Geochemical controls on lead concentrations in stream water and sediments

The equilibrium distribution of lead in solution and adsorbed on cation exchange sites in sediment theoretically may be calculated from equations representing selectivities of substrate for lead over H⁺, Ca²⁺ and Na⁺, and the stabilities of lead solute species. Such calculations include consideration of total concentrations of major ions, cation exchange capacity (CEC) of substrate, and pH, at values expected in various natural systems. Measurements of CEC and selectivity coefficients were made for synthetic halloysite, a finely divided amorphous 1:1 clay prepared by precipitation from a mixture of solutions of aluminum and silica. Where suspended sediment having the same properties is present in concentrations of 10-1,000 mg/1 at pH 6–8, more than 90% of the lead present can be adsorbed on sediment surfaces. The cation exchange behavior of lead and other minor cationic species in natural systems could be predicted by this type of model if enough other supporting information were available. Information of the type needed describing natural stream sediments, however, is presently inadequate for accurate predictions.     

Selective extraction of organically bound gold in soils, lake sediments and stream sediments

The extraction of organically bound gold from surficial materials is accomplished by using a solution of 5% sodium hypochlorite. A 10-g sample is digested at least twice with 50 mL of hypochlorite solution. The recovery of gold from the organic phase is nearly quantitative for lake sediments and stream sediments high in organic content. For stream sediments and soils low in organics, the recovery of gold is complicated by the sorption of released gold onto inorganic sample components. For these materials, recoveries as low as 70% of the organically bound gold are possible. For a stream sediment sample containing 53.4 ppb gold, the precision of the method is approximately 6% for a 10-g sample and 14% when a 5-g sample of stream sediment is used. The results obtained for all materials tested indicate that a substantial fraction of the gold found in sediments and soils exists as organically bound gold.     

Interpretation of Si concentrations in lake sediments: three case studies

 The biogenic Si concentration in a sediment can be determined as the non-detrital Si concentration by normalization of the total Si concentration with Al. This procedure is based upon the assumptions that (1) that Al exists predominantly in detrital, i.e. minerogenic, particles and (2) that biogenic Si (mainly diatom frustules) is the dominating non-detrital Si phase. This paper focuses on the reasons for the variations of the non-detrital Si concentration in these lake sediments. Data from three lakes are presented, representing three principally different cases regarding the Si concentration. The processes controlling the concentrations of detrital and non-detrital Si are discussed. Received: 10 June 2000 · Accepted: 25 November 2000     

Detrital tin patterns in stream sediments and soils in mid-cornwall

Regional stream sediment sampling in southwest England has revealed extensive anomalous Sn values outside the known Sn mining districts. A detailed investigation of some of these anomalies in the Truro-Mitchell area of Cornwall, has shown that the high Sn concentration was due to detrital cassiterite introduced during previous cycles of erosion. Soil anomalies associated with these detrital cassiterite concentrations are contrasted with those related to known Sn mineralization. Methods of differentiating significant anomalies from those of detrital type on the basis of associated elements are discussed.     

Trace metal concentrations in lake and overbank sediments in southern Norway

 As, Be, Cd, Co, Cr, Cu, Hg, Ni, Pb, V, Se and Zn concentrations were determined and compared in lake and overbank sediments from 33 catchments without local pollution sources in southern Norway. There were no significant differences in concentrations of Be, Co, Cr, Cu, Ni, and V in overbank and pre-industrial lake sediments. In areas with shallow overburden, and significant influence from long-range atmospheric pollution, concentrations of As, Cd, Hg, Pb, Se, and Zn in overbank sediments were probably modified by vertical percolating water. In such areas, we suggest using lake sediments as a better sampling medium for mapping pre-industrial concentrations. Pre-industrial lake sediments yield natural concentrations of Hg and Se, which consist of both geogenic and natural atmospheric deposition. Important covariables like organic carbon content, Fe oxides, and fine mineral fraction were generally higher in pre-industrial lake sediments as compared to overbank sediments. By adjusting for such differences overbank sediments could be used as an alternative in mapping background concentrations of trace metals in regions with few lakes. Received: 19 February 1999 · Accepted: 17 April 1999     

Regional geochemical reconnaissance in the Belgian Ardennes: Secondary dispersion patterns in stream sediments

Two geochemical stream sediment surveys were conducted in the Belgian Ardennes for Fe, Cu, Pb, Zn, Co, Ni and Mn. The geological setting is of quartzo-pelitic rocks of Early Devonian and Cambro-Ordovician age. One survey covers the Ourthe and Salm valleys, the other the Lesse and Lomme drainage basins. Results of these surveys show the geochemical similarity of the two regions. Anomalies have been ascribed to four factors: human activity, Fe-Mn enrichment, known mineralization and potential mineralization. Mean values of metal concentration and statistical dispersion are identical. This results from similar physico-chemical conditions which induce the same types of surficial metallic association. The stream environment is characterized by two main phases, an oxidizing phase where Fe-Mn oxides trap trace metals, and a reducing phase where organic matter and secondary sulphides are bound with Pb, Zn and Cu. In oxidizing conditions, Cu and Ni show a greater affinity for the Fe-oxides while Zn and especially Co are related to the Mn-oxides. Pb has only a weak relationship with the Fe-Mn oxides but is also associated with the organic matter and the secondary sulphides in the reducing environment. Consequently, one may propose the following scale of mobility: Zn > Ni-Mn > Cu-Co > Pb > Fe.     

Heavy metals in bottom sediments of great slave lake (Canada): A reconnaissance

Near the shores of Great Slave Lake, natural sources of heavy metals include gold and base metal deposits, mineralized greenstone belts and sedimentary bedrock, and uraniferous granites. Potential anthropogenic sources of heavy metals (As, Zn, Pb) include large gold and base metal extraction -processing facilities on the shores of the lake. Six sediment cores were collected on a traverse of the lake. Heavy metal concentrations and distributions are related to the regional bedrock geochemistry in the drainage basin. Higher uranium concentrations in the northernmost core are attributed to extensive uraniferous areas north of the lake. More subtle variations of concentration are related to sedimentologic characteristics and processes in the lake. The west basin is a natural sink for most of the heavy metals determined. Two centrally located west basin cores had mean zinc concentrations of 145 ppm, whereas cores closer to the north and south shores had mean concentrations of 80–110 ppm. Mn, Ni, and Pb were enriched in some of the cores from the area of shallower water near the MacKenzie River outlet, rather than in the central west basin. The enrichment is related to Mn-, Ni-, and Pb-rich amorphous coatings on quartz grains. Elevated zinc or lead levels from anthropogenic activities were not detected but elevated levels are suspected for arsenic. In the two cores from the center of the west basin, surface sediment contains up to 12 ppm arsenic, not high in comparison with noncontaminated freshwater lakes elsewhere in Canada but considerably elevated relative to concentrations of 1 ppm found deeper in the same cores. Contribution from the National Water Research Institute, Western and Northern Region. Presented at the 10th International Sedimentological Congress, Jerusalem, Israel, 1978.     

Environmental controls on branched tetraether lipid distributions in tropical East African lake sediments

Quantifying past continental temperature changes is an important aspect of paleoclimate research as it allows us to constrain the amplitude of natural variability, test predictive climate models, and provide a proper context for changes that may arise in response to anthropogenically-induced climate change. The recently developed biomarker-based methylation index of branched tetraethers/cyclization ratio of branched tetraethers (MBT/CBT) proxy shows potential as a new method for continental temperature reconstruction, but thus far it has only been applied successfully in ocean margin sediments. To assess whether this proxy is also applicable to the sedimentary record in tropical lacustrine systems, we investigated the distribution of branched glycerol dialkyl glycerol tetraethers (GDGTs) in recently deposited sediments from 46 lakes in tropical East Africa. These lakes span a substantial range in surface elevation (770-4500 m above sea level), and thus also a wide gradient of mean annual temperature. We find that, saline lakes excepted, branched GDGTs are universally abundant in the lakes investigated and can be used to predict mean annual air temperature (MAAT) with a high degree of accuracy. However, the existing global MBT/CBT calibration for MAAT based on soils predicts inaccurate temperatures when applied to our African lake dataset. This observation, together with the fact that surface water pH, and to lesser extent, lake depth appear to influence the distribution of branched GDGTs among sites, leads us to conclude that in situ production of branched GDGTs in lakes is likely. The robust relationship between branched GDGT distribution and the temperature and pH of African freshwater lakes makes these compounds suitable for paleoenvironmental reconstruction, however we urge caution in using branched GDGTs in lake sediments to infer past temperatures, unless their exact origin can be determined.     

Mercury in lake sediments of the Precambrian Shield near Huntsville, Ontario, Canada

 Long sediment cores (>1 m) were collected from eight Precambrian Shield lakes in southern Ontario, Canada and analyzed for mercury (Hg), loss-on-ignition (LOI), and a suite of 36 other elements. Results indicated at least 100-fold variation in sediment Hg concentrations between lakes in close proximity (from 450 ppb), comparable to the variation reported for lakes across the whole of Canada. Strong areal correlations between Hg concentrations and LOI (r ² =0.77), between Hg and other trace element concentrations (Pb, Zn, Cd, Sb, As, Br), and similarities in the vertical concentration profiles of Hg and LOI, all point to the importance of organic matter in the release, transport and redistribution of metals in watershed systems. The spatial pattern of Hg concentrations in deep, precolonial sediments (>20 cm) was found to mirror the pattern of Hg concentrations in modern surface sediments, an observation that was confirmed in a follow-up survey (r ² =0.85;n=25 lakes), indicating that natural processes govern the unequal distribution of Hg among these lakes. Between-lake differences in surface sediment Hg concentrations normalized to organic carbon (Hg/C) were also reflected by Hg concentrations in smallmouth bass normalized to 35 cm length (R ² =0.63;n=15 lakes). The latter relationship suggests that smallmouth bass and lake sediment indicators provide mutually supportive information regarding Hg loading to the lacustrine environment from geological sources in the watershed system. Received: 31 October 1996 · Accepted: 27 May 1997     

Particle size and abundance of gold in selected stream sediments,southern British Columbia,Canada

Abundance of gold has been determined by neutron activation in six size fractions of sediments from five streams draining known gold occurrences in southern British Columbia. Analytical results and visual examination of heavy mineral separates have then been used to calculate the number of particles of free gold in each fraction. These estimates, considered in relation to sampling statistics based on the Poisson distribution, demonstrate the difficulty of collecting an adequate field sample and reducing it to an acceptable subsample small enough to be analyzed conveniently. More favourable sampling statistics are found with the original minus 53 μ fraction, for which a 20-kg minus 4-mesh (5 mm) field sample will usually suffice to provide a representative sediment sample that can be analyzed directly without the neccessity of preparing a heavy mineral concentrate.     

Rare earth elements (REE) and yttrium in stream waters, stream sediments, and Fe-Mn oxyhydroxides: Fractionation, speciation, and controls over REE + Y patterns in the surface environment

We have collected ∼500 stream waters and associated bed-load sediments over an ∼400 km² region of Eastern Canada and analyzed these samples for Fe, Mn, and the rare earth elements (REE + Y). In addition to analyzing the stream sediments by total digestion (multi-acid dissolution with metaborate fusion), we also leached the sediments with 0.25 M hydroxylamine hydrochloride (in 0.05 M HCl), to determine the REE + Y associated with amorphous Fe- and Mn-oxyhydroxide phases. We are thus able to partition the REE into “dissolved” (<0.45 μm), labile (hydroxylamine) and detrital sediment fractions to investigate REE fractionation, and in particular, with respect to the development of Ce and Eu anomalies in oxygenated surface environments. Surface waters are typically LREE depleted ([La/Sm]_NASC ranges from 0.16 to 5.84, average = 0.604, n = 410; where the REE are normalized to the North America Shale Composite), have strongly negative Ce anomalies ([Ce/Ce^∗]_NASC ranges from 0.02 to 1.25, average = 0.277, n = 354), and commonly have positive Eu anomalies ([Eu/Eu^∗]_NASC ranges from 0.295 to 1.77, average = 0.764, n = 84). In contrast, the total sediment have flatter REE + Y patterns relative to NASC ([La/Sm]_NASC ranges from 0.352 to 1.12, average = 0.778, n = 451) and are slightly middle REE enriched ([Gd/Yb]_NASC ranges from 0.55 to 3.75, average = 1.42). Most total sediments have negative Ce and Eu anomalies ([Ce/Ce^∗]_NASC ranges from 0.097 to 2.12, average = 0.799 and [Eu/Eu^∗]_NASC ranges from 0.39 to 1.43, average = 0.802). The partial extraction sediments are commonly less LREE depleted than the total sediments ([La/Sm]_NASC ranges from 0.24 to 3.31, average = 0.901, n = 4537), more MREE enriched ([Gd/Yb]_NASC ranges from 0.765 to 6.28, average = 1.97) and Ce and Eu anomalies (negative and positive) are more pronounced.The partial extraction recovered, on average ∼20% of the Fe in the total sediment, ∼80% of the Mn, and 21-29% of the REEs (Ce = 19% and Y = 32%). Comparison between REEs in water, partial extraction and total sediment analyses indicates that REEs + Y in the stream sediments have two primary sources, the host lithologies (i.e., mechanical dispersion) and hydromorphically transported (the labile fraction). Furthermore, Eu appears to be more mobile than the other REE, whereas Ce is preferentially removed from solution and accumulates in the stream sediments in a less labile form than the other REEs + Y. Despite poor statistical correlations between the REEs + Y and Mn in either the total sediment or partial extractions, based on apparent distribution coefficients and the pH of the stream waters, we suggest that either sediment organic matter and/or possibly δ-MnO₂/FeOOH are likely the predominant sinks for Ce, and to a lesser extent the other REE, in the stream sediments.     

Regional variations in atmospheric deposition and sources of anthropogenic lead in lake sediments across the Canadian Arctic

Stable Pb isotope profiles in dated lake sediment cores were used to gauge the relative amounts and possible sources of anthropogenic Pb deposited from the atmosphere in different regions of the Canadian Arctic. A distinct north-south difference was found. In four High Arctic lakes (i.e., north of 66°N) in this study, recent Pb isotopic shifts or concentration increases attributable to anthropogenic Pb were negligible. The maximum possible contribution from anthropogenic Pb was 0 to 19% of acid-leachable Pb in the 1980s or 1990s. In contrast, two lakes in the Hudson Bay region displayed significantly lower Pb isotope ratios and threefold to fivefold increases of Pb concentrations in modern sediments, corresponding to anthropogenic Pb inputs of at least 72 to 91% of leachable Pb. Eurasian urban and industrial Pb is known to dominate the High Arctic atmosphere. A possible explanation for its negligible influence on northern lake sediments is that atmospheric Pb deposition at northern latitudes is reduced compared with southern regions and is small compared to local geological inputs. ²¹⁰Pb deposition declines with increasing latitude, apparently because of declining precipitation rates; stable Pb deposition may be similarly affected. Meteorological considerations and variations in the post-1900 Pb isotopic trajectories indicated that the predominant anthropogenic Pb source region in NW Hudson Bay was Eurasia, while in SE Hudson Bay, it was Canada and the United States, with a minor Eurasian component.     

Analcime in lake and lake-margin sediments of the Carboniferous Rocky Brook Formation, Western Newfoundland, Canada

In the Deer Lake Basin of western Newfoundland, Canada, analcime has been found within fine-grained, siliciclastic and carbonate, lake and lake-margin sediments of the Carboniferous Rocky Brook Formation. Analcime is the only zeolite observed in these unmetamorphosed and non-volcanogenic rocks. Microscopic analcime occurs as vug-filling, limpid, isotropic to anisotropic crystals. Microprobe analyses show this type of analcime to be lower in silica than most analcime in sedimentary rocks. A second textural type of finer-grained submicroscopic analcime can be detected from X-ray diffraction spectra of Rocky Brook Formation mudstones, where it is associated with phyllosilicates (illite, interstratified chlorite/smectite, smectite, chlorite), other silicates (quartz, feldspar) and carbonates (dolomite, calcite). Results of this study suggest that the analcime in the Rocky Brook Formation formed either by direct lakewater/porewater precipitation or by reaction of these waters with one or more clay mineral types or plagioclase.     

Contrasting dispersion patterns for gold in stream sediments at Timbarra,NSW, Australia

Regional geochemical exploration for granite-hosted Au mineralisation on the Timbarra Plateau has identified poor correlation between bulk cyanide leach (BCL) Au contents in the wet-screened <2000-μm fraction and aqua regia (AR) extractable Au in the wet-screened <180-μm fraction of stream sediments across most parts of the plateau.At James East, there is a weak response to mineralisation in both the BCL and <180-μm fraction Au contents of stream sediments and poor correlation between these fractions. The total Au content of the sediments is primarily contained within the organic-rich <63-μm fraction. There is little visible Au in panned concentrates from stream sediments or adjacent alluvium. Dispersion of Au appears to be predominantly hydromorphic with minor mechanical contributions from a more diffuse palaeo-colluvial source.At the nearby Surface Hill, there is a strong response to mineralisation in stream sediments with contributions to total Au from a wide range of size fractions. The BCL and <180-μm fraction Au contents are closely correlated. Visible Au is present close to bedrock sources in eluvial and alluvial accumulations. Particle size and Au distribution in the stream sediments indicates dispersion of Au to be dominantly by mechanical processes from the proximal part of a colluvial fan on the margin of the plateau.Variations in the concentration, form and distribution of Au in stream sediments at the two sites can be attributed to variations in surficial conditions and the effects of varying amounts of organic material; however, a complimentary factor is the potential loss of colloidal or organic-adsorbed Au through decanting of sieve water during wet-screening of sediments in the field.     

Assessment of metal concentrations in lake sediments of southwest Japan based on sediment quality guidelines

Geochemical analysis of the bottom sediments of Lakes Banryoko, Onbe, Hamahara and Kijima in Shimane prefecture of southwest Japan was carried out to determine their metal concentrations, and to assess the potential for ecological harm by comparison with sediment quality guidelines. The work conducted includes water quality measurement, and analyses of trace and major elements and rare earth element (REE) of sediments. Results showed that water quality of the lakes contrasts slightly between their upper and lower parts. Average abundances of As, Pb, Zn, Cu, Ni, and Cr in Banryoko sediments were 27, 33, 90, 27, 25, and 46 ppm, respectively, compared to 31, 52, 175, 44, 44, and 75 ppm at Onbe, 11, 26, 96, 13, 13, and 35 ppm at Hamahara, and 24, 43, 193, 31, 12, and 30 ppm at Kijima. These concentrations exceeded the lowest effect level that has moderate impact on aquatic organisms as proposed by the New York State Department of Environmental conservation. Pb and Cu abundances are comparable to the Coastal Ocean Sediment Database threshold, while As and Zn exceed this value, indicating the concentrations of these metals are potentially toxic. Increases in the abundances of these metals in lake sediments are probably related to the reducing condition of the sediments, anthropogenic sources and surface soil erosion. The REE patterns of sediments in the study areas are broadly comparable to the average upper continental crust, but show some contrast between lakes due to differing source litho type. Significant positive correlations between Fe₂O₃ and As, Pb, Zn, and Cu were found in the sediments, suggesting these metals may be adsorbed on Fe oxides in the lake sediments.     

Continental-scale patterns in soil geochemistry and mineralogy: Results from two transects across the United States and Canada

In 2004, the US Geological Survey (USGS) and the Geological Survey of Canada (GSC) initiated a pilot study that involved collection of more than 1500 soil samples from 221 sites along two continental transects across Canada and the United States. The pilot study was designed to test and refine protocols for a soil geochemical survey of North America. The two transects crossed a wide array of soil parent materials, soil ages, climatic conditions, landforms, land covers and land uses. Sample sites were selected randomly at approximately 40-km intervals from a population defined as all soils of the continent. At each site, soils representing 0 to 5 cm depth, and the O, A, and C horizons, if present, were collected and analyzed for their near-total content of over 40 major and trace elements. Soils from 0–5 cm depth were also collected for analysis of organic compounds. Results from the transects confirm that soil samples collected at a 40-km spacing reveal coherent, continental- to subcontinental-scale geochemical and mineralogical patterns that can be correlated to aspects of underlying soil parent material, soil age and climate influence. The geochemical data also demonstrate that at the continental-scale the dominance of any of these major factors that control soil geochemistry can change across the landscape. Along both transects, soil mineralogy and geochemistry change abruptly with changes in soil parent materials. However, the chemical influence of a soil’s parent material can be obscured by changing climatic conditions. For the transects, increasing precipitation from west to east and increasing temperature from north to south affect both soil mineralogy and geochemistry because of climate effects on soil weathering and leaching, and plant productivity. Regional anomalous metal concentrations can be linked to natural variations in soil parent materials, such as high Ni and Cr in soils developed on ultramafic rocks in California or high P in soils formed on weathered Ordovician limestones in central Kentucky. On local scales, anomalous metal concentrations recognized in soil profiles, such as high P in soils from animal confinement sites, are consistent with local anthropogenic disturbances. At a larger scale, the distribution of Hg across the west to east transect demonstrates that it can be difficult to distinguish between natural or anthropogenic contributions and that many factors can contribute to an element’s spatial distribution.Only three samples in a subset of seventy-three 0–5 cm depth soil samples from the north to south transect had organochlorine pesticides values above the method detection limit, apparently related to historic usage of the pesticides DDT and dieldrin.     

Petrographic and geochemical characteristics of organic matter associated with stream sediments in Trail area British Columbia,Canada

Fifty-six samples of stream sediments from 12 creeks in the vicinity of Trail, British Columbia, Canada were examined to determine their origin, characterize their organic matter and their relation to natural/geogenic and anthropogenic sources. The samples were initially screened by Rock-Eval® 6 pyrolysis for their TOC, HI, and OI contents and then examined by both reflected (polarized) and fluorescent light microscopy. It was found that organic matter in stream sediments is mostly from natural/biological sources from local vegetation, such as woody tissue, suberin, spores, and pollen, as well as altered natural/biological input from char formed due to forest fires. Anthropogenic organic matter, mostly coke particles, was also found in the stream sediments. The coke particles have anisotropic properties with medium grained texture formed from medium volatile bituminous coal. The occurrence of coke particles is limited to Ryan Creek located close to an area were some small gold, nickel, and lead smelting operations previously occurred. There is no evidence to indicate that the coke particles found in the creek are emitted from the lead and zinc smelter currently in operation in the area. There are no coal-bearing strata in the area that may have a direct input of coal fragments in any of the creeks.     

Heavy metals in stream sediments: Error versus concentration

A regional geochemical mapping program in southern Norway carried out by the Geological Survey of Norway is based on sampling of stream sediments at road/stream intersections. At each site two individual subsamples have been taken. The subsamples were treated independently during the chemical analysis. This procedure enables an estimation of the reproducibility of the element concentration at each sample site. The paper gives results for the average relative total, sampling and laboratory errors at various levels of metal concentration using data for HNO₃ -soluble Mn, Pb and Cu as examples. For Mn and Pb the relative total error reaches a minimum near the median concentration and increases toward both the high and low concentrations. For Cu the relative total and laboratory errors decrease with increasing concentration throughout the whole concentration range, while the sampling error is fairly constant. The results suggest that the models which are commonly used for error determination in geochemical data, are too simple as far as relationships between error and concentration are concerned.     

Standardization of metal concentrations in sediments using regression residuals: an example from a large lake in Florida, USA

 A new method of standardizing metal concentrations in sediments was tested on samples from Lake Miccosukee, a large karstic lake in north Florida. Metal concentrations were analyzed in 222 sediment samples from 26 cores representing 9 sampling sites in the lake. Measured sedimentation rates in the lake are low. Percent organic matter strongly increases upward in all the cores. The C/N ratio remains constant throughout all the samples, with a mean value of about 13, regardless of depth or location. All of the geochemical variables are at least approximately log-normally distributed; thus, log-log or semi-log scattergrams were used and the data were log-transformed before statistical calculations were performed. Some elements (Mn, Zn, Hg, Cu, and Ca) are primarily associated with the organic fraction; others (La, Cr, Sr, and Ba) are clearly related to the terrigenous fraction; others show affinities for both fractions. Consequently, no bivariate scattergrams or plots of ratio versus depth – commonly used for standardization by plotting or ratioing a reference element (such as Al) to an element of interest – were found to be adequate for standardization of this dataset. The best method for standardization was found to be one based on multivariate (trivariate) linear regression, using log Al and log C as the independent variables (reference elements representing terrigenous and organic fractions, respectively), and the log of the element of interest as the dependent variable. Residuals (deviations) from the best-fit linear surface were then plotted versus depth in the cores to accomplish the standardization. The results indicate that, with the possible exception of Mn at two sites, there is little evidence of anthropogenic input of trace elements to the lake, and most trace-element concentrations in the lake can be considered as valuable baseline information. A significant finding is that different and erroneous conclusions might have been reached if other standardization methods, not based on trivariate regression, had been employed. Received: 28 August 1997 · Accepted: 24 November 1997