Comparison of basin-scale in situ and meteoric 10Be erosion and denudation rates across a rainfall, slope, and elevation gradient at George River, northeast Tasmania, Australia

Abstract. Long-term erosion rates in Tasmania, at the southern end of Australia’s Great Dividing Range, are poorly known, yet such knowledge is critical for making informed land-use decisions and improving ecological health of coastal ecosystems. Here, we present the first quantitative, geologically-relevant estimates of erosion rates for the George River basin, in northeast Tasmania, based on in-situ produced 10Be (10Bei) measured from stream sand at two trunk channel sites and seven tributaries (average 10.5 mm kyr−1). These new 10Bei-based erosion rates are strongly related to mean annual precipitation rates and elevation, and we suggest that the current East-West precipitation gradient across George River greatly influences erosion in northeast Tasmania. This stands in contrast to erosion rates along the mainland portions of Australia’s Great Dividing Range, which are more strongly related to basin slope. We also extract and measure meteoric 10Be (10Bem) from sediment grain coatings of the stream sand at each site, which we use to estimate 10Bem-based erosion and denudation rates for George River. 10Bem based erosion and denudation metrics, particularly those from the central and eastern tributaries, are also closely related to elevation and precipitation in the same manner as 10Bei erosion rates. Although 10Bem-based denudation rates replicate 10Bei erosion rates within a factor of two, 10Bem-based erosion rates are systematically 5–6x higher than 10Bei erosion rates. 10Bem erosion and denudation metrics for the westernmost headwater catchments are significantly lower than expected and have likely been affected by intensive and widespread topsoil erosion related to forestry, which delivers large volumes of sediment rich in 10Bem to tributary streams. The 10Bei erosion rates presented in this study may be useful for land managers seeking to restore ecological health of Tasmania’s estuaries by reducing sediment input to levels prior to landscape disturbance.



Introduction
Erosion rates of river basins derived from measurements of the in-situ produced cosmogenic isotope, 10 Bei, have been used to elucidate and infer topographic, tectonic, and climate drivers of landscape evolution for thousands of individual river basins (Codilean et al., 2018;Harel et al., 2016;Mishra et al., 2019;Portenga and Bierman, 2011;Wittmann et al., 2020). Recently,30 erosion rates from individual studies have been compiled and analyzed at the scale of entire continental orogens to demonstrate https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License.

Importance of Erosion of George River and the Georges Bay Watershed
Applications of 10 Bei erosion studies in Australia often are set within the context of assessing the impact of sediment delivery to sensitive offshore coastal environments, primarily the Great Barrier Reef (Croke et al., 2015;Nichols et al., 2014). Recently, efforts to conserve and restore estuarine environments across Australia have gained significant traction, particularly because estuaries link terrestrial fluvial systems to coastal environments and act as a biogeochemical buffer and sediment trap between 55 the two environments (Creighton et al., 2015;Fitzsimmons et al., 2015;Wolanski and Ducrotoy, 2014). These restoration efforts include hundreds of Tasmanian estuaries (Coughanowr and Whitehead, 2013;Murphy et al., 2003), which suffer from centuries of human-caused degradation resulting from urbanization, introduction of invasive species, forestry, mining, fishing, agriculture, and tourism (Augustineus et al., 2010;Butler, 2006;Davis and Kidd, 2012;Ellison and Sheehan, 2014;Jones et al., 2003;Martin-Smith and Vincent, 2005;Nanson et al., 1994;Seen et 60 al., 2004). Active conservation, restoration, and monitoring efforts are underway at many Tasmanian estuaries (Beard et al., 2008;Crawford and White, 2005;Creighton et al., 2015); none quantify geologically-relevant erosion rates nor sediment delivery, despite a recognized need to lower sediment delivery in order to reduce nutrient and pollutant loads, improve water clarity, and prevent burial of hard surfaces important for marine life (Elliott et al., 2007;Geist and Hawkins, 2016;Noe et al., 2020;Verdonschot, 2013). 65 This study focuses on the Georges Bay estuary in northeast Tasmania, which is known for its oyster stocks (Mitchell et al., 2000) but has been degraded by a history of timber production, tin mining, and agriculture. Historical land-use practices have supplied >10 6 m 3 of sediment to Georges Bay's primary tributary, the George River (no "s"), since the late 19 th century (Knighton, 1991) and continue to supply pollutants to Georges Bay (Bleaney et al., 2015;Crawford and White, 2005). The 70 intensive historical industrial use of the land in the George River catchment and the threat of excess sediment delivery to the fragile estuarine environment in Georges Bay has driven state and local municipalities to focus restoration and conservation efforts on the bay. As elsewhere, the success of these efforts relies in part, on reducing sediment delivery from George River to Georges Bay (Batley et al., 2010;Crawford and White, 2005;Kragt and Newham, 2009;McKenny and Shepherd, 1999;Mount et al., 2005). 75

In situ 10 Be and meteoric 10 Be erosion and denudation metrics
The primary goal of this study is to provide background rates (over millennia) of landscape change and sediment delivery from the George River to Georges Bay using the in situ cosmogenic isotope beryllium-10 ( 10 Bei) in fluvial sediment (Bierman and Steig, 1996;Brown et al., 1995;Granger et al., 1996). 10 Bei production decreases exponentially with depth in rock and sediment 80 at Earth's surface such that 10 Bei concentrations at depths >2 m is negligible compared to that measured closer to Earth's surface (Gosse and Phillips, 2001;Lal, 1991); 10 Bei produced by muons dominates at depths >2 m (Braucher et al., 2003;Gosse and Phillips, 2001;Heisinger et al., 1997), but muogenic 10 Bei production is generally negligible when compare to https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License. spallogenic 10 Bei production, except in rapidly eroding landscapes or landscapes with steep terrain (e.g. Dethier et al., 2014;Fellin et al., 2017;Rosenkranz et al., 2018;Scherler et al., 2014;Siame et al., 2011) or in paleoerosion studies (e.g. Schaller 85 et al., 2001Schaller 85 et al., , 2004Schaller 85 et al., , 2016. Bioturbation homogenizes 10 Bei concentrations in soils (Brown et al., 1995;Granger et al., 1996;Schaller et al., 2018), and 10 Bei erosion rates are therefore considered to be insensitive to widespread shallow erosion. This insensitivity allows 10 Bei erosion rates to be a useful gauge of pre-disturbance rates of landscape change (Ferrier et al., 2005;Portenga et al., 2019;Schmidt et al., 2018;Vanacker et al., 2007); exceptions have been noted where human land use is intensive (i.e. Schmidt et al., 2016) or the effects of human land use are exacerbated by climate extremes (i.e. Rosenkranz et 90 al., 2018). Pre-disturbance 10 Bei erosion data can inform approaches to reducing sediment delivery from George River and support efforts to improve the ecological health of Georges Bay estuary and possibly other watersheds in northeast Tasmania that share similar bedrock and topographic characteristics.
Whereas 10 Bei is produced in rock and sediment, 10 Be is also produced via spallation of oxygen in the atmosphere; this 10 Be 95 rains out or falls to Earth's surface (meteoric 10 Be; 10 Bem) where it is readily adsorbed into sediment grain coatings and traditionally used to trace sediment through landscapes (Brown et al., 1988;Heikkilä and von Blanckenburg, 2015;Helz et al., 1992;Monaghan et al., 1986;Portenga et al., 2017;Reusser et al., 2010b;Valette-Silver et al., 1986). Recently derived equations allow erosion rates and denudation rates to be calculated from measurements of 10 Bem and the chemically-similar, non-cosmogenic 9 Be, which is weathered out of mineral grains ( 9 Bereac; Willenbring and von Blanckenburg, 2010;von 100 Blanckenburg et al., 2012). 10 Bem erosion and 10 Bem/ 9 Bereac denudation rates have been used to quantify landscape evolution over a variety of spatial scales for long-established river basins (Dannhaus et al., 2018;Deng et al., 2020;Harrison et al., 2021;Portenga et al., 2019;Rahaman et al., 2017;Wittmann et al., 2012Wittmann et al., , 2015 and has shown particular promise in quantifying landscape dynamics in quartz-poor landscapes (Deng et al., 2020;Rahaman et al., 2017).

105
Here, we consider erosion to be the physical mass loss from a landscape and denudation to be the sum of physical and chemical mass loss. Conceptually, and with regards to 10 Be, Portenga et al. (2019) suggested that if soil thickness approximates the zone of 10 Bei production (<2 m) and if pH values in the environment are high (>3.9, Graly et al., 2010) such that no 10 Bem desorbs from sediment grain coatings, erosion and denudation rates derived from measurements of 10 Bei and 10 Bem/ 9 Bereac should be comparable and should measure landscape dynamics similarly. Although replication between 10 Bei and 10 Bem erosion and 110 denudation rates at individual sites is poor (Portenga et al., 2019;Rahaman et al., 2017;Wittmann et al., 2015), average rates from 10 Bei and 10 Bem/ 9 Bereac erosion and denudation datasets tend to be similar in magnitude, and similar spatial patterns of landscape change emerge from both 10 Bei and 10 Bem datasets across large regions (Deng et al., 2020;Portenga et al., 2019;Wittmann et al., 2015). Further use of the 10 Bem/ 9 Bereac denudation method in landscapes where 10 Bei erosion can be measured and compared is important for evaluation the veracity of 10 Bem erosion and 10 Bem/ 9 Bereac denudation calculations. 115 https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License.
The small size and relatively uniform bedrock geology of George River basin provide an ideal location to compare 10 Bei erosion rates with erosion and denudation rates derived using 10 Bem and 10 Bem/ 9 Bereac, respectively (Willenbring and von Blanckenburg, 2010;von Blanckenburg et al., 2012); additionally, measured soil pH values in the catchment range from 4.0-5.5 (Kidd et al., 2015), thereby suggesting that 10 Bem loss to chemical weathering is not a concern in George River. Thus, a 120 secondary goal of this study is to compare 10 Bem erosion and denudation rates to 10 Bei erosion rates as a means of assessing the efficacy of the 10 Bem erosion and 10 Bem/ 9 Bereac denudation methods in a landscape that minimizes geological heterogeneity, which otherwise may introduce scatter to larger datasets covering larger, more geologically-diverse landscapes (i.e. Deng et al., 2020;Portenga et al., 2019;Rahaman et al., 2017). Although George River has a simple bedrock geology, it also has a long history of intensive lode and placer tin mining that has, in the past, disturbed its fluvial systems (Knighton, 1991;Preston, 125 2012). Given that intensive land-use histories have affected results of 10 Bem calculations elsewhere (Portenga et al., 2019), we explore how mining in George River affects our interpretations of 10 Be-based erosion and denudation calculations throughout this study.
Bedrock of George River basin comprises granodiorite and granite associated with the Blue Tier Batholith, which were contemporaneously emplaced into sediments of the Mathinna Supergroup in the Devonian ( Fig. 2; Foster et al., 2000;Gee and Groves, 1971;Gray and Foster, 2004;Higgins et al., 1985;McCarthy and Groves, 1979;Seymour et al., 2006). Siluro-140 Devonian sedimentary rocks and Neogene basalts underlie small areas, primarily along drainage divides in the central and western George River basin (Seymour et al., 2006).
George River basin is of modest size (557 km 2 ) in northeastern Tasmania with low elevation (mean = 386 m) and gentle 145 hillslopes (mean = 10°) that drain the eastern slopes of the Rattler Range, which currently has a warm, temperate climate (Kottek et al., 2006). Despite eastern Tasmania being in the rain shadow of the central Tasmanian Highlands and western coast Logarithmic relationship between mean annual precipitation and elevation; data points binned at 01' longitudinal intervals. Historic rainfall data from active and inactive rainfall gaging stations (cyan and magenta triangles, respectively; BoM, 2021) is greater than modeled WorldClim rainfall for comparable elevations (Table  1), but the overall relationship between elevation and rainfall persists, regardless of which rainfall data are used.

150
Aboriginal Australians crossed to Tasmania from mainland Australia >35 ka (Cosgrove, 1995;Cosgrove et al., 1990), possibly corresponding to subaerial exposure of the Bass Strait ~56-40 ka (McIntosh et al., 2006) and localized ice advances in the central Tasmanian highlands (Barrows et al., 2001(Barrows et al., , 2002Macintosh et al., 2006). Ecological habitat suitability models, based 155 on characteristics and locations of thousands of archaeological sites across Tasmania indicate that Aboriginal communities were located close to freshwater sources and coastal resources, such as the landscapes around Georges Bay and the lower elevations within George River tributaries . Human arrival in Tasmania has been linked to widespread erosion events in mid-elevation landscapes (McIntosh et al., 2009).

160
Historically, decades of intensive tin lode mining in isolated headwaters of some tributaries and pockets of hydraulic sluice mining for tin in lowland floodplains introduced >10 6 m 3 of tailings to George River and its tributaries (Fig. 2a), decreasing the average grain size of alluvium from 30-50 mm to 1-2 mm (Knighton, 1991). Bedload characteristics have since returned to pre-disturbance levels following widespread alluvium storage in floodplains and aggradation at the George River delta in Georges Bay (Knighton, 1991;Cheetham and Martin, 2018). Despite George River's return to pre-disturbance channel and 165 bedload characteristics, a study from an experimental forest in the Gentle Annie tributary to George River shows that sediment yields from logged plots relative to unlogged plots continues to contribute sediment to the George River system (Wilson, 1999). More recently, land use within George River basin in 2008, at the time of sample collection, consisted primarily of forestry production from relatively natural environments and secondarily of conservation land (Fig. 4); intensive land use (i.e. built structures, permanent land alteration) and agricultural production from unirrigated land occur in equal proportion, though 170 much less than the primary and secondary land uses; a small percentage of George River is used for agricultural production from irrigated lands (ABARES, 2016).  (Gallant et al., 2011) https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License.

Methods
Sediment samples for this study were collected in 2008 from several locations along the trunk (n = 2) and tributaries (n = 7) of George River (Fig. 2). At each site, sediment was collected from the streambed and/or in-channel bars to ensure active fluvial transport and mixing. Samples were sieved in the field to the 250-850 μm grain-size fraction. Although this grain-size is finer than the mean natural grain size (30-50 mm; Knighton, 1991), previous studies show that 10 Bei grain-size bias is 180 minimal or not present in small, low-elevation, low-relief, temperate landscapes where landslides are uncommon (van Dongen et al., 2019); thus, 10 Bei measured from the 250-850 μm grain-size fraction can be interpreted as geological erosion rates. 10 Bem and the weathered and in situ phases of 9 Be ( 9 Bereac, 9 Bemin, respectively) were measured from the 250-850 μm grain-size fraction from all seven tributary sites and one of the trunk channel sites. 185 10 Bei was extracted from quartz from each sample at the University of Vermont following standard methods, during which a known amount of a 9 Be carrier ( 9 Becarr) was added to each sample (Kohl and Nishiizumi, 1992;Corbett et al., 2016); no native beryllium was detected in quartz concentrates from any sample, which can otherwise lead to significant overestimates of 10 Beibased erosion rates (Portenga et al., 2015). 10 Bei/ 9 Becarr ratios were measured by accelerator mass spectrometry at the Lawrence Livermore National Laboratory CAMS facility (Table 2); 10 Bei measurements were blank-corrected (the average ratio of three 190 blanks was subtracted from the ratio of unknown sample) and normalized to the 07KNSTD3110 AMS 10 Be standard material, which has a nominal 10 Be/ 9 Be ratio of 2.85 x 10 -12 (Nishiizumi et al., 2007). 10 Bei production was averaged across all sampled basins to a single point following Portenga and Bierman (2011), and the CRONUS on-line erosion rate calculator (Balco et al., 2008) was used to derive 10 Bei erosion rates following the Lal (1991) and Stone (2000) scaling schemes (ε , Table 3); here, ε is presented in units of mm kyr -1 . As in Portenga et al. (2019), we present 10 Bei-based sediment flux rates in this study 195 (Appendix A), which we present as the factor of ε and ρ in units of Mg km -2 yr -1 , so as to compare to 10 Bem/ 9 Bereac-based denudation rates (see below). Muogenic production of 10 Bei is incorporated into 10 Bei based erosion rates; however, muogenic 10 Bei is negligible relative to spallogenic 10 Bei production given George River's post-orogenic, low-elevation, low-relief setting. 10 Bem was extracted following Stone's (1998) fusion method and a 9 Be carrier solution was added to each sample. 10 Bem/ 9 Becarr ratios of these fusion extracts were measured at the Lawrence Livermore National Laboratory CAMS facility, blank-corrected (ratio of one blank was subtracted from ratio of unknown samples; Table 2) and normalized to the 07KNSTD3110 standard 205 material (Nishiizumi et al., 2007). Sample material used to calculate 9 Bereac was first subject to strong acid leaching to remove sediment grain coatings (Greene, 2016;Portenga et al., 2019 supplement); it was then fully digested in HF and 9 Bemin was measured in that solution. Both 9 Bereac from sediment grain coatings and 9 Bemin from the remaining mineral material were measured by inductively coupled plasma-optical emission spectrometry (ICP-OES) at the University of Vermont. In this study, E and Dm are presented in units of mm kyr -1 and Mg km -2 yr -1 , respectively. 210 E and Dm presented in this study (Table 3) are calculated using values of Q that range from 8.12 x 10 5 atoms cm -2 yr -1 to 1.06 x 10 6 atoms cm -2 yr -1 (Appendix A; Graly et al., 2011). Values of Q used here are of the same order of magnitude as 10 Bem accumulation rates measured from a similar latitude in New Zealand (1.68 to 1.72 x 10 6 atoms cm -2 yr -1 ; Reusser et al., 2010a), those integrated throughout the Holocene (1.0-1.5 x 10 6 atoms cm -2 yr -1 ; Heikkilä and von Blanckenburg, 2015), and 215 atmospheric-depth integrated rates of Q (~7 x 10 5 atoms cm -2 yr -1 ; Masarik and Beer, 2009;Willenbring and von Blanckenburg, 2010). We choose to use Graly et al.'s (2011) approach to deriving values of Q for this study since they are specific to the latitude and rainfall for each basin.
We compare ε, sediment flux, E, and Dm to various topographic and climatic factors to assess dominant processes driving or 220 related to background landscape evolution in George River (Table 1). Topographic data are derived from the SRTM 90-m resolution global dataset (Gallant et al., 2011). We use mean annual precipitation data from the updated WorldClim global dataset instead of precipitation from meteorological stations because of its greater spatial coverage, but we note that while WorldClim rainfall values are nominally lower than measured precipitation, both datasets show increased rainfall at higher elevations ( Fig. 3c; BoM, 2021). Ratings of soil erosivity have been derived for Tasmania (Kidd et al., 2014(Kidd et al., , 2015 and are 225 strongly tied to hillslope angle within George River basin (Fig. 5); thus, comparing erosion and denudation metrics against ε a Integration Sed Flux E D m (mm kyr -1 ) ± 2σ duration (kyr) (Mg km -2 yr -1 ) ± 2σ (mm kyr -1 ) ± 2σ (Mg km -2 yr -1 ) ± 2σ TG-1 9.6 1. basin slope metrics provides an adequate assessment of whether models of hillslope erodibility influences erosion in George River.

10 Bei erosion rates, ε
Erosion rates, ε, based on measured concentrations of in situ 10 Be (Table 3) range from 4.8 to 24.5 mm kyr -1 (Appendix A), and we find that the average ε from tributaries (13.6 ± 1.0 mm kyr -1 ; 2σ) is greater than either of the trunk channel samples (TG-1 = 9.6 ± 1.6 mm kyr -1 ; TG-9 = 8.3 ± 1.4 mm kyr -1 ; 2σ). Tributary values for ε, and ε-based sediment flux rates are greater in the high-elevation, western headwaters of George River basin and decrease systematically, eastwards towards the lower-235 elevation coast (Fig. 6). This eastward decrease in ε also corresponds to a decrease in rainfall along the precipitation gradient (R 2 = 0.82); relationships between ε and basin relief, basin-weighted slope, and the percent of each basin that is categorized as being greater than or equal to High Erosivity are weak ( Fig. 7; R 2 = 0.39, R 2 = 0.17, R 2 = 0.05, respectively).
Taking the product of ε and the area of each catchment provides the average annual volume of sediment exported from each 240 catchment over millennia. Summing these volumes, shows that a similar volume of sediment passes through the trunk channel sample sites annually as the sum of sediment exiting sampled tributaries (TG-1 = 3.8 ± 0.7 km 3 yr -1 ; TG-9 = 3.5 ± 0.7 km 3 yr -1 ; tributaries = 3.9 ± 0.3 km 3 yr -1 ; 2σ). Trunk channel samples, TG-1 and TG-9, should also incorporate erosion from their respective subcatchmentsthe area upstream of the sample site, but downstream of all tributary sample points. Using regression equations for ε and longitude, elevation, and precipitation, each, an average modelled ε for TG-1 is 9.6 ± 3.0 mm 245 kyr -1 and an average modelled ε for TG-9 is 1.1 ± 3.1 mm kyr -1 . The average volume of sediment these subcatchments contribute to annual sediment loads, based on modelled ε data are TG-1 = 1.1 ± 0.4 km 3 yr -1 and TG-9 = 0.0 ± 0.1 km 3 yr -1 .

10 Bem erosion rates, E, and denudation rates, Dm
10 Bem-based erosion rates, E, range from 17.0 to 78.3 mm kyr -1 and replicate values for ε well in the three westernmost 280 headwater catchments, but not in the lower-elevation, center and easternmost tributaries, where E is systematically ~5-6x higher than ε (Fig. 8). 10 Be m / 9 Be reac -based denudation rates, D m , range from 22.7 to 53.7 Mg km -2 yr -1 . Except for TG-4 and TG-9, values for Dm do not replicate sediment fluxes derived from ε, although the central and easternmost tributaries plot much https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License. closer to a 1:1 line than the samples from the three western-most headwater tributaries (Fig. 8). TG-4 was collected at the mouth of Groom River, upstream of which activities at the long-closed Anchor Mine significantly altered the topography. The 285 meteoric erosion rate for the trunk channel site TG-9, E = 78.3 ± 2.2 mm kyr -1 is significantly higher than ε for the same site (10.7 ± 1.7 mm kyr -1 ), but the denudation rate at TG-9, Dm = 33.0 ± 0.9 Mg km -2 yr -1 , replicates the ε-based sediment flux (28.9 ± 4.5 Mg km -2 yr -1 ).
In general, 10 Bem-based measures of E and Dm are not significantly related to any topographic or climatic metric (Figs. 6, 7). 290 However, the observed relationships between E and Dm and longitude, elevation, and precipitation are similar to those observed with ε in the central and eastern tributaries (Fig. 6); E and Dm in the western tributaries do not follow the spatial trends that ε exhibits.

5 Discussion
Erosion in George River is strongly related to basin longitude, elevation, and mean annual precipitation east of the Rattler Range and the prominent Ben Lomond Plateau (Fig. 6), and we find no evidence to suggest that ε is related to slope in George River over millennial timescales. This result differs from most studies, which show strong correlations between ε and mean basin slope at a global scale (Portenga and Bierman, 2011), at regional scales, across the Great Dividing Range on Australia's 300 mainland ( Fig. 9; Codilean et al., 2021;Nichols et al., 2014), and despite prior assessments of George River that suggest slope imparts a large control over erosion and sediment generation in the catchment (Jerie et al., 2003;Kragt and Newham, 2009).
Instead, our finding is consistent with Mishra et al.'s (2018) suggestion that in low-slope, low-elevation, post-tectonic settings, the relationship between slope and erosion becomes secondary to precipitation, and this study presents one of the clearest examples of erosion along a topographically induced precipitation gradient in 10 Bei erosion literature. 305

315
The very strong relationship between precipitation and ε would likely not have emerged had our 10 Bei samples been affected by clast attrition (Carretier et al., 2009), deep-seated landslides (Aguilar et al., 2014;Gonzalez et al., 2016;Puchol et al., 2014), or intensive erosion associated with mining, forestry, or agriculture (Barreto et al., 2014;Neilson et al., 2017). Even intensive https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License. tin mining, which supplied >10 6 m 3 to George River over the last two centuries (Knighton, 1991) seems to not have not had a long-lasting diluting effect on 10 Bei in sampled stream sediment. It is possible that mining efforts, especially the sluice mining, 320 did not lead to 10 Bei dilution because of the homogenizing effect of 10 Bei in well-bioturbated soils (Brown et al., 1995) or that the size of George River was large enough to buffer the effects of mining efforts in a similar way that large catchments can buffer the effects of landslide material (Niemi et al., 2005;Yanites et al., 2009). It is also possible that mining activity did lead to 10 Bei dilution, but that these effects have since recovered along with bedload characteristics (Knighton, 1991) similar to the rapid, two-year recovery of 10 Bei concentrations following storm-triggered landslides in Puerto Rico (Grande et al., 2021). 325 Overall, the close relationship between 10 Bei erosion rates and mean annual precipitation across George River demonstrates how well 10 Bei erosion rates can reflect background, geologically-meaningful rates of landscape evolution on millennial timescales, even in areas with long histories of intensive human land-use.
Higher values of ε where there is more rainfall suggests that more sediment is being generated in the western portion of the 330 catchment where larger volumes of rainfall can facilitate the generation, erosion, entrainment, and delivery of sediment to trunk channels than in the eastern portion of the catchment. However, we recognize that perhaps ε is not necessarily related to precipitation, but rather ε and precipitation may both be more directly influenced by elevation. Although no part of George River was ever glaciated, cirque development and periglacial activity was active to the southwest of George River, across the Ben Lomond Plateau during the Last Glacial Maximum and previous glaciations (Barrows et al., 2002;Colhoun, 2002), in 335 which case ε may be greater at higher elevations due to greater amounts of periglacial weathering.
High 10 Bei erosion rates have been linked to greater amounts of periglacial activity elsewhere (e.g. Delunel et al., 2010;Hancock and Kirwan, 2007;Marshall et al., 2017); however, periglacial activity in northeast Tasmania was typically limited to elevations >1,100 m (Colhoun, 2002), and it is therefore unlikely that periglacial processes would have increased erosion 340 rates in George River's western tributaries, all of which are below this elevation. Alternatively, higher ε at higher elevations may be due to greater amounts of rock exhumation for inland northeast Tasmania relative to the coasts throughout the Cenozoic, interpolated from apatite fission track cooling ages across Tasmania (Kohn et al., 2002). This is also unlikely, however, as landscape lowering over millions of years has slowed from an early Cenozoic peak rates of 30-50 m Myr -1 to late Cenozoic rates of <10 m Myr -1 , and rock exhumation rates are presently comparable, if not slower than new 10 Bei based erosion 345 rates presented in this study. We therefore remain confident that the relationship between ε and elevation and rainfall in this study are real and reflective of the influence of rainfall in driving landscape evolution over millennial timescales in George River.
Since pre-disturbance stream flow and bedload conditions were re-established by the 1990s (Knighton, 1991), it appears the 350 greatest risk of future excesses of sediment flux from George River to Georges Bay comes from land-use changes involving the widespread disturbance of surficial soils, such as through forestry (Wilson, 1999). The percentage of land used for https://doi.org/10.5194/gchron-2021-23 Preprint. Discussion started: 27 August 2021 c Author(s) 2021. CC BY 4.0 License.
Production forestry in native environments has been decreasing throughout the 21 st century (Fig. 4), and while some of this land use is being supplanted by Conservation and Protected Native Land Cover, which could buffer the effects of widespread erosion, much is being replaced by grazing and agriculture, which would only serve to increase erosion, particularly in the 355 headwater catchments where geological erosion rates are naturally high (Fig. 4). Given recent land-use trends, the 10 Bei erosion rates presented here may provide a useful benchmark level of sediment delivery to George River, Georges Bay, and other fluvial systems in northeast Tasmania that share topographic and geologic characteristics similar to those at George River.

Considerations of ε for trunk channel versus tributary sites 360
A mass-balance comparison of the volume of sediment passing through the trunk channels (ε x upstream area; 3.8 ± 0.7 km 3 yr -1 at TG-1 and 3.5 ± 0.7 km 3 yr -1 at TG-9) versus the summed volume of sediment exiting tributaries (3.9 ± 0.3 km 3 yr -1 ) suggests that little erosion (and therefore addition of sediment) is occurring in the trunk channel subcatchments. Average modelled ε for trunk channel sites, calculated using the regression equations and mean longitude, the mean elevation, and mean annual precipitation values for the TG-1 and TG-9 subcatchments (9.6 ± 3.0 mm kyr -1 and 1.1 ± 3.1 mm kyr -1 , respectively), 365 however, suggests that the TG-1 subcatchment should be contributing at least ~1 km 3 of sediment more to George River annually. Given that the mass of sediment leaving the tributaries is equal to the mass of sediment passing through TG-1 and TG-9, we make the interpretation that the 10 Bei measured at TG-1 and TG-9 trunk channel locations is effectively dominated by erosion in the tributaries, with little input from the subcatchments or George River floodplain, and ε at trunk channel sites should be considered minimum estimates of erosion for the upstream contributing area. Our interpretation of erosion at trunk 370 streams being dominated by headwater input has been made elsewhere, albeit in much larger river basins (i.e. Wittmann et al. 2009Wittmann et al. , 2011Wittmann et al. , 2016. The average erosion rate we therefore present for George River is the average of ε at TG-1, TG-9, and the average of ε from the seven tributaries combined (10.5 ± 0.8 mm kyr -1 ); we do not consider ε of trunk channel samples, modelled or measured, when considering spatial statistics of erosion in George River basin.

375
Compared to measurements of ε on the Australian mainland, the mean value of ε for George River (10.5 ± 0.8 mm kyr -1 ) is of similar magnitude as the median erosion rate for all catchments draining the eastern flanks of the Great Dividing Range passive margin of mainland Australia (15.9 mm kyr -1 ; Fig. 9; Codilean et al., 2021). Average ε from George River is most consistent, however, with the erosion rates of mainland basins at the southernmost extent of the Great Dividing Range, across the Bass Strait, which are those that share similar topographic characteristics and geological histories as George River (Codilean et al., 380 2021). The similarity between the geology, climate, and topography of newly-sampled basins and derived 10 Bei erosion rates in Tasmania and those from southeast mainland Australia suggests that evolution of landscapes that share similar climatic, topographic, and geologic characteristics is driven by common forces.

Comparing 10 Bei-based and 10 Bem-based erosion and denudation metrics
The strong relationship between 10 Bei erosion rates and topographically-induced precipitation across Georges River (Fig. 6) suggests that 10 Bei erosion rates, ε, are geologically accurate and meaningful. The small, geologically-homogeneous landscape of George River, therefore allows us to test a previous hypothesis (Portenga et al., 2019) that measured 10 Bem-based erosion 390 rates, E, and 10 Bem/ 9 Bereac-based denudation rates, Dm, to replicate ε or ε-based sediment fluxes, respectively. At first glance, E replicates ε only in the headwater catchments and Dm replicates ε-based sediment fluxes relatively well in all tributaries except for the headwater catchments (Fig. 8). Overall, values of E and Dm do not replicate the spatial patterns or yield the same relationships with topographic and climate parameters that we observe with ε and ε-based sediment fluxes (Figs. 6, 7).
However, when only E and Dm from tributaries in the central and eastern areas of George River (TG-2 through TG-5) are 395 considered, a consistent relationship between E and Dm and basin longitude, elevation, and precipitation emerges and is similar to the relationships we observe between longitude, elevation, and precipitation and ε (Fig. 6). Despite this small sample subset (n = 4), we suggest that E and Dm reflect the same patterns of landscape dynamics in George River as ε. Moreover, the similarity of Dm and ε-based sediment fluxes for central and eastern tributaries provides support our hypothesis that 10 Bem/ 9 Bereac based denudation rates should more-closely replicate 10 Bei-based erosion rates in small river basins where geological heterogeneity 400 is minimized. Our findings also, generally, support the continued exploration and application of 10 Bem/ 9 Bereac denudation rates in geomorphological studies (Dannhaus et al., 2017;Deng et al., 2020;Portenga et al., 2019;Rahaman et al., 2017;Wittmann et al., 2012Wittmann et al., , 2015. Interestingly, measured values of E are systematically ~4-5x greater than ε, and thus while E may be influenced by the same geomorphological processes in George River as ε and Dm, it does not appear to reflect accurate rates of landscape change. 405 The similarity of the spatial patterns of E and Dm with longitude, elevation, and precipitation in the central and eastern tributaries with those exhibited by ε across the whole George River basin suggests that E and Dm in the headwater catchments (TG-6, TG-7, and TG-8) significantly underestimate accurate or realistic values of E and Dm for these tributaries. Of all of the variables and measurements required to derive both E and Dm, excess amounts of measured 10 Bem is the only common factor 410 that would lead to erroneously low calculated values of E and Dm. Once delivered to Earth's surface, meteoric 10 Bem concentrates in uppermost soil horizons (Graly et al., 2010;Willenbring and von Blanckenburg, 2010), and thus any disturbance and excavation of large volumes of topsoil (i.e. agriculture, forestry, wildfire erosion, or mining activities) could strip soil with the highest concentrations of 10 Bem in its grain coatings entering and introduce them into a stream's bedload, a process similar to that identified following early land-use changes and deforestation in the Chesapeake Bay and San Francisco 415 Bay (Valette-Silver et al., 1986;van Geen et al., 1999). Mining activities in George River were restricted to lower catchment areas and tributaries where 10 Bem based metrics demonstrate the same relationships to topography and climate as ε (Fig. 2), and no wildfires nor prescribed fires have burned through the headwater catchments (Fig. 4). At the time of sample collection for this study in 2008, forestry from natural environments and from production plantations 420 was the largest land use designation within George River. Elsewhere in George River, rainfall and runoff experiments carried out in the Gentle Annie experimental catchment, a tributary to TG-2 (Fig. 4), showed that rills and gullies developed in hillslope plots that were heavily disturbed by forestry machinery, yielding significantly more sediment during simulated rainfall events than soil plots that were burned and more sediment than plots where soils were left undisturbed (Wilson, 1999). Although there are no detailed records of when plots of land in George River were timbered, we invoke Wilson's (1999) findings to 425 suggest it is plausible that active forestry had disturbed soils in the headwater catchments at or shortly prior to the timing of sample collection, significantly disturbing 10 Bem-rich top soils, and delivering large volumes of sediment with excess 10 Bem to sample collection sites, which subsequently resulted in the calculated values of E and Dm that are much lower than otherwise expected based on the trends of E and Dm in other tributaries. Following this interpretation, we suggest that measures of E and Dm may reflect spatial patterns or replicate ε rates in geologically homogeneous landscapes, respectively, but caution should 430 be taken when applying 10 Bem erosion and denudation metrics in landscapes with intensive soil disturbances (Portenga et al., 2019).

Conclusions
10 Bei erosion rates throughout the George River basin, and 10 Bem erosion and denudation rates in its central and eastern 435 tributaries, are closely related to a topographically-induced East-West precipitation gradient across the catchment. Tasmanian landscapes differ from the Great Dividing Range where erosion rates and slope are closely linked. The average 10 Bei erosion in George River, 13.6 mm kyr -1 , reflects erosion in tributaries to George River where precipitation is greatest; little sediment is generated in trunk channel subcatchments. These findings support the notion that precipitation imparts more influence on landscape development in low-slope, low-elevation landscapes (Mishra et al., 2018), which often tend to be in post-orogenic, 440 passive margin settings.
Although sediment erosion associated with mining, agricultural, and forestry land-use practices occurred in the George River basin during the 19 th and 20 th Centuries, 10 Bei based erosion rates in the basin appear to reflect pre-disturbance rates of landscape change. Such rates are useful as part of Tasmania's current efforts to re-establish healthy and sustainable ecological 445 conditions in its many estuarine environments, particularly those in northeast Tasmania where estuary tributaries have similar geological and topographic characteristics to those found at George River. The pace of erosion in the George River basin is similar to that at the southern end of the Great Dividing Range on the Australian mainland, which has similar bedrock and climate characteristics. 450 10 Bem-based erosion and denudation in the central and eastern tributaries of George River generally replicates spatial patterns of 10 Bei-based erosion and denudation. Low 10 Bem-based erosion and denudation rates calculated in three headwater tributaries demonstrate the sensitivity of meteoric 10 Be-based calculation to recent and intensive land use that disturbs and erodes topsoils.

Data Availability
All data used in this study, and all data needed to reproduce our findings are presented in Tables 1-3 and the equations we use to work with data to calculate erosion rates, sediment fluxes, denudation rates, and integration times are presented in Appendix A. Data entry for calculating erosion rates from the CRONUS online erosion rate calculator, formatted for text entry, are given 465 in Appendix.

Author Contribution
The conceptual analysis of the data presented in this paper comes from LV's Undergraduate Honors Thesis (2020) at Eastern Michigan University. EWP contributed to post-thesis manuscript revisions, data analysis, and figure drafting. Samples and the 470 10 Bei data presented here were collected and facilitated by PRB and ECL in 2008. 9 Be and 10 Bem data were first presented in Sophie E. Greene's Master's Thesis (2016) at the University of Vermont; SEG declined a request to participate in the writing and publication of this paper. AJH verified Lawrence Livermore National Laboratory's measurement of beryllium at the Center for Accelerator Mass Spectrometry in 2009. This work was performed in part under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. This is LLNL-JRNL-825534. 475