Light and temperature are key ecosystem drivers, but their synchronous annual cycles typically confound partitioning of their relative influence. Arctic spring-streams, subject to extreme annual fluctuations in light but stable water temperatures, provide a rare contrast that allows the parsing of their independent effects. Over 30 months, we assessed the effects of light and temperature on ecosystem metabolism and nutrient uptake in Ivishak Spring, Alaska, USA. (latitude 698 N, water temperature range ;48-78C) using open-channel methods and short-term NH 4 þ -N, NO 3 À -N, and P additions, respectively. We predicted that rates of ecosystem respiration (ER) would mirror seasonal patterns of gross primary production (GPP), rather than temperature, due to relatively constant rates of metabolic demand year-round, resulting in carbon limitation during winter (October-March) when photosynthesis effectively ceases. Because patterns of nutrient uptake and GPP are often coupled due to assimilatory demand, we also predicted that extreme annual cycles of light would result in equally extreme cycles of nutrient uptake, with demand being relaxed during winter.In accordance with our prediction, we found that ER scaled linearly with GPP. Peak summer rates of GPP (.4.0 g CÁm À2 Ád À1 ) and ER (.5.0 g CÁm À2 Ád À1 ) were surprisingly high, being comparable to those of productive streams at temperate latitudes. Winter rates (GPP ;0.0, ER ,1.0 g CÁm À2 Ád À1 ) were low, however, and Arrhenius plots showed clear deviations from theoretical temperature dependence of GPP and ER during winter when other factors assumed primacy. For GPP, this factor was undoubtedly light availability, but for ER, carbon limitation is implicated due to low GPP. Significant nutrient uptake occurred only for NH 4 þ -N, indicating N limitation, and rates of uptake were also synchronous with cycles of light availability. Consequently, light, rather than temperature, was the major driver of annual patterns of ER and nutrient cycles in this arctic ecosystem. Synchronous light and temperature cycles are pervasive among ecosystems. The winter onset and severity of energy limitation we document highlights the importance of this synchrony and how the confounding of light and temperature obscures details of mechanisms by which these fundamental drivers affect ecosystem processes.

1. We studied the effect of substratum movement on the communities of adjacent mountain and spring tributaries of the Ivishak River in arctic Alaska (69°1¢N, 147°43¢W). We expected the mountain stream to have significant bed movement during summer because of storm flows and the spring stream to have negligible bed movement because of constant discharge. 2. We predicted that the mountain stream would be inhabited only by taxa able to cope with frequent bed movement. Therefore, we anticipated that the mountain stream would have lower macroinvertebrate species richness and biomass and a food web with fewer trophic levels and lower connectance than the spring stream. 3. Substrata marked in situ indicated that 57-66% of the bed moved during summer in the mountain stream and 4-20% moved in the spring stream. 4. Macroinvertebrate taxon richness was greater in the spring (25 taxa) than in the mountain stream (20 taxa). Mean macroinvertebrate biomass was also greater in the spring (4617 mg dry mass m )2 ) than in the mountain stream (635 mg dry mass m )2 ). Predators contributed 25% to this biomass in the spring stream, but only 7% in the mountain stream. 5. Bryophyte biomass was >1000 times greater in the spring stream (88.4 g ash-free dry mass m )2 ) than the mountain stream (0.08 g ash-free dry mass m )2 ). We attributed this to differences in substratum stability between streams. The difference in extent of bryophyte cover between streams probably explains the high macroinvertebrate biomass in the spring stream. 6. Mean food-web connectance was similar between streams, ranging from 0.18 in the spring stream to 0.20 in the mountain stream. Mean food chain length was 3.04 in the spring stream and 1.83 in the mountain stream. Dolly Varden char (Salvelinus malma) was the top predator in the mountain stream and the American dipper (Cinclus mexicanus) was the top predator in the spring stream. The difference in mean food chain length between streams was due largely to the presence of C. mexicanus at the spring stream. 7. Structural differences between the food webs of the spring and mountain streams were relatively minor. The difference in the proportion of macroinvertebrate biomass contributing to different trophic levels was major, however, indicating significant differences in the volume of material and energy flow between food-web nodes (i.e. food web function).

Abstract:We predicted that substratum freezing and instability are major determinants of the variability of stream community structure in Arctic Alaska. Their effects were conceptualized as a two-dimensional habitat template that was assessed using a natural experiment based on five stream types (mountain-spring, tundra-spring, tundra, mountain, glacier). Detrended correspondence analysis (DCA) indicated distinct macroinvertebrate assemblages for each stream type. The contribution of functional feeding groups to assemblage biomass varied systematically among stream types, indicating that structure and function are linked. Assemblage position within a DCA biplot was used to assess factors controlling its structure. Springs separated from other stream types along a gradient of nutrient concentration and freezing probability. Glacier and mountain streams separated from springs and tundra streams along a gradient of substratum instability and freezing probability. Owing to differences in sources of discharge to streams, the effects of nutrients and substratum stability could not be separated from freezing. Although many factors likely contribute to the variability of Arctic stream communities, the major determinants may be conceptualized as a template structured by gradients in (i) nutrient supply and substratum freezing and (ii) substratum instability and substratum freezing. This template provides a basis for predicting the response of Arctic stream communities to climate change. Résumé :Nous avons prédit que le gel et l'instabilité du substrat sont des causes majeures de la variabilité de la structure des communautés des cours d'eau en Alaska arctique. Nous avons conceptualisé ces effets dans un cadre bidimensionnel de l'habitat qui a été évalué au cours d'une expérience en nature dans cinq types de cours d'eau (source montagneuse, source de toundra, toundra, montagne et glacier). Une analyse de correspondance redressée (DCA) révèle des peuplements caractéristiques de macroinvertébrés dans chacun des types de cours d'eau. La contribution des diffé-rentes guildes alimentaires fonctionnelles à la biomasse du peuplement varie systématiquement selon le type de cours d'eau, ce qui montre un lien entre la structure et la fonction. La position de chaque peuplement sur le graphique DCA bidimensionnel sert à évaluer les facteurs qui contrôlent sa structure. Les sources se séparent des autres types de cours d'eau le long d'un gradient de concentration de nutriments et de probabilité de gel. Les cours d'eau de montagne et de glaciers se distinguent des sources et des cours d'eau de toundra sur un gradient d'instabilité du substrat et de probabilité de gel. À cause des différentes sources des apports vers les cours d'eau, il est impossible de distinguer les effets des nutriments et de la stabilité du substrat de ceux du gel. Bien que de nombreux facteurs contribuent vraisemblablement à la variabilité des communautés des cours d'eau arctiques, on peut imaginer les principales causes déterminantes sur un cadre structuré par de...