In a maple-beech forest in eastern New York, equal numbers of mounds, pits, and adjacent undisturbed soil microsites were censused for plant species density and cover over the growing seasons of 1978 and 1979. Several soil properties were also measured for each microsite. Detrended correspondence analysis (DCA) and chi-square goodness-of-fit tests were used to test whether the species were uniformly distributed over the microrelief positions. Comparisons of species composition, diversity, and vegetative phenologies were made between microsites and between hemlock-and hardwood-dominated forest areas. Differences in soil properties between microsites and between the two forest areas were assessed with one-tailed paired t tests and Wilcoxon rank-sum tests, respectively.Microrelief created a mosaic of soil properties and a vegetation pattern in the understory, both of which were affected by the presence of hemlock in the canopy. Ordinations (DCA) indicated two factors that affected species composition of the microsites: (I) presence of hemlock (Tsuga canadensis) as the nearest canopy tree, and (2) microtopography (whether the microsite was a mound, pit, or undisturbed-soil site). Species in non-hemlock areas had patchy distributions in the forest community: eight species grew mainly on mounds, six in pits, and five on undisturbed soil. The mounds and pits had characteristic assemblages of species. Species richness and total density and cover of plants in each microsite were constant over the growing season. In contrast, little spatial or temporal community pattern was apparent in forest areas influenced by hemlock.The effect of hemlock on species distributions was due in part to its effect on soil properties. For all microsites, forest areas with hemlock had greater soil organic matter content, available nitrogen, cation exchange capacity, and litter depth, lower soil calcium content, moisture content, temperature, pH, and A1 horizon depth; and no frost heaving.Microrelief also affected soil properties significantly. In comparison with pits, mounds were drier and poorer in nutrient content, and had a lower cation exchange capacity, less organic matter, less litter cover, a thinner A1 horizon, and less snow accumulation. In areas without hemlock, the mounds were also more acid, warmer in summer and colder in winter, and more subject to frost heaving than pits. Hemlock-influenced areas had no consistent microsite differences in pH or temperature.The different microenvironments, created by the interaction of microrelief and hemlock, resulted in patchy distributions of most understory species. This pattern was likely a result of species requirements for and tolerances of environmental conditions, tempered by competitive interactions.

The uprooting of forest trees leads to the formation of microsites on the forest floor, contributing to fine-scale heterogeneity in soil properties. We found the types of microsites formed depended on the way the tree fall occurred. Tree falls were classified as either hinge or rotational types. Hinge tree falls formed when the root mat of a tree and the surrounding soil were uplifted vertically, leaving an adjacent pit in the soil. Hinge tree falls varied as to thickness of the root mat and angle of uplift. Rotational tree falls were usually a result of a ball and socket motion of the root mat and soil, which positioned the tree bole over the newly created pit. The tree falls disrupted and redistributed surface soil organic matter and subsoil. In rotational tree falls, the surface material remained intact, covering some of the pit and the adjacent side of the mound. In hinge tree falls, the surface organic matter was deposited on the throw side of the mound, leaving subsoil on the other side and in the pit. With time, however, hinge-type pits accumulated litter and eventually had more organic matter than mounds. Old mounds from both hinge and rotational tree falls had lower concentrations of calcium and magnesium, lower pH, and less moisture than pits. The tree fall process creates long-term soil patterns and maintains microsite heterogeneity in forest communities.

In deciduous forests of central New York, treefall pits have rich soil, but low plant species richness. To test whether leaf litter limited species distributions, we removed litter in 10 randomly chosen pits in 1983 and compared them with controls over 3 years. Wire cages prevented litter accumulation in experimental plots. By year 1 (1984) litterless pits were significantly warmer and less moist than controls. From years 1 to 3 the average Coefficient of Community (CC) index between experimental mounds and pits increased 10-fold. No changes occurred in controls. Species composition of experimental pits became different from the controls. Species richness increased in litterless pits, with no change in other plots. Experimental and control mounds were similar in species composition all years, indicating no effect of the cage. Species responses to litter removal included: (i) enhanced germination, (ii) establishment, and (or) (iii) no change. All herbs showing increased germination or establishment were species previously restricted to mounds. The species that did not change were those already inhabiting pits. No species declined after litter removal. Litter removal in pits allowed several species to expand their distributions from mounds into litterless pits, without affecting existing pit species. The spatial heterogeneity accentuated by leaf litter accumulation does not facilitate coexistence of species in the forest community, but only limits species population sizes and spatial distributions.

Aim The impact of microscale frost disturbance on vegetation colonization and successionary trends was examined within patterned ground features of Little Ice Age chronosequences. The goal was to investigate and compare vegetation response to micro-site frost disturbance with that of previous studies done at a coarser landscape scale.Location The study sites occur on Little Ice Age glacier forelands within Jotunheimen, Norway (61°-62°N). The forelands of the glaciers Slettmarkbreen, Styggedalsbreen and Vestre Memurubreen have been well studied providing chronological controls for landscape studies. Sorted patterned ground features are found within the chronosequences, typically declining with frost intensity and disturbance with increasing terrain age.Methods Micro-plots (8.3 · 8.3 cm) were placed at the inner borders and centres of patterned ground features. Species were identified and per cent species cover and per cent cover of life-form category were noted. Nonparametric Kruskal-Wallis and Mann-Whitney U-tests were used to test for differences between percent cover of life-form categories within patterned ground features as well as to identify thresholds of successional change across the chronosequences. ResultsSignificant relationships between life-from groups and patterned ground positions of varying ages were deduced using nonparametric statistics. Findings were then used to discuss trends of succession within patterned ground features and across the chronosequences. Vegetation establishment occurs at the border positions of young (< 30 years) patterned ground features. With time and distance from the ice margin, vegetation encroaches inwards toward the disturbed centres. Succession within patterned ground exhibits several stages: (1) bryophytes/crusts and lichens, (2) grasses/sedges and (3) woody shrubs. The occurrence of forbs was sporadic and generally non-significant.Main conclusions Frost disturbance in patterned ground appears to delay successional trends of vegetation communities when compared with previous studies on 'stable' terrain, producing micro-site lag effects. These small patches of disturbed ground are therefore important regarding vegetation assemblages across the landscape and are unlikely to be detected at the landscape scale.

In drylands of southeastern Utah, USA, the invasive exotic grass Bromus tectorum L. occurs in distinct spatial patterns suggesting soil control of ecosystem susceptibility to invasion. To improve our understanding of these patterns, we examined performance of B. tectorum in relation to additions of water, KCl, MgO, and CaO at seventeen 1600 m 2 sites distributed across a calcareous soil gradient in Canyonlands National Park. Water additions resulted in a 57% increase in B. tectorum establishment. Fall establishment was significantly correlated with silt and clay content in wet plots but not in dry plots, suggesting that texture effects on B. tectorum establishment patterns may be greater in wet years than in dry years. Applications of MgO resulted in a 49% decrease in B. tectorum establishment, although MgO had no effect on whole-plot biomass at the end of the growing season. B. tectorum-soil relations were strongest during winter (December-March) when relative growth rates were negatively related to soil acid-neutralizing potential, sand and CaCO 3 content, and a measure of bioavailable Mg; and positively related to silt and clay content, total N, measures of bioavailable Mn, P, and K, and a measure of magnetite indicating distributional patterns of eolian dust. As soils were persistently moist during this period, we attribute strong B. tectorum-soil patterns in winter to effects of low temperature on diffusion, microbial activity, and/or production of root exudates important for nutrient mobilization and uptake. In spring, there was a reversal in B. tectorum-soil relations such that loamy soils with higher B. tectorum densities were unfavorable for growth relative to sandy soils with higher warm-season water potentials. We conclude that resource limitations for B. tectorum in this study area shift seasonally, from water limitation of fall establishment, to nutrient limitation of winter growth, and back to water limitation of spring growth. Because study sites generally were arrayed along a hillslope gradient with downslope trends in soil vtexture and nutrient content, close B. tectorum-soil relations documented in this study indicate that a geomorphic framework is useful for understanding and predicting B. tectorum invasion patterns in dryland ecosystems of this region.