| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Zoology, Southern Illinois University, Carbondale, IL 62901-6501
b Division of Biology, Ackert Hall, Kansas State University, Manhattan, KS 66506
* Corresponding author (mwhiles{at}zoology.siu.edu)
Received for publication September 7, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Along with their significance to lotic food webs, suspended particles can also represent both chronic and acute pollutants in streams and rivers. Yearly compilation of U.S. State and Tribal water quality reports (305b reports) indicate that approximately 13% of rivers and streams are impaired by suspended solids (USEPA, 2001), and suspended sediments and bedload in streams are the largest pollutants, by volume, in the United States (Fowler and Heady, 1981; USEPA, 1990; Waters, 1995).
Suspended particles can impair water quality in a variety of ways. Organic-rich particles and sediments can contribute to dissolved oxygen depletion by increasing biological oxygen demand (BOD) (Wood and Armitage, 1997). Toxic substances, particularly hydrophobic compounds, can adsorb to the surfaces of suspended solids and be transported into streams and downstream reaches (Karichkoff, 1984). Excessive sedimentation can degrade aquatic habitats and interfere with growth, reproduction, and survival of many aquatic organisms, thus compromising biotic integrity (Berkman and Rabeni, 1987; Cooper, 1993; Waters, 1995). Suspended particles can also increase costs associated with water treatment and purification.
A number of anthropogenic activities can influence seston dynamics. Watershed disturbances that involve removal of vegetative cover (e.g., row crop agriculture, logging, urban development, overgrazing) are known to increase sediment loading to surface waters (Gurtz et al., 1980; Webster and Golladay, 1984; Cooper, 1993; Waters, 1995; Strand and Merritt, 1999). Impoundments can also alter particle dynamics; heavy particles settle out in the lentic environment above impoundments and phytoplankton and zooplankton contribute to suspended materials in outflowing water (Hynes, 1970; Hesse et al., 1982; Angradi, 1993; Cattaneo, 1996). Thus, human activities may alter both quantity and quality of suspended particles.
Macroinvertebrate-based biological assessment techniques have become important for monitoring surface water quality and examining effects of a variety of pollutants (e.g., Rosenberg and Resh, 1993; Loeb and Spacie, 1994; Barbour et al., 1999). However, the use of macroinvertebrates for assessing suspended sediment pollution in rivers is not well established because of the dual role of sediments as food for filter-feeding invertebrates and as a stressor of some invertebrate species. Further, suspended sediment pollution is often confounded with physical habitat changes associated with sediment deposition, making identification of specific mechanisms difficult (Strand and Merritt, 1999). Linkages between sediment quantity and quality and structure of filter-feeding macroinvertebrate communities may be useful for assessing anthropogenic effects on seston dynamics.
In light of current issues regarding the regulation and management of suspended particle quality and quantity in U.S. streams, our objectives were to: (i) quantify patterns of seston quantity and quality across a stream size gradient in a Great Plains river system, (ii) assess the influence of regional anthropogenic activities on seston dynamics, (iii) examine seasonal patterns, and (iv) investigate relationships between seston dynamics and filter-feeding macroinvertebrate communities. We performed this investigation in a Great Plains river system that represents a mosaic of land and stream conditions, including row-crop agriculture, pristine prairie, rangeland, and suburban development. We hypothesized that agriculture, suburban development, and impoundments would alter seston dynamics and relationships with filter-feeding macroinvertebrate communities.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
On each sampling date, two samples of at least 30 L were collected from approximately 60% depth from the surface within the thalweg. Wadeable sites were sampled with a displacement sampler (Dodds and Priscu, 1988). This sampler is a 19-L bucket with a valve placed on the bottom that only allows water to flow in. The bottom of the bucket is placed at the desired sampling depth, and water is allowed to fill the bucket. Large river sites were not wadeable during periods of higher discharge, so they were sampled with a filtration system that was lowered into the river from a bridge. The system consisted of a large funnel that forced water through a flow meter and then through a series of sieves of decreasing mesh size (see below). The mouth of the funnel was lowered to the desired depth, and flow through the system was gauged with an attached velocity meter. The measured flow rate and time of submersion were used to compute total volume passing through the filtration apparatus. Mass values from the two suspended particle samples collected on each date at each site were averaged.
Water samples were processed through a wet filtration system that consisted of a series of stacked sieves with mesh sizes 754, 385, 180, and 98 µm. A portion of the sieved water (<98 µm) from each sample was then filtered through a Whatman (Maidstone, UK) GF/C filter (particle retention size = 1.6 µm). For analyses and data presentation, particles were grouped into fine (>98 µm) and very fine (<98 and >1.6 µm) size classes. Although we did not separate out coarse particles (e.g., particles > 1 mm that were occasionally retained on the 754-µm sieve), our sampling methods were not appropriate for sampling coarse materials (e.g., Wallace and Grubaugh, 1996) and particles > 754 µm never accounted for >2% of particle concentrations in samples from any site (annual average for all sites = 0.75% of total).
Seston samples were washed from the sieves in the field with deionized water and stored in sterile bottles on ice until they could be processed. Samples were processed within 12 h of collection. All size fractions were filtered onto ashed, pre-weighed GF/C filters and dried for at least 12 h at 95°C. These samples were weighed to determine total dry weight of particles. One of the duplicate filters of each size class was ashed for 4 h at 450°C, rewetted with deionized water, dried at 95°C for 12 h, and weighed to determine ash-free dry mass (AFDM), as a proxy for organic matter.
The other duplicate filter from each size class was subsampled with a cork borer for C and N analysis on a Carlo Erba (Milan, Italy) elemental analyzer. Because the smallest size fractions can contain substantial amounts of carbonates, these samples were placed in silver combustion capsules and treated with 40 µL of 50% HCl and dried at 65°C to remove carbonates prior to analysis. Particles were distributed unevenly over the filters, so it was not possible to calculate C or N contents directly with these measurements. Rather, these measurements were used to estimate C to N ratio.
Filter-Feeder Sampling and Analysis
Macroinvertebrates were sampled during October and November 1999 at each site following modified USEPA rapid bioassessment protocols (Barbour et al., 1999). A 0.5-mm mesh kick net was used to sample stable substrates (cobble, boulder, woody debris) in the main channel of each site. For each sample, an area approximately 1 m2 was disturbed and rubbed to dislodge invertebrates into a kick net held just downstream of the sampled area. At least two kick net samples were collected from each site, and these were combined in the field.
A random, fixed-count subsample of 300 macroinvertebrates was removed from each sample with a gridded pan and table of random numbers (e.g., Barbour et al., 1999). All filter-feeders were identified at least to genus, and all trichopteran filter-feeders to species, with keys provided by Merritt and Cummins (1996), Schuster and Etnier (1978), and Schefter and Wiggins (1986).
Data Analyses
Not all of our sites were gauged. To account for this, and to facilitate statistical analyses and data presentation, we assigned sites to flow classes (Table 1). Flow classes ranged from 1 (headwater streams with discharge <0.5 m3/s) to 9 (lowest site on the Kansas River with discharge >200 m3/s), and were based on annual average discharge values from United States Geological Survey (USGS) gauging stations (when available), field measurements of wetted width and velocity during baseflow conditions, and/or estimates of stream order based on Kansas Department of Health and Environment surface water information management system (SWIMS) maps.
Relationships between stream size, seston quality and quantity, and filter-feeder communities were examined with annual average values for each site, and analyzed with simple correlation techniques (p < 0.05). A two-sample t test (p < 0.05) was used for specific comparisons between sites (e.g., upstream and downstream of impoundments). All statistical procedures were performed with Systat Version 9 (SPSS, 1999).
For examination of seasonal patterns, we chose three sites that were representative of small, medium, and large streams in our system, and that were not noticeably impaired at sampling locations (e.g., not impounded). These sites were Kings Creek (flow class = 1), Mill Creek 2 (flow class = 4), and Kansas River 3 (flow class = 9). If more than one sampling event occurred in a season at a given site, these values were averaged for seasonal pattern comparisons.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
Impoundments had a profound effect on macroinvertebrate filter-feeder community structure. Filter-feeder communities upstream of impoundments had significantly higher taxa richness than those below impoundments (t = -4.93, df = 5, p = 0.004) (Fig. 6A) . Hydropsychid taxa, primarily Cheumatopsyche spp., H. venularis, and P. flava dominated filter-feeder abundance upstream of impoundments, whereas Simulium spp. (Diptera) dominated sites below impoundments (Fig. 6B).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The change in seston quality that we found below impoundments is similar to patterns observed in other regions. In general, prior investigations found enhanced seston quality directly below impoundments, and this pattern has been attributed to relatively high concentrations of algae and zooplankton derived from lentic habitats above impoundments (Oswood, 1979; Parker and Voshell, 1983; Voshell and Parker, 1985; Angradi, 1993; Cattaneo, 1996). The two reservoirs that we examined are surface release systems, and we observed abundant zooplankton on sieves while filtering samples from below impoundments. Further, these organisms were retained on sieves with >98-µm mesh sizes, the fraction with significantly lower C to N ratio. Thus, as with some other investigations, we attribute the pattern of higher seston quality below impoundments to planktonic inputs from reservoirs, coupled with the settling of particles from upstream sources in the reservoirs.
Microfine particles (0.5–10 µm), particularly those formed by aggregations, can also increase the quality of seston below lakes outlets (e.g., Wotton, 1984). However, our results suggest that very fine particles below the reservoirs we examined were nearly identical in C to N ratio to upstream reaches, although we did not account for particles < 1.6 µm. The lack of a difference in very fine particle quality above and below reservoirs may be related to biological processes. The sites we examined below reservoirs are all >2 km below the impoundments, and Simulium were extremely abundant in and upstream of our study sites (Whiles, unpublished data, 1999; see Fig. 6B). Members of the family Simuliidae commonly feed on suspended particles < 100 µm (Wotton, 1976, 1977; Wallace and Merritt, 1980), and it has been suggested that dense congregations of filter-feeding macroinvertebrates below impoundments can rapidly remove and/or alter the quality and quantity of seston (Morin et al., 1988; Malmqvist et al., 2001), even within 100 m of reservoir outlets (Richardson and Mackay, 1991). As such, it is quite possible that our sites were far enough downstream that most higher quality particles of the finest size class had already been removed and/or modified in size and quality.
The overall reduction in total seston concentration that we observed below reservoirs may have been a result of depletion by abundant filter-feeding macroinvertebrates, but could also have been influenced by settling of particles in the lentic environment of the reservoirs (Hynes, 1970). Regardless of the mechanisms, the sampling station furthest downstream of the impoundment on the Republican River (RR4) had consistently higher seston concentrations than the site closest to the impoundment (RR3), suggesting there was a "recovery" of seston concentrations with increased distance downstream from the reservoir (see Fig. 2).
In contrast to impoundments, streams draining suburban development had unusually high concentrations of seston (up to 50 times higher than unaffected, similar-sized streams), much of which was relatively low in quality. The suburban development upstream and along the Little Kitten Creek (LK) sites includes a recently constructed golf course, and the unusually high concentration of inorganic seston in LK1 was probably related to associated soil disturbances. Other studies have documented increased seston concentrations, particularly inorganic particles, and sedimentation of streams associated with human disturbances to watersheds (e.g., Gurtz et al., 1980; Webster and Golladay, 1984; Pozo et al., 1993; Waters, 1995; Quinn et al., 1997). We also observed large quantities of yard waste (grass clippings, leaves etc.) deposited in the stream above and at LK2, and this probably contributed to the unusually high organic particle concentrations at this site. Others have found higher organic seston concentrations associated with urban development and attributed the pattern to sewage inputs (e.g., Pozo et al., 1993), but sewage inputs were not a factor in the Little Kitten drainage.
Many prior studies suggest that, although spatially and temporally highly variable, seston in rivers averages approximately 25% organic (Naiman, 1983; Hart and Beckett, 1986; Cellot et al., 1991). Our results from this Great Plains system indicate that it contains a relatively high proportion of inorganic seston. The region we studied is dominated by relatively soft, sedimentary limestone, which weathers rapidly (Oviatt, 1998). Further, wind is a nearly constant force and a powerful agent for movement of particulates in this region (Self, 1978), a portion of which fall into surface waters. Combined with extremes in temperature and agricultural activities that expose mineral soils to wind and water erosion, these factors probably contribute to relatively high proportions of inorganic particle concentrations in streams. Additionally, allochthonous inputs of organic matter into streams in this predominately prairie region are much lower than in forested regions (Gurtz et al., 1988; Gray, 1997).
The proportion of ash in seston has been considered an indicator of watershed disturbance, particularly agricultural activities that remove or alter vegetation (e.g., Webster and Golladay, 1984; Young and Huryn, 1997). However, the watershed disturbances included in our study did not result in consistent patterns for this metric. For example, sites located on the Black Vermillion River (BV) drain a more intensely agricultural basin than any others that we examined, yet the percentage of inorganic particles in seston at these sites was similar to other, less disturbed sites of similar size (e.g., KC, WC, and MC sites, see Fig. 2). Kings Creek, which drains pristine tallgrass prairie, had relatively high percent inorganic seston values (annual average = 89% inorganic), and was consistently higher than all BV sites. Although inorganic seston concentrations were high directly below a suburban and golf course development (LK1, total seston approximately 96% inorganic), organic content increased further downstream in the same suburban area (LK2, total seston approximately 50% inorganic). These results suggest that this metric may vary with type of disturbance and region, and that more research regarding the reliability of percent inorganic seston as an indicator of watershed disturbance is needed.
Although many of the sites we sampled drain catchments with agricultural activities (row crops and cattle grazing), we found no evidence that these activities were substantially influencing relationships that we observed, compared with impoundments and suburban development. These results are in sharp contrast to numerous studies that have documented substantial agricultural effects on suspended particles in streams.
Agricultural activities, including row cropping, timber harvesting, and cattle grazing, have been linked to sedimentation and increased turbidity in streams (e.g., Gurtz et al., 1980; Webster and Golladay, 1984; Cooper, 1993; Waters, 1995; Strand and Merritt, 1999). However, the Flint Hills region, where most of our sample sites were located, has limited row crop agriculture because upland soils are thin and rocky. Although low areas are often cultivated, many of the streams in this region also have riparian gallery forests that can act as buffers. Further, we did not examine periods of extremely high discharge, and thus may not have accounted for agricultural inputs that are most prevalent during floods. The only sites that we examined where extensive row cropping dominates the drainage area and crops are planted to near stream margins were in the Black Vermillion system (BV). Although we did not detect unusual patterns regarding seston in BV sites, habitat condition in these sites was generally poor compared with other similar-sized streams (e.g., dominated by fine substrates).
Seasonal Patterns
Seasonal patterns of seston quality (C to N ratio) were evident and very consistent among sites, particularly those of similar size. However, sites altered by suburban development, and even more so those that were impounded, were exceptions. Changes in natural seasonal patterns of seston quality could have important effects on the biological integrity of these streams, as the life cycles of many aquatic organisms, including filter-feeding macroinvertebrates, are often linked to the timing of resource availability and quality, and filter-feeder macroinvertebrate productivity can be limited by the quantity and nutritional quality of seston (Benke and Wallace, 1980; Merritt et al., 1982).
Our inability to detect consistent seasonal patterns in seston concentrations among our study sites was probably linked to our sampling design. We chose to examine a large number of sites for 1 yr, and this made intense sampling logistically difficult. Thus, seasonal concentrations reflected conditions on the one to three sampling dates of each season and were subject to specific conditions on those dates, rather than integrating varying conditions over a season. Although relationships between organic seston concentrations and discharge are sometimes complex (Webster et al., 1987), storm flows can result in substantial increases in concentration and transport (e.g., Wallace et al., 1991), and storm flows have a seasonal pattern in this region (Gray et al., 1998). Along with discharge, the flooding history, biological processes, and storage and availability of particulate materials can influence seston concentration on a given date (Cummins et al., 1983; Webster et al., 1987; Wallace et al., 1991). Other investigations suggest that seston concentrations in temperate streams generally peak in spring and summer, and are lowest in winter (Malmqvist et al., 1978; Wallace et al., 1982; Webster et al., 1990).
Filter-Feeders and Seston
Our results indicate that filter-feeders are more diverse, and most groups are more abundant, in larger streams with higher organic seston concentrations, but this relationship was obscured by impoundments and suburban development. Impoundments depressed richness and resulted in dominance by Simulium, followed by the Asiatic clam, Corbicula fluminea. We did not directly address mechanisms underlying this shift in community structure, but it is probably linked to changes in seston dynamics, a major factor governing filter-feeder communities (Wallace and Merritt, 1980; Richardson and Mackay, 1991). Changes in temperature, hydrology, and substrates that often occur below impoundments also cannot be ruled out as factors influencing patterns we observed.
Suburban development depressed filter-feeder diversity, but did not alter the community as dramatically as impoundments. Little Kitten Sites 1 and 2 had unusually high seston concentrations for headwater streams, and filter-feeder richness was only two and three taxa in LK1 and LK2, respectively, compared with six taxa in the similar-sized, undisturbed Kings Creek. Taxa that were common in Kings Creek but not found in the LK system included Isonychia, Polycentropus (Trichoptera: Polycentropodidae), and Potamyia flava. However, regardless of differences in richness, Cheumatopsyche was the most abundant filter-feeder in all smaller sites that we examined, including Kings Creek, indicating that a complete shift in the filter-feeding community did not occur at suburban sites, as was observed below impoundments.
Patterns of macroinvertebrate filter-feeder abundance, diversity, and distribution are often difficult to interpret because many factors influencing them are confounded (e.g., food quality and quantity, stable substrates, temperature) (Wallace and Merritt, 1980; Richardson and Mackay, 1991). For example, the relationship between filter-feeders and organic seston concentration that we found was confounded with stream size, but the correlation between filter-feeder richness and organic seston concentration was stronger than the correlation with stream size. This pattern is consistent with other evidence that food (seston) dynamics is the primary factor influencing filter-feeding macroinvertebrate communities (Wallace and Merritt, 1980; Richardson and Mackay, 1991). Water velocity, water depth, and stable substrate are also often cited as factors governing filter-feeder diversity and abundance (Wallace and Merritt, 1980), and some of these factors could have interacted with organic seston concentrations in our system. However, our sampling methods were designed to integrate the range of habitat conditions available at each site, and we sampled only stable substrates.
Despite relationships with seston quantity, we found no evidence for a filter-feeder response to seston quality (C to N ratio or percent organic). Some studies have suggested that macroinvertebrate filter-feeders are limited by food quality (e.g., Richardson, 1984; Fuller et al., 1988; Schneider et al., 1998), and it is possible that our sampling design was not adequate for detecting differences, or that our array of sites did not encompass a wide enough variation in seston quality. Alternatively, these patterns may operate at the species level, and many of the species that we collected (e.g., Hydropsyche orris and H. bidens) were not found across a large enough gradient of conditions for us to detect patterns. Further, macroinvertebrate filter-feeders can selectively remove and ingest particles from their filtering structures (Wallace and Merritt, 1980), and even those that live in systems where low quality particles dominate may select higher quality particles.
Most individual filter-feeding taxa that we examined, particularly hydropsychid caddisflies, increased in abundance with organic seston concentrations, suggesting that seston quantity may be limiting. This is an important pattern, as filter-feeders, particularly hydropsychids, can represent a sizeable portion of invertebrate production in streams (e.g., Voshell, 1985; Grubaugh and Wallace, 1995; Grubaugh et al., 1997), and contribute substantially to material and energy cycling and fish production. Simulium was one of the few dominant taxa that did not show a relationship with seston quantity, and this may be related to the size of particles they filter. Simuliids filter extremely fine particles (Wotton, 1976, 1977; Wallace and Merritt, 1980), including size fractions finer than we sampled. Therefore, our sampling methods may not have accounted for the range of seston quantity available to Similium at each site.
Cheumatopsyche was the only hydropsychid that did not increase in abundance with organic seston concentration, and this may be related to specific food and/or habitat preferences. Cheumatopsyche are reported to feed primarily on animal and algal material in some streams (Coffman et al., 1971), but other studies indicate that detritus is also an important food (Sanchez and Hendricks, 1997). Similar to the pattern we observed, Sanchez and Hendricks (1997) studied Cheumatopsyche in two Virginia stream reaches and measured higher abundance and production from the site with lower organic seston concentrations. In our region, Cheumatopsyche are abundant in even the smallest of streams where other hydropsychids are not found and seston concentrations are presumably low (Fritz et al., 1999; Stagliano and Whiles, 2002). These observations suggest that factors other than seston concentration are more important determinants of Cheumatopsyche abundance.
In summary, we provide strong evidence for the presence of predictable patterns regarding stream size, seston dynamics, and filter-feeding macroinvertebrates in this Great Plains river system, and demonstrate that anthropogenic factors can have a profound influence on these patterns. Relationships with filter-feeding macroinvertebrates that we observed suggest that biological assessment may be a viable option for initial evaluation of human influences on seston dynamics. Our results also indicate that future regulations regarding suspended particles should take into account the existing patterns in a given system and how they are influenced by anthropogenic activities.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
