Behavioural Responses of Cerastoderma edule as Indicators of Potential Survival Strategies in the Face of Flooding Events

According to climate change scenarios the incidence of extreme events, such as flooding, is expected to increase worldwide. In the current climate change context, understanding behavioural responses of marine species to such stressors is essential, especially for species of high ecological and economic interest such as bivalves, which can be quite useful for future management and conservation actions. In this study, a laboratory experiment using different salinity conditions was undertaken to assess potential behavioural responses of cockles (Cerastoderma edule), as a survival strategy facing low-salinity stress during riverine flood events. Results showed consistent patterns of burrowing/emergence of cockles facing salinity variation: with high salinities the individuals were observed buried in the sediment; when salinity decreased, organisms were observed to actively emerge, and when salinity was <10, cockles were found exposed at the sediment surface. These behavioural changes may be a strategy for the survival of this species in response to flooding: once at the sediment surface, hydrodynamics may transport organisms towards areas that are more suitable


Introduction
Global climate change is one of the main challenges the world faces, seriously affecting coastal ecosystems. In particular, the increasing occurrence and intensity of extreme climatic events (e.g., floods, droughts and heat waves) often lead to abrupt changes in temperature, salinity and hydrodynamic conditions [1]. This results in impacts on organisms (e.g., physiological processes, behaviour, and mortality), leading to changes in population abundance, community structure and ecosystem functioning [2]. Since this trend is expected to continue, further impacts on aquatic ecosystems are predictable [3,4]. As such, understanding how biological communities respond to these scenarios is essential for coastal ecosystem management.
Bivalves are key species of estuarine and coastal macrofaunal communities, since they usually are: (i) long-lived and large-sized species dominating the assemblage biomass [5]; (ii) highly abundant and productive [6]; (iii) economically valuable as a food resource, being frequently harvested and produced in aquaculture [7][8][9]. Moreover, they perform essential ecological functions in the ecosystem, as they influence the benthic-pelagic interface through sediment reworking, resuspension and filtration activities, promoting nutrient regeneration and contributing to water column purification [9][10][11]. As such, they play a key role in the ecosystem food web [12].
C. edule is usually subjected to great physiological stress related to high gradients and variability of abiotic factors, typical of transition habitats [27,28]. Salinity is a key physical factor for estuarine bivalves and determine the limits of their geographic distribution and biological features [16,29,30]. As an euryhaline species, C. edule can survive in a wide range of salinities, being commonly found in habitats where salinity changes regularly (e.g., estuaries) with daily tidal cycles and seasonal freshwater inflows driven by precipitation. However, the increasing frequency and intensity of extreme climatic events is additionally demanding for these organisms [30][31][32][33], which must endure augmented osmotic stress [27,28]. Short-term (tidal) and long-term (rain periods) salinity changes generate an osmotic gradient between the ambient medium and the organisms, often leading to behavioural and physiological responses [34][35][36] and ultimately to mortality episodes [8,19,37].
In a previous laboratorial experimental study, Verdelhos et al. [16] showed that C. edule is negatively affected by low salinities, being particularly vulnerable to floods. One of the main observations was the absence of burrowing behaviour on salinity treatments <10 [16]. At first, this could be considered as a simple protection strategy: organisms submitted to low salinity close their valves and suspend their activity, shielding themselves from the osmotic stress. However, C. edule specimens in nature were observed at the sediment surface during a period of high freshwater inflow-low salinity (personal observation), which suggests an emergent movement from the sediment towards the surface. These in situ observations raised a question that motivated the present study: does C. edule have a behavioural strategy to cope with salinity declines during riverine flooding events? This work aimed to evaluate C. edule's burrowing/emergence patterns when subjected to salinity variations, observing if the individuals: (a) burrow in the sediment at higher salinity; (b) emerge with decreasing salinity; (c) re-burrow when salinity increase; (d) show consistent behaviour when salinity variation is repeated.

Study Site and Sampling
The Mondego estuary (8.6 km 2 ) is a warm-temperate intertidal ecosystem on the Atlantic coast of Portugal (40 • 08 N, 8 • 50 W). It comprises two arms separated by an alluvium-formed island ( Figure 1) and has a mean water flow of 79 m 3 .s −1 , which in rainy years can reach above 140 m 3 .s −1 , dropping to 27 m 3 .s −1 in dry years [38]. Freshwater outflow moves mainly via the northern arm, while in the southern arm water circulation is more dependent on tides and on the freshwater input from the Pranto River, a small tributary with a flow controlled by a sluice. Over the last few decades, there was an increase in the occurrence and intensity of flooding events characterized by intense freshwater flow, increased turbidity and abrupt salinity declines, with impacts on bivalve populations and macrobenthic communities [15,38].
C. edule is mainly distributed in downstream areas, specifically on an intertidal sandflat on the South arm [16,17,39] and subtidal areas on both arms [38]. Salinity is usually high in these areas, ranging between 30 and 34, during high tide, although they may be subjected to lower salinity values during low tides (ranging between 10 and 30) and high riverine freshwater flow. Adult specimens (average length 29.63 ± 1.82 mm) were collected by hand, at low tide during Spring (Water temperature = 19.2 • C; Water salinity = 20.7), on the South arm intertidal sandflat (composed by 72% fine sand and 23% medium sand, with~0.8% organic matter content [39]). They were transported in cooled boxes with local water (salinity~18-22; practical salinity scale), acclimated to laboratory conditions for 48 h (temperature = 20 • C; salinity = 20; 12 h light/dark; continuous aeration) and kept in starvation to equalize the hunger state among individuals [40]. Sediment was collected at the site, transported to the laboratory and litter, shells, boulders and large biological structures were removed by hand and sediment was sieved through a 4 mm mesh. Seawater (salinity~33-35) was collected, transported to the laboratory, and filtered using GC-50 glass fiber membrane filters. C. edule is mainly distributed in downstream areas, specifically on an intertidal sandflat on the South arm [16,17,39] and subtidal areas on both arms [38]. Salinity is usually high in these areas, ranging between 30 and 34, during high tide, although they may be subjected to lower salinity values during low tides (ranging between 10 and 30) and high riverine freshwater flow. Adult specimens (average length 29.63 ± 1.82 mm) were collected by hand, at low tide during Spring (Water temperature = 19.2 °C; Water salinity = 20.7), on the South arm intertidal sandflat (composed by 72% fine sand and 23% medium sand, with ~0.8% organic matter content [39]). They were transported in cooled boxes with local water (salinity ~18-22; practical salinity scale), acclimated to laboratory conditions for 48 h (temperature = 20 °C; salinity = 20; 12 h light/dark; continuous aeration) and kept in starvation to equalize the hunger state among individuals [40]. Sediment was collected at the site, transported to the laboratory and litter, shells, boulders and large biological structures were removed by hand and sediment was sieved through a 4 mm mesh. Seawater (salinity ~33-35) was collected, transported to the laboratory, and filtered using GC-50 glass fiber membrane filters.

Experimental Procedures and Statistical Analysis
A laboratorial experiment was conducted to assess C. edule's burrowing vs. emergence response under different salinity conditions, simulating low-salinity stress during riverine flood events. Specimens (Ntotal = 180) were exposed to different salinity variation treatments at 20 °C for 156 h (Table 1). In the Control treatment (Treatment C), salinity was constant during the test (= 20), but the water in the tank was changed every 12 h. In Treatments A and B, organisms were submitted to each salinity level for 12 h, starting at salinity = 20 and varying 5 units each time the water in the tank was changed (12 h). Treatment A consisted on: (i) initial salinity decrement from 20 to 5 (0 to 48 h); (ii) salinity increase from 5 to 20 (48 to 84 h); (iii) decrease from 20 to 5 (84 to 120 h); (iv) maintenance at 5 until 156 h. Treatment B comprised: (i) initial salinity decrement from 20 to 5 (0 to 48 h); (ii) salinity increase from 5 to 20 (48 to 84 h); (iii) decrease from 20 to 5 (84 to 120 h); (iv) final increase up to 20 until 156 h. For further analysis of the results obtained

Experimental Procedures and Statistical Analysis
A laboratorial experiment was conducted to assess C. edule's burrowing vs. emergence response under different salinity conditions, simulating low-salinity stress during riverine flood events. Specimens (N total = 180) were exposed to different salinity variation treatments at 20 • C for 156 h (Table 1). In the Control treatment (Treatment C), salinity was constant during the test (= 20), but the water in the tank was changed every 12 h. In Treatments A and B, organisms were submitted to each salinity level for 12 h, starting at salinity = 20 and varying 5 units each time the water in the tank was changed (12 h). Treatment A consisted on: (i) initial salinity decrement from 20 to 5 (0 to 48 h); (ii) salinity increase from 5 to 20 (48 to 84 h); (iii) decrease from 20 to 5 (84 to 120 h); (iv) maintenance at 5 until 156 h. Treatment B comprised: (i) initial salinity decrement from 20 to 5 (0 to 48 h); (ii) salinity increase from 5 to 20 (48 to 84 h); (iii) decrease from 20 to 5 (84 to 120 h); (iv) final increase up to 20 until 156 h. For further analysis of the results obtained in Treatment A and Treatment B "salinity variation cycles" will be considered, i.e., in Treatment A we consider a 1st cycle from 0 to 84 h consisting on a salinity decrease/increase, followed by a 2nd cycle from 84 to 120 h with a salinity decrease (until 120 h), remaining = 5 until 156 h ("salinity maintenance"); in Treatment B we consider two cycles of salinity decrease/increase: the 1st from 0 to 84 h and the 2nd from 84 to 156 h.
For each treatment (Treatment A, Treatment B and Control) three runs were performed, using two tanks per run. Each tank (26 × 36 × 22 cm) contained a 5 cm layer of sediment and 15 L of water (filtered seawater was diluted with distilled water to reach the correct salinity). In each tank, 10 specimens were individually placed in submerged perforated cups that allow water flux and enable salinity balance between the water and the sediment ( Figure 2). Every 12 h the specimens' burrowing condition (e.g., buried in the sediment vs. at the surface) was observed and registered (cockles were considered burrowed when they had totally or partially (more than 2/3) burrowed into the substratum). Tanks were continuously aerated; salinity and temperature values were monitored and adjusted if needed to maintain constant conditions; and specimens were fed daily ad libitum (with "Ocean Nutrition Microplankton" composed by phyto and zooplankton, diluted in the experimental water).  A one-way analysis of variance (ANOVA) with repeated measures was performed for each treatment to test for significant differences on the observed burrowing response (dependent variable) between the different salinity levels (independent variable), using A one-way analysis of variance (ANOVA) with repeated measures was performed for each treatment to test for significant differences on the observed burrowing response (dependent variable) between the different salinity levels (independent variable), using IBM SPSS Statistics 19.0. To avoid pseudoreplication issues, since individuals in the tank are not totally independent (pseudoreplicates), analysis was performed considering each tank as a replicate and the proportion of buried individuals in each tank (i.e., p = 1, if all the individuals were buried; p = 0.5 if five individuals were buried). Thus, six replicates were considered on each treatment (n • tanks per treatment), taking into account that the experimental conditions on each tank and each run were the same. Sphericity was tested using Mauchly's test as part of the GLM Repeated Measures procedure, and after a significant F test, differences among means were identified using the Bonferroni post hoc procedures. Moreover, the relationship between the % of buried cockles and salinity levels was estimated using a linear regression. Shapiro-Wilk and Levene tests were used to check for the normality of the data and homogeneity of the variances, respectively.
Burrowing differences between the distinct salinity levels in each treatment were assessed using an ANOVA test with repeated measures. As our data did not meet the sphericity assumption (Mauchly's test with p < 0.05), values with a Greenhouse-Geisser correction were considered (Table 3). Overall, the percentage of buried individuals was significantly different between salinity levels on treatments A and B (Table 3)

Discussion
Cerastoderma edule exhibited a burrowing vs. emergence pattern when subjected to salinity variations: (a) at the initial experimental salinity (= 20) individuals burrowed in the sediment, quite swiftly in some cases; (b) they displayed an emerging response during decreasing salinity trend; (c) specimens showed a burrowing response in the course of salinity increment; (d) when salinity variations were repeated the same behavioural responses were observed, however their magnitude was higher on the 1st salinity variation cycle when compared to the 2ndcycle. A clear relationship between changes in salinity and behaviour was observed, with cockles actively burrowing in the sediment column when water salinity is within the optimal performance range for this species [16] and emerging when salinity drop below a low salinity threshold (<10). Although this pattern was consistent throughout the experiment, the observed lower magnitude on the 2nd cycle of salinity variation may indicate an altered physiological condition of specimens subjected to experimental conditions for too long [16,29,41,42].
Salinity is one of the factors determining the biological features and spatial distribution of species on estuarine habitats [16,29,30]. In the Mondego estuary, increased mortality of the common cockle C. edule and the peppery furrow shell Scrobicularia plana was observed during flooding events, impacting the population structure and dynamics of these bivalves [8,43,44]. Low salinities also modified their behaviour (e.g., affecting burrowing/emergence behaviour, feeding activity, movement and valves opening/closure), impairing their overall activity [16]. Conversely, these species showed the best performance within a narrow optimal salinity range (20-25 for C. edule; 20-30 for S. plana) [16], and C. edule population occupy preferentially downstream intertidal and subtidal sandflats [16,39].
C. edule is better suited to higher salinity areas [7,16] and is impaired by abrupt salinity declines [16], especially during long and intense flooding episodes [8,24,43]. Under these circumstances, organisms respond initially with behavioural changes, which in the event of prolonged exposure result in mortality [16]. On previous experimental studies [16] the absence of burrowing behaviour under low salinity conditions was considered a protection mechanism related to valve closure, retraction of sensitive body parts and an overall inactivity to avoid osmotic stress [16,29,34,36,45]. In an in situ personal observation during low tide, numerous cockles were found at the sediment surface in the Mondego south arm intertidal sandflat during a period of high freshwater inflow. This suggests that buried organisms actively emerged to the sediment surface during adverse conditions, putting themselves at risk.
The pattern observed in this study seems to indicate a behavioural strategy to endure severe salinity declines. During riverine flooding events, increased freshwater inflow leads to reduced salinity and improved hydrodynamics, promoting sediment erosion and bedload transport of materials, with associated impacts on the populations of C. edule [8,43,45,46]. In these conditions, cockles can be involuntarily or passively mobilized by currents or when the sediments they live in become eroded [47][48][49][50]. With the observed behaviour, cockles emerge at the sediment surface when salinity drops (<10), becoming more likely to be dragged to further downstream areas, where salinities are higher ( Figure 5). As such, it can be viewed as an avoidance mechanism allowing organisms to avoid impacts associated with severe salinity declines. Dispersion of adult cockles has been previously observed in a field experiment study, with individuals being dragged "upslope" by the tide over distances up to 200 m [51]. Avoidance can be: (i) active-ability to detect environmental stressors and to move Avoidance can be: (i) active-ability to detect environmental stressors and to move (by swimming, flying, walking) towards less stressed environments or to change the behaviour (e.g., unburrowing) [52][53][54][55]; (ii) passive or drift-consists in the passive displacement of aquatic organisms in running waters [56]. Macroinvertebrates dispersion to downstream areas by means of drift has been widely documented [57][58][59][60] and many factors, such as predation, population density, current/discharge, photoperiod, water chemistry, lifecycle stage, genetic, food availability and quality, and contamination seem to induce their displacement [57,58,61].
How exactly C. edule can detect salinity changes, and the mechanism underlying the observed behavior, are unclear and outside the scope of this work, although we suspect it could be triggered in a physiologically similar way to what was observed by Peteiro et al. [62] when investigating drifting capability and physiological response to salinity stress. Regardless the mechanism, by adopting this behavioural strategy organisms increase their chance to find suitable conditions downstream for their survival, although at the expense of longer exposure to pelagic predators and the risk of being transported too far.
The idea that living adult C. edule can be transported by improved hydrodynamic conditions (e.g., storms), facilitated by its bulbous-shaped shells, is not entirely new and has long been documented [63][64][65]. Cadée [65] recently highlighted that adult cockles of C. edule could be eroded from a tidal flat on Texel (Wadden Sea) by storm events and be naturally transported by rolling over to a nearby beach, proving this can indeed occur in the field. As well, Anta et al. [50] reported the bedload transport of large amounts of C. edule in the Ulla River estuary (Galicia, Spain) after severe storm events. Evidence of escaping to reduce the exposure to stressors has also been reported and pointed out as beneficial in previous bivalve studies in order to avoid mass mortality episodes and recruitment failures [62].

Conclusions
Overall, our laboratory experiment highlighted a vertical migratory behaviour of cockles in the sediment column with changes in the water salinity. At low salinity, cockles were observed emerging at the sediment surface, while at high salinity they burrow in the sediment. This migration pattern could be a behavioural strategy for cockles' survival when facing flood events, consisting of avoidance of the stressful and impairing low salinity conditions allowing themselves to be dragged to higher salinity areas downstream where environmental conditions are favourable. This strategy may increase C. edule's survival probabilities under extreme scenarios, such as riverine flooding events on estuarine areas, representing an advantage on the coastal environmental change scenarios. Nevertheless, the sustainability of this behaviour and physiological consequences must be further investigated.
Findings of the present study constitute a first step to bring new insights regarding the effects of climate change stressors (floods) on adult bivalves' behaviour and to highlight the relevance of in situ monitoring to investigate the impacts of environmental change on key species. Future studies should consider and expand knowledge on species specific behavioural responses when planning and implementing management solutions for current and future bivalve fisheries and conservation/restoration projects worldwide.

Data Availability Statement:
The data presented in this study can be replicated from it.