Hydrobiologia 400: 123–128, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.

123

Processing of native and exotic leaf litter in two Idaho (U.S.A.) streams Todd V. Royer1 , Michael T. Monaghan2 & G. Wayne Minshall1 1 Department 2 Present

of Biological Sciences, Idaho State University, Pocatello, ID 83209-8007, U.S.A. address: Department of Limnology, EAWAG/ETH, CH-8600 Dübendorf, Switzerland

Received 6 October 1998; in revised form 23 March 1999; accepted 1 April 1999

Key words: exotic species, leaf litter, C/N ratios, decomposition, stream ecology

Abstract The Russian olive tree (Elaeagnus angustifolia L.) was brought to the western United States from Eurasia during the early to mid-1900s, and has since become a common member of many riparian communities in Idaho. We compared leaf chemistry and in-stream processing of Russian olive leaves (exotic) and various species of native leaves in one hardwater and one relatively softwater Idaho stream. Measurements using air-dried leaves showed that Russian olive contained the greatest concentration of nitrogen, approximately 1.6% of the dry mass, whereas the native species each contained less than 1.0% nitrogen. The C/N ratio of Russian olive was <30, whereas the natives each had C/N ratios greater than 40. Results from the hardwater stream indicated no difference in 30-day loss of AFDM between Russian olive and the native leaves (dogwood and aspen). In the relatively softwater stream, the Russian olive leaves were processed significantly slower than the native leaf species (cottonwood). The results indicate that a replacement of native riparian trees by exotics, such as Russian olive, may result in slower rates of leaf processing in Idaho streams but that the effect may vary among streams. When comparing the processing of native and exotic leaf litter, initial nitrogen concentrations and initial C/N ratios of the leaves did not appear to be accurate indicators of relative decay rates.

Introduction Many streams throughout North America receive substantial inputs of detrital carbon from the terrestrial environment, particularly in the form of deciduous leaf-litter (Cummins, 1974; Benfield, 1997). The rate at which leaves are processed in a stream is controlled by several factors, including the concentration of dissolved nutrients in the water, the thermal regime of the stream, the concentration of nitrogen (N) in the leaf material, the amount of N relative to the amount of carbon (C/N ratio) and the presence of allelochemicals in the leaves (Webster & Benfield, 1986). Those leaf species with relatively high N concentration are expected to decay more rapidly than those with relatively low N. Accordingly, litter with a relatively low C/N ratio is expected to decay more rapidly than litter with a relatively high C/N ratio. For all but the most autotrophic streams, decomposition of terrestrial leaf litter is an important ecosys-

tem process and the exclusion of leaf litter has been shown to disrupt the normal productivity of a stream (Wallace et al., 1997). A change in the species of leaves entering a stream is expected to alter the structure of the macroinvertebrate community, particularly detritivores (Cummins et al., 1989; Gregory et al., 1991). Such a change in litter type also may influence organic matter dynamics within the stream (Webster et al., 1995). Oak (Quercus spp.) replaced American chestnut (Castanea dentata) in many riparian areas of eastern North America following a fungal blight in the early 1900s that virtually eliminated chestnut. Smock & MacGregor (1988) found that oak leaves were processed more slowly than chestnut and that stream invertebrates supplied with oak leaves grew more slowly than those given native chestnut leaves. Although oak and chestnut both were native trees, the results from Smock and MacGregor (1988) suggest that the type of leaf litter entering a stream is no less important than the total amount of input.

124 The Russian olive tree (Elaeagnus angustifolia L.), native to Eurasia, was brought to the western United States during the early to the mid-1900s (Christensen, 1963). In Idaho, it was (and continues to be) planted widely and subsequently has become a common member of many riparian communities, particularly in the southern portion of the state. In addition to intentional planting, the altered flow regimes that have followed impoundment of most western rivers have led to a reduction of native riparian trees, primarily cottonwoods (Populus spp.), while favoring the establishment of Russian olive (Howe & Knopf, 1991; Shafroth et al., 1995). Because of decreased cottonwood recruitment and because Russian olive can germinate in a wider range of physical habitats than can cottonwoods (Shafroth et al., 1995), the probable outcome is extensive replacement of native cottonwoods by Russian olive. Such replacement is predicted to occur within 50– 100 years from the present (Howe & Knopf, 1991). Despite the importance of riparian leaf litter to stream function, the possible effects of Russian olive invasion on leaf processing in streams remains unstudied. The objectives of this study were to: 1. compare leaf chemistry of Russian olive leaves with native leaves; and 2. measure the 30-day processing rate of Russian olive leaves versus native riparian leaves in two Idaho (U.S.A.) streams. Additionally, the two sites allowed us to examine the processing of native versus exotic leaves under different water chemistry conditions.

Methods We conducted the study in Mink Creek, Bannock County, and the Big Wood River, Blaine County, Idaho. Mink Creek has a mixed riparian community that includes aspen (Populus tremuloides) and dogwood (Cornus stolonifera). Downstream of the study site on Mink Creek, Russian olive has become established but it does not occur at or upstream of the study site. The Big Wood River in central Idaho has a riparian community predominantly composed of cottonwood (Populus trichocarpa); Russian olive does not occur along the Big Wood. This location was of interest because the replacement of cottonwood with Russian olive is underway in similar cottonwooddominated river systems in the western U.S. The Big Wood River flows through Quaternary alluvium of

granite and quartz sands. Mink Creek is a montane hardwater stream, flowing through limestone and quartzite valley fill. Russian olive, cottonwood, aspen and dogwood leaves were gathered in October of 1996 at the time of natural abscission. All leaves gathered for the study were senescent but had not fallen from the trees. The native leaves were gathered from the riparian areas in which they were studied and the Russian olive leaves were gathered from Deep Creek in southern Idaho where they became established during the 1970s. Leaves were air dried at room temperature for >7 days and were incubated in the streams using nylon litter bags with a mesh diameter of 5 mm. Each litter bag contained approximately 4 g of air-dried leaves. The litter bags were placed in the Big Wood on 31 October 1996, and in Mink Creek on 5 November 1996, coincident with natural seasonal inputs of leaf litter. At both sites, three additional bags of each leaf type were used for correction of handling losses incurred during the transport of the leaf-packs to the streams. The litter bags were tied to bricks and placed in riffle habitats. Six litter bags of each species were removed after 30 days (n=6) and immediately placed into plastic bags, put on ice and returned to the laboratory within 2–4 h. In the laboratory, sediments and invertebrates were rinsed from the remaining leaf material; invertebrates were preserved in 5% formalin. Loss of leaf material (i.e. in-stream processing) was measured as the change in ash-free dry mass (AFDM) over the 30-day period. The AFDM was determined as the loss of mass after combustion at 550 ◦ C for 2 h. For each litter bag, initial AFDM was calculated from a regression equation relating dry mass to AFDM; equations were developed for each species using leaves collected at the same time as the leaves used to measure processing. Statistical differences in the amount of material processed were determined using one-way ANOVA for the Big Wood River and a t-test for Mink Creek; all data were arc sin (x0.5 ) transformed (Zar, 1984). Invertebrates associated with the litter bags from the Big Wood River were sorted, counted and identified to the lowest feasible taxonomic unit, typically genus; invertebrates were not examined from Mink Creek. Following identification, all invertebrates were dried at 50 ◦ C and the biomass measured on a Cahn 25 electrobalance using the 0–20 mg range. Differences in the number and biomass of invertebrates between Russian olive and cottonwood leaves were examined using a t-test on natural log (x + 1) transformed data.

125 Table 1. Physical and chemical characteristics of the Big Wood River and Mink Creek, Idaho. Measurements were collected on 31 October 1996 and 05 November 1996, respectively

Elevation (m) Stream Order Discharge (l/s) pH Specific Conductance (µS/cm) Alkalinity (mg CaCO3 /l) NO2 +NO3 (mg/l) Total PO4 (mg/l)

Big Wood River

Mink Creek

1840 5 2600 8.2 243 72 0.030 0.026

1560 3 150 8.5 519 124 0.044 0.051

The N and C concentrations were determined by drying leaves of each species at 50 ◦ C, grinding the material with a Thomas-Wiley intermediate mill, and analyzing with a Fisons Instruments elemental analyzer (model NA 1500). The % N and C/N ratio were determined both for air-dried leaves and for leaves that had been leached in distilled water for 24 h. The leaching was accomplished by placing approximately 2 g of air-dried leaves of each species in beakers containing 500 ml of distilled water (5 replicate beakers per species). The leaves were removed from the beakers after 24 h and the N and C concentrations determined as described above. Data were natural log (x + 1) transformed (Zar, 1984) and statistical differences determined using one-way ANOVA. Water temperature at each stream was recorded hourly with a datalogger (Onset Corporation, model STEB08). Total degree-days were determined by summing the mean daily temperatures over the 30 days. Discharge and water chemistry were measured on the day the litter bags were placed in the streams. Discharge was determined using an Ott current meter (type C2) as described in Platts et al., (1983). Specific conductance and pH were measured on-site with portable meters (Orion Inc.), alkalinity was determined in the laboratory using acid titration. Samples of NO2 +NO3 were filtered (0.45 µm) and frozen; total PO4 samples were kept on ice following acidification to pH <2. Samples were analyzed at the Desert Research Institute, Reno, Nevada.

Results The two streams we examined differed in elevation, size and water chemistry (Table 1). In particular, Mink

Figure 1. Initial mean (+/− 1SE) concentration of (a) nitrogen and (b) the C/N ratio of leaf types used in the study; for both variables, all leaf types were significantly different (p<0.01; n=5) from all others based on Tukey HSD. (c) The post-leaching C/N ratio for each leaf type; Russian olive leaves were significantly lower (p<0.05, Tukey HSD) than any of native species.

Creek is a hardwater stream, chemically more rich than the Big Wood River based on specific conductance, alkalinity and PO4 . Due to the higher elevation of the Big Wood River, it was considerably colder during the course of the study than was Mink Creek. Over the 30-day incubation period, Mink Creek accumulated 107 degree days and mean daily temperature ranged from 0.4 to 6.0 ◦ C, while the Big Wood River accumulated 34 degree days and mean daily temperature ranged from 0.0 to 3.0 ◦ C. The initial concentration of N in the air-dried leaves differed significantly among all four species (one-way ANOVA, p<0.001). Russian olive contained the most N, approximately 1.6% of the dry mass (Figure 1a). The native species each contained less than 1.0% N. In particular, Russian olive contained greater

126 than 2-fold as much leaf-N as did cottonwood. As with N, the C/N ratios also differed significantly among each of the four species (one-way ANOVA, p<0.001). The C/N ratio of Russian olive was <30, whereas the natives each had C/N ratios greater than 40 (Figure 1b). The greatest difference in C/N ratios was observed between Russian olive and cottonwood, with cottonwood approximately 2.5 times greater than Russian olive. Leaching of the leaves for 24 h reduced the C/N ratio of dogwood and aspen leaves, increased the ratio for cottonwood leaves, and did not affect the ratio of Russian olive leaves. Despite the changes in actual C/N ratios, the relative position of each species did not change following the 24 h of leaching (Figure 1c). Based on the N concentrations and C/N ratios, Russian olive would be expected to decay faster than any of the native riparian species we examined. Despite the differences in leaf chemistry, there was no significant difference in the amount of mass lost among dogwood, aspen and Russian olive in Mink Creek (one-way ANOVA, p=0.24). All three leaf types had 30–40% of the initial AFDM remaining after a 30-day incubation in Mink Creek (Figure 2a). Unlike the results from Mink Creek, in the Big Wood River there was a significant difference in processing rates between Russian olive and the native cottonwood (t-test, p<0.001). After 30 days in the Big Wood River, Russian olive leaves lost approximately 40% of the initial AFDM, whereas cottonwood lost >60% (Figure 2b). The mean number of invertebrates on the leaf packs in the Big Wood River was approximately 650 and 400/g AFDM for cottonwood and Russian olive, respectively. However, this difference was not statistically significant (p=0.28) because of the large amount of variability in invertebrate density among the leaf-packs. Invertebrate biomass also was highly variable among leaf-packs in the Big Wood River and did not differ statistically (p=0.15) between Russian olive (x¯ = 15 mg/g AFDM) and cottonwood (x¯ = 23 mg/g AFDM), although the means suggested that the cottonwood leaves supported a greater invertebrate population. There was no observed difference in the composition of the invertebrate taxa found on Russian olive versus cottonwood leaf packs.

Discussion Our results indicate that a replacement of native riparian trees by exotics, such as Russian olive, may

result in slower rates of leaf processing in some Idaho streams. In addition to the present study, Minshall et al. (1982) found that non-native hickory leaves (Carya tomentosa) placed at several locations in the Salmon River, Idaho, were processed more slowly than native litter, including conifer needles. Vitousek (1990) suggested that population-level changes, such as invasion by exotics, could affect ecosystem-level processes. In streams, slowed processing of leaf litter could alter the biological communities that are linked to the timing of litter inputs (Cummins et al., 1989), and could alter the amount, timing and quality of organic matter available for transport downstream. Because transported organic carbon, derived in part from the breakdown of leaf litter, provides a longitudinal linkage within stream ecosystems (Minshall et al., 1983; Cushing et al., 1993), a change in the type of litter entering a headwater stream could have effects on both local and downstream reaches. Water chemistry exerts a strong influence on leaf processing in streams. For example, Suberkropp & Chauvet (1995) reported faster leaf processing in hardwater streams than in softwater streams. The water chemistry of the Salmon River during the study by Minshall et al., (1982) was similar to that measured in the Big Wood River during the present study. Both the Salmon and Big Wood rivers showed significantly slower processing of exotic leaves than native leaves and both streams contain relatively softwater. Conversely, Mink Creek contains hardwater, relative to the Big Wood River (see Table 1), and showed no difference in processing rates between exotic and native leaves. These results suggest that replacement of native trees by exotics may differentially affect litter processing and organic matter dynamics among different streams. The role of water chemistry in these changes needs further investigation across a wider array of stream types. Initial N concentrations and C/N ratios often have been suggested as indicators of relative processing rates among different species of leaves entering streams (e.g. Kaushik & Hynes, 1971; Suberkropp et al., 1976). Based on the initial N concentration and initial C/N ratios from air-dried leaves of the four species we examined, the predicted order of processing was Russian olive > cottonwood in the Big Wood River, and Russian olive > aspen > dogwood in Mink Creek. In neither stream was the predicted order of processing observed. A similar failure of initial C/N ratios to predict relative decay rates was reported by Chauvet (1987). In the present study, native cotton-

127

Figure 2. Mean (+/− 1SE) amount of leaf material remaining after 30 days of in-stream processing. In Mink Creek (a) there was no significant difference among the three species (p=0.24). In the Big Wood River (b) significantly more Russian olive material remained compared to the native cottonwood leaves (p<0.001).

wood leaves in the Big Wood River were processed significantly faster than Russian olive leaves, despite cottonwood leaves having a more than 2-fold greater initial C/N ratio. The post-leaching C/N ratios appeared to be more accurate indicators of relative processing rates than did the initial C/N ratios. Specifically, aspen and dogwood leaves were processed at the same rate, although dogwood leaves had a significantly lower initial C/N ratio (Figure 1b; Tukey HSD p<0.001). However, there was no difference between aspen and dogwood leaves in terms of the post-leaching C/N ratios (Figure 1c; Tukey HSD p=0.232). The leaching of soluble compounds from leaves during the first 24–48 h in water is a physical process that happens independent of any biological action. Thus, when organisms actually begin to decompose leaves it is the postleaching C/N ratios that will most accurately represent the quality of the leaves. Current evidence suggests that leaching may simply be a methodological artifact of air-drying leaves (Gessner & Schwoerbel, 1989; Gessner, 1991); however, when using air-dried leaves, it appears that measures of leaf quality may be most useful if obtained after a 24–48 h incubation in water. In the Big Wood River, the density of invertebrates was not significantly different between the two leaf species, suggesting that both were equally palatable or that the invertebrates were using the leaves primarily as habitat rather than as a food resource (see Richard-

son, 1992). Thus, the significantly greater mass loss of cottonwood leaves, compared to Russian olive leaves, likely was due to microbial processing rather than invertebrate feeding. Llinares et al. (1993) found that allelochemicals in the leaves of Russian olive inhibited nitrification by soil microbes. Such allelochemicals also may have inhibited microbial colonization and/or breakdown of Russian olive leaves in the streams we examined, possibly overriding the effect of C/N ratios. Invertebrate data are not available for the leaf packs from Mink Creek, but the equal processing rates observed among the three leaf species would not likely have occurred if there was a strong invertebrate preference for any one of the leaf species.

Acknowledgments Field and laboratory assistance was provided by Kate Bowman, Christine Fischer, Douglas Ottke, Christina Relyea and Scott Relyea. This study was supported in part by a grant from the Idaho Department of Health and Welfare, Division of Environmental Quality.

References Benfield, E. F., 1997. Comparisons of litterfall input to streams. J. n. am. Benthol. Soc. 16: 104–108.

128 Chauvet, E., 1987. Changes in the chemical composition of alder, poplar and willow leaves during decomposition in a river. Hydrobiologia 148: 35–44. Christensen, E. M., 1963. Naturalization of Russian olive (Elaeagnus angustifolia L.) in Utah. Am. Midl. Nat. 70: 133–137. Cummins, K. W., 1974. Structure and function of stream ecosystems. Bioscience 24: 631–641. Cummins, K. W., M. A. Wilzbach, D. M. Gates, J. B. Perry & W. B. Taliaferro, 1989. Shredders and riparian vegetation. Bioscience 39: 24–30. Cushing, C. E., G. W. Minshall & J. D. Newbold, 1993. Transport dynamics of fine particulate organic matter in two Idaho streams. Limnol. Oceanogr. 38: 1101–1115. Gessner, M. O., 1991. Differences in processing dynamics of fresh and air-dried leaf litter in a stream ecosystem. Freshwat. Biol. 26: 387–398. Gessner, M. O. & K. J. Schwoerbel, 1989. Leaching kinetics of fresh leaf-litter with implications for the current concept of leaf-processing in streams. Arch. Hydrobiol. 115: 81–90. Gregory, S. V., F. J. Swanson, W. A. McKee & K. W. Cummins, 1991. An ecosystem perspective of riparian zones. Bioscience 41: 540–551. Howe, W. H. & F. L. Knopf, 1991. On the imminent decline of Rio Grande cottonwoods in central New Mexico. Southwestern Nat. 36: 218–224. Kaushik, N. K. & H. B. N. Hynes, 1971. The fate of the dead leaves that fall into streams. Arch. Hydrobiol. 68: 465–515. Llinares, F., D. Muñoz-Mingarro, J. M. Pozuelo, B. Ramos & F. Bermúdez de Castro, 1993. Microbial inhibition and nitrification potential in soils incubated with Elaeagnus angustifolia L. leaf litter. Geomicrobiology Journal 11: 149–156. Minshall, G. W., R. C. Petersen, K. W. Cummins, T. L. Bott, J. R. Sedell, C. E. Cushing & R.L. Vannote, 1983. Interbiome comparison of stream ecosystem dynamics. Ecol. Monogr. 53: 1–25. Minshall, G. W., J. T. Brock & T. W. LaPoint, 1982. Characterization and dynamics of benthic organic matter and invertebrate functional feeding group relationships in the upper Salmon River, Idaho (U.S.A.). Int. Rev. ges. Hydrobiol. 67: 793–820.

Platts, W. S., W. F. Megahan & G. W. Minshall, 1983. Methods for evaluating stream, riparian, and biotic conditions. United State Department of Agriculture, General Technical Report INT-138, 70 pp. Richardson, J. S., 1992. Food, microhabitat, or both? Macroinvertebrate use of leaf accumulations in a montane stream. Freshwat. Biol. 27: 169–176. Shafroth, P. B., G. T. Auble & M. L. Scott, 1995. Germination and establishment of the native plains cottonwood (Populus deltoides Marshall subsp. monilifera) and the exotic Russian olive (Elaeagnus angustifolia L.). Conserv. Biol. 9: 1169–1175. Smock, L. A. & C. M. MacGregor, 1988. Impact of the America chestnut blight on aquatic shredding macroinvertebrates. J. n. am. Benthol. Soc. 7: 212–221. Suberkropp, K. & E. Chauvet, 1995. Regulation of leaf breakdown by fungi in streams: influences of water chemistry. Ecology 76: 1433–1445. Suberkropp, K., G. L. Godshalk & M. J. Klug, 1976. Changes in the chemical composition of leaves during processing in a woodland stream. Ecology 57: 720–727. Vitousek, P. M., 1990. Biological invasions and ecosystem processes: towards an integration of population biology and ecosystem studies. Oikos 57: 7–13. Wallace, J. B., S. L. Eggert, J. L. Meyer & J. R. Webster, 1997. Multiple trophic levels of a forest stream linked to terrestrial litter inputs. Science 277: 102–104. Webster, J. R., J. B. Wallace & E. F. Benfield, 1995. Organic processes in streams of the eastern United States. In Cushing, C. E., K. W. Cummins & G. W. Minshall (eds), River and Stream Ecosystems (Ecosystems of the World, vol. 22) Elsevier Science, Amsterdam: 117–187. Webster, J. R. & E. F. Benfield, 1986. Vascular plant breakdown in freshwater ecosystems. Ann. Rev. Ecol. Syst. 17: 567–594. Zar, J. H., 1984. Biostatistical Analysis, 2nd edn. Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 718 pp.