Riparian leaf litter processing by benthic macroinvertebrates in a woodland stream of central Chile Procesamiento de detritus ripariano por macroinvertebrados bent 6 nicos en un estero boscoso de Chile central

Leaf litter input from riparian landscapes has been identified as both a major energy flow to stream ecosystems and as a food source for stream macroinvertebrates. In riparian landscapes of woodland streams of central Chile, the native deciduous hardwoods are being artificially replaced by exotic coniferous trees at a large spatial scale. It is suggested that this process has a significant impact on the stream communities of central Chile. Today, exotic plantations occur throughout central Chile, with Pinus radiata (D. Don) (Monterrey pine) accounting for about 80% of the more than I ,800,000 ha of exotic forests. The objective of this paper was to analyze the effect of the litter beds of a dominant native species (Nothofagus pumilio) and an exotic species (P. radiata) on the detritus processing carried out by benthic macroinvertebrates, in an experimental catchment of central Chile (Rucue Creek; 36° 26'00" S, 71° 35' 40" W). Results revealed that processing rates of native leaf packs are higher than rates of coniferous leaf packs, suggesting that the replacement of the native hardwoods by exotic coniferous riparian flora has an important impact on the stream energy flow in central Chile. The decay rate coefficients (k) were 0.0072 for N. pumilio, and 0.0027 for P. radiata. The greater abundance and biomass of shredders per gram of leaf pack of native Nothofagus would explain the differences in leaf processing rates, especially through the activity of two Plecoptera Gripopterygidae, Limnoperlajaffueli andAntarctoperla michaelseni.


INTRODUCTION
Riparian environments are both ecotones between terrestrial and aquatic ecosystems as well as habitat corridors across regions.These transitional environments serve diverse functions and have different values depending on their physical, biological, and cultural setting (Malanson 1995).Leaf litter input from riparian landscapes has been identified both as a major energetic component of stream ecosystems and as a food source for stream macroinvertebrates (Flory & Milner 1999).In small, heavily shaded streams, there is normally insufficient light to support substantial instream photosynthesis so energy pathways are largely supported by imported (allochthonous) energy sources (Benfield 1996).The bulk of imported energy enters such streams as autumnal leaf fall although additional leaf material may slide or blow into the stream from riparian zones during the rest of the year (Lamberty & Gregory 1996).Leaves falling into streams may be transported over short distances but usually are caught by structures in the streambed forming leaf litter beds (Flory & Milner 1999).These leaves are processed in place by components of the stream invertebrate and microbial community in a series of well documented steps (e.g., Anderson & Sedell 1979, Jones & Momot 1981, Anderson & Dan ell 1982, Wissmar 1991, Culbert & France 1995, France 1995, Benfield 1996, Flory & Milner 1999).
Leaf processing by benthic macroinvertebrates is an integrative ecosystem-level process because it links elements as leaf species, microbial activity, and physical and chemical features of the stream (Friberg & Winterbourn 1997).The major result of this linkage is that whole leaves are converted into fine particles, which are then distributed downstream and used as energy source by various consumers.Rates of leaf litter processing have been measured in a variety of stream types in the northern hemisphere (e.g., Petersen & Cummins 1974, Sedell et al. 1975, Benfield et al. 1977, Barlocher et al. 1978, Short et al. 1980).These and other studies have indicated that macroinyertebrates may play an important role in the processing of leaf material.Furthermore, leaf litter has been shown to be an important factor in the distribution of benthic macroinvertebrates in streams (Flory & Milner 1999).
Although many aspects of the structure and function of forest stream ecosystems have been investigated in recent years (Cushing et al. 1995), the effect of forest type on benthic community structure is still poorly understood (Friberg & Winterbourn 1997).Nevertheless, the replacement of native forest or afforestation of grassland catchments with coniferous tree species is an increasingly common practice in many countries.This is especially so where economic policies are resulting in a reduction of agricultural production (e.g., Kuusela 1994 ).The question therefore arises: does the catchment-scale establishment of exotic plantation forests alter the biotic communities of streams?Several investigations in the northern hemisphere indicate that conifer plantations can affect benthic communities in a negative way compared with other types of catchment vegetation, such as hardwoods (e.g., Harriman & Morrison 1982, Ormerod et al. 1989).
The present study of leaf litter processing was conducted in central Chile where conifer plantations started early in the 20th century.Today, exotic plantations occur throughout central Chile, with Pinus radiata (D.Don) (Monterrey pine) accounting for about 80 % of the more than 1,800,000 ha of exotic forests (Gonzalez et al. 1999).In riparian landscapes of woodland streams of central Chile, the native deciduous tree species (e.g., Nothofagus pumilio) are being artificially replaced by large-scale by industrial plantations of Pinus radiata, suggesting that this process of change in land cover has an important impact on stream community and energetics in central Chile.The objective of this paper was to compare the processing rate of leaf litter beds of a dominant native tree species (N.pumilio) and an exotic tree species (P.radiata) by benthic macroinvertebrates in an experimental basin of central Chile (Rucue Creek).Currently, natural forests dominate this experimental area.However, and based on current trends, forest cover dominated by N. pumilio will probably be replaced by P. radiata in the next decades, affecting downstream ecosystems.

Study site
Rucúe Creek is located in central Chile in the Biobfo river catchment.It is a third-order stream with a watershed area of 227 km 2 and 32.5 km in length.The study site is located at 36° 26' 00" S, 71 o 35' 40" W, at an elevation of about 1,050 m above sea level where the stream emerges from a narrow valley with a dense canopy of lenga (N.pumilio ), and cipres de la cordillera (Austrocedrus chilensis).Pinus radiata is a dominant tree species on industrially managed areas and in the lower part of the basin, without influencing the study site.All of the higher part of the basin is covered by native forest which influences the study site.
The study site corresponded to a transitional altitudinal zone between two types of forest cover; N. pumilio was the dominant tree species at the upper part of the site (1 ,000-1,900 m), whereas Austrocedrus chilensis was the dominant species at the lower part (700-1 ,000 m).Leaves of N. pumilio fall into the stream and are transported down to the study site.Some of them are caught by structures of the stream bed forming leaf litter beds.Leaf litter beds in the study site were mainly composed by N. pumilio (> 98 % dry-weight), followed by A. chilensis and unidentifiable detritus ( < 2% dry-weight).According to Hildebrant-Vogel et al. ( 1990), N. pumilio presents one of the most extensive latitudinal range of distribution along the temperate forest of South America, between 35o 35' and 55° 30' S (ca.2,300 km north-south distribution).In southern Chile the altitudinal range varies between 0 to 700 m (Cuevas & Arroyo 1999).However, near to the northern extreme of distribution altitude varies between 1,300 and 2,000 m (Donoso 1993 ), which correspond to the altitudinal range observed in the study area.
Stream width varied from 4 to 10 m with an average slope of 7 %.The substrate consisted mainly of rubble in riffles and a few pools.Stream flow was moderated during the study period (3.50 to 10.0 m 3 sˉ¹).Nutrients were present at low concentrations (total-P = 0.01-0.02mg 1ˉ¹; total-N, 0.11-0.14mg I-1 ).The water was soft (9.2-10.1 mg I-' CaC0 3 ), and the pH was close to neutral (7.0-7.3).Dissolved oxygen was always near 100 % saturation.

Leaf litter processing by stream fauna
Leaf processing rates were studied with leaf packs placed in mesh bags (after Benfield 1996).Leaves of the native N. pumilio and needles of the exotic P. radiata (exotic) were collected in riparian areas of Rucúe Creek in autumn 1997.Leaves were collected directly from trees just prior to abscission.Needles were collected in eight plastic traps of 0.5 m 2 in diameter.Leaves and needles were air dried to constant weight (after Benfield 1996), weighed into 5 g packs, and placed in mesh bags (3 mm mesh size).Additional packs were filled with similar sized sections of black polythene instead of leaves of N. pumilio as controls.Both leaf (and needles) and plastic packs were lashed to a brick and placed in a riffle area of the stream on winter 1997.Four packs of each type were removed from the stream in a downstream to upstream direction after 7, 27, 61, 90, and 146 days.At the time of collection, the brick was removed from the stream; the pack was cut free and placed into a plastic bag.Samples were processed within 4 h after collection to minimize possible weight loss of leaf material.
In the laboratory, each leaf pack was thoroughly rinsed with tap water to remove accumulated sediment and particulate organic matter.Macroinvertebrates were separated from the material and preserved in 80 % ethanol.The intact leaves were dried at 60 °C for 48 h and the weight determined(± 0.1 g).Absolute loss of weight and percentage of the initial weight remaining were determined.Macroinvertebrates removed from the leaf packs were identified, enumerated, and their weight determined after drying to constant weight at 60 °C.No attempt was made to correct for loss of weight due to preservation.Macroinvertebrates were assigned to functional groups according to information in Merritt & Cummins ( 1996), and by inspection of gut contents (Table 1).
A laboratory study was conducted to determine weight loss of leaf material attributable to leaching.Pre-weighed amounts (0.1 g) of air dried leaves (N.pumilio) and needles (P.radiata) were placed separately into small jars containing 150 ml of stream water.These jars were placed inside constant-temperature chambers maintained at 6 and 14 °C.Five jars of each leaf type from each temperature were removed at intervals of 24, 48, 72, and 168 h.The remaining leaf material was dried at 60 oc for 48 h, and weighed to determine loses of material.
Following Benfield (1996), fall-winter leaf processing was assumed to follow an exponential decay model of the form Yt = Y 0e-kt where Yt is the amount remaining after timet of the initial amount Y o (after leaching), and k is the loss rate or processing coefficient.Following loge transformation of the remaining percentage data, linear regression was used to estimate the processing coefficient for each leaf species after excluding leaching loss.According to Benfield (1996), processing coefficients were compared statistically using analysis of covariance (ANCOVA).

RESULTS AND DISCUSSION
Leaf litter processing revealed that native deciduous leaf packs (N.pumilio) were processed at a faster rate than exotic coniferous leaf packs (P.radiata), suggesting that the replacement of the native trees by exotic riparian flora has an important impact on the energetics of stream community in central Chile.The abundance and biomass of shredders per gram of leaf pack is a temperature on leaching between 6 and 14 °C for both species (at 24 h: N. pumilio, F(J.s) = 0.30, P = 0.60; P. radiata F(J.s) = 0.21, P = 0.66).In situ weight loss of leaf packs over time is shown in Fig. 1.Percent loss of weight after 7 days in the stream were not significantly different from the 168 h laboratory leaching experiments for each species (at 14 °C: N. pumilio, F(1,7) = 1.56,P = 0. 70; P. radiata, F(1,7) = 2. 76, P = 0.14 ), indicating that leaching was mainly responsible for the initial Loss of weight (percent loss of initial mass ± SD) of leaf material due to leaching at different temperatures Pérdida de masa (porcentaje de perdida de Ia masa inicial ± SD) del material foliar debido a lixiviaci6n a diferentes temperaturas

Nothofagus pumilio
Pinus radiata 6 oc 14 oc 6°C l4°C 12.8 ± 6.5 14.3 ± 4.1 3.6 ± 1.1 4.2 ± 1.6 20.3 ± 3.9 18.2 ± 6.3 5.2 ± 0.7 3.8 ± 0.8 15.1 ± 8.2 11.3±3.7 4.9 ± 0.9 5.0 ± 1.4 18.3 ± 4.5 13.1±4.65.0 ± 0.8 4.7 ± 0.9 (b) N. pumilio Time in place (days) loss of mass in Rucúe Creek.Following loge transformation of the "remaining percentage" data (Fig. l ), linear regressions were used to estimate the processing coefficient for each leaf species after excluding the leaching loss.Both equations were significant (P < 0.05), and the data fit well the exponential model.After leaching, N. pumilio showed the most rapid degradation.The loss rate coefficient (k) was 0.0072 and 0.0027 for N. pumilio and P. radiata, respectively.The difference between these coefficients was statistically significant (F(1,47)= 74,24, P < 0.0001);  1974, Sedell et al. 1975, Barlocher et al. 1978, Short et al. 1980; Table 3).Moderate to high densities of macroinvertebrates on leaf packs and controls were obtained after 7 days in the stream (Fig. 2).Initially, collectors exhibited the highest densities in N. pumilio, P. radiata and plastic controls, mainly due to the presence of taxa such as Orthocladius sp. and Simulium sp.In the case of Pérdida de masa de paquetes de hojas (promedio ±DE) en el estero Rucue, Chile central.(a) Especie de conffera exotica (Pinus radiata), (b) especie decidua nativa (Nothofagus pumilio ).Total number of taxa N. pumilio, the density of shredders was higher than that of collectors after 28 days, mainly due to L. jaffueli (with a maximum at 61 days).On P. radiata and plastic controls, the highest shredder density occurred at 90 days.High shredder densities were not obtained after 146 days in leaf packs and plastic controls probably due to the emergence of winter stoneflies (Fig. 3).Maximum collector densities occurred at 90 days for N. pumilio and 27 days for P. radiata and controls (Fig. 2).Generally, the same macroinvertebrate taxa colonized the two leaf pack types ( radiata packs (at 61 days, F(1,6) = 81.39,P < 0.0001), and from controls (at 61 days, F(1,6) = 79.14,P < 0.0001), which suggests thatN.pumilio leaves were used as a food source.No significant differences in shredder abundance were detected between P. radiata packs and plastic controls (at 61 days, F(t, 6 )= 0.93, P = 0.371).The biomass of shredders on N. pumilio and on P. radiata showed clear differences, particularly after 48 days (Fig. 4).  the large shredders Capnia and Nemoura in the northern hemisphere, which skeletonize leaves and thus decrease leaf weight (Brinck 1949).Hence, the rate of processing depends on the rate of conditioning (governed by the leaf species) and the intensity of attack by invertebrates.Naturally, this will vary in different streams because of differences in faunal composition (Bird & Kaushik 1981).Shredders also enhance the availability of nutrients to collectors.This has been demonstrated by Short & Maslin (1977), finding that the collectors Hydropsyche californica and Simulium arcticum accrue significantly greater amounts of 32 P in the presence of the shredder, Pteronarcys californica.Collectors, which feed on fine particulate organic matter produced by biological shredding activity or mechanical breakdown by abrasion, may increase downstream.Probably, the presence of the predator Protochauliodes sp.(Megaloptera) was facilitated indirectly by invertebrate use of leaves because of increased numbers of potential prey species.Also, needles of P. radiata and plastic substrates may accumulate periphyton and fine terrestrial matter on their surfaces, which could influence invertebrate colonization (Richardson 1991, Suren & Winterbourn 1992).Petersen & Cummins (197 4) divided leaf processing in streams into leaching, initial microbial processing (or conditioning), and animal-microbial conversion.Physical abrasion may be added to this classification (Bird & Kaushik 1981).The extent to which any of the above sub processes contributes to the overall breakdown will vary with the nature of leaves, water quality, temperature, stream substrate and colonization rates, and benthos activity (Bird & Kaushik 1981 ).Although many adult insects can disperse to Rucúe Creek from nearby streams, habitat factors probably determine successful colonization.Leaf retention has been suggested as a key factor determining shredder abundance (Rounick & Winterbourn 1982, Townsend & Hildrew 1988, Flory & Milner 1999), and the preferences of certain shredders for different leaf species, is well documented (e.g., Parkyn & Winterbourn 1997).The data presented herein make a similar case for Chilean woodland stream in that high quality leaves such as N. pumilio are rapidly processed.
The results about the native deciduous species are obviously valid only for N. pumilio and are not directly applicable to other native deciduous or coniferous species.According to the present data and the information published in the literature (e.g., Benfield 1996), the physical and chemical characteristics of the leaf litter beds are more important than the native or exotic conditions, in determining the decay rate coefficients.In this sense, the real impacts in the fluvial landscapes are derived mainly from changes in the quality and quantity of the leaf litter beds.Further investigation on the nutritive value of leaf packs of N. pumilio and P. radiata is needed to assess more clearly the impact of replacement of native species on the stream trophic structure and fluxes of matter and energy in riparian landscapes in central Chile.

TABLE 1
General classification system for aquatic insect trophic relations.This classification is applicable only to immature and adult stages that occur in the water (modified from Merrit & N. pumilio was processed at such a rapid rate that after 146 days in the stream little material remained, except for the more resistant veins and midribs.Although direct comparisons with leaf processing rates measured in other studies are difficult to interpret due to the differences in the leaf pack construction, leaf pack size, and water temperature, leaf processing rates of this study are within the reported values for conifers and deciduous trees in other temperate forest of the northern hemisphere (i.e., Petersen & Cummins

TABLE 3
Comparison of leaf litter processing coefficients (k) for species found in this study with data reported for other species in northern hemisphere streams *In the laboratory

Table 4
Simulium sp., Hydracarina and Limnoperlajaffueli.The dominant shredders were L. jaffueli and A. michaelseni .Densities of two shredder species on the leaf packs are shown in Fig.2.Densities of shredders (and total macroinvertebrates) found on the P. radiata needle packs and on plastic controls were relatively low.Shredder abundance on N. pumilio leaf packs in Rucúe Creek was significantly different from P.