Phytosociological data give biased estimates of species richness

Abstract. Large phytosociological data sets of three types of grassland and three types of forest vegetation from the Czech Republic were analysed with a focus on plot size used in phytosociological sampling and on the species‐area relationship. The data sets included 12975 relevés, sampled by different authors in different parts of the country between 1922 and 1999. It was shown that in the grassland data sets, the relevés sampled before the 1960s tended to have a larger plot size than the relevés made later on. No temporal variation in plot sizes used was detected in forest relevés. Species‐area curves fitted to the data showed unnatural shapes, with levelling‐off or even decrease in plot sizes higher than average. This distortion is explained by the subjective, preferential method of field sampling used in phytosociology. When making relevés in species‐poor vegetation, researchers probably tend to use larger plots in order to include more species. The reason for this may be that a higher number of species gives a higher probability of including presumed diagnostic species, so that the relevé can be more easily classified in the Braun‐Blanquet classification system. This attitude of phytosociologists has at least two consequences: (1) in phytosociological data bases species‐poor vegetation types are underrepresented or relevés are artificially biased towards higher species richness; (2) the suitability of phytosociological data for species richness estimation is severely limited.     

Plot sizes used for phytosociological sampling of European vegetation

Abstract. In European phytosociology, variable plot sizes are traditionally used for sampling different vegetation types. This practice may generate problems in current vegetation or habitat survey projects based on large data sets, which include relevés made by many authors at different times. In order to determine the extent of variation in plot sizes used in European phytosociology, we collected a data set of 41 174 relevés with an indication of plot size, published in six major European journals focusing on phytosociology from 1970 to 2000. As an additional data set, we took 27 365 relevés from the Czech National Phytosociological Database. From each data set, we calculated basic statistical figures for plot sizes used to sample vegetation of various phytosociological classes. The results show that in Europe the traditionally used size of vegetation plots is roughly proportional to vegetation height; however, there is a large variation in plot size, both within and among vegetation classes. The effect of variable plot sizes on vegetation analysis and classification is not sufficiently known, but use of standardized plot sizes would be desirable in future projects of vegetation or habitat survey. Based on our analysis, we suggest four plot sizes as possible standards. They are 4 m² for sampling aquatic vegetation and low‐grown herbaceous vegetation, 16 m² for most grassland, heathland and other herbaceous or low‐scrub vegetation types, 50 m² for scrub, and 200 m² for woodlands. It has been pointed out that in some situations, sampling in either small or large plots may result in assignment of relevés to different phytosociological classes or habitat types. Therefore defining vegetation and habitat types as scale‐dependent concepts is needed.     

Sampling design in large-scale vegetation studies: Do not sacrifice ecological thinking to statistical purism!

Most of the historical phytosociological data on vegetation composition have been sampled preferentially and thus belong to those ecological data that do not fulfill the statistical assumption of independence of observations, necessary for valid statistical testing and inference. Nevertheless, phytosociological data have been recently used for various ecological meta-analyses, especially in studies of large-scale vegetation patterns. For this reason, we focus on the comparison of preferential sampling with other sampling designs that have been recommended as more convenient alternatives from the point of view of statistical theory. We discuss that while simple random sampling, systematic sampling and stratified random sampling better meet some of the statistical assumptions, preferential sampling yields data sets that cover a broader range of vegetation variability. Moreover, today’s large phytosociological databases provide huge amounts of vegetation data with unrivalled geographic extent and density. We conclude that in the near future ecologists will not be able to replace the preferentially sampled phytosociological data in large-scale studies. At the same time, phytosociological databases have to be complemented with relevés of vegetation composed mostly of common and generalist species, which are under-represented in historical data. Stratified random sampling seems to be a suitable tool for doing this. Nevertheless, a methodology and input data for stratification have to be developed to make stratified random sampling an ecologically more relevant and practical method.     

The influence of sample plot size on evaluations with Ellenberg indicator values

The effect of plot size was tested on heterogeneous and homogeneous data sets that were obtained by sampling grassland and forest vegetation on plots differing in size. Mean EIV for relevés revealed no differences among data sets from various plot sizes or between homogeneous and heterogeneous data sets. This is probably due to a similar indicator value for species newly occurring in plots with increasing plot size. Using EIV is thus a robust method even for data sets associated with wide range of plot sizes.     

Random sampling does not exclude spatial dependence: The importance of neutral models for ecological hypothesis testing

Lájer (2007) notes that, to investigate phytosociological and ecological relationships, many authors apply traditional inferential tests to sets of relevés obtained by non-random methods. Unfortunately, this procedure does not provide reliable support for hypothesis testing because non-random sampling violates the assumptions of independence required by many parametric inferential tests. Instead, a random sampling scheme is recommended. Nonetheless, random sampling will not eliminate spatial autocorrelation. For instance, a classical law of geography holds that everything in a piece of (biotic) space is interrelated, but near objects are more related than distant ones. Because most ecological processes that shape community structure and species coexistence are spatially explicit, spatial autocorrelation is a vital part of almost all ecological data. This means that, independently from the underlying sampling design, ecological data are generally spatially autocorrelated, violating the assumption of independence that is generally required by traditional inferential tests. To overcome this drawback, randomization tests may be used. Such tests evaluate statistical significance based on empirical distributions generated from the sample and do not necessarily require data independence. However, as concerns hypothesis testing, randomization tests are not the universal remedy for ecologists, because the choice of inadequate null models can have significant effects on the ecological hypotheses tested. In this paper, I emphasize the need of developing null models for which the statistical assumptions match the underlying biological mechanisms.     

How reliable are our vegetation analyses?

Abstract. Two sets of 40 relevés, made independently by two observers on the same 5m x 5m sample plots, were compared to estimate the sampling error and to assess the effect of this sampling error on (1) estimates of species richness and diversity (2) results of multivariate analyses, and (3) estimation of species turnover in repeated sampling. The relevés were made according to the standard Braun-Blanquet method. The sampling error was estimated for (1) recording of species in sample plots and (2) visual estimation of the degree of cover (or of the general population size). Despite the fact that the sample plots were searched thoroughly for 30 - 40 min, the number of overlooked species was high with a discrepancy of 13% between corresponding relevés. Regarding multivariate analysis, the error caused by missing species was at least as important as the error in visual estimation of species cover. The estimates of degree of cover using the Braun-Blanquet scale are sufficiently reliable for use in multivariate analysis when they are subjected to ordinal transformation. When average cover values are used, the patterns detected are based solely on dominants. Species richness and species diversity could be reliably estimated from the relevés, but the estimates of equitability are very unreliable. The classical relevé method remains one of the most efficient survey methods for recognition of vegetation types on the macro-community and landscape scales.     

Life-history traits and habitat preferences of colonizing plant species in long-term spontaneous succession in abandoned gravel–sand pits

The following question was asked: Species of what traits and ecological demands are the best and the least successful colonizers in the course of succession in disused gravel–sand pits?Abandoned gravel–sand pits (36) were studied in the Czech Republic. Together 224 phytosociological relevés were recorded in 5 m×5 m plots in all available seral stages ranging in age from 1 to 75 years. Complete lists of species occurring in (semi-)natural habitats were surveyed up to a distance of 100 m from each relevé. The colonization success of each species was expressed by an index between 0 and 1 which was obtained as the ratio of the number of relevés with species present in a pit/the number of relevés with the species occurrence in their surroundings. Species characteristics were taken from available databases. Data were elaborated by ordination analysis and regression tree analysis. Generally, the best colonizers of gravel–sand pits were hydrophytes with the ability to vegetatively reproduce. In particular, at the beginning of succession, anemochorous species played the most important role. Stress tolerant species, possessing light diaspores, were typical of nitrogen poor and acidic habitats. Later, sciophilous and nitrophilous species of mesic habitats increased in importance. They have heavier diaspores, and are mostly phanerophytes or geophytes, often with the ability to vegetatively reproduce as well as exhibiting higher demand for pH. Plant functional traits were recognized as powerful tools to predict the colonization success of plants available in the local species pool. They may help predict vegetation succession in various human-disturbed sites and thus be used in various restoration programmes.     

A simple method for estimating homotoneity of sets of phytosociological relevés

The homogeneity of sets of phytosociological relevés or syntaxa—the homotoneity—is tested by means of a homotoneity-coefficient. This coefficient consists of the basic homotoneity-coefficient expressing the proportion of highly constant species (61–100%) in the species composition of the average relevé and the correcting factor which is proportional to the difference of species number between the richest and poorest relevé. Successive analysis of heterotoneity of sets of relevés for detecting sources of heterotoneity is proposed.     

Species richness in deciduous forests: Effects of species pools and environmental variables

Abstract. The study was conducted in deciduous forests of two Swedish regions, Öland and Uppland. It had two objectives: to (1) test the species pool hypothesis by examining if differences in small‐scale species richness are related to differences in large‐scale species richness and the size of the regional species pool, and (2) to examine the relationship between species richness and productivity and its scale‐dependence. The first data set comprised 36 sites of moderate to high productivity. In each site, we recorded the presence of vascular plant species in nested plots ranging from 0.001 to 1000 m² and measured several environmental variables. Soil pH and Ellenberg site indicator scores for nitrogen were used as estimators of productivity. The second data set included 24 transects (each with 20 1‐m² plots) on Öland in sites with low to high productivity. Species number, soil pH and relative light intensity were determined in each plot. The forest sites on Öland were more species‐rich than the Uppland sites on all spatial scales, although environmental conditions were similar. Small‐scale and large‐scale species richness were positively correlated. The results present evidence in favour of the species pool hypothesis. In the nested‐plots data set, species number was negatively correlated with pH and nitrogen indicator scores, whereas a unimodal relationship between species number and pH was found for the transect data set. These results, as well as previously published data, support the hump‐shaped relationship between species richness and productivity in Swedish deciduous forests. Two explanations for the higher species richness of the sites with moderate productivity are given: first, these sites have a higher environmental heterogeneity and second, they have a larger ‘habitat‐specific’ species pool.     

Context-dependence of diagnostic species: A case study of the central european spruce forests

In the phytosociological literature, there are numerous different approaches to the designation of diagnostic species. Frequently, this results in discrepancies between the lists of diagnostic species published for one and the same community. We examined different approaches to determining diagnostic species using as an examplePicea abies forests within the broader context of all Central European forests. Diagnostic species of spruce forests were determined from a data set of 20,164 phytosociological relevés of forests from the Eastern Alps, Western Carpathians, and the Bohemian Massif, which included 3,569 relevés of spruce forests. Phi coefficient of association was used to measure species fidelity, and species with the highest fidelities were considered as diagnostic. Diagnostic species were determined in four ways, including (A) comparison of spruce forests among the three mountain ranges, (B) comparison between spruce forests and other forests, performed separately in each of the mountain ranges, (C) simultaneous comparison of spruce forests of each of the mountain ranges with spruce forests of the other two ranges and with the other forests of all ranges, (D) comparison of spruce forests with the other forests, using pooled data sets from the three mountain ranges. The sets of diagnostic species of spruce forests yielded in comparisons A and B were sharply different; the set resulting from comparison C was intermediate between the first two and comparison D resulted in similar diagnostic species as comparison B. In comparison A, spruce forests of the Eastern Alps had a number of diagnostic species, while the spruce forests of the other two mountain ranges had only few diagnostic species. In comparison B, by contrast, the number and quality of diagnostic species decreased from the Bohemian Massif to the Eastern Alps. This exercise points out that lists of diagnostic species published in phytosociological literature are dependent on the context, i.e. the underlying data sets and comparisons: some of these lists are useful for identification of vegetation units at a local scale, some others for distinguishing units within a narrowly delimited community type over a large area. The thoughtless application of published lists of diagnostic species outside of the context for which they were intended should therefore be avoided.     

Vegetation and environmental factors during primary succession on glacier forelands: Some outlines from the Italian Alps

ABSTRACT

Relationships between plant communities and the physical environment during primary succession on recently deglaciated glacier forelands were studied in 3 areas of the Italian Alps. The aim of the research was to relate traditional phytosociological data with environmental variables. Twenty-eight phytosociological relevés were performed, each associated with twenty-six environmental variables; quantitative parameters of richness and diversity were also calculated. Species/relevés, environmental variables/relevés and species/environmental variables matrices were analyzed by cluster analysis, PCA and Spearman correlation coefficient. Three main stages of succession were identified by floristic composition and confirmed by environmental parameter evaluation. A complex of environmental variables seems to be closely correlated with terrain age and richness/diversity parameters, even though diversity decreases in late successional stages. The phytosociological significance of species is in accordance with their position in the context of succession.     

Use of Phytosociological Databases for Species Distribution Models

Species Distribution Models are key in modern ecological studies. They employ information about species locations and environmental factors to generate statistical functions that predict the potential distribution of species on the basis of landscape suitability. Although these models are powerful and useful tools, often the required information about species distribution is lacking, and the only resources are pre-collected museum data. Phytosociological databases contain a myriad of relevés with precious information, but are often considered to be the exclusive ownership of vegetation scientists. Our study tested the efficiency of a phytosociological database in the building of Species Distribution Models, including spatial autocorrelation (SAC) as a predictor to evaluate its effects on model performance. Spatial autocorrelation (SAC) is a natural characteristic of species distribution that depends on exogenous and endogenous processes. The latter??s effects could be overestimated by a subjective sample choice. We chose Festuca riccerii, an Italian endemic species. We split the whole dataset (671 relevés) into a calibration (443 relevés) and testing set (228 relevés) and performed a GLM on these data to identify the main ecological factors that lead distribution in order to build a Species Distribution Model. The dataset??s efficiency was assessed by testing the predicting power of the calibrated model on the testing subset. The phytosociological database proved to be good for building model (AUC?=?0.821), providing a useful basis for fast and low cost ecological analysis, and could be used subsequently for more detailed analyses.     

A probabilistic approach to estimating species pools from large compositional matrices

Abstract. Species pools are increasingly recognized as important controls of local plant community structure and diversity. While existing approaches to estimate their content and size either rely on phytosociological expert knowledge or on simple response models across environmental gradients, the proposed application of phytosociological smoothing according to Beals exploits the full information of plant co‐occurrence patterns statistically. Where numerous representative compositional data are available, the new method yields robust estimates of the potential of sites to harbour plant species. To test the new method, a large phytosociological databank covering the forested regions of Oregon (US) was subsampled randomly and evenly across strata defined by geographic regions and elevation belts. The resulting matrix of species presence/absence in 874 plots was smoothed by calculating Beals' index of sociological favourability, which estimates the probability of encountering each species at each site from the actual plot composition and the pattern of species co‐occurrence in the matrix. In a second step, the resulting lists of sociologically probable species were intersected with complete species lists for each of 14 geographical subregions. Species pools were compared to observed species composition and richness. Species pool size exhibited much clearer spatial trends than plot richness and could be modelled much better as a function of climatic factors. In this framework the goal of modelling species pools is not to test a hypothesis, but to bridge the gap between manageable scales of empirical observation and the spatio‐temporal hierarchy of diversity patterns.     

Maximizing plant species inventory efficiency by means of remotely sensed spectral distances

Aim Inventorying plant species in an area based on randomly placed quadrats can be quite inefficient. The aim of this paper is to test whether plant species richness can be inventoried more efficiently by means of a spectrally‐based ordering of sites to be sampled. Location The study area was a complex wetland ecosystem, the Lake Montepulciano Nature Reserve, central Italy. This is one of the most important wetland areas of central Italy because of the diverse plant communities and the seasonal avifauna. Methods Field sampling, based on a random stratified sampling design, was performed in June 2002. Plant species composition was recorded within sampling units of 100 m² (plots) and 1 ha (macroplots). A QuickBird multispectral image of the same date was acquired and corrected both geometrically and radiometrically. Species accumulation curves based on spectral information were obtained by ordering sites to be sampled according to a maximum spectral distance criterion (i.e. by ordering sampling units based on the maximum distances among them in a four‐dimensional spectral space derived from the remotely sensed data). Different distance measures based on mean and maximum spectral distances among sampling units were tested. The performance of the species accumulation curve derived by the spectrally‐based ordering of sampling units was tested against a rarefaction curve obtained from the mean of 10,000 accumulation curves based on randomly ordered sampling units. Results The spectrally‐derived curve based on the maximum spectral distance among sampling units showed the most rapid accumulation of species, well above the rarefaction curve, at both the plot and the macroplot scales. Other ordering criteria of sampling units captured less richness over most of the species accumulation curves at both the spatial scales. The accumulation curves based on other measurements of distance were much closer to the random curve and did not show differences with respect to the species rarefaction curve based on random ordering of sampling units. Main conclusions The present investigation demonstrated that spectral‐based ordering of sites to be sampled can lead to the maximization of the efficiency of plant species inventories, an activity usually driven by the ‘botanist's internal algorithm’ (intuition), without any formalized rule to drive field sampling. The proposed approach can reduce costs of plant species inventorying through a more efficient allotment of time and sampling.     

Effects of plot size on the ordination of vegetation samples

Questions: Do ordination patterns differ when based on vegetation samples recorded in plots of different size? If so, how large is the effect of plot size relative to the effects of data set heterogeneity and of using presence/absence or cover‐abundance data? Can we combine plots of different size in a single ordination? Methods: Two homogeneous and two heterogeneous data sets were sampled in Czech forests and grasslands. Cover‐abundances of plant species were recorded in series of five or six nested quadrats of increasing size (forest 49‐961 m2; grassland 1‐49 m²). Separate ordinations were computed for plots of each size for each data set, using either species presences/absences or cover‐abundances recorded on an ordinal scale. Ordination patterns were compared with Procrustean analysis. Also, ordinations of data sets jointly containing plots of different size were calculated; effects of plot size were evaluated using a Monte Carlo test in constrained ordination. Results: The results were consistent between forest and grassland data sets. In homogeneous data sets, the effect of presence/absence vs. cover‐abundance was similar to, or larger than, the effect of plot size; for presence/absence data the differences between ordinations of differently sized plots were smaller than for cover‐abundance data. In heterogeneous data sets, the effect of plot size was larger than the effect of presence‐absence vs. cover‐abundance. The plots of smaller size (= 100 m2 in forests, = 4 m² in grasslands) yielded the most deviating ordination patterns. Joint ordinations of differently sized plots mostly did not yield patterns that would be artifacts of different plot size, except for plots from the homogeneous data sets that differed in size by a factor of four or higher. Conclusions: Variation in plot size does influence ordination patterns. Smaller plots tend to produce less stable ordination patterns, especially in data sets with low ß‐diversity and species cover‐abundances. Data sets containing samples from plots of different sizes can be used for ordination if they represent vegetation with large ß‐diversity. However, if data sets are homogeneous, i.e. with low ß‐diversity, the differences in plot sizes should not be very large, in order to avoid the danger of plot size differences distorting the real vegetation differentiation in ordination patterns.     

Computer assisted floristic similarity analysis

A program for polythetic numerical evaluation of phytosociological material is described. Using Sörensen's coefficient of floristic similarity it computes the homogeneity of subjectively chosen sets of relevés, the affinity of each constituent relevé to the set as well as the similarity between any pair of sets. It is also able to plot a dendrogram of the hierarchic system obtained by agglomerating individual relevés into complex groupings. The program is suitable for processing phytosociological data arranged in any conventional table.     

Constant tree species richness along an elevational gradient of Mt. Bokor,a table-shaped mountain in southwestern Cambodia

Some previous studies along an elevational gradient on a tropical mountain documented that plant species richness decreases with increasing elevation. However, most of studies did not attempt to standardize the amount of sampling effort. In this paper, we employed a standardized sampling effort to study tree species richness along an elevational gradient on Mt. Bokor, a table-shaped mountain in southwestern Cambodia, and examined relationships between tree species richness and environmental factors. We used two methods to record tree species richness: first, we recorded trees taller than 4 m in 20 uniform plots (5 × 100 m) placed at 266–1048-m elevation; and second, we collected specimens along an elevational gradient from 200 to 1048 m. For both datasets, we applied rarefaction and a Chao1 estimator to standardize the sampling efforts. A generalized linear model (GLM) was used to test the relationship of species richness with elevation. We recorded 308 tree species from 20 plots and 389 tree species from the general collections. Species richness observed in 20 plots had a weak but non-significant correlation with elevation. Species richness estimated by rarefaction or Chao1 from both data sets also showed no significant correlations with elevation. Unlike many previous studies, tree species richness was nearly constant along the elevational gradient of Mt. Bokor where temperature and precipitation are expected to vary. We suggest that the table-shaped landscape of Mt. Bokor, where elevational interval areas do not significantly change between 200 and 900 m, may be a determinant of this constant species richness.     

A flexible multi-scale approach for standardised recording of plant species richness patterns

A sound monitoring of appropriate biodiversity indicators is necessary in order to assess the progress towards the internationally agreed target of halting the loss of biodiversity by 2010. However, existing monitoring schemes often do not address species richness as a key component of biodiversity directly or do so with insufficient methods. I provide an overview and assessment of the large variety of different sampling approaches for small-scale plant species richness. Major shortcomings of many of these are (i) non-uniform plot sizes or shapes; (ii) analysis of only one spatial scale despite the scale dependence of nearly all biodiversity parameters; (iii) lack of replication of smaller subplots; and (iv) exclusion of bryophytes and lichens despite their often large contribution to total plant diversity. Based on this review, I propose a new standardised sampling approach for plant diversity patterns at small scales that is applicable for a multitude of purposes and in any biome. In its basic variant, species composition is recorded on nested squares of 0.01 m², 0.1 m², 1 m², 10 m², and 100 m², with all smaller subplots being replicated at least 3-fold and evenly spaced within the next larger plot. Not only terricolous vascular plants, but also bryophytes, lichens, macro-algae as well as non-terricolous taxa should be recorded with the any-part system, i.e. those plants are counted within a plot whose superficial parts reach over it. This approach can be used to assess plant diversity patterns (i) of individual plots of interest, (ii) along environmental gradients, (iii) within specific vegetation types, or (iv) for landscape sectors. In the latter case, the series of nested plots must be placed randomly or systematically, but irrespective of plot homogeneity. The proposed approach allows the calculation of many meaningful biodiversity indicators, while being well compatible with a range of other sampling schemes, but avoiding their shortcomings. As this approach is not very time-consuming in its basic variant, but can easily be extended for specific purposes, I suggest its use for any kind of biodiversity studies and particularly for monitoring.     

Determination of diagnostic species with statistical fidelity measures

Abstract. Statistical measures of fidelity, i.e. the concentration of species occurrences in vegetation units, are reviewed and compared. The focus is on measures suitable for categorical data which are based on observed species frequencies within a vegetation unit compared with the frequencies expected under random distribution. Particular attention is paid to Bruelheide's u value. It is shown that its original form, based on binomial distribution, is an asymmetric measure of fidelity of a species to a vegetation unit which tends to assign comparatively high fidelity values to rare species. Here, a hypergeometric form of u is introduced which is a symmetric measure of the joint fidelity of species to a vegetation unit and vice versa. It is also shown that another form of the binomial u value may be defined which measures the asymmetric fidelity of a vegetation unit to a species. These u values are compared with phi coefficient, chi‐square, G statistic and Fisher's exact test. Contrary to the other measures, phi coefficient is independent of the number of relevés in the data set, and like the hypergeometric form of u and the chi‐square it is little affected by the relative size of the vegetation unit. It is therefore particularly useful when comparing species fidelity values among differently sized data sets and vegetation units. However, unlike the other measures it does not measure any statistical significance and may produce unreliable results for small vegetation units and small data sets. The above measures, all based on the comparison of observed/expected frequencies, are compared with the categorical form of the Dufrêne‐Legendre Indicator Value Index, an index strongly underweighting the fidelity of rare species. These fidelity measures are applied to a data set of 15 989 relevés of Czech herbaceous vegetation. In a small subset of this data set which simulates a phytosociological table, we demonstrate that traditional table analysis fails to determine diagnostic species of general validity in different habitats and large areas. On the other hand, we show that fidelity calculations used in conjunction with large data sets can replace expert knowledge in the determination of generally valid diagnostic species. Averaging positive fidelity values for all species within a vegetation unit is a useful approach to measure quality of delimination of the vegetation unit. We propose a new way of ordering species in synoptic species‐by‐relevé tables, using fidelity calculations.     

Grain size affects the relationship between species richness and above-ground biomass in semi-arid rangelands

ABSTRACT

Background: Discrepancies in the shape of the productivity–diversity relationship may arise from differences in spatial scale. We hypothesised that there is a grain size effect on the productivity–diversity relationship.

Aims: To determine the effect of three sampling grain sizes on the productivity–diversity relationship.

Methods: We applied generalised linear mixed effect models on community data from 735 vegetation plots in the Taleghan rangelands, Iran, sampled at three grain sizes (0.25, 1 and 2 m²) to ascertain plant productivity-diversity patterns, while accounting for the effects of site, plant community type, disturbance, and life form.

Results: Overall, relationships between biomass and plant species richness were unimodal at grain sizes of 0.25 and 1 m², and asymptotical at 2 m². The spurious occurrence of a single large shrub may overwhelm a small-sized sampling unit, resulting in a high estimate of the sample’s biomass relative to species richness. However, the relationship between biomass and species richness at larger grain sizes is more likely to reach an asymptote.

Conclusions: Shrubs are partly responsible for driving the relationship between plant biomass and species richness. Given that the frequency of shrubs is highly variable between small plots but not so in large plots, their presence may result in unimodal productivity–diversity relationships at small but not at large grain sizes.