Novel Application of 210Po-210Pb Disequilibria to Date Snow, Melt Pond, Ice Core, and Ice-Rafted Sediments in the Arctic Ocean

We collected surface ocean water, snow, grab ice, ice core, melt pond and ice-rafted sediment (IRS) from 5 ice stations during the Western Arctic US GEOTRACES cruise (USGCG Healy; August 10 – October 7, 2015) and analyzed for 210Po (T1/2 = 138.4 days) and 210Pb (T1/2 = 22.3 years) in dissolved and particulate phases (snow, grab ice, ice core, surface seawater) to investigate the 210Po:210Pb disequilibria in these matrices. Thirteen aerosol samples, using a large-volume aerosol sampler (PM10), from Dutch Harbor, AK to North Pole, were also collected and analyzed for 210Po/210Pb to quantify the atmospheric depositional input to the snow and surface waters. Falling snowfall is tagged with 210Po/210Pb ratio (AR) similar to that in the air column from the cloud condensation height to air-sea interface. From the measured AR in aerosol and snow, modeling the sources of 210Po and 210Pb input to the melt pond, and measured disequilibrium in ice core and ice-rafted sediment, we show 210Po/210Pb AR is a novel chronometer to date snow, ice core, melt pond, and IRS. The calculated mean ages of aerosol, snow, melt pond and IRS are 12 ± 7 (n = 13), 13 ± 11 (n = 6), 60 ± 14 (n = 4), and 87 ± 23 (n = 6) days, respectively. The average IRS age corresponds to an average drift velocity of sediment-laden ice of 0.18 ± 0.06 (n = 6) m s–1. We report highly elevated levels of 210Po and 210Pb in snow and melt pond compared to those in Arctic surface seawater and enrichment of 210Po compared to 210Pb onto particles extracted from snow, ice and melt ponds. The observed disequilibrium between 210Po and 210Pb in ice could serve as a quantitative tool in delineating multiple-year ice from seasonal ice as well as a metric in quantifying the speed of ice/snow melting and delay in autumn freeze.


INTRODUCTION
The Arctic is undergoing drastic environmental change which has manifested in decrease in the areal extent of sea ice cover, from 6.95 million km 2 in 1980 to 3.95 million km 2 in 2015, ∼45% decrease which is attributed to increase in sea surface as well as surface air temperature. Earlier melt of sea ice and later freeze in the Arctic shelves lead to a longer open-water season which has impacted the biogeochemical cycling of key trace elements and isotopes due to wind-driven vertical mixing Rutgers van der Loeff et al., 2018;Grenier et al., 2019). In addition, early retreat of sea ice edge is expected to result in higher wave action which in turn, expected to affect the amount of energy transferred from wind to surface water. Ice-free shelf waters are anticipated to result in higher wind-driven upwelling which in turn is expected to increase the amount of deeper waters onto the shelf (Carmack and Chapman, 2003). Thus, the residence time of snow, ice, and melt ponds is of great interest and has bearing on the changes in the biogeochemical cycling. The age of snow, melt pond, ice and ice-rafted sediment (IRS) in the Arctic Ocean has direct relevance to the total heat energy absorbed by surface water due to differences in their albedo and heat transfer during phase change (e.g., heat absorbed by snow/ice to become liquid water) (Uttal et al., 2002).
Polonium-210 ( 210 Po, T 1/2 = 138.4 days) and lead-210 ( 210 Pb, T 1/2 = 22.3 years), progeny of radon-222 [ 222 Rn, half-life (T 1/2 ) = 3.82 days], the heaviest and longest-lived noble gas in the U-Th series, are particle-reactive and thus have been utilized as tracers and chronometers in environmental studies (Robbins, 1978;Turekian et al., 1977;Cochran and Masque, 2003;Rutgers van der Loeff and Geibert, 2008;Baskaran, 2016). A major fraction of 210 Pb in surface ocean waters is derived from direct atmospheric deposition, which is derived from the decay of atmospheric Radon (e.g., Bacon et al., 1976). A small fraction of 222 Rn, daughter of 226 Ra, produced in rocks and mineral grains on the Earth's upper crust escapes through cracks and crevices to the atmosphere. From there, it embarks on its journey in the atmosphere via advection and diffusion and produces 12 different isotopes, including 210 Po and 210 Pb (Turekian et al., 1977). The radon emanation rates from continents decrease with increasing latitude by a factor of 5, from 1 atom cm −2 s −1 at 30 • N to 0.2 atom cm −2 s −1 at 70 • N (Conan and Robertson, 2002;Baskaran, 2011). In addition, the average 222 Rn emanation rate from the continental area (including land area covered by glaciers and permafrost with negligible 222 Rn release) of 0.75 atom cm −2 s −1 is ∼1-2 orders of magnitude higher compared to surface waters (rivers, lakes, and ocean) (e.g., Wilkening and Clements, 1975). In the absence of 210 Po from sources other than the radioactive decay of atmospheric 222 Rn-derived 210 Pb (such as volcanic, industrial release, etc.), the 210 Po/ 210 Pb activity ratio (AR) in aerosols has been utilized to determine the 'age' of aerosols (Moore et al., 1973;Robbins, 1978;Marley et al., 2000) and is generally reported to be <0.1 in the lower and middle troposphere (Turekian et al., 1977;Baskaran, 2011).
The upper end of the dating range of a radioactive daughterparent pair, where the parent half-life is much longer than the daughter half-life, is typically about 5-6 half-lives of the daughter isotope and hence 210 Po: 210 Pb disequilibrium is useful from a few days up to 2 years. From simultaneous measurements of 210 Po and 210 Pb in aerosols and precipitation, it was shown that the initial 210 Po/ 210 Pb AR was similar in rain/snow and aerosols (McNeary and Baskaran, 2007). During precipitation in the Arctic (and elsewhere), the falling snow underneath the cloud cover is tagged with the 210 Po/ 210 Pb AR of aerosols generally with values of <0.1 (Baskaran, 2011); as time elapses, this ratio in snow increases due to 210 Po ingrowth from the decay of 210 Pb. With time, the accumulated snow becomes ice or undergoes melting. During the beginning of the melt season, generally in June, meltwater from snow and ice, mostly freshwater, begins to form a pond at the surface and grow, both in areal extent and depth (Polashenski et al., 2012). From the estimated initial and measured 210 Po/ 210 Pb AR of the melt pond, its age can be determined. During storm events in early autumn, atmospherically delivered 210 Po and 210 Pb, with 210 Po/ 210 Pb AR of <0.1, are scavenged by resuspended sediments and eventually incorporated into coastal sea ice which are subsequently transported to the interior Arctic (Nürnberg et al., 1994). The in-growth of 210 Po from the decay of 210 Pb provides the time elapsed since incorporation of these radionuclides. Here, we report a novel application of disequilibrium between 210 Po and 210 Pb in dating snow, melt pond, ice core and IRS collected from the Arctic Ocean.
The primary goal of the present study, is to test the following hypotheses: (i) The in-growth of 210 Po from the decay of 210 Pb in snowfall provides a tool to determine the age of snow; (ii) From quantifying the fractional input of the sources of water to the melt pond and knowing the end-member activities of 210 Po and 210 Pb, the age of the melt ponds can be determined; (iii) Since the time range for daughter-deficient 210 Po: 210 Pb dating is up to ∼ 2 years (or 210 Po/ 210 Pb ∼0.97), multiyear ice in ice-pack can be recognized; and (iv) Most of the atmospherically delivered 210 Pb, with 210 Po/ 210 Pb AR similar to that in aerosol, is scavenged by resuspended sediment which gets incorporated into sea ice in the coastal areas and thus, we can date the IRS using 210 Po/ 210 Pb pair. The ages of snow, melt pond and ice core are relevant to the exchange of latent heat with the surrounding, changes in the albedo while ages of IRS are relevant to the transport velocity of coastal sediments and sediment-laden contaminants. In addition, transport of nutrients from coastal zone to the deep Arctic through sea ice makes a coupling between coastal and deep arctic benthic community.

Sample Collection
During US GEOTRACES Arctic cruise (HLY1502; R/V Icebreaker USCGC Healy; August 10 -October 7, 2015), the following samples were collected from 6 ice stations  Figure 1): snow, melt pond (only 5 stations) and ice core; IRS, often are sporadically distributed in the ice, from two ice stations (E-6195 and -6244); and 13 aerosol samples (Figure 1), using a large-volume aerosol sampler (PM 10 ), from Dutch Harbor, AK to North Pole, where each deployment averaged ∼30 h of run time (Figure 1 and Table 1). Aliquots of GFF aerosol filters, corresponding to filtered air volume, 72 to 463 m 3 (Table 1), were provided for analysis from the aerosol group of the U.S. GEOTRACES program. Details on aerosol collection were published in Morton et al. (2013). FIGURE 1 | Sampling stations for aerosols, ice-rafted sediment and ice stations. Green arrow shows transit with no sampling; distance between successive red dots denote aerosol sampling (starting and ending of location) for that particular sample; pumps didn't work for ST-5 and no sample. Two red lettered stations (E6244 and E6195 in red triangle) denote the dirty sea ice sample collection; Blue lettered stations 31, 33, 39, 42, 43, and 46 denote ice stations (US GEOTRACES Arctic cruise, HLY1502, R/V Icebreaker USCGC Healy).
Appropriate care was taken in preventing plume discharged from the ship's chimney when sailing. Since a large number of trace metals were also analyzed in the same aerosol filters, extreme caution was exercised in sample collection. At ice stations, bulk snow deposited on sea ice were collected into precleaned plastic bags and melted onboard to a volume of 20 L and filtered. For melt pond water samples, a hole was drilled in the frozen closed melt ponds using a TM-clean corer. A battery-powered peristaltic pump with silicone tube was used to fill a carboy with unfiltered water (Kadko and Landing, 2015; depth of the melt pond not known) and was filtered onboard. In each of the 6 ice stations, 2 ice cores next to each other, from surface to all the way down to the seawater, were collected with the Kovaks corer (9 cm ID), evenly divided into 5-6 segments approximately 20-40 cm each and same segments were combined, and melted in the onboard laboratory. The melt water from snow, ice core, and melt pond water were filtered using 0.4 µm Whatman Nuclepore track-etched membrane filter in a vacuum filtration system and the material retained in the filter were considered particulate matter. All filters were retained for particulate analysis, and filtered water samples were acidified to a pH of 2 with 6M HCl and the preconcentration was done following Fe(OH) 3 method (details in Methodology). The IRS was collected by shoveling sediment-laden ice (sediment found both at the surface and inside of the ice fragments), into pre-cleaned 20 L buckets, melted, sediments were allowed to settle, and the 'supernatant' meltwater was poured off to concentrate the sample. The sediment solution was then dried in an oven at 100 • C and used for further analysis.

Development of Chemical Leaching Method for Aerosol Filter and Ice-Rafted Sediment for 210 Po and 210 Pb
The general procedure for measuring 210 Po and 210 Pb in aerosol and involves complete digestion using concentrated hydrochloric acid, nitric acid, and hydrofluoric acid, followed by drying, and then taken in dilute HCl solution for electroplating. Since the in situ 210 Po/ 210 Pb AR in aerosol is generally less than 0.1 (Baskaran, 2011), it is a prerequisite that in situ 210 Po is separated from in situ 210 Pb soon after the sample collection (note: in 30 days from sample collection to Po plating, 14% in-growth 210 Po would take place). In order to avoid usage of concentrated hydrofluoric acid on the ship, prior to the cruise we tested a series of different leaching procedures in one aerosol filter (aerosol sample collected at Wayne State University campus, Detroit, MI, United States) by both leaching and total digestion, and compared the amount of 210 Po extracted by each method performed on separate aliquots of the same filter. The aerosol sample was filtered through one 8 × 10 QFF filter at a flow rate of 1.4 m 3 /min for approximately 24 h, similar to the arctic aerosol samples collection ( Table 1). The filter was cut into 4 quarters and were weighed. Visual observation of the filter and the weights indicated uniform aerosol mass distribution. The  Table 2). For leaching, an aliquot of the filter along with the acids ( Table 2) was taken in a centrifuge tube and placed in an ultrasonic bath at 70 • C. The filtrate was then separated from filter by vacuum filtration and the process was repeated once more. The solutions were combined and processed further for 210 Po plating (Krupp, 2017). The typical time involved are: leaching (2 h total), digestion (12 h), and drying (∼6-8 h). Additional details are given in Krupp (2017). The calculated 210 Po activities are given in Table 2. We report that leaching of the filter with 6M hydrochloric acid + 8M nitric acid, followed by drying (necessary to prevent dissolution of planchet in nitric acid medium) and dilution, provided results consistent with total digestion with HF ( 210 Po = 0.329 ± 0.013 dpm 100 m −3 vs. 0.315 ± 0.013 dpm 100 m −3 for leaching and HF digestion, respectively). The results show that 6M hydrochloric acid leaching alone is not sufficient ( 210 Po = 0.250 ± 0.009 dpm 100 m −3 ), suggesting that nitric acid is necessary as an oxidizing agent to quantitatively extract 210 Po from the aerosol filter due to possible presence of 210 Poladen organic aerosol particles. These results also suggest that the complete destruction of the quartz fiber by hydrofluoric acid is not necessary to extract 210 Po from the filter media, implying that sorbed 210 Po onto lithogenic particles can be quantitatively removed by a mixture of 6M HCl + 8M HNO 3 alone. This leaching method was used in place of total digestion on the cruise in order to eliminate the need for hydrofluoric acid and concentrated nitric and hydrochloric acid, as well as reduce the time usually needed to digest the aerosol filters, which is important when considering the time-sensitivity of the 210 Po isotope and the volume of work involved onboard the ship.
In most of the enriched 210 Po and 210 Pb are sorbed onto sediment surface, we tested a set of leaching methods. Approximately 12 L of dirty lake ice using a 20 L plastic bucket was collected on March 14-15, 2015 near the mouth of the Clinton River which discharges into Lake St. Clair in southeast Michigan. After melting (melt volume = 8.7 L), settling (after 48 h), decanting, and drying the concentrated solution in an oven (48-72 h), the total amount of dried silty-clay sediment collected was 3.617 g. To ensure fine clays were collected, the decanted ice solution was subsequently allowed to settle (48 h) and then dried in the same manner, providing a total amount of clay sediment of 0.242 g. This was combined with the other aliquot of 3.617 g. The collected sediments overall were finetextured and grayish-brown, and included a small amount of macro-organic material (grass and hairs) which was removed before leaching. About 0.2 g of the homogenized sediment sample was leached for 1 h at ∼90 • C with 10 mL of 6M HCl twice and the leachate were combined. The combined solution was dried and taken in 5 mL 1 M HCl and gamma counted. To minimize time, acid consumption and waste generation onboard, results from a leaching experiment showed that 6M HCl leaching of ice-rafted sediment collected from Lake St. Clair quantitatively extracts 210 Po and 210 Pb from IRS (Krupp, 2017).

Analytical Procedure for 210 Po and 210 Pb
Polonium-210 and 210 Pb activities of snow, ice, melt pond, seawater and aerosol were measured on the same sample. IRS samples were divided into two fractions, one for gamma spectrometry to measure 210 Pb and the other for 210 Po by alpha spectrometer. Precise determination of in situ 210 Po requires that the sample be analyzed soon after the collection to minimize ingrowth of 210 Po. The dissolved phase (filtrate of melt water of snow, ice core and melt pond) was acidified with 5 ml/L of 6M HCl (Trace-Metal grade) to prevent the loss of 210 Po and 210 Pb by sorption on to the container wall. Each acidified sample was spiked with: (i) a known amount of 209 Po US-NIST Standard Reference Material as an internal yield tracer; (ii) 20 mg stable lead via a stable lead carrier of PbCl 2 in order to determine chemical efficiency of 210 Pb recovery; and (iii) iron carrier (FeCl 3 ; 5 mg Fe/L of water) to co-precipitate 210 Po and 210 Pb with Fe(OH) 3 precipitate. After 12-24 h equilibration period and periodic vigorous shaking of the sample, ammonium hydroxide was added to samples to increase the pH from 2 to 4, and then, 1 mL of 10% sodium chromate was added to samples to increase the lead yield by the co-precipitation of lead chromate. Then, the pH was increased to ∼8. This rise in pH causes the iron to flocculate to co-precipitate 210 Po and 210 Pb with Fe(OH) 3 . Details on separation of precipitate and solution, chemical processing and plating of Po are given in Niedermiller and Baskaran (2019). After Po plating, each solution was transferred to a precleaned 60 ml Nalgene bottle and stored for further onshore 210 Pb analysis. After polonium plating, any residual 210 Po (and 209 Po) must be removed from sample solution for the future measurement of in situ 210 Pb, which is measured by in-growth of 210 Po from the decay of 210 Pb. Details on the quantitative removal of 210 Po using a column separation technique, chemical efficiency determination for 210 Pb are given in Niedermiller and Baskaran (2019). Details on decay and in-growth corrections for 210 Po and 210 Pb are given in Baskaran et al. (2013), Rigaud et al. (2013), Cookbook (2014). All particulate samples (from snow, ice, melt pond samples) processing was performed using leaching methods. The particulate sample was taken in a 50-ml centrifuge tube and was leached with 10 mL 6M Omni-trace HCl and 10 mL 8M HNO 3 mixture with agitation for 1 h at 70 • C in an ultrasonic bath. Subsequently, the solution was filtered, spiked, with a known amount of 209 Po and stable Pb, and dried. To the residue, 2 mL of HCl was added and warmed it to bring the residue to solution and then diluted to 40 ml with deionized water. This solution was used for electroplating.
The 210 Po-plated discs were assayed in an Octete PC-8 input alpha spectrometer. The background subtraction in most of the samples was less than 0.1% of the net counts for both 210 Po and 210 Pb. The activities of 210 Pb in IRS samples were measured directly in a high-resolution, high-purity Ge well detector coupled to DSA multi-channel analyzer. The gammaray spectrometer was calibrated with RGU-1, RGU-Th, both IAEA Certified Reference Materials, periodically. Details on the chemical procedures, counting methods and reagent blanks subtracted, QA/QC and intercalibration results are given below.
We analyzed Certified Reference Material (CRM) RGU-1 ( 238 U concentration 400 ± 2 ppm, with all the progeny of 238 U in secular equilibrium, including 210 Pb and 210 Po; IAEA CRM) eight different times as blind samples following the same procedure as the samples determined by alpha spectrometry to assess the accuracy. About 30 mg (weighed to a precision of ±0.1 mg) of the standard (∼9 dpm) was taken each time and about 5.5 dpm of 209 Po spike was added. The same chemical procedure was followed as was outlined in Baskaran et al. (2013). Agreement between the measured and certified value is excellent. The ratio of measured activity to certified value of RGU-1 varied between 0.96 ± 0.01 and 1.02 ± 0.00, with a mean value of 1.00 ± 0.01.
The intercalibration exercise between Wayne State University and Louisiana State University (LSU) showed that the 210 Po and 210 Pb activities on four water samples (two from GEOTRACES station in the Arctic and two collected from Gulf of Mexico collected by Dr. Kanchan Maiti, LSU) agreed within 1 standard deviation from each other and hence our 210 Po and 210 Pb data are of high quality. The results from the intercalibration was submitted to BCO-DMO data center along with all of our 210 Po and 210 Pb data obtained during 2015 GEOTRACES cruise.
The calculated depositional fluxes critically depend on the assumed V d value of 1 cm s −1 . An independent validation for V d in the Arctic can be obtained from the published F Pb values. From the summarized sedimentary inventory of F Pb in lakes, coastal regions and glaciers/ice cores, the calculated V d ranged from 0.3 to 1.6 cm s −1 (mean: 0.9 ± 0.4 cm s −1 (n = 13)), indicating V d for Arctic is similar to subarctic (Turekian et al., 1977;McNeary and Baskaran, 2003).
The residence time of aerosols (τ Po−Pb ) has been calculated from the measured 210 Po-210 Pb disequilibrium using the equation (Moore et al., 1973).  (Moore et al., 1973;Baskaran and Shaw, 2001;Baskaran, 2011). If there are additional contributions of 210 Po from sources such as volcanic eruption, transport from urban setting, and other anthropogenic sources, the estimated residence time will be an upper estimate (Kim et al., 2000(Kim et al., , 2005Su and Huh, 2002;summarized in Baskaran, 2011).
The observed higher fraction of 210 Po p in snow (and melt pond and ice cores) is intriguing as it may be due to production of higher amounts of gelatinous exopolymeric substances (EPS) by microorganisms underneath the snow which serve as agents for increased primary productivity (Krembs et al., 2011). Laboratory experiments have shown that Po strongly adsorbed on to biogenic organic matter such as chitin compared to lithogenic elements such as Pb and, with fractionation factor (=K d Po /K d Pb ratio; K d , partition coefficient = (particulate activity/dissolved activity) * (1/particulate concentration, µg/L)) of ∼4 compared to 0.2 to 1.0 for commonly occurring clay minerals (Fowler, 2011;Yang et al., 2013;Wang et al., 2019). Since no measurements of POC, macronutrients or pigments were made on collected snow or ice or melt pond samples, we cannot corroborate this hypothesis. Note that the higher fraction of 210 Po p has no bearing on the age determination, as age calculation involves only total 210 Po ( 210 Po T ) and total 210 Pb ( 210 Pb T ) (see below). The ( 210 Po/ 210 Pb) p ARs varied from 0.08 to 0.83 (mean: 0.38 ± 0.27) which are significantly higher than that in the dissolved phase (range: 0.016 to 0.128, mean: 0.062 ± 0.044; calculated from data in Figures 3A,B and Table 3).
The age of snow (and ice core, melt pond or IRS) can be obtained from eq. (3). Age of snow refers to time elapsed between the snow fall to the ground and sample collection; age of ice core is the time elapsed between the formation of ice core and sample collection; age of melt pond is the time elapsed between the formation of melt pond and the sample collection; and age of ice-rafted sediment (IRS) is the time elapsed between the incorporation of excess 210 Pb in to the sediment and the time of collection IRS. The age can be obtained from eq. (3): The ( 210 Po/ 210 Pb) i is initial AR at the time of deposition, taken to be 0.039 ± 0.026, average ( 210 Po/ 210 Pb) AR of aerosols (n = 13, Table 1); λ is 210 Po decay constant (λ = 5.01 × 10 −3 day −1 ). Instead of taking average of 13 samples, taking a subset (station 4, 6-10), the initial ratio becomes 0.031 ± 0.029 and it made very little difference on the age of snow. Since there is large spatial and temporal variability on the activities of 210 Po and 210 Pb in aerosols, we chose to use the average of all 13 samples.
The age of snow obtained from eq. (3) varied between 1.7 to 34 days (mean: 13 ± 11 days, Figure 4 and Tables 1, 3). If the analyzed snow is from multiple snow events at different times, then this is a composite age. The snowfall in the Arctic is highly patchy and no weather records are available for individual ice stations. The low-end age likely represents recent snow deposition. It is commonly assumed that once 210 Pb and 210 Po are delivered from the Arctic atmosphere to the air-sea interface through snowfall, they remain as a closed system and the changes in their activities are caused only by their radioactive decay as well as in-growth of 210 Po from the decay of 210 Pb until sample collection. This study shows generally that the age of snow is less than a month, and thus, snow deposited more than a month ago likely has undergone changes to become ice with associated exchange of heat energy with the surroundings. The implication of this study is that the changes in the ages of snow and ice due to global climate change (e.g., earlier melt and later freeze of ice) will affect the latent heat exchanged with the surroundings.
Age of Ice-Rafted Sediment Using 210 Po xs / 210 Pb xs Activity Ratio Sea ice in the Arctic plays an important role in the biogeochemical cycling of key trace metals, transport and subsequent dispersal of coastal sediments and nutrients to the deep Arctic as well as impacting radiation balance. Three proposed mechanisms for sediment entrainment into sea ice include: (i) incorporation of fine suspended sedimentary particles into frazil-ice crystals in shallow coastal areas; (ii) uplift of sediments by anchor ice; and (iii) discharge of river-borne ice-laden sediments into the sea (Reimnitz et al., 1987;Hebbeln and Weber, 1991;Nürnberg et al., 1994;Eicken et al., 1997). The concentration and composition of IRS, incorporated in coastal sea ice which are subsequently transported to the deep Arctic by the Beaufort Gyre and Transpolar Drift, has direct bearing on the albedo, long-range redistribution of contaminants, and particle flux to the deep sea (Hebbeln and Weber, 1991;Nürnberg et al., 1994;Pfirman et al., 1995;Eicken et al., 1997;Landa et al., 1997;Meese et al., 1997;Cooper et al., 1998;Charette et al., 2020).
The excess 210 Pb xs activities ( 210 Pb xs = 210 Pb total -226 Ra; average measured 226 Ra: 0.95 ± 0.18) in IRS, range between 19 and 186 dpm g −1 (mean: 96 ± 76, Table 4), similar to earlier published data and is up to 2 orders of magnitude higher than the average total 210 Pb activity, 1.90 dpm g −1 (n = 134) of Russian Arctic surface sediments (Roberts et al., 1997;Baskaran, 2005). The 210 Po xs ( 210 Po xs = 210 Po total -226 Ra) ranged from 5 to 76 dpm g −1 (mean: 40 ± 33, Table 4). This is the first 210 Po published data in IRS showing high 210 Po enrichment as well disequilibrium between 210 Po and 210 Pb. Note that enrichment of 210 Po onto particulate matter will not affect the total 210 Po/total 210 Pb activity ratio (total 210 Po = particulate 210 Po + dissolved 210 Po) of the ages calculated based on 210 Po/ 210 Pb activity ratio. The 210 Po xs / 210 Pb xs AR varied from 0.27 to 0.50 (mean: 0.38 ± 0.09, n = 6) and the differences in the AR and enrichment of 210 Pb are attributed to the differences in their extent of initial interaction with seawater, and the time elapsed since the incorporation of 210 Pb into IRS.
Assuming the incorporation of atmospherically delivered 210 Pb into IRS is a one-time event (e.g., 210 Pb-laden sediment incorporation in to sea ice sediment taking place in one freeze over a short period of time), and the 210 Po and 210 Pb in IRS remains a closed system, the calculated age ranged from 52 to 122 days (mean = 85 ± 27 days, n = 6, Table 4). Although there is no direct observation from the Arctic for this assumed one-time incorporation of radionuclides, a serendipitous observation from collection and analysis of sediment-laden ice from Lake St. Clair in southeast Michigan showed that the sediment extracted from ice during a snow storm during March 14-15, 2015 had an average 210 Pb xs activity of 555 ± 138 (n = 3) dpm g −1 (Krupp, 2017). This is ∼6 times higher than that in the Arctic and about 100 times that of surficial sediments in Lake St. Clair (Robbins et al., 1990;Jweda and Baskaran, 2011). Note that the average atmospheric depositional flux of 210 Pb in 60-80 • N belt of 0.18 dpm cm −2 year −1 is only 13% as that of Detroit, MI (1.41 dpm cm −2 year −1 ), a mid-latitude site, and such larges differences are due to low emanation rate of 222 Rn as well as low annual amount of precipitation in the Arctic (McNeary and Baskaran, 2003;Baskaran, 2011). The corresponding transport velocities (transport velocity = shortest distance between sampling site and the nearest coastal site, taking in to consideration the currents in the sampling area and back tracked the trajectory of the ice since its formation/age) varied between 0.12 ± 0.01 and 0.27 ± 0.02 m s −1 (mean: 0.18 ± 0.06 m −1 , Table 4). These velocities can be compared to 0.082-0.086 m s −1 obtained using buoys in Laptev Sea and 0.09 ± 0.04 m s −1 (range: 0.04-0.18 m s −1 , n = 23, monthly, April 1996 -February 1998) over Eastern Beaufort Sea using Acoustic Doppler Current Profiler (Melling and Riedel, 2003). Note that the calculated velocities are based on 210 Po/ 210 Pb initial AR value of 0.039 (Table 1) and if AR is higher due to significant contribution of 210 Po from resuspension of superfine sedimentary material with initial AR > 0.039, then, the age will be lower and the transport velocities will be higher.

Dating of Melt Pond Using 210 Po/ 210 Pb AR
During late spring and early summer, melt from snow (meteoric water) and surface ice (ice core salinity range: 0-5.2 ppt, surface to 169 cm, Supplementary Figure 1) mixing with a fraction of surface seawater result in the formation of melt ponds. The melt ponds are generally darker compared to ice leading to more absorption of incident radiation (lower albedo) compared to snow and ice. Furthermore, the vertical and horizontal fluxes of melt water from snow and ice play a key role in the evolution of albedo, heat transfer, and mass balance of the Arctic ice pack (Eicken, 1994;Eicken et al., 2002;Perovich et al., 2003). Surface area and depth of the melt ponds were not recorded during the field expedition; however, an average surface area of 30 m 2 and depth of 0.2 m have been reported (Eicken et al., 2002). The fractions of these three contributing sources to these melt ponds were estimated using a three end-member mixing model using salinity, δ 18 O and 7 Be data (Marsay et al., 2018; Table 5).
The activities of 210 Po p range from 0.4 to 4.0 (mean: 1.7 ± 0.2 dpm 100L −1 , n = 5) while the corresponding 210 Pb p range from 0.3 to 5.6 (mean: 2.6 ± 0.4 dpm 100L −1 , n = 5) which are comparable to values in Arctic surface waters (Moore and Smith, 1986;Smith et al., 2003;Roca-Marti et al., 2016;Bam et al., 2020). However, the 210 Po T range from 11 to 61 (mean: 28 ± 3, n = 5) dpm 100L −1 and 210 Pb T range from 25 to 205 (mean: 94 ± 8, n = 5) dpm 100L −1 , which are about 5 to 10 times higher than those found in surface seawater from the same or nearby stations (Figures 2A,B and Table 5). Such differences between particulate and total activities are attributed to differences in the sources of water to the melt ponds, with differences in the activities between the three source waters (Figures 2A,B and Table 5). The activities of 210 Po T and 210 Pb T at the time of melt pond formation were calculated using three end-member (snow, ice, and surface seawater) mixing model, as given in (eq. 4). The initial ( 210 Po T / 210 Pb T ) AR varied between 0.071 and 0.143 (n = 4); however, in ST-39, the ratio is >1.0.
and likely the assumptions are not well constrained; taking the initial AR to be the same as that of the aerosols (Table 1), we get an age of 75 ± 9 days, although no other rationale can be given for using this ratio. Assuming the melt pond was formed in a relatively short timescale compared to the age of the pond, its age calculated using eq. (3) varied between 18 ± 3 and 79 ± 6 days (mean: 60 ± 14 days, n = 4, Table 5). If this assumption is not strictly valid, then, the calculated age reported in Table 5 is an overestimate. The ages (Figure 4 and Table 5) indicate that ST-33 pond formed in late spring while 43 in early summer and ST-46 in mid-summer.
If the particulate matter present in some of the melt ponds had undergone multiple melt-freeze cycles spanning over more than a year, then, we expect secular equilibrium between 210 Po p and 210 Pb p with 210 Po p / 210 Pb p AR ∼ 1.0, under a closed system for 210 Po and 210 Pb. In two samples where ( 210 Po/ 210 Pb) p AR is >1.0, the particulate matter could have undergone multiple melt-freeze cycles although we have no 5 | *Activities of 210 Po and 210 Pb (dpm/100 L), fractional amounts snow (F snow ), ice (F ice ), and seawater (F sw ), and age** of melt ponds.  Krupp (2017). **Ages for ST-33, 42, 43 and 46 were calculated from calculated initial and measured. ( 210 Po/ 210 Pb) T AR. In the calculation for ST-39, ( 210 Po/ 210 Pb) AR was assumed to the same as in aerosols (0.039 ± 0.026) and yields an age of 75 ± 9 days, which is a lower estimate (see in the text). Fractional values of F snow , F sw and F ice are taken from Marsay et al. (2018). Calculated 210 Po (or 210 Pb) is the activity at the time of formation of melt pond = (F snow * 210 Po T−snow ) + (F sw 210 Po T−sw ) + (F ice * 210 Po T−ice ). + calculated AR value for ST 33, 42, 43 and 46 is assumed to be the initial AR at the time of formation of melt pond using eq. (4). For 210 Po/ 210 Pb AR is 1.15 ± 0.04 and is not well constrained by the 3-end-member mixing model. other supporting evidence ( Figure 3A and Table 5). Note that average particulate fraction of 210 Po is only 8.8% (range: 0.7-16.7%), and even if some of the particulate matter is derived from multiple melt-freeze cycles that will not affect the ( 210 Po/ 210 Pb) T AR-based-ages. However, in 3 samples ( 210 Po/ 210 Pb) p AR is < 1.0 and ( 210 Po/ 210 Pb) T AR is <1.0 in all 5 samples ( Table 5).

Activity of 210 Po and 210 Pb, Their AR and Dating of Ice Core
The 210 Po p,d and 210 Pb p,d activities in the upper segment of 6 ice cores varied by more than an order of magnitude, from 4.4 to 75 dpm 100L −1 for 210 Po p and 11 to 120 dpm 100 L −1 for 210 Pb p and 11 to 96 dpm 100L −1 for 210 Po d and 0.8 to 174 dpm 100 L −1 for 210 Pb d . The activities of particulate, dissolved and total phases varied over two orders of magnitude in all 31 split samples ( Table 6), similar to an earlier study (Masqué et al., 2007). Although there is no excess 210 Po (i.e., 210 Po/ 210 Pb AR < 1.0) in the inventory-based particulate or dissolved phases (Table 7), excess 210 Po d (and 210 Po T ) were observed in three discrete layers: , , and ST-39 (0-38 cm); the ( 210 Po/ 210 Pb) T AR in those layers ranged between 2.88 and 4.58 (Table 6), indicating preferential sorption of 210 Po and/or preferential loss of 210 Pb. This contrasts with observed 210 Po-210 Pb equilibrium in 87%, (within ± 1σ) of the 38 segments from two ice cores from Fram Strait, although reported associated errors in majority of the samples were high in Masqué et al. (2007).
If the accumulation and ablation of ice takes place uniformly, with accretion from the top and ablation from the bottom, the ( 210 Po/ 210 Pb) T AR profile is expected to increase with increasing depth; however, when multiple processes such as meltwater deformation, seawater congelation, and false bottom formation takes place, the measured ( 210 Po/ 210 Pb) T AR will result in a more complicated profile. If a multi-year (≥2 year) ice is present, the ( 210 Po/ 210 Pb) T AR is expected to be ∼1.0 under the assumption that the 210 Po-210 Pb has remained a closed system. We contend that ice that are >2 year old should have an 210 Po p / 210 Pb p AR of ∼1.0, because any disequilibrium broken earlier in the decay chain will adjust to radioactive secular equilibrium in about two years, and thus paving the way to delineate >1 year old ice from more recent ones. The 210 Po/ 210 Pb AR of 0.95 ± 0.07 in only one sample (ST-30, 60-80 cm, Table 6) could be a part of multi-year ice segment; thus, we report the ice samples we collected are mostly first-year ice, with 210 Po/ 210 Pb ARs ranging between 0.12 ± 0.01 and 0.78 ± 0.07 (n = 25, Table 6).
The inventory-based (= activities in each segment of the ice core, dpm/ volume of melt water, L) activities of 210 Po p, 210 Pb p , 210 Po d and 210 Pb d varied widely: 210 Po p, varied about an order of magnitude, from 1.8 to 17.4 (mean: 8.1 ± 5.8, n = 6) dpm 100L −1 ; 210 Po T varied within a factor of ∼3, from 7.9 to 26.2 (mean: 18.8 ± 6.4, n = 6, calculated from Table 7) dpm 100L −1 . The range of 210 Pb p activities is similar to 210 Po p , from 3.8 to 28.5 (mean: 14.3 ± 9.2, n = 6) dpm 100L −1 , while the corresponding values for 210 Pb T varied within a factor of ∼3, from 22 to 69 (mean: 49 ± 18, n = 6, dpm 100L −1 ). These values are almost an order of magnitude higher than those for the Arctic surface waters (Figures 2A,B and  Tables 6, 7), suggesting a significant fraction of the source of water is from snow melt and possibly some recycled component of melt ponds. During spring and summer, melt pond water can percolate through the depth of the ice floe forming a lens of super-cooled freshwater at the interface between the ice floe and surface seawater, which ultimately grows into a horizontal ice sheet known as false-bottom ice (Eicken, 1994;Eicken et al., 1997;Polashenski et al., 2012) and thus the signature of the melt pond could have been be imprinted on bottom ice. It seems this is a novel observation presenting a radioisotopebased evidence that the ice core derived some component of water from snow. 123 ± 8 72 ± 3 66 ± 3 109 ± 7 77 ± 3 * Activities of individual layers are given in Table 6. The particulate and total activities were calculated from the summation of the corresponding activity in each layer and divided by total melt volume of water for each core. * * In the calculation of the age, the average of 13 aerosol initial ( 210 Po/ 210 Pb) i value of 0.039 ± 0.026 was used.
The ( 210 Po/ 210 Pb) p AR for the whole ice core varied between 0.36 and 0.70 (mean: 0.56 ± 0.13, n = 6) which is significantly higher than the range of values and mean for ( 210 Po/ 210 Pb) T AR, 0.32 to 0.50 (mean: 0.39 ± 0.07, n = 6, Table 7). The higher ( 210 Po/ 210 Pb) p AR compared to ( 210 Po/ 210 Pb) d and ( 210 Po/ 210 Pb) T are similar to snow and melt pond, confirming enrichment of 210 Po in particulate matter. The calculated composite ages of the ice core, based on the inventory of 210 Po T and 210 Pb T and assuming initial ( 210 Po/ 210 Pb) AR = 0.039, ranged between 66 and 123 days (mean: 87 ± 23 days; Figure 4 and Table 7). If the initial ARs are higher, the calculated ages are upper limits.

CONCLUSION AND FUTURE OUTLOOK
We measured the disequilibrium between 210 Po and 210 Pb in a suite of aerosols and ice-rafted sediment as well in particulate and dissolved phases of snow, water from melt pond and surface seawater. We report, for the first time, highly elevated levels of 210 Po in biogenic particulate matter in snow and melt pond compared to 210 Pb indicating biogeochemical cycling of biogenic elements such as polonium is different in the Arctic. We also report one to two orders of magnitude higher 210 Pb and 210 Po activities in ice-rafted sediments (IRS) compared to benthic source sediments which indicate that the sea ice-sediment (i.e., IRS) is a powerful vector in the transport of land-and atmospherically delivered particlereactive contaminants from the coastal area to the deep Arctic. Furthermore, transport of ice-rafted sediments through sea ice is one of the major mechanisms by which coastal sediments are transported and dispersed in the deep Arctic (during melt season) and nutrients associated with coastal sediments serve as food resource for deep benthic organisms in the open Artic, thus indicating coupling between coastal and deep arctic benthic ecosystems. The key findings are given below: (i) From the measured 210 Po/ 210 Pb AR in snow and using the mean AR in aerosol as the initial AR in snow, the calculated age of snow collected from 6 different ice stations varied between 1.7 and 34 days (mean: 13 days); (ii) The ages of 5 different melt ponds, based on the measured 210 Po/ 210 Pb AR and three end-member mixing modeling, ranged between 18 and 79 days (mean: 60 days). The activities of 210 Po and 210 Pb in melt-ponds is about 10 times higher compared to the surface seawater in the Arctic Ocean; (iii) From the measured 210 Po-210 Pb activities in aerosols, collected from 13 different stations covering the entire 10 weeks of Western Arctic GEOTRACES cruise track from Dutch Harbor, AK to the North Pole, we show that the mean 210 Po/ 210 Pb activity ratio (AR) of 0.039 ± 0.026 (n = 13) corresponds to a residence time of 12 ± 7 days; (iv) The transport velocity of the ice-rafted sediment in the shelf and interior Arctic Ocean, estimated based on the age obtained using 210 Po/ 210 Pb AR, 0.12 and 0.27 m/s, (mean: 0.18 ± 0.06 m/s) agree with the data obtained using buoys and Acoustic Doppler current profiler; and (v) The 210 Po/ 210 Pb AR serves as a quantitative tool in delineating multiple-year ice from seasonal ice.
The residence time of snow and melt pond has direct bearing on the energy exchange between surface ocean and atmosphere as well as the heat exchange between Arctic and sub-arctic global oceans. The disequilibrium between 210 Po and 210 Pb provide not only a powerful tool in establishing chronology but also insight on the mechanism(s) of Po enrichment onto biogenic particulate matter. Routine identification of multi-year (those that survived more than one melt cycle) ice is now possible using 210 Po/ 210 Pb AR as a metric. Furthermore, longterm longitudinal study on the annual variations in the age of snow and melt pond in a particular region caused by climate change in the Arctic may enable us to quantify the radiation balance changes caused by faster melting of snow and later freeze. Highly enriched 210 Po in particulate matter in snow, ice and melt ponds suggest that the biological organisms play a key role in the enrichment of biogenic elements. The factors and processes that cause high enrichment of polonium in the particulate matter obtained from snow, ice and melt pond remains unknown. A systematic study on the distribution of 210 Po and 210 Pb in particulate and dissolved phases along with characterization of particulate matter including quantification of acid polysaccharides and other biogenic molecular compounds will yield valuable information to understand the mechanisms of Po enrichment onto particulate matter in the Arctic.

DATA AVAILABILITY STATEMENT
The original data for this study are presented in the article/ Supplementary Material. The data also can be obtained from https://www.bco-dmo.org/dataset/794064; and further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
MB developed the idea, secured the funding, oversaw the project from conception to the end, and wrote most of the manuscript. KK participated in the cruise, collected and analyzed the samples, and interpreted the data. Both authors contributed to the article and approved the submitted version.

FUNDING
The work presented was a part of a master's thesis of KK. This work was supported by the National Science Foundation grant (NSF-PLR-1434578).