Passive open-top devices have been proposed as a method to experimentally increase temperature in high-latitude ecosystems. There is, however, little documentation on the efficacy of these devices. This paper examines the performance of four open-top chambers for altering temperature at six sites in the Arctic and Antarctica. Most of the heating effect was due to daytime warming above ambient; occasional night-time cooling below ambient, especially of air temperatures, depressed mean daily temperature. The mean daily temperatures at four arctic sites were generally increased by 1.2-1.8°C; but occasionally, temperature depressions also occurred. Under optimal conditions at the antarctic site (dry soils, no vegetation, high radiation) mean daily soil temperatures were increased by ⍣2.2°C (-10 cm) to ⍣5.2°C (0 cm). Protection from wind may play a more important role than temperature per se in providing a favourable environment for plant growth within opentop devices. Wind speed had a generally negative impact on mean daily temperature. Daily global radiation was both positively and negatively related to chamber temperature response. The effect of chambers on snow accumulation was variable with the Alexandra Fjord site showing an increased accumulation in chambers but no difference in the date of snowmelt, while at Latnjajaure in a deep snowfall site, snowmelt occurred 1-2 weeks earlier in chambers, potentially increasing the growing season. Selection of a passive temperature-enhancing system requires balancing the temperature enhancement desired against potential unwanted ecological effects such as chamber overheating and altered light, moisture, and wind. In general, the more closed the temperature-enhancing system, the higher is the temperature enhancement, but the larger are the unwanted ecological effects. Open-top chambers alter temperature significantly and minimize most unwanted ecological effects; as a consequence, these chambers are a useful tool for studying the response of high-latitude ecosystems to warming.

No abstract

No abstract

We are studying the distribution, biodiversity, and abundance of nematodes in the most extreme terrestrial environment on earth, the Dry Valley region of Antarctica. Here we report that the nematode community structure of 1–3 species in two functional groups may be the simplest soil food web of any terrestrial ecosystem. Nematodes were widespread and not correlated with moisture, C, or N, factors that define soil biotic complexity elsewhere. In a field experiment, treatments increasing soil water, carbon, and temperature, alone or in combination, generally decreased the abundance of the single omnivore‐predator species and increased the abundance of its microbivorous prey species. These low‐diversity nematode communities, limited to ≤3 species, apparently lack species redundancy and appear sensitive to environmental change. Our findings suggest that Antarctic soil ecosystems are sensitive to anthropogenic disturbance.

No abstract

No abstract

Population samples of an African C4 grass, Panicum coloratum L., were collected from two locations in the Serengeti Grasslands varying in grazing intensity, one a high—grazing location (GA = grazing—adapted), the other a low—grazing location (NGA = nongrazing—adapted). Plants were cloned, put in controlled environments mimicking the natural photo—thermoperiod, and subjected to light grazing pressure by a generalist feeding North American grasshopper, Melanoplus sanguinipes. Carbon assimilation and redistribution were measured in the short term with an infrared gas analyzer and 11C—labelled CO2, coupled with a three—compartment analytical model, and by harvesting whole plants at the end of a 12—wk regrowth experiment. Results documented several significant differences between the GA and NGA samples, suggesting the evolution of physiological traits related to C assimilation, translocation, and storage in response to previous grazing history. Pregrazing net C—fixation rates, translocation rates, and C—storage pools were identical for the two ecotypes. After grazing, overall C—fixation rates were 39% higher for the GA plants, and the regrowth data suggest they remained higher than NGA rates throughout the experiment. Removal of < 10% of initial green—leaf biomass by grazing at each grazing period produced major differences in carbon flux between the two samples. Throughout the experiment GA plants produced and stored more C in leaves, stored less C in stem sinks, had higher phloem activity, and translocated more of the labile C to roots where it was stored in higher quantities. This suggests that storage of labile C reserves in sinks or pools readily accessible to the plant, which allows rapid mobilization after grazing, is an important element of adaptation to grazing. Major storage of labile reserves in stems, characteristic of the NGA plants, may be advantageous in ungrazed habitats where there is vertical growth due to canopy closure and competition for light, but such storage makes those reserves accessible to grazers. GA ecotype plants in the regrowth experiment compensated completely for the eight weekly defoliation events by the time of the 12—wk harvest; yield of the NGA ecotype was reduced 21% by the moderate level of grasshopper grazing. Increased total yield (biomass harvested plus compensation) by the GA ecotype was expressed in both above— and belowground biomass, suggesting that suppression of the latter does not contribute to compensation by the former.

In order to better understand the spatial distributions of soil trophic groups and the potential significance of these distributions to ecosystem functioning we initiated a study to describe the within—site variability of nematode feeding groups in a row—crop ecosystem. Soil cores were removed from a 48—ha corn (Zea mays) field in the U.S. Midwest prior to spring planting, and nematodes were identified by phenotypic criteria to four groups: bacterivores, fungivores, omnivores/predators, and plant parasites. Within—site variability was high for all groups; population counts spanned two orders of magnitude, with coefficients of variation ranging from 40—130% (n = 115—138 soil samples). Probability distributions were strongly lognormal. Geostatistical analysis showed that a major part of this variability was spatially dependent; variograms suggest that 70—99% of sample population variance was related to spatial autocorrelation over our geographic range of 6—80 m, except for the parasitic group, for which we detected no autocorrelation to 1200m. Maps of nonparasitic feeding groups across the field showed large multi—hectare areas of low to moderate population densities, with sub—hectare clusters of high—density populations towards one end of the site. Individual feeding groups were only weakly correlated with one another across the field (Kendall's t < 0.363, P < 0.001). Edaphic factors (bulk density, texture, pH, C availability, N availability) could collectively explain <30% of the variability in the nonparasitic groups across the area sampled. Results suggest that important soil food web components are strongly patterned at sub—hectare scales in this site. That this patterning is maintained in an ecosystem subjected to the homogenizing influences of annual soil tillage and a monoculture plant population is remarkable, and suggests that such patterning may be even more common in less—disturbed sites. Inclusion of these patterns in studies of ecosystem processes and soil community dynamics may significantly improve soil trophic models and our understanding of the relationship between soil populations and ecosystem function.