Arctic tundra has large amounts of stored carbon and is thought to be a sink for atmospheric carbon dioxide (CO(2)) (0.1 to 0.3 petagram of carbon per year) (1 petagram = 10(15) grams). But this estimate of carbon balance is only for terrestrial ecosystems. Measurements of the partial pressure of CO(2) in 29 aquatic ecosystems across arctic Alaska showed that in most cases (27 of 29) CO(2) was released to the atmosphere. This CO(2) probably originates in terrestrial environments; erosion of particulate carbon plus ground-water transport of dissolved carbon from tundra contribute to the CO(2) flux from surface waters to the atmosphere. If this mechanism is typical of that of other tundra areas, then current estimates of the arctic terrestrial sink for atmospheric CO(2) may be 20 percent too high.

1. We studied the spatial and temporal patterns of change in a suite of twenty‐one chemical and biological variables in a lake district in arctic Alaska, U.S.A. The study included fourteen stream sites and ten lake sites, nine of which were in a direct series of surface drainage. All twenty‐four sites were sampled between one and five times a year from 1991 to 1997. 2. Stream sites tended to have higher values of major anions and cations than the lake sites, while the lake sites had higher values of particulate carbon, nitrogen, phosphorous and chlorophyll a. There were consistent and statistically significant differences in concentrations of variables measured at the inlet versus the outlet of lakes, and in variables measured at upstream versus downstream sites in the stream reaches which connect the lakes. In‐lake processing tended to consume alkalinity, conductivity, H+, DIC, Ca2+, Mg2+, CO2, CH4, and NO3–, and produce K+ and dissolved organic carbon (DOC). In‐stream processing resulted in the opposite trends (e.g. consumption of K+ and DOC), and the magnitudes of change were often similar to those measured in the lakes but with the opposite sign. 3. Observed spatial patterns in the study lakes included mean concentrations of variables which increased, decreased or were constant along the lake chain from high to low altitude in the catchment (stream sites showed no spatial patterns with any variables). The strongest spatial patterns were of increasing conductivity, Ca2+, Mg2+, alkalinity, dissolved inorganic carbon and pH with lake chain number (high to low altitude in the basin). These patterns were partly determined by the effect of increasing catchment area feeding into lakes further downslope, and partly by the systematic processing of materials in lakes and in the stream segments between lakes. 4. Synchrony (the temporal coherence or correlation of response) of variables across all lakes ranged from 0.18 for particulate phosphorus to 0.90 for Mg2+ the average synchrony for all twenty‐one variables was 0.50. The synchronous behaviour of lake pairs was primarily related to the spatial location or proximity of the lakes for all variables taken together and for many individual variables, and secondarily, to the catchment to lake area ratio and the water residence time. 5. These results illustrate that, over small geographic areas, and somewhat independent of lake or stream morphometry, the consistent and directional (downslope) processing of materials helps produce spatial patterns which are coherent over time for many limnological variables. We combine concepts from stream, lake and landscape ecology, and develop a conceptual view of landscape mass balance. This view highlights that the integration of material processing in both lakes and rivers is critical for understanding the structure and function of surface waters, especially from a landscape perspective.

Phosphorus fertilization of a pristine tundra river for four consecutive summers dramatically changed biological processes and populations at all trophic levels. At the primary producer level, both algal biomass and productivity increased and chlorophyll accumulated on the river bottom in the first two summers. Diatom community composition changed little in spite oflarge chlorophyll changes. However, an increase in grazing insects prevented chlorophyll buildup in the third and fourth summers.Some microbial processes were also stimulated by the increased photosynthesis caused by fertilization. Total respiration of the epilithon, acetate uptake, and decomposition of lignin monomers were all stimulated but only in light-grown epilithon. When epilithon was grown in the dark in the fertilized region of the river, there was no increased respiration. Also, phosphorus did not stimulate the decomposition of Carex litter.Although insects grew more rapidly in the fertilized section of the river, there were community interactions that kept total insect production from appreciable change. The four most abundant large insects did increase their growth rates in response to phosphorus addition and there were increases in populations of Baetis lapponicus and Brachycentrus americanus. These increases were offset by the decline in abundance of the dominant species, the black fly Prosimulium martini, perhaps caused by competition for space from Brachycentrus.Growth of both young-of-the-year and adult grayling (Thymallus arcticus) was strongly stimulated by phosphorus addition in years 3 and 4 (not tested in years 1 and 2). Carbon and nitrogen stable isotope tracers indicated that the measured increases in insect and fish growth were largely attributable to increases in the production of epilithic algae. Overall, the results indicate a strong "bottom-up" response of the riverine food web to additions of the limiting nutrient, phosphorus. The response was modified in later years, however, by a strong "top-down" feedback of insects grazing on epilithic algae and by competitive exclusion of black flies by caddisflies.

No abstract

Partial pressures of CO 2 and CH 4 were measured directly or calculated from pH and alkalinity or DIC measurements for 25 lakes and 4 rivers on the North Slope of Alaska. Nearly all waters were supersaturated with respect to atmospheric pressures of CO 2 and CH 4 . Gas fluxes to the atmosphere ranged from -6.5 to 59.8 mmol m -2 d-1 for CO 2 and from 0.08 to 1.02 mmol m -2 d-' for CH 4 , and were uncorrelated with latitude or lake morphology. Seasonal trends include a buildup of CO 2 and CH 4 under ice during winter, and often an increased CO 2 flux rate in August due to partial lake turnover. Nutrient fertilization experiments resulted in decreased CO 2 release from a lake due to photosynthetic uptake, but no change in CO 2 release from a river due to the much faster water renewal time. In lakes and rivers the groundwater input of dissolved CO 2 and CH 4 is supplemented by in-lake respiration of dissolved and particulate carbon washed in from land. The release of carbon from aquatic systems to the atmosphere averaged 24 g C m-2 y-1, and in coastal areas where up to 50 % of the surface area is water, this loss equals /5 to /2 of the net carbon accumulation rates estimated for tundra.

Lakes and streams in the foothills near Toolik Lake, Alaska, at 68"N have been studied since 1975 to predict physical, chemical and biological impacts of future global change. Experimental manipulations include whole lake and continuous stream fertilization as well as removal and addition of predators (copepods, lake trout, grayling, sculpin). Based on our evidence the following scenario is likely. Warming thaws the upper layers of permafrost and streams and lakes become enriched with phosphorus. Streams respond quickly with higher production of diatoms but animal grazers keep biomass changes to a minimum. Fish productivity also increases. If phosphorus levels are too high, mosses become the dominant primary producer and sequester all of the nutrients. Growth of Arctic grayling under the present conditions only occurs in summers with higher than average stream flow. The present population would be stressed by warmer temperatures. When higher phosphorus levels reach lakes and cause slight eutrophication, the number of trophic levels will increase, especially within the microbial food web. Warmer lake temperatures increase stratification and, combined with eutrophication, could decrease oxygen in the hypolimnion. Oxygen levels will also decrease in winter under the ice cover. Eventually this habitat change will eliminate the lake trout, a top predator. Removal of lake trout results in a striking increase in abundance and productivity of smaller fish, including small lake trout, and the emergence of burbot as an alternate top predator. Large species of zooplankton will become virtually extinct.

In situ radon-222 flux experiments conducted in benthic chambers in Cape Lookout Bight, a small marine basin on the North Carolina coast, reveal that enhanced chemical transport across the sediment-water interface during summer months is caused by abiogenic bubble tube structures. Transport rates for dissolved radon, methane, and ammonium more than three times greater than those predicted on the basis of molecular diffusion occur when open tubes are maintained by semi-diurnal low-tide bubbling.

No abstract