| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a School of Forestry and Wildlife Sciences, 108 M.W. Smith Hall, Auburn University, Auburn, AL 36849-5418
b Arboretum, Auburn University, Auburn, AL 36849. Sponsoring organizations: USDoD/EPA/DoE, Strategic Environmental Research and Development Program and Fort Benning Military Reservation
* Corresponding author (lockabg{at}auburn.edu)
Received for publication March 5, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
A key question is how susceptible the water filtration function may be to ecosystem stresses of anthropogenic origin. The impact of land use on sediment export into streams and elevated deposition in wetlands may represent one type of stress. Already causing a very serious water quality problem throughout the world, expanding human populations and resulting landscape changes will continue to heighten the severity of the sediment export issue (Novotny, 2003). Higher sediment loads in streams imply greater sediment accumulation within riparian wetlands, as those systems respond as filters to the increased loadings. Elevated deposition rates have been observed in riverine forests within watersheds that have significant levels of agriculture or urbanization (Thom et al., 2001; Hupp, 2000; Kleiss, 1996; Howarth et al., 1991).
Filtration of sediments from floodwaters is a normal function of riparian forests (Meyer, 1990) and, depending on geomorphic position and natural characteristics of associated streams, may result in substantial accumulation of sediment on the floodplain. One of the primary determinants of retention capacity is surface roughness, as it relates to reduction in flow velocity. In riparian forests, the principle component associated with roughness is the vegetation, both standing and dead. While some sediment accumulation reflects a natural process in depositional topographic positions (Hupp and Morris, 1990), there exists little understanding of critical levels beyond which further accumulation becomes a stress in regard to ecosystem integrity and function (Jurik et al., 1994; Adamus and Brandt, 1990).
Although floodplain forests are naturally subject to varying degrees of alluviation, upstream or upslope land use may increase sedimentation rates to the point that the riparian vegetation is severely damaged (Bazemore et al., 1991). Severe damage may affect roughness and result in a reduction of the capacity of the system to trap sediment (Hupp and Osterkamp, 1995; Hupp and Bazemore, 1993). Additionally, excess nutrients, trace elements, and other hydrophobic contaminants are often physically bound to fine sediment (Johnston et al., 1984; Hupp et al., 1993) and forested riparian areas function to remove these environmental contaminants. Consequently, excessive sedimentation has the potential to impair the filtration function and negatively affect downstream water quality (Hupp, 1992; Boto and Patrick, 1979).
At a general perspective, some functional changes may be hypothesized as responses to elevated rates of deposition. If sediment influx exceeds sustainable loadings, tree roots may be subjected to increasingly anoxic conditions resulting in growth declines and, ultimately, mortality (Cavalcanti and Lockaby, 2003; Junk and Piedade, 1997; Broadfoot and Williston, 1973; Kennedy, 1970). In long-term scenarios, a vegetation transition may occur in which species capable of tolerating new site conditions (e.g., developing new roots as alluvium aggrades) increase in dominance.
Changes in vegetation caused by excessive sedimentation may result from and/or drive biogeochemical changes including an alteration of decomposition and mineralization patterns. In some cases, burial of foliar litter by sediment has resulted in subtle changes in mass loss rates (Wang et al., 1994; Mayack et al., 1989). Working near blackwater streams in the Coastal Plain of South Carolina, Mayack et al. (1989) found that sediment burial of sweetgum (Liquidambar styraciflua L.) foliage caused alterations in microarthropod dynamics compared with patterns observed in unburied litter. Although microarthropod activity was generally suppressed by litter burial, the activity of shredders was enhanced. Consequently, annual mass loss rates were similar for both categories of litter.
In contrast, Herbst (1980) found substantial decreases in mass loss rates when maple (Acer spp.) and cottonwood (Populus spp.) leaves were buried. Although Wang et al. (1994) found that initial phases of decomposition were slower for buried leaves, no significant effects of sediment burial were apparent after 17 mo. Herbst (1980) and Mayack et al. (1989) speculated that litter burial decreased organic matter export via streams and dampened the seasonal pulses of nutrient inputs to floodplain soils.
The degree to which sediment deposition can promote temporal uniformity in nitrogen supply may be dependent on many factors, such as the textural composition of sediment. Pinay et al. (1995) demonstrated that floodplain retention of nitrogen was far lower if deposition was dominated by coarse-textured vs. fine-textured sediment. Thus, accumulation of sandy alluvium might be expected to cause reductions in available levels of soil nutrients, such as nitrate (NO3) and potassium (K), that are easily leached.
Many of the forest sedimentation studies have taken place in the floodplains of higher-order rivers and streams. However, riparian zones of headwaters are crucially important in defining down-basin water quality (Brinson, 1993), but are much less studied (Wardrop and Brooks, 1998). Depending on the magnitude of sediment-driven changes within headwater riparian systems, there is great potential for alterations in the filtration function there and, consequently, greater risk of water quality impairment in the associated, higher-order streams. However, there is no information available regarding how ephemeral riparian forests might respond to anthropogenic stresses and, in particular, how nutrient circulation in any riparian forest might be altered by sedimentation. Consequently, the objectives of this investigation were to examine relationships between key biogeochemical attributes and sediment accumulation in ephemeral riparian forests and to identify levels of sediment accumulation beyond which nutrient cycling attributes change markedly.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The main source of sediment to riparian areas at Fort Benning is unpaved roads and trails that are used by military traffic. The roads are generally established along ridgelines and, consequently, the sandy substrate on road edges and near turn-outs erodes quite readily. Sediment moves downslope primarily in channelized flow through ephemeral streams and accumulates where channels widen and surface roughness increases. Roughness is mainly attributable to woody stems and, to a lesser extent, coarse woody debris. Almost no herbaceous vegetation is present in these erosion features.
Riparian forests of the study site consist of uneven-age, primarily deciduous cover with common occurrences of red maple (Acer rubrum L.), tag alder [Alnus serrulata (Aiton) Willd.], flowering dogwood (Cornus florida L.), sweetgum (Liquidambar styraciflua L.), yellow poplar (Liriodendron tulipifera L.), sweet bay (Magnolia virginiana L.), wax myrtle (Myrica cerifera L.), black gum (Nyssa sylvatica Marshall), loblolly pine (Pinus taeda L.), water oak (Quercus nigra L.), red oak (Quercus rubra L.), and post oak (Quercus stellata Wang), among other species. This is in contrast to the dry uplands dominated by mixed pine forests. Soils consist of poorly drained Bibb (coarse-loamy, siliceous, active, acid, thermic Typic Fluvaquents) and Chastain (fine, mixed, semiactive, acid, thermic Fluvaquentic Endoaquepts) series within riparian areas and well- to excessively drained Troup (loamy, kaolinitic, thermic Grossarenic Kandiudults), Lakeland (thermic, coated Typic Quartzipsamments), and Cowarts (fine-loamy, kaolinitic, thermic Typic Kanhapludults) soils on upper slopes and ridges. Deposition of sandy alluvium in upper reaches of some catchments may have created better-drained conditions than originally existed.
Nine ephemeral streams were selected to cover a range of disturbance and sedimentation conditions and included two reference catchments and seven that were disturbed in the upper portions. The presence of alluvial fans and buried lower portions of tree stems were used as evidence of sedimentation. In reference catchments, these traits were absent. Paired, circular plots (0.04 ha) were established within each ephemeral catchment with one delineated in the upper portion near the stream origin where sediment was most likely to be received. Another plot was installed lower within each catchment beyond visual evidence of sediment deposition. The latter served as relative (within-catchment) controls for comparisons with upper plots (Fig. 1) .
|
|
Decomposition was studied using the litter bag approach (Swift et al., 1979). Foliar litter was collected in traps (0.5 m2, three per plot), air-dried, and sorted by species on an individual plot basis. Eight 13- x 13-cm litterbags composed of nylon mesh (6-mm opening on the upper side and 2-mm opening on the side in contact with the soil) were filled with a mix of litter that reflected the tree species composition of each plot. Litterbags on two of the plots were destroyed by fire and animals. Bags were collected at intervals of 0, 2, 4, 10, 16, 25, 36, and 48 wk between April 2002 and March 2003. The time 0 collection was used to estimate handling loss and starting conditions in terms of C and N content. After removal from the field, bags were placed inside paper sacks for transport to the laboratory. Litter was removed from the bag, gently washed if necessary to remove extraneous material, oven-dried at 70°C for 48 h, and weighed. Total C and N were determined on litter samples using thermal combustion on a PerkinElmer (Wellesley, MA) 2400 Series II CHNS/O analyzer.
Nitrogen mineralization was estimated using the in situ methodology outlined by Hart et al. (1994). Seven sampling collections (at approximately 3-mo intervals) were made between April 2002 and July 2003. During each sampling, four soil samples (approximately 150 mL each) were collected from each plot to a depth of 7.5 cm and two were bagged and put on ice for transportation to Auburn University laboratories. These were extracted for NH4–N and NO3–N to ascertain initial levels of mineral N prior to in situ incubations. The additional two samples were placed in polyethylene bags and returned to the soil where they were buried to the same depth. The latter samples were collected after 30 d and extracted to estimate the amount of N mineralized (sum of NH4–N and NO3–N) per day over that time period.
Microbial biomass was determined using the chloroform-fumigation technique (Vance et al., 1987; Brooks et al., 1985). Soil sampling occurred in tandem with N mineralization collections and consisted of two pairs of soil samples (7.5 cm deep, 75 mL each) being gathered with the use of a push probe from each plot. Samples were bagged, put on ice, and transported to laboratories for analyses. One sample within each pair was subjected to chloroform fumigation and then all samples were extracted with K2SO4. Extracts were analyzed for total C using a Dohrmann DC-80 total organic carbon analyzer (Rosemount-Dohrmann Analytical, Santa Clara, CA) and total Kjeldahl N (Bremner, 1996).
During March and September 2003, 0.1-m2 samples of unconfined forest floor were collected from upper plots in all catchments and sampled for arthropod populations. The Berlese–Tullgren funnel extraction technique (Seastedt and Crossley, 1980) was used. Litter samples were subjected to constant light and heat for 5 d and all extracted arthropods were stored in 70% ethyl-alcohol, counted, and subsampled for identification to the order level.
Two temperature sensors (HOBO; Onset Computer Corporation, Pocasset, MA) were placed on the soil surface of upper and lower plots within one reference catchment and three disturbed. All sensors were encased in PVC capsules. Temperatures were recorded hourly and downloaded once a month.
Statistical Analyses
Regression analysis was used to estimate decomposition rates (i.e., y = e–kt, where y = proportion of original carbon remaining, k = rate coefficient, and t = weeks) separately for each plot across the 48-wk period. Regression relationships between sedimentation rates vs. decomposition rate coefficients, microbial biomass C and N, mineralized nitrogen, and arthropod counts were examined using either the nonlinear model previously listed, simple linear, or natural log transformations. Comparisons of upper vs. lower plots were made within each catchment for N mineralized and microbial C and N using analyses of variance. All analyses were conducted using SAS (SAS Institute, 1985) and all probability levels are reported and considered significant if less than P < 0.10.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Temperatures
Average temperatures at 0200 h were identical for upper and lower plots in the D1 through D7 catchments and averaged 21°C annually. The lower plot in the reference catchments (R1 and R2) averaged 19°C at 0200 h vs. 21°C on the upper. Temperatures at 1400 h exhibited more variation among plots with upper locations in disturbed catchments averaging 27°C and corresponding lower plots averaging 25°C. Temperature averages on upper and lower locations within the reference were 25 and 23°C, respectively. The range of temperatures showed greater divergence between upper and lower plots in disturbed catchments (i.e., 7 and 4°C, respectively). Corresponding values for the reference catchments were both 4°C. The wider range of temperatures and slightly higher maximums on upper plots in disturbed catchments may create harsher conditions there for some microbial populations (Swift et al., 1979).
Decomposition
Decomposition rates for foliar litter on upper and lower plots averaged 0.95 and 0.91 yr–1, respectively, and were not significantly different. Rates within this range are near the average of 1.01 noted for wetland forests in the U.S. Southeast (Lockaby and Walbridge, 1998) and well within ranges noted for wetland forests in the temperate zone (Brinson, 1990).
The regression relation between decomposition rates and long-term sedimentation rates was significant (r2 = 0.62, P = 0.04) (Fig. 3) . Apparently, a rapid decrease in decomposition rates occurred even with the smallest sedimentation levels observed (i.e., 0.20–0.32 cm yr–1). Rates of carbon loss appear to approach a reduced equilibrium at sediment accumulations above 0.50 cm yr–1. In comparing decomposition among several floodplain forests in the southeastern United States, Baker et al. (2001) observed the lowest rates of foliar litter decomposition on the Cache River floodplain in Arkansas, a system characterized by high rates of sediment accumulation (Hupp, 2000).
|
The relationship between net mineralized N and sedimentation rates on upper plots was significant (r2 = 0.41, P < 0.0001) (Fig. 4) . Comparisons of upper vs. lower plots within catchments revealed only one significant difference (Table 1). Net mineralized N levels on the lower plot within catchment R2 were significantly higher than those of the upper. The lack of differences between upper and lower plots may be in part due to generally high levels of variation associated with mean estimates (Table 1).
|
Comparisons between upper and lower plots within catchments suggested that, at higher sedimentation rates (i.e., above 0.32 cm yr–1), microbial C and N were reduced on upper plots (Table 2). In terms of microbial C, there were significant differences between upper and lower plot values on all catchments where upper plots had received sediment at rates higher than 0.32 cm yr–1. There were no significant differences within the reference catchments or D6 and D7, the two having the lowest sediment accumulation rates apart from the references. In all cases where there were differences, the lower plot values were higher. Similarly, catchments with sedimentation rates above 0.32 (with the exception of D3) displayed significant differences between upper and lower plots in terms of microbial N (Table 2). As was the case with microbial C, where significant differences did occur, lower plot values were higher.
|
|
|
Regardless of sedimentation level, samples were dominated by orders Acarina, Collembola, Coleoptera, and Diptera. Acarina often dominates in decomposition studies because they are the most abundant in nature (Seastedt et al., 1989). Collembola also was a dominant order in the Atchafalaya.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Relationships between long-term (25-yr average) sedimentation rates and biogeochemical indices were stronger than for current year rates. This suggests that some biogeochemical alterations are driven by processes that occur over longer time periods such as changes in litter quality reflecting vegetation transitions. Clarification of the nature of the long-term influence would be a worthy goal of future studies.
Sedimentation rates as low as 0.20 cm yr–1 are not uncommon in floodplain forests (Hupp, 2000). This is particularly true in catchments with significant amounts of disturbance such as urbanization and agriculture. These results imply that proper road maintenance should receive a very high priority on military installations and other areas where highly erodible soils occur. Given the projected increases in human populations, a greater proportion of river basins throughout the world will likely be subjected to significant levels of disturbance. Since even relatively low levels of sediment accumulation can alter some aspects of biogeochemical cycling in riparian forests, key functions such as water filtration might be vulnerable as well.
Although we did not measure surface roughness, our field observations on impacted sites suggest that higher levels of sediment may physically bury coarse woody debris, forest floors, and perhaps induce mortality in small vegetation. This would result in reduced roughness and, theoretically, less effective filtration since surface runoff velocities would be slowed to a lesser degree. Consequently, surface runoff containing sediment would be more likely to reach streams.
During restoration of riparian forests that are likely to receive high sediment inputs, this possibility should be taken into account and consideration should be given to selection of vegetation species that are tolerant of sediment accumulation.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  R. L. Jolley, B. G. Lockaby, and G. G. Cavalcanti Productivity of Ephemeral Headwater Riparian Forests Impacted by Sedimentation in the Southeastern United States Coastal Plain J. Environ. Qual., March 25, 2009; 38(3): 965 - 979. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
