Quantifying tolerance indicator values for common stream fish species of the United States

The classification of fish species tolerance to environmental disturbance is often used as a means to assess ecosystem conditions. Its use, however, may be problematic because the approach to tolerance classification is based on subjective judgment. We analyzed fish and physicochemical data from 773 stream sites collected as part of the U.S. Geological Survey's National Water-Quality Assessment Program to calculate tolerance indicator values for 10 physicochemical variables using weighted averaging. Tolerance indicator values (TIVs) for ammonia, chloride, dissolved oxygen, nitrite plus nitrate, pH, phosphorus, specific conductance, sulfate, suspended sediment, and water temperature were calculated for 105 common fish species of the United States. Tolerance indicator values for specific conductance and sulfate were correlated (rho = 0.87), and thus, fish species may be co-tolerant to these water-quality variables. We integrated TIVs for each species into an overall tolerance classification for comparisons with judgment-based tolerance classifications. Principal components analysis indicated that the distinction between tolerant and intolerant classifications was determined largely by tolerance to suspended sediment, specific conductance, chloride, and total phosphorus. Factors such as water temperature, dissolved oxygen, and pH may not be as important in distinguishing between tolerant and intolerant classifications, but may help to segregate species classified as moderate. Empirically derived tolerance classifications were 58.8% in agreement with judgment-derived tolerance classifications. Canonical discriminant analysis revealed that few TIVs, primarily chloride, could discriminate among judgment-derived tolerance classifications of tolerant, moderate, and intolerant. To our knowledge, this is the first empirically based understanding of fish species tolerance for stream fishes in the United States.     

Classification of vegetation sequences in Toohey Forest,Queensland

In this paper we consider one method of mapping larger units identified from the spatial pattern of sequences of vegetation types. The basic data were presence/absence data for 6450 stands arranged in 90 transects. A second set of data was derived by averaging the species occurrences in non-overlapping groups of 5 stands. A divisive numerical classification was used to determine the primary vegetation units. In all, 5 different sets of primary types were derived, using different species suites, different sample sizes and different numerical methods. We briefly discuss the types identified and their spatial patterns in the area.Each of these types was then used to define a string of type-codes for every transect so that each transect represents a sample from the landscape containing information on the frequency and spatial distribution of the primary vegetation types. The transects may be classified using a Levenshtein dissimilarity measure and agglomerative hierarchical classification, giving 5 analyses of transects, one for each of the primary types discussed above. We then examine these transect classifications to investigate the stability of the vegetation landspace patterns under changes in species used for the primary classification, in size of sample unit and in method of primary classifications. There is a considerable degree of stability in the results. However it seems with this vegetation that the tree species and non-tree species have considerable independence. We also indicate some problems with this approach and some possible extensions.     

Selecting discriminant function models for predicting the expected richness of aquatic macroinvertebrates

1. The predictive modelling approach to bioassessment estimates the macroinvertebrate assemblage expected at a stream site if it were in a minimally disturbed reference condition. The difference between expected and observed assemblages then measures the departure of the site from reference condition. 2. Most predictive models employ site classification, followed by discriminant function (DF) modelling, to predict the expected assemblage from a suite of environmental variables. Stepwise DF analysis is normally used to choose a single subset of DF predictor variables with a high accuracy for classifying sites. An alternative is to screen all possible combinations of predictor variables, in order to identify several ‘best’ subsets that yield good overall performance of the predictive model. 3. We applied best‐subsets DF analysis to assemblage and environmental data from 199 reference sites in Oregon, U.S.A. Two sets of 66 best DF models containing between one and 14 predictor variables (that is, having model orders from one to 14) were developed, for five‐group and 11‐group site classifications. 4. Resubstitution classification accuracy of the DF models increased consistently with model order, but cross‐validated classification accuracy did not improve beyond seventh or eighth‐order models, suggesting that the larger models were overfitted. 5. Overall predictive model performance at model training sites, measured by the root‐mean‐squared error of the observed/expected species richness ratio, also improved steadily with DF model order. But high‐order DF models usually performed poorly at an independent set of validation sites, another sign of model overfitting. 6. Models selected by stepwise DF analysis showed evidence of overfitting and were outperformed by several of the best‐subsets models. 7. The group separation strength of a DF model, as measured by Wilks’Λ, was more strongly correlated with overall predictive model performance at training sites than was DF classification accuracy. 8. Our results suggest improved strategies for developing reliable, parsimonious predictive models. We emphasise the value of independent validation data for obtaining a realistic picture of model performance. We also recommend assessing not just one or two, but several, candidate models based on their overall performance as well as the performance of their DF component. 9. We provide links to our free software for stepwise and best‐subsets DF analysis.     

Non-random co-occurrence of native and exotic plant species in Mediterranean grasslands

Invasion by exotic species in Mediterranean grasslands has determined assembly patterns of native and introduced species, knowledge of which provides information on the ecological processes underlying these novel communities. We considered grasslands from Spain and Chile. For each country we considered the whole grassland community and we split species into two subsets: in Chile, species were classified as natives or colonizers (i.e. exotics); in Spain, species were classified as exclusives (present in Spain but not in Chile) or colonizers (Spanish natives and exotics into Chile). We used null models and co-occurrence indices calculated in each country for each one of 15 sites distributed along a precipitation gradient and subjected to similar silvopastoral exploitation. We compared values of species co-occurrence between countries and between species subsets (natives/colonizers in Chile; exclusives/colonizers in Spain) within each country and we characterised them according to climatic variables. We hypothesized that: a) the different coexistence time of the species in both regions should give rise to communities presenting a spatial pattern further from random in Spain than in Chile, b) the co-occurrence patterns in the grasslands are affected by mesoclimatic factors in both regions. The patterns of co-occurrence are similar in Spain and Chile, mostly showing a spatial pattern more segregated than expected by random. The colonizer species are more segregated in Spain than in Chile, possibly determined by the longer residence time of the species in the source area than in the invaded one. The segregation of species in Chile is related to water availability, being species less segregated in habitat with greater water deficit; in Spain no relationship with climatic variables was found. After an invasion process, our results suggest that the possible process of alteration of the original Chilean communities has not prevented the assembly between the native and colonizer species together.     

A comprehensive framework for the study of species co‐occurrences,nestedness and turnover

Binary presence–absence matrices (rows = species, columns = sites) are often used to quantify patterns of species co‐occurrence, and to infer possible biotic interactions from these patterns. Previous classifications of co‐occurrence patterns as nested, segregated, or modular have led to contradictory results and conclusions. These analyses usually do not incorporate the functional traits of the species or the environmental characteristics of the sites, even though the outcomes of species interactions often depend on trait expression and site quality. Here we address this shortcoming by developing a method that incorporates realized functional and environmental niches, and relates them to species co‐occurrence patterns. These niches are defined from n‐dimensional ellipsoids, and calculated from the n eigenvectors and eigenvalues of the variance–covariance matrix of measured environmental or trait variables. Average niche overlap among species and the spatial distribution of niches define a triangle plot with vertices of species segregation (low niche overlap), nestedness (high niche overlap), and modular co‐occurrence (clusters of overlapping niches). Applying this framework to temperate understorey plant communities in southwest Poland, we found a consistent modular structure of species occurrences, a pattern not detected by conventional presence–absence analysis. These results suggest that, in our case study, habitat filtering is the most important process structuring understorey plant communities. Furthermore, they demonstrate how incorporating trait and environmental data into co‐occurrence analysis improves pattern detection and provides a stronger theoretical framework for understanding community structure.     

OptimClass: Using species‐to‐cluster fidelity to determine the optimal partition in classification of ecological communities

Question: Community ecologists are often confronted with multiple possible partitions of a single set of records of species composition and/or abundances from several sites. Different methods of numerical classification produce different results, and the question is which of them, and how many clusters, should be selected for interpretation. We demonstrate a new method for identifying the optimal partition from a series of partitions of the same set of sites, based on number of species with high fidelity to clusters in a partition (faithful species). Methods: The new method, OptimClass, has two variants. OptimClass 1 searches the partition with the maximum number of faithful species across all clusters, while OptimClass 2 searches the partition with the maximum number of clusters that contain at least a preselected minimum number of faithful species. Faithful species are determined based on the P value of the Fisher's exact test, as a measure of fidelity. OptimClass was tested on three vegetation datasets that varied in species richness and internal heterogeneity, using several classification algorithms, resemblance measures and cover transformations. Results: Results from both variants of OptimClass depended on the preselected threshold P value for faithful species: higher P gave higher probability that a partition with more clusters was selected as optimal. Good partitions, in terms of OptimClass criteria, involved flexible beta clustering, and also ordinal clustering. Good partitions were also obtained with TWINSPAN when the required number of clusters was small, or UPGMA when the required number of clusters was large. Poor partitions usually resulted from classifications that used resemblance measures and cover transformations emphasizing differences in species cover; this is not unexpected because OptimClass uses a presence/absence‐based fidelity measure. Conclusions: If the aim of a classification is to obtain clusters rich in faithful species, which can be subsequently used as diagnostic species for identification of community types, OptimClass is a suitable method for simultaneous choice of the optimal classification algorithm and optimal number of clusters. It can be computed in the JUICE program.     

Character selection in relation to the numerical taxonomy of some male Diaspididae (Homoptera: Coccoidea)

A quantitative comparison of the classifications of 24 species of male Diaspididae by principal component and principal coordinate analysis was extended to allow the comparison of classifications based on selected subsets of the 101 available characters. It was shown that subsets of 25–28 characters could be chosen numerically and used to construct classifications that resembled very closely the classification based on the reference set of characters. The conclusions are discussed in relation to other views on character-weighting and in relation to classifications of the Diaspididae based on traditional taxonomic characters provided by the females.     

The influence of seasonal and taxonomic factors on the ordination and classification of running-water sites in Great Britain and on the prediction of their macro-invertebrate communities

SUMMARY. 1. Macro-invertebrate samples were collected from 268 running-water sites in Great Britain in each of three seasons (spring, summer and autumn). A combined seasons’treatment was generated by amalgamating the individual seasons’data. These four seasonal options were each subjected to four distinct taxonomic analyses differing in level of identification and whether the data were quantitative or qualitative. Thus sixteen data-sets were available for analysis. Environmental data on physical and chemical variables, macrophyte cover and date of sampling were also recorded for each site. 2. All sixteen data-sets were ordinated by detrended correspondence analysis and classified by two-way indicator species analysis. There were strong correlations between the sixteen ordinations and significant concordance between classifications. 3. The relationships between ordination scores and single environmental variables were investigated. Muhiple discriminant analysis was used to fit environmental data to eight selected classifications covering the full range of seasonal and taxonomic treatments. The environmental variables most useful in distinguishing between rivers were substratum characteristics, alkalinity and total oxidized nitrogen. Within-river differences were often highly correlated with discharge, distance from source, width and depth. Slope and altitude contributed strongly to both between-river and within-river distinctions. 4. Between-site variation (beta diversity), eigenvalues of ordination, the reliability of classifications, the proportion of sites correctly assigned to their biological group using environmental data and the standardized similarity between observed and predicted fauna were all higher when identifications were taken to species level, rather than one of three family treatments. Qualitative data on a reduced list of families gave comparable or better results than more detailed family treatments. 5. Combined seasons’data enabled better categorization and prediction than single season's. 6. The values of the Czekanowski Index of Similarity between the observed and predicted fauna of test sites were close to realistic maximum values. 7. Recommendations are made concerning potential usages of the various classifications. The species level classification has uses in the field of conservation and in the prediction of biological response to environmental change. The family level classifications have value in developing local site inventories and in the interpretation of pollution surveillance programmes.     

Using repeated measurements to validate hierarchical gene clusters

MOTIVATION: Hierarchical clustering is a common approach to study protein and gene expression data. This unsupervised technique is used to find clusters of genes or proteins which are expressed in a coordinated manner across a set of conditions. Because of both the biological and technical variability, experimental repetitions are generally performed. In this work, we propose an approach to evaluate the stability of clusters derived from hierarchical clustering by taking repeated measurements into account. RESULTS: The method is based on the bootstrap technique that is used to obtain pseudo-hierarchies of genes from resampled datasets. Based on a fast dynamic programming algorithm, we compare the original hierarchy to the pseudo-hierarchies and assess the stability of the original gene clusters. Then a shuffling procedure can be used to assess the significance of the cluster stabilities. Our approach is illustrated on simulated data and on two microarray datasets. Compared to the standard hierarchical clustering methodology, it allows to point out the dubious and stable clusters, and thus avoids misleading interpretations. AVAILABILITY: The programs were developed in C and R languages.     

Improving the assessment of species compositional dissimilarity in a priori ecological classifications: evaluating map scale,sampling intensity and improvement in a hierarchical classification

Question: Can species compositional dissimilarity analyses be used to assess and improve the representation of biodiversity patterns in a priori ecological classifications? Location: The case study examined the northern‐half of the South‐east Queensland Bioregion, eastern Australia. Methods: Site‐based floristic presence–absence data were used to construct species dissimilarity matrices (Kulczynski metric) for three levels of Queensland's bioregional hierarchy – subregions (1:500 000 scale), land zones (1:250 000 scale) and regional ecosystems (1:100 000 scale). Within‐ and between‐class dissimilarities were compiled for each level to elucidate species compositional patterns. Randomized subsampling was used to determine the minimum site sampling intensity for each hierarchy level, and the effects of lumping and splitting illustrated for several classes. Results: Consistent dissimilarity estimates were obtained with five or more sites per regional ecosystem, 10 or more sites per land zone, and more than 15 sites per subregion. On average, subregions represented 4% dissimilarity in floristic composition, land zones approximately 10%, and regional ecosystems over 19%. Splitting classes with a low dissimilarity increased dissimilarity levels closer to average, while merging ecologically similar classes with high dissimilarities reduced dissimilarity levels closer to average levels. Conclusions: This approach demonstrates a robust and repeatable means of analysing species compositional dissimilarity, determining site sampling requirements for classifications and guiding decisions about ‘lumping’ or ‘splitting’ of classes. This will allow more informed decisions on selecting and improving classifications and map scales in an ecologically and statistically robust manner.     

A preliminary classification of running-water sites in Great Britain based on macro-invertebrate species and the prediction of community type using environmental data

SUMMARY. 1. Macro-invertebrate species lists were obtained for 268 sites on forty-one river systems throughout Great Britain by qualitative sampling in spring, summer and autumn. Information on twenty-eight environmental variables was also collated for each site. The sites were ordinated on the basis of their species content using detrended correspondence analysis (DCA) and classified by two-way indicator species analysis (TWINSPAN). Correlation coefficients between ordination scores and single environmental variables indicated that Axis 1 distinguished between types of rivers and Axis 2 reflected variation along the length of rivers. A preliminary classification of sites into sixteen groups has been proposed, together with a key which allows new sites to be classified. Information on the species and environmental features which characterize each group is also presented.
2. Multiple discriminant analysis (MDA) was employed to predict the group membership of the 268 sites using the twenty-eight environmental variables. 76.1% of sites were classified correctly. An independent assessment of predictive ability using forty test sites yielded a 50% success rate. Predictive ability was higher for the classification presented in this paper than in fifteen additional classifications produced using data from single seasons and/or different taxonomic treatments.
3. TWINSPAN and MDA were found to be useful approaches to the classification of running-water sites by their macro-invertebrate fauna and the prediction of community type (as indicated by the occurrence of species in the sites comprising the group) using environmental variables. Extension of the scope of the classification, coupled with the use of additional environmental variables to increase predictive ability, is now in progress.     

Heterogeneity‐constrained random resampling of phytosociological databases

Aim: Phytosociological databases often contain unbalanced samples of real vegetation, which should be carefully resampled before any analyses. We propose a new resampling method based on species composition, called heterogeneity‐constrained random (HCR) resampling. Method: Many subsets of the source vegetation database are selected randomly. These subsets are sorted by decreasing mean dissimilarity between pairs of the vegetation plots, and then sorted again by increasing variance of these dissimilarities. Ranks from both sortings are summed for each subset, and the subset with the lowest summed rank is considered as the most representative. The performance of this method was tested using simulated point patterns that represented different levels of aggregation of vegetation plots within a database. The distributions of points in the subsets resulting from different resampling methods, both with and without database stratification, were compared using Ripley's K function. The mean of random selections from an unbiased sample was used as a reference in these comparisons. The efficiency of the method was also demonstrated with real phytosociological data. Results: Both stratified and HCR resampling yielded selection patterns more similar to the reference than resampling without these tools. Outcomes from the resampling that combined these two methods were the most similar to the reference. The efficiency of the HCR resampling method varied with different levels of aggregation in the database. Conclusions: This new method is efficient for resampling phytosociological databases. As it only uses information on species occurrences/abundances, it does not require the definition of strata, thereby avoiding the effect of subjective decisions on the selection outcome. Nevertheless, this method can also be applied to stratified databases.     

A comparison of DNA‐based methods for delimiting species in a Cretan land snail radiation reveals shortcomings of exclusively molecular taxonomy

We compared the results of different approaches for delimiting species based on single‐locus DNA sequences with those of methods using binary multilocus data. As case study, we examined the radiation of the land snail genus Xerocrassa on Crete. Many of the methods based on mitochondrial sequences resulted in heavy under‐ or overestimations of the species number. The methods using AFLP data produced classifications with an on average higher concordance with the morphological classification than the methods based on mitochondrial sequences. However, the percentage of correct species classifications is low even with binary multilocus data. Gaussian clustering produced the classifications with the highest concordance with the morphological classification of all approaches applied in this study, both with single‐locus sequences and with binary multilocus data. There are two general problems that hamper species delimitation, namely rarity and the hierarchical structure of biodiversity. Methods for species delimitation using genetic data search for clusters of individuals, but do not implement criteria that are sufficient to distinguish clusters representing species from other clusters. The success of morphological species delimitation results from the potential to focus on characters that are directly involved in the speciation process, whereas molecular studies usually rely on markers that are not directly involved in speciation. © The Willi Hennig Society 2011.     

Comparing agglomerative clustering and three weed classification frameworks to assess the invasiveness of alien species across spatial scales

To prioritize weed management at the catchment scale, information is required on the species present, their relatively frequency, abundance, and likely spread and impact. The objective of this study was to classify the invasiveness of alien species that have invaded the Upper Burdekin Catchment in Queensland, Australia, at three spatial scales. A combination of three published weed classification frameworks and multivariate techniques were employed to classify species based on their frequency and cover at a range of spatial scales. We surveyed the Upper Burdekin Catchment for alien species, and for each species determined the following distribution indices — site frequency, total cover, transect frequency per site frequency and quadrat frequency per site frequency, cover per quadrat when present, cover per transect when present, and cover per site when present. These indices capture the effect of species abundance and frequency between sites (site frequency and total cover), within sites (transect frequency per site and cover per transect when present), and within transects (quadrat frequency per site frequency and cover per site). They were used to classify the species into seven groups using a hierarchical cluster analysis. The relationship between the indices was explored to determine how effective the small scale, site‐specific indices were at predicting the broader, landscape‐scale patterns. Strong correlations were observed between transect frequency per site and frequency (r² = 0.89) and cover per transect when present and total cover (r² = 0.62). This suggests that if a weed is abundant at the site level, it has the potential to occupy large areas of the catchment. The species groupings derived from the application of the three published weed classification frameworks were compared graphically to the groupings derived from the cluster analysis. One of the frameworks classified species into three groups. The other two frameworks classified species into four groups. There was a high degree of subjectivity in applying the frameworks to the survey data. Some of the data were of no relevance to the classification frameworks and were therefore ignored. We suggest that the weed classification frameworks should be used in conjunction with existing multivariate techniques to ensure that classifications capture important natural variations in observed data that may reflect invasion processes. The combined use of the frameworks and multivariate techniques enabled us to aggregate species into categories appropriate for management.     

Semi‐dry grasslands along a climatic gradient across Central Europe: Vegetation classification with validation

Question: What is the variation in species composition of Central European semi‐dry grasslands? Can we apply a training‐and‐test validation approach for identifying phytosociological associations which are floristically well defined in a broad geographic comparison; can we separate them from earlier described associations with only a local validity? Location: A 1200 km long transect running along a gradient of increasing continentality from central Germany via Czech Republic, Slovakia, NE Austria, Hungary to NW Romania. Methods: Relevés with > 25% cover of Brachypodium pin‐natum and/or Bromus erectus were geographically selected from a larger database. They were randomly split into two data sets, TRAINING and TEST, each with 422 relevés. Cluster analysis was performed for each data set on scores from significant principal coordinates. Different partitions of the TRAINING data set were validated on the TEST data set, using a new method based on the comparison of % frequencies of species occurrence in clusters. Clusters were characterized by statistically defined groups of diagnostic species and values of climatic variables. Results: Species composition changed along the NW‐SE gradient and valid clusters were geographically well separated. Optimal partition level was at 11 clusters, six being valid: two clusters Germany and the Czech Republic corresponded to the Bromion erecti; two clusters from the Czech Republic and Hungary to the Cirsio‐Brachypodion, and two clusters were transitional between these two alliances. Conclusion: The training‐and‐test validation method used in this paper proved to be efficient for discriminating between robust clusters, which are appropriate candidates for inclusion in the national or regional syntaxonomic overviews, and weak clusters, which are specific to the particular classification of the given data set.     

Object-based classification of land cover and tree species by integrating airborne LiDAR and high spatial resolution imagery data

We evaluated the effectiveness of integrating discrete return light detection and ranging (LiDAR) data with high spatial resolution near-infrared digital imagery for object-based classification of land cover types and dominant tree species. In particular we adopted LiDAR ratio features based on pulse attributes that have not been used in past studies. Object-based classifications were performed first on land cover types, and subsequently on dominant tree species within the area classified as trees. In each classification stage, two different data combinations were examined: LiDAR data integrated with digital imagery or digital imagery only. We created basic image objects and calculated a number of spectral, textural, and LiDAR-based features for each image object. Decision tree analysis was performed and important features were investigated in each classification. In the land cover classification, the overall accuracy was improved to 0.975 when using the object-based method and integrating LiDAR data. The mean height value derived from the LiDAR data was effective in separating “trees” and “lawn” objects having different height. As for the tree species classification, the overall accuracy was also improved by object-based classification with LiDAR data although it remained up to 0.484 because spectral and textural signatures were similar among tree species. We revealed that the LiDAR ratio features associated with laser penetration proportion were important in the object-based classification as they can distinguish tree species having different canopy density. We concluded that integrating LiDAR data was effective in the object-based classifications of land cover and dominant tree species.     

Distribution patterns of boreal marshland birds: modelling the relationships to land cover and climate

Aim The aims of this work were (1) to study how well land‐cover and climatic data are capable of explaining distribution patterns of ten bird species breeding and/or feeding primarily on marshes and other wetlands and (2) to compare the differences between red‐listed and common marshland species in explanatory variables, and to study the predictability of their distribution patterns. Location Finland, northern Europe. Methods The data of the bird atlas survey carried out in 1986–89 using a 10 × 10 km uniform grid system in Finland were used in the analyses. Land‐cover data based on CORINE (Coordination of Information on the Environment) classification and climatic variables were compiled using the same 10 × 10 km grid. Generalized additive models (GAM) with a stepwise selection procedure were used to select relevant explanatory variables and to examine the complexity of the response shapes of the different species to each variable. The original data set was randomly divided into model training (70%) and model evaluation (30%) sets. The final models of common and red‐listed bird species richness were validated by fitting them to the model evaluation set, and the correlation between observed and predicted species richness was calculated. We assessed the discrimination ability of the binary models (single species) with the area under the curve (AUC) of a receiver operating characteristic (ROC) plot and the Kappa coefficient. Results Cover of marshland, shoreline length and mean temperature in April–June were significantly (P < 0.01) related to the common marshland species richness. Cover and clumping of marshland and mean temperature and precipitation in April–June were selected in the model of red‐listed marshland species richness. The level of discrimination in our single species models varied in ROC from fair to excellent (AUC values 0.70–0.95). Cover of marshland was included in all GAM models built for the target species, but clumping of marshland, shoreline length and cover of mires also appeared as important predictors in single species models. Seven species had statistically significant relationships with climatic variables in the multivariate GAMs. Cover of marshland was highest in squares in which the red‐listed bittern Botaurus stellaris, marsh harrier Circus aeruginosus and great reed warbler Acrocephalus arundinaceus and the water rail Rallus aquaticus were observed. Main conclusions Cover of marshland was the only variable which was included in all the models, reinforcing the close connection between the studied species and marshlands. Broad‐scale clumping of marshlands was important for the red‐listed species, probably due to the much lower population sizes of red‐listed species than those of common species. Land‐cover data produced in CORINE seems to be well suited for modelling the distribution patterns of marshland birds. Although climatic variables also strongly affect the studied marshland birds, habitat availability plays a crucial role in their occurrence. The distribution patterns of marshland birds at the scale of 10 × 10 km reflect the interplay between habitat availability and direct climatic variables.     

Methodologically induced differences in oak site classifications in a homogeneous tree-ring network

Spatiotemporal variations in tree growth are induced by varying environmental conditions. Different methods like variants of the principal component analysis and the hierarchical cluster analysis are commonly applied in dendroecology to separate subsets of growth patterns within large tree-ring datasets. To seek for methodological differences in classification techniques and their specific characteristics, we compared three standard methods using a homogeneous oak (Quercus spp.) network from temperate forests in Central-West Germany. Classifications of the original dataset consisting of 46 oak ring-width sites, carried out with the varimax rotated principle component analysis, Ward's method and the average linkage method, reveal differences in the classification of approximately 20% of the sites. Analyses with modified datasets are calculated to evaluate effects of dataset extension, different time periods and different tree-ring detrendings. The application of the principal component analysis generally leads to the most stable site classifications, whereas the most sensitive response to changes in the dataset is obtained by Ward's method. The average linkage method separates single sites in the classification and thus emphasises outliers within the tree-ring network.