Selective foraging by moose on hardwoods and avoidance of conifers alters community composition and structure, which in turn can affect nutrient cycles and productivity. The effect of moose browsing on the nutrient cycles of boreal forests was studied using three 40—yr—old exclosures on Isle Royale, Michigan. Two alternative mechanisms by which moose affect ecosystems were tested: (1) moose depress both the quantity and quality of litter return to the soil, and hence N mineralization and net primary productivity, by browsing on hardwoods and avoiding conifers; (2) moose stimulate N mineralization, and hence net primary productivity, by opening the canopy and by dropping fecal pellets. Soil nutrient availability and microbial activity, including exchangeable cations, total carbon and nitrogen, nitrogen mineralization rates, and microbial respiration rates, were uniformly higher in exclosures than outside. These differences were more significant where browsing intensity was high and less often significant where browsing intensity was low. N mineralization in browsed plots declined with increasing moose consumption rates. Net primary production in exclosures and browsed plots was strongly correlated with N mineralization. N mineralization in turn was positively correlated with litter N return and negatively correlated with litter cellulose content. These differences in litter quantity and quality were caused by an increased abundance of unbrowsed spruce outside the exclosures. Moose pellets alone mineralized less N but more C than soil alone, but pellets combined with soil stimulated N and C mineralization more than the sum of the two separately. However, this did not appear to be sufficient to offset the depression in nitrogen and carbon mineralization in soil resulting from the increased abundance of unbrowsed spruce. We conclude that, in the long term, high rates of moose browsing depress N mineralization and net primary production through the indirect effects on recruitment into the tree stratum, and subsequent depression of litter N return and litter quality. These results suggest that the effects of herbivores on ecosystems may be amplified by positive feedbacks between plant litter and soil nutrient availability.

Net aboveground production (4.1—9.5 Mg°ha—1) across a series of edaphic climax forests was highly correlated with field measurements of soil N mineralization (26—84 kg°ha—1°yr—1; r2 = 0.902, P < .001) and with soil silt + clay content (5—74%; r2 = 0.883, P < .001). Soil N mineralization was positively correlated with litter production and N and P return in litter. Soil N mineralization was negatively correlated with litter C:N and C:P ratios and with efficiency of P use in litter production. Efficiency of N use in litter production declined with increasing N mineralization except for inefficient use of N in a hemlock stand at low N mineralization. Changes of litter quality across the mineralization—soil texture gradient were due to species replacement rather than changes in litter quality within each species. Nitrification was not correlated with aboveground production. Both mineralization and nitrification were highly correlated with humus P content. Differences in nitrification among the soils appeared to be related to PO4—P supply in the spring and early summer and to NH4—N supply in mid— to late summer.

No abstract

Large-scale changes in climate may have unexpected effects on ecosystems, given the importance of climate as a control over almost all ecosystem attributes and internal feedbacks. Changes in plant community productivity or composition, for example, may alter ecosystem resource dynamics, trophic structures, or disturbance regimes, with subsequent positive or negative feedbacks on the plant community. At northern latitudes, where increases in temperature are expected to be greatest but where plant species diversity is relatively low, climatically mediated changes in species composition or abundance will likely have large ecosystem effects. In this study, we investigated effects of infrared loading and manipulations of water-table elevation on net primary productivity of plant species in bog and fen wetland mesocosms between 1994 and 1997.We removed 27 intact soil monoliths (2.1 m 2 surface area, 0.5-0.7 m depth) each from a bog and a fen in northern Minnesota to construct a large mesocosm facility that allows for direct manipulation of climatic variables in a replicated experimental design. The treatment design was a fully crossed factorial with three infrared-loading treatments, three water-table treatments, and two ecosystem types (bogs and fens), with three replicates of all treatment combinations. Overhead infrared lamps caused mean monthly soil temperatures to increase by 1.6-4.1ЊC at 15-cm depth during the growing season (May-October). In 1996, depths to water table averaged Ϫ11, Ϫ19, and Ϫ26 cm in the bog plots, and 0, Ϫ10, and Ϫ19 cm in the fen plots.Annual aboveground net primary production (ANPP) of bryophyte, forb, graminoid, and shrub life-forms was determined for the dominant species in the mesocosm plots based on speciesspecific canopy/biomass relationships. Belowground net primary production (BNPP) was estimated using root in-growth cores.Bog and fen communities differed in their response to infrared loading and water-table treatments because of the differential response of life-forms and species characteristic of each community. Along a gradient of increasing water-table elevation, production of bryophytes increased, and production of shrubs decreased in the bog community. Along a similar gradient in the fen community, production of graminoids and forbs increased. Along a gradient of increasing infrared loading in the bog, shrub production increased whereas graminoid production decreased. In the fen, graminoids were most productive at high infrared loading, and forbs were most productive at medium infrared loading. In the bog and fen, BNPP:ANPP ratios increased with warming and drying, indicating shifts in carbon allocation in response to climate change.Further, opposing responses of species and life-forms tended to cancel out the response of production at higher levels of organization, especially in the bog. For example, total net primary productivity in the bog did not differ between water-table treatments because BNPP was greatest in the dry treatment whereas ANPP was greatest in the wet treatmen...

No abstract

No abstract

Decomposition and changes in nitrogen and organic—chemical content of six types of forest litter were studied for 2 yr in five adjacent Wisconsin forests. The five forests were floristically dissimilar, being dominated respectively by sugar maple (Acer saccharum), white oak (Quercus alba), bigtooth aspen (Populus grandidentata), white pine (Pinus strobus), and hemlock (Tsuga canadensis), Nitrogen mineralization rates in the five stands ranged from 29 to 125 kg°ha—1°yr—1. Decomposition rates of transplanted sugar maple leaves and red maple (Acer rubrum) wood were not correlated with nitrogen mineralization rates in all five stands, indicating that nitrogen mineralization rates do not affect initial decomposition rates. However, mineralization rates were correlated with decomposition rates of the native dominant foliage litter. Nitrogen first accumulated in all litters, but by the end of the 2—yr incubation period nitrogen release had begun in all foliage litters. Nitrogen concentrations increased approximately linearly with cumulative mass loss but eventually declined in some foliage litters. Neither maximum amount of nitrogen accumulated nor amount accumulated per gram of litter mass loss was related to the rate of soil nitrogen mineralization. Chemical composition of litter affected decomposition rates and patterns. Soluble substances and litters relatively rich in solubles disappeared rapidly during early stages of decomposition. Eventually, slowly disappearing acid—soluble and acid—insoluble substances dominated the pattern of mass loss in all litters.

Northern wetlands may be a potential carbon source to the atmosphere upon global warming, particularly with regard to methane. However, recent conclusions have largely been based on short‐term field measurements. We incubated three wetland soils representing a range of substrate quality for 80 wk in the laboratory under both aerobic and anaerobic conditions at 15° and 30°C. The soils were obtained from a Scirpus‐Carex‐dominated meadow in an abandoned beaver pond and from the surface and at 1 m depth of a spruce (Picea)‐Sphagnum bog in Voyageurs National Park, Minnesota. Substrate quality was assessed by fractionation of carbon compounds and summarized using principal components analysis. Nitrogen and carbon mineralization, the partitioning of carbon between carbon dioxide and methane, pH, and Eh were measured periodically over the course of the incubation. The responses of nitrogen mineralization, carbon mineralization, and trace gas partitioning to both temperature and aeration depended strongly on the substrate quality of the soils. Sedge meadow soil had the highest nitrogen and carbon mineralization rates and methane production under anaerobic conditions, and carbon mineralization under aerobic conditions, but the surface peats had the highest nitrogen mineralization rates under aerobic conditions. Methanogenesis was highest in the sedge soil but less sensitive to temperature than in the peats. A double exponential model showed that most of the variation in nitrogen and carbon mineralization among the soils and treatments was accounted for by differences in the size and kinetics of a relatively small labile pool. The kinetics of this pool were more sensitive to changes in temperature and aeration than that of the larger recalcitrant pool. Principal components analysis separated the soils on the basis of labile and recalcitrant carbon fractions. Total C and N mineralization correlated positively with the factor representing labile elements, while methanogenesis also showed a negative correlation with the factor representing recalcitrant elements. Estimates of atmospheric feedbacks from northern wetlands upon climatic change must account for extreme local variation in substrate quality and wetland type; global projections based on extrapolations from a few field measurements do not account for this local variation and may be in error.