Introduction

Salt marsh sediments are useful recorders of environmental change at the terrestrial-marine interface, allowing us to unravel the occurrence, magnitude, and associated feedbacks of anthropogenic, climatic, and tectonic processes. Studying these environmental archives, however, requires accurate and precise age-dating of sediments. 137Cs is one of the most commonly used short-lived radioisotopes for dating recent salt marsh sediments (Drexler et al., 2018; Ritchie & McHenry, 1990). 137Cs is an anthropogenic radionuclide produced during nuclear fission and is present in the environment globally due mainly to nuclear weapons testing, with detectable atmospheric fallout beginning in 1954 CE and peaking in 1963 CE (Pennington et al., 1973). 137Cs sediment records may also be impacted by local/regional events (e.g., the Chernobyl disaster in 1986 CE; Appleby et al., 2023). Identification of these event horizons in sediment stratigraphies serves as a method for age-dating the last ~ 70 years of sediment accumulation in estuaries (and other marine environments) and generating bulk estimates of accumulation rates over this time (DeLaune et al., 1978). Conveniently, 137Cs can be measured simultaneously with excess 210Pb, which is also commonly used to create age-depth models spanning the last century through non-destructive gamma spectrometry (Foucher et al., 2021).

Yet the utility of the 137Cs method has long been questioned. Issues include the following: large sand fractions result in low 137Cs activities (He & Walling, 1996); 137Cs may be particularly susceptible to post-depositional remobilization, especially in saline and anoxic environments (Foster et al., 2006; Johnson-Pyrtle & Scott, 2001; Leslie & Hancock, 2008); and, as with any event horizon,137Cs may be subject to physical mixing (e.g., due to bioturbation) that displaces the peak either upward or downward in the sediment column (Andersen et al., 2000). Considering the relatively short half-life of 137Cs (30.17 years) and the time since the peak emissions from Cold-War-era atmospheric weapons testing, modern measurements of down-core activities are now approaching the background detection limits of most passive gamma detection laboratories. Indeed, 137Cs profiles often lack clear peaks and/or display downward migration in sediment profiles (Drexler et al., 2018). Despite the dubious nature of many 137Cs profiles, the method is commonly employed and often chosen over other available methods (e.g., excess 210Pb) in contemporary studies, likely given that methods of its interpretation are generally more simplistic. These limitations motivate the need for alternative radionuclide tracers with longer half-lives and reduced mobility in saline and anoxic environments.

241Am has been proposed as a convenient alternative to 137Cs (Appleby et al., 1991). 241Am is a daughter product of 241Pu (half-life = 14.7 y) and is found in sediments as a direct consequence of nuclear fallout similar to 137Cs. One key difference between 137Cs and 241Am is that 137Cs was immediately produced during nuclear explosions while 241Am has ingrown over time from 241Pu; thus, activities of 137Cs are decreasing through time while activities of 241Am are increasing. The use of 241Am in dating is hindered, however, by activities that are still relatively low because 241Pu fallout was only ~ 17% the activity of 137Cs (Appleby et al., 1991), and the 241Am yield by gamma spectrometry is less than that of 137Cs (35.7% at 59.5 keV vs. 85.2% at 661.7 keV). However, as 137Cs quickly approaches detection limits, 241Am with its 432.2 y half-life may quickly become a preferred age-dating method. By modeling the decay of 241Pu, Boyd and Sommerfield (2017) found that activities of 241Am would peak in the mid 2030s, decaying to extinction in the late 4100s; by contrast, only 10% of 137Cs inventory will remain by 2075. 241Pu and 241Am may also be less chemically mobile in brackish/saline sediments than 137Cs (Boyd & Sommerfield, 2017; Dyer et al., 2002), and 241Am may be especially preferable in soils with K limitations in which 137Cs may be more biologically reactive (Kaste et al., 2021; Landis, 2025). Although 241Am is mostly absent in equatorial latitudes due to 10x lower atmospheric fallout (Landis, 2025), there has been success applying it in high latitude environments, the sediments of which are notoriously tricky to date (e.g., Heldal et al., 2002; Roos et al., 1994). Indeed, 241Am has been used to date or trace other aquatic sediments (e.g., Appleby et al., 1991; Beasley et al., 1982; Block et al., 2023; Carpenter & Beasley, 1981; Koide et al., 1980, 1981; Koide & Goldberg, 1982; Proskurnin et al., 2021; Robbins et al., 2000; Szarłowicz et al., 2022; Volante et al., 2024), though examples from estuarine sediments are more scarce (see exceptions Dyer et al., 2002; Rahman & Plater, 2023; Thomson et al., 2002), and those from North American salt marshes are limited to the US East Coast (Boyd & Sommerfield, 2017; McGauley, 2024; Wright et al., 2017).

We reassessed gamma energy spectra from salt marsh sediments from Oregon, U.S. to identify and characterize 241Am depth profiles. The core data were originally presented by Peck et al. (2020), who relied on excess 210Pb to calculate accretion rates rather than 137Cs profiles, which were less consistently detectable than excess 210Pb and often exhibited signs of post-depositional remobilization, where activities extended many cm below the peak indicating chemical migration. Through this reassessment we seek to expand current examples of depth profiles of 241Am, drawing attention to this radionuclide as a potential age-dating technique into the future, especially in coastal sediments. Our primary goal is to determine whether 241Am would be an appropriate alternative for 137Cs in salt marsh sediments. We also aim to assess the conditions under which 241Am may be more reliable.

Methods

Sediment Core Collection & Analysis

Peck et al. (2020) collected 72 sediment cores (10-cm diameter, 1.5- to 3-m in length) from intertidal zones (tide flat, low marsh, high marsh, and scrub-shrub wetlands) using a push coring approach between 2010 and 2018 in seven Oregon, US estuaries: Youngs Bay, Nehalem Bay, Tillamook Bay, Netarts Bay, Salmon River Estuary, Alsea Bay, and Coquille River Estuary. An additional three cores were collected in 2019 from Sand Lake Estuary. Elevation was measured at all sites, except Sand Lake, by RTK-GPS and relative elevations (z*) were calculated as

$$\:{z}^{*}\:=\:\frac{elevation-mean\:tide\:level}{mean\:higher\:high\:water-mean\:tide\:level}$$

where tidal datums were determined from the closest NOAA tide gauge (Table S1).

Cores were scanned by x-ray computed tomography (CT) to visualize stratigraphy and calculate dry bulk densities with depth (sensu Davey et al., 2011; see Peck et al., 2020 and the supplement for more detail). Cores were then split lengthwise, and half of the core was sectioned at 2-cm increments, freeze-dried, and disaggregated using a mortar and pestle. Sediment organic and mineral matter contents were measured by loss on ignition (Heiri et al., 2001).

All radionuclides were measured via gamma spectrometry using two Canberra GL2020RS LEGe planar detectors. Samples weighed 24 g on average (range = 3.1–71 g) and were packed into polypropylene counting jars (70-mm diameter, 32-mm height). Sediments were counted for ≥ 24 h. Excess 210Pb was calculated as the difference in total 210Pb measured at 46.5 keV (yield = 4.24%) and 214Pb measured at 351.9 keV (yield = 35.8%). 241Am and 137Cs were detected at 59.5 keV (yield = 36.3%) and 661.7 keV (yield = 85.1%), respectively. Samples were counted to a depth at which excess 210Pb approached background activities. Gamma detectors were energy calibrated at least once per season using a mixed source material. All activities were blank and geometry corrected, as well as decay corrected to the date of core collection.

Accretion Rates & Age-depth Models

Constant initial concentration (CIC) age-depth models and accretion rates were originally computed in Peck et al. (2020) on sediment cores that displayed a log-linear decay of excess 210Pb with a high coefficient of determination (R2 > 0.75). In this study, we used the constant rate of supply (CRS) model when interpreting excess 210Pb as it is well suited to estimating depth/time-varying rates, and as such the timeframe of interest can be adjusted for comparison to that of 241Am and 137Cs (see Abril-Hernández, 2025 for a review of other models for interpreting excess 210Pb).

The CRS model assumes a constant flux of excess (unsupported) 210Pb to the marsh surface; age is calculated as (Arias-Ortiz et al., 2018):

$$\:{I}_{i}=I\times\:{e}^{-\lambda\:t}$$

where Ii and I are the excess 210Pb inventories (Bq m− 2) below depth i and over the entire core, respectively, λ is the excess 210Pb decay constant (0.03114 y− 1), and t is the elapsed time since the accumulation of section i. Inventories were calculated as the cumulative sum of the entire excess 210Pb activity profile (Ai; Bq kg− 1) with depth multiplied by the thickness (hi) and dry bulk density (ρi) of the depth intervals (sensu Cochran et al., 1998):

$$I = \sum\limits_{i} (\rho_{i} h_{i} A_{i})$$

CRS model ages are impacted by the “old-date error” (described by Binford (1990), whereby excess 210Pb inventories are underestimated because, though they exponentially decay with depth never reaching 0 Bq kg− 1, an instrumental detection limit makes detection beyond a certain depth impossible. Although the excluded activity is small, it results in high age inaccuracies at the deepest depths where the underestimated activity represents a large fraction of Ix. Inventories used in the CRS age-depth model calculations were corrected for the old-date error by calculating a logarithmic decay fit, estimating excess 210Pb activities to a depth twice that of the deepest measured activity, and adding the summed modeled activity to the inventories at all measured depths. Dates were calculated only to the depth at which the deepest excess 210Pb activity was detectable. Errors (st2) associated with the CRS age-depth model were calculated using the method described by Binford (1990):

$$\:{s}_{t}^{2}={\lambda\:}^{-2}{I}_{}^{-2}{s}_{{I}_{}}^{2}+{\lambda\:}^{-2}{I}_{i}^{-2}{s}_{{I}_{i}}^{2}$$

Where \(\:{s}_{{I}_{}}^{2}\) and \(\:{s}_{{I}_{i}}^{2}\) are the variances of the total excess 210Pb profile and the excess 210Pb at depth i, respectively.

Age-depth models, as well as bulk accretion rates, were calculated by interpretation of the 137Cs and 241Am activity profiles where the surface of the sediment core is assigned the year of collection, the peak in the radionuclide is assigned the year 1963 CE when atmospheric fallout peaked, and the deepest depth of the radionuclide is assigned the year 1954 CE when atmospheric fallout first became detectable. When activity uncertainties in the 137Cs and 241Am activities resulted in insensitivity between depths around the peak, a middle depth was assigned to calculate ages and accretion rates. Errors associated with the bomb produced ages and accretion rates are based on the 2-cm sampling thickness (or greater when peak activities spanned multiple depth cohorts). Total inventories of both bomb-produced radionuclides were calculated, as well.

Bulk excess 210Pb accretion rates were calculated on the CRS age-depth models above and below the 137Cs and 241Am peaks for comparison of rates on temporally equivalent timeframes. Mineral accumulation rates were also calculated over the CRS ages vs. cumulative mineral mass depth (kg of mineral matter m− 2 y− 1) – calculated using mineral matter masses based on loss on ignition – for the whole sediment core as well as above and below the 137Cs and 241Am peaks.

Fallout Inventories

Measured inventories of all three radionuclides were compared to atmospheric fallout inventories. Excess 210Pb inventories were determined for each estuary using a relationship determined by Sedighi et al. (2020) for undisturbed soils based on mean annual precipitation in different latitudes and climates. Mean annual precipitation from 1990 to 2020 for each estuary was modeled using PRISM (2025). Decay corrected inventories for 137Cs and 241Am (based on in-situ decay of 241Pu) were calculated for each sediment core based on histories of 90Sr fallout at the closest Environmental Measurements Laboratory (EML 2025) stations (Forks, WA and Medford, OR) converted using previously determined ratios between 137Cs:90Sr (1.45; Bowen et al., 1974; Boyd & Sommerfield, 2017) and 241Pu:137Cs (0.177; Appleby et al., 1991; Boyd & Sommerfield, 2017) (see Supplementary Information for more details on this method).

Statistical Comparisons

All statistical calculations were performed in MATLAB 2025a. Inventories across elevation zone (tide flat, low marsh, high marsh, and scrub-shrub wetland) and from the high marshes of the eight Oregon estuaries were statistically compared using Kruskal-Wallis (KW) tests (α = 0.05). Depths of 137Cs and 241Am peaks were also compared by KW, as well as CRS accretion rates pre and post 1963 as defined by the radionuclide bomb peaks. Simple linear regression (α = 0.05) was used to assess correlation between measured inventories across z*, measured inventories and activities of the bomb radionuclides, and accretion rates. The accuracy of the 137Cs and 241Am methods were assessed through comparison to the excess 210Pb CRS age-depth models considering ranges in uncertainties associated with each method. Analysis of variance for within-subject effects in a repeated measures model (RANOVA, α = 0.05) was used to assess whether accretion and mineral accumulation rates measured by different methods were significantly different considering the effect of estuary on accumulation rate. The Youngs core was excluded from these analyses as the 241Am depth profile seemed impacted by local sources (discussed more below).

Fig. 1
Fig. 1
Full size image

Example radionuclide depth profiles (including excess 210Pb, 137Cs, and 241Am) with error bars for three high marsh sediment cores from Netarts Bay, (a-c): NT09, NT05, and NT03. Blue and orange arrows indicate the deepest depth with measurable 137Cs and 241Am activities, respectively. Note that each profile has a different range in radionuclide activities to ensure easy visualization of trends rather than relative magnitudes

Results

Minimum detectable activities for excess 210Pb, 137Cs, and 241Am were 0.0082, 0.99, and 0.065 Bq kg− 1, respectively (see Fig. S1 for ranges; Fig. 1 for example profiles). Detectability varied by isotope, with excess 210Pb measurable in 97% of cores, 137Cs in 95% and 241Am in 60% (Fig. 2). Excess 210Pb profiles were interpretable in 80% of scrub-shrub wetland, 96% of high marsh, 50% of low marsh, and 10% tide flat zones (Fig. 2). Accretion rates and age-depth models were not calculated for cores with poorly fitting log-linear decays in excess 210Pb with depth per Peck et al. (2020). 137Cs was found at activities above the detection limit in 100%, 98%, 92%, and 80% of scrub-shrub wetland, high marsh, low marsh, and tide flat environments, respectively, whereas 241Am was found most often in high marsh (81%) cores, with only 40%, 25%, and 10% of scrub-shrub wetland, low marsh, and tide flat cores having measurable activities.

Across all elevational zones and estuaries, median 241Am/137Cs inventory ratios were 0.013 ± 0.023, similar to the modeled inventory ratio of atmospheric fallout of ~ 0.012. Excess 210Pb inventories increased significantly with elevational zone, peaking in scrub-shrub wetlands (Fig. 2a; KW, df = 73, p < 0.01). High marsh excess 210Pb inventories were most similar to fallout predicted based on precipitation and latitude (Table S1S2). 137Cs inventories also tended to be highest at higher elevations, though inventory ranges overlapped more, especially between tide flat, low marsh, and high marsh sediments, which were most similar to predicted inventories from atmospheric fallout (Fig. 2c; KW, df = 73, p = 0.02). Trends with elevational zone were more difficult to discern for 241Am inventories as the radionuclide was not detectable in the majority of cores, except those from high marsh (Fig. 2d; KW, df = 73, p < 0.01).

Fig. 2
Fig. 2
Full size image

Inventories of (a-b): excess 210Pb, (c-d): 137Cs, and (e-f): 241Am grouped by intertidal zones plotted as box plots (boxes represent first and third quartile, center lines indicate medians, whiskers are minimums and maximums, and O’s are outliers; Kruskal-Wallis test results are presented above) and across relative elevations (z*) in the high marsh (black line indicates a statistically significant linear regression model, dashed black line indicates the 95% confidence bound; statistics are presented above). For each zone, red numbers indicate the percent of cores with measurable radionuclides, and dark blue numbers indicate the perfect of cores with interpretable excess 210Pb profiles. Gray patches indicate the range of modeled atmospheric fallout for each radionuclide (Table S2)

Given that the focus of the original field sampling scheme was on high marsh and other intertidal zones were cored inconsistently, we focus much of the rest of the analysis on the 48 high marsh cores, 38 of which had all three radionuclides measurable. Within the high marsh zone, inventories increased with increasing z*, though this relationship was only significant for excess 210Pb (Fig. 2b; R2 = 0.22, n = 35, p < 0.01) and 241Am (Fig. 2f; R2 = 0.34, n = 35, p < 0.01; 137Cs: R2 = 0.078, n = 35, p = 0.1). High marsh inventories of the three radionuclides were only weakly significantly different across the eight estuaries (Fig. 3; KW, df = 47, p = 0.04 for all). All high marsh cores had measurable excess 210Pb (Fig. 3b), though two cores from Nehalem were excluded from accretion rate measurements (NB10 and NB12), which displayed non-steady state accumulation. Almost all cores across estuaries had 137Cs (Fig. 3c). All high marsh cores from Salmon and Sand Lake and most high marsh cores from Netarts (92%) and Alsea (86%) had measurable 241Am (Fig. 3d). Tillamook (67%), Coquille (60%), Nehalem (57%), and Youngs (50%) were less likely to have cores with measurable 241Am. Amongst high marsh cores, 241Am and excess 210Pb inventories were most well correlated (Fig. S2a; R2 = 0.39, n = 38, p < 0.01), though 137Cs and 241Am inventories were also significantly correlated (Fig. S2c, R2= 0.18, n = 38, p < 0.01). None of the nuclides were significantly correlated with organic or (its inverse) mineral matter (excess 210Pb: R2 = 0.025, n = 38, p = 0.3; 137Cs: R2 = 0.018, n = 38, p = 0.4; 241Am: R2 = 0.043, n = 38, p = 0.2).

Fig. 3
Fig. 3
Full size image

(a) Map of the Oregon coast (U.S. Geological Survey GTOPO30) (inset displays the state boundary in gray) with estuaries labeled and high marsh inventories of (b): excess 210Pb, (c): 137Cs, and (d): 241Am grouped by estuary and plotted as box plots (boxes represent first and third quartile, center lines indicate medians, whiskers are minimums and maximums, and O’s are outliers; Kruskal-Wallis test results are presented above). Red numbers indicate the percent of cores with measurable radionuclides. Gray patches indicate the range of modeled atmospheric fallout for each radionuclide (Table S2). Estuaries are organized by sediment load with numbers below indicating the value in x 103 t y− 1 (Wise & O’Connor, 2016). Youngs load is likely higher given the input from the Columbia River

When both are detectable in high marsh sediment cohorts, 137Cs and 241Am activities correlate (Fig. S3; R2 = 0.32; n = 106; p < 0.01). In high marsh sediment cores, 241Am profiles tended to display narrower peaks (median FWHM = 3.6 ± 1.4 cm) than 137Cs (median FWHM = 4.2 ± 1.4 cm), though 241Am profiles had more examples of double peaks (e.g., Fig. 1c) and analytical uncertainties associated with 241Am activities are greater, resulting in greater instances of activity overlap near peaks (e.g., Fig. 1b). 137Cs and 241Am peak depths are not significantly different (KW, df = 75, p = 0.8). We also more frequently observed 137Cs present in sediments at depths well below the peak (Fig. 1), with the median distance between the peak and depth of onset 10.0 ± 4.8 cm compared to 3.0 ± 2.2 cm for 241Am. Assuming the age of the bomb peaks is 1963 CE (i.e., disregarding age-depth models calculated by excess 210Pb), the downward mobility of 137Cs makes the depth of onset, which should be 1954 CE, of 1920 CE ± 54 y too old on average, whereas the median age of onset in the 241Am profiles is more accurate at 1953 CE ± 36 y.

Fig. 4
Fig. 4
Full size image

Box plots (boxes represent first and third quartile, center lines indicate medians, whiskers are minimums and maximums, and O’s are outliers) and histograms of 137Cs and 241Am peaks dated by the excess 210Pb constant rate of supply (CRS) model. The dashed black line indicates 1963

The excess 210Pb CRS age-depth models calculated for high marsh cores dated the 137Cs and 241Am peaks to 1963 CE ± 18 y and 1959 CE ± 19 y on average, respectively (Fig. 4). 137Cs and 241Am peak ages calculated by the CRS age-depth model plus associated uncertainty ranges were dated to 1963 in only 18 (47%) and 22 (58%) cores for each radionuclide, respectively. Peak ages for both radionuclides were predicted accurately by the CRS age-depth model in 16 cores (42%) (Table S3).

Fig. 5
Fig. 5
Full size image

Box plots (boxes represent first and third quartile, center lines indicate medians, whiskers are minimums and maximums, and O’s are outliers) of (a) accretion rates calculated by different methods (including 137Cs, 241Am, and excess 210Pb by the constant rate of supply (CRS) model) and (b) mineral accumulation rates calculated using the CRS excess 210Pb data above and below the 137Cs and 241Am peaks compared across estuaries (sample sizes are indicated above each box plot)

Accretion rates determined by assuming peaks in 137Cs and 241Am represent 1963 are not significantly different (Fig. 5a; RANOVA, df = 30, p = 0.3); however, these bomb peak accretion rates tend to be faster than rates calculated by the excess 210Pb CRS model (Fig. 5a; Table S4). The three accretion rate methods are significantly different when variations across estuaries are considered (Fig. 5a; RANOVA, df = 60, p < 0.01). When excess 210Pb accretion rates are calculated post-1963 as defined by the 137Cs and 241Am peaks, sediment accumulates significantly faster in younger/shallower sediments than in older/deeper sediments (Table S5; RANOVA, 137Cs: df = 30, p < 0.01; 241Am = df = 30, p < 0.01) with estuary classification not significantly effecting differences in pre- and post-1963 accretion rate (Table S5; RANOVA, 137Cs: df = 30, p = 0.2; 241Am = df = 30, p = 0.2). Comparison of these more recent CRS accretion rates, calculated from the surface to the depth of the bomb peaks, to accretion rates determined by the 137Cs and 241Am bomb peaks resulted in no significant difference in the methods (Table S4; Table S5; RANOVA, 137Cs: df = 30, p = 0.2; 241Am: df = 30, p = 0.05), when accounting for the significant difference between estuaries (Table S4; Table S5; RANOVA, 137Cs: df = 30, p = 0.03; 241Am: df = 30, p = 0.03). Mineral accumulation rates calculated above and below the 137Cs peak are only slightly significantly different (Fig. 5b; Table 6 S; RANOVA, df = 30, p = 0.02) with no significance between estuaries (Fig. 5b; Table 6 S; RANOVA, df = 30, p = 0.2). This significance is driven by the two Tillamook cores, as without them there is no significant difference in accretion rates before and after the 137Cs peak (Fig. 5b; Table 6 S; RANOVA, df = 29, p = 0.2). Mineral accumulation rates above and below the 241Am peaks are also not significantly different (Fig. 5b; Table 6 S; RANOVA, df = 30, p = 0.06) even accounting for differences between estuaries (Fig. 5b; Table 6 S; RANOVA, df = 30, p = 0.5).

Excess 210Pb accretion rates do not correlate with excess 210Pb inventories (Fig. S4; linear regression, R2 = 0.04, n = 38, p = 0.02). Excess 210Pb accretion rates calculated above the 137Cs peak correlate with 137Cs inventories (Fig. S4; linear regression, R2 = 0.15, n = 38, p = 0.02), and excess 210Pb accretion rates calculated above the 241Am peak correlate with 241Am inventories (Fig. S4; linear regression, R2 = 0.13, n = 38, p = 0.02); however, in both cases, the relationship is not strong.

Discussion

Our primary goal was to assess whether 241Am is a robust geochronometer in salt marsh sediments and whether it might now be preferable to 137Cs after having had time to in-grow through decay of 241Pu. Through this effort we also aimed to add to the limited number of studies that use 241Am for dating in salt marsh depositional environments (Boyd & Sommerfield, 2017; Dyer et al., 2002; Thomson et al., 2002). Our findings show that while geochronologies derived from 241Am currently have limitations, the isotope’s stratigraphic behavior, future trajectory for ingrowth, and comparative performance (versus 137Cs) suggest substantial potential for sediment dating within high elevation areas of intertidal zones. In particular, 241Am can be a valuable event horizon for confirmation of excess 210Pb accretion rates and age-depth models in high marsh sediments and may replace 137Cs as a bomb-produced tracer.

Detectability of 241Am & 137Cs

Oregon intertidal zone sediments have lower activities of 241Am than 137Cs (and excess 210Pb) making it more difficult to detect. Regardless, a majority of sediment cores (60% of all cores, 81% of high marsh) had detectable activities of 241Am with which to reliably calculate sediment ages and accumulation rates. It is important to note that the relative detectability of 241Am globally is improving. As 241Pu continues its in situ decay, 241Am will peak in the mid 2030s and the isotope will remain relatively abundant for centuries given its long half-life; conversely, 137Cs with its short half-life is rapidly approaching undetectability (Boyd & Sommerfield, 2017). Moreover, as the ability to detect 137Cs diminishes, the need for gamma detectors with energy ranges up to 661.7 keV will diminish and those wishing to detect 210Pb (46.5 keV as well as 214Pb at 351.9 keV) and 241Am (59.5 keV) can choose detectors optimized to measure lower energy gamma emissions, while reducing the impact of high background signals through methods such as improved shielding and Compton suppression. Moreover, those specifically aiming to measure 241Am may use methods that separate and concentrate it in sediments (described by Ristic et al., 2002) or use α spectrometry (as suggested by Boyd & Sommerfield, 2017; see also Block et al., 2023).

Detectability Across Elevation in the Tidal Frame

Within the intertidal zone, 241Am is most frequently detectable in higher elevation environments, especially high marsh settings. Within the high marsh, 241Am inventory increases with increasing relative elevation. Limited ability to measure accretion rates in low elevation coastal sediments is not unique to the 241Am method. Mixing of sediments by turbulent flows and bioturbation and coarse grainsizes hinder detection and interpretation of any radionuclide in these tide flat and low marsh sediments. Indeed, although excess 210Pb and 137Cs are more often detectable in low elevation zones than 241Am, inventories of all radionuclides are low compared to modeled fallout inventories, indicating a more erosive environment. Moreover, depth profiles of these radionuclides are often difficult to interpret, precluding calculation of accretion rates with suitable uncertainties. This pattern underscores that limitations in 241Am detection are not unique but instead reflect broader constraints on geochronology in low-elevation erosional environments. In tide flat and low marsh sediments, excess 210Pb may be preferable to either 241Am or 137Cs because mixing and non-steady-state accretion rates are more easily identified in depth profiles, and these processes may potentially be corrected or accounted for in the accumulation rate model (Arias-Ortiz et al., 2018). Eidam et al. (2024), who focused on measuring accretion rates in tide flats of an Oregon estuary, were able to determine accretion rates in 8 of 14 (57%) sediments cores using the excess 210Pb method, further highlighting the relative success of the excess 210Pb method in these more dynamic environments lower in the tidal frame.

While the general difficulty of dating sediments in low elevation intertidal zones is lamentable, studies that leverage coastal sediments as recorders of tectonic events, climatic change, and anthropogenic activities should focus on assessment of high marsh archives when not explicitly concerned with measuring processes in low intertidal elevation zones. Not only are these sediments typically easier to date, but researchers frequently focus efforts on assessing accumulation rates in high marsh sediments given that these zones are assumed in dynamic equilibrium with mean sea level. Because of this dynamic equilibrium, high marshes are useful recorders of changes in relative sea level as well as good biogeochemical archives given their relatively consistent and reliable accumulation rates.

Detectability Across Estuaries

In high marsh sediments, 241Am is more detectable and inventories tend to be higher in estuaries with low sediment loads (Netarts, Sand Lake, Salmon, and Alsea) than in those with higher sediment loads (Nehalem, Tillamook, and Coquille), though Youngs Bay is an exception discussed below (Fig. 3d). A primary source of 241Am to these marshes may be atmospheric, yet a number of sediment cores from estuaries without large rivers displayed 241Am inventories greater than modeled atmospheric fallout. This apparent enrichment of 241Am may be resuspended estuarine and marine sediment deposited on high marsh platforms. Addition of resuspended material may be evidenced by double peaks in 241Am depth profiles, especially apparent in Netarts cores (Fig. 1c). It is possible that estuarine and marine redistributed sediments in sediment-poor systems may compensate for fluvial sediment limitations as marshes accrete to keep pace with rising seas. Peck et al. (2025) found that allochthonous sediment organic matter tended to be more marine sourced in these Oregon high marshes with low fluvial sediment supply, corroborating this hypothesis.

Youngs Bay salt marshes represent an important exception: radionuclide inventories there may represent an additional source of 241Am, highlighting the need to assess the environmental context of a site prior to using the radionuclide as a dating tool. Youngs Bay salt marshes are hydrologically connected to the Columbia River Estuary, which is downstream of the decommissioned Hanford Nuclear Reservation where plutonium was produced from 1943 until its shutdown between 1964 and 1971; and thus, these Youngs Bay sediments may have received 241Pu, the parent of 241Am, over this timeframe. Indeed, the sediment cores with detectable 241Am from Youngs showed a wider peak, dated from the present to 1977 by the excess 210Pb method. While 241Am might be a useful source tracer of radioactive contaminants in this environment, caution is urged before using it as a faithful geochronometer.

Excess 210Pb in high marsh sediments is also typically assumed to be predominately derived from atmospheric deposition (e.g., Cochran et al., 1998). While excess 210Pb inventories were less variable across estuaries and did not generally exceed modeled atmospheric fallout values, it did exhibit similar behavior as 241Am. Specifically, the highest inventories of excess 210Pb were observed in Netarts Bay, which has no major fluvial sediment source. When combined with the observation that 241Am and excess 210Pb inventories were significantly correlated – more strongly than even 241Am and 137Cs – it appears that similar mechanisms supply both 241Am and excess 210Pb to marsh sediments.

By way of contrast, 137Cs trends across marshes were more difficult to discern with the highest inventories in Coquille and Youngs, which receive the highest fluvial sediment loads. 137Cs may be less particle reactive on suspended sediment in the marine environment, where other cations are preferentially adsorbed to clays (Foster et al., 2006). As such, more terrestrially influenced marshes seem to be sites of 137Cs focusing, whereas more marine marshes receiving resuspended estuarine sediments may not accumulate more 137Cs beyond what is supplied through atmospheric fallout.

Estimated Accuracy

If the first step to assessing the usefulness of a radionuclide as a dating tool is to determine whether and under what conditions 241Am is detectable, the second step is to assess its accuracy. When it is detectable, 241Am appears less mobile than 137Cs as evidenced by narrower peaks and limited downward migration. Others previously made similar observations that 241Pu, the parent of 241Am, has a slower downward diffusion than 137Cs in sediments (Milan et al., 1995). The mobility of 137Cs in coastal sediments is well described, and frequently attributed to replacement of sorbed Cs+ by NH4+ and other cations, especially K+ and Na+, in sea water under anoxic conditions (Dyer et al., 2002; Foster et al., 2006; Johnson-Pyrtle & Scott, 2001; Leslie & Hancock, 2008; Olsen et al., 1981).

Despite the mobility of its tail, the median 137Cs peak was dated to 1963 CE ± 18 y based on the excess 210Pb CRS model. As the median 241Am peak was similarly dated to 1959 CE ± 19 y, we conclude that on average both methods are comparably reliable. However, it is important to highlight the wide distribution of bomb peak ages determined by the CRS model and that peak ages encompassed 1963 in approximately half of cores for both radionuclides. These results call for caution when using either of the bomb-related radionuclides for age dating without an independent geochronometer, such as excess 210Pb.

Moreover, the excess 210Pb accretion model impacts our interpretation of the accuracy of the 137Cs and 241Am dating method. The CRS model produces accretion rates between sampling intervals, thereby representing shorter-term, approximately decadal processes. Additionally, CRS accretion rates can be calculated for a similar timespan (1963 to the present) as the 137Cs and 241Am derived accretion rates, making it the more appropriate method for comparison than other models like the constant initial concentration (CIC) model that typically integrate over numerous depths.

When the CRS model is used to calculate shallower, more recent rates of sediment accumulation, the rates tend to be faster due perhaps to the absence of self-weight compaction and more limited time for decomposition. It is also possible that these faster rates of accretion in the more modern sediments reflect the Sadler Effect, which stipulates that accretion rates in depositional environments tend to decrease with increased period of observation related to the increased likelihood for incorporating a hiatus or erosional event (Sadler, 1981). Investigations on the presence of the Sadler Effect in modern-Holocene sediment archives and for estuarine sediments at or above wave base are rare, but there are a handful of examples demonstrating similar patterns of decreased accretion when increasingly old sediments are included in analyses (e.g., Eidam et al., 2024; McKee et al., 1983; Rodriguez et al., 2020; Sommerfield, 2006). Interestingly, Sommerfield (2006) found decreasing stratigraphic completeness with increasing suspended sediment concentration, which aligns with our observations of low radionuclide inventories at low tidal elevations and amongst estuaries of high fluvial sediment supply. Sommerfield (2006) attributed this somewhat counterintuitive trend to the episodic nature of deposition, erosion, and redistribution of sediments amongst shallow, muddy coastal systems.

However, a number of lines of evidence indicate that the Sadler Effect may not be fully reflected in these Oregon high marsh sediment cores. First, it is proposed that low radionuclide inventories indicate post-depositional erosion (i.e., stratigraphic incompleteness) whereas high inventories indicate sediment accumulation and possibly sediment focusing (Gallagher & Ross, 2018; Ravens et al., 2009). Yet high marsh sediment core inventories of excess 210Pb, 137Cs, and 241Am do not strongly statistically increase with increasing accretion rates. Perhaps most importantly, sediment mineral accumulation rates calculated across sediment cores by the excess 210Pb CRS method on depths below and above the 137Cs and 241Am peaks (i.e., representing pre- and post-1963) are not significantly different. Further, mineral accumulation rates – which account for changes with bulk density and organic matter accumulation with depth – lack significant depth trends. This absence of a depth trend in mineral accumulation rates indicates that the observed faster rates of vertical accretion in shallower depths may be a product of the absence of self-weight compaction and more limited decomposition in these surficial sediments. Our results therefore do not support the presence of the Sadler Effect in high marsh sediments. It is possible that the strong feedbacks that maintain the dynamic equilibrium between the high marsh environment’s elevation relative to mean sea level prevent substantial erosion on decadal to centennial timescales. More work is clearly needed to assess the mechanisms underpinning depositional intermittency in intertidal sediments, but radionuclide inventories, especially for less mobile 241Am and excess 210Pb, may help identify the degree of stratigraphic completeness.

Conclusion

Here we provide a list of practical recommendations for incorporating 241Am into measurements and interpretations of short-lived radionuclide data to calculate accretion rates and age-depth models in coastal sediments.

  • When performing gamma detection, 241Am (along with 137Cs) should be used to validate the excess 210Pb method through comparison to the CRS modeled ages and accretion rates. As long as 137Cs is still detectable, 241Am may be somewhat less favorable due to its low activities and high analytical uncertainties, but as 137Cs nears its obsolescence, 241Am, with its much longer half-life, has great potential as a marker horizon dating method. Labs seeking to measure accumulation rates of recent sediments by gamma detection could focus on optimizing their detector configuration for the detection of low energy excess 210Pb and 241Am.

  • Moreover, sampling at a finer depth resolution than at the 2-cm increments as was performed here likely would have improved the resolution of the 137Cs and 241Am peaks, resulting in lower age and accretion rate uncertainties. Sampling at a finer depth resolution, of course, is more time and resource intensive, can result in insufficient masses to achieve suitable activities, and would likely increase the necessary detection time. When appropriate, workers could section at a fine depth resolution, perform detection on radionuclides at the fine depth resolution in the stratigraphic section where a 137Cs/241Am peak is suspected, and combine depths to increase masses at depths above and below where excess 210Pb is the primary radionuclide of interest.

  • Excess 210Pb is more often present across intertidal habitats than either 241Am or 137Cs, and its ~semi continuous deposition allows for more reliable interpretation of activity depth profiles. Excess 210Pb is therefore a preferable method for age-dating recent coastal sediments and should generally be preferentially measured to either bomb produced radionuclide if a choice must be made, such as can be the case with α detection.

  • When making comparisons between accretion rates and ages determined by different models, it is important to compare across similar timeframes (as others have recommended: e.g., Sommerfield, 2006). Often accretion rates measured by 137Cs over the last ~half century and excess 210Pb over the last ~century or longer are interpreted together. Our results demonstrate that accretion rates measured over these timeframes are not directly comparable and that excess 210Pb should be validated through comparison of 241Am and 137Cs with accretion rates measured over a similar timeframe using the CRS model. Alternatively, workers might consider making direct comparisons using mineral accumulation rates as these seem less impacted by soil forming processes like compaction and decomposition. Ultimately when reporting accretion rates, the method and its associated timeframe should be dictated to best align with the timescale of the processes under study.