allegre C. J. (2001) Erosion
timescales derived from U-decay series measurements in rivers.
Earth Planet. Sci. Lett. 193, 549–563.
Vigier N., Burton K. W., Gislason S. R., Rogers N. W., Duchene
S., Thomas L., Hodge E. and Schaefer B. (2006) The relation-
ship between riverine U-series disequilibria and erosion rates in
a basaltic terrain. Earth Planet. Sci. Lett. 249, 258–273.
Wagai R., Mayer L. M. and Kitayama K. (2009) Extent and nature
of organic coverage of soil mineral surfaces assessed by a gas
sorption approach. Geoderma 149, 152–160.
Weintraub M. N. and Schimel J. P. (2003) Interactions between
carbon and nitrogen mineralization and soil organic matter
chemistry in arctic tundra soils. Ecosystems 6, 129–143.
Weintraub M. N. and Schimel J. P. (2005) The seasonal dynamics
of amino acids and other nutrients in Alaskan Arctic tundra
soils. Biogeochemistry 73, 359–380.
Wright J. S. (1995) Glacial comminution of quartz sand grains and
the production of loessic silt: a simulation study. Quat. Sci. Rev.
14, 669–680.
Wright J. (2001) Making loess-sized quartz silt: data from
laboratory simulations and implications for sediment transport
pathways and the formation of ‘desert’ loess deposits associated
with the Sahara. Quat. Int. 76–7, 7–19.
Wu T. H. (1984) Soil movements on permafrost slopes near
Fairbanks, Alaska. Can. Geotech. J. 21, 699–709.
Zimov S. A., Schuur E. A. G. and Chapin F. S. (2006) Permafrost
and the global carbon budget. Science 312, 1612–1613.
Zimov N. S., Zimov S. A., Zimova A. E., Zimova G. M.,
Chuprynin V. I. and Chapin, III, F. S. (2009) Carbon storage in
permafrost and soils of the mammoth tundra-steppe biome:
role in the global carbon budget. Geophys. Res. Lett., 36
Associate editor: Miryam Bar-Matthews
S.A. Ewing et al. /Geochimica et Cosmochimica Acta 152 (2015) 143–165 165