Patterns of Mekong Mollusc Biodiversity: Identification of Emerging Threats and Importance to Management and Livelihoods in a Region of Globally Significant Biodiversity and Endemism

The Lower Mekong Basin (LMB) is a key biodiversity hotspot. To facilitate conservation and management, we examine mollusc biodiversity patterns and distribution along LMB’s longitudinal gradients, identify environmental drivers, and discuss the importance of these drivers to management. Cluster analysis, redundancy analysis (RDA), and variation partitioning were conducted using mollusc data collected from 63 sampling sites. Results indicated that species diversity is dominated by gastropods (61%) and bivalves (39%) and feeding trait diversity by scrapers (52%) and filter-collectors (37%). Only 48 species (49%) out of 98 taxa have been assessed by the International Union for Conservation of Nature (IUCN) including a growing number of invasive species. The lack of complete, up-to-date information highlights the need for more research on both native and alien species. Cluster analysis revealed a clear mollusc biodiversity structure along the LMB’s longitudinal segments. Diversity was lowest in upstream tributaries, increased in upstream main channels, and was highest in downstream channels and the Mekong delta, the exception being the observed high gastropod abundance in Chi-Mun river mouth and Luang Prabang areas. The RDA and variation partitioning demonstrated that combined physical–chemical and climatic conditions are the key drivers of biodiversity patterns. Given the potential spread of invasive alien species and increasing anthropogenic impacts, further ecological research, regular monitoring, and adaptive management are needed to sustain mollusc biodiversity and associated ecosystem services, which contribute to food security, nutrition, and livelihoods in the LMB. Water 2020, 12, 2619; doi:10.3390/w12092619 www.mdpi.com/journal/water Water 2020, 12, 2619 2 of 16

. The sampling sites belonging to each cluster projected on the map (a) and the dendrogram showing the three distinct clusters using mollusc data collected from 63 sampling sites from the Lower Mekong Basin (b). See Results (Section 3) for details.

Land Cover Data
At each sampling site, surface area and land cover data were calculated using a Geographic Information System (ArcGIS 10.4, ESRI). Geographical data (ArcGIS shapefiles) of the LMB (land cover types, river networks, basin boundaries, and sub-catchments derived from topographical maps

Land Cover Data
At each sampling site, surface area and land cover data were calculated using a Geographic Information System (ArcGIS 10.4,ESRI). Geographical data (ArcGIS shapefiles) of the LMB (land cover types, river networks, basin boundaries, and sub-catchments derived from topographical maps in 2003) were provided by the MRC, available at https://portal.mrcmekong.org/land-cover/land-cover-types.

Climatic Data
Bioclimatic data (19 variables) were derived from the WordClim database, available at http: //www.worldclim.org, describing climatic conditions for the period 1950-2000 with a spatial resolution of about 1 km 2 . The mean value of the mentioned period from each sampling site coordinate (as the centre point of the 1 km 2 grid) was used for the analysis.

Data Treatment and Analysis
In this study, we examined community spatial variability and environmental variables across sampling sites in the LMB. Median values from biological and environmental data were used to represent each sampling. The median values were used due to unequal sampling efforts [27]; that is, unequal and different number of samples at each site during the 5-year sampling period.
Mollusc biodiversity was first summarized based on the number of classes, families, genera, species, functional feeding groups (FFGs), and their IUCN (International Union for Conservation of Nature) conservation status using the IUCN Red List of Threatened Species website: https://www.iucnredlist.org. Then, mollusc community abundance data from the 63 sampling sites (site-by-species) were grouped into different clusters, based on the Bray-Curtis dissimilarity distance of their community, using the Ward's hierarchical clustering method [28]. Each cluster is a group of sampling sites that are characterized by taxonomic similarity, and therefore sampling sites with similar mollusc community composition are grouped into the same cluster [29]. To analyse the spatial variation of mollusc biodiversity, their taxonomic and functional richness, abundance, and diversity index were compared between the clusters (i.e., defined by the Ward's clustering method), using either a one-way analysis of variance (ANOVA) for normal distributed data or a Kruskal-Wallis test for the not-normal distributed data. Both tests were carried out using the kruskal.test() and aov() functions of the stats package [30]. p values < 0.05 were considered statistically significant.
Redundancy analysis (RDA, [28]) was used to gain a comprehensive understanding of factors driving biodiversity patterns. The community data were Hellinger-transformed before conducting the RDA, while the environmental variables were normalized using the zero minimum because the value of each variable highly varies from one to another (different measuring scales) [31]. In this normalization approach, the normalized values are based on the standard deviation of each variable while at the same time keeping the range of all variables constant. The strength of each variable in explaining the mollusc communities is depicted by the length of its arrow in the ordination map; the longer the arrow length, the stronger the correlation/association will be [28]. Forward selection of the RDA model was then performed to retain only the most significant environmental variables. To identify which types of environmental variables (i.e., physical-chemical vs. land cover types vs. climatic variables) is/are the main driver(s) of biodiversity patterns, variation partitioning [28] was performed using the community transformed data and the most significant environmental variables yielded by the RDA forward selection. All statistical analyses were performed in R statistical programming language [30].
Among all taxa, the functional feeding groups scrapers, filter-collectors, and gatherer-collectors represented 52% (51 taxa), 37% (37), and 5% (5), respectively. Forty-eight of the 69 identified species had been assessed based on the IUCN Red List: Data Deficient (DD, 10 species), Least Concern (LC, 37), and one Near Threatened (NT) species, Pachydrobiella brevis. Moreover, five species are alien to the LMB, four of which (i.e., Angulyagra polyzonata, Corbicula fluminea, Sermyla riqueti, Sinotaia aeruginosa) are native to the river basin in China and/or Hong Kong, Korea, and one of which (Neritina rubida) is native to Madagascar and French Polynesia [32]. For further details on all taxonomic information, see Supplementary Table S1.

Community Variation
Three regional clusters were identified based on biodiversity patterns using mollusc community dissimilarity (Bray-Curtis) and hierarchical analysis (Figure 1a). Cluster 1 was made up mainly from sites located in the 3S (Sesan, Sekong, Srepok) rivers in northeastern Cambodia and Vietnam; Cluster 2, from sites in Mun, Chi, and Tonle Sap River basins, and also sites along the main channel of the Cambodian Mekong; and Cluster 3, mainly from sites in the Mekong delta ( Figure 1b).
A significant difference between clusters and a generally increasing trend in richness, abundance, and diversity from Cluster 1 (the tributaries) to Cluster 3 (Mekong delta) was observed for most of the computed metrics (Figures 2 and 3; Table 1). An exception to this trend was found for gastropod and scraper abundance, which was the highest in the upstream main channel (Figures 3 and 4), i.e., Chi-Mun river mouth and Luang Prabang, where the confluence of the Mekong and Nam Khan rivers occurs, and the lowest in the tributaries (Table 1).
Among all taxa, the functional feeding groups scrapers, filter-collectors, and gatherer-collectors represented 52% (51 taxa), 37% (37), and 5% (5), respectively. Forty-eight of the 69 identified species had been assessed based on the IUCN Red List: Data Deficient (DD, 10 species), Least Concern (LC, 37), and one Near Threatened (NT) species, Pachydrobiella brevis. Moreover, five species are alien to the LMB, four of which (i.e., Angulyagra polyzonata, Corbicula fluminea, Sermyla riqueti, Sinotaia aeruginosa) are native to the river basin in China and/or Hong Kong, Korea, and one of which (Neritina rubida) is native to Madagascar and French Polynesia [32]. For further details on all taxonomic information, see Supplementary Table S1.

Community Variation
Three regional clusters were identified based on biodiversity patterns using mollusc community dissimilarity (Bray-Curtis) and hierarchical analysis (Figure 1a). Cluster 1 was made up mainly from sites located in the 3S (Sesan, Sekong, Srepok) rivers in northeastern Cambodia and Vietnam; Cluster 2, from sites in Mun, Chi, and Tonle Sap River basins, and also sites along the main channel of the Cambodian Mekong; and Cluster 3, mainly from sites in the Mekong delta ( Figure 1b).
A significant difference between clusters and a generally increasing trend in richness, abundance, and diversity from Cluster 1 (the tributaries) to Cluster 3 (Mekong delta) was observed for most of the computed metrics (Figures 2 and 3; Table 1). An exception to this trend was found for gastropod and scraper abundance, which was the highest in the upstream main channel (Figures 3  and 4), i.e., Chi-Mun river mouth and Luang Prabang, where the confluence of the Mekong and Nam Khan rivers occurs, and the lowest in the tributaries (Table 1).
Water 2020, 12, x FOR PEER REVIEW 7 of 16

Drivers of Biodiversity Patterns
The RDA model characterized the differences of community clusters, revealing detailed environmental drivers of each cluster ( Figure 5). The first two axes explained 27.6% of the total community variance (axis 1: 16.2%, and axis 2: 11.4%). The Monte Carlo test of the generated axes of the RDA was significant (p = 0.005). Along axis 1, sites in cluster 1 were on the left part of the ordination map, and associated with evergreen forests, and cluster 3 were on the right part, associated with deep rivers with a large watershed area, high water temperature, and annual mean temperature. Along axis 2, most of sites in cluster 2 were on the lower part of the ordination, associated with a high value of water conductivity, urban area, maximum temperature of the warmest month, temperature annual range, and precipitation seasonality.   The variation partitioning indicated that each type of environmental variable explained (≤6%) the diversity patterns of molluscs less than the joint contribution of physical-chemical and climatic variables (16%), physical-chemical and land cover variables (9%), and climatic and land cover variables (8%). However, the proportion explained by the combination of the three components is very limited ( Figure 6). All of the variables combined equalled a model adjusted R 2 of 0.38.

Mollusc Biodiversity, Conservation Status, and Threats
Mollusc biodiversity in the LMB has not been well-studied. Existing information is restricted to taxonomy, distribution, and general macroinvertebrate patterns. More information is needed on species, family, and order diversity; species functional feeding groups; and conservation status. Here, we fill this gap and thus improve the existing knowledge, focusing on mollusc biodiversity patterns and the key environmental drivers of these patterns [6,13,20,21].

Mollusc Biodiversity, Conservation Status, and Threats
Mollusc biodiversity in the LMB has not been well-studied. Existing information is restricted to taxonomy, distribution, and general macroinvertebrate patterns. More information is needed on species, family, and order diversity; species functional feeding groups; and conservation status. Here, we fill this gap and thus improve the existing knowledge, focusing on mollusc biodiversity patterns and the key environmental drivers of these patterns [6,13,20,21].
At the species level, five species including Corbicula leviuscula, Corbicula tenuis, Corbicula lamarckiana, Limnoperna siamensis, and Stenothyra koratensis holosculpta were dominant, as reflected by their widespread distribution and high abundance [13,33]. Therefore, they are the key species representing mollusc communities in the LMB. Other Near Threatened (i.e., Pachydrobiella brevis) and less common species (e.g., Clea scalarina, Hyriopsis (Limnoscapha) desowitzi, Adamietta housei) may be more vulnerable to environmental changes including climate change and threats from alien species.
At a higher taxonomic level, the mollusc communities in the LMB are mainly characterized by two common families of bivalves (Unionidae and Corbiculidae) and three families of gastropods (Viviparidae, Pomatiopsidae, and Stenothyridae). The sampling design used in this study (e.g., much of the river course of the LMB were not sampled, e.g., Cambodian-Lao and Lao-Thai borders, the Tonle Sap and its tributaries, etc.) may affect the resulted diversity of this study. The other caveat for the study is that it was conducted only in March in the dry season. Some species belonging to Corbiculidae are highly abundant in dry season when there is less precipitation and higher temperature, while others such as Pila spp. (e.g., Pila virescens) are more abundant in the wet season with more precipitation and lower temperature [34]. Regarding the dominant family Unionidae found in this study, its members mainly occur in both forested and grassy riparian zones [1,35], whereas Corbiculidae has various reproductive strategies and spreading abilities [22]. These life strategies and biological traits may explain their wide distribution across the basin. However, the two families were more abundant in terms of biomass and richness downstream, which indicated their preferred habitats (e.g., large rivers with slow water flow) similar to that found by previous studies [36,37]. The commonness of these families could be also due to their wide distribution in both freshwater and estuary habitats and good dispersal strategies [14]. Among the five common families, Unionidae, Viviparidae, and Pomatiopsidae were more diverse in terms of their member genera. However, some genera of the three families were found at only one sampling site (e.g., Hyriopsis, Scabies, Uniandra, Anulotaia, Lacunopsis). The remaining two families Corbiculidae and Stenothyridae are entirely dominated by the genera Corbicula and Stenothyra, respectively. Because of their high diversity, widespread distribution, and varied reproductive strategies, these families are less likely impacted by habitat invasion and competition from the other species [22,38,39]. Regarding the less common families, most of them were found with only 1 occurrence of 1 individual, e.g., Clea scalarina (Buccinoidae), Adamietta housei (Pachychilidae), and Lymnaea swinhoei (Lymnaeidae) or 2 occurrences of 2 individuals, e.g., Neritina violacea (Neritidae), Cochliopa riograndensis (Cochliopidae), and Neoradina prasongi (Thiaridae). These species' distributions and abundances have not been assessed from the LMB, and therefore their population statuses are unknown [32]. Some exceptions were for Adamietta housei, of which 28 individuals were recorded from two stations in Thailand (Sri-Aroon, 2007), and for C. scalanina, of which 8 specimen-based occurrences are recorded in the Global Biodiversity Information Facility (GBIF) database [40]. In this regard, further studies on these species are needed because they already appear rare and may be more vulnerable to environmental change, high anthropogenic stress, climate change, or ecological alteration [32].
Compared to other basins, the species richness occurrence range reported in the LMB (14-58 species) [6] is relatively similar to that in other tropical river basins such as Yangtze, Pearl, Huaihe, and the Southeast rivers in China (8-158 species) [41]. However, studies dating back to the 2000s (e.g., Bogan, 2008;Strong et al., 2008) reported that the LMB supports the most diverse freshwater mollusc species (140 gastropod and 39 bivalve species) in the world and highest level of endemism, second after the Mobile Bay Basin in the southeastern United States (271 bivalve and 118 gastropod species) [1,14]. Our study, which does not cover the entirety of the Tonle Sap, Mun, and Chi basins, in the mid-reaches of the Mekong near the border between Cambodia, Laos, and Thailand (e.g., Kong Island) that are thought to be hotspots of pomatiopsids and stenothyrids [42,43], also reports a higher maximum species richness of bivalves per site (16 species) than that reported from European river basins (10 species) [44], and a higher total number of gastropod taxa (60 taxa) than those in Northern Australia Monsoonal wetlands (~56 taxa), Lower Uruguay River and Rio de la Plata, Argentina-Uruguay-Brazil (~54), Western lowland forest of Guinea and Ivory Coast (~28), and Zrmanja river basin in Croatia (~16) [14]. In this regard, and given the fact that much of the river course of the LMB (e.g., the Cambodian Tonle Sap Lake and River and its main tributaries such as Sangke, Pursat Stung Sen, and Chinith Rivers) has not been investigated, the LMB still appears to be the hotspot area for molluscs, as illustrated by Köhler et al. [6]. The main channels (the Mekong and Bassac) most likely support a higher mollusc species diversity than that indicated in Köhler et al. [6]. This can be due to the fact that the highest species richness of the present study (34 taxa/site) is recorded along these main channels and that another 20 additional species (see Quang et al. [45]), which are different from the recorded species of the present study, have been also reported from the main channels and its tributaries in the delta [45]. This suggests that the main channels downstream of the LMB and its tributaries deserve a re-assessment as mollusc biodiversity hotspots.
Despite the high biodiversity of molluscs and their importance for the ecosystems as well as the livelihoods of local people in the region [6,46], this resource is facing critical threats. Overharvesting, habitat degradation (e.g., land use change), and changes of water regime are among the major factors negatively impacting mollusc biodiversity [34]. These impacts have already been observed in the rivers of North America [47,48]. Other anthropogenic disturbances, e.g., agricultural wastes such as chemical fertilizers, herbicides and pesticides, and household wastes, are also a great threat, leading to the decline of molluscs and other biodiversity. Alien invasive species are the other emerging threat; they compete, alter the habits of the native species [49]. Although there is limited research on alien mollusc species in the LMB, which reflects the necessity of up-to-date assessment, several of them have already become established in the LMB, as in the case of this study (i.e., Angulyagra polyzonata, Corbicula fluminea, Sermyla riqueti, Sinotaia aeruginosa, and Neritina rubida) [32]. Many more alien species are suspected to occur in the LMB. In this Anthropocene epoch, the current impact of climate change in the LMB can facilitate the spread of alien species. Since the LMB is a transboundary river in which its channel is inter-connected to other neighbouring ecosystems, the species most likely move (e.g., via water flow) or are transported in and out of the LMB when facing the on-going climate change and the current anthropogenic activities. Limited or regulated flow in the LMB, due to hydropower development, also has a negative effect on the species turnover (i.e., the presence of rare species in the community) in the LMB [13,50], and this indeed benefits alien species as they are considered to have high ability to compete and adapt to different environmental changes. All of these factors consequently and inevitably will have strong negative impacts on the native mollusc and other biodiversity in the LMB.

Environmental Gradients as Drivers of Mollusc Diversity
Environmental gradient is the key factor driving biodiversity patterns in lotic ecosystems [51,52]. Here we found that environmental heterogeneity of the LMB, compared to other factors, had a greater influence on mollusc biodiversity patterns. The overall mollusc community, sub-groups, and FFG diversity, by means of richness, abundance, and Shannon diversity H, increased following the longitudinal gradients of the LMB. This result is different from the finding of Köhler et al. [6], who reported mollusc species richness was the highest in the main channel and large floodplain tributaries (e.g., Mun and Chi rivers) of the LMB upstream. The difference could be due to the fact that the study of Köhler et al. [6] counted the species richness at the catchment scale while this study measured the diversity at a sampling site scale and did not include a large portion of the Mun, Chi, and Tonle Sap Lake basins. Nonetheless, we found that mollusc communities at the site-specific scale are less diverse in the tributaries and the mainstream upper channels but more diverse in the delta (lower floodplains) downstream.
Tributary rivers such as the 3S river system (Cluster 1) of the LMB are smaller, narrower, and shallower than the main channels and the downstream. Moreover, they are covered by a large proportion of forests or evergreen forests ( Figure 5). These conditions are mostly preferred by insect shredders [13] but not molluscs because such conditions constrain the abundance of phytoplankton and reduce organic materials produced via photosynthesis in the water, due to limited penetration of solar radiation [13,53]. Therefore, in these tributary rivers, the consumption outpaces the production and thus limits the diversity of molluscs, in particular the collectors. Only some mollusc scrapers prefer these habitats.
Mollusc diversity increases in the main channels of the basin (except for the delta) ( Figure 2). The habitats of mollusc communities in this cluster mainly located at/nearby the main channels and in key large sub-river basins such the Chi, Mun, and Tonle Sap rivers [15]. The structure of these river habitats is characterized by a large floodplain, varied aquatic plants, different rock sizes and riffles, and deciduous forests. These conditions lead to a higher production than the consumption, and so play an important role in supplying organic matter (e.g., periphyton and other autochthonous), which is preferred by mollusc scrapers (snails) [34,54]. Therefore, it is not surprising that the highest gastropod and scraper abundance were found in Cluster 2.
We found the highest diversity of molluscs in the downstream of the LMB (Cluster 3). The Mekong delta is the mouth of the Mekong basin with a large flux of particulate materials. The large watershed and large and open rivers in the delta support diverse, abundant mollusc populations/communities. This is because the high flux of particulate materials and high temperature, due to solar radiation, provide the most suitable conditions for phytoplankton [53,55], and as a result, phytoplankton feeders and filter/gatherer molluscs proliferate. These findings also support previous study results in other tropical regions such as the Guadiana River basin (SW Iberian Peninsula) [36] and Southern Brazil wetlands [56].
Drivers contributing to biodiversity patterns are more than just simple or single factor. Shared contribution of drivers is often reported to shape biodiversity in the river systems [57][58][59].
As we found in this study, the combined contribution of physical-chemical and climatic conditions explained a larger proportion of variation of mollusc biodiversity, compared to single and other combined factors ( Figure 6). Climatic variables such as the annual mean temperature (bio1), maximal temperature of warmest month (bio5), temperature annual range (bio7), precipitation seasonality (bio15), and precipitation of driest quarter (bio17) were the key factors driving the patterns of mollusc biodiversity. These variables are interrelated with the physical-chemical and land cover conditions. High temperature at some degree can lead to decreased dissolved oxygen and increased electrical conductivity [13,60], whereas the precipitation depends on the forest land cover and vice versa [61]. Precipitation affects the seasonal water flow and thus the movement of nutrients along the longitudinal segments of the rivers and flood-pulse regime (e.g., the Tonle Sap) [52,62]. For instance, higher temperatures (climatic factor), coupled with higher nutrients inputs (physical-chemical factors) in the Delta, lead to higher mollusc diversity. All of these combined factors and interrelations are therefore the main drivers shaping the biodiversity patterns of the LMB biota including the molluscs.

Management Implications
The LMB is a biodiversity hotspot supporting one of the most diverse mollusc communities in the world. Appropriate planning and management of the basin are vital for protecting molluscs, which depend on healthy aquatic habitats and support people's livelihoods. In the basin, molluscs are an important ecosystem element as they can change the biotic diversity in the basin, due to their high among-site taxonomic and FFG variation (i.e., high beta diversity) across the LMB [33]. Moreover, molluscs are considered as ecosystem engineers, especially gastropods and bivalves, by means of shell production [63]. Architectural complexity and heterogeneity in the benthic environments, as the results of shell aggregation, are important in determining other species' colonization success and composition within the habitat. This is because the complexity and heterogeneity routinely modify the resources available to other aquatic organisms. For instance, the shell cavity of a gastropod can be a microhabitat of other organisms (e.g., crabs), where they can avoid temperature extremes, hydrodynamic forces, or predation. All of these processes can therefore positively influence the whole basin (or landscape-) level species richness.
In the context of ecosystem services, molluscs are an excellent water filtering organism. Bivalves can filter out large amounts of phytoplankton, bacteria, and organic materials and can also absorb pollutants such as heavy metals [1,46]. In terms of provisioning services, molluscs play a key role in food web dynamics as they are the food sources of various organisms (fish, turtles, rodents, birds etc.). Moreover, molluscs are the source of protein intake for the local people sharing the LMB. In some regions, e.g., the Tonle Sap Lake in Cambodia, they are the second most important fisheries after fish [34], from which local people can generate significant income.
Anthropogenic disturbance is changing and degrading the LMB. Modification of rivers' flow (e.g., artificial island created along the rivers, ports development), degradation of habitats (e.g., forest clearance, erosion, agricultural and industrial wastes), and hydropower development are contributing causes to biodiversity changes and river alteration. For instance, the current development of the hydropower dam in the Lower Sesan 2 leading to the creation of large reservoir could create favourable conditions for some invasive species including Pomacea spp. to thrive as they can tolerate a high nutrient load in the reservoir. The environmental changes induced by dams can also alter the distribution of some native pomatiopsid and stenothyrid species, which are the intermediate hosts of paragonimiasis and schistosomiasis [64,65]. These parasitic worms, which can infect local people, are thought to benefit from the regulation and modification of the Mekong and its tributaries. Moreover, as discussed earlier, the LMB is already home to several invasive species. In addition to the five reported alien species in this study, the golden apple snail (Pomacea canaliculata) is also of great threat to biodiversity [59], in particular to native mollusc species. For instance, the golden apple snail can outcompete the food resources and available habitats of the native snails (e.g., Pila spp.). The golden apple snail is a globally well-known invasive species, wide-spread and highly adaptable to different environments. Overharvest is another potential threat, especially in the Tonle Sap Lake and Mekong Delta, which can affect the overall food web, reducing food available for higher tropic level organisms [34]. All of these factors consequently impact the biodiversity sustainability and thus raise a global concern for the long-term persistence of Mekong molluscs and their availability as a food resource in the basin.
Given the factors driving the mollusc biodiversity, the LMB resources require a higher attention for management and conservation planning such as habitat/resource protection and enhancement and resource use coordination between competing users. Good conservation efforts and effective management of the basin and its sub-catchment, e.g., Mun, Chi, and Tonle Sap River basins, and the Mekong Delta (where numerous mollusc taxa and species indicators occupy [13]), will enhance biodiversity and maintain the system integrity, functions, and services and consequently contribute to food security, employment, and socioeconomics of the people sharing this important river ecosystem.

Conclusions
In this study, we revealed that main channel and large tributary mollusc biodiversity in the LMB is dominated by gastropods and bivalves, consisting mainly of scrapers and collectors. Among the 98 taxa, 48 species (49%) have already been assessed in the IUCN RedList, one of which is Near Threatened, 10 are Data Deficients, and 5 Aliens. This suggests a need for a more up-to-date assessment of the species and existing and potential threats, including new alien species, by setting up regular monitoring on aquatic invertebrates including the molluscs and environmental DNA. Regarding their biodiversity patterns, this study showed mollusc diversity structured along environmental gradients of the river. Their diversity increases from the upstream tributaries to the upper main channels and then to the Mekong delta, except for the high abundances of gastropod and scraper found in the upstream. Combined physical-chemical and climatic conditions, such as river width, river depth, surface area of watersheds (physical-chemical), and temperature, temperature annual range, and precipitation (climatic), are the key drivers of the observed patterns. Other important variables such the nutrient and sediment loads, and micro/meso habitat types should be considered in the analysis as well in further studies. Nevertheless, given high levels of endemism, lack of relevant and up-to-date data, potential invasion, and anthropogenic impacts, further ecological research, regular evaluation, management planning, and legislative support are needed to sustain mollusc biodiversity and associated ecosystem services, especially in the context of people's livelihood in the LMB and global change.