Environmental magnetic and magnetic fabric studies in ...

Journal of Paleolimnology 29: 61–78, 2003.  2003 Kluwer Academic Publishers. Printed in the Netherlands.

61

Environmental magnetic and magnetic fabric studies in Lake Waynewood, northeastern Pennsylvania, USA: Evidence for changes in watershed dynamics M.T. Cioppa

1,2,

* and K.P. Kodama 1

1

Department of Earth and Environmental Sciences, Lehigh University, 31 Williams Drive, Bethlehem, PA 18015, USA; 2 Current address: Earth Sciences, University of Windsor, Windsor, Ontario N9 B 3 P4, Canada; * Author for correspondence (e-mail: mcioppa@ uwindsor.ca) Received 10 September 2001; accepted in revised form 25 June 2002

Key words: Environmental magnetism, Magnetic fabric, Watershed dynamics

Abstract An environmental magnetic and magnetic fabric study of sediments collected from Lake Waynewood, a postglacial lake in the Pocono Mountains of Pennsylvania, USA, provided a history of the lake’s watershed dynamics over the past 3500 years. Two 5 m long, Mackereth cores of lake sediments and three watershed soil profiles were analyzed magnetically. Paleosecular variation and 14 C measurements allowed timing of changes in the lake’s watershed which are documented by changes in ARM, SIRM,x, S-ratio, SIRM / x, ARM / x and ARM / SIRM downcore. Prior to 2900 years BP, there is little evidence for allogenic influx. Dramatic changes in mineral magnetic properties and a strong AAR magnetic fabric appear approximately 2900 years BP, suggesting major changes in watershed conditions, either in the hydrologic regime or in vegetative cover. Between 2900 and 1800 yrs BP, large variations in magnetic mineralogy are apparent, whereas about 1800 years BP, a single sediment source began to dominate the magnetic mineralogy. About 100–200 years ago, conditions again changed, probably due to clearcutting and settlement of the watershed. Topsoil erosion appears to have dominated the magnetic signal. S and Mn concentration downcore indicate that there is little evidence for reduction diagenesis having caused the changes observed in magnetic mineral type and concentration, except in the top 10 centimeters of the sediment column.

Introduction The term ‘watershed dynamics’ identifies a systems approach to the changes within a lake ecosystem by examining the flow of materials and energy between different reservoirs in the watershed. Paleomagnetic and rock magnetic techniques provide a way of timing and identifying sediment flux between the soil, bedrock and lake sediment reservoirs within a lake watershed ecosystem framework (Thompson and Oldfield 1986; Versosub and Roberts 1995; Maher and Thompson 1999). The mineral magnetic properties measured downcore in lake sediments may reflect either anthropogenic land use changes or longer term climatic changes, and may be modified by authigenic, diagenetic and biologic magnetic minerals produced

during sediment transport, within the lake basin, and during post-depositional processes. Sediment flux is determined by assessing the allogenic (detrital) magnetic minerals within the lake sediments, and identifying their origin through an examination of possible sources, such as watershed soils and bedrock. Additional information on sediment deposition and flux can be determined if depositional magnetic fabrics are present since they reflect the action of gravitational and hydrodynamic forces on the detrital grains being deposited (Tarling and Hrouda 1993). In still water, the depositional fabric is a simple, strongly oblate fabric confined to the bedding plane. Deposition on an inclined surface may result in a superimposed lineation, parallel to the maximum slope (Tarling and Hrouda 1993). The

62 presence of water currents also result in a weak lineation superimposed on a strong foliated compactional fabric, where the minimum axis is tilted in the flow direction, and the maximum axes are lineated parallel to the flow direction (Rees and Woodall 1977) Paleomagnetic, mineral magnetic parameter and magnetic fabric data can therefore be used to assess changes in watershed dynamics. Paleosecular variation data can constrain the timing of events, mineral magnetic parameter data can be used to determine the variation in concentration and type of magnetic minerals and the sources of magnetic materials in lake sediments, and the magnetic fabric of the sediments can provide information about the hydrodynamic regime. Lake Waynewood, in Pennsylvania (Figure 1), was considered to be an excellent site for examining the effects of watershed dynamics through the combination of lake sediment paleomagnetic, mineral magnetic parameter and magnetic fabric data and possible source paleomagnetic and mineral magnetic parameter data.

Geologic setting Lake Waynewood is a eutrophic lake located in the Pocono Mountains of northeastern Pennsylvania. It is one of three core lakes of Lehigh University’s Pocono Comparative Lakes Program (PCLP), an interdisciplinary research and educational program, designed to investigate the lakes’ biological, ecological, chemical, and physical properties (Moeller et al. 1995). The lakes are located between 418209 and 418309 north latitude, and between 758 and 768 west longitude and were formed during deglaciation in the Valley and Ridge Province and the Appalachian Plateau in northeastern Pennsylvania at 16,400 years BP (Cotter 1983). Lake Waynewood has the largest watershed of the three PCLP lakes and is the only lake with an inlet stream. The present-day topography of the Lake Waynewood watershed indicates the presence of a second, smaller, inlet stream that is not currently flowing. Its history is unknown. Sediment accumulation does not appear to have started until 15,000 yrs BP (Cotter 1983). The Devonian Catskill Formation underlies the Lake Waynewood watershed, and consists of grey, olive grey, and greenish-grey sandstones and conglomerates and minor redbeds (Berg et al. 1980). Lake Waynewood has a drainage area of 7.28

km 2 , an area of 0.28 km 2 , a maximum depth of 12.5 m, and a mean depth of 6.0 m(Moeller et al. 1995) Three types of soil dominate the Lake Waynewood watershed: the Holly silt loam (Ho), the Wellsboro (We) channery loam and the Volusia stony silt loam (V3B). The Holly silt loam is described as a poorly to very poorly drained silt loam found on the flood plains of major rivers and creeks (Martin 1985). It would be the most direct source of sediment into the inlet stream and therefore the lake. The Wellsboro channery loam is a gently sloping moderately well drained soil, developed on reddish glacial till (Martin 1985). The Volusia stony silt loam is a level to gently sloping, poorly drained soil formed in greyish and brownish till (Martin 1985).

Methodology Sampling Two sites were cored in Lake Waynewood, the first (WAY1) within 200 m of the inlet on the northwestern side of the lake, and the second (WAY2) about 200 m to the south, in the deepest part of the lake (Figure 1). A pneumatic Mackereth piston corer was used to collect two non-azimuthally oriented 7. 6-cm diameter cores (5.2 m of sediment recovered in core WAY1, 4.2 m in core WAY2) from Lake Waynewood. The relative orientations of subsamples down core were maintained by inscribing a line the length of the core prior to sectioning. A Geotek MultiSensor Core Logger was used to determine whole core susceptibility, sonic density and gamma ray count. The cores were then split lengthwise and one half was archived. Three non-oriented soil profiles and five bedrock samples were collected, characteristic of the predominant soil types and bedrock in the watershed. The soils were sampled using an AMS soil auger (Figure 1). Twenty-seven subsamples, selected from horizons within the soil profiles were used to characterize the magnetics of the three soil profiles. Pre-weighed 8 cm 3 plastic boxes were used to subsample the lake and soil cores continuously. After all magnetic measurements were completed, the samples were dried and weighed for normalization of the magnetic measurements. Small (10 mm 3 20 mm) plastic tubes of sediment were used for high field (.1.3 T) IRM acquisition experiments: in this case, the plastic box samples were subsampled.

63

Figure 1. Bathymetry and sampling localities in Lake Waynewood, Pocono Mountains, northeastern Pennsylvania. Note the two inlet streams on the northwestern side of the lake. The northernmost stream, with the well-developed Ho loam floodplain is the only permanent stream flowing into the lake today and is the inlet stream discussed in the paper. Localities of cores WAY1 and WAY2 are denoted by * and 1 respectively. Soil sampling localities are denoted by squares and the soil type distribution northwest of the lake is shown. Soil types: WeB, Wellsboro channery loan,V3B,Volusia silt loam; Ho Holly silt loam; MoB, Mardin stony silt loam; OxB Oxford very stony loam. Inset map: North America, showing the location of Lake Waynewood near the east coast.

Magnetic analysis Reconnaissance susceptibility measured on the

Geotek MST indicated a good downcore correlation between the two lake cores; therefore mineral magnetic measurements were conducted predominantly

64 on core WAY1. Measurements on pilot samples every 20 cm included: susceptibility (x), alternating field (AF) demagnetization, anhysteretic remanent magnetization in a peak AF field of 100 mT and a dc biasing field of 0.01 mT (ARM), acquisition of partial anhysteretic remanent magnetization in 10 mT AF steps up to 100 mT (pARM), acquisition of isothermal remanent magnetization (IRM), saturation isothermal remanent magnetization (SIRM), and application of a back IRM at 0.3 T (-IRM -0.3T ). The following magnetic parameter ratios were also determined for the samples: S ratio (-IRM 20.3T / SIRM), SIRM / x, ARM / x and ARM / SIRM. AAR fabric measurements were also made for lake sediment samples. This involved imparting a pARM in nine positions (demagnetizing at 100 mT between each). The ARM was imparted over an AF field of typically 0–60 mT. In a few cases, two sets of ARM measurements were run on the same sample, in fields of 0–50 mT and 50-100 mT. The DC biasing field was 0.01 mT, and the McCabe et al. (1985) technique was used to analyze the data. A total of 39 measurements of AAR were made on core WAY1 and 10 measurements on core WAY2. Susceptibility, ARM, SIRM (in a 4.1 T field), and back IRM at 20.3 T were measured for the remaining samples. Magnetic measurements were also made on soil samples. Susceptibility, ARM, pARM, IRM acquisition, SIRM and back IRM at 0.3 T were made on 17 pilot samples. Measurements on twenty more soil samples included only SIRM (4.1 T) and back IRM, as the S-ratio was determined to be the most characteristic of the properties measured. At the Institute for Rock Magnetism (University of Minnesota), the temperature dependence of susceptibility (2938K - 9738K), temperature dependence of SIRM (208K-3008K) and the frequency dependence of susceptibility (x fd ) were determined on lake sediment and soil samples. Standard stepwise thermal demagnetization (to 9538K) of orthogonal IRMs (applied in 0.3 and 1.3 T fields) was performed on 3 lake sediment samples, two soil samples and a bedrock (Devonian Catskill Formation) sample. Colgate University’s Geofyzika Kappabridge was used to measure magnetic susceptibility for the samples in the plastic boxes. All remanence measurements were made on Lehigh University’s 2-axis CTF Superconducting Magnetometer. A Schonstedt TSD-1 was used for thermal demagnetization, and a Schonstedt GSD-5 tumbling alternating field demagnetizer, with a maximum demagnetizing field of 100 mT,

modified to produce bias fields of 0.01 mT, was used for alternating field demagnetization and pARM application. An ASC Scientific IM-10–30 was used to apply magnetic fields up to 5.5 T for IRM acquisition, SIRM and back IRM measurements. At the Institute for Rock Magnetism, a lakeshore Cryotonics AC Susceptometer was used for high temperature susceptibility measurements and a Quantum design MPMS was used for low temperature SIRM measurements. In order to determine the paleosecular variation downcore for age dating purposes, the sample ChRM directions were determined from the AF demagnetization results using principal component analysis (Kirschvinck 1980). Pilot specimens were demagnetized in 10 mT steps up to 100 mT. Blanket demagnetization of the remaining samples was conducted at 60 mT. Organic carbon analysis Sediment samples (,5 g) from various horizons in the sediment cores were placed in pre-weighed clean ceramic crucibles and dried at 758C. The mass loss on ignition values were determined by ashing the dried sediment samples at 5508C, and calculating the weight of the remaining sediment. Organic carbon content was then determined through the use of a predetermined relationship between loss on ignition and percent organic carbon (Dean 1974): % organic carbon content 5 % ignition loss / 2.13. 14

C Analysis

After determining their organic carbon content, four samples from core WAY1, from depths of 2.0–2.15 m, 2.55–2.65 m, 3.09–3.20 m, and 4.40–4.51 m were sent to Beta Analytic Inc. (Miami, FL) to be 14 C dated. The samples ranged from 200–225 g in size. At Beta Analytic Inc., the samples were pretreated using an acid wash, which dispersed the sediments to increase their surface area and to ensure that no carbonates were present. Conventional radiocarbon dating methods were used. The methods described in Snedecor and Cochran (1989) were used to determine the linear regression equations for calculating age (and error of age) as a function of depth in a core. The program Analyseries (Paillard and Labeyrie 1993) was used to calculate correlation coefficients for time series data, and the program Kaleidagraph (v. 3.0, Synergy Software) was used to calculate linear corre-

65 lation coefficients on intra- and inter-core mineral magnetic properties. Elemental analyses In order to check if reduction diagenesis had affected the magnetic properties of the lake sediments, samples (8 cm 3 ) were collected at depths of 8, 52, 100, 200, 324, 350 and 374 cm for S and Mn analyses. Samples were sent to Chemex Labs, Inc. for analysis. Samples were prepared for analysis by HNO 3 – aqua regia digestion. Mn (in ppm) was determined by atomic absorption spectroscopy and S (in ppt) by Leco furnace-infrared detector.

Results Secular variation analysis AF demagnetization of samples typically revealed a single component magnetization removed between 50 and 80 mT. The magnetically cleaned direction, determined both through AF demagnetization at the 60 mT and through principal component analysis, did not differ from the original NRM direction by more than 58. The secular variation record obtained from the two Lake Waynewood cores (WAY1 and WAY2) are in general agreement, however, there are minor discrepancies (Figure 2). Core WAY1 has a coherent paleosecular variation signal downcore from the sediment-water interface, whereas in core WAY2, the NRM is disturbed in the top 30 cm of sediment, suggesting that the top of this core may have not been recovered intact. Below 1.0 m, the inclinations in WAY2 are shallower than in WAY1 by as much as 208, although the changes in declination are similar, suggesting that one (or both) core liners entered the sediment at an angle. The declination records from the two cores (Figure 2b) are shown in sample coordinates, and the directional trends are in general agreement below 0.3 m. Below 4.2 m in WAY1 and 3.7 m in WAY2, both inclination and declination are very scattered. 14

C results

The 14 C dates on the WAY1 core show that oldest sample (4.45 m depth) is 3260690 yrs BP, and the youngest (2.05 m) is 1860670 yrs BP (Table 1). Calendar calibration data are included in Table 1. A

linear regression analysis of the 14 C ages indicates that the sedimentation rate during this interval was |0.2 cm / yr, and that the regression line intercepts the age axis at approximately 800 years BP.

Mineral magnetic properties of lake sediments The concentration-dependent mineral magnetic parameters from core WAY1 all show very similar trends downcore, as do many of the calculated ratios (Figure 3) (Cioppa 1997). However, SIRM / x shows little variation downcore with values between 1500 and 2000 Am 2 / kg. Using the other mineral magnetic properties, which do vary significantly downcore, the core is divided into five zones. Zone I (0–0.2 m) is characterized by high S-ratios (0.7), high SIRM values, and intermediate ARM values. The ARM / x and ARM / SIRM ratios are low. The boundary between Zones I and II is characterized by sharp decreases downcore in the S-ratio, susceptibility, and SIRM parameters. Zone II (0.2–0.7 m) is characterized by intermediate, decreasing S-ratios (0.3–0.1), low SIRM values, medium to low ARM values, and relatively high ARM / SIRM and ARM / x ratios. Zone III (0.7–2.5 m) is characterized by low Sratios (0.1–0.2), and low SIRM, ARM, ARM / x and ARM / SIRM ratios. The Zone II / Zone IV boundary is characterized by an increase in the S-ratio, susceptibility, ARM and SIRM values. Zone IV (2.5–4.2 m) is characterized by high-amplitude, long wavelength (20–30 cm) variations in S-ratios (0.35–0.75), ARM and SIRM values. The changes in all mineral magnetic parameters are in phase: when the S-ratio is high, so are the SIRM, ARM, ARM / x and ARM / SIRM ratios. The Zone IV/ Zone V transition is marked by sharp decreases in SIRM and ARM. Less significant decreases in susceptibility and S-ratio are observed. Zone V (4.2–5.2 m) is characterized by intermediate S-ratios (0.5–0.6), SIRM, and ARM values. ARM / SIRM ratios are relatively high, and ARM / x ratios are intermediate to high. All magnetic parameters show little variation. Paleosecular variation could not be determined. IRM acquisition curves (Figure 4a) indicate that a low-coercivity mineral saturates at fields , 0.3 T, however, full saturation is not achieved until 4.1 T indicating the presence of an antiferromagnetic mineral. The pARM spectra (Figure 5a) show that the

66

Figure 2. Paleosecular variation in cores from Lake Waynewood. (top) Inclination as a function of depth in cores WAY1 and WAY2 compared to reference inclination from Lac St. Croix. (bottom) Normalized declination in cores WAY1 and WAY2 as a function of depth compared to reference declination from Lac St. Croix.

67 Table 1. Summary of radiocarbon data from core WAY1 Sample

Depth (m)

WA WB WC WD

2.0–2.11 2.55–2.65 3.09–3.20 4.40–4.50

14

C Age (yrs BP) 1860 2110 2400 3260

Error (yrs)

Calendar age

Error (yrs)

80 70 70 90

145 AD 115 BC 410 BC 1515 BC

1235 / 2150 1170 / 2250 140 / 2370 1135 / 2220

Note: Analyses performed by Beta Analytic Inc. (Miami, FL). Error on calendar age is the 95% confidence range for the calendar calibrated age, as calculated by Beta Analytic, Inc.

greatest proportion of low-coercivity magnetic minerals in the lake have coercivities between 10 and 30 mT, indicating grain sizes of 2–5 microns (Jackson et al. 1988), independent of depth in the core. x fd measurements on samples from different zones in WAY1 show Zones I and III have no frequency dependence of susceptibility and thus no superparamagnetic content, whereas Zones II, IV and V do show the presence of superparamagnetic grains. Low temperature SIRM results showing the Verwey transition indicate the presence of magnetite in all five zones (Figure 6). The Morin transition in hematite is observed in Zones III and IV. The presence of goethite is suggested by a sharp peak in the temperature dependence of susceptibility measurements in all zones, similar to one observed in a synthetic goethite / organic carbon mixture (Hanesch 1995). This may indicate the growth of a ferrimagnetic phase from a goethite or lepidocroicite precursor.

Mineral magnetic properties of soils The Ho silt loam has very low S-ratios (0.1–0.2) in the top 20 cm, which increase to about 0.25 to 0.4 below this, however there is a considerable amount of variation (Table 2). The S-ratios in the We soil tend to range from 0.2 to 0.6 with no consistent trend observed. The S-ratios in the V3B soils have high values (0.5–0.7) above 40 cm and decrease to extremely low values (.0) below 40 cm (Table 2). The SIRM values do not decrease as drastically as the S-ratio does. The pARM spectra for the soil samples indicate that the magnetic minerals in the We and V3B samples have coercivities below 30 mT (Figure 5b). However, Ho spectra show that much of Ho’s magnetic material has coercivities between 10 and 70 mT or even higher, indicating a higher proportion of finegrained (, 2 microns) material or the presence of hematite and / or goethite. The IRM acquisition spectra (Figure 4b) indicate that the magnetic minerals

Figure 3. Mineral magnetic properties as a function of depth in core WAY1. Zonation and 14 C depths and ages are shown on the right. Note that all magnetic properties and ratios, except for SIRM / x exhibit very similar properties downcore.

68

Figure 4. Isothermal remanence acquisition as a function of applied field (T) in samples from (A) core WAY1 and (B) watershed soil profiles. Saturation of all samples occurred by 4 T. Magnetization at 4.1 T was used as the SIRM value in all calculations.

present in the various soils saturate in fields between 1 and 4 T. x fd measurements of samples from the Ho topsoil

and subsoil and the V3B topsoil show some frequency dependence of susceptibility, indicating the presence of superparamagnetic minerals, while the

69

Figure 5. Partial anhysteretic acquisition as a function of the AF window over which the field was applied in samples from (A) core WAY1 and (B) watershed soil profiles. Note that all samples in the core, regardless of depth, indicate that the greatest proportion of magnetic grains have coercivities between 10 and 30 mT.

V3B subsoil shows none. Low temperature SIRM spectra from both Ho and both V3B samples do not show strong Verwey transitions (Figure 6). The transition may be repressed by oxidation (Ozdemir and Dunlop 1993). Hematite’s Morin transition was simi-

larly not apparent in the soil samples. High temperature-susceptibility results indicate that the Ho subsoil and topsoil both show the goethite peak at T . 4008C, whereas this peak is muted to absent in the V3B topsoil and subsoil.

70

Figure 6. Low temperature dependence of SIRM in samples from (top) the WAY1 core, and (bottom) watershed soil profiles. The Verwey phase transition in magnetite occurs at |1208K, and the Morin phase transition in hematite occurs at 2608K.

Thermal demagnetization Thermal demagnetization of two orthogonal IRMs (0.3 T and 1.3T) in lake sediment samples from Zones I, III, and IV in core WAY1, the Catskill bedrock and from soil samples (Figure 7) revealed the presence of three magnetic components. A low-coercivity component (IRM applied in 0.3 T field) was unblocked below 2508C in the lake sediment and soil samples. A second low coercivity component was unblocked

between 500 and 5508C in the sediment, soil and bedrock samples, and a higher coercivity component was unblocked between 600 and 6808C (IRM applied in 1.3 T field) in the sediment and soil samples. The 2508C component was seen in all three lake sediment zones. The 5508C phase was most clearly seen in Zone I, and was least apparent in Zone III. All three lake sediment zones showed the high unblocking temperature component. The Catskill Formation bedrock from this area carried only the 5508C phase (7c).

71 Table 2. S-ratios in soil profiles from the Lake Waynewood watershed Floodplain (Ho) Depth (cm) 2 6 10 12 16 22 28 34 40 44

Woodland (V3B) S-ratio 0.21 0.18 0.07 1.09 0.14 0.41 0.17 0.25 0.25 0.39

Depth (cm) 2 4 6 8 12 18 22 30 34 40 44 48

S-ratio 0.76 0.66 0.34 0.55 0.75 0.43 0.39 0.57 0.60 20.01 20.04 20.03

The woodland topsoil (V3B) carries the 5508C phase and a higher coercivity (6508C) phase whereas the subsoil has only the high coercivity 6508C phase. The 2508C phase is clearly apparent in the Ho topsoil (7b). The Ho subsoil carried predominantly the low coercivity 5508C phase.

Anisotropy of anhysteretic remanence (AAR) results There are three major changes in the degree of anisotropy downcore (Figure 8; see Cioppa (1997) for a more complete description). From the sediment-water interface down to 2.5 m depth, the anisotropy generally ranges between 2.5 to 7.5%. From 2.5 to 4.2 meters, the anisotropy is higher, ranging from 7 to 11%. Below 4.2 m, the anisotropy is very low, ranging from 1.5 to 3%. The shape of the fabric is somewhat variable as well. Over half of the measurements in the lowermost section have a prolate fabric (Q . 1), as do the measurements in the upper meter. Between 1 and 4 meters, the fabric is strongly oblate. Figure 9 shows the AAR principal axis directions for the WAY1 core, in geographic coordinates. The fabric in the lowermost section, equivalent to Zone V (Figure 9a), appears to be random. In a few of the samples, the minimum axes are vertical, and may indicate the presence of a very weak compactional fabric. In the middle section (Figure 9b), equivalent to Zone IV, the fabric is consistently oblate: the K min axes are vertical to subvertical, and the K max and K int axes are horizontal, with no overall consistent orienta-

tion. However, a subdivision of the data in this section shows that the samples with lower S-ratios tend to have K max axes that lie in the southwest / northeast quadrant. In the top section (Figure 9c), equivalent to Zones I, II and III, both oblate and prolate fabrics are present, resulting in a strong preferred orientation for the K max axes (lineation).

Organic carbon results The organic carbon content of the sediments ranges from 8 to 14% percent (Figure 10). Organic carbon content increases between 15 and 25 cm (the Zone I / Zone II boundary), and shows a spike over the Zone IV/ Zone V boundary (between 410 and 430 cm).

Elemental analysis results The elemental analyses show nearly constant concentrations of S from depths of 52 cm downcore to 374 cm with values of approximately 0.2% (Figure 11). Mn concentrations vary from 740 ppm at a depth of 52 cm to a peak of 1200 ppm at 324 cm. These data do not support a paleoredox horizon at depths below 52 cm in the core. However, high S concentrations near the sediment water interface (8 cm depth – 0.63% S) and low Mn concentrations (440 ppm) suggest that reduction diagenesis is currently occurring in the sediments within centimeters of the lake bottom.

Discussion

Age determination Paleosecular variation records obtained from the two cores (Figure 2) may be correlated to Lund (1996) summary of Holocene secular variation records across North America to determine ages and sediment accumulation rates. The paleosecular variation record in core WAY1 remains coherent in the higher water content sediments near the sediment-water interface, suggesting that the magnetic minerals carrying the record must be depositional or formed very soon after deposition. Using the separate inclination and declination records (Figure 2), various features in the

72

Figure 7. Thermal demagnetization of IRMs applied in 0.3 and 1.3 T fields for (a) lake sediments from Zones I, III, and IV; (b) the Ho loam topsoil; (c) the Catskill Formation bedrock; and, (d) the V3B stony silt loam topsoil and subsoil. Note the presence of three magnetic phases at 2508C, 500–5508C and 600–6508C.

paleosecular variation records in core WAY1 can be correlated to the age markers of Lund (1996) (Table 3). An age model for core WAY1 was calculated, using all age markers (PSV declination, PSV inclination, and 14 C), The model has with a correlation coefficient of 0.92. However, we note that the 14 C ages generally give ages about 200 years older than the PSV inclination-derived ages from the same parts of the core, and that the 14 C intercept on the age axis occurs at approximately 800 yrs BP. There are several possible reasons for this discrepancy. The most likely possibility is that sediment focusing, in which material from near shore is resuspended and incorporated into younger sediments deeper in the lake could redeposit ‘old’ organic carbon and result in 14 C ages older than the actual age of the sediment. A second possibility is that a magnetization lock-in

depth significantly below the sediment-water interface would cause the PSV record to lag the depositional record, as is observed. However; in core WAY1, the PSV record is coherent within centimeters of the sediment-water interface. In core WAY2, while the NRM is disturbed in the top 30 cm, the NRM directions below this depth are not parallel to the present day field, but correlate to approximately the same depth in the WAY1 core. The coherent PSV signal near the surface in core WAY1 implies that the PSV determined ages should not differ from the 14 C ages by more than the time it took to deposit several centimeters of sediment (,100 years). Since at least 10 cm of sediment was used to determine each 14 C age, the temporal effects of a subsurface lock-in depth should be incorporated within the averaged 14 C age. A third possibility is that small amounts of inorganic

73 of older and younger organic material, causing a ‘smearing’ of the 14 C ages. The depositional magnetization would be destroyed and a post-depositional remanence would record the geomagnetic field directions, which would then appear to be younger than the 14 C age of the sediment. However, anoxic conditions dominate Lake Waynewood’s bottom water, making it unlikely that bioturbation would cause the age discrepancy. The most likely cause of the discrepancy between PSV and 14 C ages is therefore sediment focusing or the inclusion of ‘old carbon’ in the sediment model. The age model, incorporating both PSV and 14 C age determinations, allows ages to be assigned to the boundaries between the zones (Table 4). The extrapolated age at the bottom of the core (5.2 m) is |3550 (6 480) yrs BP. The strong correlation between age and depth shown by the fit of the linear regression model (r 2 5 0.92) implies that the sediment accumulation rate has remained close to constant though the past 3700 yrs, at approximately 0.16 cm / year. The 210 Pb determinations (Lott et al. 1994) suggest a near-surface sedimentation rate of 0.2–0.3 cm / year. Comparison of the linear regression age model to the data reveals that the fit of the regression model is poor in the uppermost meter, and that the age axis intercept is |200 yrs BP. It is likely that the combination of dating methods (PSV and 14 C, which probably contained ‘old carbon’) resulted in underestimates of the sedimentation rate. The 210 Pb age determinations are probably more accurate than our age model for the uppermost section of the core.

Magnetic mineralogy, granulometry and source determination Figure 8. Anisotropy of anhysteretic remanence (AAR) magnetic fabric parameters. (A) Percent anisotropy as a function of depth in core WAY1. Note that three major sections are delineated by changes in the percent anisotropy: between 0 and 2.5 m depth, between 2.5 and 4.2 m depth, and below 4.2 m depth. (B) Flinn plot (Flinn 1965) of lineation vs. foliation.

precipitated calcium carbonate could bias the 14 C ages. Since the pH of Lake Waynewood is generally greater than 7 (Moeller et al. 1995) this could have occurred, although the acid treatment during preparation for 14 C dating should have minimized this possibility. Finally, bioturbation may have caused mixing

Thermal demagnetization isolated three magnetic minerals in the lake sediment: a low-coercivity phase demagnetized by 2508C and another low-coercivity phase, demagnetized by 500–5508C, and a high coercivity phase unblocked at 6808C (Figure 7a). IRM acquisition data indicate that both low coercivity ferrimagnetic and high coercivity antiferromagnetic minerals are present. Low temperature SIRM data show the presence of both the Verwey and Morin transitions in the lake sediments. All these results suggest that magnetite and hematite are present in the lake sediments. Greigite (Fe 3 S 4 ) is the most likely candidate given its low coercivity, however, SIRM / x ratios are low through the core (1500–2000 Am 2 / kg)

74

Figure 9. Anisotropy of anhysteretic remanence magnetic fabric principal directions in core WAY1. Stereonets are shown for each of the anisotropy zones in Figure 8. Circles, crosses and squares indicate the K min , K int and K max axes, respectively. (A) data from 4.2–5.2 m depth. The minimum axes are scattered and no dominant fabric is present. (B) Data from 2.5 to 4.2 m depth. Minimum axes are vertical to subvertical, indicating a depositional / compactional fabric. (C) Data from 0.0 –2.5 m depth. Minimum axes are vertical to subvertical, but the % anisotropy is less than in the section below.

and are more typical of multidomain magnetite (Thompson and Oldfield 1986) than greigite (Snowball and Thompson 1988; Reynolds et al. 1999). Given the large watershed of Lake Waynewood, the presence of an inlet stream, and the depositional magnetic fabric observed throughout most of the core, detrital MD magnetite may dominate the SIRM / x ratio of the lake sediments. Alternatively, a high paramagnetic content may have affected the SIRM / x ratios. Susceptibility versus temperature results suggest the presence of some goethite in the lake sediments, but this is not supported by the high coercivity (1.3 T –IRM) thermal demagnetization data. It is possible that 1.3 T was not a high enough field to give goethite an IRM strong enough to be observed during thermal demagnetization.

In Zone I, the magnetic mineral characteristics include high S-ratios, high SIRM values (Figure 3) as well as very high susceptibility values. The sharp decrease downcore in magnetic concentration parameters (SIRM, x ARM) at the Zone I / Zone II boundary (Figure 3) coincides with an increase in the organic carbon content, suggesting that dissolution due to reduction diagenesis may be occurring near this boundary. The variations in Mn and S concentrations also support this interpretation. The peak in the ARM / x ratio at the Zone I-II boundary may suggest the reduction diagenesis has removed some of the finestgrained magnetic minerals in the top 20 cm of the core. However, the high magnetic mineral concentrations in Zone I are probably due to increased settlement of the watershed in the past 100–200 years

75

Figure 10. Organic carbon content in samples from the WAY1 core. The zone boundaries are marked. Note that the Zone I / Zone II boundary and Zone IV/ Zone V boundaries appear to be delineated by changes in organic matter content. Numbers are calculated from mass loss in ignition at 5508C.

(Lott et al. 1994), although an increase in air pollutant-derived iron cannot be ruled out. Reduction diagenesis has not been able to offset the increase in magnetic mineral input. The unusual ARM / x peak at the Zone I –II boundary could indicate that reduction diagenesis has only recently become important in the surface lake sediments as settlement has caused increased eutrophication (Lott et al. 1994). In Zone II, the low S-ratio indicates the presence of the high coercivity mineral hematite and possibly goethite. Both are allogenic, since they form in oxidizing conditions of soils. The AAR fabric ellipses of samples in Zone II are both prolate and oblate, but the minimum axes are always vertical or subvertical. When the mean declination azimuthal orientation is rotated to coincide with magnetic north, the K max direction is horizontal and oriented N258E. This horizontal lineation is sub-perpendicular to the southeasterly stream flow direction. The bathymetry (Figure 1) indicates that the lake bottom is sloping to the southeast at the sampling site. The lineation direction may therefore result from a combination of current flow and deposition on an inclined surface (Tarling and Hrouda 1993). The presence of a depositional fabric in the AAR further supports a depositional origin for Zone II’s magnetic minerals. In Zone III, the low S-ratios, susceptibility and

Figure 11. S and Mn elemental concentrations downcore in WAY1. Note that S concentrations do not change below 50 cm depth, suggesting no paleoredox boundary at depth in the core. Both Mn and S concentration variations in the top 10 cm of the core suggest that reduction diagenesis is currently occurring at the top of the sediment column.

SIRM values suggest that hematite and possibly goethite are the dominant magnetic minerals. The pARM (Figure 5) spectra suggest that there must be also be a minor contribution from low coercivity material since the spectra peak between 20 and 30 mT. The AAR magnetic fabric throughout this zone is strongly indicative of a depositional origin for the magnetic mineralogy (Figure 9). Although the watershed was not sampled extensively, the analyses on the dominant types of soil in the watershed suggested that the most likely source for the magnetic minerals in Zone III is the Ho loam topsoil and the V3B stony silt loam subsoil, since both have low S-ratios, similar to that observed in Zone III lake sediments. Furthermore, the temperature dependence susceptibility indicate the presence of some goethite in the soils, while thermal demagnetization of the Ho loam topsoil indicates that it also contains greigite or another magnetic iron sulfide. In Zone IV, the magnetic record is dominated by

76 Table 3. Correlation of paleosecular variation features in the WAY1 core to the North American Holocene PSV compilation record of Lind (1996) Depth (m) Inclination 0.6 1.8 3.0 3.5 4.2 Declination 1.4 2.5 3.0 3.3

Correlation feature

Age (yrs BP)

I1 I2 I4 I5 I6

8706180 12506180 18406210 21906230 27356380

D1 D4 D5 D6

9656180 1670660 10606180 25306300

Notes: Correlation Feature is the number given by Lund (1996) to specific correlatable features in the North America Holocene PSV record. Age is the average age of those features (radiocarbon years BP) determined by Lund (1996)

wide swings in all mineral magnetic properties. At one extreme (referred to as E1 periods), the low Sratios, SIRM and susceptibility resemble the mineral magnetic properties in Zone III. At the other extreme (referred to as E2 periods), the high S-ratios, high SIRM and high susceptibility values resemble the mineral magnetic properties in Zone V. Both E1 and E2 periods have depositional AAR fabric and an accurate secular variation record indicating that allogenic input to the lake dominated the magnetic minerals. Percent magnetic anisotropy also increased from Zone IV to Zone III, showing that depositional processes may have become more firmly established as the lake watershed passed from Zone IV to Zone III time. Based on the similarity of magnetic properties, the most obvious sediment source for the E2 periods is the V3B topsoil. The similarity in mineral magnetic properties in Zone III and the E1 intervals suggests a common source, most probably the Ho topsoil. AAR fabric provides additional evidence for an

allogenic source of magnetic minerals during Zone IV time. The grouping of the K max axes in the low-S-ratio sections (E1) show a lineation superimposed on the dominant foliation. Assuming the mean PSV declination should be true north, the fabric was rotated into geographic coordinates, yielding a K max oriented approximately 3458, and a K min oriented approximately 1758, but dipping steeply. Since the inlet stream is located to the northwest of the sampling site, the lineation is subparallel to the stream flow direction, and the minimum axes are tilted in the flow direction (SE). This pattern implies that the fabric observed in the E1 sections is current induced (Rees and Woodall 1977). The fabric in the E2 sections is still oblate, but no preferred orientation of the K max axes is apparent. The lack of a depositional fabric in Zone V and the cessation of a coherent secular variation signal (Figure 2) suggest little detrital input of magnetic minerals in Zone V. The x fd results indicate the presence of superparamagnetic minerals while the presence of a Verwey transition in the low temperature SIRM spectra (Figure 6) strongly suggests the presence of magnetite. The high ARM / SIRM ratio indicates that fine-grained, single domain magnetic minerals are also present in this zone. Therefore, it is likely that the dominant magnetic mineral is authigenic or biogenic magnetite. The Zone IV/ Zone V transition is marked by sharp increases in the values of susceptibility, SIRM, ARM, and S-ratio and a spike in organic carbon content (from 10 to 13%). The boundary could be a paleoredox front, below which the detrital magnetic minerals have been dissolved away. However, organic carbon content does not change appreciably above or below the organic carbon spike and the S concentration data show little change across this boundary. Both data sets offer little support for a paleoredox boundary. The change in magnetic mineral properties is probably due to a major source change in sediment, or a change in the source of the magnetic minerals The organic carbon spike may have been caused by a flushing of the watershed as a hew hydrologic regime was established.

Table 4. Ages of lake sediment magnetic zone boundaries Zone boundary

Dept (m)

Age (yrs BP)

90% confidence limits (yrs)

Zone I / II Zone II / III Zone III / IV Zone IV/ V

0.2 0.7 2.55 4.2

326 648 1807 2901

483 464 430 455

Watershed dynamic model The mineral magnetic parameter data, when combined with the AAR fabric data, suggest the following watershed dynamic model for Lake Waynewood:

77 1. 3550–2900 yrs BP Apparently, little allogenic input into the lake occurred during this interval. The magnetic signal is dominated by authigenic minerals, perhaps bacterially-produced magnetite. Given the presence of superparamagnetic grains, as well as SD grains, the signal may be predominantly carried by bacterial induced mineralization (BIM) magnetite (Bazylinski and Moskowitz 1997). The low allogenic input may have been due to lack of an active inlet stream. 2. 2900–1800 yrs BP Starting about 2900 years ago, conditions in the lake watershed changed dramatically. The detrital component of sedimentation at the sampling site increased, resulting in a depositional fabric with strong anisotropy and a good record of PSV. A change in sedimentation rate is not apparent, but cannot be ruled out given the resolution of the age data. Possible causes for the watershed dynamical change include: 1) establishment of an inlet stream near the coring site or 2) increase in the input stream’s sediment load, perhaps due to stream piracy. The large amplitude variations in mineral magnetic property values during this interval can be explained by intermittent behaviour of the lake’s input stream and / or large fluctuations in lake level. During E1 period conditions, the sediment influx into the lake basin would be derived from the stream as in the succeeding Zone III period when the Ho loam was the principal source. During E2 periods, the Ho loam would no longer be the immediate source of sediments, and the forest topsoil (V3B) would have dominated the lake sediments. The change may have resulted from variations in the flow of the inlet stream: during E2 periods, either the stream stopped flowing or the secondary inlet stream, which no longer functions, may have contributed sediment, decreasing the overall Ho soil contribution to the lake. Alternatively, low lake levels during E1 periods would result in the stream contribution dominating the allogenic signal, whereas the E2 intervals of higher lake levels would result in increased sediment input from the surrounding area. The E1–E2 periods were approximately 100–200 years long, so gradual climatic or vegetation changes in the watershed are an unlikely explanation for the abrupt short term changes in sediment source. 3. 1800–600 yrs BP At about 1800 years BP, the watershed conditions appear to have stabilized. The allogenic input from the stream is dominated

by a single high-coercivity magnetic mineral source, apparently derived from the Ho loam. A depositional magnetic fabric is still present, but in the upper section of the core, it becomes dominated by a prolate fabric, which crosscuts the direction of flow from the current input stream. The fabric may reflect the formation of an inclined delta surface in the lake. The stabilization of magnetic mineral properties may be either due to stability of climactic conditions or the watershed’s vegetative cover. The stability of these conditions allowed firm establishment of the inlet stream floodplain and the Ho loam. 4. 600–100 yrs BP (1350–1900 AD) The paleoenvironmental signal within this interval is obscured by the likelihood that some dissolution of magnetic minerals has occurred. However the magnetic fabric remained the same, reflecting a strong depositional signal and implying that not all of the magnetic minerals were dissolved away. 5. 100 – 0 yrs BP (1900 AD-present day) At the beginning of the 20th century, the land surrounding Lake Waynewood was clearcut (Lott et al. 1994), resulting in high erosional flux, increased sedimentation rate (2–3 cm / yr), and a major change in the magnetic mineral properties of the sediment deposited in the lake.

The watershed dynamics model proposed for Lake Waynewood can only be generally correlated with regional climate change. The lake was probably responding, through the idiosyncrasies of its watershed, to the gradual change from the warm, dry Mid Holocene (8000–5000 yrs BP) to the cooler wetter Late Holocene. This gradual change culminated in the Little Ice Age (600 years BP). Pollen records show that, in northeastern North America, an increase in boreal forest occurred between 5000 and 1000 yrs BP, probably as a result of climate cooling, although in Pennsylvania, the boreal forest expansion appears to have occurred somewhat earlier than Zone IV time (Davis 1983). In western North America, a mountain glacial advance started |3000 yrs BP and ended |2000 yrs BP (Mayewski et al. 1981), similar to Zone IV time. The coincidence of the Zone III-II transition with the start of the Little Ice Age (14 th to 18 th centuries AD) may suggest that colder, wetter conditions ended Zone III time in the lakes’s watershed. This study shows that a combination of mineral magnetic property and magnetic fabric data can pro-

78 vide a sensitive record of changes in the dynamics of a lake’s watershed.

Acknowledgments This work was completed as part of M. Cioppa’s doctoral dissertation at Lehigh University. Stacy Ensminger and Xiaidong Tan conducted parts of the mineral magnetic measurements on core WAY1 as part of a graduate level course at Lehigh University. Discussion with R. Moeller greatly improved the formulation of our ideas about the data.

References Cited Bazylinski D.A. and Moskowitz B.M.Banfield J.F., Nealsons K.H. (eds) 1997. Microbial biomineralization of magnetic iron minerals: Microbiology, magnetism and environmental science. Geomicrobiology: Interactions between microbes and minerals. Min. Soc. Amer. 181–223. Berg T.M., Edmunds W.E., Heyer A.W., Glover A.D., Hoskins D.M., MacLachlan D.B. et al. 1980. Geologic map of Pennsylvania, Pennsylvania Geologic Survey. Cioppa M.T. 1997. Paleo- and environmental magnetic investigations in the Appalachians of Pennsylvania., PhD, Lehigh University pp. 178. Cotter J.F.P. 1983. The minimum age of the Woodfordian deglaciation of northeastern Pennsylvania and northwestern New Jersey, phd, Lehigh University pp. 150. Davis M.A. 1983. Holocene vegetational history of the eastern United States. In: Wright H.E.J. Jr. (ed.), The Holocene. University of Minnesota Press, Minneapolis, pp. 166–181. Dean W.E. 1974. Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition: comparison with other methods. J. Sed. Petr. 44: 242–248. Flinn D. 1965. On the symmetry principle and the deformation ellipsoid. Geol. Mag. 102: 36–42. Hanesch M. 1995. Magnetische Eigenschaften von Boeden (Magnetic Properties of Soils), MSc, Institut fur Allgemaine und Angewandte Geophysik der Ludwig Maximilians Universitaet. pp. 168. Jackson M., Gruber W., Marvin J. and Bannerjee S.K. 1988. Partial anhysteretic remanence and its anisotropy: applications and grain size-dependence. Geophys. Res. Lett. 15: 440–443. King J.W., Banerjee S.K. and Marvin J. 1983. A new rock magnetic approach to selecting sediments for geomagnetic paleointensity studies: Application to paleointensity for the last 4000 years. J. Geophys. Res. 88: 5911–5921. Kirschvinck J.L. 1980. The least-squares line and plane analysis of

paleomagnetic data. Geophys. J. Royal Astron. Soc. 62: 699– 718. Lott A.M., Siver P., Marsciano L.J., Kodama K.P. and Moeller R.E. 1994. The paleolimnology of a small waterbody in the Pocono Mountains of Pennsylvania, U.S.A.: reconstructing 19 th –20 th century specific conductivity trends in relation to changing land use. J. Paleolim. 12: 75–86. Lund S.P. 1996. A comparison of Holocene paleomagnetic secular variation records form north America. J. Geophys. Res. 101: 8007–8024. Maher B.A. and Thompson R. 1999. Quaternary Climates, Environments and Magnetism. Cambridge University Press, Cambridge, pp. 390. Martin G.D. 1985. Soil survey of Wayne County, Pennsylvania. United States Department of Agriculture, Soil Conservation Service. Mayewski P.A., Denton G.H. and Hughes T.J. 1981. Late Wisconsin ice sheets in North America. In: Hughes T.J. (ed.), The Great Ice Sheets. John Wiley and Sons, New York, pp. 67–178. McCabe C., Jackson M. and Ellwood B.B. 1985. Magnetic anisotropy in the Trenton limestone: results of a new technique, anisotropy of anhysteretic susceptibility. Geophys. Res. Lett. 12: 333–336. Moeller R.E., Williamson C,E., Hargraves B.E. and Morris D.P. 1995. Limnology of Lakes Lacawac, Giles and Waynewood 1989–1993: An Introduction to the Core Lakes of the Pocono Comparative Lakes Program. Department of Earth and Environmental Sciences, Lehigh University, pp. 34. Ozdemir O. and Dunlop D.J. 1993. The effect of oxidation on the Verwey transition in magnetite. Geophys. Res. Lett. 20: 1671– 1674. Paillard D. and Labeyrie L. 1993. A ‘‘user-friendly’’ MacIntosh software for rapid correlations of paleomagnetic signals and treatments. Centre des Faibles Radioactivities, PNEDC / INSU. Rees A.I. and Woodall W.A. 1977. The magnetic fabric of some laboratory deposited sediments. Earth Planet Sci. Lett. 25: 121– 130. Reynolds R.L., Rosenbaum J.G.,Van metre P., Tuttle M., Callender E. and Goldin A. 1999. Greigite (Fe3S4) as an indicator of drought – the 1912–1994 sediment magnetic record from White Rock Lake, Dallas, Texas, USA. J. Paleolim. 21: 193–206. Snedecor G.W. and Cochran W.G. 1989. Statistical Methods. Iowa States University Press, Ames, OH, pp. 503. Snowball I. and Thompson R. 1988. An occurrence of greigite in the sediments of Loch Lomond. J. Quat. Sci. 4: 121–125. Tarling D. and Hrouda F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London, pp. 217. Thompson R. and Oldfield F. 1986. Environmental Magnetism. Allen & Unwin, London, pp. 227. Versosub K.L. and Roberts A.P. 1995. Environmental magnetism: Past, present and future. J. Geophys. Res. 100: 2175–2192. Wright H.E.J. 1983. Introduction. In: Wright H.E.J. (ed.), Late Quaternary Environments in the United States, The Holocene Vol. 2. University of Minnesota Press, Minneapolis xi–xvii.