Two simulated coenoclines and a real data set were differently recoded with respect to the Braun-Blanquet coding (including presence/absence) and analysed through the most common multidimensional scaling methods. This way, we aim at contributing to the debate concerning the nature of the Braun-Blanquet coding and the consequent multidimensional scaling methods to be used. Procrustes, Pearson, and Spearman correlation matrices were computed to compare the resulting sets of coordinates and synthesized through their Principal Component Analyses (PCA). In general, both Procrustes and Pearson correlations showed high coherence of the obtained results, whereas Spearman correlation values were much lower. This proves that the main sources of variation are similarly identified by most of used methods/transformations, whereas less agreement results on the continuous variations along the detected gradients. The conclusion is that Correspondence Analysis on presence/absence data seems the most appropriate method to use. Indeed, presence/absence data are not affected by species cover estimation error and Simple Correspondence Analysis performs really well with this coding. As alternative, Multiple Correlation Analysis provides interesting information on the species distribution while showing a pattern of relevés very similar to that issued by PCA.
The identification of differentiating species, also termed indicator species, is a key issue in striving for group pattern in vegetation samples. With this objective in mind Jancey (1979) proposed species ranking on a variance ratio (F-ratio) criterion, the mechanism of analysis of variance. Almost 20 years later Dufrêne and Legendre (1997) presented their indicator value analysis with the same objective in mind but based on different reasoning. This raises the question if (i) the results of the two approaches are equivalent or (ii), if one of the measures of performance is superior in predicting specific properties of ecosystems, such as site conditions, biodiversity, succession or other. Because the outcome of both methods strictly relies on the strength of vegetation pattern reflected by the data sets used as well as the quality of classifications, we compare results from a small and also a large real-world data set and we evaluate the effect of the number of groups involved when clustering sites. In a subsequent step, the ranking of indicator species identified by either of the methods is compared with a ranking obtained by correlating species with measured environmental factors. The results confirm that the outcome of ranking by maximum indicator value (IndVal) according to Dufrêne and Legendre (1997) is very similar to the ranking devised by Jancey (1979). Rank correlation reaches a maximum of r > 0.95 when the data set is large and group number in clustering is low. In our examples Jancey’s method is generally more closely related to the environmental predictive power of species, outperforming IndVal when applied to continuous variables measured in the field. We conclude that the potential of Jancey’s method is generally underrated. As expected, the agreement of results between the two approaches depends on the strength of similarity patterns inherent in the data sets analysed. The method of Dufrêne and Legendre (1997) is well adapted to issues of phytosociology where classification is frequently based on expert knowledge. If the resolving power of species is used as a surrogate for patterns and processes of plant-environment systems, then ranking by variance ratio may be the more promising approach.
In this paper, we examine the application of a particular approach to induction, the minimum message length principle and illustrate some of the problems that can be addressed through its use. The MML principle seeks to identify an optimal model within some specified parameterised class of models and for this paper we have chosen to concentrate on a single model class, that of mixture separation or fuzzy clustering. The first section presents, in outline, an MML methodology for fuzzy clustering. We then present some applications, including the nature of the within-cluster model, examination of the univocality of results for different groups of species and the effectiveness of presence data compared to purely quantitative data. Finally, we examine some possibilities of extending MML methodology to include within-class correlation of species, the existence of dependence between observed samples and the comparison of different classes of models.
Among other objectives, forest inventories are aimed to identify ecological communities and to correlate community composition with environmental variables. The identification of different communities would require several forest inventories, each covering small sampling areas with relatively homogeneous environmental conditions. The multiple plot sampling method, traditionally used in local inventories of tropical forests, cannot assure such homogeneity, since even small sampling areas would show environmental heterogeneity influencing vegetation. In this paper we assessed the consequences of this heterogeneity for sampling by quantifying the variability of species abundance ranks for species sampled with 10 or more individuals in a set of plots covering a small sampling area. The species reference abundance ranks were obtained from a sample of 100 plots of 10 m × 10 m each randomly set in a sampling area of 6.5 ha in a tropical forest fragment (Southeastern Brazil). For each species we used resamplings (30 trials) to obtain the species abundance ranks in sub-samples, considering different sampling intensities (n = 25, 50 and 75 plots), and compared these ranks with the species reference rank (n = 100 plots). Then, we compared the species ranks in sub-samples of 50 plots (10.000 trials) with the reference rank and assessed the frequency and extent of rank displacements. Species rank was highly variable across resampling trials for the sampling intensities of n = 25 and n = 50, but decreased considerable with a sampling intensity of n = 75 plots. The mean rank variability and especially the maximum displacement raised significantly from the seventh most abundant species on, and some species occupied quite discrepant abundance ranks in up to 10% of the 10.000 resampling trials. This high internal variability of forest samples may impair the search for floristic patterns as scale lessens, say, to the meso-scale (1–100 km2). We discussed some possible ways to increase internal homogeneity of tropical forest samples with the multiple plot sampling method. Among these, objective entitation, based on an ancient phytosociological procedure, is suggested as the most appropriate for use on the hilly relieves of the Atlantic forest biome.
Ihm, P. and H. van Groenewoud. 1972A multivariate ordering of vegetationdata based on Gaussian type gradient response curves J. Ecol. 63: 767-777.
multivariate ordering of vegetationdata based on Gaussian type gradient
This paper demonstrates a possible application of large historical vegetation data sets as reference to reveal natural trends. Phytosociological releves re-sampled after 3–6 decades were used to detect and interpret long-term plant compositional changes of seven rock grassland communities in Hungary. Altogether 151 re-established plots were subject of the study. Data analyses were designed to minimize the negative effects arising from the application of historical information. Principal coordinates analysis was used to discover general compositional changes. With the help of ecological indicator values and species attributes, vegetation state trends were evaluated. Principal coordinates analysis reveals a uniform displacement of plot averages in the ordination space. Ecological indicator values for nitrogen requirement of vascular plants show a significant increase in mesotrophic categories. Occurrences of species typical of rock grasslands decreased significantly. Natural pioneers, disturbance-tolerant and weed species increased in number. Nevertheless, their amount is relatively low and the natural constituents of the communities still dominate, which is a sign of only a minor disturbance. Considering the wide geographical distribution of the sample plots, general changes seem to indicate pressures operating at large scales. These include elevated nitrogen deposition, increased rates of erosion and trampling caused by overpopulated ungulates and more frequent summer drought events. Acidification only occurs in silicate grasslands as calcareous soil types have higher buffering capacities. The global tendency of biotic homogenisation with the increase of common species was also detected over time. The study showed that the use of historical vegetation data enables us to estimate long-term trends in vegetation state.
Authors:P. B. Drewa, P. B. Drewa, G. E. Bradfield, and G. E. Bradfield
The influence of scale on the discernment of plant community patterns was examined using vegetation-environment data collected from a subalpine wet meadow in south-coastal British Columbia. Species cover data were recorded in 225, 0.25 m2quadrats systematically located at 5m intervals in a 40 m x 120 m sampling grid. Environmental data consisted of quadrat elevations as well as measured and kriged estimates of five soil variables (carbon content, pH, electrical conductivity, percent sand, and percent clay). Sampling scale was adjusted by aggregating neighbouring quadrats into composite sampling units; analytical scale was altered by varying the intercept level in dendrograms from minimum increase of sum of squares cluster analysis of the vegetation data corresponding to the different sampling scales. The resulting classifications were evaluated for their ability to explain variation in the vegetation data and in the environmental data. The vegetation variation explained by the classifications was highest at the smallest sampling scale indicating that vegetation heterogeneity is fine grained. In contrast, the environmental variation explained was higher for the classifications based on the larger composite sampling units implying a coarser scaling of abiotic conditions within the study area. These results were consistent with the recognition of three main zones along a drainage gradient within the sampling grid. upper mixed-forb, middle heath, and lower sedge. There was also evidence that the orientation of rectangular sampling units parallel to the drainage gradient leads to higher levels of explained variation. This study reaffirms the need for careful consideration of alternatives both in field sampling and analytical phases of vegetation research to ensure that description and interpretation of patterns adequately address study objectives and that vegetation-environment relationships are more completely investigated from a hierarchical perspective.
There are several community-based bioindicator systems that use species presence or abundance data as proxies for environmental variables. One example is the Ellenberg system, whereby vegetation data are used to estimate environmental soil conditions. Despite widespread use of Ellenberg values in ecological research, the correlation between bioindicated values and actual values is often an implicit assumption rather than based on empirical evidence. Here, we correlate unadjusted and UK-adjusted Ellenberg values for soil moisture, pH, and nitrate in relation to direct environmental measures for 50 woodland sites in the UK, which were subject to repeat sampling. Our results show the accuracy of Ellenberg values is parameter specific; pH values were a good proxy for direct environmental measures but this was not true for soil moisture, when relationships were weak and non-significant. For nitrates, there were important seasonal differences, with a strong positive logarithmic relationship in the spring but a non-significant (and negative) correlation in summer. The UK-adjusted values were better than, or equivalent to, Ellenberg’s original ones, which had been quantified originally for Central Europe, in all cases. Somewhat surprisingly, unweighted values correlated with direct environmental measures better than did abundance-weighted ones. This suggests that the presence of rare plants can be highly important in accurate quantification of soil parameters and we recommend using an unweighted approach. However, site profiles created only using rare plants were inferior to profiles based on the whole plant community and thus cannot be used in isolation. We conclude that, for pH and nitrates, the Ellenberg system provides a useful estimate of actual conditions, but recalibration of moisture values should be considered along with the effect of seasonality on the efficacy of the system.