Cattle influence grassland dynamics in three ways: herbage removal, dung deposition and trampling. The objective of this study was to assess the effects of these factors, separately or in combination, and to compare them with cattle grazing over a one year period in a field experiment conducted in the Jura Mountains of northwestern Switzerland.

A set of controlled treatments simulating the three factors was applied in a fenced area: (1) repeated mowing – three levels; (2) intensive trampling – two levels; (3) manuring with a liquid mixture of dung and urine – three levels. All treatments were applied homogeneously to the entire surface of each of the 40 plots inside the exclosure. Additionally, ten plots outside the fenced area represented reference plots with regular cattle pasturing. The multivariate response of species composition was assessed three times with the point-intercept method: in spring before the treatments, in autumn after one season of treatments and at the beginning of the following year after winter rest.

Multivariate analyses of vegetation data in the first year showed an overwhelming seasonal shift and significant differences induced by treatments. Abandoned and manured plots showed the largest deviation from the cattle grazed reference. Herbage removal, simulated by repeated mowing, appeared to be the most important factor for maintaining vegetation texture. Seasonal treatment effects were only partially carried over to the next spring, showing an unexpected resilience of the plant community, probably due to life-history traits and competition release following climatic disturbance in winter.

Nomenclature: Tutin et al. (1964–1980).

Abbreviation: PCA = Principal Component Analysis; RDA = Redundancy analysis.

Patterns of plant species composition and their relationships to soil and topographic variables were investigated in tropical dry forests across the north central Yucatan, Mexico. Seven sites were studied in the oldest accessible forests along a 200-km transect oriented northwest to southeast; an eighth site was located in a little-disturbed area located 75 km northeast of the transect. Two of the sites were on Mayan ruins. All sites were sampled using 9 - 24, 10 m × 20 m plots (n = 132) for woody stems ≥ 3.0 cm diameter breast height. The important natural forest species were Bursera simaruba, Caesalpinia gaumeri, Gymnopodium floribundum, Piscidia piscipula, and Thouinia paucidentata. The two most important woody species in ruin woodlands were Brosimum alicastrum and Croton lundellii. Forest plots (n =108) had 17 species on average, ruin plots (n = 24) nine species. Mean basal area of stems at the forest plots (20.7 m².ha⁻¹) was lower than in ruin plots (28.4 m².ha⁻¹). Detrended Correspondence Analysis generally placed plots by site along the geographic transect. Natural forest plots and sites were separated from the plots on ruin sites. The five soil and topographic variables (slope, soil depth, percent surface rock, soil pH, total soil organic matter) differed significantly among sites. Plot values were correlated with DCA axe scores. Intersite floristic variation reflects an overall west to east environmental gradient affected by climate.

Nomenclature: White & Darwin (1995) and the floras and monographs used therein.

We calculated annual mean stem volume increment (AMSVI) and total litter fall to produce forest net primary production (NPP) maps at 1-km² and half-degree resolutions in Finland and Sweden. We used a multi-scale methodology to link field inventory data reported at plot and forestry district levels through a remotely sensed total plant biomass map derived from 1-km² AVHRR image. Total litter fall was estimated as function of elevation and latitude. Leaf litter fall, a surrogate for fine root production, was estimated from total litter fall by forest type. The gridded NPP estimates agreed well with previously reported NPP values, based on point measurements. Regional NPP increases from northeast to southwest. It is positively related to annual mean temperature and annual mean total precipitation (strongly correlated with temperature) and is negatively related to elevation at broad scale. Total NPP (TNPP) values for representative cells selected based on three criteria were highly correlated with simulated values from a process-based model (CEVSA) at 0.5° × 0.5° resolution.

At 1-km² resolution, mean above-ground NPP in the region was 408 g/m²/yr ranging from 172 to 1091 (standard deviation (SD) = 134). Mean TNPP was 563 (252 to 1426, SD = 176). Ranges and SD were reduced while the mean values of the estimated NPP stayed almost constant as cell size increased from 1-km² to 0.5° × 0.5°, as expected. Nordic boreal forests seem to have lower productivity among the world boreal forests.

Abbreviations: ABIO = Above-ground biomass; AMSVI = Annual mean stem volume increment; ANPP = Above-ground net primary production; AVHRR = Advanced Very High Resolution Radiometer; NPP = Net primary production; RS = Remote sensing; SBIO = Stem biomass; TNPP = Total net primary production; TPB = Total plant biomass.

Lake Uddjaur in northern Sweden was formed as a consequence of non-uniform glacio-isostatic uplift in which a forested valley was gradually flooded and high elevation areas became islands. We hypothesized that small islands in Lake Uddjaur burnt through lightning strike more frequently when they were part of a large forested area compared to when they became true islands, and that this reduction in fire impact has enhanced the domination of late successional species and humus accumulation. Fire history and vegetation dynamics were studied by analysis of macroscopic charcoal (> 0.5 mm) and pollen in humus profiles from two islands. According to a model of isostatic uplift, the islands became gradually isolated from the mainland between ca. 2000 to 1000 BP, i.e. during the same time that fire impact declined. Prior to that, both islands were part of a Pinus-Betula forest landscape affected by fires from ca. 5800 to ca. 1500 BP. Thereafter fire influence ceased and the islands became more strongly characterized by late successional species, e.g. Picea. This change was associated with substantial humus accumulation. The decreased fire influence on these islands contrasts with the regional increase in fire influence during the last 1000 yr. Long-term influence of wildfire is important in vegetation dynamics and humus accumulation and, thus, post-glacial isostatic land uplift can indirectly have a substantial influence on ecosystem development. Consequently, this effect should be further considered in long-time ecosystem studies of areas with large, non-uniform land uplift such as those found in northern Fennoscandia and eastern Canada.

Abbreviations: BP = Radiocarbon based age expressed as Before Present = before 1950.

Nomenclature: Lid (1985); Koponen et al. (1977).

Vascular plant species compositional patterns of the low forest, scrub, and herbaceous vegetation on white sand soils and sandstone substrates were studied at six sandstone plateaus in Colombian Amazonia, by means of a field survey according to the Braun-Blanquet relevé method. Canonical Correspondence Analysis (CCA) was applied to separate effects of habitat and spatial configurations of the plateaus on species patterns. Also, information on dispersal ability and phytogeographic affinity of species was used to test explanations for between plateau differences.

Low trees, shrubs and herbs were the main species recorded in 212 relevés. The main gradients in the species patterns were linked to the spatial configuration of the plateaus. Spatially controlled species patterns were mainly related to soil depth and soil organic matter. The association between phytogeographic affinity and the habitat controlled spatial link of species pointed at insufficient sampling at one plateau. Dispersal ability did not explain the habitat controlled spatially distributed occurrences of plant species. This might indicate a low frequency of local extinctions at the sandstone plateaus, especially of the poorly dispersed species, possibly because plant populations survive fire or drought disturbances in local sheltered places. Space and habitat controlled species patterns at one plateau were quite distinct from patterns at the other plateaus. This might be due to unmeasured habitat factors (e.g. unrecorded soil variation or human disturbance history) or the preferential, surveyor biased sampling procedure.

Nomenclature: Arbeláez & Callejas (1999).

Abbreviation: UTM = Universal Transverse Mercator.

Patterns of moss and liverwort species diversity – species richness and species turnover (β-diversity) – in three conifer-dominated boreal forest stands of northern Alberta, Canada are described. We examined the relationship between bryophyte species diversity and micro-environment at two sample grains, the microsite – substrate types for moss colonization: logs, stumps, tree bases, undisturbed patches of forest floor (dominated by feather moss species), and disturbed patches of forest floor – and the mesosite (25 m × 25 m plots). Microsite type and properties (e.g. decay class, hardwood vs softwood, pH) were the principal predictors of bryophyte species diversity and not micro-environment variation among mesosites. Microsite type was the strongest predictor of microsite species richness and β-diversity was higher among microsites and types and within microsites than among mesosites or stands. Microsite properties were significant predictors of species richness for all microsite types. Log and stump decay classes, influenced also by hardwood vs softwood predicted species richness of woody microsite types and soil pH and moisture predicted species richness of forest floor microsites. β-diversity was highest for tree bases and disturbed patches of forest floor and lowest for logs. Mesosite β-diversity was lower than that among microsites, and mesosite species richness was not well explained by measured environmental parameters. Results suggest that in conifer-dominated boreal stands, species richness of microsites is only negligibly influenced by within-stand variation at the mesosite grain and that substrate characteristics are the most important predictors of bryophyte species diversity in this ecosystem.

Nomenclature: Anderson et al. (1990) for bryophytes except for Sphagnaceae (Anderson1990) and Hepaticae (Stotler & Crandall-Stotler 1977); Orthotrichum elegans is recognized as a distinct species from Orthotrichum speciosum (Vitt & Darigo 1997).

In present day European landscapes many forest plant species are restricted to isolated remnants of a formerly more or less continuous forest cover. The two major objectives of this study were (1) to determine the relative importance of habitat quality (mainly in terms of soil parameters), habitat configuration (patch size and isolation) and habitat continuity for the distribution of herbaceous forest plant species in a highly fragmented landscape and (2) to examine if groups of species with different habitat requirements are affected differently. Deciduous forest patches in northwestern Germany were surveyed for the presence of a large set of forest species. For each patch, habitat quality, configuration and continuity were determined. Data were analysed by Redundancy Analysis with variation partitioning for effects on total species composition and multivariate logistic regression for effects on individual species, for two different data sets (base-rich and base-poor forest patches). Overall, we found strong effects of habitat quality (particularly of soil pH, water content and topographic heterogeneity in the base-rich forest patches; and of calcium content and disturbance in the base-poor patches), but only relatively weak effects of habitat configuration and habitat continuity. However, a number of species were positively affected by patch area and negatively affected by patch isolation. Furthermore, the relative importance of habitat configuration tended to be higher for species predominantly growing in closed forests compared to species occurring both in the forest and in the open landscape.

Abbreviations: GIS = Geographic information system; RDA = Redundancy analysis.

Nomenclature: Wisskirchen & Haeupler (1998).

Many rare plant species occur in Dutch wet dune slacks, particularly in the Junco baltici-Schoenetum nigricantis. For nature management it is important to understand the processes controlling the presence of these basiphilous early successional communities, which is why we investigated vegetation and soil development during succession in coastal dune slacks. We compared 12 chronosequential stages of 0, 2, 4, 9, 10, 13, 25, 30, 43, 60, 70 and 85 yr in five different dune slack systems. In four of these locations turf had earlier been removed in order to restore the basiphilous pioneer stage.

The main variation in the vegetation is related to the acidification/soil enrichment gradient and the salinity/maritime gradient. During succession, organic matter accumulates and acidification takes place. Maritime influence can buffer the soil and postpone the succession of basiphilous pioneer communities for many years. A significant correlation with age was found for 18 variables. Multiple regressions predicted changes in the vegetation (dependent variables: biomass, cover of Salix repens, Calamagrostis epigejos and Schoenus nigricans) as a function of acidification, organic matter accumulation, increase in available P and presence of Na in the soil.

We conclude that natural ageing of the vegetation and the associated process of accumulation of biomass drive succession in this hydrosere. The underlying soil processes are acidification and organic matter accumulation. During succession dominance shifts from S. nigricans to S. repens or C. epigejos. Maintenance of the pioneer character of the habitat is only possible by local intervention or by natural or man-induced dune forming. The effect of sod-stripping depends on the environmental conditions and, in case of acidification, success is limited. Succession can be postponed by mowing.

Abbreviations: ROL = Radial oxygen loss.

Nomenclature: van der Meijden (1990) for vascular plants; Touw & Rubers (1989) for mosses; Gradstein & van Melick (1996) for liverworts; Schaminée et al. (1995, 1996, 1998) for syntaxa. Agrostis stolonifera was probably identified as A. canina in some cases. The liverworts Moerckia hibernica and Pellia endiviifolia may have been confused but probably co-occurred in most plots.

Climate-related increases in terrestrial vegetation activity in the northern regions of the Northern Hemisphere have been identified by recent satellite based studies. However, evidence for this increase from ground observations is very limited. In the current study, we used a time series data set for the Normalized Difference Vegetation Index (NDVI) for the growing season (April to October) from 1982 to 1999, along with historical climate data, to analyse year to year variations in vegetation activity and to explore the relationship between the growing season NDVI and climatic variables in China. Vegetation activity, as measured by NDVI, increased in 81% of the study area, with significant gains in 27% of the region. The magnitude of the mean growing season NDVI for the 1980s and the 1990s was not significantly different. The increase in NDVI corresponded to an increase in temperature on the national scale, while regional variations in NDVI appeared to be related to precipitation. The NDVI trend showed a large spatial heterogeneity, possibly associated with changes in regional climate, land use and vegetation type. Our study suggests that agricultural practices caused an increase in NDVI in some regions, and rapid urbanization on the east coast resulted in a sharp decrease in NDVI since the 1980s.

Abbreviations: AVHRR = Advanced Very High Resolution Radiometer; GIMMS = Global Inventory Monitoring and Modelling Studies; MVC = Maximum NDVI Value Composite; NDVI = Normalized Difference Vegetation Index; NOAA = National Oceanic and Atmospheric Administration.

We investigated some of the factors influencing exotic invasion of native sub-alpine plant communities at a site in southeast Australia. Structure, floristic composition and invasibility of the plant communities and attributes of the invasive species were studied. To determine the plant characteristics correlated with invasiveness, we distinguished between roadside invaders, native community invaders and non-invasive exotic species, and compared these groups across a range of traits including functional group, taxonomic affinity, life history, mating system and morphology. Poa grasslands and Eucalyptus-Poa woodlands contained the largest number of exotic species, although all communities studied appeared resilient to invasion by most species. Most community invaders were broad-leaved herbs while roadside invaders contained both herbs and a range of grass species. Over the entire study area the richness and cover of native and exotic herbaceous species were positively related, but exotic herbs were more negatively related to cover of specific functional groups (e.g. trees) than native herbs. Compared with the overall pool of exotic species, those capable of invading native plant communities were disproportionately polycarpic, Asteracean and cross-pollinating. Our data support the hypothesis that strong ecological filtering of exotic species generates an exotic assemblage containing few dominant species and which functionally converges on the native assemblage. These findings contrast with those observed in the majority of invaded natural systems. We conclude that the invasion of closed sub-alpine communities must be viewed in terms of the unique attributes of the invading species, the structure and composition of the invaded communities and the strong extrinsic physical and climatic factors typical of the sub-alpine environment.

Nomenclature: Australian Plant Name Index (APNI);  http://www.anbg.gov.au/cgi-bin/apni

Abbreviations: KNP = Kosciuszko National Park; MRPP = Multi response permutation procedure; VE = Variance explained.

We investigated the responses of two co-existing Mediterranean trees with different regeneration strategies (Phillyrea latifolia seedlings and Quercus ilex sprouts) to experimental drought below the forest canopy. We considered different recruitment stages and used leaf isotopic discrimination to estimate water use efficiency (WUE) and nitrogen availability and use. Drought decreased the emergence and survival of seedlings and sprouts. Survival and growth of older saplings were not influenced by drought. Seedling emergence of P. latifolia was higher than Q. ilex sprout production, but Q. ilex sprouts had higher survival and growth rates. These differences disappeared in the sapling stage. Carbon isotopic discrimination suggested that Q. ilex sprouts had higher WUE than P. latifolia seedlings. Drought increased WUE of recruits, particularly in Q. ilex. Water use regulation increased with ageing, particularly in P. latifolia. Q. ilex had higher δ¹⁵N values than P. latifolia; these were also higher under drier soil conditions. Current year seedlings had higher δ¹⁵N than saplings, particularly in P. latifolia, suggesting they exploit superficial soil layers. These results suggest that sprouts obtain benefit from resources stored in parent plants. At earlier stages, they perform better than seedlings. This response is not coupled to adult vulnerability to drought for these species, revealing the difficulty of predicting species dynamics during climate change.

Abbreviations: TDR = Time domain reflectometry; WUE = Water use efficiency.

Nomenclature: Bolòs et al. (1990).

We examined forest structure and regeneration in a 350-ha forest dominated by Pinus sylvestris 31 yr after a wildfire in the Vienansalo wilderness, Russian Karelia. In most parts of the area, the 1969 fire was not stand replacing but had left larger trees alive so that the area generally remained forest covered. In some localities, however, all trees apparently died and distinct gaps were formed, suggesting that the fire severity varied considerably, contributing to increased variation in stand structure. Living and dead wood volumes were similar, 112 and 96 m³.ha⁻¹, respectively. The tree species proportions of dead vs living wood indicated that prior to fire disturbance Picea was more common in the area. Regeneration was abundant (saplings, ca. 14 000 ind.ha⁻¹, height 20 - 200 cm) and tree seedling recruitment had occurred over a long period of time. Regeneration density was highest on the mesic Vaccinium-Myrtillus forest site type, decreasing towards nutrient-poor site types. The most common regeneration microsites were level ground (56% of saplings), immediate surroundings of decayed wood (23%) and depressions (11%). The high proportion of saplings on level ground suggests that after the fire regeneration conditions have been favourable across the whole forest floor. Nevertheless, the areas in the vicinity of decayed wood have been particularly important microsites for seedling establishment. The results provide an example of the effects of wildfire on forest structure in a natural Pinus sylvestris dominated forest, demonstrating the non stand replacing character of fire, high variability in stand structure and the abundance of post-fire regeneration.

Nomenclature: Hämet-Ahti et al. (1986).

Abbreviations: CV = Coefficient of variation; CWD = Coarse woody debris; ECT = Empetrum-Calluna type; EVT = Empetrum-Vaccinium type; GOMT = Geranium-Oxalis-Myrtillus type; VMT = Vaccinium-Myrtillus type.

We examined the distribution of tree species across five habitats in 69 small plots within a single watershed of the Gunung Palung National Park, West Kalimantan, Indonesia (GPNP). The spatially complex distribution and close proximity of habitats provided an opportunity to test habitat specificity of tree species across strong environmental gradients, in a situation where dispersal into ‘inappropriate’ habitat should not be a rare event. Habitat had a weak influence on community structure, although species diversity was lower in the alluvium and peat habitats. Association tests based on two randomization models (spatially independent and explicit) were used to examine habitat distribution of 55 ‘common’ and 142 ‘frequent’ taxa. The general patterns were similar in the two models but the interpretation of specific patterns depended greatly on assumptions about dispersal ability. A majority (67%) of the common species was significantly associated with a single habitat, while few were restricted to one habitat. A small proportion (16%) of the species appear to be habitat generalists. The peat habitat had the most profound effect on species distribution. Overall, a large amount of variation was found in the degree of habitat specificity, even within speciose groups. No obvious evolutionary or ecological correlates with degree of habitat specificity were found. These results suggest that a mixture of stochastic and deterministic processes determine species distribution even across strong environmental gradients.

Abbreviations: GPNP = Gunung Palung National Park.; SEMI = spatial dependence; SPI = spatial independence.

Nomenclature: Ng (1978, 1989); Soepadmo & Wong (1995); Soepadmo et al. (1996); Whitmore (1972, 1973b).

Ice storms cause periodic disturbance to temperate forests of eastern North America. They are the primary agents of disturbance in some eastern forests. In this paper, a forest gap model is employed to explore consequences of ice storms for the long-term dynamics of Tsuga canadensis-northern hardwoods forests. The gap model LINKAGES was modified to simulate periodic ice storm disturbance in the Adirondack Mountains of New York. To adapt the gap model for this purpose, field data on ice storm disturbance are used to develop a polytomous logistic regression model of tree damage. The logistic regression model was then incorporated into the modified forest gap model, LINKADIR, to determine the type of damage sustained by each simulated tree.

The logistic regression model predicts high probabilities of bent boles or severe bole damage (leaning, snapping, or uprooting) in small-diameter trees, and increasing probability of canopy damage as tree size increases. Canopy damage is most likely on gentle slopes; the probability of severe bole damage increases with increasing slope angle.

In the LINKADIR simulations, tree damage type determines the probability of mortality; trees with severe bole damage are assigned the highest mortality rate. LINKADIR predicts Tsuga canadensis dominance in mesophytic old-growth forests not disturbed by ice storms. When ice storms are simulated, the model predicts Acer saccharum-dominated forests with higher species richness. These results suggest that ice storms may function as intermediate disturbances that enhance species richness in forested Adirondack landscapes.

Nomenclature: Magee & Ahles (1999).

Abbreviation: GDD = Growing degree day.

Although many methods have been proposed for analysing point locations for spatial pattern, previous methods have concentrated on clumping and spacing. The study of anisotropy (changes in spatial pattern with direction) in point patterns has been limited by lack of methods explicitly designed for these data and this purpose; researchers have been constrained to choosing arbitrary test directions or converting their data into quadrat counts and using methods designed for continuously distributed data. Wavelet analysis, a booming approach to studying spatial pattern, widely used in mathematics and physics for signal analysis, has started to make its way into the ecological literature. A simple adaptation of wavelet analysis is proposed for the detection of anisotropy in point patterns. The method is illustrated with both simulated and field data. This approach can easily be used for both global and local spatial analysis.

Logistic Multiple Regression, Principal Component Regression and Classification and Regression Tree Analysis (CART), commonly used in ecological modelling using GIS, are compared with a relatively new statistical technique, Multivariate Adaptive Regression Splines (MARS), to test their accuracy, reliability, implementation within GIS and ease of use. All were applied to the same two data sets, covering a wide range of conditions common in predictive modelling, namely geographical range, scale, nature of the predictors and sampling method.

We ran two series of analyses to verify if model validation by an independent data set was required or cross-validation on a learning data set sufficed. Results show that validation by independent data sets is needed. Model accuracy was evaluated using the area under Receiver Operating Characteristics curve (AUC). This measure was used because it summarizes performance across all possible thresholds, and is independent of balance between classes.

MARS and Regression Tree Analysis achieved the best prediction success, although the CART model was difficult to use for cartographic purposes due to the high model complexity.

Abbreviations: AUC = Area under the ROC curve; CART = Classification Regression Trees; FN = False negative; FP = False positive; GAM = Generalized Additive Model; GIS = Geographic Information System; GLM = Generalized Linear Model; LMR = Logistic Multiple Regression; MARS = Multivariate Adaptive Regression Splines; NDVI = Normalized Difference Vegetation Index; PCR = Principal Components Regression; ROC = Receiver Operating Characteristics.