Ecosystems (2005) 8: 748–759 DOI: 10.1007/s10021-005-0004-y Aquatic Terrestrial Linkages Along a Braided-River: Riparian Arthropods Feeding on Aquatic Insects Achim Paetzold,1* Carsten J. Schubert,2 and Klement Tockner1 1Department of Limnology, EAWAG/ETH, P.O. BOX 611, 8600 Du¨bendorf, Switzerland; 2Department of Surface Water, EAWAG, Limnological Research Center, Seestrasse 79, 6047 Kastanienbaum, Switzerland ABSTRACT Rivers can provide important sources of energy for ant Manica rubida fed mainly on terrestrial sources. riparian biota. Stable isotope analysis (d13C, d15N) The proportion of aquatic insects in the diet of ly- together with linear mixing models, were used to cosid spiders changed seasonally, being related to quantify the importance of aquatic insects as a food the seasonal abundance of lycosid spiders along the source for a riparian arthropod assemblage inhab- stream edge. The degree of spatial and seasonal iting the shore of the braided Tagliamento River aggregation of riparian arthropods at the river edge (NE Italy). Proportional aquatic prey contributions coincided with their proportional use of aquatic to riparian arthropod diets differed considerable subsidies. The results suggest that predation by among taxa. Carabid beetles of the genus Bembidion riparian arthropods is a quantitatively important and Nebria picicornis fed entirely on aquatic insects. process in the transfer of aquatic secondary pro- Aquatic insects made up 80% of the diet of the duction to the riparian food web. dominant staphylinid beetle Paederidus rubrothorac- icus. The diets of the dominant lycosid spiders Arc- Key words: boundary; Carabidae; food web; tosa cinerea and Pardosa wagleri consisted of 56 and Formicidae; Lycosidae; riparian; stable isotopes; 48% aquatic insects, respectively. In contrast, the Staphylinidae; subsidy. INTRODUCTION aquatic biota in rivers is widely recognized (Fisher and Likens 1973; Vannote and others 1980; Wallace Ecosystem dynamics are rarely confined to a par- and others 1999), less is known about the energy ticular habitat and the spatial flow of matter and flow from aquatic to terrestrial systems. However, organisms between habitats can substantially affect reciprocal flows of matter and organisms from the local population, community, and food web stream into the riparian zone can be important dynamics (Polis and others 1997). The exchange of energy sources to terrestrial consumers (Power and energy and nutrients between the river channel and Rainey 2000; Nakano and Murakami 2001; Naiman its riparian zone is an important process in lotic and others 2002; Sabo and Power 2002). These ecosystems (Ward 1989; Naiman and De´camps spatial subsidies seem to be particularly important 1997; Fisher and others 1998; Helfield and Naiman for consumers living in less productive habitats that 2001). Whereas the role of allochthonous inputs receive inputs from adjacent habitats of higher (for example, leaf litter) as an energy source for production (Bustamante and others 1995; Polis and Hurd 1996; Stapp and Polis 2003). Received 8 January 2004; accepted 1 June 2004; published online In braided rivers, the primary channel splits into 21 October 2005. several channels that flow around low productivity *Present address: Catchment Science Center. The University of Sheffield North Campus. Kroto Building, Broad Lane Sheffield S3 71IQ. UK areas of exposed gravel. These gravel banks are *Corresponding author; e-mail: [email protected] colonized by carnivorous riparian arthropods, pri- 748 Aquatic-terrestrial Feeding Linkages 749 marily spiders, staphylinid beetles, ground beetles, pods differ in their proportional consumption of and ants (Manderbach and Hering 2001; Framenau aquatic insects, and does this change seasonally? Is and others 2002; Sadler and others 2004). The the degree of seasonal and spatial aggregation of strong productivity gradient (river to gravel bank), riparian arthropod taxa along the land–water combined with a predator dominated arthropod interface related to the respective contribution of community, suggests the importance of aquatic aquatic insects to their diets? Does riparian subsidies for the ground-dwelling arthropods of arthropod predation represent a quantitatively gravel banks. For instance, gut analysis of riparian important pathway in the transformation of sec- ground beetles in a braided river indicated a high ondary aquatic production to the terrestrial food proportion of aquatic insects (Hering and Plachter web? 1997). Recent stable isotope studies from other rivers showed that riparian spiders (Collier and others 2002; Sanzone and others 2003) and grass- METHODS hoppers (Bastow and others 2002) can also feed Study Site extensively on aquatic resources. However, we need to extend our knowledge of aquatic subsidies Extensive studies of riparian arthropods along five from a small subset of consumers to entire com- braided rivers (Switzerland and Italy) exhibited munities to estimate prey fluxes from the river to similar community composition, spatial distribu- the riparian zone. The contribution of emerged tions and densities of riparian arthropods (Uhl- aquatic insects to bird diets has been quantified for mann 2001; A. Paetzold, unpublished data). For a riparian forest bird assemblage (Nakano and the detailed investigation of trophic linkages across Murakami 2001). To estimate aquatic prey fluxes the aquatic–terrestrial interface in this study, we through different pathways of the riparian food selected the site that was least modified by human web (ground predators versus aerial predators), we activities. The riparian arthropod assemblage along also need to quantify the role of aquatic insects as a the stream edge consisted mainly of lycosid spiders food source for entire riparian arthropod assem- (Lycosidae), ground beetles (Carabidae), rove bee- blages. This requires a quantitative knowledge of tles (Staphylinidae), and ants (Formicidae). the use of aquatic insects for all dominant riparian The study was conducted from April to October arthropods and their relative dominance in the 2002 along a 400 m gravel bank of a side-channel assemblage. The role of aquatic subsidies for ripar- in a braided reach of the Tagliamento River. The ian consumers can also vary strongly among sea- Tagliamento River is a seventh order gravel-bed sons (Nakano and Murakami 2001). Therefore, we river located in NE-Italy (46°N, 12°30ÕE) with a need to understand the interrelationship between catchment of 2580 km2 (see detailed description in seasonal changes in aquatic subsidies and the spa- Ward and others 1999; Tockner and others 2003). tio-temporal distribution of riparian arthropods The river corridor is fringed by continuous riparian along the river edge. woodland (Gurnell and others 2001). The Taglia- Natural abundances of stable isotopes have been mento River is characterized by a flashy flow re- successfully applied to assess flows of nutrients and gime, with highest average discharge in autumn by matter across the aquatic terrestrial interface (for torrential rains and smaller peak flows in spring example, Ben-David and others 1998; Thorp and from snowmelt runoff (Arscott and others 2002). others 1998). As the turnover time for stable iso- Despite local water abstraction and a channelized topes in small arthropods can be relatively short downstream section, the river retains an essentially (Tieszen and others 1983; Ostrom and others pristine morphological character and flood 1997), stable isotopes also can be used to detect dynamics (Ward and others 1999). The active flood seasonal changes in the food sources of an animal. plain width was 0.8 km in the island-braided reach. Stable isotope methods in ecosystem studies have The reach included a complex channel network the advantage over traditional methods in that they and exposed gravel bars represented the largest reveal an integrated feeding history of an animal proportion of the landscape cover (van der Nat and and what source of food is actually used for tissue others 2002). growth (Rounick and Winterbourn 1986). We used The 400-m gravel bank of the study section stable isotope analysis to examine aquatic–terres- represented the dominant type of river-riparian trial feeding linkages for the dominant taxa of a interface along the entire braided-river corridor riparian arthropod assemblage. (Petts and others 2000). The channel had an In the present study, we addressed the following average width of 20 m at low water level and primary questions: Do dominant riparian arthro- comprised pool, run, riffle, and backwater sections. 750 A. Paetzold and others The gravel bar was up to 60 m wide and bordered they were larger than most riparian arthropods. by riparian forest on the upslope side. It included At least ten specimens of each family were used steep (eroding) and shallow (depositional) banks. for isotope analysis. Ground cover was predominantly gravel along the We inferred possible terrestrial food sources of stream edge with sand and patches of grass and riparian arthropods indirectly from predaceous woody vegetation in higher, less frequently flooded ground-dwelling arthropods living far from the habitats. Woody vegetation consisted of Populus stream (terrestrial arthropods) because it can be nigra and various willow species (Karrenberg and assumed that they integrated isotope signals from a others 2003). wide variety of terrestrial food sources that are of- ten unknown, small, and difficult to sample in the Sampling for Isotope Analysis field. To test