Lake morphometry controls the remobilization and long-term geochemical imprint of distal tephra deposition

Abstract

Deposition of distal tephra-fall deposits in lake systems with low watershed sediment fluxes can represent large geochemical anomalies in sediment elemental composition. We examine the persistence of compositional effects of fallout from the 7.63 kyr BP Mt. Mazama caldera-forming event and other tephra deposits in the micro-laminated abyssal sediments of Pettit Lake, central Idaho, which received ~50 kg m⁻² of Mazama tephra-fall material. Persistent post-depositional dominance of the Mazama material is reflected in elemental composition and grain size distribution and is related to lake hypsometry and the presence of a large in-lake sediment reservoir subject to remobilization. Decoupling of recycled clastic mineral material with Mazama compositional affinity from both biogenic matter and from clastic material derived from within the watershed is shown by coherence between dated primary tephras and a radiocarbon chronology.

Introduction

Widely distributed tephra-fall deposits occur in lakes throughout the western North American cordillera, and commonly provide valued chronological control for Holocene sediment sequences. Beyond their chronological and stratigraphic usefulness, thicker tephras from major eruptions may represent a substantial proportion of the allogenic sediment deposited over the course of lake history, delivered effectively instantaneously (as fallout) or over relatively short periods of post eruptive erosion from watershed land surfaces. The delivery of such an event deposit can have substantial transient effects on both terrestrial (Long et al. 2014) and aquatic productivity and ecosystem structure (Bradbury et al. 2004; Gregory-Eaves et al. 2004; Telford et al. 2004; Hughes et al. 2013). Distal tephras also provide elemental tracers of allogenic material, and a means to trace watershed processes and the persistence of tephra-fall effects on lacustrine clastic sediment character through erosion from watershed storage, in-lake resuspension, and bioturbation (Thompson and Bradshaw 1986; Boygle 1999). In an advantageous depositional environment with high temporal resolution, elemental composition determined by X-ray fluorescence (XRF) scanning can yield remarkably detailed histories of recurrent volcanic activity (Van Daele et al. 2014). If compositional effects are long lived, event deposits may also pre-condition lake systems to sensitively record subsequent limnogeologic conditions by sediment elemental composition, as material is remobilized and deposited in response to changing wave base and/or lake stage. In lakes preserving fine-scale sediment structure, highly resolved records of compositional variation, enabled by prior allogenic events and remobilization of event-sediment reservoirs, can be linked to watershed hydroclimate.

Using elemental analysis obtained from core-scanning XRF supported by grain size analysis, we examine the long-term compositional effects and the record of subsequent remobilization of the regionally distributed Mazama tephra (7.63 kyr BP; Zdanowicz et al. 1999) within the basin of Pettit Lake, a persistently stratified, moraine dammed subalpine lake in Idaho’s Stanley Basin that received distal (600 km transport) Mazama fallout amounting to ≥6 cm depositional thickness and approximately 50 kg m⁻² loading. Pettit Lake occupies the western margin of the Stanley Basin at an elevation of 2,136 m. The lake lies on the hanging wall of the Sawtooth Fault, an active basin and range province normal fault bounding the Sawtooth Range to the west and exhibiting 2–3 m of Holocene offset (Thackray et al. 2013). Bedrock lithologies in the Pettit Lake drainage basin comprise mainly biotite granodiorite of the Atlanta lobe of the late Cretaceous Idaho Batholith (Worl et al. 1991; Gaschnig et al. 2011), a suite of rocks with relatively homogeneous whole-rock elemental composition (Gaschnig et al. 2011). Pettit Lake is impounded by one of a complex of post-LGM, late-Glacial end moraines, with a minimum age near 14,000 cal yrs BP (Thackray et al. 2004). Basal sediment ages in lacustrine headwater cirque lakes of the adjacent drainages (Mijal 2008) indicate rapid deglaciation of up-gradient basins. We hypothesize that major element compositional contrasts between Atlanta lobe rocks and Mazama tephra (Table 1) will allow us to assess the persistence of a single allogenic deposit in controlling subsequent sedimentary character. If so, characteristic compositional trends will provide a record of sediment remobilization and ultimately lacustrine transport energy, e.g., storminess and wave base depth.

Table 1 Major elemental compositions of clastic source materials in the Pettit Lake watershed

Lacustrine environment

Pettit Lake is a 162 hectare, 4.5 × 10⁷ m³ water body (Fig. 1) with a maximum depth near 50 m, a morphometrically simple main basin containing >95 % of lake volume and a low-volume secondary basin (defined by a drowned end moraine) at the outlet (eastern) end of the lake. Within the main lake basin, water-column stratification has been persistent, but not permanent, during year-round monitoring occurring over a 20-year period (Budy et al. 1995; Griswold unpublished data). During this time, hypolimnetic dissolved oxygen (DO) concentrations of <31.5 μmol l⁻¹ (0.5 mg l⁻¹) have typically been observed below water depths of 35–40 m. Occasionally, short term fully-mixed water column conditions have been reported (Griswold unpublished data; Fig. 2). Previous research activity on Pettit Lake centers on its ecosystem structure and status as a former nursery lake for the endangered upper Snake River Sockeye salmon (Selbie et al. 2007; Griswold et al. 2011).

Fig. 1

Bathymetry and coring locations in Pettit Lake, Idaho. Closed symbols represent coring sites of 2007, compiled here. Open symbol shows eastern basin coring location of Thackray et al. (2004). Arrows indicate surface water inflow and outflow locations. Water depth in meters, from Budy et al. (1995), with shading indicating typical position of oxycline near 40 m depth. Inset shows location of Pettit Lake (triangle) and of Crater Lake, Oregon, the Mt. Mazama source caldera (circle). Contours showing depositional thickness of the Mazama tephra are adapted from Young (1990)

Fig. 2

Dissolved oxygen concentrations in Pettit Lake during a typical multi-seasonal period of persistent stratification (upper series, 1998–1999) and a less-characteristic period of dominantly mixed conditions (lower series, 1995–1996). Unpublished data from Griswold

Methods

Core recovery and sampling

We collected three piston cores spanning Holocene and latest Pleistocene sediment of Pettit Lake in June of 2007, using a Kullenberg coring system and platform and the logistical support of the University of Minnesota LacCore staff. Piston core recovery ranged from ~2.5 to 3.5 m, with cores ending in laminated glaciolacustrine silts and sands (Fig. 3). Companion surface cores captured the sediment–water interface (SWI) with minimal disturbance and overlap the Kullenberg sequences. We carried out documentation of sediment character and stratigraphy at the LacCore facility (http://lrc.geo.umn.edu/laccore/), including high-resolution line scans and magnetic susceptibility measurements, detailed macroscopic sediment description, and petrographic analysis and classification (Schnurrenberger et al. 2003) of smear slides from all lithologic units. Continuous sampling on 5-mm centers of a composite sequence drawn from three overlapping, highly correlative core sections provides the framework for sediment physical properties, isotopic composition of organic matter (OM), biogenic opal concentrations, and carbon content.

Fig. 3

Composite lithologic section of Pettit Lake central basin. Lithofacies A, very finely laminated silty ooze and diatomaceous silt with numerous well-sorted thin (<0.2 cm) white silt lamellae, is the dominant component of Unit 1. Lithofacies B, normally graded event deposits, is a minor component of Unit 1 and the dominant component of the late-Pleistocene Unit 2. Lithofacies C, repetitively laminated silt and fine sand, defines the glaciolacustrine sequence of Unit 3. Lithofacies D, tephra, includes four distinct deposits: the Glacier Peak ‘G’ series (D_GP), the Mazama fallout deposit (D_MAZ), the Mt. St. Helens ‘Y’ series (D_MSH), and an unidentified olivine-bearing mafic tephra (D_MAF). Data curves show centered log ratios (clr) of Ti counts, showing modal Ti increase in post-Mazama relative to pre-Mazama sediment

Elemental analysis

Elemental data are derived from continuous XRF scans run on the Cox Analytical Systems Itrax™ core scanner operated by the Large Lakes Observatory at the University of Minnesota-Duluth. Pettit Lake XRF scans used a 0.5 mm step, 30 s dwell time and a Mo radiation source. Intensity data were processed by the following procedure: Elements with known instrument sources and those with a significant number of ‘0’ values were deleted. Rare ‘0’ values in the remaining 15-element matrix were eliminated by deleting the entire depth increment where zero values appear, resulting in the clipping of a total of 287 intervals from the original 6,262-interval data matrix. Intensity values were then normalized to the total emission count for that interval. These nonzero, count-normalized intensity data were then transformed by the centered log ratio (clr) and isometric log ratio (ilr) techniques (Pawlowsky-Glahn and Egozcue 2006; Weltje and Tjallingii 2008). The interpolated age model was then tied to the 0.5-mm depth increment, giving an A * B matrix of 5,974 interpolated time steps and 15 elements over the latest Pleistocene and Holocene record of sedimentation. Descriptive ternary and quaternary plots of elemental compositions were generated using the analytical package CoDaPack (Comas-Cufí and Thió-Henestrosa 2011; http://imae.udg.edu/codapack), with elemental data points classified within one of seven principal lithostratigraphic sub-units recognized in the Pettit Lake post-glacial sequence: (1) laminated glaciolacustrine silt, (2) a late Pleistocene turbidite sequence, (3) the Glacier Peak ‘G’ tephra, (4) pre-Mazama diatomaceous silt and silty ooze punctuated by thin (<0.5–2 mm) white silts, (5) the Mazama tephra, (6) the Mt. St. Helens ‘Y’ series tephra, and (7) post-Mazama diatomaceous silt and silty ooze punctuated by thin white silts.

Grain size analysis

We carried out analyses of grain size distribution in selected intervals using a Beckman-Coulter LS-232 laser analyzer with PIDS lamp. Sample preparation was adapted from the method of Jiilavenkatesa et al. (2001), with diatom frustules removed by 1 M NaOH and OM removed by 35 % H₂O₂. The intervals analyzed included 32 complete silt lamellae (7 pre-Mazama and 25 post-Mazama) identifiable visually and by XRF signatures, and 30 ‘background’ samples (6 pre-Mazama and 24 post-Mazama) taken as 5- or 10-mm sections of laminated sediment lacking recognizable clastic pulses. Pre-Mazama and post-Mazama grain size distributions were stacked by first normalizing the binned amplitudes to the total aggregate curve area for that sample (to normalize for differences in sample quantity), summing the normalized amplitudes, and dividing the aggregate binned amplitudes by the number of samples in the class.

Sediment chronology and age modeling

The chronology for the Pettit Lake sediments we analyze here is based on (1) ten AMS radiocarbon analyses of <250 μm charcoal (three samples) and terrestrial macrofossil remains (seven samples; Table 2) (2) two regionally distributed silicic tephras from the Cascade Range, the Mazama tephra and the Glacier Peak G series tephra, initially identified by stratigraphic position and geochemically confirmed by electron microprobe analysis of major elements in glass shards, and (3) a visual count of sand/silt couplets, inferred to represent annual glaciolacustrine cycles, found in the lowermost lithostratigraphic unit recovered by our cores. The SWI, present intact in our surface cores, is assigned an age of −57 calibrated years BP (CE 2007), the year of core recovery. Radiocarbon samples sieved from 1- to 1.5-cm sediment intervals are assumed to have been evenly distributed across sample intervals and are assigned the mid-depth of the interval in age models. The Mazama tephra is assigned an age of 7,627 cal yrs BP (Zdanowicz et al. 1999) and the presumptive Glacier Peak G series tephra an age of 13,550 cal yrs BP (Kuehn et al. 2009). The Mt. St. Helens Y series tephra is also present, but is not used explicitly in this chronology, due to some dating uncertainty and the availability of a radiocarbon analysis from immediately beneath the tephra. Event deposits (tephras and four short, normally graded or massive sections) are assigned constant ages based on the interpolated age of basal contacts.

Table 2 Results and accession numbers (Center for Accelerator Mass Spectrometry, Lawrence Livermore Laboratory) for AMS radiocarbon analyses of selected Pettit Lake material

The interpolated age-depth model and probability distributions used are generated by the Bayesian statistical model Bacon (Blaauw and Christen 2011), running in R 2.15.1 under MacOSX 10.7.5. The weighted mean calendrical age time series generated by Bacon closely approximates manual interpolations based on CALIB 6.0 (Stuiver et al. 2005) calibrations of the ¹⁴C data set. Statistical age-depth modeling is confined to the post-glacial sediments; sedimentation rate and age of the laminated glaciolacustrine sequence assumed the lowermost sediments are approximated by annual couplet formation and added the calculated years to the down-core end of the modeled age series.

Results

Sediment lithology and structure

Deep-basin Pettit Lake sediments consist of four identified lithofacies (Fig. 3): (A) very finely laminated (sub-mm) diatomaceous slightly silty to silty ooze. Image based counts of visible lamellar structure, which varies in degree of preservation, indicate that although very fine, these are multi-year features. Counts suggest average frequencies of 0.33–0.5 yr⁻¹ for these lithofacies A lamellae. This lithofacies is also punctuated episodically by thin (2-mm to fractions of a millimeter), prominent, clean white silt intervals. Bulk sediment composition of this dominant lithofacies is 15–20 % OM, 25–35 % biogenic silica, and 45–55 % clastic mineral and glassy material, with occasional trace amounts of endogenic calcite. (B) Normally graded sequences (<10 cm thick) of fine sand and silt with white very fine silt caps, interpreted as turbidites generated by rare delta-front failure basinward of the main lake inlet stream. (C) Light grey, very fine silt repetitively interlaminated at a scale of 1–10 mm with fine sand or coarse silt, interpreted as glaciolacustrine varves deposited during glacial recession. (D) Air-fall tephra.

Lithofacies A dominates Lithostratigraphic Unit 1 (U1, Fig. 3), accounting for approximately 2.4 m of latest Pleistocene and Holocene section in central Pettit Lake. White, sorted silt lamellae incorporating subordinate diatom frustules, rare sponge spicules and OM occur episodically up-section away from contacts with underlying clastic material of lithofacies C and D. Variably preserved sub-mm scale laminated structure persists throughout.

Lithofacies B occurs as a 2.6-cm event deposit within U1, dating to near 5,050 cal yrs BP, and as the dominant lithology of Lithostratigraphic Unit U2, where three distinct event deposits totaling ~12 cm cumulative thickness occurred shortly prior to the deposition of the presumptive Glacier Peak G tephra near 13,550 cal yrs BP. Lithofacies C makes up Lithostratigraphic Unit U3, with an unknown thickness (>55 cm) deposited prior to 13,600 cal yrs BP. Lithofacies D occurs as four recognized tephra-fall deposits within Lithostratigraphic Unit U1. These are the Mt. St. Helens Y series and Mazama deposits, confirmed by microprobe analysis of 50 (MSH) and 175 (Mazama) glass shards from the two deposits (J. Addison unpublished data), the presumptive Glacier Peak G series tephra (confirmed geochemically in nearby Upper Cramer Lake; Mijal 2008) and a thin (<1 mm) olivine-bearing mafic tephra of undetermined source dating to near 6,000 cal yrs BP by our radiocarbon chronology (Fig. 4).

Fig. 4

Sediment depth-age models for Pettit Lake abyssal composite section (solid symbols in Fig. 1). Dotted curve and shaded region show weighted mean and 95 % confidence intervals calculated by the Bayesian modeling package Bacon (Blaauw and Christen 2011; http://chrono.qub.ac.uk/blaauw/manualBacon_2.2.pdf) using section length of 1 cm. MAZ, Mazama tephra; GP, Glacier Peak ‘G’ series tephra; GV, laminated glaciolacustrine section. Markov Chain Monte Carlo iterations (upper left) indicate a stationary distribution, consistent with a robust result. Prior (dashed curve) and posterior (shaded) estimates of accumulation rate correspond closely (upper center). Using this short section length, default priors of memory give large overestimates of posterior distributions

Grain size distributions

Normalized pre-Mazama silt lamellae and bulk clastic material within organic-rich silts share a dominant mode near 3 μm, with subordinate modes near 15 and 40 μm (Fig. 5). Post-Mazama sediments display less sorted and more complex grain size distributions, with medium silt fractions (16–31 μm) constituting an important mode of the clastic fraction in both the silt lamellae and in organic-rich silts throughout the post-Mazama period of deposition. In post-Mazama bulk silts, the normalized abundance of the ~15 μm mode reaches an amplitude similar to the ~3 μm fraction dominating the pre-Mazama sediment.

Fig. 5

Pre-Mazama and post-Mazama grain size distributions for discrete silt lamellae (a) and for low-clastic laminated organic sediments (b). Individual grain size distributions determined by laser diffraction were stacked by first normalizing the binned amplitudes to the total aggregate curve area for that sample (to normalize for differences in sample quantity), summing the normalized amplitudes, and dividing the aggregate binned amplitudes by the number of samples in the class

Elemental composition

Figure 6 shows normalized elemental relationships of lithologically and temporally distinguished units in K–Ti–Ca–S elemental space, chosen to exploit known contrasts in Ti:K and Ti:Ca between dominant source lithologies, and to highlight diagenetic compositional variability in S that we anticipate may be indirectly related to OM production through its effect on redox potential. Seasonally laminated glaciolacustrine sediments, low in OM and comprising erosional products of the intrusive Sawtooth Range core, plot in a relatively low Ti:K field also occupied by a 9-cm turbidite sequence derived from remobilization of the glaciolacustrine material shortly after glacial retreat from Pettit Lake. Higher-OM silt from the pre-Mazama section (~13.5–7.63 ky BP) extends along an overlapping field that in this elemental space follows a trend between the late-Pleistocene compositions and a relatively high-S sediment component. The graded primary deposit of Mazama ash (approximately 6 cm in thickness) defines a high Ti:Ca, high Ti:K field relative to the locally derived Idaho Batholith compositions (Fig. 6; Table 2). Post-Mazama silts define a genetically related trend plotting as a variable sulfur contribution to the Mazama composition, similar to the pre-Mazama compositional trend but now with an end member composition concordant with the Mazama tephra rather than the Idaho Batholith. Mt. St. Helens Y ash (MSH, approximately 3.7 kyr BP, 2 cm in thickness) intervals are similar to Mazama in this compositional space, and a MSH memory effect distinct from Mazama is therefore not apparent. The Glacier Peak tephra (an ~8-mm deposit in profundal Pettit Lake sediments) appears to exert little persistent compositional influence over the subsequent sediments, as measured by XRF.

Fig. 6

Isometric log-ratios (ilr) of normalized XRF elemental intensities, classified by lithostratigraphic sub-units and centered on the system Ca–K–Ti–S. Inset is the same plot, scaled to the full ternary axes

Discussion

Shifting of the post-Mazama stacked grain size distributions relative to the pre-Mazama distributions reflects a persistent coarsening of the allogenic mineral fraction, sustained over ~7,500 years. Based on grain size alone, this coarsening could reflect either diminishment of the very fine silt (<8 μm) contribution, or the sustained contribution of material in the 10–50 μm size range. As we show below, this observation is best explained as the persistent textural effect of event deposition, i.e., the Mazama tephra-fall, a normally graded deposit which by petrographic observation of the Pettit Lake section ranges in grain size from ≤5 to 100 μm.

Normalized XRF intensities reflect the underlying affinities of quiescently deposited sediments with preceding event deposits emplaced in the Pettit Lake basin, overprinted by biologically mediated sediment components. In both the pre-Mazama and post-Mazama sequences, we interpret the relatively high-S trends as reflecting redox mediated sulfur cycling and Fe-sulfide formation, likely related to varying OM concentrations. This interpretation of sulfur variability gains support from trends in Ca–K–Ti–Fe space (not shown), which exhibit varying Fe enhancement (predictably less apparent than for S, since secondary Fe sulfides are one of several mineral reservoirs for Fe but are likely the only inorganic reservoir of significance for S). Petrographic observations showing wide variation in diagenetic sulfide abundances within discrete lithostratigraphic units also support this idea.

Storage and remobilization of Mazama material

Persistent compositional effects of the Mazama tephra deposit indicate ongoing remobilization of a tephra-dominated clastic reservoir or reservoirs throughout the 7,600 years of post-Mazama lake history. Reservoir candidates include soil and sediment of the Pettit Lake watershed, and remobilized sediment from within the lake basin itself. Petrographic examination of Pettit Lake sediment shows that secondary Mazama silt lamellae are more strongly associated with diatoms (commonly including many benthic forms), sponge spicules, and amorphous OM than they are with terrestrial organic debris. Furthermore, the well preserved lamellar structure of profundal Pettit Lake sediments, supported by observations of typical DO conditions at depth, dictates against bioturbation as an important mechanism for the compositional memory observed. We therefore interpret the Mazama-dominated clastic flux to be primarily the result of intra-lake remobilization of material stored in shallow-water reservoirs, in some contrast with the emphasis given to watershed remobilization in Icelandic lacustrine sediments by Thompson and Bradshaw (1986) and Boygle (1999), although an element of watershed remobilization is also implied by the thickness of the Mazama fallout and high relief terrain of the Pettit Lake watershed. By this model, the pulsed character of the silt flux (Fig. 3) largely reflects episodic deepening of wave base resulting in resuspension and lake-wide distribution of material that includes a prominent medium silt size fraction.

Two limnogeologic factors promote the persistence of a recycled allogenic signal in Pettit Lake. First is a very low overall clastic sediment flux during the Holocene. Central-basin post-glacial linear sediment accumulation (1.8 × 10⁻² cm yr⁻¹) and net mass accumulation (7.2 × 10⁻³ g cm⁻² yr⁻¹) rates are notably low in comparison to rates in nearby Redfish and Hell Roaring lakes. Importantly, the bathymetry of Pettit Lake (Fig. 1) provides an extensive shallow-water reservoir for the temporary storage and remobilization of fine-grained sediment delivered as tephra-fall material, in the form of the lake’s shallow eastern basin. With a low rate of lake-wide sediment loading from the watershed, and the main lake basin acting as an efficient sediment trap, the time-averaged Holocene clastic sediment flux in the eastern basin (which receives no direct stream inflow) may be dominated by rare tephra inputs. Cores recovered from this eastern basin in 2000 (Lundeen 2001; Thackray et al. 2004) display low overall Holocene accumulation rates and very thin (15–30 cm) post-Mazama sections. We infer that these features are due to episodic re-entrainment of sediment, including OM and stored Mazama clastic material, and advection of suspended material back into the main lake basin. Core-based estimates of clastic sediment balance support the plausibility of this sediment storage and transfer mechanism. The ~35 ha eastern basin received at least 4.5 cm of tephra fallout (Lundeen 2001). Visually identifiable Mazama-rich clastic lamellae amount to ~4.2 cm of cumulative section in profundal cores collected from the main lake basin. The ~27 ha of east basin shoal with <4 m water depth could thus provide much of the material for the remobilized silt lamellae observed in the 53 ha profundal basin below the 40 m contour. Littoral areas of the main lake basin provide a secondary reservoir from which tephra likely has been remobilized.

The hypothesized re-entrainment of Mazama material from storage in the eastern lake basin, plausible from a sedimentological standpoint, can also be tested against sediment suspension mechanics and local climatology. Wind speed and direction data from the nearby Stanley Ranger Station (NCDC site GHCND:USW00004112) are available from 1998 to the present, and provide insight into the plausibility of this re-entrainment mechanism for advecting stored east basin sediment into abyssal waters of the main lake basin. We estimated boundary-layer shear stresses according to the formulations of Sheng and Lick (1979) and Laenen and LeTourneau (1996), using a range of wind speeds bounded by smoothed daily averages of the midsummer windy season (~200 cm s⁻¹) and 400 cm s⁻¹; daily 2-min sustained maxima for the windy season are 700–800 cm s⁻¹ with an average azimuth of 225°. Assuming a roughness coefficient of 0.05 (Reynolds number = 20) for an OM-bearing silt composition and fetch approximated by the distance of open lake surface along a 225° trajectory into the eastern basin, boundary-layer shear stress may exceed 0.5 dynes cm⁻² at water depths from ~1 to ~5 m, depending on wind speed in the range we considered (Fig. 7). In Lake Erie, 0.5 dynes cm⁻² approximates an entrainment threshold for silt-dominated surficial sediments (Sheng and Lick 1979). If a similar threshold applies to fine and medium silt in Pettit Lake, wind speeds in the range of recent observation should result in entrainment of sediment across a large fraction of the eastern basin. Suspension mechanics, sediment supply and mass balance considerations thus are all consistent with our storage and transport model for the persistence of this compositional memory effect (Fig. 8).

Fig. 7

Calculated boundary-layer shear stress for Pettit Lake’s eastern basin, following Hamilton and Mitchell (1996) and Laenen and LeTourneau (1996). Water depths cover the range shown for the eastern Pettit Lake basin in Fig. 1. Fetch is measured from lake shoreline along a bearing of 225°, approximating the average wind direction for daily 2-min maxima at the Stanley Ranger Station (NCDC site GHCND: USW00004112). Wind speeds shown bracket the smoothed (7-day window) daily average for the season of maximum winds of about 200 cm s⁻¹. Daily averages of 400 cm s⁻¹ are typically achieved a few times each year during the ice-free season. Daily 2-min maxima average ~500 cm s⁻¹ and occasionally exceed 1,000 cm s⁻¹ during all months of the year. A possible threshold value near 0.5 dynes cm⁻² for re-suspension of silt-dominated materials is suggested by data of Sheng and Lick (1979). Re-suspension across the majority of the eastern basin is likely under a range of expected environmental conditions

Fig. 8

Schematic entrainment and re-distribution of stored sediment dominated by Mazama fallout. Shaded areas are ≤5 m water depth (Budy et al. 1995). Most lakebed shoal area subject to silt entrainment is within the eastern lake embayment

Paleolimnological implications

In Pettit Lake and other lacustrine systems with similarly low watershed clastic sediment delivery, a single event deposit of the magnitude of the distal Mazama tephra can persistently dominate elemental compositions and clastic sediment flux for at least several millennia following the event. This is considerably longer than the duration of most observed effects on diatom community composition and inferred nutrient status (although Bradbury et al. (2004) infer aqueous geochemical effects persisting for a millennium following deposition of the Mazama fallout). The result in Pettit Lake is an effectively modal, before-and-after transformation of elemental sediment composition and grain size distribution. In a setting such as Pettit Lake, this long-term ‘event memory’ underlies highly structured, well-preserved sediment that records limnogeologic history at relatively high temporal resolution, generally with little post-depositional bioturbation. Moreover, the Pettit Lake radiocarbon chronology, derived from selected terrestrial plant remains and fine-grained charcoal, is both systematic and coherent with independently dated, regionally distributed tephras. In this case, then, our moderate-resolution chronology gives no hint of the sediment memory recorded by elemental composition. This lack of a chronological effect can be attributed to selectivity in choosing material for radiocarbon analysis and/or to physical decoupling of the inorganic and organic clastic sediment through transport processes. Nevertheless, analysis of high-frequency Holocene variability in sediment composition must account for modal changes in sediment source composition that result from this extended memory of event deposition in the minerogenic fraction. Our analysis suggests that the interplay of factors including allogenic clastic flux, lake bathymetry, and re-suspension mechanics determine a lake’s sensitivity to these memory effects in the wake of episodic event deposition.