A global synthesis of ecosystem services provided and disrupted by freshwater bivalve molluscs

ABSTRACT Identification of ecosystem services, i.e. the contributions that ecosystems make to human well‐being, has proven instrumental in galvanising public and political support for safeguarding biodiversity and its benefits to people. Here we synthesise the global evidence on ecosystem services provided and disrupted by freshwater bivalves, a heterogenous group of >1200 species, including some of the most threatened (in Unionida) and invasive (e.g. Dreissena polymorpha) taxa globally. Our systematic literature review resulted in a data set of 904 records from 69 countries relating to 24 classes of provisioning (N = 189), cultural (N = 491) and regulating (N = 224) services following the Common International Classification of Ecosystem Services (CICES). Prominent ecosystem services included (i) the provisioning of food, materials and medicinal products, (ii) knowledge acquisition (e.g. on water quality, past environments and historical societies), ornamental and other cultural contributions, and (iii) the filtration, sequestration, storage and/or transformation of biological and physico‐chemical water properties. About 9% of records provided evidence for the disruption rather than provision of ecosystem services. Synergies and trade‐offs of ecosystem services were observed. For instance, water filtration by freshwater bivalves can be beneficial for the cultural service ‘biomonitoring’, while negatively or positively affecting food consumption or human recreation. Our evidence base spanned a total of 91 genera and 191 species, dominated by Unionida (55% of records, 76% of species), Veneroida (21 and 9%, respectively; mainly Corbicula spp.) and Myoida (20 and 4%, respectively; mainly Dreissena spp.). About one third of records, predominantly from Europe and the Americas, related to species that were non‐native to the country of study. The majority of records originated from Asia (35%), with available evidence for 23 CICES classes, as well as Europe (29%) and North America (23%), where research was largely focused on ‘biomonitoring’. Whilst the earliest record (from 1949) originated from North America, since 2000, annual output of records has increased rapidly in Asia and Europe. Future research should focus on filling gaps in knowledge in lesser‐studied regions, including Africa and South America, and should look to provide a quantitative valuation of the socio‐economic costs and benefits of ecosystem services shaped by freshwater bivalves.


I. INTRODUCTION
The concept of Ecosystem Services (ESs), i.e. 'the contributions that ecosystems make to human well-being' (Haines-Young & Potschin, 2018, p. 9), has proven instrumental in galvanising public and political support for biodiversity conservation (MEA, 2005). Based on the latest Common International Classification of Ecosystem Services (CICES; Haines-Young & Potschin, 2018), ESs include the provisioning of material and energy needs (provisioning ESs), non-material characteristics of ecosystems that affect physical and mental states of people (cultural ESs), and regulation and maintenance of the environment for humans (regulating ESs).
Although the importance of ESs extends beyond their economic value, estimates indicate an ESs worth of €171,521 million in the European Union for 2012 (Vysna et al., 2021), at least US$ 250 billion/year for pollination services globally (IPBES, 2019), and an expected worldwide loss of US$ 9.87 trillion by 2050 due to environmental change (Roxburgh et al., 2020). Even though much progress has been made (e.g. Green & Elmberg, 2014;Rodrigues et al., 2020;Brock, Cini & Sumner, 2021), understanding how different animal groups contribute to ESs is far from complete, particularly with respect to freshwater invertebrate taxa (Collier, Probert & Jeffries, 2016;IPBES, 2019). Freshwater bivalves (FBs) are among the most abundant groups of invertebrates in freshwater ecosystems globally and can make up more than 90% of the benthic (i.e. bottom-dwelling) biomass (Okland, 1963;Sousa et al., 2008). FBs comprise >1200 species spread across taxonomically and biologically distinct groups, with most species (72%) belonging to the strictly freshwater order Unionida . Due to their (semi)infaunal, suspension-feeding life habit and complex life cycle, involving a parasitic larval stage, Unionida are particularly sensitive to anthropogenic habitat degradation and represent one of the most threatened taxonomic groups. Forty species are already presumed extinct IUCN, 2021), including eight species recently declared extinct by the U.S. Fish and Wildlife Service (The Center for Biological Diversity, 2021). By contrast, several notoriously invasive species are known within the Unionida (e.g. Sinanodonta woodiana) and more commonly, non-unionid FB orders, including the Ponto-Caspian myoid zebra and quagga mussels (Dreissena polymorpha, Dreissena bugensis), the veneroid Asian clam (Corbicula fluminea) and the mytiloid golden mussel (Limnoperna fortunei), native to East and Southeast Asia . These species often displace or replace native FB species in invaded regions, which has resulted in severe population losses of Unionida in, for example, the Laurentian Great Lakes and large stretches of European mainland river basins (Strayer, 1999;Sousa et al., 2014). Owing to their different life histories and ecology, invasive FBs often attain much greater biomass than native FBs and can cause severe ecological and economic damage (Strayer, 1999;Sousa et al., 2014;Haubrock et al., 2022).
Research and conservation of FBs is commonly justified by the important ecosystem functions and services that they provide . However, the available scientific evidence for FB-ES associations has never been comprehensively reviewed and systematised. Adopting the ES designations of the Millennium Ecosystem Assessment (2005), Vaughn (2018) identified three provisioning (i.e. food for humans, food for other species, products from shells), two cultural (i.e. cultural and existence value), one regulating (i.e. biofiltration) and four supporting ESs (i.e. nutrient cycling and storage, habitat/habitat modification, environmental monitoring, food webs) provided by FBs (see also Vaughn & Hoellein, 2018). Whilst these works represent an important first step towards assessing the ESs of FBs, they are not comprehensive, largely restricted to a single continent (North America) and taxonomic group (Unionida), and strongly focused on supporting services, which are not regarded as ESs per se by CICES (Haines-Young & Potschin, 2018). In addition, the available evidence on FBs disrupting rather than providing ESs has never been reviewed to date, although attempts have been made to quantify both ES provisions and disruptions, particularly of non-native FBs (Limburg et al., 2010).
A comprehensive, systematic review of the global scientific evidence is needed not only to provide a more complete understanding of the ESs provided and disrupted by FBs, but also to understand temporal and geographic trends across specific ESs and taxa, and identify gaps in our current knowledge and areas of importance for future research. Considering the ongoing and predicted future spread of non-native FBs in many regions of the world (Gallardo et al., 2018;Petsch et al., 2021), a better understanding of the differences and similarities in ESs provided and disrupted by native and nonnative species is therefore of particular urgency.
This study provides a systematic review of the available evidence for ESs associated with FBs. Its specific objectives were to (i) synthesise ESs that are associated with FBs; (ii) quantify the available evidence across temporal and geographic scales, taxonomic groups, native versus non-native species, and types of ESs; and, on that basis; (iii) derive novel insights into the importance of FBs to humans; and (iv) identify current shortcomings in our knowledge and recommend future directions for research.

II. LITERATURE SEARCH AND REVIEW
(1) Data collection Data on published evidence of associations between FBs and ESs were derived from a scientific literature search conducted in ISI Web of Science (ISI WOS; http:// webofknowledge.com/) and Scopus (https://www.scopus. com/) search engines in June 2020 and updated in March 2021, using an exhaustive compilation of search terms (Appendix S1). Search terms referring to FBs included common terms/names [e.g. 'freshwater' AND ('mussel' OR 'clam')], and scientific names of all known bivalve genera as well as species within predominantly marine genera that complete their life cycles in fresh water as provided by Graf & Cummings (2021). Search terms referring to ESs included general terms (e.g. 'environment* function*' OR 'ecosystem service') as well as terms referring to specific ESs that are potentially associated with FBs, including provisioning (e.g. 'food' OR 'material'), regulating (e.g. 'water quality' OR 'biological control') and cultural (e.g. 'ornamental' OR 'recreation') ESs (Haines-Young & Potschin, 2018). The search string started with the most general terms on FBs and ESs, followed by step-by-step addition of new, more specific terms. If the new search resulted in a higher number of outputs, the new term was retained, otherwise it was removed. Studies written in English, German, French, Italian, Romanian, Portuguese, Spanish, Czech and Slovakian, and published up to and including 31 December 2020 were considered.
Published studies retrieved via ISI WOS and Scopus searches were combined, and duplicates removed. Relevant records (i.e. pieces of evidence for an association between FB(s) and a specific ES) were identified by applying the following inclusion and exclusion criteria. In general, only records that reported primary evidence for an association between FBs (i.e. species/genera that complete their life cycles in fresh water) and ESs were retained. Therefore, records based on secondary evidence (e.g. in literature reviews or meta-analyses) were excluded to avoid doublecounting. Definition of ESs followed CICES (Haines-Young & Potschin, 2018) and thus, evidence for ecological functions of FBs without explicit implications on human well-being were excluded.
To allow reproducibility, ESs reported in the literature were classified following CICES (Haines-Young & Potschin, 2018) employing a three-level hierarchical system (Table 1). In addition, we extracted data on the geographic location (i.e. country/region and continent; Appendix S2) and taxonomic name(s) (i.e. order, genus and species) of FBs from each record where available. Based on that information, the FB taxon/taxa of each record was/were categorised as native or non-native to the country of study using information on native species distributions from Graf & Cummings (2021) and CABI (2021). Finally, information was extracted from each record on whether the FB(s) was/ were reported by the authors as a provider and/or disruptor, i.e. promoting and/or diminishing the quality or quantity of the respective ES.
(2) Data visualisation The number of records per year of publication was plotted as smoothing curves (averages for 2-year time periods) between 1949 (the first case study reported in our data set) and 31 December 2020. Sankey diagrams were generated to illustrate the relative quantity of and linkage among records on different ES categories based on different continents of study, status (native versus non-native), effect (provider versus disrupter) and/or taxonomic order of study species using the R-package 'd3Network' (Gandrud, 2015).

III. GENERAL DESCRIPTION OF THE DATA SET
A total of 6745 published studies were retrieved using our search terms. After the application of exclusion and inclusion criteria, the final data set comprised 684 studies (see Appendix S3 for a full list of studies) and 904 records (i.e. pieces of evidence for association between FBs and specific ES classes as defined in Table 1). In total, we identified evidence for 24 CICES classes of provisioning, cultural and regulating ESs that are associated with FBs (Table 1). A total of 146 studies provided evidence for more than one ES class, with up to five records being extracted from a single publication. Cultural ESs were the most represented (N = 491), followed by regulating (N = 224) and provisioning services (N = 189) ( Fig. 1A; Table 1). Evidence was strongly skewed towards FBs as providers (91%) rather than disrupters (9%), with evidence for disruption being largely restricted to provisioning (5%) and regulating ESs (4%; Fig. 1A).

Ecosystem services of freshwater bivalves
(2) Geographical trends Evidence was collected from a total of 69 countries, with 18% of records originating from the USA, followed by China (13%), France and Canada (each 5%). On the continent level, most records originated from Asia (35%), Europe (29%) and North America (23%), whilst records from Africa (2%), Australasia (3%) and South America (6%) were scarce (with 2% of records not providing this information; Figs 1A and 2A). Research has been strongly focused towards cultural ESs in Africa, Europe, North America, South America and Australasia (48-89% of overall records), whilst in Asia, provisioning, regulating and cultural ESs contributed to records in almost equal amounts ( Fig. 2A). Asia also provided the widest range of evidence, with available evidence for 23 of the 24 CICES classes (i.e. all except 'Social education'; Table 1) compared with lower coverage from Europe (19), North America (16), South America (14), Australasia (9) and Africa (7). Records from Asia, Africa and Australasia were almost exclusively focused on native species (>95%), whilst a considerable proportion of records from Europe, North America and South America related to non-native species (<52% natives) (Fig. 1A).

(3) Temporal trends
Annual publication of studies reporting FB-ES associations generally increased through the years, with the earliest found paper published in 1949 ( Fig. 2B,C). The first evidence for non-native species was published in 1980 (Fig. 2C). The rate of acquisition of evidence for non-native species increased markedly in the early 1990s and has been increasing relatively steadily since (Fig. 2C). The rate of record publication on native species and Unionida increased considerably in the mid 2000s (Fig. 2C). Since the turn of the century, annual record output remained relatively stable for the Myoida (dominated by usually non-native Dreissena spp.) and Sphaerioida, but increased notably for Veneroida (largely dominated by Corbicula spp.; Fig. 2C).
Whilst the first record from North America dates from 1949, first records from the other continents are more recent, spanning from 1977 (Asia) to 1982 (Europe), 1983 (Africa), 1989 (Australasia) and 1995 (South America) (Fig. 2B). Annual publication of ES records from North America has been relatively constant for the past 20+ years, while records from Europe have increased over the same period, particularly with regard to cultural ESs (Fig. 2B). Rate of publication for Asia has been increasing rapidly for the past ten years across all three ES sections (Fig. 2B).
(1) Nutrition Nutritional PESs fulfilled by FBs include the direct use of FBs captured in the wild (30%) or cultivated (8%) as a food source for humans, and the association of FBs with the production of other edible organisms supporting human diets (18%) ( Table 1).

(b) Food production
The effect of FBs on other edible organisms or other food can be beneficial as well as disruptive for human well-being. Examples for beneficial ES effects include the use of FBs, including non-native species, as chicken feeds ( as diffusers of antibiotic resistance in trout aquaculture (Sicuro et al., 2020), andin the form of shell powderas antimicrobial agents in soybean curd preservation . Disruptive effects in this respect mostly refer to non-native FBs, including the observed decrease of commercially harvested fish following introduction of Dreissena spp. to habitats in North America (Fera, Rennie & Dunlop, 2017;Hansen et al., 2020) and reduced productivity of shrimp farms in Colombia due to the invasive myoid Mytilopsis trautwineana (Aldridge et al., 2008).

(2) Materials
Records refer to the use of FBs or their parts in the production of other materials (11%) or tools (7%) without explicit or exclusive ornamental value (Table 1).

(a) Material production
Notable examples include studies investigating the potential application of FB shells or their derivatives as filling material in construction (Li et al., 2013b;Chakraborty et al., 2020), as adsorbents of dyestuffs (Figueiredo, Loureiro & Boaventura, 2005) or purifiers of peat water (i.e. via hydroxyapatite synthesised from Corbicula moltkiana shells; Alif, Aprillia & Arief, 2018). Evidence for actual use of FBs in material production is relatively scarce, with the use of nacre in button production arguably being the most significant, particularly with respect to historical communities (Haag, 2012;Strack, 2015;Sakalauskaite et al., 2019) but also including contemporary societies, such as fishermen in the Tocantins River, Brazil (Beasley, 2001).

(b) Tools
Evidence for the use of FBs as tools is completely restricted to historical communities, pertaining to the use of shells and pearls as, for example, knives, scrapers (e.g. to skin animal hides, scaling of fish, scraping plant material for string/ fishnet production) and tools for shaping pottery vessels during the Pleistocene and pre-historic times across all continents (Leechman, 1949;Jackson & Jackson, 2008;Romanus et al., 2008;Debruyne, 2010;Weston, Szab o & Stern, 2017;M arg arit, Mirea & Radu, 2018).

(3) Medicinal
Records pertaining to the use of FBs for medicinal purposes refer to genetic/protein information (15%) and biophysical products (10%) ( Table 1), and originate almost exclusively (89%) from Asia. A large number of studies in this category are based on C. fluminea from Eastern Asia (China, Taiwan, Japan), where this species has been used in Traditional Chinese Medicine and has been shown to exhibit antiinflammatory (Yeh et al., 2017), neuroprotective (Hsieh et al., 2017), hypotensive (i.e. reducing blood pressure; Tsai et al., 2006), hepato-protective and antioxidant effects (Bai et al., 2020), and the ability to reduce cholesterol levels (Chijimatsu et al., 2011) and improve wound healing (Peng et al., 2017). Some notable examples of medicinal ESs of other FB species include therapeutic effects of Pisidium coreanum and Lamellidens marginalis on bone diseases (osteoporosis and arthritis, respectively) (Chakraborty et al., 2010;Choi et al., 2019), and properties of S. cumingii concerned with wound repair (Dai et al., 2010) and immunoenhancement (Qiao et al., 2016). An example for the medicinal use of a biophysical FB product is the use of shells or their derivatives in dental implants (Wang et al., 2006;Zhu et al., 2011).

(4) Energy
Finally, two studies from Asia provided evidence for the potential use of FBs in energy provision, including the suitability of S. woodiana shells as an economic catalyst for biodiesel (Chinese tallow oil) production (Hu, Wang & Han, 2011).

V. CULTURAL ECOSYSTEM SERVICES
The vast majority of records for cultural ESs (CESs) are concerned with the acquisition of knowledge from FBs (using FBs as instruments rather than objects of research; 88%), whilst evidence for their relevance to aesthetics (6%), people's attitudes and other interactions (5%), and physical interactions (1%) is comparatively scarce (Fig. 1C, Table 1). Representation of non-native species was largely confined to knowledge acquisition, whilst evidence for FBs providing benefits in terms of attitudes, aesthetics and other interactions was almost exclusively based on native species (Fig. 1C).

(a) Biomonitoring
Due to their high filtration capacity and consequently, bioaccumulation rates, coupled with often high population densities that facilitate the collection of adequate replicate samples (Vaughn, 2018), FBs are popular organisms in biomonitoring studies that assess spatial and temporal patterns of water quality, including the location of pollution sources and quantification of the effects of remedial (e.g. establishment of a water treatment plant) or destructive actions (e.g. pesticide spill). Non-native species, in particular, D. polymorpha, D. bugensis and C. fluminea in Europe and North America, are commonly used in this context due to their ability to survive in poor water quality habitats .

(b) Palaeoenvironment
Due to their calcareous shells, FBs can also be used as archives of past environmental conditions (Vaughn, 2018). In their simplest form, these palaeoenvironmental reconstructions are based on data on the presence or relative abundances of FB species, which are used to infer the prevalent climatic, hydrological or other environmental conditions at the time and site. Many of these records refer to relatively recent, i.e. Holocene (0.01-0 Mya) and Pleistocene (2.6-0.01 Mya), deposits, including those from Turkey (Kuzucuo glu et al., 1997), Lithuania and Belarus (Sanko, Gaigalas & Yelovicheva, 2011), the Arabian Peninsula (Matter et al., 2015) and China (Noda et al., 2007). Other records refer to older strata, starting from the Pliocene [5-2.6 Mya; e.g. Omo-Turkana Basin, Kenya (Van Bocxlaer, 2020)], to the Miocene [23-5 Mya; e.g. Austria (Harzhauser & Tempfer, 2004) (Pierce & Constenius, 2001;Good, 2004;Montgomery & Barnes, 2012) and the British Isles (Andrews & Walton, 1990). In a few cases, authors have used intraspecific variation in morphological characteristics, such as the shape of the shell (Eagar, 1948) or relative growth rate (Black et al., 2010), to reconstruct palaeoenvironments. However, considering the incomplete understanding of how shell shape, size and sculpture is related to specific environmental conditions and functions in and across contemporary species (Zieritz & Aldridge, 2009;Levine, Hansen & Gerald, 2014), the value of fossilised specimens in this respect is currently restricted.
A rather widespread approach is the use of stable isotope ratios (mostly of oxygen and/or carbon) in (sub)fossil FB shells to reconstruct palaeoenvironmental (mostly palaeoclimatic) conditions. These usually refer to the Holocene, e.g. from Egypt (Hassan et al., 2012), Hungary (Schöll-Barna et al., 2012), Italy (Baroni et al., 2001), Argentina (Pérez et al., 2020) and the USA (Tevesz et al., 1997;Yu et al., 2008), and less frequently, to earlier epochs (e.g. Eocene USA; Buskirk et al., 2016). The regularity of annual shell formation of FBs in non-tropical regions, leading to the formation of visible annual shell rings, further provides the opportunity to evaluate changes in e.g. isotope composition at high temporal (i.e. annual or even intra-annual seasonal) resolution. Such a 'sclerochronological' approach has been applied on (sub)fossil shells, e.g. from Holocene Syria (Çakirlar & Şeşen, 2013), Holocene Turkey (Lewis et al., 2017) and Miocene Amazon (Kaandorp, Wesselingh & Vonhof, 2006), but also on contemporary specimens, such as Margaritifera margaritifera, which can attain lifespans of >200 years (Schöne et al., 2020). In a combined sclerochronological analysis of archaeological, recent and contemporary FB shells (Amblema plicata and Quadrula quadrula), Fritts et al. (2017) constructed an environmental archive of the Illinois River, USA, spanning the past 1000 years.
Evidence of the use of FBs in palaeoenvironmental reconstructions via other indicators includes the reconstruction of past metal concentrations (Binkowski et al., 2019), and the reconstruction of palaeobasins through phylogeographical analyses (Hewitt et al., 2018).

(c) Archaeology
Evidence for FBs providing information on historical human societies and communities is based on the presence, abundance or isotopic signature of species/specimens. Several studies provide evidence for the historic use of FBs as food, tools, ornaments and/or objects of spiritual, symbolic or Biological Reviews 97 (2022) (Li et al., 2013a); Unio sp. in Eneolithic Romania (M arg arit, 2020)]. In other cases, information on other aspects of past human societies are provided through FBs. For example, the isotopic signature (carbon and oxygen) of ornamental disc beads made from Unio shells from several Neolithic burials in Poland revealed that these originated from both riverine and lacustrine environments, which suggests changes in the shell source over time and thereby provided evidence for the existence of a regional exchange network (Apolinarska & Kurzawska, 2020). In Burleigh (1983), an Egyptian tomb was dated to 5000 ya using radiocarbon data from an Etheria elliptica shell found in the tomb. Finally, a number of studies utilise palaeoenvironmental and palaeoclimatic information reconstructed through stable isotope data from FBs to infer information on ancient human ecologies. An example is the study by Çakirlar & Şeşen (2013), who used sclerochronological isotope data from archaeological Unio sp. shells to show that continuous perennial flow of the Jaghjagh River, Syria, throughout the transition from the third to the second millennium BC was responsible for the continuous occupation of the Tell Mozan.

(d) Biomimicry
Records providing evidence for the use of FBs in biomimicry (see Table 1 for a description of the term) were scarce, restricted to Asia and focused on improving understanding of the mechanical behaviour of nacre for the design of nacreinspired synthetic materials (Jiao et al., 2019;Liu et al., 2020).

(e) Social education
Evidence for the value of FBs in social education in our database is confined to only two records, including a collaborative FB translocation initiative between the Matauranga Maori tribe (New Zealand) and western scientists, which aimed not only to renature this freshwater ecosystem but also to reconnect this indigenous tribe with their land and water (Michel et al., 2019).

(2) Attitudes and other interactions
Records relate to a symbolic, spiritual or religious meaning of FBs (3%) and their importance for the cultural heritage or traditions of human communities (2%).

(a) Spiritual, symbolic and religious
The majority of these records relate to FB shells and pearls, mostly from the order Unionida, as burial objects used by ancient human communities in e.g. Egypt (Burleigh, 1983), Mongolia (Kiryushin et al., 2011), Russia (Korolev et al., 2018) and Argentina (Fabra, Gordillo & Piovano, 2012). More recently, from the 18th to the early 20th centuries, harvested freshwater pearls, particularly from M. margaritifera, have been used as decoration on various objects of religious significance (Strack, 2015).

(b) Tradition
Evidence on the relevance of FBs to cultural heritage and traditions in our database are restricted mainly to their use in traditional foods and particularly that of Corbicula spp. in traditional Asian cuisine, including soups [e.g. in China and Japan (Ke et al., 2011;Horikoshi, 2020)] or as smoked etok salai in Malaysia (Rak et al., 2020a). FBs are also considered traditional ('first') foods of indigenous tribes in North America, Australia and New Zealand (Brim Box et al., 2006;Noble et al., 2016).

(3) Aesthetic experiences
Evidence for FBs providing aesthetic experiences to humans is restricted to the ornamental use of living FBs, their shells or pearls (Fig. 3B). Many of these records refer to archaeological FBs (Burleigh, 1983;Apolinarska & Kurzawska, 2020;M arg arit, 2020). In North America and Europe, intense 'pearl fishing' over the past century has led to a steep decline and even extirpations of unionid populations but is now largely banned (Bauer, 1988;Anthony & Downing, 2001;Strack, 2015). Freshwater pearl culture for ornamental purposes is an important economy especially in China and other parts of Asia (Janakiram, 2003;Fiske & Shepherd, 2007), with considerable ongoing research efforts aimed at optimising the culture of high-quality pearls and nacre . Finally, live FBs are commonly used in the ornamental pet trade (Erdo gan & Erdo gan, 2015; Ng et al., 2016).

(4) Physical interactions
Evidence for FBs affecting human physical interactions is restricted to leisure and recreational activities. Most of these records refer to the effects of non-native species, showing both their roles as providers (e.g. D. polymorpha increasing the recreational value of an English lake by improving water clarity; Mansfield et al., 2014) as well as disrupters of these ESs (e.g. C. fluminea shells after massive die-off events reducing the recreational value of beaches; Fig. 3F; Ilarri et al., 2011).

VI. REGULATING ECOSYSTEM SERVICES
Evidence for associations between FBs and regulating ESs (RESs) concerns the regulation of organisms, i.e. algae, bacteria and/or diseases (39% of RES records), mediation of human inputs, such as sequestration or decomposition of harmful wastes of human origin (34%), and physico-chemical regulation, including water clearance (27%). RES records were disproportionately commonly associated with nonnative species (compare Fig. 1D with Fig. 1A-C), particularly D. polymorpha and D. bugensis in North America and Europe, and L. fortunei in South America. Evidence for disruption of RESs was predominantly based on non-native rather than native species (81 versus 19% of records). Different taxonomic orders contributed evidence to each of the three CICES Biological Reviews 97 (2022)  Ecosystem services of freshwater bivalves groups unequally, with Myoida and Unionida contributing particularly to biological and physico-chemical regulation, and Veneroida and Unionida to the mediation of human inputs (Fig. 1D).
(1) Regulation of organisms Records relate mostly to FBs affecting the prevalence and concentration levels of algae in the water (28%; Fig. 3D) and, to a lesser extent, those of bacteria (8%) and diseases (4%), thereby affecting human health, security and/or amenity value of the respective waterbodies. Evidence for an effect on humans of regulation of organisms by FBs is available from all continents apart from Africa, and is based on both native and non-native species.

(a) Algae
Evidence is exclusively based on the effects of (selective) biofiltration by FBs of phytoplankton, which is often accompanied by an effect on water clarity (Fig. 3D). Many of these studies are based on laboratory experiments that investigate the ability and rate at which FBs can ingest (and, in some studies digest) and thus, remove (certain compartments of) phytoplankton from the water column, including harmful derivatives, such as hepatotoxic microcystins and nodularins (Pham et al., 2016;Buelow & Waltham, 2020;Silva et al., 2020). A number of studies provide evidence for these mechanisms in situ, often through mesocosm experiments. Notable examples on the provision of this ES by native species include the alleviation of the negative effects of cyanobacterial blooms on submerged macrophyte growth by the unionid S. cumingii in Lake Taihu, China (He et al., 2014), and effective bioaccumulation of microcystins and/or nodularins by unionid species in Japanese and Latvian lakes (Park, Yokoyama & Okino, 2001;Barda et al., 2015). Much of the respective literature from Europe and North America is focused on non-native, invasive Dreissena species, which have caused reductions and, in some cases, extirpation of local, native unionid populations (Strayer, 1999;Sousa et al., 2014). The prevailing evidence indicates that this shift in FB species composition results in a reduction of overall phytoplankton biomass, which is, however often accompanied by an increase in nuisance algae, such as the green alga Cladophora glomerata (e.g. Lake Ontario, Canada; Ozersky et al., 2009), and/or harmful cyanobacteria, such as Microcystis spp. (e.g. Laurentian Great Lakes; Vanderploeg et al., 2001). The negative knock-on effects on human health have been illustrated by Jones (2019), who showed that a steep drop in microcystin levels caused by a sudden, massive die-off of D. polymorpha in Gull Lake, Michigan, USA, led to improved infant health (measured as e.g. instances of low birth weight, length of gestation). That said, the effects of dreissenid invasions appear to depend on the prevailing conditions, including phosphorus concentrations, at the site (Raikow et al., 2004). For example, data from Lake Michigan, USA, indicate that invasion by D. polymorpha has led not only to an increase of cyanobacteria but also to increased overall phytoplankton density (De Stasio et al., 2008). On the other hand, D. bugensis has been shown to control phytoplankton, including cyanobacteria, in urban ponds in the Netherlands (Waajen et al., 2016).

(b) Bacteria
Scientific interest in the potential use of FBs as a bioremediation tool for bacterial contamination of water bodies has gained traction only relatively recently. One of the first studies providing strong evidence for FBs significantly reducing concentrations of bacteria of human importance is that by Bianchi et al. (2014), showing that the unionid Diplodon chilensis can reduce bacterial loads in sewage water from a Patagonian lake. In situ experiments by Ismail et al. (2015Ismail et al. ( , 2016 revealed the ability of the native unionid Anodonta californiensis and non-native veneroid C. fluminea to reduce concentrations of Escherichia coli in human-impacted lakes and rivers in California, USA. Laboratory experiments in Italy provided similar results for the nonnative D. polymorpha, which almost completely removed E. coli from the water (Mezzanotte et al., 2016). Also in Italy, mesocosm experiments showed that integration of the non-native unionid S. woodiana can reduce the environmental impacts of inland trout farming in terms of reducing total bacterial concentrations by up to 72% (Sicuro et al., 2020).

(c) Diseases
Evidence for FBs as potential bioremediation tools for human pathogenic viruses and bacteria and their vectors originated mainly from Asia and Europe. Laboratory experiments in Italy showed that non-native D. polymorpha can significantly reduce concentrations of rota-and polioviruses in water (Mezzanotte et al., 2016). In Japan, the native, pearlproducing unionid Sinohyriopsis schlegelii was found effectively to deplete oocysts of gastroenteritis-inducing Cryptosporidium parvum in the final settling pond of a sewage plant (Izumi et al., 2012). A study from Indonesia showed that FB shell powder can be used as an environmentally friendly alternative to organophosphate pesticides in controlling larvae of Anopheles and Aedes aegypti mosquitos, which are vectors of malaria and dengue fever (Sorontou & Agussalim., 2016).

(2) Mediation of human inputs
In the vast majority of records on FBs acting as remediators of harmful organic or inorganic substances of human origin, evidence relates to filtration and/or sequestration of these substances (32%), whilst evidence for (subsequent) transformation or decomposition into less harmful substances is rare (2%). L. marginalis, which was shown to significantly alleviate Cdcontamination in waters receiving industrial effluents (Jana & Das, 1997;Das & Jana, 2003). As FBs are also readily eaten in this region and thus can pose a health hazard, the findings of Cd-biosorptive properties of FB shells (Ismail, Aris & Latif, 2014a;Hossain, Bhattacharyya & Aditya, 2015) and applied by Ghosh et al. (2020b) are particularly promising.
In Europe, North America and South America, research on the potential application of FBs as bioremediation tools for anthropogenic pollution of fresh waters is largely focused on non-native species. In the USA, a flow-through mussel filter designed by Diggins et al. (2002) using D. polymorpha and D. bugensis was found to clear up to 96% of suspended particles (to which many pollutants readily adsorb) from effluents before discharge. A similar design with 40,000 D. polymorpha individuals was piloted in the largest wastewater treatment plant of Milan, Italy, resulting in the efficient removal of various heavy metals as well as pharmaceuticals and drugs of abuse ( Fig. 3E; Binelli et al., 2014;Magni et al., 2015). In Portugal, C. fluminea was shown to be able to assist in the remediation of acid mine drainage, as well as olive oil mill and winery wastewaters (Rosa et al., 2014;Pipolo et al., 2017;Ferreira et al., 2018;Domingues et al., 2020). One of the few such records on native unionid species from outside Asia is that of D. chilensis alleviating environmental effects of a fish aquaculture in Chile by removing nutrients and organic matter from the water (Parada et al., 2008).

(b) Transformation
Evidence for the ability of FBs to transform substances of human origin is scarce and comes mainly from L. fortunei. In its native China, laboratory experiments showed that both living specimens and, to a lesser extent, shells, can reduce the concentrations of the plant growth regulator forchlorfenuron in a process involving bacteria-associated nitrification and denitrification reactions (Zhang, Cui & Huang, 2015). In laboratory experiments in Argentina, where L. fortunei is invasive, herbicide (including glyphosate) and pesticide concentrations in the water decreased by 40% under live L. fortunei presence and by 25% in empty shell treatments, again involving transformation due to mineralisation by microbial communities (Di Fiori et al., 2012).

(3) Physico-chemical regulation
Evidence is available from all continents except South America for FBs decreasing or increasing concentrations of organic or inorganic substances (of non-human origin) (17%) as well as affecting the physical properties of water quality (11%) with implications for human health and/or recreation.

(a) Storage and excretion
Particularly in China, there has been growing interest in using native unionid FBs in the restoration of freshwater ecosystems. L. Wang et al. (2017) showed that when combined with stocking of planktivorous fish and replanting submerged macrophytes, S. cumingii can effectively remove nutrients from eutrophicated waterbodies (as well as supress phytoplankton growth). Similarly, Lamprotula leai in combination with the annelid Tubifex tubifex has been demonstrated to be effective in promoting plant growth and nutrient absorption within 'Constructed Wetlands', an ecological engineering solution for removing and transforming pollutants from wastewater (Kang et al., 2018). In North America and Europe, evidence comes almost exclusively from non-native species and predominantly D. polymorpha and D. bugensis. Research has focussed mainly on their effects on nutrient cycling (especially of phosphorus and nitrogen), which are complex and not uniform across ecosystems, rendering these invaders both providers and disruptors of this RES. In North America, a considerable body of research has been conducted on this topic in the Laurentian Great Lakes. In Lake Huron, it was shown that sequestration of phosphorus by D. polymorpha and D. bugensis led to reduced primary productivity with knock-on effects on secondary producers and fish (Cha et al., 2011). In Lake Erie, modelling showed that the invasion of these FBs decreased ecosystem resistance to eutrophication, necessitating increased phosphorus management to preserve lake ESs (Roy et al., 2010). In Lake Simcoe, dreissenid FBs have been shown to remineralise and thus increase bioavailability of phosphorus and nitrogen, whilst their shells represent a longterm sink for phosphorus, nitrogen and calcium (Ozersky, Evans & Ginn, 2015). On the other hand, decline in primary production and increase in water clarity caused by a dreissenid invasion has resulted in the Great Lakes becoming a significant CO 2 net emitter (>7.7 Tg-C annually) (Lin & Guo, 2016). Despite their disruptive effects on RESs, cultivation of D. polymorpha has been suggested as a tool for managing nutrient-enriched reservoirs in the UK (McLaughlan & Aldridge, 2013) and has successfully reduced nutrient exports from the Szczecin/Oder Lagoon to the open Baltic by up to 3500 t N and 420 t P per year (Schernewski, Stybel & Neumann, 2012;Friedland et al., 2019).

(b) Water clearance
Evidence for FBs improving clarity and other physical properties of water quality with implications for humans (e.g. in terms of health or recreational activities) is mostly associated with the regulation of organisms, e.g. phytoplankton, sequestration and/or transformation of substances of human origin, and/or removal of naturally occurring substances (Holland, 1993;Schernewski et al., 2012;Waajen et al., 2016;. Many of these studies concern non-native dreissenids in North America and Europe, and whilst their effects on increased water clarity are generally regarded as an ES provision, they have also been shown to disrupt ESs provided by the physical properties of water, e.g. by causing unpleasant odour and taste due to increased cyanobacterial growth in the St. Lawrence River, Canada (Watson & Ridal, 2004

VII. DISCUSSION (1) Temporal and geographic trends
Our systematic review of ESs provided and disrupted by FBs has identified a number of patterns over time and space. Our data set suggests that occasional scientific interest on FB-ES associations from the late 1940s to the 1980s was followed by an exponential increase in attention. This trend coincides with increased global interest in ecology and biodiversity conservation since the 1980s (Stork & Astrin, 2014), with particular interest focussed on non-native species since the 1990s (Vaz et al., 2017). The rapidly increasing rate of annual record publication on FB-ESs in the 2000s may be explained by the popularisation of the ES concept, especially following publication of the Millennium Ecosystem Assessment (MEA, 2005) and subsequent initiatives, such as The Economics of Ecosystems and Biodiversity, the Ecosystem Services Partnership and the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services.
Despite the general increase in publications over time, the rate and amount of evidence published has not been homogenous across geographical areas. Asia stands out in our data set as the current epicentre of FB-ES research in terms of the quantity of available evidence (35% of records originate from Asia), breadth of research (23 of 24 CICES-classes) and increase in annual evidence output in recent years. Four CICES classes were based exclusively on evidence from Asia, i.e. the use of genetic or protein information for medicinal purposes, use in the provision of energy, use in the design and production of other materials, and relevance to cultural heritage and traditions (although evidence for latter class also exists at least for North America and Australasia; Noble et al., 2016). Other CICES classes and groups were strongly dominated by evidence from Asia, and particularly East Asia, including the use of FBs or their products for food and material production, and the restoration of freshwater ecosystems. If anything, this dominance of Asia is likely underrepresented in our data set due to our exclusion of publications written in Asian languages. The prevalence of records on FB-ES associations from Asia may reflect the continent's size (29% of land area and 60% of global population; UN, 2021), social and biogeographical heterogeneity, and diverse FB fauna, particularly in the Indotropics Zieritz et al., 2018b). However, research into other aspects of FB biology and ecology in Asia is considerably lagging behind that in North America and Europe, despite numerous conservation issues Zieritz et al., 2018b).
Long-term protection of global FB diversity and functionality will require a good understanding of both their ecology and their ESs. For example, in China, where FB products ('commodities'), such as ornamental pearls or medicines, are particularly popular, knowledge gaps on species diversity and conservation status should be urgently addressed to avoid unsustainable exploitation of rare and/or threatened species. By contrast, in North America and Europe, ES research has so far strongly focused on the use of FBs in biomonitoring, with other potential ESs having been largely ignored. In North America particularly, significant resources have also been dedicated to generate a better understanding of the functional roles of FBs within their ecosystems (Vaughn & Hakenkamp, 2001;Vaughn, 2018), but the majority of these studies lack a clear relevance to humans. FB conservation in these regions would benefit from more explicitly recognising the contributions that FBs make to human well-being. That said, there is evidence for this acquired knowledge on FB-ESs already pushing conservation initiatives in these regions, including the Mussels for Clean Water Initiative, which aims to restore native FB populations in the Delaware and other northwestern US river basins 'to promote cleaner water and healthier aquatic ecosystems' (The Partnership for the Delaware Estuary, 2017). Finally, in the southern hemisphere, all aspects of FB research, including ecology, conservation and ESs, are notoriously understudied.
Some avenues of ES research that are popular in one region may not be attractive in other regions due to differences in socio-economic conditions and/or characteristics of the FB fauna, but efforts should be made to identify the complete range of (potential) ESs of FBs globally. For example, whilst in the western world, FBs are rarely eaten any more, their potential as a bioremediation tool for disturbed ecosystems appears to be severely underexplored. In Africa, research into the use of FBs as a source of food and biomonitoring tool may be of particular relevance (Akélé et al., 2015;Labuschagne et al., 2020). Future research would benefit greatly from international collaborations to facilitate exchange of knowledge, know-how and technologies. Ultimately, this should be accompanied by quantification and economic valuation of ESs over space and time (Strayer, 2017), including data from unpublished and grey literature, as well as cost-benefit analyses under alternative management and environmental change scenarios.
(2) Synergies and trade-offs among FB ecosystem services From our database, 146 studies provided evidence for multiple FB-ES associations, indicating that the provision/ disruption of one ES coincided with the provision/disruption of one or more other ESs (synergistic and trade-off effects; Turkelboom et al., 2016). Common examples for ES synergies in our database referred to knowledge acquisition (e.g. biomonitoring), regulation of organisms (e.g. algae), mediation of human inputs (e.g. heavy metals) and/or physico-chemical regulation (e.g. through water clearance). All of these ESs, which provide benefits to human health and improved amenity value of waterbodies, derive from the considerable filtration and bioaccumulation capacity of FBs (Vaughn, 2018). The same traits are also responsible for FBs providing nutritious food (Ke et al., 2011;Horikoshi, 2020), and supporting the production of (shell)fish  Sicuro et al., 2020).
The effect of a particular FB on one or more ESs is often context specific, which can result in trade-offs among the provision and disruption of specific ESs. For example, the remediation of polluted water by FB filtration may be offset by the adverse effects on human health when these FBs are consumed. Based on our data set, non-native FBs appear to act more commonly as disrupters of ESs than native FBs, e.g. by reducing the recreational value of waterbodies by causing cyanobacterial blooms (Vanderploeg et al., 2001;Jones, 2019). Furthermore, invasive FBs are known to cause a range of other ecological and economic impacts, including the complete alteration of ecosystem structure and functioning, and the fouling of industrial intakes and boats (Strayer, 1999;Sousa, Gutiérrez & Aldridge, 2009;Sousa et al., 2014), with massive economic costs (63.8 billion US$ between 1980 and 2020 according to Haubrock et al., 2022). On the other hand, there is evidence for invasive FBs providing ESs, most notably through biomonitoring and nutrient removal in eutrophicated ecosystems (Richman & Somers, 2010;McLaughlan & Aldridge, 2013). Future research should aim to provide a more complete understanding and quantification of how the provisioning of one ES may reduce the provisioning of another or result in additional ecological and/or economic costs, particularly with regard to native versus non-native species.

VIII. CONCLUSIONS
(1) Our systematic review of 684 published studies identified 904 records providing evidence for an association between FBs and 24 specific ES classes. Records originated predominantly from Asia, Europe and North America, with poor representation of countries from the southern hemisphere. About one in ten records referred to a disruption rather than provision of ESs.
(2) Temporal trends in the publication of FB-ES associations reflected increased interest in ecology and conservation since the 1980s, increased attention on non-native species since the 1990s and popularisation of the ES concept since the 2000s.
(3) Evidence on provisioning ESs, and particularly the provisioning of food, materials and medicines by Corbicula spp. and unionids, was prominent in Asian countries, yet this is a region where species diversity, distribution and ecology remain understudied.
(4) In North America and Europe, evidence was primarily focused on cultural ESs, and predominantly the use of native and non-native FBs for biomonitoring. Other cultural CICES groups referred to the ornamental use of FBs, their religious, spiritual or traditional meaning, and their effects on physical interactions. (5) Regulating ESs comprised the regulation of organisms, including (harmful) algae and bacteria, the regulation of substances and physico-chemical water properties, and the mediation of harmful substances of human origin. Records were commonly based on non-native species (mainly Dreissena spp.).
(6) Multiple FB-ES associations within single studies were often associated with the biofiltration and bioaccumulation capacity of FBs, linking various provisioning (e.g. source of micronutrients), cultural (e.g. biomonitoring) and regulating (e.g. of algae) ESs. Trade-offs among ESs provided and disrupted by FBs were particularly commonly observed for non-native FBs.
(7) As the global community of FB researchers grows, we recommend that future efforts are directed towards the continued documentation of FB-ESs, particularly in less wellstudied geographic regions, such as Africa and South America. Attention should be given to ESs and FBs beyond those commonly reported. Progress should be made in the economic quantification of FB-ESs, so that trade-offs can be properly enumerated and informed decisions can be made about effective ecosystem management, conservation and restoration programs.