Layout 1 INTRODUCTION The field of biodiversity-ecosystem functioning (BEF) studies the impact that changes in biological diversity can have on the functioning of ecological systems. For in- stance, it depicts the consequences of biodiversity loss in terms of basic functions such as nutrient uptake, respira- tion, primary production and nutrient recycling, among others. In BEF research, diversity is quantified generally as the number of species (i.e., richness) or using other richness-based metrics that include species’ abundances (e.g., Simpson index, Shannon index, Evenness). Overall, and despite some exceptions, it is well established that most ecosystem functions and their temporal stability in- crease as the number of species increases (Hooper et al., 2005; Cardinale et al., 2012; Duffy et al., 2017). Given the low explanatory power of such richness-based diver- sity metrics and the absence of a proper mechanistic elu- cidation, BEF research is increasingly adopting a trait-based perspective (Flynn et al., 2011; Cardinale et al., 2012; Krause et al., 2014; Gagic et al., 2015). Trait variability and the resulting ecological differentiation among species are considered as major determinants of the nature and strength of species interactions and conse- quently are expected to have a direct strong influence on ecosystem functioning. However, determining and quan- tifying the traits that are relevant for ecosystem function- ing is not straightforward. Despite the importance of phy- toplankton for global scale processes such as oxygen production and primary production, trait-based BEF stud- ies with phytoplankton remain rare. Here, I review exist- ing studies linking trait-based diversity to ecosystem functioning in freshwater lentic systems, summarize their major findings and provide some ideas for future devel- opment of this underexplored line of research. STUDY SELECTION I collated all the published empirical studies on the re- lationship between freshwater phytoplankton diversity in lentic systems (lakes, reservoirs, ponds) and any aspect of their functioning. I first collected all previous reviews and meta-analyses on the topic of biodiversity and ecosys- tem functioning (BEF), irrespective of the organism in- cluded and checked for references on freshwater phytoplankton (Hooper et al., 2005; Srivastava and Vel- lend, 2005; Balvanera et al., 2006; Cardinale et al., 2009; Cardinale et al., 2011; Cardinale et al., 2013; Gross et al., 2014; Duffy et al., 2017). This was supplemented with a search of the ISI Web of Science database using the key- word sequence combining (freshwater OR lake OR pond OR reservoir) AND (phytoplankton* OR alga* OR diatom Advances in Oceanography and Limnology, 2017; 8(2): 179-186 REVIEW DOI: 10.4081/aiol.2017.7207 This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). Biodiversity ecosystem functioning research in freshwater phytoplankton: A comprehensive review of trait-based studies Patrick Venail1,2 1Department F.-A. Forel for Environmental and Aquatic Sciences, University of Geneva; 2Institute for Environmental Sciences, ISE, Geneva, Switzerland ABSTRACT In an effort to reach a clearer mechanistic understanding of the influence of biological diversity on ecosystem functioning, research in the field is increasingly applying a trait-based approach. In this comprehensive review, I searched for and analyzed studies that focused on the relationship between biodiversity and ecosystem functioning (BEF) using a trait-based approach in freshwater phytoplankton from lentic systems (lakes, ponds, reservoirs). I found that this type of studies is very rare and included a plethora of traits, diversity metrics, sta- tistical analyses and study locations that contributed to the high variability in the results they obtained. Overall, trait-based diversity is not a very good predictor of ecosystem functioning in freshwater lentic ecosystems. Null relationships between trait-based diversity and ecosys- tem functioning in freshwater lentic systems were the more frequent outcome. When significant, the amount of variation in ecosystem functioning explained by trait-based diversity was small. Still, trait-based research remains a promising approach to increase the mechanistic understanding of BEF relationships. For this purpose, studies directly testing the underlying mechanistic rationale, exploring diversity effects on the temporal stability of ecosystem functions, including multiple functions at a time, focusing more in cell size and shape and confirming the relative importance of individual trait variation for ecosystem functioning are needed. Key words: Biodiversity; freshwater; functioning; phytoplankton; traits. Received: November 2017. Accepted: December 2017. Non -co mmerc ial us e o nly P. Venail180 OR cyano*) AND (trait* OR function*) AND (diversity OR richness) AND (community OR ecosystem) AND (function* OR product* OR biomass OR biovolume OR resource use). In this review, I only included studies that statistically analyzed the link between any trait-based metric of diversity and ecosystem functioning using fresh- water phytoplankton. I excluded studies in lotic systems (i.e., rivers and streams) and studies in which phytoplank- tonic organisms were classified into functional groups, such as major algal groups or Reynold’s classification be- cause they did not include a clear trait-based diversity measure (Schmidtke et al., 2010; Behl et al., 2011; Borics et al., 2012; Fernandez et al., 2014; Abonyi et al., 2017). STUDIES I found only six studies that explored the relationship between trait-based diversity and ecosystem functioning using freshwater phytoplankton from lentic systems (Tab. 1). These studies are recent, with the oldest published only seven years ago (Vogt et al., 2010), revealing that, on average, less than one study per year has been published in this topic. The six studies can be separated into two cate- gories: field and laboratory studies. The former includes four studies in which both trait-based diversity and ecosystem functioning data were obtained from natural conditions (Vogt et al., 2010; Pälffy et al., 2013; Santos et al., 2014; Fontana et al., 2017). In the other two studies, diversity was directly manipulated under highly controlled conditions in the laboratory (Shurin et al., 2014; Steudel et al., 2016). The main conclusions of these studies regarding the link between trait-based diversity and functioning are quite variable (Tab. 1). Below, I explore those differences more in detail. ORGANISMS AND FRESHWATER SYSTEMS The six studies reported a wide variety of organisms from all major phytoplankton groups including mostly chlorophyta (green algae), chrysophyta (golden algae), bacillariophyta (diatoms) and cyanophyta (cyanobacte- ria). Phytoplankton from other major groups such as glau- cophyta and heterokontophyta were less frequent. The richness of taxa within each study was rather variable. The laboratory studies included 16 (Shurin et al., 2014) and 64 (Steudel et al., 2016, only chlorophytes) taxa respec- tively, whereas field studies reported 212 (Vogt et al., 2010) and 412 species (Santos et al., 2014). The other two field studies did not report the number of species ana- lyzed. Field data were collected in a variety of freshwater systems around the world: Vogt et al. (2010) included data from 65 lakes in Canada, Pälffy et al. (2013) from one single lake in Hungary, Santos et al. (2014) included 19 reservoirs from Brazil, whereas Fontana et al. (2017) re- ported data from 28 lakes in total, 2 from Switzerland and 26 from the Danube delta in Romania. TRAITS A total of 33 traits were included in the six studies about the effect of trait-based phytoplankton diversity on ecosystem functioning in freshwater lentic systems (Tab. 2). The authors selected such traits based on their supposed ecological relevance for competitive interac- tions, reproduction, predator avoidance, resource acqui- sition and/or bioenergy production. Most of these traits were measured at the species level, meaning that they rep- resent an average value obtained by measuring and recording traits in some representative individuals or pop- ulations from each species. Different to all other studies, Fontana et al. (2017) recorded individual level data for seven traits, meaning that species’ identification was not required prior to trait measurement and the reported trait values may reveal both intra and inter-specific variability in the phytoplankton community. Traits reported belong to three different categories: de- mographic, morphological or physiological. Demographic Tab. 1. List of trait-based BEF studies in freshwater lentic systems, type of study and their main conclusions regarding trait-based diversity effects on ecosystem functioning. Reference Type of study Main conclusion Vogt et al., 2010 Field Trait-based diversity was positively associated with total community biovolume Pälffy et al., 2013 Field Significant negative correlations between total biomass, functional group diversity and functional group evenness Santos et al., 2014 Field A positive relation between productivity and diversity, except for functional evenness for which the relation was negative Fontana et al., 2017 Field Trait evenness exhibited a robust negative relationship with biomass Shurin et al., 2014 Laboratory Biomass yield exceeded the component monocultures in polycultures consisting of species with highly divergent traits Steudel et al., 2016 Laboratory Functional diversity was positively correlated with biomass overyield BEF, biodiversity and ecosystem functioning. Non -co mmerc ial us e o nly Biodiversity ecosystem functioning research in freshwater phytoplankton: A comprehensive review of trait-based studies 181 (7 traits), also named life history traits, are all continuous and include population growth parameters such as r or K (measured by using chlorophyll a as a proxy for growth; Shurin et al., 2014). In another study, Vogt et al. (2010) included five demographic traits reported as “response traits” because they are based on how different environ- mental parameters influence population growth. These in- clude: optimal growth conditions regarding total nitrogen, total phosphorous, pH, dissolved organic carbon and dis- solved CO2. Morphological (10 traits), include continu- ous, categorical or binomial traits. In this category, we find cell size, also reported in some studies as cell volume, greatest axial linear dimension (GALD), maximal linear dimension (MLD) or maximum length. This is the more frequently used trait, included in five different studies (Tab. 2). While cell size is normally reported as a contin- uous variable, Palffy et al. (2013) reported it as a categor- ical variable with three size classes. Other morphological traits that are reported in multiple independent studies are growth form (referring to either colonial or single cell or- ganisms), presence of gas vacuoles (referring to buoyancy control capabilities), and the presence of flagella, that re- lates to motility. The other six traits in the morphological category were only reported in one independent study and can be either binomial or continuous (Tab. 2). Physiolog- ical (16 traits) represent the largest array of features among the three categories and can be either continuous, categorical or binomial and each was reported in only one study. These traits relate to minimum resource require- ments (light, nitrogen, phosphorous, silica), cellular chemical content, biochemistry or stoichiometry (lipids, fatty acid, carbon, nitrogen, phosphorous), resource ac- Tab. 2. List of traits included in BEF studies with freshwater phytoplankton. Demographic ER Type Reference Exponential growth rate, r* BP, CI Continous Shurin et al., 2014 Asymptotic density, K* BP, CI Continous “ Total Nitrogen optimal concentration Continous “ Total Phosphorous optimal concentration Continous “ pH optimal Continous “ Dissolved Organic Carbon optimal Continous “ Dissolved CO2 optimal Continous “ Morphological Cell volume/size/GALD/MLD/max length PA, RA, RE, BP, CI Continous/categorical all but Steudel et al., 2016 Growth form/body form/complexity PA, RA, RE Categorical Pälffy et al., 2013; Santos et al., 2014 Surface to volume ratio, s/v Continous Pälffy et al., 2013 Presence of aerotopes/gaz vacuoles/buoyancy PA, RA Binomial/categorical Pälffy et al., 2013/Santos et al., 2014 Presence of flagella/motility PA, RA Binomial “ Presence of mucilage PA, RA Binomial Santos et al., 2014 Presence of siliceous exoskeletal structures PA, RA Binomial “ Presence of heterocysts PA, RA Binomial “ Frontal shape of particle RA Continous Fontana et al., 2017° Cell rugosity/internal structure/gas vesicle/thylacoids PA, RA Continous “ Physiological Cellular lipid concentration BP, CI Continous Shurin et al., 2014 Cellular C:N ratio BP, CI Continous “ Cellular C:P ratio BP, CI Continous “ Minimum light requirement, L* BP, CI Continous “ Minimum nitrogen requirement, N* BP, CI Continous “ Minimum phosphorous requirement, P* BP, CI Continous “ Fatty acid composition CI Continous Steudel et al., 2016 Photosynthetic pigment composition Categorical Pälffy et al., 2013 Fluorescence chlorophyl a RA Continous Fontana et al., 2017° Fluorescence phycoerythrin RA Continous “ Fluorescence accesory pigments RA Continous “ Eveness in the distribution of pigments within cell RA Continous “ Hability to fix nitrogen Binomial Pälffy et al., 2013 Phagotrophic potential Binomial “ Motility/buoyancy Categorical “ Presence of toxins PA Binomial Santos et al., 2014 BEF, biodiversity and ecosystem functioning; ER, ecological relevance as explicitly claimed by the authors; RA, resource acquisition; RE, reproduction; PA, predator avoidance; BP, bioenergy production; CI, competitive interactions; *growth using fluorescence (chl a); °individual level traits. Non -co mmerc ial us e o nly P. Venail182 quisition (pigments, phagotrophy, nitrogen fixation) and toxin production. TRAIT-BASED DIVERSITY VARIABLES AND METRICS The six studies reviewed here include 29 trait-based variables that were associated to ecosystem functioning afterwards (Tab. 3). These variables can be classified in three categories: functional group based variables, trait- based diversity metrics and trait-based non-diversity met- rics. The functional group category includes well-known diversity metrics that are traditionally used to quantify species-level diversity such as richness, Shannon index, Simpson index and Evenness (the three latter incorporate information on species’ abundances). For this, species are first classified into functional groups such as those pro- posed by Kruk (Kruk et al., 2010) or in major algae groups (e.g., chlorophytes or cyanophytes). Then, the dif- ferent metrics were calculated based on group richness’ information. Kruk’s classification is based on morpholog- ical aspects and consequently corresponds to actual trait- based quantification of diversity; classifications based on major algae groups or Reynold’s groups are not trait- based only (Reynolds et al., 2002) and as such should not be considered as formal traits-based diversity metrics. The other two categories of trait-based metrics require collect- ing trait information on species (but see Fontana et al., 2017 for individual level trait metrics). Species’ traits are aggregated according to the taxa present in the natural community or artificial assemblage. This aggregation may include averaging, calculating distances or variation among species and other more sophisticated aggregation methods. The trait-based diversity metrics are the more common and diverse in the phytoplankton BEF literature, as I recorded up to 20 different metrics in five studies (Tab. 3). Some metrics incorporate information of one sin- gle trait (11 in total) at a time while others include up to six (Vogt et al., 2010), seven (Fontana et al., 2017), eight (Santos et al., 2014) or nine traits simultaneously (Shurin et al., 2014). This type of metrics can also be weighted by species’ abundances. None of the twenty different diver- sity metrics based on traits were used in more than one study, revealing a large variability in the methodology of trait-based BEF studies. Up to six different metrics were used in one single study (Vogt et al., 2010). Tab. 3. List of trait-based variables associated to ecosystem functioning in freshwater phytoplankton. Functional groups Reference Functional group richness (Kruk’s groups, taxonomic) Santos et al., 2014 Functional group diversity (Shannon Hf) Pälffy et al., 2013 Functional group diversity (Evenness Jf) “ Functional group diversity (Simpson) Santos et al., 2014 Trait-based diversity metric Difference in PCA vector (on 9 traits) Shurin et al., 2014 Functional dispersion, Fdis (on 9 traits) “ Difference in C:N ratios between 2 species “ Difference in cell volume between 2 species “ Difference in minimun light requirement L* between 2 species “ Fatty acid composition similarity, FTD Steudel et al., 2016 Fatty acid composition similarity, FD “ Variance of species in total nitrogen optima, TV Vogt et al., 2010 Variance of species in total phosphorous optima, TV “ Variance of species in pH optima, TV” Variance of species in dissolved organic carbon optima, TV “ Variance of species in CO2 optima, TV” Sum branch length dendrogram (on 6 traits) “ Functional richness based on distances, FR (on 8 traits) Santos et al., 2014 Functional evenness based on distances, Feve (on 8 traits) “ Functional divergence based on distances, MFD (on 8 traits) “ Functional divergence weigthed by density, MFDDens (on 8 traits) “ Trait diversity richness, TOP (on 7 traits) Fontana et al., 2017 Trait diversity evenness, TED (on 7 traits) “ Trait diversity divergence, Fdis (on 7 traits) “ Other trait based gradients (not variation) Average PCA vector (on 9 traits) Shurin et al., 2014 Average cell volume of 2 species “ Average C:P ratios of 2 species “ Average C:N ratios of 2 species “ Average minimum phosphorous requirement P* of 2 species “ Non -co mmerc ial us e o nly Biodiversity ecosystem functioning research in freshwater phytoplankton: A comprehensive review of trait-based studies 183 Finally, the third category includes trait-based metrics that do not represent diversity per se because they are just average traits among species and do not include information in the variation of a trait (Shurin et al., 2014). Some studies combine multiple of these metrics together in one single sta- tistical analysis to determine the combinations of metrics de- scribing better the variation in ecosystem functioning among communities (Santos et al., 2014; Fontana et al., 2017). ECOSYSTEM FUNCTIONING Five different variables were documented as measures of ecosystem functioning and can be separated in two cate- gories: the biomass related and the non-biomass related. The two controlled laboratory studies (Shurin et al., 2014; Steudel et al., 2016) focused on biomass related ecosystem functioning variables. For this, they started by measuring the biomass of both polycultures and monocultures using optical density as a proxy. Then, two log-ratios were calcu- lated: one between the biomass of the polyculture to the av- erage of constitutive monocultures (i.e., Net Biodiversity Effect, NBE) and another between the biomass of the poly- culture to the more productive monocultures (i.e., Overyielding, OY). This method can only be applied to con- trolled laboratory experiments because it requires monocul- ture’s biomass estimations. The third biomass related variable, used in two field studies was total community bio- mass (Vogt et al., 2010; Fontana et al., 2017). The two non- biomass related ecosystem functions included in the other field studies are chlorophyll a concentration (Pälffy et al., 2013; Santos et al., 2014) and a proxy for resource use effi- ciency calculated as the ratio between total biomass and available total phosphorous (Fontana et al., 2017). BIODIVERSITY AND ECOSYSTEM FUNCTIONING RELATIONSHIPS A key step in every BEF study is to relate biodiversity (the explanatory variable) to ecosystem functioning (the response variable). In controlled laboratory studies, this link infers causality because all the observed variations in ecosystem functioning result from changes in either the diversity and/or the composition of the species assem- blages being tested. In field studies, given the possibility of abiotic and biotic changes among sites or dates, the link between diversity and ecosystem functioning is just cor- relational. A plethora of statistical methods have been used to relate diversity to ecosystem functional as causal- ity effects or correlational links, including correlations, linear regressions and linear mixed effect models. The lat- ter allows combining multiple diversity metrics in one sin- gle statistical model (Steudel et al., 2016; Fontana et al., 2017). A total of 190 relationships between trait-based di- versity and ecosystem functioning have been established so far for freshwater phytoplankton (Tab. 4). I classified them into either positive, null or negative based on the statistical analyses directly reported by the authors. Half of them showed no influence of trait-based diversity on ecosystem functioning, meaning that variations in func- tioning are independent from variations in trait diversity among freshwater phytoplankton. This higher prevalence of null relationships, compared to the significant ones, was consistent in both field and laboratory studies with 42% and 53% of total BEF relationships being null re- spectively. Positive BEF relationships, meaning that ecosystem functions considered increase as phytoplankton trait diversity increases, were present in nearly 40% of the experiments. The authors suggest some potential mecha- nisms to explain this positive effect of diversity. A larger functional trait-based diversity may embrace a wider range of optimal growth responses in face of changing abiotic lake conditions, leading to higher biomass produc- tion (Vogt et al., 2010). Also, species with different traits would be more complementary in resource use or com- pete less strongly against each other (Santos et al., 2014; Shurin et al., 2014). Negative relationships were present in 11.5% of cases, which means a decreasing ecosystem functioning as trait diversity increased. As an explanation for this pattern, authors suggest that under some condi- Tab. 4. Summary table of the effect of trait-based metrics on ecosystem functioning. Type of study Reference Sign of trait based diversity effects on function Positive Null Negative Field Vogt et al., 2010 19 11 0 Pälffy et al., 2013 0 0 4 Santos et al., 2014 2 1 1 Fontana et al., 2017 6 19 11 Laboratory Shurin et al., 2014 8 2 6 Steudel et al., 2016 40 60 0 Total 75 93 22 Percentage 39.5% 49% 11.5% Non -co mmerc ial us e o nly P. Venail184 tions one single productive taxa with particular traits might dominate, leading to low functional diversity cou- pled to high biomass (Pälffy et al., 2013;Santos et al., 2014). Unfortunately, none of the studies reviewed here explicitly tested the suggested mechanisms and remained purely conjectural. Some authors also suggest that other forces (such as resource scarcity) might simultaneously influence both trait diversity and ecosystem functioning, resulting in a negative or positive pattern that is not me- diated directly by diversity (Fontana et al., 2017). SOURCES OF VARIATION IN BEF RELATIONSHIPS The reviewed papers allowed determining a series of factors influencing the relationship between trait-based diversity and ecosystem functioning in experiments with freshwater phytoplankton. Vogt et al. (2010) found big differences in sign and strength of the BEF relationship depending on: the trait, the number of traits, the metric of diversity and the organisms included in the analysis. For instance, whereas functional diversity had a positive effect on total biovolume in the benthic algae, this effect was not present in the planktonic compartment. The inverse happened for functional diversity based on one single trait (i.e., CO2 optimal). Fontana et al. (2017) showed that the relationship between trait-based diversity and ecosystem functioning may also vary among locations. For instance, a combination of three trait-based diversity metrics de- scribes very well variations in biomass in Lake Greifensee but this same set of metrics describes less well total bio- mass in Lake Zurich or the Danube delta. Steudel et al. (2016) showed that the effects of trait-based diversity also depend on the number of interacting species, with a ten- dency for higher influence of trait-based diversity as the number of species increases from two to sixteen. BEST TRAIT-BASED DESCRIPTORS OF ECOSYSTEM FUNCTIONING One purpose of BEF studies is to determine which trait (or set of traits) and which metric (or set of metrics) de- scribes better variations in ecosystem functioning. In case of significant effects (P<0.05), either positive or negative, one wants to know which trait is the best predictor using coefficients of variance (R2 values) and to compare mod- els using for instance the Akaike information criterium (AIC). This would reveal which trait matters the most for ecosystem functioning in freshwater lentic ecosystems. One may as well want to know which traits do not influ- ence ecosystem functioning. I collected data on the per- centage of variation in ecosystem functioning explained by trait-based diversity and found that this percentage ranged from 1.8% to 90%, with an average of 34.7% (n = 51). The model that explained the most (90%) of the variation in ecosystem functioning included one single metric of diversity (i.e., trait diversity evenness, TED), which was based on individual level trait variation and in- cluded seven different traits (Fontana et al., 2017). In studies focused on species level trait variation (5 out of 6 studies), the best single metric describing ecosystem func- tioning included information from six traits and explained 54% of variation in total biomass production in benthonic diatom communities (Vogt et al., 2010). The same study showed that the capacity to predict biomass production depended on the number of traits considered. This is, in- cluding two or three traits, rather than five or just one, in- creased the predictive power. This result suggests that different traits may incorporate different information but also that some traits might be functionally redundant and their use may not lead to a better description of biomass variation among communities. GENE-BASED DIVERSITY AS A PROXY FOR TRAIT-BASED DIVERSITY One study explored the capacity of gene-based diver- sity to predict freshwater phytoplankton functioning (Steudel et al., 2016). The underlying rationale is that gene differentiation among species (i.e., phylogenetic di- vergence) may relate to trait differentiation (assuming phylogenetic signal or phylogenetic niche conservatism; Blomberg and Garland 2002; Losos 2008; Wiens et al., 2010). Overall, the results show that the explanatory power is higher for trait-based diversity metrics than for gen-based diversity metrics. Moreover, at high richness level (16 species) both types of metrics have contrasting effects on biomass production. Trait-based metrics had a positive effect on biomass whereas gene-based metrics had a negative effect on biomass. This suggests that gene- based metrics should not be considered as proxies of trait- based ones in freshwater communities. IDEAS FOR FUTURE DEVELOPMENT In the biodiversity-ecosystem functioning context, sta- bility refers to the capacity of an ecological system to per- form ecosystem functions despite variations (e.g., perturbations) in the abiotic or biotic conditions over time. Overall, diversity is expected to have a strong positive im- pact on ecosystem functioning stability given the capacity of a diverse set of organisms to cope with the different en- vironmental conditions (Hooper et al., 2005; Tilman et al., 2006). Freshwater lentic ecosystems are steadily ex- posed to changing environmental conditions such as tem- perature, light intensity, resource input, etc. No study has Non -co mmerc ial us e o nly Biodiversity ecosystem functioning research in freshwater phytoplankton: A comprehensive review of trait-based studies 185 explored the influence of trait-based diversity on the tem- poral stability of ecosystem functioning. Trait-based BEF studies in freshwater phytoplankton focused on single functions. However, ecological systems often perform multiple functions at a time and some of these functions are expected to be directly related, such as resource uptake and biomass production. It has been suggested that diversity effects on ecosystem functioning might be stronger when multiple functions are considered simultaneously (Byrnes et al., 2014; Lefcheck et al., 2015). Such a multifunctional approach using trait-based diversity in freshwater lentic systems is missing. Cell size and shape are considered as key traits for phytoplankton, directly influencing resource acquisition, reproduction, predator avoidance and species interactions (Litchman and Klausmeier, 2008; Finkel et al., 2010). Whereas some trait-based BEF studies reported here in- cluded phytoplankton cell size and/or shape in the esti- mation of trait-based diversity, only one study (Shurin et al., 2010) explored directly the effect of these traits on ecosystem functioning. More studies manipulating phy- toplankton cell size and/or shape are required to provide more consistent conclusions on this topic. Individual trait information may increase the descrip- tive power of ecosystem functioning (Fontana et al., 2017). More studies incorporating and comparing the ex- planatory capacity of individual vs. species trait-based metrics of diversity on ecosystem functioning are also re- quired to determine the generality of this effect. Finally, motivation to include trait-based information into BEF research supposes that trait variability among species reflects their ecological differentiation and thus de- termine the nature and strength of species interactions that ultimately influence ecosystem functioning. Trait-based studies reported to date did not explicitly test this rationale and the proposed mechanistic interpretations remain purely conjectural. Studies directly testing the underlying mecha- nisms such as the traits involved in the prevalence of com- petition or facilitation should be further developed. IN A NUTSHELL This first comprehensive review of studies linking trait- based freshwater phytoplankton in lentic systems to ecosys- tem functioning revealed the scarce research conducted in this topic, with only six published studies over the last seven years. Two of these studies were conducted under controlled laboratory conditions and the other four studies reported field collected data, in which other external vari- ables might be influencing biodiversity-ecosystem func- tioning relationships. A total of 33 traits and 29 diversity metrics have been reported. Traits are either demographic, morphological or physiological. Some diversity metrics are based in one single trait but the majority include several traits simultaneously. No empirical evidence suggests that variation in one specific trait or group of traits improves the predictability of ecosystem functioning in freshwater lentic systems. Similarly, including multiple traits simulta- neously or including multiple trait-based diversity metrics together does not necessarily make BEF relationships stronger. Nevertheless, a plethora of traits, diversity met- rics, statistical analyses and study locations contributed to the high variability in the results obtained. Null relationship between trait-based diversity and ecosystem functioning in freshwater lentic systems was the more frequent outcome, accounting for nearly half of the experiments. When statis- tically significant, positive effects of trait-based diversity on ecosystem functioning were nearly four times more common than negative ones. In these studies, the amount of variation in ecosystem functioning explained by trait- based diversity was variable but rather small. Overall, this means trait-based diversity is often not a very good predic- tor of ecosystem functioning in freshwater lentic ecosys- tems. The capacity to improve our mechanistic understanding of biodiversity-ecosystem functioning rela- tionships in freshwater lentic ecosystems using trait infor- mation has not been fully exploited. Studies directly testing the underlying mechanistic rationale are required. Other ideas for further development in this field include studying diversity effects on the temporal stability of ecosystem functions, exploring multiple functions at a time (multi- functionality), focusing exclusively in cell size and shape as master traits and confirming the relative importance of individual trait variation for ecosystem functioning. REFERENCES Abonyi A, Horváth Z, Ptacnik R, 2017. Functional richness out- performs taxonomic richness in predicting ecosystem func- tioning in natural phytoplankton communities. Freshwater Biol. Doi: 10.1111/fwb.13051. [Epub ahead of print]. Balvanera P, Pfisterer AB, Buchmann N, He JS, Nakashizuka T, Raffaelli D, Schmid B, 2006. Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecol. Lett. 9:1146-1156. Behl S, Donval A, Stibor H, 2011. The relative importance of species diversity and functional group diversity on carbon uptake in phytoplankton communities. Limnol. Oceanogr. 56:683-694. Blomberg SP, Garland JrT, 2002. Tempo and mode in evolution: phylogenetic inertia, adaptation and comparative methods. J. Evol. Biol. 15:899-910. Borics G, Tóthmérész B, Lukács BA, Várbíró G, 2012. Func- tional groups of phytoplankton shaping diversity of shallow lake ecosystems. Hydrobiologia 698:251-262. Byrnes JEK, Gamfeldt L, Isbell F, Lefcheck JS, Griffin JN, Hec- tor A, Cardinale BJ, Hooper DU, Dee LE, Duffy E, 2014. Investigating the relationship between biodiversity and ecosystem multifunctionality: challenges and solutions. Methods Ecol. Evol. 5:111-124. Non -co mmerc ial us e o nly P. Venail186 Cardinale BJ, Duffy JE, Gonzalez A, Hooper DU, Perrings C, Venail P, Narwani A, Mace GM, Tilman D, Wardle DA, Kinzig AP, Daily GC, Loreau M, Grace JB, Larigauderie A, Srivastava DS, Naeem S, 2012. Biodiversity loss and its im- pact on humanity. Nature 486:59-67. Cardinale BJ, Gross K, Fritschie K, Flombaum P, Fox JW, Rixen C, van Ruijven J, Reich PB, Scherer-Lorenzen M, Wilsey BJ, 2013. Biodiversity simultaneously enhances the produc- tion and stability of community biomass, but the effects are independent. Ecology 94:1697-1707. Cardinale BJ, Matulich KL, Hooper DU, Byrnes JE, Duffy E, Gamfeldt L, Balvanera P, O’Connor MI, Gonzalez A, 2011. The functional role of producer diversity in ecosystems. Am. J. Botany 98:572-592. Cardinale BJ, Srivastava DS, Duffy JE, Wright JP, Downing AL, Samkaran M, Jouseau C, Cadotte MW, Carroll IT, Weis JJ, Hector A, Loreau M, 2009. Effects of biodiversity on the functioning of ecosystems. A summary of 164 experimental manipulations of species richness. Ecology 90:854. Duffy JE, Godwin CM, Cardinale BJ, 2017. Biodiversity effects in the wild are common and as strong as key drivers of pro- ductivity. Nature 549:261-264. Fernandez C, Cáceres EJ, Parodi ER, 2014. Phytoplankton de- velopment in a highly eutrophic man-made lake from the Pampa plain of Argentica - a functional approach. Int. J. En- viron. Res. 8:1-14. Finkel ZV, Beardall J, Flynn KJ, Quigg A, Rees TAV, Raven JA, 2010. Phytoplankton in a changing world: cell size and ele- mental stoichiometry. J. Plankt. Res. 32:119-137. Fontana S, Thomas MK, Moldoveanu M, Spaak P, Pomati F, 2017. Individual-level trait diversity predicts phytoplankton commu- nity properties better than species richness or evenness. ISME J. Doi: 10.1038/ismej.2017.160. [Epub ahead of print] Flynn DFB, Mirotchnick N, Jain M, Palmer MI, Naeem S, 2011. Functional and phylogenetic diversity as predictors of bio- diversity ecosystem-function relationships. Ecology 92: 1573-1581. Gagic V, Bartomeus I, Jonssin T, Taylor A, Winqvist C, Fischer C, Slade EM, Steffan-Dewenter I, Emmerson M, Potts SG, Tscharntke T, Weisser W, Bommarco R, 2015. Functional identity and diversity of animals predict ecosystem function- ing better than species-based indices. Proc. R. Soc. B 282: 20142620. Gross K, Cardinale BJ, Fox JW, Gonzalez A, Loreau M, Polley HW, Reich PB, van Ruijven J, 2014. Species richness and the temporal stability of biomass production: a new analysis of recent biodiversity experiments. Am. Nat. 183:1-12. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M, Naeem S, Schmid B, SetalaH, Symstad AJ, Vandermeer, Wardle DA, 2005 Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecol. Monog. 75:3-35. Krause S, Le Roux X, Niklaus PA, Van Bodegom PM, Lennon JT, Bertilsson S, Grossart H-P, Philippot L, Bodelier PL, 2014. Trait-based approaches for understanding microbial biodiver- sity and ecosystem functioning. Front. Microbiol. 5:251. Kruk C, Huszar VLM, Peeters ETHM, Bonilla S, Costa L, Lür- ling M, Reynolds C, Scheffer M, 2010. A morphological classification capturing functional variation in phytoplank- ton. Freshwater Biol. 55:614-627. Lefcheck JS, Byrnes JEK, Isbell F, Gamfeldt L, Griffin JN, Eisenhauer N, Hensel MJS, Hector A, Cardinale BJ, Duffy E, 2015. Biodiversity enhances ecosystem multifunctional- ity across trophic levels and habitats. Nature Comm. 6:6936. Litchman E, Klausmeier CA, 2008. Trait-based community ecology of phytoplankton. Annu. Rev. Ecol. Evol. Syst. 39:615-639. Losos JB, 2008. Phylogeneitc niche conservatism, phylogenetic signal and the relationship between phylogenetic relatedness and ecological similarity among species. Ecol. Lett. 11:995- 1007. Pälffy K, Présing M, Vörös L, 2013. Diversity patterns of trait- based phytoplankton functional groups in two basins of a large, shallow lake (Lake Balaton, Hungary) with different trophic state. Aquat. Ecol. 47:195-210. Reynolds CS, Huszar V, Kruk C, Naselli-Flores L, Melo S, 2002. Towards a functional classification of the freshwater phyto- plankton. J. Plankt. Res. 24:417-428. Santos AMC, Carneiro FM, Cianciaruso MV, 2014. Predicting productivity in tropical reservoirs: the roles of phytoplankton taxonomic and functional diversity. Ecol. Ind. 48:428-435. Schmidtke A, Gaedke U, Weithoff G, 2010. A mechanistic basis for underyielding in phytoplankton communities. Ecology 91:212-221. Shurin JB, Mandal S, Abbott RL, 2014. Trait diversity enhances yield in algal biofuel assemblages. J. Appl. Ecol. 51:603-611. Srivastava DS, Vellend M, 2005. Biodiversity-ecosystem func- tion research: is it relevant to conservation. Annu. Rev. Ecol. Evol. Syst. 36:267-294. Steudel B, Hallmann C, Loremz M, Abrahamczyk S, Prink K, Herrfurth C, Feussner I, Martini JWR, Kessler M, 2016. Contrasting biodiversity-ecosystem functioning relation- ships in phylogenetic and functional diversity. New Phytol- ogist 212:409-420. Tilman D, Reich PB, Knops JM, 2006. Biodiversity and ecosys- tem stability in a decade-long grassland experiment. Nature 441:629-632. Vogt RJ, Beisner BE, Prairie YT, 2010. Functional diversity is positively associated with biomass for lake diatoms. Fresh- water Biol. 55:1636-1646. Wiens JJ, Ackerly DD, Allenn AP, Anacker BL, Buckley LB, Cornell HV, Damschen EI, Davies TJ, Grytnes J-A, Harrison SP, Hawkins BA, Holt RD, McCain CM, Stephens PR, 2010. Niche conservatism as an emerging principle in ecol- ogy and conservation biology. Ecol. Lett. 13:1310-1324. Non -co mmerc ial us e o nly