Colonization of mine tailings by marine invertebrates

Marine Environmental Research 51 (2001) 301±325 www.elsevier.com/locate/marenvrev

Colonization of mine tailings by marine invertebrates E.R. Kline *, M.S. Stekoll Juneau Center, School of Fisheries and Ocean Sciences, University of Alaska Fairbanks, 11120 Glacier Highway, Juneau, AK 99801, USA Received 27 June 1999; received in revised form 4 December 1999; accepted 19 March 2000

Abstract An experiment was conducted to determine if a tailings substrate would inhibit recolonization of benthic macrofauna upon closure of a submarine tailings disposal (STD) operation. Trays of defaunated marine sediment, serving as a reference, and trays of tailings from a proposed gold mine were placed at 21 m depth on the ocean ¯oor to allow colonization via settlement from the water column. Trays of reference sediment and tailings, and cores of ambient sediment were collected 9, 17, and 22 months after tray placement. Probable tray e€ects, which were desirable given the objectives of this study, precluded direct comparison of ecological succession in the tray sediments to the ambient assemblage. The ambient macrofauna assemblage was distinguishable from the reference sediment and tailings assemblages throughout the experiment and displayed more pronounced seasonality. Di€erences between the reference sediment and tailings assemblages were generally insigni®cant, including total abundance, total biomass, number of taxa, average size of individuals, numerically dominant taxa, abundance by ecological guilds, and overall community composition. Upon cessation of STD, recolonization of a benthic macrofauna community should not be inhibited by the presence of these tailings as a benthic substrate. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Benthic ecology; Mining; Ocean disposal; Submarine tailings disposal; Underwater tailings placement; STD; STP; Mine tailings; Colonization; Sediment bioassay

* Corresponding author. Kline Environmental Research, LLC, 1731 50th Street, Somerset, WI 54025, USA. Tel. : +1-715-247-4466; fax: +1-715-247-4466. E-mail address: [email protected] (E.R. Kline).

0141-1136/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0141-1136(00)00105-7

302

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

1. Introduction Tailings are a byproduct of hard rock metal mining operations. They are ®negrained, crushed rock from which valuable metals have been removed through a milling process. Submarine tailings disposal (STD) involves discharging tailings as a slurry, usually below the photic zone, along the bottom of a near-shore, marine habitat (Caldwell & Welsh, 1982; Ellis, Pedersen, Poling, Pelletier & Horne, 1995a; Hay, Murray, & Burling, 1983). The US Environmental Protection Agency determined that STD was prohibited under New Source Performance Standards of the Clean Water Act because the regulations prohibit the discharge of process water into waters of the United States from, among others, gold mines using froth ¯otation. STD was thereby prohibited because all process water cannot be removed from tailings. Exemptions have been considered in the USA and STD is employed in other countries (Ellis, Poling & Bear, 1995b; Hesse & Ellis, 1995; Jones & Ellis, 1995). The primary potential for STD in the USA is in Alaska due to abundant coastal mineral reserves combined with other required coastal features (Coldwell & Gensler, 1993). Numerous questions pertaining to the physical, chemical, and biological aspects of STD must be answered in order to predict the site-speci®c environmental impacts. Potential negative consequences of STD include bioaccumulation of metals (Johansen, Hansen, Asmund & Nielsen, 1991), toxicity from milling reagents and dissolved metals (Leduc, Ruber & Webb, 1975), increased photic zone turbidity (Ellis et al., 1995a), and habitat alteration (Poling, Ellis, Pelletier, Petersen & Hesse, 1993). The risks of these consequences occurring are negligible if: (1) tailings are chemically stable in the receiving environment; (2) toxicity of the liquid portion of the tailings slurry is low; (3) tailings are discharged at depth and settle rapidly; (4) tailings are placed on a soft-bottom habitat; and (5) bottom currents are not strong enough to resuspend tailings (Ellis, Poling & Pelletier, 1994; Kline, 1994; Poling, 1982). However, an unavoidable impact of STD is burial and displacement of benthic organisms (e.g. Kathman, Brinkhurst, Woods & Je€eries, 1983). Most members of the marine invertebrate benthos are infauna, living within or partially within benthic substrate. The majority of marine benthos undergo a meroplanktonic larval stage prior to settlement and metamorphosis to the juvenile and adult forms. Infauna ingest, burrow in, and construct dwellings from sediment. They function to irrigate sediment, retain and process organic matter (Rhoads & Germano, 1986), and serve as the primary food source of many bottom-feeding species (Arntz, 1978; Virnstein, 1977). Because of their intimate exposure, low mobility, and constant contact with the bottom for most of their lives, benthic invertebrates are sensitive indicators of ocean ¯oor disturbances and sediment quality. The extent of burial and displacement of benthos resulting from STD can be on the order of several square kilometers, depending mainly on the rate of discharge, the duration of the operation, and the path of the tailings density current as it meanders over the ocean ¯oor. At any given time during operation, only a small portion of an entire area of deposition is actively impacted. Elsewhere, benthos begin to colonize areas that have stabilized (Ellis et al., 1995a). Benthic impacts

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

303

generally decrease with distance from the discharge point (Brinkhurst, Burd & Kathman, 1987, Ellis et al., 1995a). Upon cessation of an STD operation, the entire tailings deposit is available for colonization. The initial mode of colonization is mainly through settlement of meroplankton (Levin, 1984; Santos & Simon, 1980). The rate of colonization is a function of the natural recruitment potential of an ecosystem and the ability of benthos to inhabit tailings. Informative studies have been conducted on benthic colonization during and after STD using monitoring techniques (Kathman, Brinkhurst, Woods & Cross, 1984; Ellis et al., 1995a, b). Another valuable study compared benthic colonization in tailings relative to marble sand (Taylor, 1986). However, without experimental comparison of colonization in natural marine sediment and tailings, it is dicult to attribute possible post-operational di€erences between benthic communities at impacted and unimpacted sites to the presence of tailings. Rather, a benthic community at an impacted site may re¯ect altered ecological succession (Ellis, 1998; Hughes, 1984; Zajac & Whitlatch, 1982a, b) attributable to the physical disturbance of STD. Experimental comparisons of di€erent sediment types are often conducted in the laboratory by monitoring responses of individuals of several benthic species to each sediment type. Each species is taken to be representative of a broader taxonomic group, and relevance of the response of individuals to populations or communities is assumed. While laboratory experiments are less expensive and less time consuming, in situ experiments that address colonization of di€erent sediment types can be more informative. In situ colonization experiments do not address responses of individual organisms. Rather, they address the in¯uence of responses of individuals on species assemblages. In situ experiments also allow assessment of habitability for more species and life stages than would be feasible in the laboratory. Furthermore, if an in situ experiment is conducted near a proposed STD site, colonizing species are likely to be indigenous to the proposed STD site either as existing benthos or as opportunistic species that would be present to initially colonize a tailings deposit. Finally, in situ experiments o€er added realism through incorporation of ecological interactions and natural variables, and the ability to conduct longer term experiments. These advantages, however, can lead to high variability, requiring numerous replicate experimental units to produce statistically powerful results. The objective of this study was to determine whether the presence of a tailings substrate would inhibit recolonization of benthos upon cessation of an STD operation. This study did not directly address the impact at the onset of STD, which was assumed to be complete defaunation in the path of the tailings density current. Benthic macrofauna were selected as a component of the marine ecosystem that may be among the most sensitive to the presence of tailings substrate due to their direct and constant contact with bottom material. An in situ approach was used to maximize both realism and the number of species that could be used to assess the habitability of the tailings. Methods were purposely biased toward individuals that settle from the water column as larvae, to better simulate the main mode of recolonization after STD and to increase the proportion of organisms that were exposed to tailings through all of their benthic life stages.

304

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

Assemblages of macrofauna that colonized reference sediment and tailings were compared to determine the relative habitability of the tailings. Assemblages that colonized reference sediment and tailings were also compared to the ambient assemblage to determine the successional stage of the tray assemblages relative to ambient conditions and to assess tray e€ects. 2. Materials and methods 2.1. Study site A study site was selected in Auke Bay (134 39.870 W, 58 21.330 N), north of Juneau, Alaska (Fig. 1). The study site was approximately 65 km from the proposed Kensington gold mine. Pilot run tailings that were used in this study were produced using ore from the Kensington mine. Auke Bay is too shallow for modern STD and is not a candidate site. The study site location within Auke Bay was selected because

Fig. 1. Locations of the Auke Bay study site and the Kensington gold mine, the source of tailings for the Auke Bay colonization experiment.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

305

it was accessible, soft-bottomed, low in contaminants relative to industrialized areas and other sites in Auke Bay (Karinen, 1985; Short, Rice, Brodersen & Stickle, 1989), and nearly ¯at at 21 m, a depth that allowed work to be conducted using scuba. A mooring buoy was put in place at the site. Auke Bay has been the subject of extensive physical and biological oceanographic research (Ziemann & Fulton-Bennett, 1990). The climate of Auke Bay is temperate and maritime (Coyle & Shirley, 1990). Numerous streams drain into Auke Bay, contributing to temperature and salinity strati®cation during the non-winter months. Bottom topography is irregular. Depth is approximately 40 m over most of the bay and the mean tidal range is approximately 4 m. 2.2. Collection and preparation of reference sediment Approximately 1500 l of bottom sediment were dredged during 27±28 August, 1994, near the study site, for use as reference sediment. The sediment was processed to remove resident organisms, to homogenize the total volume, and to approximate the grain size of the tailings. The sediment was frozen for 1 week at ÿ20 C and allowed to become anoxic for 1 week at room temperature. The sediment was then suspended in freshwater, sieved through 1-mm mesh, suspended in sand-®ltered seawater, and allowed to settle for 1 day. The seawater was decanted and the sediment was mixed, placed in colonization trays (see later), and refrigerated at 4 C for 6 weeks. To avoid disruption during placement, the trays of reference sediment were frozen at ÿ20 C for 1 day prior to placing them underwater at the study site (Taylor, 1986). 2.3. Production and preparation of tailings Tailings were produced by N.A. Degerstrom, Inc., Spokane, Washington, USA, during August, 1994 from a representative sample of ore from the proposed Kensington gold mine, located 72 km north of Juneau, Alaska (Fig. 1). The Kensington mine has been identi®ed as a candidate for STD (Coldwell & Gensler, 1993). Potassium amyl xanthate and methyl isobutyl carbinol were used as reagents in a pilot scale, froth ¯otation process. The tailings used in this study were from the froth ¯otation process and did not include tailings from the processing of gold mineral concentrate because o€-site processing of concentrate is used for some mines. If concentrate is processed on-site, concentrate tailings may be handled separately from ¯otation tailings. Tailings were shipped in a plastic lined drum and received at the Juneau Center, School of Fisheries and Ocean Sciences, University of Alaska Fairbanks on 27 October, 1994. The drum of tailings was stored outside for 3 weeks and then held at room temperature for 6 days. During the outside storage period, the average daily temperature was 1.1 C, ranging from ÿ4.4 to 5.6 C (National Weather Service, Juneau Forecast Oce). After the storage period, two parts tailings were suspended in ®ve parts sand-®ltered seawater, by volume, to mimic dilution and particle separation during STD. Water and remaining suspended tailings were decanted after

306

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

20 min. The settled tailings were placed in colonization trays (see later). To avoid disruption during placement, the trays of tailings were frozen for 1 day at ÿ20 C prior to placing them underwater at the study site (Taylor, 1986). 2.4. Colonization trays and experimental design Polyethylene trays (8 cm deep, 15 cm diameter) were ®lled to 7 cm depth with reference sediment or tailings and placed within larger trays (10 cm deep, 28 cm diameter; Fig. 2). The space in the larger tray, surrounding the inner tray of reference sediment or tailings, was ®lled to 7 cm depth with reference sediment (439 cm2 outside of inner tray). Lids were placed on the inner and outer trays prior to freezing and underwater placement at the study site. The reference sediment in the outer portion of reference sediment trays was intended to be compared to the reference sediment in the outer portion of tailings trays, in an attempt to detect avoidance behavior. The hypothesis was that if some species avoided the tailings, higher densities of those species may occur in the reference sediment that surrounded the inner tailings trays. The outer trays also served to increase isolation of the inner tray sediment from the ambient sediment with the intent of inhibiting colonization by ambient, adult organisms. Sediment traps were placed among the colonization trays to determine the depth of sediment that accumulated on the reference sediment and tailings during the

Fig. 2. Trays and lids used for the Auke Bay colonization experiment. Trays were placed on the ocean ¯oor at 21 m depth. Lids were placed on trays during tray placement and retrieval to minimize disruption of tray sediment.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

307

experiment. Sediment traps consisted of colonization trays ®lled with cement to the same level as tailings or reference sediment trays. The trays of frozen reference sediment or tailings and the sediment traps were lowered to the sea ¯oor at the study site. Forty-eight trays of tailings and 48 trays of reference sediment were placed at 21 m depth using scuba around the circumference of a 30 m diameter circle. Starting from an initial random point on the circle, the trays were systematically arranged in alternating pairs of each sediment type, forming two concentric circles. At the same time, two sediment traps were placed in pairs at a random location within each quarter section of the circle, bringing the total to 104 trays. Trays were spaced an average of 1.8 m apart from tray center to tray center. Trays were not pressed into the ambient sediment since the intent was to impede colonization in the trays by ambient, adult macrofauna. The lids were removed from all trays on December 11, 1994. Trays were retrieved during three periods: 10±22 September, 1995; 11±17 May, 1996; 29 September±6 October, 1996, approximately 9, 17, and 22 months, respectively, after the lids were removed. Retrieval periods were shortly prior to or shortly after the spring and summer periods of high meroplankton abundance in Auke Bay (Coyle & Paul, 1990). The intent was for the spring samples (17 months) to re¯ect macrofauna that were able to survive over winter in the trays, whereas the fall samples (9 and 22 months) were assumed to consist of proportionally more recently settled macrofauna (Bonsdor€ & OÈsterman, 1985). Immediately prior to retrieval, a lid was placed on the inner and outer tray. Both lids had a 2 cm hole for water displacement that was covered with 250-mm mesh (Fig. 2). Cores of ambient sediment were collected during each sampling period. An unused inner tray (Fig. 2) was used as a coring device to maximize comparability between inner trays and ambient core samples. The coring device, ®tted with 250-mm mesh over a 2 cm hole in the bottom for water displacement, was gently placed, bottom up, on the sediment and pounded until it was ¯ush with the sediment surface. The surrounding sediment was then excavated and the core and coring device were sealed in a plastic bag and carried to the surface. At least 12 each of reference sediment trays, tailings trays, and ambient sediment cores, and two sediment traps were collected during each sampling period. Reference sediment and tailings trays were collected in groups of three or more each along with three cores of nearby, ambient sediment from each quadrat of the circle. One core was collected from inside the circle of trays and two from the outside in each quadrat. 2.5. Sample processing Ten samples of each sediment type (reference, tailings, and ambient) were processed for each sampling period, for a total of 90 samples. Samples were sieved, initially through 250-mm mesh, and ®xed and stained in 10% bu€ered formalin with rose bengal dye. The outer and inner trays were processed separately. Each ®xed sample was resieved through 500-mm mesh and preserved in 70% ethanol. The sediment in the sediment traps was allowed to resettle in the laboratory. The depth

308

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

of sediment in each sediment trap was then measured at several regularly spaced points in each tray. Sediment that had accumulated on the surface of the tailings was processed separately from the underlying tailings for the 17-months samples. This was intended to allow estimation of the portion of macrofauna in direct contact with the tailings, as opposed to macrofauna that were restricted to the ¯occulent, surface layer of settled, natural sediment. Separation was achieved by gently rinsing the ¯occulent layer o€ the tailings after tray retrieval until it was visibly apparent that the remaining sediment consisted almost entirely of tailings. The tailings had settled ®rmly and did not rinse away easily. Since the reference sediment did not settle ®rmly, this technique would not have been e€ective for the reference sediment and was not performed. Except when stated otherwise, data from the ¯occulent layer and the underlying tailings were combined for data presentation and statistical analyses. All animals retained on a 500±mm sieve (macrofauna) were sorted under a dissecting microscope at 64 and counted. Each individual was identi®ed to species when practical using the keys of Blake, Hilbig and Scott, (1996), Fauchald (1977), and Kozlo€ (1987). A reference collection was maintained and will be archived at the University of Alaska Fairbanks. Identi®cations of abundant or problematic polychaete species were con®rmed by L. Harris, Curator of Polychaetes at the Los Angeles County Museum, Los Angeles, CA. Identi®cations of abundant or problematic bivalves were con®rmed by T. Rice, Curator of the Sea and Shore Museum, Port Gamble, Washington, USA. Biomass was determined for the combined individuals of each taxon from each sample. Macrofauna were placed on pre-weighed foil, dried to constant weight at 60 C, combusted at 550 C for 5 h and re-weighed. The loss in weight after combustion was used as an estimate of ash-free dry weight biomass (Crisp, 1971). 2.6. Sediment analysis Tailings and reference sediment were analyzed for organic content and grain size prior to in situ placement. Grain size of ambient sediment was determined for 9-months samples. Organic content of ambient sediment was determined for 17months samples. Reference sediment, tailings, and ambient sediment were dried to constant weight at 60 C and combusted at 475 C for 2 h. The loss in weight of dry sediment after combustion was used as an estimate of organic content (Buchanan, 1984). Grain size distribution was determined by weighing the sediment retained on a graded series of sieves (Buchanan, 1984) followed by analysis of the settling velocity of the particles passing through a 75-mm sieve (American Society for Testing and Materials, 1990). 2.7. Data analysis Assemblage attributes among sediment types for each period were compared using single classi®cation, model I ANOVA. When di€erences were signi®cant (P40.05), Tukey's tests were conducted for the three pairwise comparisons. Assemblage

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

309

attributes that were analyzed in this manner included total abundance, total biomass, number of taxa, average size of an individual (biomass/abundance), average size of individuals of the four most abundant taxa, abundance by feeding type, and abundance by sediment association type. Abundance and biomass data were transformed to (x+0.5)0.5 and log10(x+1), respectively, prior to analysis. Back transformed means and 95% con®dence intervals were reported. These same statistical methods were used to analyze di€erences in sediment organic content. Organic content data were transformed to arcsin (x)0.5 prior to analysis. Data transformations were standard for analysis of abundance, biomass, and percentages (Elliot, 1977; Zar, 1996). Cumulative frequency plots were made and interpreted to assess data normality after transformation. The transformations improved the normality of the data considerably but to varying degrees. In the interest of treating and analyzing data in a consistent manner, the above transformations and statistical methods were adhered to. Correspondence analysis was used to graphically display the relationship among the assemblages in the three sediment types for each period based on raw abundance in a taxon by sample matrix. The distance between points, each point representing a sample, re¯ected the overall similarity in the taxonomic composition, weighted by abundance (Greenacre & Underhill, 1982; Greenacre & Vrba, 1984). As such, large di€erences in abundance of a few numerically dominant taxa, or numerous di€erences in the presence of rare taxa, could each have contributed to discriminating between samples. Sample identi®cation was used only to indicate the sediment type of the plotted points. Otherwise, no distinction of sediment type was made in the analyses. After the analysis for each period, 95% con®dence regions were generated around the plotted points for each sediment type. All statistical analyses were performed using SYSTAT 7.0 for Windows (SPSS Inc., 1997). For classi®cation of taxa into functional groups, published literature and taxonomic keys were searched for general ecological descriptions of each taxon. Feeding types included subsurface deposit, surface deposit, suspension, carnivore, or combinations of these feeding types for taxa that switch feeding strategies. Herbivores and scavengers were grouped as other, due to low abundance. Sediment association types included burrowers, tube builders, and surface dwellers. When species information was not found, generalizations pertaining to genus or family were used, provided the description pertained to soft-bottom dwelling members of the taxon. Taxa that were described as burrowing in addition to displaying surface association or errant behavior were classi®ed as burrowers. Taxa for which no functional information was found were grouped as unclassi®ed. The relationships between statistical power and minimum detectable di€erence and between power and sample size were based on the formula: s 2ks2 2 ˆ n where d is the minimum detectable di€erence, k is the number of groups (3), s2 is the mean square error using raw data, n is the sample size (10), and  is a quantity

310

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

incorporating k, n, and the probabilities of committing a Type I and Type II error (Zar, 1996). The quantity d was divided by the raw reference mean and multiplied by 100 to obtain the minimum detectable di€erence as a percent of the reference mean. 3. Results 3.1. Sediment analysis The organic content of tailings was signi®cantly lower (P<0.001) than that in ambient sediment and reference sediment. For reference sediment, tailings, and ambient sediment, the mean percent organic content by weight (and standard error) were 2.06 (0.15), 0.27 (0.02), and 1.52 (0.06), respectively (n=3). Another di€erence between reference sediment and tailings, that was not quanti®ed, was compaction. Settled, saturated reference sediment was easily penetrated with a ®nger whereas settled, saturated, tailings were not. Compared to tailings, reference sediment consisted of proportionally more particles that were 440 mm in diameter (Fig. 3). Ambient sediment had a coarser and a ®ner component than reference sediment and tailings. The coarser component of ambient sediment consisted of shell fragments and particles sometimes exceeding 5 cm in diameter. Reference sediment was prepared from ambient sediment. The coarser particles were removed from ambient sediment during the sieving process while preparing reference sediment. The ®ner component of ambient sediment was removed with the decant water during the homogenization and rinsing procedures while preparing reference sediment. Based on median grain size (Fig. 3), reference sediment, tailings, and ambient sediment were classi®ed on the Wentworth scale as medium silt, coarse silt, and very ®ne sand, respectively (Wentworth, 1922).

Fig. 3. Grain size distributions of reference sediment and tailings prior to in situ placement, and ambient sediment from the 9 months sampling period.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

311

3.2. Whole assemblage attributes All results are for the inner trays. The outer tray sediment was not analyzed. The outer trays of reference sediment were intended to detect avoidance behavior. Mediomastus californiensis was the only candidate taxon to test the method since it was the only taxon that was abundant in reference sediment and consistently less abundant in tailings. Higher densities of M. californiensis in the outer tailings trays would have indicated that avoidance contributed to reduced M. californiensis abundance in tailings. However, the labor involved to process the outer tray sediment was not deemed worthwhile. Had there been more candidate taxa, the chances of detecting avoidance behavior would have been greater. Total abundance was signi®cantly lower in tailings than in reference sediment only for the 17-months samples (Fig. 4a). There were no signi®cant di€erences between reference sediment and tailings in total biomass (Fig. 4b). The relative di€erences between reference sediment and tailings were greater for total abundance than for total biomass for the 17- and 22-months samples because the average size of individuals (biomass/abundance) in tailings was slightly larger than in reference sediment (Fig. 4c). Di€erences between reference sediment and tailings in biomass/abundance and number of taxa were not signi®cant for each period (Fig. 4c, d).

Fig. 4. Whole assemblage attributes of macrofauna. Biomass is based on ash-free dry weight. Bars represent means and 95% con®dence intervals for entire samples (177 cm2). Groups of bars for each attribute within a sampling period that have letters but do not share a letter were signi®cantly di€erent (Tukey's, P40.05).

312

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

The ambient assemblage di€ered from tailings and reference sediment. Total abundance was signi®cantly lower in ambient sediment than in tailings or reference sediment for the 17-months samples (Fig. 4a). Total abundance in ambient sediment doubled over the course of the study, compared to an increase in total biomass of 32% (Figures 4a, b). Biomass/abundance in ambient sediment was smallest in the 22-months samples (Fig. 4c). The number of taxa was signi®cantly lower in ambient sediment than in reference sediment and tailings for all periods (Fig. 4d). The minimum detectable di€erence as a percent of the reference mean, progressing from lowest to highest was generally: number of taxa; total abundance; total biomass; biomass/abundance (Fig. 5). The sample size of 10 nearly maximized the ability to detect a 25% di€erence in number of taxa relative to the reference sediment (Fig. 6). Power for the other whole assemblage attributes would have been substantially increased with a larger sample size, with the exception of abundance for the 22-months samples, which neared the asymptote at n=10. Power was greatest for all whole assemblage attributes for the 22-months samples. This was due, in part, to presentation of the minimum detectable di€erence as a percent of the reference mean and because reference sediment means were highest for the 22-months samples (Fig. 4). The reference sediment and tailings assemblages for each period were indistinguishable using correspondence analysis, whereas the ambient sediment points were clearly separate from the reference sediment and tailings points for each period

Fig. 5. Minimum detectable di€erence of ANOVA as a percent of the reference mean for whole assemblage attributes of macrofauna (=0.05, n=10). Vertical dashed lines indicate statistical power (1ÿ ) of 0.8.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

313

Fig. 6. Relation of sample size and statistical power of ANOVA for whole assemblage attributes of macrofauna (=0.05, 1ÿ =0.8, minimum detectable di€erence=25% of reference mean). B/A for 17 months was below the range of power that was included. Vertical dashed lines indicate the sample size used for this study (n=10).

(Fig. 7). The variation explained by correspondence analysis ranged from 28 to 36% for dimension 1 and was 9% for dimension 2 for each period. The greatest contributions to both dimensions were di€erences in the abundance of abundant taxa, including M. californiensis and Galathowenia oculata. 3.3. Flocculent surface layer The majority of macrofauna in the 17 months tailings tray samples were in direct contact with tailings, beneath the ¯occulent surface layer (Table 1). The abundance of macrofauna in the ¯occulent layer, as a percent of the total, was greater than biomass, re¯ecting the relatively small size of the surface fauna. Approximately 5, 6, and 11 mm of sediment had settled in the sediment traps for the 9, 17, and 22months samples, respectively. 3.4. Species level comparisons With the exception of unidenti®ed nematodes, all of the numerically dominant taxa were polychaetes, each of a di€erent family (Table 2). For the entire study, reference sediment and tailings shared the three most abundant taxa: M. californiensis;

314

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

Fig. 7. Relation of macrofaunal assemblages for each sediment type using correspondence analysis. The distance between points indicates the overall similarity of taxonomic composition, weighted by abundance. Each point represents a replicate reference sediment tray, tailings tray, or ambient sediment core. Replicates for each sediment type are surrounded by 95% con®dence regions, generated after the analysis. Table 1 Whole assemblage attributes of macrofauna in the ¯occulent surface layer and in the underlying tailings for the 17 months samples (n=10)a Layer

Floc on tailings Tailings w/o ¯oc

Abundance

Biomass (mg)

Number of taxa

Biomass/abundance (mg/individual)

Mean

SE

Mean

SE

Mean

SE

Mean

SE

75.5 (15.3) 417.1 (84.7)

13.4 31.2

5.3 (2.7) 188.9 (97.3)

1.72 15.22

17.7 32.5

1.21 2.10

0.051 0.416

0.020 0.048

a Values in parentheses are percent of total. Percent of total was not calculated for number of taxa because the two sediment layers shared some taxa.

G. oculata; Prionospio steenstrupi, although the rank order varied. These three taxa were also the three most abundant in the 9 months ambient samples. In the 17 and 22 months ambient samples, G. oculata was replaced among the top three by Lumbrineris luti and Nephtys cornuta, respectively. Other taxa that were commonly found in the trays and ambient cores included amphipods, cumaceans, bivalves, and gastropods.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

315

Table 2 Most abundant taxa of macrofaunaa Taxon

Sampling period (months)

Percent of total abundance Reference

Tailings

Ambient

Mediomastus californiensis

9 17 22

25.2 30.2 22.6

18.5 18.4 18.0

14.6 17.9 10.2

Galathowenia oculatab

9 17 22

14.5 18.8 17.8

20.5 26.6 22.0

10.3 4.3 6.8

Prionospio steenstrupi

9 17 22

7.8 4.1 18.3

9.1 3.9 20.2

13.2 12.4 24.3

Pholoe glabra

9 17 22

8.6 8.5 4.9

10.3 9.1 4.0

9.3 3.9 7.3

Nephtys cornuta

9 17 22

5.2 6.5 6.1

3.7 8.2 3.5

9.4 7.1 11.9

Lumbrineris luti

9 17 22

1.4 2.0 3.3

1.1 1.2 2.9

10.0 16.2 7.6

Unidenti®ed Nematoda

9 17 22

0.8 1.0 1.0

0.8 0.6 0.8

4.3 7.9 7.8

Glycinde polygnatha

9 17 22

3.7 1.6 1.7

4.9 1.8 1.8

2.4 1.5 1.6

Clymenella torquata

9 17 22

1.0 6.2 1.8

1.5 4.2 2.3

0.4 2.1 0.4

b

a Values are percent of the total for each sediment type and sampling period. Taxa are listed in descending order. Taxa whose abundances were >2% of the total abundance across sediment types and periods were included. All taxa were polychaetes with the exception of Nematoda. b Identi®cation con®rmed by L. Harris, Curator of Polychaetes at the Los Angeles County Museum, Los Angeles, California, USA.

The nine most abundant taxa (Table 2), when combined, comprised 68±79% of the total abundance by period and sediment type. These same taxa, when combined, comprised 19±45% of the total biomass by period and sediment type. This disparity between the percentage of total abundance and biomass indicated that individuals of the most abundant taxa tended to be smaller than the average-sized individual. This ®nding clearly applied to M. californiensis, whose percentage of total abundance exceeded its percentage of total biomass by >10.

316

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

Table 3 Macrofauna taxa having the greatest total biomassa Taxon

Sampling period (months)

Percent of total biomass Reference

Tailings

Laonome kroyeri

9 17 22

6.2 16.2 8.0

13.8 33.8 22.5

0.0 0.0 0.0

Lumbrineris luti

9 17 22

13.5 6.7 8.2

12.0 3.9 4.6

30.8 34.1 31.6

Ophelina acuminata

9 17 22

36.1 21.8 11.3

16.0 1.4 2.4

6.4 3.3 1.3

Galathowenia oculatab

9 17 22

2.3 9.4 14.4

3.3 10.6 14.2

1.5 1.6 2.0

Tellina sp.

9 17 22

2.0 3.2 10.3

3.1 7.3 4.8

0.1 11.2 0.1

Amphicteis scaphobranchiatab

9 17 22

0.7 4.9 3.9

1.3 2.3 7.1

0.0 0.0 0.0

Dipolydora socialis

9 17 22

4.6 1.0 2.3

5.1 11.0 3.1

0.3 0.1 0.1

Terebellides stroemib

9 17 22

0.3 1.8 6.3

0.3 1.3 3.2

0.1 4.0 0.2

Axiothella sp.

9 17 22

0.0 0.1 2.7

0.0 0.0 0.1

7.4 13.0 7.2

Ophiura sp.

9 17 22

0.0 0.0 0.0

0.0 0.5 8.9

0.0 4.5 0.0

Prionospio steenstrupi

9 17 22

4.3 0.8 2.3

5.1 1.0 2.8

1.7 1.5 4.3

Nephtys ciliata

9 17 22

0.9 1.7 0.6

1.4 3.7 1.8

8.7 0.0 3.8

a

Ambient

Values are percent of the total for each sediment type and sampling period. Taxa are listed in descending order. Taxa whose biomass was >2% of the total biomass across sediment types and periods were included. All taxa were polychaetes with the exception of Tellina sp. (mollusca) and Ophiura sp. (echinodermata). b Identi®cation con®rmed by L. Harris, Curator of Polychaetes at the Los Angeles County Museum, Los Angeles, California, USA or T. Rice, Curator of the Sea and Shore Museum, Port Gamble, Washington, USA.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

317

When combined, the 12 taxa having the greatest total biomass (Table 3) comprised 51±77% of the total biomass by period and sediment type. Of these 12 taxa, L. luti, G. oculata, and P. steenstrupi were also among the most abundant taxa (Table 2). Ten of the 12 species constituting the largest portions of the total biomass were polychaetes. Individuals of three of the four most abundant taxa were signi®cantly larger in ambient sediment than in the trays in the 9-months samples (Fig. 8). By 22 months, these same taxa were signi®cantly smaller in ambient sediment than in the trays. The four most abundant taxa were generally of similar size in reference sediment and tailings with the exception of M. californiensis and P. steenstrupi. Minimum detectable di€erences for biomass/abundance of numerically dominant taxa ranged from 26 to 108% of the reference mean (=0.05, 1ÿ =0.8). 3.5. Functional groups The distribution of abundance by feeding type was similar for reference sediment and tailings compared to ambient sediment (Fig. 9). Di€erences in abundance of

Fig. 8. Average size (biomass/abundance) of numerically dominant macrofauna taxa. Biomass is based on ash-free dry weight. Bars represent means and 95% con®dence intervals. Groups of bars for each species within a sampling period that have letters but do not share a letter were signi®cantly di€erent (Tukey's, P40.05).

318

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

Fig. 9. Abundance of macrofauna by feeding type. Bars represent means and 95% con®dence intervals for entire samples (177 cm2). Feeding type: ssd, subsurface deposit; sd, surface deposit; sd/s, surface deposit/suspension; sd/c, surface deposit/carnivore; c, carnivore; o, other; u, unclassi®ed. Groups of bars for each feeding type within a sampling period that have letters but do not share a letter were signi®cantly di€erent (Tukey's, P40.05).

subsurface deposit and surface deposit/suspension feeders were attributable mainly to M. californiensis and G. oculata. Di€erences in the abundance of carnivores re¯ected the abundance of Pholoe glabra and N. cornuta. Burrowers were generally less abundant in tailings than in reference sediment (Fig. 10), owing largely to the abundance of M. californiensis. Di€erences in the abundance of surface dwellers was due largely to the abundance of N. cornuta. For each sampling period, there were signi®cantly fewer tube builders in ambient sediment than in reference sediment and tailings, due mainly to the abundance of G. oculata. Minimum detectable di€erences for functional groups ranged from 26 to 151% of the reference mean (=0.05, 1ÿ =0.8). 4. Discussion This study was designed to assess the ability of macrofauna to colonize a tailings deposit upon cessation of a hypothetical STD operation for the Kensington mine,

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

319

Fig. 10. Abundance of macrofauna by sediment association type. Bars represent means and 95% con®dence intervals for entire samples (177 cm2). Groups of bars for each sediment association type within a sampling period that have letters but do not share a letter were signi®cantly di€erent (Tukey's, P40.05).

located near Juneau, Alaska. While this was a case study, the feasibility of STD for the Kensington mine has broader interest because STD is currently prohibited in the USA. Analysis of macrofauna was intended to be a conservative measure of habitability since exposure to the tailings was intimate and prolonged. Given the ®nding of similar assemblages in reference sediment and tailings, it may be inferred that the presence of a tailings substrate would not inhibit colonization of macrofauna after STD, although the actual rate could not be determined based on the experimental results alone. The experiment was conducted at 21 m, compared to depths of past and potential STD sites along the coasts of British Columbia and Alaska of 50±370 m (Coldwell & Gensler, 1993; Ellis et al., 1995a). Benthos tend to recolonize disturbed areas at a slower rate with increasing depth (Grassle, 1977; Levin & Smith; 1984, Smith & Hessler, 1987). The Kitsault mine, located in British Columbia near the border of Alaska, used STD for 18 months in a fjord that exceeded 370 m in depth. Recolonization of macrofauna in terms of abundance, number of taxa, and proportions of major taxonomic groups was complete 4 years after the operation, although di€erences in

320

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

the proportional representations of species at impacted versus unimpacted sites were still detected (Brinkhurst et al., 1987). The Island Copper mine on Vancouver Island, British Columbia, used STD from 1970 until closure in 1995. The tailings deposit extended approximately 16 km along the axial length of a fjord and covered the majority of its 1.8 km width (Poling et al., 1993). During operation, benthic samples collected nearest the 50 m deep outfall generally contained between one and ®ve species of macrofauna (Ellis et al., 1995a). An in situ experiment revealed that tailings from the Island Copper mine and marble sand were colonized by similar taxonomic assemblages, although tailings retarded the rate of community development (Ellis & Taylor, 1988; Taylor, 1986). Two years after closure of the operation, the number of species and total number of macrofauna at locations that were nearly defaunated by depositing tailings were within the range of unimpacted locations, although assemblages at the previously impacted locations contained proportionally more opportunistic species than at unimpacted sites (Ellis, 1998). Several of the most abundant taxa in the trays of reference sediment and tailings from the Auke Bay experiment: Mediomastus sp.; G. oculata; N. cornuta; L. luti, were also among the most abundant taxa that colonized tailings after the Kitsault and Island Copper STD operations (Brinkhurst et al., 1987; Ellis & Hoover, 1998). Polychaetes were the numerically dominant taxon in the Auke Bay experiment, as they were in the fjords that received tailings from the Kitsault and Island Copper mine, and at depth adjacent to the Kensington mine in Lynn Canal (Kline, 1998). Polychaetes are also a diverse taxon that occupy a variety of functional roles. Polychaetes are therefore an appropriate taxon on which to place an emphasis for the Auke Bay experiment. Much of the data analysis and presentation was for whole assemblage attributes, thereby incorporating information on all of the taxa. Data on other taxa could be extracted from the Auke Bay database and analyzed separately. However, these data are likely to be more variable since taxa other than polychaetes were less abundant. Ideally, this study would have been conducted with sediment collected from an area where Kensington tailings could conceivably be placed, such as Lynn Canal (Fig. 1). However, comparison of colonization in tailings to any naturally occurring, uncontaminated, near-shore, ®ne-grained, marine sediment would have provided perspective since minor di€erences in sediment characteristics may not in¯uence faunal composition (Wu & Shin, 1997). The lack of di€erences in reference sediment and tailings assemblages indicated that physical and chemical di€erences between the two sediment types were inconsequential. A related study demonstrated that pore water concentrations of Cd, Cu, Pb, and Zn were similar in reference sediment and tailings while the concentration of Fe was slightly higher and the concentration of Mn was four to ®ve times lower in tailings pore water, relative to reference sediment (Lambeth & Drake, 1998). More apparent di€erences between reference sediment and tailings were relatively low organic content and high compaction of tailings. These latter factors may have been the cause of reduced abundance of subsurface deposit feeders (Fig. 9) and burrowers (Fig. 10) since organic content and compaction

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

321

should be most likely to impact subsurface taxa. However, tube builders were equally abundant in reference sediment and tailings, and were found throughout the depth of the tailings trays. Since subsurface taxa were able to colonize the tailings, bioturbation should have decreased compaction over the course of the experiment. Also, bioturbation, combined with the addition of natural, settled sediment, should have increased the organic content of the tailings. Therefore, possible di€erences in the tray assemblages resulting from high compaction and low organic content of tailings should have been most prevalent in the 9-months samples. Most of the data from the Auke Bay experiment were analyzed by abundance rather than biomass since abundance data were less variable. This approach may have emphasized the di€erences between reference sediment and tailings assemblages since total biomass in reference sediment and tailings was more similar than total abundance (Fig. 4). In addition, since the ANOVA comparisons were done within sampling periods, the Type I error rate of all comparisons taken together was greater than =0.05. Hence, some of the di€erences deemed signi®cant at P40.05, when considered together with other ANOVA comparisons, may not have been signi®cant. An example of this problem may be total abundance in tailings versus reference sediment in the 17-months samples (Fig. 4), which were marginally di€erent (P=0.03). The ecological relevance of a 25% di€erence in abundance such as this for marine benthos is not known. However, 67% of the excess abundance in reference sediment was due to lower abundance of M. californiensis in tailings. M. californiensis was not among the greatest contributors to total biomass (Table 3) and, while less abundant than in reference sediment trays, was the second most abundant taxon in the tailings trays. Experimental conditions, including tray e€ects, were equal for reference sediment and tailings, isolating sediment type as the only variable. Tray e€ects were assumed to be a potentially in¯uential factor when comparing the tray and ambient assemblages. However, the ambient sediment was sampled because recovery of benthic assemblages in sediment trays is commonly assessed through comparison to an ambient assemblage (Grassle & Morse-Porteous, 1987; Levin & Smith, 1984). For many attributes in the current study, the ambient assemblage was signi®cantly different from the tray assemblages throughout the 22-months period of study. This could be interpreted as evidence that the tray assemblages had not attained the successional stage of the ambient assemblage, however, the tray assemblages matched or exceeded the ambient assemblage in total abundance, biomass, size of individuals, and number of taxa by the 9 or 17 months sampling periods (Fig. 4). The persistent di€erences between the tray and ambient assemblages suggests that there were tray e€ects. A retrieval date prior to 9 months may have resulted in more valuable information pertaining to succession rather than extending the study to or beyond 22 months. However, the ability to detect di€erences between the tray sediments, based on statistical power (Figs. 5 and 6), was greatest for the 22 months sampling period. Tray e€ects, such as prohibiting lateral colonization by ambient fauna (Smith & Brumsickle, 1989), were desirable since the objective was not to mimic the

322

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

natural sea ¯oor, but to mimic recolonization of a defaunated sea ¯oor. STD can produce tailings deposits in the order of several square kilometers, although total defaunation is limited to areas that are being directly impacted by the tailings density current (Ellis et al., 1995b). Recolonization of large, defaunated areas is mainly through dispersal and settlement of meroplankton (Santos & Simon, 1980, Taylor, 1986). The Auke Bay experiment was intended to mimic this mode of recolonization. In addition to prohibiting lateral colonization, the trays may have reduced predation by larger, mobile benthos such as ¯at®sh and larger species of crab, although shrimp and star®sh were often observed on the trays. Predator exclusion may have been somewhat realistic with respect to the behavior of megabenthos during STD since large predators have been found to avoid azoic tailings under laboratory conditions (Johnson, Rice & Moles, 1998). However, bottom trawling during the Kitsault STD operation (Environment Canada, 1983) and crab pot ®shing 1 year after closure (Sloan, 1985) indicated that large benthic predators will remain on or near tailings deposits during STD. In general, the tray assemblages varied together over the 22 months study period. There were slight depressions in abundance and number of taxa during the spring, and a consistent increase in biomass and in the average size of individual macrofauna (Fig. 4). The ambient assemblage also displayed seasonality in abundance and number of taxa, except to a greater degree than the tray assemblages. Ambient trends in total biomass and the average size of individuals di€ered from the trays in that the average size of individuals peaked in the spring, countering the depressed spring abundance and leading to relatively constant total biomass. The di€erences in trends between the tray and ambient assemblages may largely be accounted for by G. oculata. This polychaete was among the greatest contributors to total abundance and total biomass in the tray assemblages (Tables 2 and 3). Its individual size increased throughout the study in the trays but displayed seasonality in the ambient sediment. 5. Conclusions This study was an e€ective means to predict the ability of benthic macrofauna to recolonize an area upon closure of an STD operation. The relationship between the assemblages in trays of reference sediment and trays of tailings was similar for samples that were collected after 9, 17, and 22 months colonization periods, however, statistical power was greatest for the 22-months samples. Probable tray e€ects, which were desirable given the objectives of this study and were equal for tailings and reference sediment, precluded direct comparison of ecological succession in the tray sediments to the widely varying ambient assemblage. Based on the comparison of the assemblages that colonized trays of tailings and reference sediment, recolonization by macrofauna after STD should not be inhibited by the tailings used in this study. Recolonization should be similar to that in barren, natural sediment, and controlled by natural recruitment processes.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

323

Acknowledgements We thank the US Bureau of Mines, Echo Bay Mines, the University of Alaska Fairbanks, and the University of Alaska Southeast for funding this project, Rick Richins and Coeur Alaska, Inc. for providing tailings, Robert Lambeth and Roger Baer for their technical advice, Drs. Lewis Haldorson, Stanley Rice, Thomas Shirley, Douglas Woodby, and Peter Chapman for critiquing the study design and editing an earlier version of this manuscript, Aaron Baldwin for his taxonomic services, and Kathleen Lyden, Guy Lewis, Christopher Rooper, Bradley Wheeler, Michelle Bourassa, and Carol Kline for providing technical assistance. References American Society for Testing and Materials. (1990) Test method for particle size analysis of soils. Annual book of ASTM standards, designation: D 422-63. Philadelphia: ASTM. Arntz, W. E. (1978). The ``upper part'' of the benthic food web: the role of macrobenthos in the western Baltic. Rapp. P.-v. Reun. Cons. int. Explor. Mer, 173, 85±100. Blake, J. A., Hilbig, B., & Scott, P. H. (Eds.) (1996). Taxonomic atlas of the benthic fauna of the Santa Maria Basin and the Western Santa Barbara Channel, Vol. 6, the Annelida, Part 3 Ð Polychaeta: Orbiniidae to Cossuridae. Santa Barbara: Santa Barbara Museum of Natural History. Bonsdor€, E., & OÈsterman, C. (1985) The establishment, succession and dynamics of a zoobenthic community Ð an experimental study (pp. 287±297). In Proceedings, 19th European Biology Symposium, 16±21 September 1984, Plymouth, Devon, United Kingdom, New York: Cambridge University Press. Brinkhurst, R. O., Burd, B. J., & Kathman, R. D. (1987) Benthic studies in Alice Arm B.C., following cessation of mine tailings disposal, 1982 to 1986. Canadian Technical Report of Hydrography and Ocean Sciences, No. 89, Institute of Ocean Sciences, Department of Fisheries and Oceans, Sydney, BC, Canada. Buchanan, J. B. (1984). Sediment analysis. In N. A. Holme, & A. D. McIntyre, Methods for the study of marine benthos (2nd ed.) (pp. 41±65). Oxford: Blackwell Scienti®c Publications. Caldwell, J. A., & Welsh, J. D. (1082). Tailings disposal in rugged, high precipitation environments: an overview and comparative assessment. In D. V. Ellis, Marine tailings disposal (pp. 5±62). Ann Arbor: Ann Arbor Science Publishers. Coldwell, J. R., & Gensler, E. C. (1993) Potential for submarine tailings disposal to a€ect the availability of minerals from United States coastal areas. Number OFR 101-93, US Bureau of Mines, Juneau. Coyle, K. O., & Paul, A. J. (1990). Interannual variations in zooplankton population and biomass during the spring bloom in an Alaskan subarctic embayment. In D. A. Ziemann, & K. W. Fulton-Bennett, APPRISE Ð interannual variability and ®sheries recruitment (pp. 179±228). Honolulu: The Oceanic Institute. Coyle, K. O., & Shirley, T. C. (1990). A review of ®sheries and oceanographic research in Auke Bay, Alaska and vicinity, 1966±1985. In D.A Ziemann, & K. W. Fulton-Bennett, APPRISE Ð interannual variability and ®sheries recruitment (pp. 1±73). Honolulu: The Oceanic Institute. Crisp, D. J. (1971). Energy ¯ow measurements. In N. A. Holme, & A. D. McIntyre, Methods for the study of marine benthos (2nd ed.) (pp. 263±265). Oxford: Blackwell Scienti®c Publications. Elliot, J. M. (1977) Some methods for the statistical analysis of samples of benthic invertebrates (2nd ed. reprinted with minor corrections 1983, 4th impression 1993). Freshwater Biological Association, Scienti®c Publication No. 25. Ambleside, UK: The Ferry House. Ellis, D. V. (1998) Ecological criteria for determining the sustainability of restoration. Presented at the ``Helping the Land Heal'' Conference, Victoria, November 5±8, 1998. Ellis, D. V., & Hoover, P. M. (1990). Benthos recolonizing mine tailings in British Columbia fjords. Marine Mining, 9, 441±457.

324

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

Ellis, D. V., & Taylor, L. A. (1988) Biological engineering of marine tailings beds. In W. Salomons & U. FoÈrstner, Environmental management of solid waste (pp. 185±207). Berlin: Springer-Verlag. Ellis, D., Poling, G., & Pelletier, C. (1994) Case studies of submarine tailings disposal: Volume II Ð worldwide case histories and screening criteria. Number OFR 37-94, US Bureau of Mines, Juneau. Ellis, D. V., Pedersen, T. F., Poling, G. W., Pelletier, C., & Horne, I. (1995a). Review of 23 years of STD: Island Copper Mine, Canada. Marine Georesources and Geotechnology, 13, 59±99. Ellis, D. V., Poling, G. W., & Baer, R. L. (1995b). Submarine tailings disposal (STD) for mines: an introduction. Marine Georesources and Geotechnology, 13, 3±18. Environment Canada. (1983) Environmental studies in Alice and Hastings Arm, Part V Ð Baseline and initial production period Ð Amax/Kitsault mine Ð submersible observations and otter trawls, 1980± 1982. Environmental Protection Service, Paci®c Region no. EPS-PR-83-05. Fauchald, K. (1977). The polychaete worms. Los Angeles: Natural History Museum of Los Angeles. Grassle, J. F. (1977). Slow recolonisation of deep-sea sediment. Nature, 265, 618±619. Grassle, J. F., & Morse-Porteous, L. S. (1987). Macrofaunal colonization of disturbed deep-sea environments and the structure of deep-sea benthic communities. Deep Sea Research, 34, 1911±1950. Greenacre, M. J., & Underhill, L. G. (1082). Scaling a data matrix in a low-dimensional Euclidean Space. In D. M. Hawkins, Topics in applied multivariate analysis (pp. 183±266). Cambridge: Cambridge University Press. Greenacre, M. J., & Vrba, E. S. (1984). Graphical display and interpretation of antelope census data in African wildlife areas, using correspondence analysis. Ecology, 65, 984±997. Hay, A. E., Murray, J. W., & Burling, R. W. (1983). Submarine channels in Rupert Inlet, British Columbia: II. Sediments and sedimentary structures. Sedimentary Geology, 36, 317±339. Hesse, C. A., & Ellis, D. V. (1995). Quartz Hill, Alaska: a case history of engineering and environmental requirements for STD in the USA. Marine Georesources and Geotechnology, 13, 135±182. Hughes, R. G. (1984). A model of the structure and dynamics of benthic marine invertebrate communities. Marine Ecology Progress Series, 15, 1±11. Johansen, P., Hansen, M. M., Asmund, G., & Nielsen, P. B. (1991). Marine organisms as indicators of heavy metal pollution Ð experience from 16 years of monitoring at a lead zinc mine in Greenland. Chemistry and Ecology, 5, 35±55. Johnson, S. W., Rice, S. D., & Moles, D. A. (1998). E€ects of submarine mine tailings disposal on juvenile yellow®n sole (Pleuronectes asper): a laboratory study. Marine Pollution Bulletin, 36, 278±287. Jones, S. G., & Ellis, D. V. (1995). Deep water STD at the Misima gold and silver mine, Papua, New Guinea. Marine Georesources and Geotechnology, 13, 183±200. Karinen, J. F. (1985) Distribution of heavy metals and hydrocarbons in Auke Bay sediments and in biota near active marinas. In Final environmental impact statement, Auke Bay breakwater and related marina developments (pp. 1±45 Appendix). Alasica Region: US Army Corps of Engineers. Kathman, R. D., Brinkhurst, R. O., Woods, R. E., & Je€ries, D. C. (1983) Benthic studies in Alice Arm and Hastings Arm, B.C., in relation to mine tailings dispersal. Canadian Technical Report of Hydrography and Ocean Sciences, No. 22. Kathman, R. D., Brinkhurst, R. O. Woods, R. E., & Cross, S. F. (1984) Benthic studies in Alice Arm, B.C., following cessation of mine tailings disposal. Canadian Technical Report of Hydrography and Ocean Sciences, No. 37. Kline, E. R. (1994) Potential biological consequences of submarine mine-tailings disposal: a literature synthesis. No. OFR 37-94, US Bureau of Mines, Juneau. Kline, E. R. (1998) Biological impacts and recovery from marine disposal of metal mining waste. PhD thesis, University of Alaska Fairbanks. Kozlo€, E. N. (1987). Marine invertebrates of the Paci®c northwest. Seattle: University of Washington Press. Lambeth, R. H., & Drake, P. L. (1998) An on-site simulation of submarine tailings disposal: pore water chemistry and biological recovery. Preprint No. 98-153, Presented at the Society for Mining, Metallurgy, and Exploration, Inc. annual meeting Proceedings, 9±11 March 1998, Orlando. Leduc, G., Ruber, H., & Webb, M. (1975) Toxicity studies in relation to mining in the Northwest Territories, No. ALUR 73-74-34. Department of Indian A€airs and Northern Development, Ottawa.

E.R. Kline, M.S. Stekoll / Marine Environmental Research 51 (2001) 301±325

325

Levin, L. A. (1984). Life history and dispersal patterns in a dense infaunal polychaete assemblage: community structure and response to disturbance. Ecology, 65, 1185±1200. Levin, L. A., & Smith, C. R. (1984). Response of background fauna to disturbance and enrichment in the deep sea: a sediment tray experiment. Deep Sea Research, 31, 1277±1285. Poling, G. W. (1982). Characteristics of mill tailings and their behavior in marine environments. In D. V. Ellis, Marine tailings disposal (pp. 63±84). Ann Arbor: Ann Arbor Science Publishers. Poling, G. W., Ellis, D. V., Pelletier, C. A., Pedersen, T. F., & Hesse, C. A. (1993) Case studies of submarine tailings disposal: Vol. I Ð North American examples. No. OFR 89-93, US Bureau of Mines, Juneau. Rhoads, D. C., & Germano, J. D. (1986). Interpreting long-term changes in benthic community structure: a new protocol. Hydrobiologia, 142, 291±308. Santos, S. L., & Simon, J. L. (1980). Marine soft-bottom community establishment following annual defaunation: larval or adult recruitment? Marine Ecology Progress Series, 2, 235±241. Short, J. W., Rice, S. D., Brodersen, C. C., & Stickle, W. B. (1989). Occurrence of tri-n-butyltin-caused imposex in the North Paci®c marine snail Nucella lima in Auke Bay, Alaska. Marine Biology, 102, 291± 297. Sloan, N. A. (1985). Life history characteristics of fjord-dwelling golden king crabs Lithodes aequispina. Marine Ecology Ð Progress Series, 22, 219±228. Smith, C. R., & Brumsickle, S. J. (1989). The e€ects of patch size and substrate isolation on colonization modes and rates in intertidal sediment. Limnology and Oceanography, 34, 1263±1277. Smith, C. R., & Hessler, R. R. (1987). Colonization and succession in deep-sea ecosystems. Trends in Ecology and Evolution, 2, 359±363. SPSS, Inc. (1997). SYSTAT 7.0. Ann Arbor: Ann Arbor Science Publishers. Taylor, L. A. (1986) Marine macrobenthic colonization of mine tailings in Rupert Inlet, British Columbia. MS thesis, University of Victoria. Virnstein, R. W. (1977). The importance of predation by crabs and ®shes on benthic infauna in Chesapeake Bay. Ecology, 58, 1199±1217. Wentworth, C. R. (1922). A scale of grade and class terms for clastic sediments. Journal of Geology, 30, 377±392. Wu, R. S. S., & Shin, P. K. S. (1997). Sediment characteristics and colonization of soft-bottom benthos: a ®eld manipulation experiment. Marine Biology, 128, 475±487. Zajac, R. N., & Whitlatch, R. B. (1982a). Responses of estuarine infauna to disturbance. I. Spatial and temporal variation of initial recolonization. Marine Ecology Progress Series, 10, 1±14. Zajac, R. N., & Whitlatch, R. B. (1982b). Responses of estuarine infauna to disturbance. II. Spatial and temporal variation of succession. Marine Ecology Progress Series, 10, 15±27. Zar, J. H. (1996). Biostatistical analysis (3rd ed.). Upper Saddle River: Prentice-Hall. Ziemann, D. A., & Fulton-Bennett, K. W. (Eds.) (1990). APPRISE-interannual variability and ®sheries recruitment. Honolulu: The Oceanic Institute.