Reintroduction of freshwater mussels (Bivalvia, Unionida) directly after channel dredging can serve as an effective measure in mitigation conservation

Abstract

This study is based on a natural experiment carried out in the Biebrza National Park, Poland. The study site was a channel inhabited by Anodonta anatina, A. cygnea, Unio pictorum and U. tumidus. The deepening of the channel to restore ecosystem connectivity provided an opportunity to conduct this study. Mussels were collected before dredging, held in captivity for 48 h, measured, individually tagged and released post-dredging to the same 5-m channel sections they originated from. They were subsequently monitored for three consecutive years. Mussel survival remained high throughout the study, and no increased mortality in the year following reintroduction was observed. There was no growth retardation. Mussel mobility was low, with most individuals remaining in the same channel section in which they were released. Recolonisation patterns were consistent with the composition of mussel communities in adjacent unaffected habitats. Although dredging drastically changes mussel habitat, some characteristics: microclimate, water chemistry, nutrient availability and host fish can remain adequate. Our study shows that reintroducing mussels to the same site can serve as an effective mitigation conservation measure and can be preferable to translocation, particularly when carried out under time pressure with limited possibilities of assigning appropriate destination sites.

Introduction

Freshwater mussels (Bivalvia, Unionida) play a key role in the functioning of freshwater habitats through water filtration, bottom bioturbation, nutrient cycling and provision of habitats¹, and can provide important ecosystem services². They are, however, extremely vulnerable to anthropogenic impacts^3,4, and belong to the most imperilled taxonomic groups globally⁵. Habitat loss and modifications are among the primary drivers of those declines^6,7. Bottom dredging, a common habitat modification, is a routine management practice for rivers and channels, but it has detrimental and often catastrophic effects on mussel populations^8,9. Considered a small-scale and low-impact intervention, it usually does not require licensing, or licensing is obtained without appropriate evaluation of the environmental impacts, especially outside protected areas. In effect, many unionid mussel populations are decimated. As in all long-lived and slow-growing organisms, their recovery is notoriously slow¹⁰.

Translocation of freshwater mussels in response to habitat degradation is an increasingly used conservation practice^11–13. Its success depends largely on careful planning, with the choice of appropriate relocation sites being of crucial importance^11,13. In the case of bottom dredging, especially of small streams and channels, there is often no publically available information on its being planned, and thus if a conservation action is possible at all, it is often a last-moment intervention, with all constraints inherent to the “collecting individuals in front of the bulldozers” practice. While an inevitable result of inaction is the destruction of affected mussel populations^9,14, the option of translocating mussels to appropriate sites is often unfeasible due to the time and funding limitations. Translocating mussels to random or only apparently appropriate sites poses a high risk of their death due to unsuitable habitat conditions¹⁵. Other major concerns related to translocations include the genetic consequences for the donor and recipient populations and the risk of transferring pathogens or invasive species^16–18.

This study is based on a natural experiment carried out to evaluate the conservation potential of collecting and reintroducing mussels to exactly the same site directly after dredging. Although dredging drastically changes mussel habitat, some characteristics, such as local microclimate, water chemistry, nutrient availability and the presence of host fish, necessary for the metamorphosis of parasitic unionid mussel larvae¹⁹, can remain essentially unchanged. Thus, despite the large scale of the disturbance, many aspects of mussel habitats can still be suitable. Given the high number of mussel translocations carried out, this conservation practice is possibly not unprecedented, but as far as we know, the results of such approach have not been published before.

This study was carried out in the Biebrza National Park in north-eastern Poland (Fig. 1). The study site was a channel inhabited by four species of unionid mussels: Anodonta anatina, A. cygnea, Unio pictorum and U. tumidus. The channel provides an important connection between the Biebrza River and the system of moats of the fortifications of the Osowiec Fortress. Although the moats are artificial water bodies, they have become naturalized in the 140 years since their construction and are now an important habitat and breeding ground for many species of aquatic fauna. Due to insufficient water flushes during a series of hot and dry summers prior to 2015, large amounts of sediments accumulated on the bottom of the channel; its reconstruction was aimed at restoring the connection and enabling water exchange (Fig. 2). We used this opportunity to set up an experiment in rescuing and releasing mussels directly after bottom dredging.

Figure 1.

Map of the study area and location of channel sections (S01–S15). The map was generated using the QGIS 3.28.4-Firenze software (Free and Open Source Software (FOSS); Free Software Foundation, Inc., USA; www.qgis.org).

Source for the orothophotomap: goeportal.gov.pl (Terms and conditions: https://www.geoportal.gov.pl/regulamin).

Figure 2.

The study site before (A) and after (B) dredging.

Before dredging, we collected all unionid mussels we found during an eight person-hour search. Mussels were held in captivity for 48 h while the channel was dredged. They were measured, individually tagged and released post-dredging to the the same 5-m channel sections they originated from. They were subsequently monitored for three consecutive years. We tested the following hypotheses: (1) reintroduction would negatively affect the survival and (2) cause growth retardation, (3) mussels would respond with increased mobility, (4) different species would respond differently to the intervention, and Anodonta cygnea, the most vulnerable species, would be most affected. We also hypothesised that (5) recolonization patterns would be consistent with the composition of mussel communities in adjacent unaffected habitats.

Results

Species composition, recovery rates and survival probabilities

A total of 392 mussels were collected prior to channel dredging in 2015. Anodonta anatina contributed to 62%, A. cygnea to 27%, U. pictorum to 5%, and Unio tumidus to 6% of the mussel community. In the Biebrza River adjacent to the channel section S01, A. anatina contributed to 35%, A. cygnea to 8%, U. pictorum to 10%, and U. tumidus to 24% (N = 65). In the moat, adjacent to the channel section S15, A. anatina contributed to 51%, A. cygnea to 39%, U. pictorum to 5%, and U. tumidus to 5% (N = 18).

Over the three year monitoring period, the recovery rates ranged from 55% in A. anatina in 2016 to 20% in A. cygnea in 2017, calculated in relation to the initial number of mussels released in 2015 corrected for the number of dead individuals found the previous year (Table 1). Interactions between recovery rates, study years and species were statistically significant (log-linear model for multidimensional contingency tables, G₁₇ = 51.23, P < 0,001), with recovery rates dependent on the interaction between year and species (G₁₁ = 51.19, P < 0.001). In all species, the recovery rates were lower in 2017 than in 2016 and 2018 (Fisher’s exact test, P_2016–2017 < 0.0001; P_2017–2018 = 0.007, Bonferroni corrected significance level, P = 0.016). In all study years, the recovery rates were higher in A. anatina than A. cygnea (Fisher’s exact test, P < 0.0001, Bonferroni corrected significance level, 0.0083), but there were no other statistically significant differences between species.

Table 1.

Mussel recovery rates (%) in relation to the initial number of released mussels after accounting for dead individuals recovered the previous year (N).

	A. anatina		A. cygnea		U. pictorum		U. tumidus
	%	N	%	N	%	N	%	N
2016	54.7	243	36.5	104	45.0	20	48.0	25
2017	36.7	240	20.4	103	36.8	19	45.8	24
2018	48.3	230	31.6	98	42.1	19	33.3	24

The probability of survival for the reintroduced mussels remained high throughout the study, reaching about 0.95 in all species one year after reintroduction and ranging from 0.69 in U. tumidus to 0.95 in U. pictorum two years later (Table 2). Differences between species were not statistically significant (log-rank test, χ₃² = 2.28, P = 0.52).

Table 2.

Kaplan–Meier estimates of survival probabilities (± SE) for unionid mussels reintroduced in 2015.

	A. anatina	A. cygnea	U. pictorum	U. tumidus
2016	0.988 ± 0.007	0.990 ± 0.010	0.950 ± 0.049	0.960 ± 0.039
2017	0.937 ± 0.017	0.918 ± 0.033	0.950 ± 0.049	0.960 ± 0.039
2018	0.814 ± 0.030	0.798 ± 0.054	0.950 ± 0.049	0.686 ± 0.119

Changes in growth rates

Mussels collected and reintroduced in 2015 showed a slight trend for decreasing shell growth rates over time, consistent among species but not statistically significant in any of them. First young mussels, predominantly A. anatina, were found two years after the channel dredging. They grew significantly faster than the mussels first collected in 2015 (Table 3). Shell length measurements of all individuals are given in Supplementary Table S1 online.

Table 3.

Shell-length increase (%) over consecutive study periods.

Species	Study period	Mean shell-length increase ± SD (%)	Range (%)	N
A. anatina	2015–2016 (2015) A	4.3 ± 7.5	0.0–41.8	133
	2016–2017 (2015) A	4.8 ± 4.8	0.0–19.4	59
	2017–2018 (2015) A	1.2 ± 1.5	0.0–6.9	65
	2017–2018 (2017) B	28.4 ± 12.1	0.0–68.3	146
A. cygnea	2015–2016 (2015) A	16.7 ± 17.6	0.0–54.7	38
	2016–2017 (2015) A	9.1 ± 7.4	0.0–23.2	8
	2017–2018 (2015) A	3.7 ± 4.6	0.0–15.7	10
	2017–2018 (2017) A	32.1 ± 26.8	0.0–70.8	9
U. pictorum	2015–2016 (2015) A	6.5 ± 7.0	0.0–18.9	9
	2016–2017 (2015) A	4.7 ± 4.7	0.0–9.0	4
	2017–2018 (2015) A	2.0 ± 1.3	1.2–3.9	4
	2017–2018 (2017) A	2.9 ± 2.8	0.0–9.5	17
U. tumidus	2015–2016 (2015) A	5.7 ± 7.9	0.0–22.2	12
	2016–2017 (2015) A	4.1 ± 4.6	0.0–13.1	8
	2017–2018 (2015) A	4.7 ± 6.0	0.0–15.2	5
	2017–2018 (2017) A	2.4 ± 3.2	0.0–9.3	7

The year of first collection is given in parenthesis. Letters A, B denote significant differences in the Kruskal–Wallis test with Dunn's pairwise comparisons, Bonferroni corrected significance level P = 0.0083.

Mobility

Before dredging, in the drying out channel mussels were actively moving in search of more suitable habitats, leaving long traces on the surface of the bottom sediments. After the dredging, the reintroduced mussels remained largely stationary. Of a total of 355 mobility records, 57% indicated low mobility, 39%—medium mobility, and 4%—high mobility (Table 4).

Table 4.

Mobility of the mussels reintroduced after channel dredging.

Species	Study period	Low %	Medium %	High %	N
A. anatina	2015–2016	60.2	39.8	0.0	133
	2016–2017	69.5	28.8	1.7	59
	2017–2018	46.2	41.5	12.3	65
A. cygnea	2015–2016	50.0	47.4	2.6	38
	2016–2017	75.0	25.0	0.0	8
	2017–2018	10.0	80.0	10.0	10
U. pictorum	2015–2016	55.6	33.3	11.1	9
	2016–2017	75.0	25.0	0.0	4
	2017–2018	50.0	25.0	25.0	4
U. tumidus	2015–2016	50.0	50.0	0.0	12
	2016–2017	50.0	25.0	25.0	8
	2017–2018	100.0	0.0	0.0	5

Low mobility, mussels found in the same channel section in which they were released the previous year; medium, mussels found in adjoining channel sections; high, mussels found one or more sections away.

Interactions between mobility rates, study years and species were statistically significant (log-linear model for multidimensional contingency tables, G₂₈ = 57.63, P < 0.001), and mobility rates were dependent on the interaction between year and species (G₂₂ = 48.34, P < 0.001). Fisher’s exact test did not reveal differences in mobility rates between species (P > 0.05 in all comparisons), but differences between study periods were statistically significant. For all species considered together, high mobility rates were more frequent in 2017–2018 (12%) than 2015–2016 (1%), (Fisher’s exact test, P = 0.0002, Bonferroni corrected significance level, P = 0.016) and low mobility rates were more frequent in 2016–2017 (68%) than 2017–2018 (45%), (Fisher’s exact test, P = 0.004). The largest documented distance over which mussels moved during this study was four channel sections (approximately 20 m), covered by four A. anatina individuals between 2015 and 2018. Channel sections in which individual mussels were recovered over the study period are given in SI Table 1.

Recolonisation

Recolonisation proceeded from both sides of the channel (Fig. 3). Relative frequencies of the recolonising mussels were consistent with the relative frequencies in the adjoining habitats. Recolonisation by A. anatina and A. cygnea proceeded from both sides of the channel, but was more pronounced from the moat. Recolonisation by U. pictorum and U. tumidus proceeded mostly from the Biebrza River. Active migration of adult individuals contributed to the observed patterns, but the presence of young mussels confirmed that new recruitment occurred in all the species.

Figure 3.

Number of mussels recovered in the 5-m channel sections (S01–S15). Grey bars, mussels tagged and released in 2015 and subsequently recaptured in 2016–2018; black bars, mussels captured for the first time in 2017 and 2018. Note different scales on the y-axes.

Discussion

In this study, we assessed the suitability of re-introducing mussels collected from a waterbody before bottom dredging and releasing them post-dredging to the same site as a possible mitigation conservation measure. Mussels were individually tagged and monitored over three years. This timeframe allowed us to reliably capture the effects of the reintroduction and contribute much-needed data on how the restoration project performed^20,21. Mussel survival remained high throughout the study. No increased mortality in the year following reintroduction was observed, and declines in survival probabilities over time were consistent with natural mussel mortality. There was no differential survival among species. Although our results are based on recovery rates that were relatively low compared to PIT-tagged mussels^22,23, they clearly indicate a long-term survival of a large proportion of reintroduced mussels in spite of drastically changed habitat conditions. We did not find evidence of growth retardation following mussel reintroduction. Rather, growth rates were decreasing over time in all species, as might be expected in an ageing mussel cohort. The effect of age was highlighted by the rapid growth observed in the A. anatina sample collected two years after channel dredging and consisting largely of young individuals. A similar pattern was observed in A. cygnea but without statistical significance, probably due to a small sample size. The reintroduced mussels displayed low mobility, with the majority of mobility records indicating that an individual remained in the same channel section in which it was released. While a slightly increased mobility was observed in the third year after reintroduction, we did not find evidence of elevated mobility rates directly after reintroduction.

Recolonisation of river sections after restoration or disturbance has been shown to depend on the species pool in the immediate surroundings^24,25. The recolonisation patterns in our study, consistent with the composition of mussel communities in adjacent unaffected habitats, suggest that the newly settled mussels were brought to the site as larvae by fish from the adjacent habitats and underscore the importance of connectivity of freshwater ecosystems. It was not possible to determine whether the reintroduced mussels successfully reproduced. Recruitment of unionid mussels translocated in other studies was observed downstream of their release site^26,27, and it is likely that mussels reintroduced here also served as a source of juveniles that settled elsewhere. With this intervention, the continuity of populations was recovered. In the case of unionid mussels characterized by limited dispersal capacities and patchy distribution, salvaging individuals from a place of habitat disturbance and reintroducing them afterwards can serve as in situ genetic resource conservation and prevent the elimination or mixing of locally adapted ecotypes^16,18.

According to The IUCN Guidelines for Reintroductions and Other Conservation Translocations, any proposed translocation should have a comprehensive risk assessment with due consideration given to the focal species, their associated communities and ecosystem functions in both source and destination areas¹⁷; potential transmission of pathogens is of growing concern^12,28. Possibly the most challenging aspect is the assignment of appropriate receiving sites. In the case of unionid mussels it is the most decisive factor for the survival and establishment of the populations^29–31. While careful planning is recommended^17,18, it is often not possible in emergency-driven translocations. Our study addresses such emergency-driven translocations and poses that returning mussels to their site of origin directly after the disturbance can be a viable option. This approach, although not risk-free, offers some benefits compared to translocations to new sites. While certain aspects of the habitat changed drastically, others remained largely unchanged, including microclimate, water chemistry, nutrient availability and host-fish community. Appropriateness of the founders in terms of genetic provenance, morphology, physiology and behaviour was not an issue, disease and parasite transmission did not need to be considered, and there was no associated invasion risk. The risk from inaction was clearly higher than from this conservation action. Our study also matched the reality of mitigation translocations in that the decisions had to be taken on short notice and the intervention carried out under time pressure, with the preferred options often not available. For example, providing sufficient amounts of river water was practically impossible, so we kept the mussels in containers with tap water, which proved to be sufficient for their short-term survival.

Our study was based on a natural field experiment, with limitations inherent to this approach. For example, environmental conditions fluctuated over the study period. In the first study year, an extreme drought resulted in low water levels and high mussel mortality, which was one of the incentives for the restoration work. In contrast, during the monitoring years, water levels were high due to above-average rainfalls. Other limitations might include fluctuations in mussel populations due to unrecognized biotic interactions. Taken together, however, our results indicate that reintroducing mussels to the same site directly after disturbance can serve as an effective measure in mitigation conservation and can be preferable to translocation, particularly when carried out under time pressure limiting the possibilities of assigning appropriate receiving sites.

The findings of this paper indicate that when dredging is unavoidable, it should be carried out in short interspaced stretches ensuring the survival of donor populations, and at time intervals sufficient for the re-establishment of populations affected by the disturbance. Our study also highlights the need to include all native unionid mussels in national protected species lists. As bottom dredging usually occurs outside protected areas, the protection of the mussels can provide a legal framework for permissions for bottom dredging being conditional on the assessment of environmental impact and restoration with a focus on unionid mussels.

Methods

Study area and study design

The study site was a 75 m long channel connecting a former moat with the Biebrza River. Before dredging, the free-flowing water was approximately 50 cm wide and up to 20 cm deep. The underling sediments were approximately 100 cm deep. After dredging, the channel was 5 m wide and 3 m deep. Prior to the intervention, we divided the channel into 15 sections, each 5 m long. From each section, we collected all live mussels we found during an eight person-hour search and kept them in separate mesh bags. We transported the mussels to a field station nearby (approximately 15 min drive) and kept them in plastic containers filled with tap water at ambient temperature. We measured the shell length with electronic callipers and marked all mussels with individual shellfish tags (Hallprint, Australia) using superglue to affix them (Supplementary Figure S1 online). The dredging was completed the next day in the evening. We waited overnight for the water to clear a little and returned the mussels to the channel approximately 48 h after collecting them. Each individual was released to the same section of the channel from which it was collected. We also carried out timed searches in the sections of the Biebrza River and the moat adjacent to the channel (20 person-minutes scuba diving in each of those habitats) to estimate the relative mussel frequencies. We monitored the mussels released to the channel for three consecutive years, in the summer of 2016, 2017 and 2018. In each of those years mussels were searched for on one occasion for three hours by two scuba divers and two persons collecting from the channel bank (a total of 12 person-hours). Mussels from each channel section were kept separately and after measurements and tagging of newly found individuals were returned to the channel sections from which they were recovered.

Data analysis

Recovery rates were calculated relative to all released mussels after accounting for individuals found dead the previous year. Survival probabilities were estimated with the Kaplan–Meier analysis. Mussels were treated as alive until the last re-capture date and then censored. The equality of the survival functions between species was tested with the log-rank test.

Mobility was assessed for each individual recovered in at least two consecutive years; only individuals alive when recovered were included. Mobility rates were considered low for mussels recovered in the same channel section in which they were released the previous year, medium for mussels recovered in an adjoining section, and high for mussels recovered in one or more sections away from the one in which they were released.

Differences in mussel recovery and mobility rates between species and study years were compared with log-linear models for multidimensional contingency tables, followed by the Fisher's exact test with Bonferroni correction for multiple comparisons. Shell growth was calculated as a percentage increase of shell length in individuals for which measurements were available for at least two consecutive years; only individuals alive when recovered were included. Differences in shell length and growth rates were compared with the Kruskal–Wallis test, followed by Dunn's multiple pairwise comparisons with Bonferroni correction for multiple comparisons. The log-linear models were calculated according to Zar³², the remaining analyses were carried out in XLStat 2020 (Lumivero, France).

Author contributions

Conceptualisation: M.O., M.U., A.K.; Formal analysis: M.O.; Investigation: M.O., M.U., U.B.B., P.M., K.T., A.K.; Methodology: M.O., M.U.; Visualisation: M.O., A.K.; Writing—original draft: M.O.; Writing—review and editing: M.O., M.U., U.B.B., P.M., K.T., A.K.

Funding

Polish Ministry of Science and Higher Education (Grant No. WZ/WB-IIŚ/3/2023 to Bialystok University of Technology).

Data availability

All data generated or analysed during this study are included in this published article and its Supplementary Information files.

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-67836-7.