We propose new measures to expand the conceptual and methodological framework of life–environment diversity. Our approach combines the variability in community structure with species’ traits, ecological niches, and genetic distances to measure the overall diversity of life. It also combines, for the first time, the heterogeneity in ecological communities’ structure with that in their environment and their spatial patterns.
Evolutionary and adaptive signals can explain the structure of ecological communities better than the environment and space. Therefore, they provide important insights into community assembly and may reveal the adaptive characteristics developed during the evolution of species13. Thus, although traits and ecological niches are frequently used interchangeably76,77, their separate use as independent sources of information has been advocated before13. Measures of ND were previously shown not to be redundant with measures of TD or FD, but to explain in a complementary manner the responses of communities to environmental gradients13. Our sensitivity analysis revealed that although the shared effect of the three facets of diversity considered as predictors prevailed (with almost half of the explained variation), their unique effects were significant and comparable (Fig. 2), showing that these measures are not redundant in their contribution to ODP’. The same is true when a fourth measure, PD, was included in the ODP’. Although PD and TD typically covary closely, their relationship may vary due to different patterns of speciation and migration78, resulting in a different prioritization of communities for conservation.
FD is increasingly included beside TD in the methodology for environmental monitoring, assessment of human impact, and ecological restoration79,80,81,82. However, the species’ niche characterizes the outcome of their phylogeny, adaptations, and relationships with both biotic and abiotic factors at the species level, revealing their significance at the community level, and it refers to relations between species, environment, and more. These are some of the features that distinguish niches from traits. Therefore, whereas FD is specific to life itself, ND is a measure of life-environment interaction, proving to be more sensitive to human impact than other diversity measures13. This suggests that ND can be a valuable tool in environmental monitoring and ecological management and an indicator when evaluating success in conservation projects13.
Here, we argue that the combination of these measures in an overall diversity measure would provide a more complex and complete approach to biodiversity and aid more effectual conservation. When the product (ODP) should be used instead of the summation (ODS) depends on the question asked, whether the variables are correlated, and the given values. The summation of diversity measures (ODS) may be used when the constituents are independent (uncorrelated), when the higher value in one measure can compensate for a lower value in another, or when a diversity measure is 0. We advocate for using the product (ODP) preferably because, to the best of our knowledge, diversity measures are correlated, they frequently interact, and a higher value of one is counterbalanced (and not compensated) by a possible lower value of another (e.g., a low species richness will not be compensated by a high evenness), while the result is still sensitive to all constituents and their scale. Besides, products might also be preferred for sensitivity analysis, related measures, and quests for simple and conditional effects of predictors.
Detecting the relative contributions of input variables to the uncertainty of model output can be done by a global sensitivity analysis that allows input descriptors to vary within their whole range62. When a model is established, one can assess the effect of input (explanatory) variables on the output (response) variables at a single point. By this, a local sensitivity analysis is performed. The sensitivity analysis, which has numerous theoretical and practical applications, can be done on independent or correlated variables63 and uni- and multivariate data sets61. As a proxy for the usual variance-based models, we used the variation partitioning procedure, which is well-suited for detecting simple and conditional effects. This might be used for synthetic or combined diversity measures, either as response variables or as predictors in a reverse analysis.
Using our empirical data set, we got similar results using different approaches to estimating the mean and variance of ODP’. However, based on the simulation results, the differences between the variances of ODP’ calculated using the two approaches differed significantly in all simulated datasets, indicating the need to use our proposed approach when estimating the uncertainty of diversity products. Based on the results of the regression analyses, our approach is especially useful in communities with high TD and ND. The use of one measure (ODP’) instead of the correlated TD, ND, FD, and PD is legitimate. It provides a better understanding of underlying conditions and events (the correlation matrix and the related probabilities are given in ESM1, Table S13).
Combining different categories of diversity into synthetic metrics has been proposed earlier. Among the most influential works are those of Cadotte et al.18 (who combined functional and phylogenetic species dissimilarities introducing new metrics, traitgrams, and testing community assembly patterns, among others) and de Bello et al.19 (who decoupled functional and phylogenetic dissimilarities, computing unique and shared components, introducing novel metrics, used multivariate analysis, improving inference on environmental filtering, etc.). While sharing certain similarities (such as integrating multiple dimensions and data sources for diversity, sensitivity to metric interaction and shared effects, development of composite novel metrics, and methods of their illustration), our approaches also present important differences. Among these, we highlight the use of a novel metric (ND), which increases the multidimensionality, the use of geometric and multiplicative integration (compared to additive integration mainly used in the references), the geometric interpretation of the diversity hypervolume, the different variants provided for our measures, with explicitly scaled and normalized components, and the extended use in multivariate canonical ordination space. Thus, our approach shares the same motivation by integrating diversity dimensions, but it is complementary by expanding the scope, mathematical formulation, interpretability, and including variants for different goals. It builds on the cited foundation while adding new structures and applications.
The novel lambda diversity measures the step-by-step changes in species composition along a gradient while accounting for the information change represented by what is conserved, new, or lost along the gradient. There are similarities between the lambda diversity and the species turnover rate (beta diversity) in that both compare the rates of lost and gained species. In some formulae, the length of the gradient (e.g., time or census interval, spatial dimension) is also considered. In our sequential measures, we also account for species’ repeated disappearance and reappearance, characteristic of metacommunity dynamics. Loss and gain of taxa are considered frequent, probabilistic, and recursive events caused by the continuous dynamics of the environment.
The proposed measure for lambda diversity relies on incidence-based data, but it can be adapted to use abundance data, which would give more insight into the amplitude of change in community composition along the explored gradient. In addition, if more than one gradient is concerned, our lambda diversity index could be adapted to suit changes that co-occur in 2D, 3D, or even nD spaces.
The longitudinal gamma analysis could be further developed to evaluate whether the levels of alpha and beta (or lambda) diversities may have been obtained by chance and to identify, by ordination methods, the groups of biological entities (species belonging to the same taxa or life form) that are mostly responsible for the overall diversity, as it is done (e.g.,) for TD, FD, and PD in the adiv package in R83.
When measuring species richness or turnover, a source of bias is represented by undersampling rare species. A small sample size increases the underestimation of true alpha diversity, and the differences in sampling effort hinder comparison among communities. Several solutions for this problem are available. These render estimates that account for unobserved taxa and range from the rarefaction, extrapolation from species-area relationship, and nonparametric estimators84 to methods based on the variance of the estimates in measurement error models used to compare diversity across communities85. We could expand our methods by incorporating the uncertainty given by low sampling effort or species rarity into our approach and metrics, allowing a more reliable ecosystem comparison.
The newly introduced lambda diversity and its extension towards the relative sequential cumulative gamma diversity show certain similarities with previously introduced variants of directional beta diversity measures44,86,87,88, to which they are fundamentally connected. The common feature is tracking species composition directional changes along gradients. Our approach shares some features with the quoted sources, such as using sequential comparisons with common (dis)similarity measures. However, it also uses some peculiar innovations, such as the modified squared similarity index, cumulative structures (triangular matrices summarizing cumulative species change along a gradient, allowing for a non-pairwise structured analysis of the community shift), extended and original introduced metrics (Λ, gmd, cgmd, Acgm, Rlgm, Drgm, γρ), which aim to assess gradual change in information (i.e., fragmentation vs. continuation of change along a gradient), as well as benchmark simulations (M0 and M1) which establish reference extreme cases of total turnover vs. stability and uniformity. The last provides a framework for comparing real data to extreme limit scenarios and hence reveals the meaning of the shifts and trends, including recursive events. We also mention the combined use of matrix algebra, interpolation functions, integral and differential calculus for assessing the newly introduced metrics. All these are illustrated directly and clearly in Mathcad (ESM 3), and a variant is also given in R, for a different case study (ESM 5). Thus, lambda diversity is connected to directional beta diversity, and they may be used interchangeably in relation to the asked research questions and data availability. However, the highlighted distinctions, especially those on cumulative structures, calculus, novel geometric summaries, and the relation of the real data to the benchmark simulated extreme scenarios that serve as interpretable baselines, enable lambda diversity to serve as a summarized descriptor of species (dis)continuity and turnover building upon the directional beta diversity, but extending beyond its scopes, expanding its inquiring possibilities.
Relying on Rao’s entropy measure65, the xi diversity combines and expresses unitarily the species heterogeneity and richness, the heterogeneity in environment, space, and potentially others (such as sequential cumulative changes along a dominant gradient, phylogenetic signals, niches’ variability, and dissimilarities). In our case study the correlation between the two xi diversities was weak, and they responded independently to human impact and the other environmental predictors, indicating the need to consider space in the evaluation of ecological diversity. Previously, the importance of spatial patterns was revealed in the partitioning of ecological and evolutionary processes that influence community assembly, as spatial signals found in traits that are not correlated with the environment indicate that some key environmental processes might have been overlooked45.
We exemplified the calculation of xi diversity using a longitudinal gradient dataset—communities sampled along a river—but the proposed measure may be applied to any data type, regardless of the sampling design. Our xi diversity measure incorporates the spatial dimension of species assemblages, which we propose to include in other diversity analyses as well. For instance, the zeta diversity combinatorics could be related to the distance between the sites for testing the spatial structure of environmental features and their effect on diversity patterns.
A variant of the xi measures presented here is to use, instead of a distance (or dissimilarity) matrix, the Community Weighted Means (CWM) by combining the community data (usually a site-by-species matrix) with a data table of functional traits (usually a species-by-traits matrix). Then, the dissimilarities are computed on CWM instead of the original data, and by this, traits are also included in the xi ecological diversities. This could be a step forward from the methods and illustrations of CUVARP (“Cumulative variation partitioning with multiple response and predictor matrices”) and CWM-AVARP (“Community-weighted means average variation partitioning with multiple response and predictor matrices”) introduced by Sîrbu et al.2.
With the transition from representing the community as a species assemblage to an interaction network, diversity measures were adapted to ecological networks by incorporating the probability (or strength) of interaction between species or guilds89. As zeta diversity was successfully applied in evaluating interaction turnover90, its extensions, lambda and xi diversities, may also be adapted to accommodate interspecific relationships along with community composition.
Recent advances in Bayesian modeling have also been incorporated into the methodological development of diversity research, enabling the inclusion of uncertainty in diversity estimation and partitioning. This provides an easy statistical framework for testing ecological questions91 and could be further developed for our proposed methods.
Being aware that by combining various diversity measures, their individual meaning might be diminished or lost, we proposed some geometrical ways of illustration, which could preserve both the synthetic value and the individual contributions of each implied measure. When three standardized indices are considered, they define a parallelepiped (Fig. 1) embedded within a cube of maximal volume 1, and ODP may be defined as the ratio of the real state to the maximal possible value of diversity. When more diversity measures are involved (as we did by including PD), geometric representation of the nD volume is problematic, but 2D radar graphs (Fig. 3a) or ordination diagrams may be used. The latter can be either indirect, containing only the diversity measures and depicting their relations, or direct (canonical) (Fig. 3b).
An alternative approach, which we propose to be used especially with a higher number of diversity measures, is the computation of SADDI based on the calculus of distance or dissimilarity matrix determinants. A general interpretation is that the determinant of a distance matrix assesses the geometric dispersal, variance, or independence of objects in a multidimensional space. The higher its value, the broader the volume defined by more widely distributed, well-spread, independent, or complementary diversity measures, deriving from a greater ecological complexity. A large value indicates non-redundant diversity components, reflecting distinct features, ecological heterogeneity, and multiple or complex functions. It is a matter for future studies to analyze the wide range of implications and significance of its variation. SADDI can be used, as we did, to compare ecological systems multidimensionally and multicriterially. It can also be viewed as a useful tool in assessing conservation effectiveness, ecological monitoring, and management.
Delineating hypervolumes and measuring their characteristics and use in niche- and trait-based ecology has already been addressed by Blonder et al.20,21,23 and other sources, posing a certain overlap with our approach. The quoted methods and associated R package are mainly statistical and data-driven, the multidimensional hypervolume being defined in terms of shape, size, and overlap of empirical traits or niches, relying on statistical geometry. In contrast, our methods use deterministic methods (algebraic operations, determinants of distance matrices), and the hypervolume is regarded as an abstract geometric parallelotope in an nD diversity space. The compared approaches are complementary but differ in concept, mathematics, and purposes: while the statistical approach of the hypervolume is better suited for empirical modeling of traits and ecological niches, we consider our approach suitable for meta-analysis of different categories of diversities, allowing for comparative and theoretical studies across different scales and spaces. The methods proposed by Blonder et al.20,21,23 produce many detailed representations and measures of different facets of traits and niches, whereas our framework works at a higher abstraction level using various synthetic measures. Thus, the two approaches are not redundant or concurrent but complementary, and they can be combined in a synergistic way. For instance, Blonder’s detailed probabilistic facets of the hypervolume can be used as building blocks for our synthetic model, e.g., the hypervolume size might enter as a functional diversity input in our ODP, or the dissimilarity metrics of the quoted sources might be used in spatial or environmental heterogeneity in our lambda or xi diversities, and hence forging a holistic biodiversity synthesis. Certain overlaps also exist with other approaches, such as the quantification of hypervolumes in defining functional diversity22. This method focuses on detailed functional trait-space analysis, requiring comprehensive data on functional traits, being highly valuable in functional ecology research. Again, we stress the complementarity of the two frameworks since our measures could incorporate and benefit from indices developed by the mentioned work, providing more robust insights into the ecological patterns and their dynamics. For instance, the kernel.alpha—measuring the functional richness as the total hypervolume—might be a justified input in our ODP, while the kernel.beta—assessing functional dissimilarity—or kernel.dispersion—evaluating trait dispersion22, might be useful in the SADDI framework. Developing an inclusive approach and harmonizing all these models and frameworks is a necessity and a step forward in future diversity research.
Searching for integrative approaches and theories that combine various facets and fields of science is not a new endeavor. Maybe the most popular and well-known quest is for the unified field, which aims to combine the four forces in physics and reconcile the quantum world with the macro world or the universe described by the theory of relativity. There were opponents to this quest for the unified field, as expressed by the physicist Wolfgang Pauli—‘may no man be able to bring together again what God has separated’ as quoted by Kaku92—before he turned to searching again for ‘bringing together’ the separated forces and theories. We are confident that our attempt to bring together different aspects, methods, and meanings of diversity, combining the richness and heterogeneity of life with those of space, environment, and others, will elicit criticism and reluctance. We are aware of the drawbacks of our attempt to combine distinct conceptual and mathematical sides of diversity. We agree that the meanings of individual sides or measures are lost by multiplication, summation, or other algebraic procedures. However, we also state that we don’t want to replace, but rather add our perspective to those already defined and commonly agreed upon.
The freshwater mollusk communities from the middle Olt River served as a model group and case study also earlier for introducing ND and the VADOC diagrams (variation partitioning in double-constrained ordination analysis)13 and for analyzing the differential responses of native and alien components of assemblages to environmental gradients, next to the introduction of cumulative variation partitioning in constrained analysis with multiple response and predictor matrices2. As stated in the mentioned sources, the middle Olt River was heavily impacted in the second half of the XXth century by damming and the construction of a row of reservoirs along its course, to which the pollution characteristic of the period of centralized economy was added. Hence, diversity reached a minimum at the very beginning of the 1990s’. Later, communities recovered following pollution reduction, lasting effects of hydro-technical works, heterogeneity of life conditions, and alien species colonization. Since dams and reservoirs still constitute the main source of pressure, one would expect low diversity due to their lasting impact, but this is no longer the case. The ongoing trend is illustrated in Figs. 4–6. Only the xi diversity responds negatively to increased impact, all the other measures (except for the SADDI, which we will address later) are directly related to stronger modified habitats. The reason is linked to improved living conditions and habitat heterogeneity at a medium geographical scale. The row of reservoirs disrupted the continuous flow, creating a broad range of habitat categories, from lentic to lotic, from (very) soft and fine to hard substratum, from deep to shallow waters, etc. In shallow waters, luxurious hydro- and hygroflora have developed, favoring conditions also for macrophytophylous pulmonate gastropods and other elements that thrive near the lake edges. Hence, diversity (TD, ND, FD, PD) increased (with impact, as shown in Fig. 4, 5, 6). Rheophylous elements survive in lotic environments while lentiphylous communities thrive in reservoirs, moving (harmonica effect) with water levels fluctuating seasonally but also regulated anthropogenically in relation to energetic demands and water supply. To the native elements, a set of alien species has been added mainly during the last two to three decades. The species richness increased over time, native species following a linear model, while alien taxa exhibited a polynomial pattern. The functional diversity showed a non-linear trajectory, initially decreasing and reaching a minimum during the major physical changes to the riverbed, after which it recovered and surpassed the initial level2. Over the last 150 years, the river has proven to rely on a low to medium resistance but a high resilience. The value of the ODP’ including the geometric mean of TD, FD, ND, and PD, showed a moderate average multidimensional diversity, balanced (no extreme values) across dimensions, the sensitivity analysis (in our study expressed as variation partitioning) sustaining environmental filtering and niche differentiation as major drivers of communities’ structure. Since ND contributes most to variation in ODP’ it highlights the importance of species sorting, species-habitat matching, and environmental filtering in rivers altered by reservoirs, supporting the paradigm of species-sorting coined by Leibold et al.93,94 and of the importance of environmental filtering in fragmented rivers95, which is joined by dispersal limitations. Fragmentation of rivers has huge repercussions on the dispersal of aquatic organisms and even on their ability to survive. For instance, Dias et al.96 found significant effects of fragmentation by dams on the extinction of native resident fishes, in sharp contrast with the lack of effect proved for many other factors commonly viewed as threats. The effects of dams might be reduced by enhancing connectivity through lateral channels and passive dispersion, with one main mechanism being the pathways related to human activity. In the middle Olt River, the dams and the row of reservoirs have undoubtedly not hindered the upstream (for alien species introduced in lower reservoirs) and downstream (for rheophylic native species recolonizing the middle regulated river sector) dispersal, one main mechanism being the anthropogenic vectors, both active and passive (e.g., boats, balast excavations, constructions, equipments, fishing, tourism). Downes et al.97 evidenced the prevailing importance of dispersal in structuring communities in anthropogenically impacted rivers, distinguishing between effects of degraded environmental conditions and barriers.
The cumulative sequential lambda and its associated measure of gamma diversity illustrate the gradient-related stepwise community change, detecting species turnover, and reflecting fragmentation and recurrence. The relative gamma diversity change was γρ = 0.33, showing a low to moderate species turnover, i.e., about one third of the maximal value (on a 0 to 1 scale), lower than expected in a regulated river with dispersal barriers. Our interpretation focuses on gradual shifts in mollusk communities’ structure along the gradient, despite fragmentation by dams and reservoirs, with alternating environmental conditions reflected in the peculiarities of non-random communities’ dynamics. The metacommunity seems to be dynamically stable along the gradient (Fig. 4), existing within a narrow band that avoids extremes, and able to respond to habitat changes through correlated abilities to recolonize patches, proving a certain degree of resilience with moderate spatial and taxonomic sorting. Environment features explained almost the double amount of variation (24%, unique effect 12.3%) in the complement of the lambda matrix (Fig. 5) compared to spatial eigenfunctions (about 14.8%, unique effect 3.1%), while the shared effect was about half the total explained adjusted variance. These values indicate that spatial structuring and related processes (sensu Heino et al.98 act as significant compositional drivers, but are about one quarter as important as environmental conditions, which drive species sorting, with dispersal limiting the barrier effects of dams and other hydrotechnical works. According to the model developed by Heino et al.98, the interaction of spatial and environmental predictors in our study indicates that species sorting overrides both mass effects and dispersal limitations, the intermediate spatial extents lowering dispersal so that communities are not homogenized, allowing species composition to be explained primarily by environmental features across sites. According to the quoted source98, most studies on rivers have concluded that environmental control prevails over spatial constraints, but their balance is tuned to various factors including taxonomic group, evolutionary history, functional traits, dispersal strategy, external factors, spatial extent, and stochastic factors, among others. However, some criticism about the statistical methods mostly used in this kind of study was also raised. For instance, dispersal abilities were mostly assessed by proxies and low variation explained by the spatial component could also be related to the mixing of species with various dispersal capabilities98, something that can be buffered by using traits instead of only taxa. When functional traits of species are considered together with environmental predictors, these may be more informative. For example, in the double-constrained correspondence analysis (dc-CA)2, the selected traits explained more (adjusted R2 = 43%) than the environmental predictors (adjusted R2 = 34%) in freshwater mollusk species composition, proving the importance of traits as evidence of evolutionary signals and adaptations to local conditions. Treating native species separately from alien invasive species has also yielded contrasting responses, adaptations, and relationships with the different classes of predictors2.
The xi diversity measures, coupling community composition with environmental (ξ1) and spatial (ξ2) heterogeneity, are weakly correlated in our case study, showing a certain incongruence of spatial and environmental structuring, summarized in the integrative ecological diversity (ξ3). The latter was assessed at 0.213 (normalized value), hinting towards a moderate-to-low coupling between heterogeneity of life, environment, and space, being directly related to (and increasing with) lotic conditions, flow, and distance to the nearest downstream dam, while proving an inverse relation with the human impact (Fig. 6). Reservoirs, dams’ proximity, and altered flow lead to local homogenization of communities and life conditions, decreased xi diversity, and possibly dispersal control over the environment. This is in line with the species sorting interpretation that in our study, local processes (e.g., environmental factors and species interactions) are more important than regional processes (mainly dispersal). However, the overlap between local and regional controls could assess the degree of spatial autocorrelation and hence the level of disturbance99. The large overlap between spatial and environmental predictors, in both lambda and xi analysis, indicates that the disturbance is almost as great as the unique effect of the environment, but much larger than the uniqe effect of space (and as such of the regional proccesses, i.e. dispersal). This is another argument that our xi (and especially ξ3) diversity links functional and environmental heterogeneity to spatial processes, establishing a tool for monitoring and bioassessment, as well as for other uses of metacommunity theory in ecological and environmental applications, as described by Brown et al.99. It offers the possibility to evaluate resilience and whether this is environmentally (locally) or spatially (regionally) driven. The dams and reservoirs are affecting an increasing number and percentage of riverine systems worldwide. The assessment of their effects on metacommunity dynamics and further on the implications in ecosystem services, might surely benefit from models of intermitent rivers and ephemeral streams (IRES), which is an emerging field of study with multiple socio-ecological implications, especially in the frame of climate change and aridization100,101,102. In return, the IRES models might benefit and be enhanced by integrating our novel synthetic diversity measures.
Reflecting the multidimensional spread of diversities distinctivness, the SADDI metric has higher values in either more natural or impacted river sectors, extreme habitats, close or distant to the dams, indicating heterogeneous communities with no redundant diversity components (Fig. 7). Lower values of SADDI indicate transitional or ecotone conditions, inhabited by more generalist and tolerant species, communities characterized by redundant diversity and similar functional characteristics of different taxa. As such, SADDI is mirroring, being inversely linked to, the dynamics expected by the Intermediate Disturbance Hypothesis (IDH) coined by Connel103, which relates the mode or maximal value of the diversity response curve to disturbance at intermediate range. Increasing values of SADDI towards impacted sites, as well as to opposite ends of the gradients, might illustrate the newly assembled functionally diverse communities, while in mixed zones, convergence and contraction of niche space might prevail. Thus, ecological responses to the magnitude and types of altered flow regime, as reviewed by Poff and Zimmerman104, cannot be just divided into increased, decreased, or mixed, but a combination of all these, intertwined in a complex, non-linear dynamic response finely tuned to time, space, and the analyzed gradients. We conclude that our novel metrics capture the multidimensional responses of biological and ecological diversity to environment, space, and human pressure, its dynamics being shaped by a complex interplay between species richness and turnover, traits, niches, and external features, quantifying and capturing communities’ dynamics and sorting, as well as the resilience of the ecological systems.
In defining and separating the ecological properties of stability (degree of fluctuation around specific states) from resilience, Holling105 characterized the latter by the ability of systems to persist by maintaining relations and absorb changes in their components. We have defined earlier another concept complementing the original terms of Holling, namely the ecological reliability, as “a measure of the capacity of an ecological system to maintain or to recover its functions when exposed to perturbing factors, or after such a perturbation has altered its functional responses”106. Our term defines a functional side of resilience by characterizing the abilities of systems to maintain or recover functions with altered or restructured composition, while allowing for different future trajectories. Hence, it is related to, but distinct from other widely used terms such as ecological stability, resilience, functional redundancy, ecosystem integrity, or resistance. Since our synthetic and multidimensional measures of diversity are directly related to ecological systems’ capacities of retaining and regaining complexity and functions when confronted with perturbances, we hope that in the future, ecological reliability will also gain a formal definition and a valid framework for quantification. We suggest that reliability in ecological context is directly related to ODP’ (which captures internal diversity structure, indicating a potential integrity and functional compensation after a perturbation), xi metrics (which couple diversity of life to environment and space) and SADDI (multidimensional distinctness), while being inversely related to the lambda diversity (tracking community turnover as response to altered conditions which might hint towards functional recovery). The intimate relations between these terms and metrics are still under study and have to be formally developed.
When we discuss the biodiversity crisis, everyone understands the exact meaning related to species extinction, habitat debasement, range splitting, etc. But there is, unfortunately, another meaning represented by the drop in experts capable of biodiversity assessment and their ability to communicate with non-experts, such as politicians, stakeholders, managers, and the general public, who are the key players on the stage. One neglected problem is what has been defined as the ‘identification crisis’, meaning the decrease in the number of experts in species identification and the decline of species-based biodiversity research107. For example, Hungary lacks experts in about half of its known fauna, while for another 25%, there are only one or two active experts107. In our opinion, the reality in other countries (e.g., Romania) is much worse, and there is little evidence that this issue will be mitigated. Relying only on technical novelties, such as environmental genomics, might mitigate but not solve the biodiversity (identification) crisis related to the human factor. Another rising problem is the communication between (fewer and fewer) experts or producers of diversity information and (the increasing number of) consumers (managers, politicians, stakeholders, the general public). Expressing trends in many different diversity measures will certainly not support communication, but synthesizing dynamics of an overall or global measure of diversity might help. This might be another upside for our attempt to combine and express diversity also in a synthetic way. Avoiding technicalities and subtle details while describing the general trend of biological and ecological diversity by a measure that states the decline, constancy, or increase could facilitate collaboration among people with diverse interests or backgrounds and establish the stage for a common understanding and cooperation.
Biodiversity, and especially ecological diversity, is inherently multidimensional. Therefore, composite indices are designed to summarize holistically their components into a coherent mathematical fashion. Our synthetic composite indexes (xi or SADDI, to mention the most typical) are not meant to replace but to complement the use of individual measures, which may be used or not depending on the research question, data availability (and experts involved), the number of dimensions, the correlation matrix between them (at last in the case of ODP’, its use is more effective when, as it happened in our case, individual classical diversity measures are already correlated, positively and significantly), but also to whom the report is helpful, and the purpose it serves. Reduction in complexity improves comparisons between ecological systems and processes, and it can enhance tools for detecting features of multidimensional structures. The synthetic measures we have introduced are not merely averages but also capture the geometry, multidimensional metrics, dispersion, and divergence. They are also helpful in detecting dependence and relations between components and hidden structural features that are not covered by univariate measures. Using synthetic measures improves communication between ecologists and non-experts, including policymakers and stakeholders. They can also facilitate comparative studies, elucidating trends, simplifying model inputs, multi-criteria decision-making in monitoring, management, modeling, citizen science, or public debates. When convenient or necessary, we also strongly recommend using both individual and composite synthetic measures in a context-dependent and responsible manner.
In the final chapter of their book, which is dedicated to frontiers in the measurement and assessment of biodiversity, McGill and Magurran108 pleaded for the search for means to link its many different approaches and measures. The last of the eight trends they identified in biodiversity research was related to the scarcity or lack of ties between the processes that control and the metrics that measure biodiversity. Adding in our article a few tools to what the quoted source called the process-based trend, was also one of our guiding principles. Paraphrasing McGill and Magurran, we strived to add some value to the methodological link between descriptive and process-based ecological research, summarize the panoply of models and measures more synthetically, and try to identify some measures that might prove to be more fundamental. While recognizing that the methodology related to biodiversity becomes complex and sophisticated, the quoted authors stated that work on this issue has to be guided by some broader principles, the last of them being “keeping firm the reality and measurability of biodiversity to the general public” (idem). Our present work aligns with these statements, including and highlighting the last one in particular.
Our novel measures, especially the xi integrative ecological diversity and the SADDI index, aim to define and measure characteristic, synthetic features of the system defined by intersections between Biology – Ecology – Environmental Sciences – Environmental Engineering – and – the (Human) Society (which we call the BEES System) (Fig. 8). The BEES system represents the holistic level and concept where mitigations and solutions to complex problems faced by humanity, the environment, and life must be pursued and found. One reason for failing to stop the environmental and biodiversity crisis and its effects on human society is the oversimplification and unilateral nature of measures taken, which underestimate the dimensionality of issues. Recognizing and approaching the BEES system through a complex, synthetic, and multidimensional framework, including measures and related solutions, might bring the dawn of establishing tools and hope for the future.
Overview of our contribution to the development of a multifaceted diversity framework in the context of the Biology – Ecology – Environmental Sciences – Environmental Engineering – and – (Human) Society (BEES) System. The figure summarizes multiple dimensions of diversity, including established metrics: TD—taxonomic diversity, FD—functional diversity, ND—niche-based diversity, PD—phylogenetic (or genetic-based) diversity, α—alpha diversity, γ—gamma diversity, ε—epsilon diversity (diversity within large biogeographic areas), ζ—zeta diversity, β—beta diversity, δ—delta diversity (change between large biogeographic areas), and newly introduced metrics: λ—lambda diversity, γρ— relative sequential cumulative gamma diversity, ξ— xi ecological diversities (ξ1—environment-life diversity, ξ2—life-space diversity, ξ3—integrative ecological diversity), OD’—overall diversities (ODP’—product-based, ODS’—sum-based expressions), SADDI—Standardized Average Diversity Distinctness Index. For clarity, only the primary relationship—through SADDI—is illustrated; other interrelations between diversity measures and the BEES System components are not shown.
Instead of evaluating and reporting different aspects of diversity and decomposing it, we adopt an opposing view, namely, combining and synthesizing information. Communication between experts, managers, laypeople, stakeholders, and policymakers can be enhanced, and environmental monitoring and related fields may also benefit, contributing to better prioritization of objectives in both theoretical and applied sciences. Using and interpreting the synthetic framework and its measures will sustain trans-, meta-, and interdisciplinary research and communication, linking people and their different interests and knowledge to nature and its values. It will also presumably contribute to interdisciplinary educational programs because science experts and students might be motivated to train within environmental and ecological sciences while also considering applied mathematics and IT, thus enhancing the general intellectual and cultural heritage, with benefits for society, environment, and life.
