TFJF069-03-60812.tex Paleolimnological assessment of human-induced impacts on Walden Pond (Massachusetts, USA) using diatoms and stable isotopes Dörte Köster,1∗ Reinhard Pienitz,1∗ Brent B. Wolfe,2 Sylvia Barry,3 David R. Foster,3 and Sushil S. Dixit4 1Paleolimnology-Paleoecology Laboratory, Centre d’études nordiques, Department of Geography, Université Laval, Québec, Québec, G1K 7P4, Canada 2Department of Geography and Environmentals Studies, Wilfrid Laurier University, Waterloo, Ontario, N2L 3C5, Canada 3Harvard University, Harvard Forest, Post Office Box 68, Petersham, Massachusetts, 01366-0068, USA 4Environment Canada, National Guidelines & Standards Office, 351 St. Joseph Blvd., 8th Floor, Ottawa, Ontario, K1A 0H3, Canada ∗Corresponding authors: E-mail: doerte.koster.1@ulaval.ca, reinhard.pienitz@cen.ulaval.ca Multi-proxy analysis of a sediment core spanning 1600 years from Walden Pond, Massachusetts (USA), reveals substantial changes in the nutrient status over the past ∼250 years resulting from anthropogenic impacts on the lake and watershed. Following a period of environmental stability from about 430 AD to 1750 AD, the abundance of the diatom Cyclotella stelligera increased, the chrysophyte cyst to diatom ratio decreased, organic content declined, bulk organic δ13C decreased, and bulk organic δ15N increased. These changes coincided with logging in the watershed, and are mainly attributed to an increase in detrital input of inorganic sediment and delivery of dissolved soil decomposition products from the watershed. With the beginning of intensive recreational development of Walden Pond in the early 20th century, oligotrophic diatom species were largely replaced by disturbance indicators and the diatom-inferred lake pH increased by 0.5 units, while the bulk organic carbon and nitrogen stable isotope composition markedly shifted to lower and higher values, respectively. These changes reflect inorganic inputs from erosion related to trails, beaches, and construction, as well as increased nutrient inputs by wastewater seepage into groundwater and extensive recreational usage. Diatom-inferred total phosphorus increased only slightly, probably because oligotrophic species still persist during spring and autumn, when Walden Pond has lower nutrient concentrations due to reduced recreational activity. During the last 25 years, diatom assemblages stabilized, suggesting that management measures have been effective in reducing the rate of eutrophication. Notably, the changes observed over the past 250 years are well beyond the range of natural variability of the past 1600 years, yet the pre-disturbance record provides a useful target for developing additional restoration and conservation measures to ensure future environmental protection of this historical site. Keywords: New England, chrysophytes, pH, total phosphorus, land-use history Introduction The evaluation of anthropogenic impacts on ecosys- tems and the development of effective and appropriate restoration and conservation management strategies re- quire knowledge of natural pre-disturbance ecosystem structure. In New England, lakes have been subject to watershed disturbance since European settlement in the 117 Aquatic Ecosystem Health & Management, 8(2):117–131, 2005. Copyright C© 2005 AEHMS. ISSN: 1463-4988 print / 1539-4077 online DOI: 10.1080/14634980590953743 118 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 17th century, when woodlands were cleared for pasture, housing, and wood harvest (Foster, 1995). Logging, agriculture and urban development in the catchment as well as recreational use have affected water quality in many New England lakes (Siver et al., 1996; Dixit et al., 1999; Francis and Foster, 2001). Walden Pond, Massachusetts, and its catchment have experienced multiple historical anthropogenic im- pacts. The site has attracted much attention since the publication of the American classic Walden by the natu- ral philosopher Henry David Thoreau (Thoreau, 1854; Whitney and Davis, 1986). Despite the public interest in the protection of the lake as a natural and histori- cal monument, few studies exist on the limnology of Walden Pond (Deevey, 1942; Winkler, 1993; Colman and Friesz, 2001) and no comprehensive study of the pre- and post-disturbance limnology of Walden Pond has been conducted. The comparison of historical and recent limnological data suggests that Walden Pond has become more eutrophic since the early 20th century, but there is insufficient data to estimate whether the trophic level has reached a steady state or eutrophica- tion is continuing (Colman and Friesz, 2001). Paleolimnological studies offer a well-established approach for acquiring pre-historical data on lake wa- ter quality (Pienitz and Vincent, 2003). The physical, chemical and biological information preserved in lake sediments provide insight into past events that have oc- curred within the catchment and their effects on the lake environment. Fossil diatoms, which are well preserved in lake sediments, are useful indicators of past envi- ronmental change, including lake trophic status (Moser et al., 1996). By relating limnological data such as to- tal phosphorus (TP) and pH to modern diatom assem- blages, diatom inference models have been developed and subsequently applied to fossil samples in order to reconstruct quantitatively past lake properties (Hall and Smol, 1999). Changes in sedimentary diatom compo- sition have been shown to reflect nutrient enrichment following logging and subsequent urban development in the catchments of New England lakes (Davis and Norton, 1978; Engstrom et al., 1985). Diatom-based quantitative reconstructions have been used to infer in- creased P and nitrogen (N) concentrations (Siver et al., 1999; Dixit et al., 1999) and rising pH and alkalinity (Davis et al., 1994) in New England lakes since Euro- pean settlement. The remains of chrysophytes, a group of fossilizing planktonic algae, are also well preserved in lake sediments and are useful indicators of water quality (Zeeb and Smol, 2001). They are used in pale- olimnological studies to investigate lake acidification and eutrophication (e.g., Siver et al., 1999). Lake sediment carbon and N elemental and stable isotope stratigraphy have been used extensively for as- sessing changes in lake water nutrient balance in re- sponse to the impact of human activities on watersheds. Forest clearance, agricultural activity and urban devel- opment in watersheds, for example, frequently lead to lake eutrophication, which is recorded by changes in the carbon (C) and N elemental and stable isotope compo- sition in lake sediment cores (e.g., Schelske and Hodell, 1991, 1995; Hodell and Schelske, 1998; Brenner et al., 1999; Teranes and Bernasconi, 2000). In some cases, these data also reveal the extent of lake recovery if measures have been taken to reduce nutrient loading, thus serving to evaluate the effectiveness of manage- ment strategies (e.g., Schelske and Hodell, 1991, 1995; Hodell and Schelske, 1998). Here we use a multi-proxy paleolimnological ap- proach to assess the impact of human disturbances and of recently adopted management measures on the nu- trient balance of Walden Pond within the context of natural conditions over the past 1600 years. Our efforts focus on 1) the pollen record to document the human impact on the regional vegetation, 2) lake sediment analysis of C and N elemental and isotope geochem- istry to qualitatively assess nutrient balance history, 3) the fossil diatom record and diatom-based quantitative inferences for pH and TP to assess the biological re- sponse of autotrophic organisms to disturbance, and 4) an evaluation of current management strategies for continued environmental stewardship of Walden Pond based on the results of this study. Study site and land-use history Walden Pond is a 25-ha kettle hole lake in eastern Massachusetts, USA (42◦26.3′N, 71◦20.4′W; Figure 1a). The present climate is temperate with warm summers (mean July temperature: 21.9◦C) and win- ters allowing the formation of ice cover (mean January temperature: −4.1◦C). Presently, 86% of the watershed is covered with pine-oak forest, dominated by Pinus strobus and P. rigida, as well as Quercus rubra, Q. velutina and Q. alba. The lake consists of three main basins, the eastern shallow basin (max. depth: 18 m), the central basin (max. depth: 20 m) and the deep west- ern basin (max. depth: 30.5 m) (Figure 1b). The wa- tershed (38 ha) is small in relation to the surface of the lake and is characterized by sandy glacial outwash soils and glacial till on granitic bedrock (Barosh, 1993). The main water supply (55%) is groundwater, enter- ing the lake at the eastern end and draining an area of 62.2 ha including a small lake. Precipitation contributes Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 119 Figure 1. a) Map of the study region. b) Bathymetric map of Walden Pond with coring site. Modified from Coleman and Friesz (2001). 120 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 45% to the water balance (Colman and Friesz, 2001). The thermocline during summer stratification develops at ca. 15 m water depth. The forested, steep banks of the Walden Pond watershed and a short fetch help to keep the stratification until October or November, when turnover starts. The area was settled at the beginning of the 17th cen- tury. The forest around Walden Pond was partly logged for timber harvest, with increasing intensity from the 17th to the 19th century (Brian Donahue, Brandeis University, pers. comm.). A slight decline in cutting occurred from 1850 to 1920 and a dramatic decline followed due to the donation of 80 acres surround- ing the pond by the owners to the Commonwealth of Massachusetts for conservation and open space recre- ation. A railroad was built in 1844 near the south- western end of the lake. In 1866, an excursion park was built on the shore, which burned down in 1902 and was not rebuilt. In the 1920s, a beach was created at the eastern shore, a bathhouse opened, an amusement park was established, and the number of visitors rose to 500,000 per year. In 1975, the Massachusetts Depart- ment of Environmental Management assumed respon- sibility for managing the park and implemented protec- tion measures including the closure of the amusement park, trail improvement, control of access and improve- ment of facilities. Presently, the watershed is contains a well-developed series of hiking trails, and a public beach on the western shore is intensively used, attract- ing up to 100,000 visitors per month in the summer. According to recent limnological surveys (Table 1), Walden Pond has been characterized as a mostly cir- cumneutral, oligotrophic to mesotrophic lake with low conductivity (Colman and Friesz, 2001). The high- est total phosphorus (TP) concentrations in Walden Pond were measured during summer (e.g., 140 µg l−1, 6 September 1989 and 60 µg l−1, 27 May 1998). No- tably, these high values were measured on combined eastern and western basin samples, whereas the high- est TP values measured between 1997 and 1999 were 22 µg l−1 (23 May 1998) and 19.5 µg l−1 (16 July 1998) in the western basin, thus indicating that the eastern basin has the highest TP concentrations. Nu- trient budgets for the lake indicate that N inputs are dominated by groundwater, which is N-enriched by plume water from the septic leach field of the Walden Pond State Reservation headquarters and the bathhouse (30%), both located on the eastern shore of the lake, and recreational usage (34%) (Colman and Friesz, 2001). Phosphorus inputs are dominated by atmospheric dry deposition, background groundwater, and recreational usage, the latter representing more than 50% of the P load during summer (Colman and Friesz, 2001). The P budget calculations indicate that the present TP load- ing is classified between ‘permissible’ and ‘critical load’ according to the Vollenweider model (Colman and Friesz, 2001). The budgets include ground wa- ter, atmospheric deposition, swimmers, waterfowl, fish stocking, and direct runoff, but exclude internal nutri- ent recycling. Yet, high N and P concentrations near the bottom (max. 0.6 mg l−1and 50 µg l−1, respec- tively) as well as anoxic conditions in the hypolimnion during stratification reveal that nutrients are recycled from the sediments. Comparison with historical dis- solved oxygen data collected in 1939 (Deevey, 1942), which recorded an oxygenated hypolimnion, indicates that eutrophication has occurred since then. In the early 1990s, there have been reports on algae accumulation at the surface and bloom-like conditions (Baystate En- vironmental Consultants, 1995). Methods Sampling and dating On 24 February 2000, a 143 cm long sediment core was taken from the western basin of Walden Pond at a depth of 30 m (Figure 1b), using a simple-tube sur- face corer. The core was cut into 1 cm thick slices and sub-samples were used subsequently for the analy- ses of different proxy indicators and dating. The 210Pb chronology (Binford, 1990) was established for the upper 14 cm of the core and by 14C-AMS dating of bulk sediment (gyttja) at four horizons downcore. The European settlement horizon (ESH) was determined by pollen analysis and has been assigned to the historical recorded date 1635 AD for the foundation of the town Concord (Wheeler, 1967). The complete chronology was calculated by linear interpolation between the 14C dates and by exponential trend lines between the oldest 210Pb date, the ESH and the youngest 14C date. Sediment preparation of sub-samples for pollen analysis followed standard procedures (Faegri and Iversen, 1975). Pollen was counted to a total of 500 tree and shrub grains at 400× magnification. Pollen per- centages are based on total terrestrial pollen grains, ex- cluding aquatics, but including agricultural grains (i.e., herbs). Identification is based on standard taxonomic keys (McAndrews et al., 1973; Moore et al., 1991). Organic matter was measured at 1 cm intervals us- ing standard loss-on-ignition procedures (Dean, 1974). Approximately 1 cm3 of wet sediment was dried overnight at 95◦C and then weighed before and after combustion at 550◦C for one hour. Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 121 Table 1. Limnological characteristics of Walden Pond. Data for 1994 were obtained at four sampling dates (January, May, June, and August, 1994). Samples were taken from the epilimnion (1 m depth) and the metalimnion (10 m depth) and combined from the eastern and western basins (Baystate Environmental Consultants 1995). Data for 1997–99 are measurements in the western basin: epilimnion and metalimnion (0–14 m depth), taken at 11 sampling dates (May–August, 1997; March-September, 1998; February, 1999) (Colman and Friesz, 2001). For calculating minimum (min), maximum (max) and mean values, each sample was used separately, including samples from different depths. The mean N:P ratio for 1997–1999 was calculated from mean total N and mean total P, while minimum and maximum were not calculated because of different sample numbers and depths. The 2001 measurement represents surface water. 1989 Sample Date 12/2001 1997–1999 1994 West & East Sampled Basin West West West & East (combined) No. of Samples 1 >44 16 8 Characteristic Unit Min Max Mean Min Max Mean Min Max Mean Conductivity µS cm−1 93 83 92 n.d. 65 87 75 79 87 84 PH 6.9 6.4 9.4 7.7 6.5 7.3 6.8 6 7.7 7 Turbidity NTU n.d. n.d. n.d. n.d. 0.43 1.7 1.1 n.d. n.d. n.d. Alkalinity µg l−1 n.d. n.d. n.d. n.d. 134 182 153 11 12 11 Total phosphorus µg l−1 5.3 1 22 8.4 10 60 19 10 140 60 Month May April September Ammonia-N mg l−1 0.02 n.d. n.d. n.d. <0.01 0.14 <0.03 0.02 0.02 0.02 Nitrate-N mg l−1 0.03 0.01 0.12 0.04 <0.01 0.05 <0.03 0.02 0.02 0.02 Total Nitrogen mg l−1 0.33 0.1∗ 0.57∗ 0.23∗ 0.21∗∗ 1.1∗∗ 0.41∗∗ n.d. n.d. n.d. Secchi Depth M 8.5 2.5 7.5 5.2 4 9.3 6.8 n.d. n.d. n.d. N:P Ratio 68:1 n.d. n.d. 28:1 8.7:1 112:1 22:1 n.d. n.d. n.d. Chlorophyll a µg l−1 2.2 0.1 8.2 2 2.3 3.73 3.1 n.d. n.d. n.d. ∗Including ammonia. ∗∗Excluding ammonia. n.d. = no data. Sediment sub-samples were analysed at 4 cm in- tervals for bulk organic C and N elemental and sta- ble isotope composition. Samples were acid-washed in 10% HCl to remove carbonates, rinsed with de-ionized water and freeze-dried. Coarse-grained (>500 µm) sediments were removed by sieving. The fine-grained fraction was analysed for organic C and N per- cent and stable isotope composition using a Micro- mass Isochrom continuous flow mass spectrometer equipped with an elemental analyser at the Univer- sity of Waterloo—Environmental Isotope Laboratory. Carbon and N isotope composition are reported in δ-notation, such that δ = [Rsample/Rstd − 1] × 1000 where R is the 13C/12C and 15N/14N ratios in the sample and VPDB and AIR standards, respectively. Analytical uncertainty is ±0.03% for δ13C and δ15N based on re- peated analyses of several samples. Samples for diatom analysis were processed us- ing standard procedures (Pienitz et al., 1995). A min- imum of 500 valves per sample were counted and identified according to Camburn and Charles (2000), Krammer and Lange–Bertalot (1986–1991), and Fallu et al. (2000). The taxonomy of the fossil and the modern samples was matched by personal communication with S. Dixit and by comparing our photographs with figures presented in Camburn et al. (1984–86), as this was the main taxonomic reference used by Dixit et al. (1999) for the training set. Chrysophyte cysts were counted, but not identified, and the scales of synurophytes (a group of chrysophytes) were counted and partly identified on the same slides as the diatoms. The ratio of chrysophyte cysts and scales to diatom frustules are expressed as the percentage of the respective remains related to the sum of diatom frustules (= half of the diatom valve count) and the total count of chrysophyte remains. Fossil diatom assemblages were subdivided into stratigraphic zones by optimal partitioning using the computer program ZONE (S. Juggins, unpublished program; Stephen.Juggins@ncl.ac.uk). The significant number of zones was estimated by the broken stick model (Bennett, 1996). However, this zonation tech- nique separates groups at the midpoint of gradual 122 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 Table 2. Performance of the pH and TP models used for paleolim- nological reconstructions in this study, based on diatom and envi- ronmental data from 82 New England lakes (Dixit et al., 1999). max. Gradient Method r2 RMSE r2jack RMSEP bias pH WA inv. 0.87 0.24 0.82 0.30 0.37 TP WA inv. 0.65 0.19 0.40 0.25 0.36 WA inv. = Weighted averaging with inverse deshrinking; RMSE = root mean squared error of prediction; RMSEP = jacknifed RMSE; r2 = jacknifed coefficient of determination. changes, which does not always reflect the exact tim- ing of the onset of change. Therefore, we assigned the border between Zone II and Zone III to the level where the shift began (12 cm instead of the calculated 10 cm), in order to relate the start of this change to historically recorded human disturbances. For quantitative reconstructions of pH and TP, we used transfer functions based on limnological and sub- fossil diatom data (Dixit et al., 1999) from 82 lakes located in the states of Vermont, New Hampshire, Massachusetts, and Connecticut (Köster et al., 2004). Details concerning numerical analyses of the data set, model development, and comparison of quantitative re- constructions with the instrumental record of Walden Pond are presented in Köster et al., 2004). Diatom- based inference models based on weighted averaging with inverse deshrinking (WAinv) were used for the re- constructions in this study (Table 2), as they represented the simplest models with good statistical performance (Köster et al., 2004). Model development, environmen- tal reconstructions and calculation of sample-specific errors for the fossil samples were carried out using the program C2 (Juggins, 2003). Errors were estimated by jacknifing, an intensive cross-validation procedure. In order to test how well the fossil diatom assem- blages are represented in the model data set, two meth- ods of analog measures were applied. The coefficients of dissimilarity using chord distance (Overpeck et al., Table 3. 14C dates for Walden Pond including AMS dates in years BP (ybp), calibrated dates (cal ybp) and calibrated dates converted to years BC/AD. 2 Sigma 14C Converted to Years BC/AD Depth (cm) ybp Calibrated ybp Lab Error Error Cal ybp + 50 (2000 − Cal ybp + 50) 44–45 785 689 35 80 739 1261 ± 80 60–61 1084 970 35 110 1020 980 ± 110 90–91 1301 1261 36 120 1311 689 ± 120 124–125 1517 1396 38 96 1446 554 ± 96 1985) were calculated using the program ANALOG (H.J.B. Birks, University of Norway, Bergen, Norway and J.M. Line, Cambridge University, Cambridge, UK, unpubl. program). Based on the mean minimum dis- similarity coefficient (DC) within the model data, the 75 and 95% confidence intervals were calculated (Laird et al., 1998). Fossil samples with a DC lower than the 75% confidence interval were deemed to have good analogs in the calibration set, whereas DCs between 75 and 95% indicated poor analogs and DCs outside the 95% interval indicated no analogs (Laing et al., 1999). Furthermore, the ‘goodness of fit’ of the fossil assem- blages to the reconstructed variables pH and TP was evaluated by a CCA with the first axis constrained to one variable and defining the modern and fossil samples as active and passive, respectively. Fossil samples hav- ing residual distances to the first axis outside the 95% confidence interval of the modern samples’ distances were considered to have very poor fit (Birks, 1990). Results The chronology of the sediment core, based on 210Pb and 14C dates as well as the European settlement hori- zon (Table 3, Figure 2), indicates a sedimentation rate of about 0.13 cm yr−1 from ca. 430 to 1200 AD. From ca. 1200 to 1700 AD, sediment accumulation declined to 0.04 cm yr−1, rising to an average rate of 0.09 cm yr−1 over the last 300 years. The organic content of the sediments, inferred from loss-on-ignition, remained relatively constant with val- ues around 40% from ∼430 AD to ∼1500 AD, except at ∼670 AD (100 cm) where a value of 49% was ob- tained (Figure 4a). From ∼1600 AD, the percentage of organic matter began to decline, until it reached about 25% at ∼1960 AD. Afterwards, the proportion of or- ganic matter fluctuated between 25% and 40% with a tendency to higher levels in the most recent sediments. Before European settlement, Quercus (oak), Pinus (pine) and Betula (birch) dominated the pollen stratig- raphy, with about 15% aquatic macrophytes, and no Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 123 Figure 2. Age-depth model for Walden Pond.Dates were established by the 210Pb and 14C methods.The ESH (European settlement hori- zon) was determined by pollen analysis and assigned to the historical recorded date 1635 AD. significant changes were observed (Figure 3). After settlement in the early 17th century, the proportion of Pinus increased from 15 to 25% and the amount of agricultural weeds (including Poaceae, Ambrosia, Rumex, etc.) started to increase. Between ca. 1840 and 1900 AD, the abundance of Quercus decreased by 20 to 30%, Pinus declined and agricultural weeds increased until they reached peak proportions of 28% at 1900 AD. In the early 20th century, Betula showed a maximum of 28% and the herbs declined sharply. Shortly afterwards, declines in Betula coincided with increases of Quercus and Pinus. Fagus (beech) and Tsuga (hemlock) pollen declined gradually from the early 18th century on until present. Between 10 and 8 cm (ca. 1940–1960 AD), the relative abundance of aquatic macrophytes, mainly Isoëtes (Quillworth), declined rapidly from 11 to 1.5%. The C and N elemental and isotope composition of Walden Pond sediments was generally stable until marked changes occurred at the top of the core above ca. 24 cm depth (after ca. 1770 AD) corresponding with the main period of human disturbance in the catchment (Figure 4). Organic C and N contents in sediments be- low these strata were near constant with values of 20 and 2%, respectively (Figure 4b,c). Above this horizon, C and N rapidly declined to values as low as 16.8 and 1.2%, respectively, and then increased to values similar to those observed in the pre-1770 AD interval. A brief increase in % C at 12.5 cm depth (ca. 1910 AD) inter- rupted this general trend and corresponds to a horizon rich in charcoal fragments. Similarly, δ13C and δ15N were also generally stable in the lower strata with val- ues of about −26.5 and 1.5‰, respectively (Figure 4e, f). Around 1770 AD, δ13C values began to decline sub- stantially to −30‰, whereas δ15N values markedly be- gan to increase to 5‰. Unlike the trends observed in the elemental profiles, the C and N isotope values did not return to values observed in the pre-1770 AD interval. Three major periods with differing diatom assem- blages were distinguished for the last ∼1600 years, in- cluding one zone preceding ∼1750 AD and two zones after ∼1750 AD, the latter two corresponding to human disturbance in the watershed (Figure 5). Zone I (140–26 cm, ca. 430 to 1750 AD): In Zone I, the oligotrophic to mesotrophic and circumneutral to acidophilic species Tabellaria flocculosa (Roth) Kütz. str. IIIp sensu Koppen and Cyclotella stelligera (Cleve and Grunow) Van Heurck dominated. The acidobion- tic diatom Asterionella ralfsii var. americana Körner was present at low abundance throughout this period. Fragilaria nanana Lange-Bertalot and Cyclotella bo- danica aff. lemanica (Müller ex Schröter) Bachmann were present with abundances less than 10%. The ratio of synurophyte scales to diatom frustules varied be- tween 9 and 29% (mean: 18%, Figure 5). The ratio of chrysophyte cysts to diatoms ranged from 7 to 21% (mean: 13%), with the highest values (18–20%) around 1700 AD. Zone II (26–12 cm, ca. 1750 to 1920 AD): Around 1750 AD, the abundance of T. flocculosa str. III p started to decline, while C. stelligera increased by about 30%. Cyclotella bodanica aff. lemanica de- creased from around 10% to very low abundance at ∼1860 AD. The synurophyte scales to diatom frus- tules ratio remained constant (mean: 21%), however, the mean relative abundance of chrysophyte cysts de- creased from 13 to 8%. Zone III (12–0 cm, ca. 1920 to 2000 AD): In the early 20th century, diatom assemblage composition showed a striking change. The abundances of the disturbance indicators Asterionella formosa Hassal and Fragilaria nanana increased from about 5% in the late 19th cen- tury to 50 and 35% respectively in the 1960s. Simul- taneously, C. stelligera decreased rapidly, from a max- imum abundance of 77% at ∼1900 AD to an average abundance of 3% after ∼1930 AD whereas C. bodan- ica aff. lemanica and Tabellaria flocculosa str. IIIp re- mained relatively constant at ca. 5% and 15%, respec- tively. Asterionella ralfsii var. americana disappeared around 1940 AD and was not subsequently recorded. From about 1980 to 2000 AD this new diatom assem- blage stabilized. The mean synurophyte scales to diatom frustules ratio increased to 43%, with a maximum of 64% in ∼1998. This increase was mainly caused by much higher abundances of scales of Mallomonas cras- sisquama (Asmund) Fott. The ratio of chrysophyte cysts to diatom frustules remained low with values 124 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 Figure 3. Pollen and spores stratigraphy of selected taxa in Walden Pond sediments.Abundances of aquatic taxa were calculated as percentage of total non-aquatic pollen and were not included in the calculation of pollen percentages.Dashed horizontal lines indicate the major diatom zones (see Figure 5). around 8% until about 1997, and then increased to 10– 20% in the upper three levels. Two major periods with differing analog situations prevail in the sediment core (Figure 5). The samples between 140 cm and 9 cm (ca. 430 to 1950 AD) have dissimilarity coefficients lower than the 75% confidence limit, indicating that the fossil diatom species are well represented in the model data set. With the exception of the samples at 4, 3 and 0 cm, the recent samples between 8 and 0 cm (ca. 1950 to 2000 AD) have dissimilarity coefficients between the 75 and 95% critical values, indicating poor analogs (Figure 5). Similar patterns were obtained by the ‘good- ness of fit’ analysis using CCA. The samples from 140 to 9 cm have a good fit to pH and TP, whereas the levels 8 to 0 cm lay outside the 95% confidence limit, indicating a poor fit of those samples to both variables. Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 125 Figure 4. Loss-on-ignition (a), bulk organic carbon and nitrogen elemental (b, c, d) and stable isotope stratigraphy (e, f) for Walden Pond.Dashed horizontal lines indicate the major diatom zones (see Figure 5). The diatom-inferred pH (DI-pH, Figure 5) remained stable from ca. 430 to 1880 AD, with values fluctuating around 7.3. Afterwards, the DI-pH sharply increased to 7.8 at ca. 1970 AD, after which values stabilized until the present day. Figure 5. Diatom stratigraphy of Walden Pond with major zonations, ratios of chrysophyte scales and cysts to diatom frustules, diatom- inferred pH (DI-pH) and TP (DI-TP) and analysis of dissimilarity using the program ANALOG. The sample-specific error for DI-pH and DI-TP (for TP back-transformed to µg l−1) is indicated by horizontal bars. Dashed vertical lines in the ANALOG graph indicate the 75% and 95% confidence limits.Values lower than the 75% limit indicate good analogs, values between the 75% and the 95% limits indicate poor analogs. The diatom-inferred TP (DI-TP, Figure 5) followed similar trends as DI-pH, but increased with a time lag of about 50 years. DI-TP remained relatively stable with values around 6 µg l−1 in the lower levels until ca. 1930 AD. Thereafter, DI-TP continuously increased in 126 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 recent sediments, reaching a maximum of 10 µg l−1 in the 1970s, and remained stable afterwards. The mag- nitude of these inferred changes, however, remained inside the error range of lower core samples. Discussion Major patterns in sedimentary pollen assemblages (Figure 3) reflect the regional land-use history. High abundances of agricultural weeds and parallel declines of tree species between ca. 1800 and 1900 AD are likely associated with the regional maximum deforestation. Pollen abundances of Fagus and Tsuga also decreased since the 18th century due to land clearance and fire. A peak of the pioneer tree birch in the early 20th century and a subsequent increase of dominant pre-disturbance species of the genera Quercus and Pinus reflect the regional re-growth of forest. The clear decline of aquatic macrophytes (Isoëtes) spores between 1940 and 1960 AD may be related to changes in water transparency. Today, water trans- parency is moderate to high (Secchi depth: 2.5–9 m) thereby permitting growth of aquatic macrophytes, mainly the macroalga Nitella, between 6 and 13 m (Colman and Friesz, 2001). Deevey (1942) observed high water transparency in 1939, with aquatic mosses found at 15.7 m water depth. Therefore a moderate de- cline in transparency may have taken place since then. Fundamental to the interpretation of organic geo- chemical records in lake sediments is knowledge of the origin of the organic matter. In the Walden Pond sediments, the fine-grained organic matter fraction is dominantly of aquatic origin based on the C/N ratios of near 10 throughout the record (cf., Meyers and Lallier- Verges, 1999). The only exception is the charcoal-rich horizon at 12 cm (∼1910 AD), which correspondingly has an elevated C/N ratio of 18.2 and is consistent with organic matter derived from terrestrial sources (Figure 4d). Human disturbance since about 1770 AD evidently has had a dramatic impact on the nutrient balance of Walden Pond. Large changes observed in the C and N elemental and stable isotope stratigraphy are clearly beyond the range of natural environmental variability observed over the last ∼1600 years. Initial removal of catchment vegetation by logging appears to have resulted in an influx of primarily inorganic sediment, based on the decline in % C and % N (Figure 4b, c). Although the decline in organic content could also be interpreted as a reduction in lake productivity, this is less likely given the increase in catchment-derived nu- trient loading that probably occurred during European settlement and which has been suggested elsewhere in the region (e.g., Kaushal and Binford, 1999; Hilfinger et al., 2001). The subsequent increase in organic content at the top of the core may reflect a very recent increase in lake pro- ductivity. Alternatively, erosion protective measures in the catchment may be responsible for a reduction in the amount of inorganic sediment entering the lake or this sample may represent freshly deposited sediment that has not undergone degradation. A number of factors can influence the C and N sta- ble isotope composition of organic matter. For C, these are primarily processes that determine the C isotope composition of lake water dissolved inorganic carbon (DIC), the source of C for aquatic plants. These pro- cesses include the lake productivity-respiration bal- ance, CO2 exchange between the lake water DIC and the atmosphere, and the C isotope composition of DIC derived from soil respiration and bedrock weather- ing in the catchment (Boutton, 1991). The C isotope fractionation between the C source and aquatic plants can also vary depending on dissolved CO2 concentra- tions and HCO−3 uptake (Hodell and Schelske, 1998). Similarly, the N isotope composition of organic mat- ter is dependent on several factors, including produc- tivity, N transformations in the water column (e.g., denitrification, ammonium volatilization), and the N isotope composition of dissolved nutrients (Talbot, 2001). Changes in C and N stable isotope records of lakes that have undergone catchment disturbance are fre- quently attributed to shifts in C and N cycling re- sulting from increased nutrient loading (Teranes and Bernasconi, 2000; Wolfe et al., 2000; Herczeg et al., 2001). Indeed, trends to lower δ13C and higher δ15N values in the initial post-1770 AD interval (Figure 4e, f ) are also most likely associated with increased nutrient supply from removal of vegetation in the catchment and, in particular, increased delivery of 13C-depleted dissolved CO2 and 15N-enriched NO−3 from soil de- composition processes. Alternatively or in addition, productivity-driven enrichment of the dissolved inor- ganic nitrogen (DIN) with 15N may have contributed to the positive shift in δ15N in the lake sediment record. Evidently, similar productivity-driven enrichment of the DIC did not occur in Walden Pond perhaps due to a larger lake water reservoir of DIC compared to DIN. During the past 70 years, these trends have accelerated well beyond the range of natural variability observed over the past 1600 years and are possibly related to 13C- depleted and 15N-enriched wastewater seepage into the lake, although additional analysis of C and N isotope Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 127 composition of the wastewater at Walden Pond is re- quired to test this hypothesis. The changes in fossil diatom and chrysophyte assemblages observed in Walden Pond sediments (Figure 5) are likely related to the human-induced changes in nutrient balance at Walden Pond over the past two centuries, as evidenced by comparison of bios- tratigraphic changes with those revealed by the C and N elemental and stable isotope record. In Zone I (140–26 cm, ca. 430 to 1750 AD) the dominance of the diatom taxa C. stelligera and Tabel- laria flocculosa str. III p and the presence of Asteri- onella ralfsii and C. bodanica aff. lemanica indicate that Walden Pond was an oligotrophic, circumneutral to slightly acidic lake between ca. 430 and 1750 AD. The low nutrient content can be explained by the natu- rally nutrient-poor soils in the catchment and the rela- tively small watershed from which not much terrestrial material is washed into the lake, as reflected by low C/N ratios. The fossil diatom flora consists mainly of planktonic species, which may be expected in a core taken from the deepest part of the lake. As the coring site is located at 30 m depth, there is likely a large area around the coring site that does not receive enough light to support a benthic diatom population. In Zone II (26–12 cm, ca. 1750 to 1920 AD) the de- crease of Tabellaria flocculosa str. III p and the increase of C. stelligera between ca. 1750 and 1900 AD coin- cide with the first known anthropogenic disturbance by timber harvesting in the Walden Pond watershed. The proportion of herbal pollen increased significantly at the same time, from 5 to 30% (Figure 3), indicating an opening of the forest vegetation in the region. Addi- tionally, the decreases in sediment organic matter, % C, and % N (Figure 4a–c) suggest higher mineral inputs through soil erosion caused by logging activities. Cyclotella stelligera similarly increased following forest clearance in the watershed of British Columbia lakes (Laird and Cumming, 2001), in Michigan lakes (Fritz et al., 1993) and in a Swiss lake (Lotter, 2001) as well as in Adirondack lakes (Rhodes, 1991). Logging in the watershed can lead to nutrient enrichment of lakes (Lott et al., 1994; Carignan et al., 2000). Engstrom et al. (1985) and Stoermer et al. (1985) related higher Cy- clotella abundances, including C. stelligera, to nutrient enrichment. Other studies have indicated that C. stellig- era may reflect re-oligotrophication (Clerk et al., 2000) and Brugam (1989) reported its decline after almost complete deforestation of Washington lake watersheds. These results suggest that the increasing abundance of C. stelligera after ca. 1750 AD reflects the removal of vegetation in the catchment of Walden Pond by logging activities, but that its abundance may decline with further nutrient enrichment. The decrease in the chrysophyte cyst : diatom frus- tule ratio between ca. 1750 and 1920 supports the hy- pothesis of a slight nutrient enrichment. Decreasing abundances of chrysophyte cysts relative to diatoms in lake sediments have been proposed to accompany lake eutrophication (Smol, 1985). In Zone III (12–0 cm, ca. 1920 to 2000 AD) the striking change from a diatom assemblage associated with oligotrophic conditions and with moderate catch- ment disturbance, such as discussed above,to one dom- inated by disturbance indicators, such as Asterionella formosa and Fragilaria nanana, is likely related to ad- ditional nutrient supply derived from increased activity in the watershed (i.e., trail and beach development, rail- way construction, wastewater seepage, and other recre- ation activities) over much of the past century. Nutrient- enriched lakes are often characterized by increased abundances of Asterionella formosa (e.g., Pennington, 1981; Bennion et al., 2000) and larger Fragilariaceae, such as Synedra delicatissima W. Smith (Brugam, 1978), Fragilaria crotonensis Kitton (Reavie, 2000; Lotter, 2001), Fragilaria nanana(Reavie et al., 1995), and S. nana Meister (Bennion et al., 2000). The diatom species shift found in recent sediments probably relates to changes in summer conditions. Ac- cording to a phytoplankton record of Walden Pond in 1995, Asterionella formosa and Synedra spp. (likely synonymous with Fragilaria spp.) grow during sum- mer in this lake (Baystate Environmental Consultants, 1995). However, Asterionella and Fragilaria are also common during spring and autumn in lakes through- out North America. The explanation for this difference may be found in the combination of preservation and recreational use of Walden Pond, which is unusual for this region. During most of the year, the lake is al- most undisturbed because of the forested watershed and the absence of private houses, whereas during few summer months there are thousands of visitors to the lake, who contribute approximately 27% of the annual P load and 64% of the annual N load (Colman and Friez, 2001). Continuing nutrient (e.g., P) availabil- ity and sufficient epilimnetic mixing throughout the summer may permit diatoms to form a richer summer population (Sommer et al., 1986). Maximum nutrient concentrations at Walden Pond were measured in sum- mer in the eastern basin, close to the beach (Table 1). Therefore it is possible that recreational activities in summer were the source for nutrients that support a summer diatom population. This was likely different under the oligotrophic conditions that prevailed during 128 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 periods preceding the anthropogenic disturbances, be- cause P would have been largely consumed by the spring blooms. Cyclotella stelligera is also a summer species, yet it is out-competed by Asterionella formosa and Fragilaria spp. which are good competitors for P (Van- Donk and Kilham, 1990) and which have been shown to be favoured by high total N:total P ratios (Inter- landi et al., 1999). The high N:P ratio of 40:1 (average January to August values in the epilimnion and metal- imnion, Baystate Environmental Consultants, 1995), indicates that phytoplankton growth is P-limited in Walden Pond. The proposed association between di- atom disturbance indicators with summer conditions in Walden Pond (see above) requires testing from compre- hensive phytoplankton surveys of the lake over several years and including all seasons. Also, as most nutri- ent inputs enter the lake on the eastern shore, a more detailed analysis of the eastern basin is necessary to completely characterize the lake’s limnology. The higher relative abundance of synurophyte scales in the recent sediments (Figure 5), mainly produced by Mallomonas crassisquama, supports the hypothesis of a nutrient enrichment of Walden Pond. This species also proliferated following nutrient addition to an Ex- perimental Lake in western Ontario, Canada (Zeeb, 2004). Preliminary counts of synurophyte scales in an- other core, taken at Walden Pond in 1979, showed strik- ing changes in species composition at about 1900 AD (John P. Smol, Queen’s University, Kingston, ON, pers. comm.), coincident with the major change in the diatom assemblages in zone III. Taxa associated with undis- turbed conditions and that are now quite rare, such as M. lelymene Harris and Bradley and M. allorgei (Deflandre) Conrad, dominated the pre-disturbance sediments. In the recent sediments, they were replaced by taxa indicating higher nutrient availability, such as M. elongata Reverdin and M. caudate Iwanoff. Thus, the changes in chrysophyte populations also reflect the nutrient enrichment of Walden Pond. The chrysophyte cyst proportions remained lower during the 20th century than during pre-settlement times, indicating a higher trophic state. The upper three samples (1995–2000 AD) showed increasing propor- tions, indicating decreasing nutrient concentrations. The stability in chrysophyte cyst proportions during the early 20th century seems to contradict the increas- ing synurophyte scale:diatom ratio at the same time. The explanation may be that only the Synurophyceae bear scales, such as the dominant M. crassisquama in the recent sediments of Walden Pond. However, cysts are formed by scaled and non-scaled taxa from several chrysophyte groups. It is therefore possible that scales were more abundant due to this scaled species follow- ing nutrient enrichment, but that overall chrysophyte and cyst abundances (consisting mainly of oligotrophic species) remained low. During the last ca. 20 years, diatom assemblages seem to have stabilized, probably indicating a reduced rate of eutrophication. As first management measures for water quality protection were implemented in 1975, these measures may have indeed reduced the nutrient loading to Walden Pond during summer. Analog and Goodness-of-fit analyses have shown that the reconstructions are reliable for all levels, with the exception of levels 0–8 cm. Poor analogs and fit were detected for these uppermost levels, although all dominant fossil species are present in the model. The poor analogs result from high abundances of few species in our fossil samples (e.g., Fragilaria nanana, Asterionella formosa), which is not the case in the train- ing set lakes, where these species are present at lower abundances. This is likely due to higher relative abun- dances of benthic taxa in the training set lakes, which are generally shallower than Walden Pond. We assume that useful ecological parameters (species optimum and tolerance) have been estimated for the models and that they are therefore appropriate for inferring pH and TP in our sediment core. From ca. 430 to 1880 AD (diatom zones I and II), diatom-inferred pH and TP remained constant, reflect- ing the stable diatom assemblages (Figure 5). The DI- TP of 5–6 µg l−1 indicated oligotrophic conditions. The circumneutral character of the dominant species was reflected by the DI-pH, which fluctuates between 7.2 and 7.3. A rapid increase in DI-pH by 0.5 units occurred between ca. 1920 and 1960 AD, reflecting the abrupt change in species composition from zones II to III. A rise in pH may indicate higher primary productivity in the lake. Phytoplankton and macrophytes remove CO2 from the water column by photosynthetic assimilation, thereby driving the bicarbonate balance towards the alkaline end. PH measurements in recent years have shown peak values of 9 and more in the epilimnion during summer (Colman and Friesz, 2001), suggesting a temporally alkaline environment generated by high primary productivity. Alternatively, pH can increase due to higher base cation concentrations (i.e., higher alkalinity). The source of such ions can be diverse: soluble inorganic material (Mg2+, Ca2+) may be washed into the lake by runoff over eroded soils, as suggested by lower organic N and C contents in the sediments beginning in zone II Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 129 and persisting through zone III. An explanation for an increase in alkalinity may also be the higher in-lake productivity attributable to higher nutrient levels from logging and recreational use as discussed above. In- creased deposition of organic material causes oxygen depletion in the hypolimnion of Walden Pond dur- ing late summer and autumn (Baystate Environmental Consultants, 1995; Colman and Friesz, 2001; Köster, unpubl. data) permitting sulphate reduction and den- itrification in the hypolimnion, as well as release of base cations from sediments (Schindler, 1986; Psen- ner, 1988). Manganese, Fe, P and ammonia concentra- tions were high in the hypolimnion near the sediments in late summer (Colman and Friesz, 2001), indicating that alkalinity was in fact generated in the anoxic hy- polimnion and at the sediment/water interface. Higher DI-TP values in the top of the core sup- port the evidence of rising nutrient levels at Walden Pond, which was provided by stable isotopes and chrys- ophytes. However, this change is subtle with 5 µg l−1. Given the drastic changes in diatom species composi- tion, and changes in geochemistry and stable isotope composition that were attributed to an increase in nu- trient loading, the small increase in diatom-inferred TP is surprising. However, the reconstructions corre- spond well with the instrumentally recorded mean an- nual TP levels in the western basin that remained below 10 µg l−1 between 1997 and 1999 (Table 1). Yet, some summer samples during this time show a deviation from the mean (ca. 20 µg l−1; Table 1). The shift in the diatom community of the western basin to one dominated by species typical of nutrient-enriched waters is probably due to these moderate seasonal P maxima, as discussed above. During seasons when Walden Pond is not sub- ject to intense recreational use (e.g., spring and autumn) the lake water returns to lower nutrient concentrations (<10 µg l−1), probably allowing proliferation of di- atoms with low TP optima, such as Cyclotella bodan- ica and Tabellaria flocculosa str. III p. For quantitative reconstructions based on fossil samples that integrate all seasons, this means that species with low-TP optima may partly balance out the effects of species with high TP optima growing during summer. Although the lake is presently considered to be oligo- to mesotrophic and its clear water suggests a ‘natural state,’ the geochemical and diatom records in the sediments of Walden Pond indicate substantial changes in the nutrient balance of the lake due to human impact during the last three centuries. The low suscep- tibility of the lake to exhibit evident signs of eutroph- ication, such as turbid water caused by phytoplankton blooms may be due to its great depth, its small water- shed, and the fact that it is mainly fed by groundwater. Human-derived nutrients added to the lake in the sum- mer may be used up by planktonic organisms, which subsequently transfer the nutrients by sedimentation to the lake bottom. However, increased nutrient inputs to Walden Pond through erosion and recreational use dur- ing much of the 20th century have left a signature in the lake sediments, which have the potential of being re- cycled into the lake water column. Active bacteriolog- ical decomposition of organic matter leads to oxygen- depleted conditions in the hypolimnion of the lake in late summer and autumn, permitting recycling of sed- imentary nutrients to the water column and up-welling during full circulation periods. If high nutrient inputs during summer persist, past and present nutrient loads may lead to further eutrophication of Walden Pond. Stable diatom assemblages and increasing chryso- phyte cyst proportions in the uppermost sediments sug- gest that initial management practices applied in the 1970s have halted the process of eutrophication. Cur- rent preventive measures to reduce nutrient loading in- cluding the monitoring of groundwater down gradient from the septic leach field and of oxygen content in the hypolimnion, storm water runoff control, and public awareness activities (Colman and Friesz, 2001) should therefore continue. Given the difficult interpretation of the available limnological records due to highly vari- able sampling protocols, continued monitoring using standardized sampling and analysis methods are of great importance. If signs of further eutrophication, such as more frequent anoxia in the hypolimnion, are observed, or if nutrient contamination of the groundwa- ter starts to increase, more effective wastewater treat- ment and reduced recreational usage may be neces- sary to protect Walden Pond from further water quality deterioration. Acknowledgements This study was supported by NSF and NSERC grants to D. Foster and to R. Pienitz, respectively, and by logistical support from the Centre d’études nordiques and Harvard Forest. We are grateful for the very help- ful comments of the members of the Paleolimnology- Paleoecology Laboratory at Université Laval and of several anonymous reviewers on earlier versions of the manuscript. We thank Susan Clayden for counting the pollen and John Smol for providing information about the chrysophytes. References Baystate Environmental Consultants, Inc., 1995. Study of trophic level conditions of Walden Pond, Concord, Massachusetts. Com- monwealth of Massachusetts, Department of Environmental 130 Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 Management, Division of Resource Conservation. East Longmeadow, MA. Barosh, P. J., 1993. Bedrock Geology of the Walden Woods. In: E. A. Schofield and R. C. Baron, (Eds. ), Thoreau’s World and Ours: a Natural Legacy. North American Press, Golden, Colorado, pp. 212–259. Bennett, K. D., 1996. Determination of the number of zones in a biostratigraphical sequence. New Phytol. 132, 155–170. Bennion, H., Monteith, D., Appleby, P., 2000. Temporal and geo- graphical variation in lake trophic status in the English Lake District: evidence from (sub)fossil diatoms and aquatic macro- phytes. Freshwat. Biol. 45, 394–412. Binford, M. W., 1990. Calculation and uncertainty analysis of 210Pb dates for PIRLA project lake sediment cores. J. Paleolim. 3, 253– 267. Birks, H. J. B., Line, J. M., Juggins, S., Stevenson, A. C., ter Braak, C. J. F., 1990. Diatoms and pH reconstruction. Philos. T. Roy. Soc. B 327, 263–278. Boutton, T. W., 1991. Stable carbon isotope ratios of natural mate- rials: II. Atmospheric, terrestrial, marine, and freshwater envi- ronments. In: D. C. Coleman and B. Fry (Eds. ), Carbon Isotope Techniques, pp. 173–185. Elsevier, New York. Brenner, M., Whitmore, T. J., Curtis, J. H., Hodell, D. A., and Schelske, C. L., 1999. Stable isotope (δ13C and δ15N) signa- tures of sedimented organic matter as indicators of historic lake trophic state. J. Paleolimn. 22, 205–221. Brugam, R. B., 1978. Human disturbance and the historical devel- opment of Linsley Pond. Ecology. 59, 19–36. Brugam, R. B., Vallarino, J. 1989. Paleolimnological investigation of human disturbance in Western Washington lakes. Arch. Hydro- biol. 116, 129–159. Camburn, K. E., Charles, J. C., 2000. Diatoms of Low-alkalinity Lakes in the northeastern United States. The Academy of Natural Sciences of Philadelphia, Philadelphia. Camburn, K. E., Kingston, J. C., Charles, D. F., 1984–86. Paleoeco- logicaI Investigation of Recent Lake Acidification, PIRLA Diatom Iconograph. Indiana University, Bloomington, USA. Carignan, R., D’Arcy, P., Lamontagne, S., 2000. Comparative im- pacts of fire and forest harvesting on water quality in Boreal Shield lakes. Can. J. Fish. Aquat. Sci. 57, 105–117. Clerk, S., Hall, R., Quinlan, R., Smol, J. P., 2000. Quantitative in- ferences of past hypolimnetic anoxia and nutrient levels from a Canadian Precambrian Shield lake. J. Paleolim. 23, 319–336. Colman, J. A., Friesz, P. J., 2001. Geohydrology and Limnology of Walden Pond, Concord, Massachusetts. U. S. Geological Survey. Water-Resources Investigations Report 01–4137. Northborough, MA. Davis, R. B., Anderson, D. S., Norton, S. A., Whiting, M. C., 1994. Acidity of twelve northern New England (U. S. A. ) lakes in recent centuries. J. Paleolim. 12, 103–154. Davis, R. B., Norton, S. A., 1978. Paleolimnologic studies of human impact on lakes in the United States, with emphasis on recent research in New England. Pol. Arch. Hydrobiol. 25, 99–115. Dean, Jr, W. E., 1974. Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on igni- tion: comparison with other methods. J. Sediment. Petrol. 44, 242–248. Deevey, Jr, E. S., 1942. A re-examination of Thoreau’s “Walden”. Q. Rev. Biol. 17, 1–11. Dixit, S. S., Smol, J. P., Charles, D. F., Hughes, R. M., Paulsen, S. G., Collins, G. B., 1999. Assessing water quality changes in the lakes of the northeastern United States using sediment diatoms. Can. J. Fish. Aquat. Sci. 56, 131–152. Engstrom, D. R., Swain, E. B., Kingston, J. C., 1985. A paleolim- nological record of human disturbance from Harvey’s Lake, Vermont: geochemistry, pigments and diatoms. Freshwat. Biol. 15, 261–288. Faegri, K., Iversen, J., 1975. Textbook of Pollen Analysis, 3rd edition. Hafner Press, Denmark. Fallu, M. -A., Allaire, N., Pienitz, R., 2000. Freshwater diatoms from northern Québec and Labrador (Canada). Bibliotheca Diatomo- logica, vol. 45. J. Cramer, Berlin/Stuttgart. Foster, D. R., 1995. Land-use history and four hundred years of veg- etation change in New England. In: B. L. Turner (Ed.), Global Land Use Change. A Perspective from the Columbian Encounter, pp. 253–321. Consejo Superior de Investigaciones Cientificas, Madrid. Francis, D. R., Foster, D. R., 2001. Response of small New England ponds to historic land use. Holocene 11, 301–311. Fritz, S. C., Kingston, J. C., Engstrom, D. R., 1993. Quantitative trophic reconstruction from sedimentary diatom assemblages: a cautionary tale. Freshwat. Biol. 30, 1–23. Hall, R. I., Smol, J. P., 1999. Diatoms as indicators of lake eutroph- ication. In: E. F. Stoermer and J. P. Smol (Eds.), The Diatoms: Applications for the Environmental and Earth Sciences, pp. 128– 168. Cambridge University Press, Cambridge. Herczeg, A. L., Smith, A. K., Dighton, J. C., 2001. A 120 year record of changes in nitrogen and carbon cycling in Lake Alexand- rina, South Australia: C:N, δ15N and δ13C in sediments. Appl. Geochem. 16, 73–84. Hilfinger, M. F., Mullins, H. T., Burnett, A., Kirby, M. E., 2001. A 2500 year sediment record from Fayetteville Green Lake, New York: evidence for anthropogenic impacts and historic isotope shift. J. Paleolim. 26, 293–305. Hodell, D. A., Schelske, C. L., 1998. Production, sedimentation, and isotopic composition of organic matter in Lake Ontario. Limnol. Oceanogr. 43, 200–214. Interlandi, S. J., Kilham, S. S., Theriot, E. C., 1999. Responses of phytoplankton to varied resource availability in large lakes of the Greater Yellowstone Ecosystem. Limnol. Oceanogr. 44, 668– 682. Juggins, S., 2003. C2 User Guide. Software for ecological and pale- oecological data analysis and visualisation. University of New- castle, Newcastle upon Tyne, UK. Kaushal, S., Binford, M. W., 1999. Relationship between C: N ratios of lake sediments, organic matter sources, and historical defor- estation in Lake Pleasant, Massachusetts, USA. J. Paleolim. 22, 439–442. Krammer, K., Lange-Bertalot, H., 1986–1991. Bacillariophyceae. Gustav Fischer Verlag, Stuttgart/Jena/New York. Köster, D., Racca, J. M. J., Pienitz, R., 2004. Diatom-based infer- ence models and reconstructions revisited: methods, transfor- mations, and species response models. J. Paleolim. 32, 233– 246. Laing, T. E., Rühland, K., Smol, J. P., 1999. Past environmental and climatic changes related to tree-line shifts inferred from fossil diatoms from a lake near the Lena River Delta, Siberia. Holocene 9, 547–557. Köster et al. / Aquatic Ecosystem Health and Management 8 (2005) 117–131 131 Laird, K. R., Cumming, B., 2001. A regional paleolimnological as- sessment of the impact of clear-cutting on lakes from the central interior of British Columbia. Can. J. Fish. Aquat. Sci. 58, 492– 505. Laird, K. R., Fritz, S. C., Cumming, B., 1998. A diatom-based re- construction of drought intensity, duration, and frequency from Moon Lake, North Dakota: a sub-decadal record of the last 2300 years. J. Paleolim. 19, 161–179. Lott, A. -M., Siver, P. A., Marsicano, L. J., Kodama, K. P., Moeller, R. E., 1994. The paleolimnology of a small water body in the Pocono Mountains of Pennsylvania, USA : reconstructing 19th- 20th century specific conductivity trends in relation to changing land use. J. Paleolim. 12, 75–86. Lotter, A. F., 2001. The paleolimnology of Soppensee (Central Switzerland), as evidenced by diatom, pollen, and fossil pigment analyses. J. Paleolim. 25, 65–79. McAndrews, J. H., Berti, A. A., Norris, G., 1973. Key to the Qua- ternary Pollen and Spores of the Great Lakes Region. The Royal Ontario Museum Publications in Life Sciences, Toronto. Meyers, P. A., Lallier-Verges, E., 1999. Lacustrine sedimentary or- ganic matter records of Late Quaternary paleoclimates. J. Pale- olim. 21, 345–372. Moore, P. D., Webb, J. D., Collinson, M. E., 1991. Pollen analysis (2nd edition). Blackwell Scientific Publishers, Oxford. Moser, K. A., MacDonald, G. M., Smol, J. P., 1996. Applications of freshwater diatoms to geographical research. Prog. Phys. Geog. 20, 21–52. Overpeck, J. T., Webb, T., Prentice, I. C., 1985. Quantitative inter- pretation of fossil pollen spectra: dissimilarity coefficients and the method of Modern Analogs. Quat. Res. 23, 87–108. Pennington, W., 1981. Records of a lake’s life in time: the sediments. Hydrobiologia 79, 197–219. Pienitz, R., Vincent, W. F., 2003. Generic Approaches towards water quality monitoring based on paleolimnology. In: M. Kumagai and W. F. Vincent (Eds.), Freshwater Management: Global versus Local Perspectives, pp. 61–83. Springer Verlag, Heidelberg, New York. Pienitz, R., Smol, J. P., Birks, H. J. B., 1995. Assessment of freshwater diatoms as quantitative indicators of past climatic change in the Yukon and Northwest Territories, Canada. J. Paleolim. 13, 21– 49. Psenner, R., 1988. Alkalinity generation in a soft-water lake: wa- tershed and in-lake processes. Limnol. Oceanogr. 33, 1463– 1475. Reavie, E. D., Smol, J. P., Carmichael, N. B., 1995. Post-settlement eutrophication histories of six British Columbia (Canada) lakes. Can. J. Fish. Aquat. Sci. 52, 2388–2401. Reavie, E. D., Smol, J. P., Sharpe, I. D., Westenhofer, L. A., Roberts, A. -M., 2000. Paleolimnological analyses of cultural eutrophica- tion patterns in British Columbia lakes. Can. J. Bot. 78, 873–888. Rhodes, T. E., 1991. A paleolimnological record of anthropogenic disturbances at Holmes Lake, Adirondack Mountains, New York. J. Paleolim. 5, 255–262. Schelske, C. L., Hodell, D. A., 1991. Recent changes in productivity and climate of Lake Ontario detected by isotopic analysis of sediments. Limnol. Oceanogr. 36, 961–975. Schelske, C. L., Hodell, D. A., 1995. Using carbon isotopes of bulk sedimentary organic matter to reconstruct the history of nutrient loading and eutrophication in Lake Erie. Limnol. Oceanogr. 40, 918–929. Schindler, D. W., 1986. The significance of in-lake production of alkalinity. Water, Air, Soil Pollut. 30, 931–944. Siver, P. A., 1999. Development of paleolimnologic inference models for pH, total nitrogen and specific conductivity based on plank- tonic diatoms. J. Paleolim. 21, 45–59. Siver, P. A., Canavan IV, R. W., Field, C. K., Marsicano, L. J., Lott, A.- M., 1996. Historical changes in Connecticut lakes over a 55-year period. J. Environ. Quality 25, 334–345. Siver, P. A., Lott, A. -M., Cash, E., Moss, J., Marsicano, L. J., 1999. Century changes in Connecticut, U. S. A., lakes as inferred from siliceous algal remains and their relationships to land-use change. Limnol. Oceanogr. 44, 1928–1935. Smol, J. P., 1985. The ratio of diatom frustules to Chrysophycean statospores: a useful paleolimnological index. Hydrobiologia 123(3): 199–204. Sommer, U., Gliwicz, Z. M., Lampert, W., Duncan, A., 1986. The PEG-model of seasonal succession of planktonic events in fresh waters. Arch. Hydrobiol. 106, 433–471. Stoermer, E. F., Wolin, J. A., Schelske, C. L., Conley, D. J., 1985. An assessment of ecological changes during the recent history of Lake Ontario based on siliceous algal microfossils preserved in the sediments. J. Phycol. 21, 257–276. Talbot, M. R., 2001. Nitrogen isotopes in palaeolimnology. In: W. M. Last and J. P. Smol (Eds.), Tracking Environmental Change Using Lake Sediments: Physical and Chemical Techniques, Develop- ments in Paleoenvironmental Research, Volume 2, pp. 401–439. Kluwer Academic Publishers, Dordrecht. Teranes, J. L., Bernasconi, S. M., 2000. The record of nitrate uti- lization and productivity limitation provided by δ15N values in lake organic matter—A study of sediment trap and core sedi- ments from Baldeggersee, Switzerland. Limnol. Oceanogr. 45, 801–813. Thoreau, H. D., 1854. Walden; or Life in the Woods. Ticknor and Fields, Boston. VanDonk, E., Kilham, S. S., 1990. Temperature effects on silicon- limited and phosphorus-limited growth and competitive interac- tions among three diatoms. J. Phycol. 26, 40–50. Wheeler, R., 1967. Concord: Climate for Freedom. The Concord Antiquarian Society, Concord. Whitney, G. G., Davis, W. C., 1986. From primitive woods to cul- tivated woodlots: Thoreau and the forest history of Concord, Massachusetts. J. Forest His. 30, 70–81. Winkler, M. G., 1993. Changes at Walden Pond during the last 600 years. In: E. A. Schofield and R. C. Baron (Eds.), Thoreau’s World and Ours: a Natural Legacy, pp. 199–211. North American Press, Golden, Colorado. Wolfe, B. B., Edwards, T. W. D., Duthie, H. C., 2000. A 6000- year record of interaction between Hamilton Harbour and Lake Ontario: quantitative assessment of recent hydrologic disturbance using 13C in lake sediment cellulose. Aquat. Ecosyst. Health Manage. 3, 47–54. Zeeb, B. A., Smol, J. P., 2001. Chrysophyte scales and cysts. In: J. P. Smol, H. J. B. Birks and W. M. Last (Eds.), Tracking Envi- ronmental Change Using Lake Sediments Volume 3. Terrestrial, Algal, and Siliceous Indicators, pp. 203–223. Kluwer Academic Publishers, Dordrecht. Zeeb, B. A., Christie, C. E., Smol, J. P., Findlay, D. L., Kling, H. J., Birks, H. J. B., 2004. Responses of diatom and chrysophyte as- semblages in Lake-227 sediments to experimental eutrophica- tion. Can. J. Fish. Aquat. Sci. 50, 2300–2311.