1. The influence of 11 environmental variables on benthic macroinvertebrate communities was examined in seven glacier‐fed European streams ranging from Svalbard in the north to the Pyrenees in the south. Between 4 and 11 near‐pristine reaches were studied on each stream in 1996–97. 2. Taxonomic richness, measured at the family or subfamily (for Chironomidae) levels for insects and higher levels for non‐insects, increased with latitude from Svalbard (3 taxa) to the Pyrenees (29 taxa). 3. A Generalized Additive Model (GAM) incorporating channel stability [Pfankuch Index (PFAN)], tractive force, Froude number (FROU), water conductivity (COND), suspended solids (SUSP) concentration, and maximum temperature explained 79% of the total deviance of the taxonomic richness per reach. Water temperature and the PFAN of stability made the highest contribution to this deviance. In the model, richness response to temperature was positive linear, whereas the response to the PFAN was bell‐shaped with an optimum at an intermediate level of stability. 4. Generalized Additive Models calculated for the 16 most frequent taxa explained between 25 (Tipulidae) and 79% (Heptageniidae) of the deviance. In 10 models, more than 50% of the deviance was explained and 11 models had cross‐validation correlation ratios above 0.5. Maximum temperature, the PFAN, SUSP and tractive force (TRAC) were the most frequently incorporated explanatory variables. Season and substrate characteristics were very rarely incorporated. 5. Our results highlight the strong deterministic nature of zoobenthic communities in glacier‐fed streams and the prominent role of water temperature and substrate stability in determining longitudinal patterns of macroinvertebrate community structure. The GAMs are proposed as a tool for predicting changes of zoobenthic communities in glacier‐fed streams under climate or hydrological change scenarios.

1. Macroinvertebrate assemblages were studied in the glacial river West‐Jökulsá, originating from the Hofsjökull Ice Cap in central Iceland at an altitude of 860 m. Sampling sites were distributed from the source to 45 km downstream at 160 m a.s.l. Comparative studies were carried out on non‐glacial rivers and tributaries in the area, at similar altitudes and distances from the glacial source. 2. Detrended correspondence analysis (DCA) demonstrated that species composition of benthic macroinvertebrates was related to the distance from the glacier. Assemblages at sampling sites furthest from the glacier were similar in species composition to sites in non‐glacial rivers. Temporal variation was small compared with longitudinal zonation. 3. Based on canonical correspondence analysis (CCA) of data from the main glacial river, distance from the glacier, altitude, bryophyte biomass and the Pfankuch Index of channel stability were the measured explanatory variables having a significant effect on the structure of macroinvertebrate assemblages, accounting for 31% of the total variation in the data set. When data from all the rivers were analysed, altitude, bryophyte biomass, channel slope, suspended sediment concentration and maximum water temperature explained 21% of the variance. 4. Macroinvertebrate communities were in general agreement with the predictions of the conceptual model of Milner & Petts (1994) for the upstream reaches. The assemblages consisted mainly of Orthocladiinae and Diamesinae (Chironomidae), although other taxa such as Simuliidae, Plecoptera and Trichoptera were also found in low numbers. Shredders were lacking from the benthic communities, apparently because of continued glacial influence in the river even 45 km downstream from the glacier and lack of allochthononus inputs from riparian vegetation.

No abstract

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. This content downloaded from 169.230.243.252 on TueAdalsteinsson, H., J6nasson, P.M. and Rist, S. 1992. Physical characteristics of Thingvallavatn, Iceland. -Oikos 64: 121-135.The basin of subarctic Thingvallavatn (64010'N, 21?10'W) was originally formed by tectonic subsidence and glacial erosion. Since the glacial recession some 10,000 yr ago, the basin has been strongly modified by volcanic activity. The catchment area is characterized by lava flows, of which four entered the lake, and one dammed it. One eruption occurred within the lake. Fault movements have resulted in horizontal and vertical displacements. The lake, area 83 km2, is at 100.5 m a.s.l. It has a mean depth of 34 m, and a maximum of 114 m. The catchment area is 1000 km2, of which ca. 50% is postglacial lava. Interglacial lavas and subglacially formed mountains rising to 800-1100 m, glaciers and glacial deposits account for another 50%. The mean discharge of the last 50 yr is 100 m3 s-1. The precipitation, estimated at 2000-2500 mm yr-1, is mostly percolated; hence the lake is ca. 90% spring-fed. The main spring areas in the N show a constant temperature of 2.8-3.5?C. Warmer ground water (7.5-10.7?C) enters the lake in the SW owing to the nearby Hengill geothermal area. A discrepancy between potential water balance and discharge is explained by an underestimate of the amount of precipitation in the mountainous areas, by evapotranspiration and underground flow from the Langjokull glacier bed from outside the topographic catchment area. The water level has been regulated since 1960, and the pattern of water level fluctuations before and after are discussed. Irradiance (PAR) measurements since 1982 reveal strong diel fluctuations and a reflection of 49% from the lake surface. Values from PAR measurements and global radiation estimates show fairly consistent results. The euphotic layer varies between 20 m in spring to 40 m in summer. Temperature measurements cover 30 yr. The lake is dimictic and the temperature commonly reaches 10-11?C in August with a gradient of 2-3?C over the thermocline. On an average, the ice-cover lasted 95 d during the period 1974 to 1990.

No abstract