Limnetica, 32 (2): 159-174 (2013). DOI: 10.23818/limn.32.14 Limnetica, 29 (2): x-xx (2011) c Asociación Ibérica de Limnología, Madrid. Spain. ISSN: 0213-8409

Biomonitoring for the 21st Century: new perspectives in an age of globalisation and emerging environmental threats

Guy Woodward1,ClareGray1,2 and Donald J. Baird3,∗ 1 Grand Challenges in and the Environment. Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, Berkshire SL5 7PY, UK. 2 School of Biological and Chemical Sciences, Queen Mary University of London, London E1 4NS.UK. 3 Environment Canada @ Canadian Institute, Department of Biology, University of New Brunswick, 10 Bailey Drive, PO Box 4400, Fredericton, NB, E3B 5A3, Canada.

∗ Corresponding author: [email protected] 2 Received: 14/1/13 Accepted: 25/9/13

ABSTRACT Biomonitoring for the 21st Century: new perspectives in an age of globalisation and emerging environmental threats As we move deeper into the Anthropocene, the scale and magnitude of existing and emerging anthropogenic threats to freshwater ecosystems become evermore apparent, yet we are still surprisingly poorly equipped to diagnose causes of adverse change in freshwater ecosystems. Our main aim in this perspectives and opinion piece is to suggest some new approaches to biomonitoring that could improve on the currently limited capabilities of existing schemes. We consider how biomonitoring might develop in the future as “Big Data” and next generation sequencing (NGS) approaches continue to revolutionize all branches of , with a particular emphasis on the need to consider not just nodes in the , but their interactions too, and also to look beyond our current reliance on the Latin binomial system of describing biological entities as “species”, when this concept is largely meaningless for many branches of the tree of life. We highlight the possible scope for enriching existing datasets to start assembling reasonable facsimiles of food webs, the need to collect and share more data more widely, and the value of metagenomics and metagenetics approaches to characterizing biodiversity in situ in a far more complete way than has been possible previously. Finally, we explore how these new approaches could provide a better marriage between structure and functioning than we have at present, but which is demanded increasingly by environmental legislation.

Key words: Biomonitoring, metagenomics, food webs, next generation sequencing, big data, functional genes, operational taxonomic units, multiple stressors, anthropogenic stress.

RESUMEN Biomonitoreo para el siglo 21: nuevas perspectivas en la era de la globalización y las amenazas ambientales emergentes A medida que avanzamos más en la Antropocena, la escala y la magnitud de las amenazas antrópicas actuales y futuras para los ecosistemas de agua dulce son más evidentes, sin embargo, seguimos estando sorprendentemente mal preparados para diagnosticar las causas de los impactos en estos ecosistemas. Nuestro principal objetivo en esta perspectiva y artículo de opinión es sugerir algunos nuevos enfoques para el control de la calidad biológica que podrían mejorar la limitada capacidad de los esquemas existentes. Consideramos como el biomonitoreo podría desarrollarse en el futuro como lo ha hecho el "Big Data" o la secuenciación de nueva generación (NGS), enfoques que continúan revolucionando todas las ramas de la ecología, con especial énfasis en la necesidad de tener en cuenta no sólo los nodos de la red trófica, sino también sus interacciones, y también para mirar más allá de nuestra actual dependencia de la nomenclatura binomial del latín para describir entidades biológicas como “especies”, a pesar de que este concepto carece en gran parte de sentido para muchas ramas del árbol de la vida. Destacamos el posible alcance que tendría, para enriquecer las bases de datos existentes, comenzar a ensamblar facsímiles razonables de redes tróficas, reunir y compartir datos de forma más amplia, y el valor de la metagenómica y metagenética para la caracterización de la biodiversidad in situ de una manera mucho más completa de lo que ha sido posible hasta ahora. Por último, se explora cómo estos nuevos enfoques podrían proporcionar una mejor conexión entre estructura y función de la que tenemos actualmente, y que cada vez más es exigida por la legislación ambiental.

Palabras clave: Biomonitoreo, metagenómica, redes tróficas, secuenciación de nueva generación, “big data”, genes fun- cionales, unidades taxonómicas operacionales, estresores múltiples, estrés antrópico.

14974 Limnetica 32(2), pàgina 159, 21/11/2013 160 Woodward et al.

INTRODUCTION Biomonitoring is a deeply conservative area of applied science, which has been slow to embrace Bioassessment and monitoring of freshwaters new theories and emerging technologies. It is has never been more important, as we move often based on obsolete ecological notions: the further into the so-called “Anthropocene” and widespread reliance on co-called its associated degradation of biological systems typologies, which are anachronistic relics of (Steffen et al., 2011). Freshwaters are especially the long-abandoned superorganism concept vulnerable to a host of anthropogenic stressors, (Clements, 1936), is one such case in point. some of which have been present for centuries It is also supported by a bewildering array of (e.g., organic pollution), whereas others are sampling and analytical methods (Friberg et more recent (e.g., acidification) or are only now al., 2011, Demars et al., 2012), most of which appearing on the horizon (e.g., nanoparticles). are essentially footnotes to the ’saprobic index’ Despite this evermore complex environmental developed over 100 years ago to classify organic context, biomonitoring still relies on pollution in central European rivers. Just as a techniques developed decades ago and remains mere four base pairs form the template for the focussed on a narrow range of stressors (see entire planet’s genetic biodiversity, the same reviews by Bonada et al., 2006, Friberg et al., basic biomonitoring template has been endlessly 2011) which are typically considered in isolation trawled to produce hundreds of biotic indices, (e.g. Extence, 1981, Bengtsson et al., 1986, all aimed at answering the question: “is a test Pestana et al., 2010), rather than recognising site in a natural state?”. There are now over that many may be operating simultaneously 300 such indices in use, many of which have and, potentially, synergistically (Ormerod et appeared in the last decade (Bonada et al., 2006). al., 2010). It is no surprise then that ecosystem Almost all of these are derived from changes in degradation is associated with multiple stressors, relative abundances of taxa within assemblages although it is often difficult or impossible to and largely, it is asserted, in response to ’organic gauge the degree to which each is responsible pollution’ and/or (e.g. Trophic because complex responses are possible – as Diatom Index, Kelly & Whitton, 1995; BMWP has been shown repeatedly in field studies and scores, Hawkes, 1998). experiments (e.g. Hughes & Connell, 1999, Culp Within this general biomonitoring frame- et al., 2005, Culp & Baird, 2006). This generally work there are two dominant approaches, the unacknowledged fact challenges biomonitoring “RIVPACS-style” (Wright, 2000) and “typology- scientists to develop solutions which can deal based” methods (Forbes & Richardson, 1913, with complex stressor situations, yet the standard Thienemann, 1920, Thienemann, 1959), with biomonitoring schema applied in most countries the latter underpinned by ideas that were jet- remains wedded to an historical solution to tisoned by mainstream ecology many decades a “single-stressor” phenomenon: point-source ago. The former is arguably more flexible and effluent. In addition, increasing global- more grounded in modern ecological theory, isation of trade and mass tourism has increased using continuous responses to assess community the connectedness of natural ecosystems to structure in response to environmental gradients, a level unprecedented in the post-glacial era rather than fixed categories designated apriori (Hulme, 2009), further complicating the task by “expert knowledge” (Friberg et al., 2011). of assessing, prioritising, and linking emerging In the last decade both approaches have been threats at local, regional and global scales. The extended beyond their original biogeographi- current set of assessment tools and techniques cal context, and are now being used to assess used in freshwater biomonitoring are inadequate stressors other than organic pollution, although for these increasingly complex tasks, and it is our such extrapolations are often applied without view that we must start to look to other approaches fully validating whether or not they work in their and emerging technologies to help fill the gaps. new contexts. Indeed in a recent comparative

14974 Limnetica 32(2), pàgina 160, 21/11/2013 Biomonitoring for the 21st Century 161

analysis, a typology-based approach performed identification, and patchy, inconsistent only marginally better than a null model resolution of those that are included. Moreover, (Armanini et al., in press). current routine biomonitoring programs often The general, almost universal, schema in lack true functional measures and integration current use has been described recently as with structural measures (but see also Environ- “Biomonitoring 1.0”, with calls for a radical, ment Canada/Govt of Alberta, 2012), despite rather than incremental, rethink to develop the ready availability of suitable approaches. “Biomonitoring 2.0” to deal with the realities of All biomonitoring schemes employ arbitrarily pollution in a 21st Century environment (Baird defined subsets of the community (e.g., di- & Hajibabaei, 2012). Here, we describe how atoms, macroinvertebrate families, chironomid emerging techniques in ecology and genomics exuviae, rooted macrophytes), which require can be harnessed to reboot biomonitoring sci- highly-skilled taxonomists to derive the basic ence, as we enter a new age of ecoinformatics, data, often in a painstakingly slow manner. All metagenomics, and other molecular-based such schemes also apply the taxonomy of their “omics” approaches. In doing this, we build chosen assemblage in a haphazard fashion, with on previous critiques (e.g. Woodward et al., the more easily identified taxa (often the larger, 2010a, Friberg et al., 2011, Demars et al., 2012), more charismatic taxa) being identified to a identifying current gaps in our knowledge, and higher level of resolution (e.g., Ephemeroptera / pointing the way towards how biomonitoring Plecoptera / Trichoptera [EPT] taxa described to might look in the future. species, genus or family) than the more obscure Key areas of biomonitoring science that are or challenging taxa (e.g. chironomids and many ripe for development include quantifying the “other Diptera” lumped into a single category). roles of multiple stressors and their potential In this way, significant amounts of data and use- synergies with one another, and considering ful information about the ecosystem are either true higher-level responses that link community discarded or neglected, for practical rather than structure to ecosystem functioning more ex- scientific reasons. This is especially true for the plicitly than is currently the case. In particular, more speciose groups of freshwater there is a need to integrate these approaches (e.g. bacteria, fungi, protists) at the base of the with molecular to assess both food web, where most of the biodiversity and, the true biodiversity and biogeochemistry of by extension the greatest potential range of ecosystem functioning, rather than relying on responses to stressors are located. Even lump- the more proxy measures involving a narrow ing chironomids together, as is done routinely range of taxa and single processes which have because they are difficult to identify using mi- become increasingly popular in recent years, croscopy, can reduce the power and sensitivity of such as leaf- decomposition assays (e.g., biomonitoring schemes considerably as species Hladyz et al., 2011a, Woodward et al., 2012b). within this group can have markedly different “Biomonitoring 1.0” has so far failed to deal responses to stressors (e.g. Ruse, 2010). These with these issues and, given its cumbersome problems with biases are further compounded and rigid framework, we argue that it is better when samples are compared through time or to start afresh, rather than attempt to modify across systems, as information is lost when taxa the existing system with innumerable minor ad- have to be aggregated due to inconsistencies in justments (e.g. Czerniawska-Kusza, 2005). New identification and/or to avoid the introduction of approaches, unencumbered with the traditions duplicated taxa (Friberg et al., 2011). and methodological and philosophical baggage Another particular drawback is the biogeo- of the past, offer far greater scope for the future. graphical constraint inherent in each scheme, A brief summary of the major failings of which is region-specific. Generality is therefore Biomonitoring 1.0 include a bias towards cer- lacking at large spatial scales, necessitating the tain taxa, necessitated by the practicalities of use of complex intercalibration measures. The

14974 Limnetica 32(2), pàgina 161, 21/11/2013 162 Woodward et al. resulting “metrics” that emerge after convoluted BIOMONITORING 1.5: INCORPORATION rounds of stratified sampling, subsampling, OF BIOGEOGRAPHY, GEOMATICS, data transformations, downweighting and other FUNCTIONAL TRAITS AND SPECIES statistical adjustments are so far removed from INTERACTIONS INTO BIOMONITORING the raw data as to be largely unintelligible as SCIENCE ecological descriptors, difficult to communicate to stakeholders, and render the highly-derived One way to resolve problems associated with metrics largely unusable for other purposes. biogeographical and methodological inconsisten- Sample properties are also constrained to a cies in data collection is through intercalibration, lowest common denominator (e.g. samples which is essentially back-calculating from ex- containing larvae which are too small to iden- isting methods to a supposedly lowest common tify reliably to the same resolution as in the denominator. This is far less justifiable or statis- other samples; 1-minute kick samples versus tically robust than using a common approach in 3-minute kick samples etc), leading to yet more the first place, and reflects more an appeasement information loss during attempts to standardise of different and often deeply-entrenched local the data across studies/regions (Friberg et al., historical traditions than it does a logical and 2011). Building a scientific framework on such scientific optimisation of effort (Woodward et weak foundations compromises the integrity of al., 2010b). Unfortunately, because so many “Biomonitoring 1.0”, until it is eventually no human and financial resources have already longer fit-for-purpose for gauging the emerging been devoted to achieve the current state-of-play suites of threats faced by freshwaters in the 21st in biomonitoring, there is huge inertia against Century. The question, then, is what might the changing the status quo, but this is a challenge alternative(s) look like? that cannot just be delayed ad infinitum as the We put forward some suggestions here as problems outlined above are not going to disappear to how we might develop new approaches by by simply ignoring them (Friberg et al., 2011). moving beyond the traditional scales and levels Several important steps that can be made to of organisation under investigation to develop help standardise approaches include considera- a more holistic perspective (Table 1). We em- tion of alternative descriptors of biological com- phasise, however, that we are not claiming these munity structure, such as biological traits rather to be either entirely exclusive or exhaustive than Linnean taxonomy (e.g. Townsend & Hil- alternatives to all other possible methods: rather, drew, 1994, Bonada et al., 2006, Menezes et al., our suggestion is they simply provide novel 2010, Culp et al., 2011). Focusing on traits of- ways of assessing stressors that go beyond the fers two key advantages: first, it removes much capabilities of current schemes. Some of these of the biogeographic constraints on data aggre- suggestions involve overhauling existing large gation, allowing a more universal approach to biomonitoring databases by enriching them with be developed in different parts of the world and, additional (often pre-existing) information on second, it facilitates future linkage to functional functional traits and species interactions, which measures. Biomonitoring in running waters is could be done relatively easily in the short-term, still dominated by structural rather than func- which we could call “Biomonitoring 1.5”. We tional measures, as it has been for over a century. also take a longer-term and more radical view in The question, though, is: which functional traits which powerful new molecular techniques could do we need to measure? Part of this requires at develop an entirely new way of assessing and least some ecological insight as to what might monitoring the biota: i.e. “Biomonitoring 2.0”. be important in relation to the stressor or inter- We will start with the former, shorter-term view and est (e.g. respiratory mode in systems suffering then extend our horizons by considering the latter. from oxygen sags).

14974 Limnetica 32(2), pàgina 162, 21/11/2013 Biomonitoring for the 21st Century 163 et et et et et al., et al., , et al., 2008, et al., et al., 1999; et al. 2010a, Cross 2011, Layer , 2012a. et al., 2010. 2012b et al., to the majority of 2010a, 2010b, et al., 2003, Bonada et al. 2011; Demars 2011a; Hladyz et al., 2005; Baird & et al., et al., 2008; Vila-Costa 2006. et al., , 1998; Rosi-Marshall & , 2001; Oberdorff et al., 2004; Riipinen 2011. et al., et Al., ,1999, Doledec et al., et al., et al., et al. et al. , 1985; Bunn et al. et al., et al., , 2011, Friberg et al. 2012. 2010; Morales & Holben. 2011. , 2011, Woodward , 2007, Menezes Example references Wallace, 2002, Layer et al. Yvon-Durocher Demars Sweeney, 2011; Hajibabaei, 2012; Yu al., al., Forbes & Richardson, 1913; Thienemann, 1920; Thienemann, 1959. Oberdorff Dangles Bott Young & Huryn, 1999; Uzarski 2001; Glud, 2008, Young Doledec Handelsman, 2004, Edwards & Rohwer, 2005; Tringe Frias-Lopez 2002; Pont al. 2010; Boyero 2011, Hladyz 2011b; Woodward 2000, Gayraud al. Pros y contras de los diferentes sistemas if such data are either typically collected or si no lo están, reconociendo al mismo tiempo þ ý ?Townsend ý ý ý ý ý þ þ þ Community completeness ere a que están incluidos la mayoría de taxones (microbianos y no ? ? ? ? ý ý þ þ ý ý þ þ Population abundances si esos datos están recogidos o con ? ý ý ý ý ý þ þ þ þ Resolved to species ? ? ? ? ? ? ý ý þ þ þ þ þ þ predictive ? ? ý ý þ þ þ þ þ þ cos, en lugar de centrarse en comunidades u otras agrupaciones. fi ý ý þ þ þ þ þ þ Structure Function A priori non-microbial) taxa being included, across trophic levels, rather than focusing on assemblages or other groupings. + Typical pros and cons of current and emerging biomonitoring approaches. These categories are represented with if not, whilst recognising there are exceptions. Where there is less of a clear distinction “?” is used to denote this. Community completeness referes ý (microbial actuales y emergentes de biomonitoreo. Estas categorías están representadas con Table 1. que hay excepciones. Cuandomicrobianos), no a hay través de una los certeza niveles clara, tró se utiliza “?”. Community completeness sefi re Typologies approaches RIVPACS-style approaches Litter breakdown Ecosystem metabolism Trait-based approaches Food webs Metagenomics / metagenetics Metatranscriptomics

14974 Limnetica 32(2), pàgina 163, 21/11/2013 164 Woodward et al.

One obvious property of species and individuals as well as the fluxes of and nutrients is their body size, which acts as a useful proxy through the web (e.g. Reuman & Cohen, 2005). for many other very important ecological traits The food web therefore forms a logical bridge and attributes, such as longevity, fecundity, that connects biodiversity to the ecosystem pro- biomass production, trophic status and nutrient cesses, goods, and services that are increasingly cycling rates (Woodward et al., 2005a). It is the focus of attention in , in also sensitive to many forms of disturbance, both aquatic and terrestrial systems (Reiss et as larger species tend to be most strongly neg- al., 2009, Mulder et al., 2012). Many of these atively impacted by a wide range of stressors are system-level attributes: for instance, com- (Raffaelli, 2004). Examples of stressors in which munity resilience, the ability of an ecosystem to the loss of larger organisms are widely reported sequester carbon, and the maintenance of viable include: drought, warming, acidification, and fisheries are just three examples of key services habitat loss and fragmentation (e.g. Layer et al., provided via the food web. 2010b, Layer et al., 2011, Yvon-Durocher et al., Many existing biomonitoring databases could 2011, Dossena et al., 2012, Ledger et al., 2012, be enriched easily with additional data, by infer- Woodward et al., 2012a). Organic pollution often ring typical body masses and diets from the liter- appears to have more subtle non-linear effects, ature, to construct relatively realistic food webs yet these still appear to be predictable: the (or subwebs of assemblages) from the data al- largest species often predominate in moderately- ready available. Although these would not be as enriched conditions, with smaller taxa at either highly resolved as some of the more quantita- extreme where nutrient limitation or toxic ef- tive food webs that have been described from re- fects come into play (Woodward et al., 2012b). peated and intensive studies conducted in a few Body size is not only a useful proxy for model systems (e.g. Cohen et al., 2003, Wood- many autecological traits, but it is also a key ward et al., 2005b, Gilljam et al., 2011), these po- determinant of the synecology of a species tential shortcomings would be counterbalanced within the food web, in terms of its number by the huge volume of data that would be gen- and strength of interactions and, by extension, erated. This would allow broad macroecological its role in the transfer of biomass, energy and responses to environmental gradients to be dis- key nutrients. The inclusion of data on size and cerned across many hundreds or thousands of trophic interactions therefore allows us to take a sites, adding a whole new dimension to existing more whole-system approach to understanding biomonitoring data. stressors and ecological responses to environ- Developing such an approach would represent mental change (e.g. Layer et al., 2010a, Ledger a considerable advance over the current meth- et al., 2012). Since it moves beyond characteris- ods that ignore trophic interactions, despite the ing the nodes to include more nuanced food-web fact that we know they are key to many higher- properties of relevance to bioassessment, ad- level responses to perturbations, such as trophic ditional high-level emergent phenomena and cascades and catastrophic secondary extinctions. functional attributes can be explored, such as the For instance, alternative stable states in shallow dynamic stability of the community and biomass represent a textbook example of how the flux through the ecosystem as a whole (Layer et food web can mould the community, via trophic al., 2010a, Mulder & Elser, 2010, Layer et al., feedbacks, even under otherwise identical envi- 2011, Ledger et al., 2012, Mulder et al., 2012). ronmental conditions (Jones & Sayer, 2003), and The patterning and strength of interactions are there are plenty of analogous examples from run- known from experiments and modelling to be ning waters (e.g. Power, 1990). Resilience, resis- key determinants of community stability and tance and persistence of communities are all the propensity of a system to lose species when functional attributes of ecosystems that emerge exposed to perturbations (e.g. Emmerson et al., from the coupling of pattern and process in the food 2005, Layer et al., 2010a, Layer et al., 2011), web, and therefore they cannot be fully understood

14974 Limnetica 32(2), pàgina 164, 21/11/2013 Biomonitoring for the 21st Century 165

from relative abundance data without information as envisaged here, could also provide important on trophic interactions. Since the food web explic- new insights by moving beyond the reliance on itly connects structure to function, it can provide coarse functional feeding groups (FFGs) which a new focus for the development of potentially are often used as dietary or trophic proxies, but more sensitive and relevant monitoring metrics. which are based on foraging modes rather than In simple operational terms, one way in diet per se (Woodward, 2009, Lauridsen et al., which this step towards a network-based 2012, Layer et al., 2013). approach could be achieved is to build a re- In addition to using empirical data from the gional database of pairwise trophic interactions literature to assign interactions in order to create recorded in the literature, which could then inferred food webs, ecological theory and math- be used to assign trophic interactions between ematical models could be used where there are species in local food webs where the nodes have gaps in the data. Some of these approaches have been identified. Since it would be impossible to been used to predict food web structure from construct directly-observed food webs in their first principles with a high degree of accuracy entirety for each individual system, due to the (e.g. > 90 % of links assigned correctly) in fresh- enormous amount of sampling effort required waters (e.g. the Allometric Diet Breadth Model, (Ings et al., 2009), a sensible balance should sensu Petchey et al. (2008), Woodward et al. be struck between resolution and volume of (2010a)). Finally, the use of advanced comput- data. Links could be assigned by a combination ing techniques, such as machine-learning, which of direct observation, extrapolation from other have recently been developed in disciplines as systems and/or based on theoretical predictions disparate as the social sciences and molecular bi- (e.g. via body size scaling relationships between ology, could also be applied to ecological data putative consumers and resources). It would to construct food webs in silico (Bohan et al., therefore be desirable to “ground truth” these 2011). Such artificial networks could be chal- synthetic food webs by comparing them with lenged and validated repeatedly with real data, a subset of those that have already been con- as is being done in studies on networks of gene structed directly and solely from exhaustive gut regulation (Walhout, 2011). This iterative feed- contents analysis, in which yield-effort curves back between data and models, each of which is for links and species are close to their respective revised and refined as the cycles proceed, could asymptotes (e.g. Ings et al., 2009, Woodward et offer a powerful and efficient new way of gen- al., 2010a). Further, the quality of the data could erating network data from simple species lists, be further enriched and refined iteratively by thereby adding entirely new dimensions to exist- ground-truthing new subsets of the database, for ing biomonitoring databases. instance by randomly selecting particular species and characterising their diet directly to compare with the inferred data. This would not only help BIOMONITORING 2.0: GENOMICS TO validate the accuracy of inferred dietary data THE RESCUE? but could feed back into the global database on pairwise trophic interactions to improve the Taxonomic identification is the sine qua non next round of inference, essentially applying of biomonitoring, and yet ironically has always a systems biology approach to network-level been a rate-limiting step, as taxonomists remain biomonitoring. Care would obviously need to in short supply (e.g. McClain, 2011). Moreover be taken to avoid introducing circularities in the procedure of separating and identifying the database, such that different sets are used specimens is expensive, and, coupled with time for testing and training, but this could easily be constraints, can hamper the scope of biomonitor- avoided by employing algorithms to screen for ing programs. This reliance on microscopy and such scenarios. The construction of comprehen- detailed taxonomic expertise has not changed for sive interdigitating species and links databases, decades, yet the recent advances in molecular

14974 Limnetica 32(2), pàgina 165, 21/11/2013 166 Woodward et al. ecology seem certain to challenge the current library-building across many taxonomic groups hegemony. These “big data” approaches orig- (e.g. Webb et al., 2012), the value of ecoinfor- inally emerged from microbial ecology as a matics data collected today will inevitably in- means of describing the massively diverse and crease over time, as we become better able to cryptic assemblages found in soils, but they assign taxonomic identities to currently cryptic are increasingly being applied to macrofauna “operational taxonomic units” (OTUs). This will (e.g. Hajibabaei et al., 2011). The field of allow us to return repeatedly to the same datasets metagenomics (aka environmental genomics, to identify new potential indicator species and re- ecogenomics, community genomics), which sponses to a wider array of stressors than may seeks to characterise the species complement be apparent at present. In the interim these OTUs from environmental samples, is revolutionising can still act as useful indicators even if we cannot molecular microbial ecology and its potential in yet put a species name to them, so long as they other fields is now starting to be recognised, in- have recognisable and characteristic signatures cluding the characterisation of food webs (Purdy and respond to environmental gradients. In fact, et al., 2010, Hajibabaei, 2012). Pyrosequencing it seems likely that metagenomics will force us and associated next-generation DNA sequencing to question our over-reliance on the Latin bino- (NGS) technologies can generate colossal quan- mial as a principal building block in community tities of data at a rapidly reducing cost – even ecology (Raffaelli, 2007), as the species concept to the stage that the Human Genome Project itself starts to break down in certain branches of (Collins et al., 2004) could now be replicated in the tree of life: for instance, many aquatic macro- a modest sized laboratory in a matter of weeks. phytes hybridise readily and have different chro- Another big step forward would be to mea- mosome numbers even within the same “species” sure gene expression in the metatranscriptome, (Lansdown, 2009). Remaining wedded to these enabling us to quantify genetic responses to stres- more traditional views of how we aggregate bi- sors across assemblages or functional groupings ological entities based primarily on morphology of species within the food web. Characterising rather than molecular data may be slowing our both the metagenome and the metatranscriptome progress in general ecology (Raffaelli, 2007). In- simultaneously could also be used to measure re- deed, there is an intriguing tension developing dundancy, and hence the potential for resilience, between different ways of perceiving the biota, in the system. Comparing RNA- versus DNA- with taxonomic-morphological, functional, and based metrics in response to stressors should tell molecular techniques all in current biomonitor- us both what is present and what is active within ing usage, albeit with the strongest emphasis still the metagenome (e.g. Mason et al., 2012). Some placed on the former. of these capabilities are already with us, whereas A major advantage of NGS is that biodiversity others are on the horizon – but given the rapid and and functional diversity can be measured simul- accelerating pace of development in the field they taneously in the same sample, by characterising seem sure to revolutionise our view of the natural both the community metagenome whilst also tar- environment and how we monitor it. geting key genes of functional importance that, in We should always be wary of excessive evan- turn, can be mapped onto ecosystem functioning gelising, but we have already entered a new era of (e.g., linking certain respiratory genes to whole unprecedented data generation in ecology, and if system metabolism, or particular enzymes to car- even a small portion of the full potential of NGS bon metabolism as could be revealed in pheno- is realised we will have made huge advances. type microarray plates to measure substrate util- We are close to having the capacity to carry out isation). This could provide a far better match true community-level analysis by including all between structural and functional measures than trophic levels, from single-celled prokaryotes at can be achieved using current methods, which of- the base of the food web to the largest metazoan ten rely on proxies for the latter (e.g. leaf-litter top predators. By allying metagenomics to DNA breakdown, BOD5 [Woodward et al., 2012b]).

14974 Limnetica 32(2), pàgina 166, 21/11/2013 Biomonitoring for the 21st Century 167

Indeed, recent research has suggested that research, and its application in biomonitoring many of the current functional approaches do not could drastically reduce costs while increasing the necessarily translate in a simple linear fashion to speed and coverage of bioassessment. However, a given environmental stressor, or to structural its misuse or premature application could create measures of “ecological integrity”. For example, an ecological Tower of Babel (Caterino et al., in a study of 100 European across a 2000), and thus we must explore this promis- 1,000-fold gradient in nutrient concentrations ing technique in the field of biomonitoring with a complex three-dimensional space-filling rela- some caution. What is clear is that the charac- tionship was evident between two nutrients and teristics of the data generated through NGS dis- leaf-litter decomposition rates (Woodward et al., covery of molecular taxonomic composition do 2012b): rates were always low at the extremes not always align directly with traditional biomon- of SRP and DIN concentrations, but low-to-high itoring methods (e.g. Hajibabaei et al., 2011) and rates were manifested in moderately enriched the genes employed may yet under-report bio- systems. In a subset of 10 % of these streams diversity in a sample (e.g. Tang et al., 2012). a unimodal relationship between breakdown Nevertheless, it is our view that these challenges and a classic BMWP (Biological Monitoring – inherent in any radically new technique – can be Working Party score) gradient in community overcome, and that in years to come, molecular structure in response to organic pollution was identification of biodiversity from environmental evident, with the abundance of large shredders samples will be routine in the near future. explaining most of the variance (i.e. once again highlighting the structural-functional role of body size). This shows that functional measures CONCLUSIONS on their own can be extremely difficult or im- possible to interpret (e.g. pristine low nutrient While ’Biomonitoring 1.0’ is practised in many waters had identical breakdown rates to those countries of the world, it is often viewed as that were grossly polluted), and that they provide a luxury rather than a necessary component complementary information to that supplied by of ecosystem monitoring and assessment. De- structural measures (Woodward et al., 2012b). spite its clear advantage over other assessment Another similarly large-scale study in a second approaches because it directly measures bi- set of 100 European streams, this time comparing otic responses, it remains stuck in the starting litter breakdown in streams with native riparian blocks, such that the patterns observed are vegetation with those that had human-altered often difficult to interpret, and at best are nu- riparian zones, found idiosyncratic responses to anced versions of simple binary outcomes (i.e. the perturbation: rates in the impacted sites could impacted/unimpacted). While multimetric ap- range from being slower, to equivalent, to faster proaches claim to be a step beyond simple binary than in the reference sites (Hladyz et al., 2011a). outcomes, these can be difficult to communicate The direction and magnitude of the response to stakeholders, and in many cases, illusory, appeared to be largely contingent on the type and their apparent complexity often masks an of alteration imposed and the biogeographical inability to truly capture diverse ecosystem setting. One of the more consistent effects that outcomes in the face of multiple stressors. This emerged, however, was that the contribution said, there are a number of significant new to total breakdown rates by macroinvertebrates ’Biomonitoring 1.5’ innovations which offer relative to microbes was generally lower in im- hope for the immediate future: the application of pacted systems – but what those microbes might null models, the development of stressor-specific be remains unknown, and in need of molecular metrics, the use of geomatics information in characterisation. habitat description and the development of data Molecular identification of biological species interoperability standards all can contribute to- could usher in a new golden age of ecological wards the development of more robust, scalable,

14974 Limnetica 32(2), pàgina 167, 21/11/2013 168 Woodward et al. broadly-applicable biomonitoring models with ACKNOWLEDGMENTS the possibility of diagnosis beyond pass/fail. However, it is towards the ’Biomonitoring 2.0’ This paper was based on presentations given by approaches that we must look for the devel- GW and DJB at the Limnologia 2012 Conference opment of truly diagnostic tools in the future in Guimarães, Portugal in July 2012. The authors (Table 1). While Biomonitoring 2.0 is a leap into gratefully acknowledge the Iberian Limnologi- the unknown in the best possible sense, we con- cal Association, and in particular, the conference cede that it is unlikely that these techniques can organisers Fernanda Cássio and Cláudia Pascoal be mainstreamed in ecosystem assessment in the for their hospitality and support. We would also short term (i.e. next 5 years) as they will require a like to thank Jessica Orlofske for useful com- longer lead-in time to achieve widespread accep- ments on an earlier draft of this paper. tance (5-10 years). The application of food web theory, coupled with high-throughput genomics and high performance computing systems for REFERENCES analysis of ’Big Data’ has the potential to gener- ate high-resolution biodiversity information for ARMANINI, D. G., W. A. MONK, L. CARTER, monitoring purposes on an unprecedented scale. D. COTE & D. J. BAIRD. (in press) Towards Such data offer an opportunity for ecosystem generalised reference condition models for envi- monitoring scientists to become engaged in ronmental assessment: a case study on rivers in a rapidly developing field, with Bio2 studies Atlantic Canada. and generating data with broad value beyond local Assessment. monitoring concerns, and could foster new col- BAIRD, D. J. & M. HAJIBABAEI. 2012. Biomoni- laborations between applied and basic research toring 2.0: a new paradigm in ecosystem assess- and between academic, government, industry ment made possible by next-generation DNA se- and other environmental stakeholders. Also, quencing. Molecular Ecology, 21: 2039–2044. the scale and extent of observations will allow BAIRD, D. J. & B. W. SWEENEY. 2011. Applying scientists to align them more closely with other DNA barcoding in benthology: the state of the sci- observation and prediction systems (i.e. remote ence. Journal of the North American Benthological Society, 30: 122–124. sensing; satellite observation, climate models), particularly in the light of new automated robotic BENGTSSON, G., T. GUNNARSSON & S. RUND- sampling systems (e.g. Harvey et al., 2012). GREN. 1986. Effects of metal pollution on the earth-worm Dendrobaena rubida (Sav.) in acidi- We are aware that advocating a move away fied soils. Water, Air, & Soil Pollution, 28: 361– from the reassuring traditional bedrock of 383. Linnean taxonomy based on individual type- BOHAN, D. A., G. CARON-LORMIER, S. MUG- specimens towards a gene sequence-based GLETON, A. RAYBOULD & A. TAMADDONI- unit of observation is not without some risks. NEZHAD. 2011. Automated Discovery of Food However, we believe that the potential benefits Webs from Ecological Data Using Logic-Based offered by this new paradigm could completely Machine Learning. Plos One, 6 (12): e29028. revolutionise the monitoring and management doi:10.1371/journal.pone.0029028. of environmental risks within just a few years. BONADA, N., S. DOLEDEC & B. STATZNER. The ability to monitor the ecosystem in toto 2007. Taxonomic and biological trait differences rather than as a series of isolated components, of macroinvertebrate communities between andtodothisinmuchgreaterdetail,with mediterranean and temperate regions: implications improved precision and accuracy, will allow us for future climatic scenarios. Global Change to move towards a truly generic approach in Biology, 13: 1658–1671. ecosystem science. This will greatly facilitate BONADA, N., N. PRAT, V. H. RESH & B. STATZ- globally-applicable and compatible observation, NER. 2006. Developments in modelling and prediction of natural ecosystems. biomonitoring: A comparative analysis of recent

14974 Limnetica 32(2), pàgina 168, 21/11/2013 Biomonitoring for the 21st Century 169

approaches. Annual Review of Entomology, 51: CULP, J. M. & D. J. BAIRD. 2006. Establishing 495–523. cause-effect relationships in multi-stressor envi- BOTT, T. L., J. T. BROCK, C. S. DUNN, R. J. NAI- ronments. . In: Methods in Stream Ecology. G. A. MAN, R. W. OVINK & R. C. PETERSEN. 1985. Lamberti & F. R. Hauer(eds): 835–854. Academic Benthic community metabolism in four temperate Press. stream systems: An inter- comparison and CULP, J. M., K. J. CASH, N. E. GLOZIER &D. J. evaluation of the continuum concept. Hydro- BAIRD. 2005. Insight into pollution effects in biologia, 123, 3–45. complex riverine habitats: a role for food web BOYERO, L., R. G. PEARSON, M. O. GESSNER, experiments. In: Dynamic Food Webs. P. De L. A. BARMUTA, V. FERREIRA, M. A. S. GRA- Ruiter, V. Wolters & J. Moore(eds): 354–368. CA, D. DUDGEON, A. J. BOULTON, M. CA- Academic Press. LLISTO, E. CHAUVET, J. E. HELSON, A. BRU- CULP, J. M., D. G. ARMANINI, M. J. DUNBAR, J. DER, R. J. ALBARINO, C. M. YULE, M. ARU- M. ORLOFSKE, N. L. POFF, A. I. POLLARD, NACHALAM, J. N. DAVIES, R. FIGUEROA, A. A. G. YATES & G. C. HOSE. 2011. Incorporating S. FLECKER, A. RARNIREZ, R. G. DEATH, T. traits in aquatic biomonitoring to enhance causal IWATA, J. M. MATHOOKO, C. MATHURIAU, diagnosis and prediction. Integrated environmental J. F. GONCALVES, M. S. MORETTI, T. JIN- assessment and management, 7: 187–197. GGUT, S. LAMOTHE, C. M’ERIMBA, L. RAT- CZERNIAWSKA-KUSZA, I. 2005. Comparing mod- NARAJAH, M. H. SCHINDLER, J. CASTELA, ified biological monitoring working party score L. M. BURIA, A. CORNEJO, V. D. VILLANUE- system and several bioiogical indices based on VA & D. C. WEST. 2011. A global experiment macroinvertebrates for assessment. suggests climate warming will not accelerate litter Limnologica, 35: 169–176. decomposition in streams but might reduce carbon DANGLES, O., M. O. GESSNER, F. GUEROLD & sequestration. Ecology Letters, 14: 289–294. E. CHAUVET. 2004. Impacts of stream acidifica- BUNN, S. E., P. M. DAVIES & T. D. MOSISCH. tion on litter breakdown: implications for assessing 1999. Ecosystem measures of river health and their ecosystem functioning. Journal of Applied Ecol- response to riparian and catchment degradation. ogy, 41: 365–378. , 41: 333–345. DEMARS, B. O. L., J. R. MANSON, J. S. OLAFS- CATERINO, M. S., S. CHO & F. A. H. SPERLING. SON, G. M. GISLASON, R. GUDMUNDSDOT- 2000. The current state of insect molecular system- TIR, G. WOODWARD, J. REISS, D. E. PICH- atics: A thriving Tower of Babel. Annual Review of LER, J. J. RASMUSSEN & N. FRIBERG. 2011. Entomology, 45: 1–54. Temperature and the metabolic balance of streams. CLEMENTS, F. E. 1936. Nature and Structure of the Freshwater Biology, 56: 1106–1121. Climax. Journal of Ecology, 24: 252–284. COHEN, J. E., T. JONSSON & S. R. CARPENTER. DEMARS, B. O. L., J. M. POTTS, M. TREMO- 2003. Ecological community description using the LIERES, G. THIEBAUT, N. GOUGELIN & V. food web, species abundance, and body size. Pro- NORDMANN. 2012. River macrophyte indices: ceedings of the National Academy of Sciences of not the Holy Grail! Freshwater Biology, 57: the United States of America, 100: 1781–1786. 1745–1759. COLLINS, F. S., E. S. LANDER, J. ROGERS, R. H. DOLEDEC, S., B. STATZNER & M. BOURNARD. WATERSTON & INT HUMAN GENOME SE- 1999. Species traits for future biomonitoring QUENCING, C. 2004. Finishing the euchromatic across : patterns along a human- sequence of the human genome. Nature, 431: 931– impacted river. Freshwater Biology, 42: 737–758. 945. DOLEDEC, S., J. M. OLIVIER & B. STATZNER. CROSS, W. F., C. V. BAXTER, K. C. DONNER, E. 2000. Accurate description of the abundance of J. ROSI-MARSHALL, T. A. KENNEDY, R. O. taxa and their biological traits in stream inverte- HALL, H. A. W. KELLY & R. S. ROGERS. 2011. brate communities: effects of taxonomic and spa- Ecosystem ecology meets adaptive management: tial resolution. Archiv Fur Hydrobiologie, 148: 25– food web response to a controlled flood on the 43. Colorado River, Glen Canyon. Ecological Appli- DOSSENA, M., G. YVON-DUROCHER, J. GREY, J. cations, 21: 2016–2033. M. MONTOYA, D. M. PERKINS, M. TRIMMER

14974 Limnetica 32(2), pàgina 169, 21/11/2013 170 Woodward et al.

& G. WOODWARD. 2012. Warming alters com- HAJIBABAEI, M., S. SHOKRALLA, X. ZHOU, G. munity size structure and ecosystem functioning. A. C. SINGER & D. J. BAIRD. 2011. Environ- Proceedings of the Royal Society B-Biological Sci- mental Barcoding: A Next-Generation Sequencing ences, 279: 3011–3019. Approach for Biomonitoring Applications Us- EDWARDS, R. A. & F. ROHWER. 2005. Viral meta- ing River . PLoS ONE, 6(4): e17497. genomics. Nature Reviews Microbiology, 3: 504– doi:10.1371/journal.pone.0017497 510. HANDELSMAN, J. 2004. Metagenomics: Applica- EMMERSON, M., J. MONTOYA & G. WOOD- tion of genomics to uncultured microorganisms. WARD. 2005. Body size, interaction strength and Microbiology and Molecular Biology Reviews, 68: food web dynamics. Dynamic Food Webs: Mul- 669–685. tispecies Assemblages, Ecosystem Development HARVEY, J. B. J., J. P. RYAN, R. MANN, C. M. and Environmental Change, 167–178. PRESTON, N. ALVARADO, C. A. SCHOLIN & EXTENCE, C. A. 1981. The effect of drought on ben- R. C. VRIJENHOEK. 2012. Robotic sampling, in thic invertebrate communities in a lowland river. situ monitoring and molecular detection of marine Hydrobiologia, 83: 217–224. . Journal of Experimental Marine Bi- FORBES, S. A. & R. E. RICHARDSON. 1913. Stud- ology and Ecology, 413: 60–70. ies on the Biology of the Upper Illinois River. Bul- HAWKES, H. A. 1998. Origin and development of letin of the Illinois State Laboratory of Natural the Biological Monitoring Working Party score History. Vol. IX, Art. X. system. Water Research, 32: 964–968. HLADYZ, S., K. ABJORNSSON, E. CHAUVET, FRIAS-LOPEZ, J., Y. SHI, G. W. TYSON, M. L. M. DOBSON, A. ELOSEGI, V. FERREIRA, T. COLEMAN, S. C. SCHUSTER, S. W. CHIS- FLEITUCH, M. O. GESSNER, P. S. GILLER, V. HOLM & E. F. DELONG. 2008. Microbial com- GULIS, S. A. HUTTON, J. O. LACOURSIERE, munity gene expression in ocean surface waters. S.LAMOTHE, A. LECERF, B. MALMQVIST, B. Proceedings of the National Academy of Sciences G. MCKIE, M. NISTORESCU, E. PREDA, M. P. of the United States of America, 105: 3805–3810. RIIPINEN, G. RISNOVEANU, M. SCHINDLER, FRIBERG, N., N. BONADA, D. C. BRADLEY, M. S. D. TIEGS, L. B. M. VOUGHT & G. WOOD- J. DUNBAR, F. K. EDWARDS, J. GREY, R. B. WARD. 2011a. Stream Ecosystem Functioning HAYES, A. G. HILDREW, N. LAMOUROUX, M. in an Agricultural Landscape: The Importance TRIMMER & G. WOODWARD. 2011. Biomon- of Terrestrial-Aquatic Linkages. Advances in itoring of Human Impacts in Freshwater Ecosys- Ecological Research, 44: 211–276. tems: The Good, the Bad and the Ugly. Advances HLADYZ, S., K. ABJORNSSON, P. S. GILLER & in Ecological Research, 44: 1–68. G. WOODWARD. 2011b. Impacts of an aggres- GAYRAUD, S., B. STATZNER, P. BADY, A. HAY- sive riparian invader on community structure and BACHP, F. SCHOLL, P. USSEGLIO-POLATERA ecosystem functioning in stream food webs. Jour- & M. BACCHI. 2003. Invertebrate traits for the nal of Applied Ecology, 48: 443–452. biomonitoring of large European rivers: an initial HUGHES, T. P. & J. H. CONNELL. 1999. Multiple assessment of alternative metrics. Freshwater Bi- stressors on coral reefs: A long-term perspective. ology, 48: 2045–2064. and Oceanography, 44: 932–940. GILLJAM, D., A. THIERRY, F. K. EDWARDS, D. HULME, P. E. 2009. Trade, transport and trouble: FIGUEROA, A. T. IBBOTSON, J. I. JONES, R. managing invasive species pathways in an era of B. LAURIDSEN, O. L. PETCHEY, G. WOOD- globalization. Journal of Applied Ecology, 46: 10– WARD & B. EBENMAN. 2011. Seeing Double: 18. Size-Based and Taxonomic Views of Food Web INGS, T. C., J. M. MONTOYA, J. BASCOMPTE, Structure. Advances in Ecological Research: The N. BLUTHGEN, L. BROWN, C. F. DORMANN, Role of Body Size in Multispecies Systems, 45: 67– F. EDWARDS, D. FIGUEROA, U. JACOB, J. I. 133. JONES, R. B. LAURIDSEN, M. E. LEDGER, H. GLUD, R. N. 2008. Oxygen dynamics of marine sed- M. LEWIS, J. M. OLESEN, F. J. F. VAN VEEN, iments. Research, 4: 243–289. P. H. WARREN & G. WOODWARD. 2009. Eco- HAJIBABAEI, M. 2012. The golden age of DNA logical networks – beyond food webs. Journal of metasystematics. Trends in genetics. 28: 535–537. Animal Ecology, 78: 253–269.

14974 Limnetica 32(2), pàgina 170, 21/11/2013 Biomonitoring for the 21st Century 171

JONES, J. I. & C. D. SAYER. 2003. Does the fish- MCCLAIN, C. 2011. Op-Ed: The mass extinction of invertebrate-periphyton cascade precipitate plant scientists who study species. In: Wired Science, loss in shallow lakes? Ecology, 84: 2155–2167. Wired Magazine. January 2011, Web http://www. KELLY, M. G. and WHITTON, B. A. 1995. The wired.com/wiredscience/2011/01/extinction-of-ta- Trophic Diatom Index - a new index for monitor- xonomists/. ing eutrophication in rivers. Journal of Applied MENEZES, S., D. J. BAIRD & A. SOARES. 2010. Phycology, 7: 433–444. Beyond taxonomy: a review of macroinvertebrate LANSDOWN, R. V. 2009. A Field Guide to the River- trait-based community descriptors as tools for ine Plants of Britain and Ireland. Including Se- freshwater biomonitoring. Journal of Applied lected Vascular Plants, Bryophytes, Lichens and Ecology, 47: 711–719. Alga. Ardeola Environmental Services, Stroud. MORALES, S. E. & W. E. HOLBEN. 2011. Linking LAURIDSEN, R. B., F. K. EDWARDS, M. J. bacterial identities and ecosystem processes: can BOWES, G. WOODWARD, A. G. HILDREW, A. ’omic’ analyses be more than the sum of their T. IBBOTSON & J. I. JONES. 2012. Consumer– parts? Fems Microbiology Ecology, 75: 2–16. resource elemental imbalances in a nutrient-rich MULDER, C. & J. J. ELSER. 2010. Soil pH, ecolog- stream. Freshwater Science, 31: 408–422. ical stoichiometry, and allometric scaling in soil biota. Available from Nature Proceedings: http:// LAYER, K., A. HILDREW, D. MONTEITH & G. dx.doi.org/10.1038/npre.2010.4217.1. WOODWARD. 2010a. Long-term variation in the MULDER, C., A. BOIT, S. MORI, J. ARIE VONK, littoral food web of an acidified mountain . S. D. DYER, L. FAGGIANO, S. GEISEN, A. Global Change Biology, 16: 3133–3143. L. GONZÁLEZ, M. KASPARI & S. LAVO- LAYER, K., J. O. RIEDE, A. G. HILDREW & G. REL. 2012. Distributional (In) Congruence of WOODWARD. 2010b. Food Web Structure and Biodiversity-Ecosystem Functioning. Advances in Stability in 20 Streams Across a Wide pH Gradi- Ecological Research, 46: 1–88. ent. Advances in Ecological Research: Ecological OBERDORFF, T., D. PONT, B. HUGUENY & D. Networks, 42: 265–299. CHESSEL. 2001. A probabilistic model character- LAYER, K., A. G. HILDREW, G. B. JENKINS, J. izing fish assemblages of French rivers: a frame- O.RIEDE,S.J.ROSSITER,C.R.TOWNSEND work for environmental assessment. Freshwater & G. WOODWARD. 2011. Long-Term Dynamics Biology, 46: 399–415. of a Well-Characterised Food Web: Four Decades OBERDORFF, T., D. PONT, B. HUGUENY, J. BEL- of Acidification and Recovery in the Broadstone LIARD, R. B. D. THOMAS & J. P. PORCHER. Stream Model System. Advances in Ecological Re- 2002. Development and validation of a fish-based search, 44: 69–117. index (FBI) for the assessment of “river health” in LAYER, K., A. G. HILDREW & G. WOODWARD. France. Bulletin Francais De La Peche Et De La 2013. Grazing and detritivory in 20 stream food Pisciculture, 365–366: 405–433. webs across a broad pH gradient. Oecologia, 171: ORMEROD, S. J., M. DOBSON, A. G. HILDREW 459–471. & C. R. TOWNSEND. 2010. Multiple stressors in LEDGER, M. E., L. E. BROWN, F. EDWARDS, A. freshwater ecosystems. Freshwater Biology, 55: 1–4. M. MILNER & G. WOODWARD. 2012. Drought PESTANA,J.L.T.,S.LOUREIRO,D.J.BAIRD& alters the structure and functioning of complex A. M. V. M. SOARES. 2010. Pesticide exposure food webs. Nature Climate Change. 3: 223–227. and inducible antipredator responses in the zoo- MASON, O. U., T. C. HAZEN, S. BORGLIN, P. S. grazer, magna Straus. Chemo- G. CHAIN, E. A. DUBINSKY, J. L. FORTNEY, sphere, 78, 241–248. J. HAN, H. Y. N. HOLMAN, J. HULTMAN, R. PETCHEY, O. L., A. P. BECKERMAN, J. O. RIEDE LAMENDELLA, R. MACKELPRANG, S. MAL- & P. H. WARREN. 2008. Size, foraging, and food FATTI, L. M. TOM, S. G. TRINGE, T. WOYKE, web structure. Proceedings of the National Aca- J. H. ZHOU, E. M. RUBIN & J. K. JANSSON. demy of Sciences of the United States of America, 2012. Metagenome, metatranscriptome and single- 105: 4191–4196. cell sequencing reveal microbial response to Deep- PONT, D., B. HUGUENY, U. BEIER, D. GOF- water Horizon . Isme Journal, 6: 1715– FAUX, A. MELCHER, R. NOBLE, C. ROGERS, 1727. N. ROSET & S. SCHMUTZ. 2006. Assessing

14974 Limnetica 32(2), pàgina 171, 21/11/2013 172 Woodward et al.

river biotic condition at a continental scale: a meiofauna. Proceedings of the National Academy European approach using functional metrics and of Sciences of the United States of America, 109: fish assemblages. Journal of Applied Ecology, 43: 16208–16212. 70–80. THIENEMANN, A. 1920. Die Grundlagen der POWER, M. E. 1990. Effects of in River Food Biocoenotik und Monards faunistische Prinzipien. Webs. Science, 250: 811–814. Festschrift für F. Zschokke Nr. 4: 6. PURDY, K. J., P. J. HURD, J. MOYA-LARANO, THIENEMANN, A. 1959. Erinnerungen und Tage- M. TRIMMER, B. B. OAKLEY & G. WOOD- buchblatter eines Biologe. Schweitzerbart, Stutt- WARD. 2010. Systems Biology for Ecology: gart. From Molecules to Ecosystems. Advances in TOWNSEND, C. R. & A. G. HILDREW. 1994. Ecological Research, 43: 87–149. Species traits in relation to a habitat templet for RAFFAELLI, D. 2004. How Extinction Patterns Af- river systems. Freshwater Biology, 31: 265–275. fect Ecosystems. Science, 306: 1141–1142. TOWNSEND, C. R., R. M. THOMPSON, A. R. RAFFAELLI, D. 2007. Food Webs, Body Size and MCINTOSH, C. KILROY, E. EDWARDS & M. the Curse of the Latin Binomial. In: From Ener- R. SCARSBROOK. 1998. Disturbance, resource getics to Ecosystems: The Dynamics and Structure supply, and food-web architecture in streams. of Ecological Systems. N. Rooney, K. S. McCann Ecology Letters, 1: 200–209. & D. L. G. Noakes (eds.): 53–64. Springer Nether- TRINGE, S. G., C. VON MERING, A. KOBAYASHI, lands. A. A. SALAMOV, K. CHEN, H. W. CHANG, REISS, J., J. R. BRIDLE, J. M. MONTOYA & G. M. PODAR, J. M. SHORT, E. J. MATHUR, J. WOODWARD. 2009. Emerging horizons in biodi- C. DETTER, P. BORK, P. HUGENHOLTZ & E. versity and ecosystem functioning research. Trends M. RUBIN. 2005. Comparative metagenomics of in Ecology & Evolution, 24: 505–514. microbial communities. Science, 308: 554–557. |rm REUMAN, D. C. & J. E. COHEN. 2005. Esti- UZARSKI, D. G., T. M. BURTON & C. A. STRI- mating relative energy fluxes using the food web, CKER. 2001. A new chamber design for measur- species abundance, and body size. Advances in ing community metabolism in a Michigan stream. Ecological Research, 36: 137–182. Hydrobiologia, 455: 137–155. RIIPINEN, M. P., T. FLEITUCH, S. HLADYZ, G. VILA-COSTA, M., J. M. RINTA-KANTO, S. L. WOODWARD, P. GILLER & M. DOBSON. SUN, S. SHARMA, R. PORETSKY & M. A. MO- 2010. Invertebrate community structure and eco- RAN. 2010. Transcriptomic analysis of a marine system functioning in European conifer plantation bacterial community enriched with dimethylsulfo- streams. Freshwater Biology, 55: 346–359. niopropionate. Isme Journal, 4: 1410–1420. ROSI-MARSHALL, E. J. & J. B. WALLACE. 2002. WALHOUT, A. J. M. 2011. Gene-Centered Regula- Invertebrate food webs along a stream resource tory Network Mapping. In: Caenorhabditis Ele- gradient. Freshwater Biology, 47: 129–141. gans: Molecular Genetics and Development, Sec- RUSE, L. 2010. Classification of nutrient impact on ond Edition. J. H. Rothman & A. Singson (eds.): lakes using the chironomid pupal exuvial tech- 271-288. Methods in Cell Biology. Elsevier Aca- nique. Ecological Indicators, 10: 594–601. demic Press Inc, San Diego. STEFFEN, W., Å. PERSSON, L. DEUTSCH, J. WEBB, J. M., L. M. JACOBUS, D. H. FUNK, X. ZALASIEWICZ, M. WILLIAMS, K. RICHARD- ZHOU, B. KONDRATIEFF, C. J. GERACI, R. E. SON, C. CRUMLEY, P. CRUTZEN, C. FOLKE, DEWALT, D. J. BAIRD, B. RICHARD, I. PHI- L. GORDON, M-MOLINA, V. RAMANATHAN, LLIPS & P. D. N. HEBERT. 2012. A DNA Bar- J. ROCKSTROM, M. SCHEFFER, H. J. code Library for North American Ephemeroptera: SCHELLNHUBER, & U. SVEDIN. 2011. The Progress and Prospects. Plos One, 7(5): e38063. Anthropocene: From global change to planetary doi:10.1371/journal.pone.0038063 stewardship. Ambio, 40: 739–761. WOODWARD, G. 2009. Biodiversity, ecosystem TANG, C. Q., F. LEASI, U. OBERTEGGER, A. functioning and food webs in fresh waters: assem- KIENEKE, T. G. BARRACLOUGH & D. bling the jigsaw puzzle. Freshwater Biology, 54: FONTANETO. 2012. The widely used small 2171–2187. subunit 18S rDNA molecule greatly underesti- WOODWARD, G., B. EBENMAN, M. EMMER- mates true diversity in biodiversity surveys of the SON, J. M. MONTOYA, J. M. OLESEN, A.

14974 Limnetica 32(2), pàgina 172, 21/11/2013 Biomonitoring for the 21st Century 173

VALIDO & P. H. WARREN. 2005a. Body size ZO, G. RISNOVEANU, M. SCHINDLER, in ecological networks. Trends in Ecology & VADINEANU, L. B. M. VOUGHT & E. CHAU- Evolution, 20, 402–409. VET. 2012b. Continental-Scale Effects of Nutrient WOODWARD, G., D. C. SPEIRS & A. G. HIL- Pollution on Stream Ecosystem Functioning. DREW. 2005b. Quantification and resolution of a Science, 336: 1438–1440. complex, size-structured food web. Advances in WRIGHT, J. F. 2000 An introduction to RIVPACS, Ecological Research, 36: 85–135. Freshwater Biological Association, The Ferry WOODWARD, G., G. PAPANTONIOU, F. ED- House, Far Sawrey, Ambleside, Cumbria, LA22 WARDS & R. B. LAURIDSEN. 2008. Trophic 0LP, UK. trickles and cascades in a complex food web: im- YOUNG, R. G. & A. D. HURYN. 1999. Effects of pacts of a keystone predator on stream community land use on stream metabolism and organic matter structure and ecosystem processes. Oikos, 117: turnover. Ecological Applications, 9: 1359–1376. 683–692. YOUNG, R. G., C. D. MATTHAEI & C. R. WOODWARD, G., J. BLANCHARD, R. B. LAU- TOWNSEND. 2008. Organic matter breakdown RIDSEN, F. K. EDWARDS, J. I. JONES,D. and ecosystem metabolism: functional indicators FIGUEROA, P. H. WARREN & O. L. PETCHEY. for assessing health. Journal of 2010a. Individual-Based Food Webs: Species the North American Benthological Society, 27: Identity, Body Size and Sampling Effects. 605–625. Advances in Ecological Research, 43: 211–266. YU, D. W., Y. Q. JI, B. C. EMERSON, X. Y. WANG, WOODWARD, G., N. FRIBERG & A. G. HIL- C. X. YE, C. Y. YANG & Z. L. DING. 2012. Bio- DREW. 2010b. Science and non-science in the diversity soup: metabarcoding of arthropods for biomonitoring and conservation of fresh waters. rapid biodiversity assessment and biomonitoring. In: Freshwater Ecosystems and Aquaculture Re- Methods in Ecology and Evolution, 3: 613–623. search. F. De Carlo and A. Bassano (eds.). Nova YVON-DUROCHER, G., A. P. ALLEN, J. M. Science Publishing. MONTOYA, M. TRIMMER & G. WOODWARD. WOODWARD, G., L. E. BROWN, F. K. ED- 2010a. The Temperature Dependence of the WARDS, L. N. HUDSON, A. M. MILNER, D. Carbon Cycle in Aquatic Ecosystems. Advances C. REUMAN & M. E. LEDGER. 2012a. Climate in Ecological Research, 43: 267–313. change impacts in multispecies systems: drought YVON-DUROCHER, G., J. I. JONES, M. TRIM- alters food web size structure in a field experiment. MER, G. WOODWARD & J. M. MONTOYA. Philosophical Transactions of the Royal Society 2010b. Warming alters the metabolic balance of B-Biological Sciences, 367: 2990–2997. ecosystems. Philosophical Transactions of the Royal WOODWARD, G., M. O. GESSNER, P. S. GILLER, Society B-Biological Sciences, 365: 2117–2126. V. GULIS, S. HLADYZ, A. LECERF, B. YVON-DUROCHER, G., J. M. MONTOYA, M. MALMQVIST, B. G. MCKIE, S. D. TIEGS, H. TRIMMER & G. WOODWARD. 2011. Warming CARISS, M. DOBSON, A. ELOSEGI, V. FER- alters the size spectrum and shifts the distribution REIRA, M. A. S. GRACA, T. FLEITUCH, J. of biomass in freshwater ecosystems. Global O. LACOURSIERE, M. NISTORESCU, J. PO- Change Biology, 17: 1681–1694.

14974 Limnetica 32(2), pàgina 173, 21/11/2013 14974 Limnetica 32(2), pàgina 174, 21/11/2013