Ecological Indicators 111 (2020) 106031 Contents lists available at ScienceDirect Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind Original Articles Sublethal responses of four commercially important bivalves to low salinity a,⁎ a b b a Rula Domínguez , Elsa Vázquez , Sarah A. Woodin , David S. Wethey , Laura G. Peteiro , Gonzalo Machoa, Celia Olabarriaa T a Departamento de Ecoloxía e Bioloxía Animal, Facultade de Ciencias do Mar, Campus As Lagoas-Marcosende s/n and Centro de Investigación Mariña, Universidade de Vigo, Illa de Toralla s/n, 36331 Vigo, Spain b Department of Biological Sciences, University of South Carolina, 715 Sumter Street, Columbia, SC 29208, United States A R T I C LE I N FO A B S T R A C T Keywords: Bivalves Small-scale fisheries Salinity fluctuations Valve closure Scope for growth Burrowing The abilities of estuarine species to respond to salinity fluctuations by behavioural and physiological responses can determine the maintenance of populations, particularly in the context of climate change. The native clams Ruditapes decussatus and Venerupis corrugata, the native cockle Cerastoderma edule and the introduced clam Ruditapes philippinarum are important resources in Galician (NW Spain) coast. As inhabitants of estuaries, these species are exposed to frequent salinity fluctuations as a result of heavy rains. This study investigated the shortterm sublethal effects of salinity drops on their physiological (scope for growth, SFG) and behavioural (valve closure and burrowing activity) responses. Bivalves were exposed to simulated tidal cycles and similar salinities to the field conditions, i.e., four salinity ramps (5–20, 10–25, 15–30 and 30–30) during six days over three different periods of the year (autumn, winter and spring). The overall response was the same for all species under the lower salinities (5, 10 and 15), with a dramatic reduction of pumping activity, SFG and burrowing. Results differed among species under the higher salinities (20, 25 and 30). While C. edule was the most affected species in autumn showing no recovery despite having higher SFG compared to the venerids, R. decussatus was more resistant in all seasons despite having the lowest SFG compared to the rest of species. In winter, V. corrugata was the most vulnerable due to lower SFG at the lowest salinities. All species showed a compensation pattern in spring that led to non-recovery of individuals. Burrowing ability had similar patterns to SFG in autumn and winter but differed in spring, when recovery was the general pattern. The decrease of burrowing ability at lower salinities during stress seen to some degree in all species can increase vulnerability to predation. Results suggest that differential responses of lower activity over time could be related to the physiological condition and habitat preferences of each species and should be taken in consideration for management plans in the context of climate change. The results drive a discussion of the usefulness of SFG as the metric with which to assess salinity stress in adult bivalves and the need in future research to increase frequency and duration of stresses in concert with variables such as food availability. 1. Introduction Bivalve shellfisheries are an important component of the world’s fishery production in Europe (FAO, 2018). Mostly because of their sensitivity to environmental conditions and changes in climate, these fisheries experience high spatial and temporal variability of catches in many parts of the world (reviewed by McLachlan and Defeo, 2018) and particularly in Spain (Juanes et al., 2012; Parada et al., 2012; Morgan et al., 2013; Aranguren et al., 2014). Most shellfish beds are located in estuarine areas in the inner part of rias (sensu von Richthofen, 1886) and thus exposed to large fluctuations of salinity due to heavy rains and locally managed freshwater releases from river dams, both associated ⁎ with increased river runoffs (Nogueira et al., 1997; Parada and Molares, 2008; Parada et al., 2012). Of special concern are the extreme precipitation events that are expected to increase in frequency in the near future as estimated by IPCC current predictions of global and regional climate change (Hoegh-Guldberg et al., 2018). Such episodes of heavy rainfall occurred often in Galicia, being the most dramatic during autumn-winter 2000–2001, which caused massive mortality of bivalves (100% mortality in several areas), preventing these resources from being exploited for 1 year (Parada et al., 2012). Episodes of mass mortality and recruitment failures in bivalve populations can thus lead to huge socio-economic costs (Sobral and Fernandes, 2004; Parada et al., 2012; Verdelhos et al., 2015), as well as ecological effects as Corresponding author. E-mail address: ruladominguez@uvigo.es (R. Domínguez). https://doi.org/10.1016/j.ecolind.2019.106031 Received 12 May 2019; Received in revised form 14 November 2019; Accepted 17 December 2019 1470-160X/ © 2019 Elsevier Ltd. All rights reserved. Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. exposures (Molares et al., 2008), although a more recent study using 6 day exposures reported mortalities only at salinities below 7 (Carregosa et al., 2014). In R. philippinarum, mortality was observed at salinities below 14 after 6 days and rates of mortality were greater than those for R. decussatus with further salinity reductions. These data are consistent with field distributions where R. decussatus is more abundant in fresh water influenced areas than R. philippinarum (Juanes et al., 2012). A series of short-term mesocosm experiments simulating field salinity stress were run in sediment with tidal cycles. Exposures to salinities below 20 occurred during ebb and low tide, periods when river flow and runoff predominate in the field. Despite previous studies on these and other bivalve species (Sunoko, 1997; Resgalla et al., 2007; Sará et al., 2008; Wang et al., 2011; Guzmán-Agüero et al., 2013), the effects of salinity changes of different magnitude on physiology and behaviour of these four species are poorly understood (Verdelhos et al., 2015; Gharbi et al., 2016; Peteiro et al., 2018), especially when considering realistic scenarios (Nossier, 1986). We tested the cumulative effect of exposure to low salinities over six consecutive days and two days of recovery by evaluating scope for growth and activity in terms of valve closure and burrowing activity of the four species. The hypothesis tested was that lower salinities would have measurable impacts: increasing valve closure, decreasing scope for growth and decreasing burrowing activity compared to bivalves in the control salinity. We expected different responses among species related to their different physiological tolerances and behavioural strategies. Based on mortalities from continued stress experiments and field data, the tolerance to low salinity was expected to be lowest for V. corrugata and C. edule, intermediate for R. philippinarum and highest for R, decussatus (Molares et al., 2008; Carregosa et al., 2014; Verdelhos et al., 2015). To check for consistency of responses over time, the experiment was repeated in autumn, winter and spring, when low salinity events occur in Galician shellfish beds, as responses are likely to vary depending on the physiological conditions of organisms (Beninger and Lucas 1984; Sará et al., 2008; Aníbal et al., 2011). disruptions of food webs, biogeochemical cycles or pelagic-benthic coupling that result in changes in structure and functioning of communities (Morgan et al., 2013; Haider et al., 2018). Sudden salinity drops can also occur in combination with temperature changes, parasite infections or variations in the food supply, enhancing their impact on species (Aranguren et al., 2014; Villalba et al., 2014; Macho et al., 2016). Salinity fluctuations not only cause mortality but also important sublethal effects, such as altering feeding, respiration, growth, osmoregulation, behaviour, reproduction, and parasite-disease interactions (Gosling, 2015), even in euryhaline species such as the studied venerids and cockles (Parada and Molares, 2008; Carregosa et al., 2014). Bivalves have the ability to adapt to small changes in salinity in a few days, but sudden unforeseen changes produce negative effects, quantified depending on the authors, as loss of 1 to 24% of the energy acquired (Gosling, 2015). Salinity drops can be either regular or episodic and induce different types of responses in marine organisms; behavioural responses may be the primary strategy under stress (Lockwood, 1976). Among them, valve closure for short periods is common for bivalves frequently exposed to salinity fluctuations over cyclic periods to avoid osmotic shock (Nossier, 1986; Verdelhos et al., 2015). These species often can withstand hypoxia for short periods, minimizing the need for energetic adjustments due to osmotic stress (Shumway, 1977); if closure is extended in time, filtration activity and therefore energy acquisiton are reduced, while excretion products are accumulated. In contrast, if tissues are not isolated by valve closure, changes in salinity can cause reductions in activity and energy acquisition due to an increase in energy demand to maintain cell volume and avoid osmotic shock (Burton, 1983; Berger and Kharazova, 1997; Gosling, 2015). Burrowing behaviour can be also affected by salinity fluctuations because it is an activity of high energetic cost, so after repeated disturbance burrowing may become limited when the energy reserves are exhausted below a certain threshold, with potential implications for predation risk (Haider et al., 2018). Changes in energy balance due to salinity stress can be assesed by the calculation of scope for growth, SFG, i.e. energy available for growth (Griffiths and Griffiths, 1987; Sará et al., 2008; Wang et al., 2011), which is an useful indicator of bivalve physiological responses in short term laboratory experiments (Widdows, 1978; Bayne, 1998; Widdows and Staff, 2006; Resgalla et al., 2007; Guzmán-Agüero et al., 2013). It integrates the difference between the energy absorbed from the food (clearance rate and absorption efficiency) and the energy loss via metabolic energy expenditure (respiration and ammonium excretion). In addition, valve closure, monitored during physiological measurements (Stenton-Dozey et al., 1994; Sobral and Fernandes, 2004; Carregosa et al., 2014) can be considered as a different response variable since, while bivalves remain closed, their activity in terms of feeding and aerobic respiration is theoretically stopped (Griffiths and Griffiths, 1987). The present study investigated the behavioral and physiological responses of the four most relevant bivalve species for the traditional fishery in Galicia (northwestern Spain), to salinity decreases of different intensity over a short-term period. The species selected include the native venerids Venerupis corrugata (Gmelin 1791) (pullet carpet shell), Ruditapes decussatus (Linnaeus 1758) (grooved carpet shell), the introduced Ruditapes philippinarum (Adams and Reeve 1850) (Japanese carpet shell), and the native cockle Cerastoderma edule (Linnaeus 1758) (common edible cockle). The four species represented 78% of the landings in 2018 and yielded 9046 tonnes of bivalves worth ~84 million € (own elaboration based on www.pescadegalicia.gal, access January 2019). They differ in their salinity tolerance ranges and optima. Venerupis corrugata has low tolerance to low salinity (almost 100% mortality at < 10) (Molares et al., 2008) while Cerastoderma edule shows a similar pattern to V. corrugata with an optimum salinity range of 20–25 and high mortality at < 10 (Verdelhos et al., 2015). Salinities below 20 cause mortality in R. decussatus under long 2. Materials and methods 2.1. Clam collection and maintenance Three experiments were performed in a mesocosm system at Estación de Ciencias Mariñas de Toralla (ECIMAT) (www.ecimat.uvigo.es) of the Universidade de Vigo (Spain), during autumn 2015 (Dec15), winter and spring 2016 (March16 and May16, respectively). These are the seasons when heavy rains tend to occur in this region, with a high impact on the shellfish beds (Parada et al., 2012). The day before each experiment, adult clams were manually collected in intertidal fishing beds by shellfishers at Ría de Arousa (42° 30′ 55″ N, 08° 48′ 53″ W) and adult cockles at Ría de Noia (42° 47′ 0″ N, 8° 53′ 0″ W), and transported to the laboratory in refrigerated boxes. Both locations are located at the Rías Baixas, characterized by its temperate and humid weather, influenced by the upwelling regime typical of this coast due to their SW- NE orientation (Méndez and Vilas, 2005). Shellfish beds are influenced by the mesotidal and semidiurnal tidal regime and river discharges because of their location in the inner part of the rías at shallow depths around 5–10 m (Alvarez et al., 2005). Once in the laboratory, bivalves were immediately placed in seawater and afterwards they were individually marked and measured (mean ± SD for Dec15, March16 and May16 respectively: 40.5 ± 2, 37.5 ± 1.8 and 39.8 ± 1.7 mm for V. corrugata; 42.4 ± 1.3, 45.02 ± 1.5 and 42.6 ± 1.5 mm for R. decussatus; 41.8 ± 1.1, 42.7 ± 1.2 and 41.4 ± 1.4 mm for R. philippinarum; 30.4 ± 1.1, 30 ± 1.2 and 31.3 ± 1.5 mm for C. edule), placed on sediment surfaces in 16 L tanks (17 cm tall × 26 cm width × 36 cm length) and allowed to burrow. Those that did not burrow within 8 h were discarded and replaced by new individuals before the experiments started. 2 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. The experimental treatment salinities were monitored by AT controllers (AquaMedic) in Dec15 and miniCTDs (Star ODDI) in March16 and May16. Data for a 24 h period in 30 min intervals indicated the effectiveness of salinity treatments (Fig. 2A and B for Dec15 and March16, respectively). Tides were simulated only during the day in Dec15, while in March16 and May16 there were two tides each 24 h. For each treatment the lower salinity occurred during the ebb and low tide period and the higher salinity during the flood and high tide period. Note that maximum water depth was the same during the periods called ‘low tide’ and ‘high tide’ in this work. After six days of exposure to these treatments, all tanks were maintained for two more days at constant salinity (30) to measure bivalve recovery ability (three days for recovery in Dec15). At initiation 28 (Dec15) or 30 (March16 and May16) individuals were placed on the sediment surface in the small 16 L tanks, one tank per species per larger tank. Densities were reduced by 6 on each day of measurement including day 0. As of day 0 the densities in the small tanks approximated those found in the shellfish beds (220 ind m−2). Fig. 1. Experimental design including the salinity ramps and blocks. 1: water pumps. 2: head tanks with fresh and salt water. 3: mixing tanks for salinity treatments (5–20, 10–25, 15–30, 30–30). 4: eight big tanks (two per salinity ramp). 5: four small tanks, one per species, within a big -tank, each small tank initially with 30 individuals. 6: Inlet. 7: tank overflow. 8: Burrowing buckets. 2.3. Measured variables A total of 2880 animals comprising the four species were used in the three experiments. Oxygen consumption, clearance rate, ammonium excretion, absorption efficiency, valve closure and burrowing behaviour were measured on 6 individuals per species per block and salinity treatment. Measurements for each block were taken on day 0 (prestress), day 4 (under stress), day 6 (last day of stress) and day 8 (day 9 in Dec15) (recovery period), during the last half of the flood tide, with the exception of day 4 in Dec15 when physiological variables were measured during the lowest salinity of each experimental ramp (ebbtide conditions) in order to test for the physiological limits of individuals. The same individuals were used consecutively for all physiological and behavioural measurements on the same measuring day and then removed from the experiment. During physiological measurements the activity of organisms, measured as valve closure, was checked every 5 min (every 20 min while individuals were in ammonium excretion beakers) and scored as follows: 0 valves closed, 1 valves open with or without siphons visible. Individual oxygen consumption rate (mgO2 h−1) was estimated by measuring oxygen fluxes every 60 s inside cylindrical respirometers (125 ml) connected to luminescent dissolved oxygen probes (Hach Lange HQ40D). Respirometers were filled with aerated 50 µm-filtered seawater at ambient temperature, 15–18 °C, depending on the season, using seawater at the treatment salinity, mixed gently with a magnetic stirrer during the measuring period. Dissolved oxygen concentration was recorded until it declined ≈ 20% from the initial value or after 20 min. Estimates were normalized by the volume of the seawater inside the respirometer. In addition, controls with only shells of each species maintained in each experimental treatment served as blanks to correct for respiration rates of bacterial film in the shell. Individual clearance rate (CR; L h−1) was measured after oxygen consumption measurements. Each individual was placed in a glass beaker with 500 ml of 50 µm-filtered seawater with aeration at each salinity treatment. After 5 min of acclimation, each individual was fed with Isochrysis galbana (≈ 50 000 cells ml−1 to avoid the production of pseudofeces, Sobral and Widdows, 1997). Water samples were taken at 5 and 30 min and preserved with lugol. Concentration of particles was measured in an electronic particle Coulter Counter (Beckman Coulter Multisizer 3). Two glass beakers without individuals, but with the same microalgal concentration were also sampled in each run of measurements to correct for particle loss due to sedimentation. Individual ammonium excretion rate (mg NH4-N h−1) was calculated following clearance rate measurements. Each individual was gently transferred to a new beaker with 200 ml of 50 µm-filtered sea water of each salinity treatment, and left for 2.5 h without aeration. The feces were collected by pipetting and all water was stored and frozen The 16 L tanks were placed inside eight 480 L (50 cm tall × 80 cm width × 120 cm length) experimental tanks with running 50 μm-filtered seawater (salinity approximately from 35 to 37 depending on the season) in a room with temperature set at 18 °C. Then animals were fed and kept overnight at those conditions until the experiment started the following day. Animals were fed in the evenings during the experiments with a microalgae mixture of Isochrysis galbana, Tetraselmis suecica, Chaetoceros gracilis and Rhodomonas lens. Several liters (volume varied daily depending on the concentration of the phytoplankton cultures) of the algal mix were added to each tank to reach a 1% maintenance diet for the clams’ density and size (mean individual dry weight of 0.68 g). 2.2. Experimental setup Four small plastic tanks, one per species, (16 L, 17 cm tall × 26 cm width × 36 cm length) were placed inside each of 8 big tanks (480 L, 50 cm tall × 80 cm width × 120 cm length) with bottom drains (Fig. 1) that were randomly located in the experimental room. There were two 480 L tanks for each salinity treatment (four salinity ramps, see below), one per block. Blocks 1 and 2 were measured on consecutive days due to time constraints for physiological measurements. The 16 L tanks had four 2 cm bottom orifices covered with 80 μm mesh to avoid sediment loss but to allow water flux through the column of sediment. They were filled to the top with sediment collected from the intertidal at Canido (42° 11.68′ N; 8° 47.81′ W) where clams live (median grain size of 0.19 mm). To produce salinity profiles similar to those experienced by bivalves in the field (www.intecmar.gal), treatments consisted of four salinity ramps in which salinity varied from 5 to 20, 10 to 25, 15 to 30 and 30 to 30, i.e. control treatment. Salinity ramps were created automatically through the use of timers controlling dual bellows pumps that mixed dechlorinated fresh water and 50 μmfiltered sea water at ambient temperature at different proportions (see Fig. 1). Water entered via inlets in the bottom of the large tanks and exited via ~ 30 cm tall standpipes. Total water exchange took ~90 min at 1 L min−1. During daylight hours the ebb tides were simulated every day in each tank by drainage followed by the automatic change in water source and thus salinity resulting in a rapid salinity change. During night tides the automatic change in water source occurred but was not preceded by drainage so salinity change occurred more slowly (Fig. 2). This was particularly true for salinity changes from a higher to a lower salinity because the incoming water was of lower density than the water in the tank. 3 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 2. Salinity profiles for each treatment during experiments. A) data for December 2015 (AT controller), B) data for March 2016 (miniCTD). Solid line represents the S30-30 treatment, dashed line represents the S15-30 treatment and dotted-dashed line and dotted line represent the S10-25 and S5-20 treatments, respectively. Data for May 16 are omitted because they followed a similar pattern to March16. Missing values in March16 are due to the period that the miniCTD was out of the water during the daytime tide. the difference between the energy absorbed from the food (food consumption × absorption efficiency) and energy loss via metabolic energy expenditure was determined following the equation given by Warren and Davis (1967): for analysis by electrometric determination (Carpenter and Capone, 1983). Afterwards, feces together with three replicates of food samples (20 ml) previously taken were rinsed with 0.5 M ammonium formiate to dissolve salt crystals and filtered throughout 25-mm prewashed and weighed Whatman GF/C glass filters. Those filters were dried for 24 h at 60 °C, weighed, then left in a muffle furnace for 2 h at 450 °C to determine ash free weight. Content of organic matter assimilated was calculated for each bivalve as the difference between ash free dry weight and dry weight. After physiological measurements were finished, individuals were carefully placed on the surface of buckets previously filled with clean sediment from the same source as that used in treatments to measure their burrowing activity. The sediment was prepared by exchanging the porewater several times followed by aeration to ensure that both the porewater and the overlying water were at the corresponding salinity treatment. Burrowing activity was recorded every 15 min during 2 h. Animals were scored as burrowed if ≥75% of the body was within the sediment. Finally, bivalve tissues were removed from their shells and dried to obtain their dry weight for the standardization of all estimates to 1 g dry weight. Scope for Growth P = A − (R + U), where A (energy absorbed) = (C) × food absorption efficiency. C (energy consumed or ingested) = [maximum clearance rate: L g−1h−1] × [mg POM L−1] × [18.74 J mg−1 POM] R (energy respired) = (µmoles O2 g−1 h−1) × 0.456 U (energy excreted) = (µmoles NH4-N g−1 h−1) × 0.349 Energy equivalents used to convert rates of oxygen uptake, clearance rate and excretion to joules were 1 µmol O2 = 0.456 J (Gnaiger, 1983), 1 mg algal cells or particulate organic matter = 18.74 J (Whyte, 1987) and 1 µmol NH4-N = 0.349 J (Elliott and Davidson, 1975). All the values of the energy components were expressed in Joules per hour per gram of dry weight (J h−1 gdw−1). 2.5. Statistical analysis All analyses were performed for each species and experiment separately due to intrinsic differences among species and seasons, as revealed by preliminary analyses. Generalized Linear Models (GLMs) were used when exploratory analysis showed a linear pattern in data, and Generalized Additive Models (GAMs) for non-linear patterns. Changes in valve closure activity of organisms (0 = inactive, 1 = active) and burrowing activity (0 = not burrowed, 1 = burrowed) were analysed through regression analyses fitting GLMs with a binomial distribution of the error term and a logit link function, i.e. logistic regression, with Salinity treatment as a fixed factor (S, 4 levels: 5–20, 10–25, 15–30 and 30–30). Two orthogonal a priori contrasts, i.e. S30-30 vs all others, and S5-20 vs S10-25 and S15-30) were made to evaluate differences among low salinity treatments and control for valve closure analysis. The relative importance of each variable (i.e. clearance rate, respiration, absorption efficiency and ammonium excretion) to the scope for growth was also calculated based on a bootstrapping analysis with 95% confidence intervals of a linear model including all variables. Changes in SFG were analysed through GAMs with Salinity treatment as 2.4. Data calculation Valve closure activity during the physiological measurements was identified for each individual and data subsequently analysed. Individuals with 0 activity, with valves closed during the clearance rate measurement were classified as non-active. The physiological rates of oxygen consumption, clearance rate and absorption efficiency were calculated according to Widdows and Staff (2006). Oxygen consumption was estimated by regressing oxygen concentration (mg L-1) in the respirometer over time using only the steepest part of the regression slope. To correct for activity of individual animals, clearance rate (CR; L h−1) was multiplied by the average activity based on the 6 observations of valve closure during filtration, i.e. a score of 0 was given to animals with valves closed at the beginning and end of each 5 min interval, 0.5 to animals with open valves at the beginning and closed at the end or the opposite behaviour, and 1 was given if valves were opened both at the beginning and the end. The SFG formula that integrates all these calculations and represents 4 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. smoothed fixed factor. Prior to analyses, normality and homogeneity of variances were checked by visual inspection of Q-Q and residual plots as well as by Shapiro-Wilk and Levene’s tests, respectively. For data collected on day 4 in the Dec15 experiment, analyses were carried out with untransformed data because no transformation removed nonnormality and heterogeneity of data. Analyses were made using the car package (Fox and Weisberg, 2011) and modEvA package (Barbosa et al., 2013) for GLM and mgcv (Wood 2011) for GAM in R version 3.6.1 (R Development Core Team, 2010). Package multcomp (Hothorn et al., 2008) was used to do a priori contrasts and package relaimpo (Grömping, 2006) was used to determine the relative importance of each variable to the SFG. Package ggplot2 (Wickham, 2016) was used to produce graphs with 95% confidence intervals fitting a GLM or GAM function. Average data are reported as means ± standard deviation. Table 1 Total number (N) and proportion of active individuals (PA) in the experiments. The significant effects of a priori contrasts of salinity treatments are also shown (Sg): * represents the contrast S30 vs all others, + represents the contrast S5-20 vs S10-25 and S15-30. D0, D4, D6, D8/9 represent the days of the measurement. Note that measurements on day 4 of the Dec15 experiment were done under different conditions, i.e. at the lowest salinity of each ramp. Species V. corrugata Day D0 D4 D6 3. Results D8/9 As expected with short term exposures to low salinity, no mortality was observed for any of the species after the acclimatization period or during the low salinity treatments. R. decussatus 3.1. Activity (valve closure) Most individuals remained active during all measurements with some remarkable exceptions, particularly when the measurements were made at the lowest salinity of each experimental ramp on day 4 of the Dec15 experiment, when significant differences were found for all species (Table 1). Individuals were significantly more likely to be inactive in the S5-20, S10-25 and S15-30 treatments than in the control. Both V. corrugata and C. edule however had active individuals at salinity 5 while no individuals of R. decussatus and R. philippinarum were active. Only V. corrugata had active individuals in all salinity treatments below 30. In March16 and May16 measurements were made at the highest salinity of each ramp on both day 4 and 6. In all measurements on day 4 with R. decussatus there was significantly less activity in the S5-20 treatment than in the S10-25 and S15-30 treatments (Table 1). Surprisingly in two cases there were fewer active individuals in the control compared to the others: for R. decussatus on day 8 in May16, and for R. philippinarum on day 4 in March16, marked by statistically significant differences. D0 D4 D6 D8/9 R. philippinarum D0 D4 D6 D8/9 3.2. Scope for growth C. edule 3.2.1. Responses under ebb-tide conditions Results of measurements under the lowest salinities of each ramp, i.e. S5, S10 and S15 after 4 days of stress in Dec15 indicated lower activity of bivalves at the lowest salinities relative to controls (Fig. 3-A, B, C, D, Table 2). The SFG models explained a high percentage of deviance (between 36% and 81%) and showed significant differences among treatments for all species (Table 2) with lower values in all reduced salinity treatments compared to the controls. This pattern was driven by a dramatic reduction in clearance rate which was not compensated by a reduction on respiration in the lower salinity treatments (Fig. 3-A,B,C,D). D0 D4 D6 D8/9 Salinity Dec15 N PA 30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 17 12 12 11 12 11 11 11 12 12 10 12 12 0.88 0.33 0.25 0.27 1 1 1 1 1 0.92 1 0.92 0.92 30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 16 12 12 13 13 11 12 12 12 12 12 12 12 0.69 0 0 0.46 0.92 1 1 1 1 1 1 1 1 30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 15 12 12 12 13 12 12 12 12 12 12 12 12 0.94 0 0.08 0 0.92 0.92 1 0.75 1 0.75 1 1 0.83 30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 5–20 10–25 15–30 30–30 16 12 11 12 13 12 12 12 12 12 12 12 12 0.88 0.5 0 0.33 0.92 1 1 1 1 1 1 1 1 Mar16 Sg * + * * + * N PA 16 12 12 12 12 12 12 12 12 12 12 12 12 1 1 1 1 1 1 1 1 1 0.92 1 1 0.92 16 10 12 12 12 12 12 12 11 12 12 12 12 1 0.83 1 1 1 1 1 1 0.92 1 1 1 1 16 12 11 12 9 12 12 11 11 12 12 12 11 1 1 0.92 1 0.75 1 1 0.92 0.92 1 1 1 0.92 15 12 12 12 12 12 12 11 12 11 11 12 12 0.94 1 1 0.92 0.92 1 1 1 1 0.92 0.92 1 1 May16 Sg + * N PA 15 11 12 12 12 11 12 12 12 12 12 12 12 0.94 0.92 1 1 1 0.92 1 1 1 1 1 1 1 16 10 12 12 12 12 12 12 12 12 12 12 7 1 0.83 1 1 1 1 1 1 1 1 1 1 0.78 14 12 12 11 12 10 12 12 12 12 12 12 12 0.88 1 1 0.92 1 1 1 1 1 1 1 1 1 14 12 12 12 12 12 12 12 12 12 11 12 12 0.88 1 1 1 1 1 1 1 1 1 0.92 1 1 Sg + * March16, models fitted the data better, with ~20% of deviance explained. The SFG varied significantly among treatments all days, although only marginally on day 8 (recovery). However, while on day 4, the SFG increased almost linearly with salinity, on days 6 and 8 the SFG followed a similar pattern with the lowest values at S5-20 and the highest at S10-25 (Fig. 4B). The SFG curve on day 8 was above those of days 4 and 6 in the reduced salinity treatments, which indicated a recovery of individuals after the stress. Differences in SFG were mainly driven by clearance rate that showed strong differential patterns among treatments and days compared to respiration. In May16, the models 3.2.2. Responses under flood-tide conditions The examined species showed different patterns of variation in SFG after exposure to the salinity treatments and responses differed across experiments, i.e. Dec15, March16 and May16, in most cases. Venerupis corrugata had the lowest mean basal SFG of all species in Dec15 (2.8 ± 12.2 Jh−1g−1), it was higher in March16 (17.7 ± 20.3 Jh−1g−1) and then negative in May16 (-10.1 ± 13.1 Jh−1g−1) (Fig. 4A-C, Table 2). In Dec15, no significant differences were found on days 6 and 9 (Table 2), and no patterns were identified (Fig. 4A) with a very low percentage of deviance explained by the models (< 2%). In 5 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 3. A,B,C,D. Scope for growth, clearance rate and respiration rate (Jh−1g−1) of A: Venerupis corrugata, B: Ruditapes decussatus, C: Ruditapes philippinarum and D: Cerastoderma edule measured on day 4 of the Dec15 experiment under low tide conditions. Line is the predicted line of the salinity model with 95% confidence band. The dots are the raw data for each salinity treatment. Note the different scales used for R. philippinarum and C. edule in Clearance rate and SFG. 25 and S15-30 below those of days 4 and 6 indicated no recovery of individuals at those treatments. Patterns were again clearly driven by clearance rates that differed strongly from respiration. Ruditapes decussatus had a mean basal SFG of 8.2 ± 17.5 Jh−1g−1 in Dec15, 3.7 ± 6.4 Jh−1g−1 in March 16, and 7.1 ± 17.1 Jh−1g−1 in May 16 (Fig. 5A-C, Table 2). In Dec15, the model explained 22% of explained the greatest percentage of deviance (> 39% on all days) and showed again significant differences among treatments all days. In this case, a similar parabolic pattern was seen for day 4 and 6 with lowest values at S5-20, increasing at S10-25 and S15-30 and decreasing for the controls (Fig. 4C). On day 8, the SFG increased almost linearly from the lowest values at S5-20 to the highest at control. The SFG curve at S106 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Table 2 Summary of results of the GAM analyses to test the effect of salinity treatments on the Scope for Growth (SFG). Values in bold are statistically significant (p < 0.05) and for those graphical patterns are presented by: + when the expected pattern of lower SFG for lower salinities was apparent in the graphical analysis, − when the inverse pattern was apparent and 0 when neither of these patterns was apparent. Ref. df: reference degrees of freedom; DE: % of deviance explained by the model. a: Note that measurements on day 4 of the Dec15 experiment were done under different conditions, i.e. at the lowest salinity of each ramp. Species V. corrugata Exp Dec 15 Mar 16 May 16 R. decussatus Dec 15 Mar 16 May 16 R. philippinarum Dec 15 Mar 16 May 16 C. edule Dec 15 Mar 16 May 16 Day Ref.df F, pvalue % DE Patterns in SFG 5–20 lower 30–30 higher 0 + + + + + + 0 + – – – – + 4a 6 9 4 6 8 4 6 8 2.709 1 1.318 1 2.947 2.37 2.707 2.798 2.237 7.955, 0.028, 0.619, 8.278, 4.608, 2.426, 11.01, 13.72, 12.32, 4a 6 9 4 6 8 4 6 8 2.561 2.99 1 2.951 1.042 1.209 2.705 2.408 1 9.998, p = 6.46e-05*** 3.914, p = 0.0163* 0.272, p = 0.604 4.495, p = 0.00493** 0, p = 0.98 2.096, p = 0.173 4.031, p = 0.0101* 2.267, p = 0.143 10.85, p = 0.00194** 38.2 22.1 0.589 27.5 0.092 5.1 24.4 12.1 20.2 0 0 + 0 + – + – + + 4a 6 9 4 6 8 4 6 8 2.996 1 1.511 2.99 2.243 2.458 2.141 2.939 2.398 62.48, p < 2e-16*** 0.266, p = 0.609 0.886, p = 0.498 5.606, p = 0.00191** 1.54, p = 0.208 3.593, p = 0.0393* 1.168, p = 0.362 7.561, p = 0.00117*** 6.507, p = 0.00263** 81.3 0.574 3.08 29.7 9.48 17.7 6.57 33.6 26.8 0 + 0 0 + – + 0 – + 4a 6 9 4 6 8 4 6 8 2.982 1 1 2.16 2.383 2.96 1 2.998 2.352 26.73, p = 9.88e-12*** 0.092, p = 0.763 4.933, p = 0.0312* 2.94, p = 0.0621. 0.934, p = 0.536 2.379, p = 0.0615. 1.13, p = 0.293 9.519, p = 5.41e-05*** 2.686, p = 0.0592. 65.5 0.199 9.68 13.8 5.72 17.2 2.4 40 15.2 0 + + + 0 – 0 + 0 + p p p p p p p p p = = = = = = = = = 0.000269*** 0.867 0.619 0.00601** 0.0159* 0.0939. 5.45e-05 *** 5.37e-06 *** 2.69 e-05 *** 36.2 0.0612 1.7 15.3 23 13.6 40.6 46.8 39.1 salinity treatments was above those of the other days indicating recovery of individuals. In May 16, the models for days 6 and 8 explained 34% and 27% of deviance, respectively, and indicated highly significant differences among treatments, but with contrasting patterns driven by clearance rates (Fig. 6C). While on day 6, there was a parabolic curve with the highest values at S10-25, similar to those of the other venerids, on day 8 there was an increase of SFG at S15-30 and control. The SFG curve of that day was below those of the other days except for control, indicating no recovery of individuals. Cerastoderma edule had the highest basal values compared to the other species with a mean value of 89.4 ± 70.3 Jh−1g−1 in Dec15, then lower in March16 and May16 48.2 ± 36.5 Jh−1g−1 and 5.8 ± 33.69 Jh−1g−1, respectively (Fig. 7A-C, Table 2). In Dec15 there were significant differences in SFG among treatments on day 9, with models explaining around 10% of deviance and the SFG increasing almost linearly with salinity treatments (Fig. 7A). The pattern was driven by clearance rate. In March16, the SFG varied among treatments with marginally lower values at S5-20. On day 8, the model explained 17% of deviance and indicated significant differences with the highest values at S15-30 compared to the rest of treatments (Fig. 7B). Such patterns were driven by both clearance rate and respiration. In May16, the SFG varied significantly among treatments on days 6 and 8, with models explaining 40% and 15% of deviance, respectively. Such patterns were mainly driven by clearance rate (Fig. 7C). On day 6, the SFG showed higher values at S5-20 and a peak at S10-25 perhaps reflecting deviance on day 6, and showed significant differences among treatments due to higher values at S10-25. On day 9, the model was not significant (Fig. 5A). In March16 only on day 4, the model showed significant variation in SFG among treatments and explained ~ 27% of deviance, with the highest values at S15-30 (Fig. 5B). In May16, the response was similar to that of V. corrugata although less pronounced (Fig. 5C). The SFG varied significantly among salinity treatments on day 4 and 8, both models explaining higher deviance (~24%) compared to the model for day 6 (12%). On days 4 and 6, the SFG followed a parabolic pattern with highest values at S10-25 and S15-30, whereas on day 8 the SFG increased almost linearly with salinity, but the curve was below those of days 4 and 6 indicating no recovery of individuals. The pattern of SFG resembled that of clearance rate. Ruditapes philippinarum had a mean basal SFG of 24.1 ± 26.5 Jh−1g−1 in Dec15, 10.4 ± 14.4 Jh−1g−1 in March16 and then in May16 mean basal SFG was negative −8.6 ± 6.4 Jh−1g−1 (Fig. 6A-C, Table 2). On days 6 and 9 in Dec 15, models explained a very low percentage of deviance (< 4%) and did not show significant differences among treatments (Fig. 6A). In March16, the models for days 4 and 8 explained 30% and 18% of deviance, respectively, and showed significant differences among treatments, but with very different patterns mainly driven by clearance rate. While on day 4 the SFG decreased at S10-25 and then increased at S15-30 following a sinusoidal pattern, on day 8 there was a parabolic response with values increasing at the intermediate salinities (Fig. 6B). The SFG curve on day 8 in the reduced 7 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 4. Scope for growth, clearance rate and respiration rate (Jh-1g−1) of Venerupis corrugata in A: Dec15, B: March16 and C: May16 experiments, measured under flood tide conditions. Lines are the predicted lines of the salinity model with 95% confidence bands. The symbols are the raw data for each salinity treatment. Yellow solid circles: day 4, red rhombus: day 6, blue crosses: day 8/9. compensation activity apparent both in clearance rates and respiration. On day 8, the SFG was similar at all treatments except for controls and values were below those on day 6, indicating no recovery of individuals. decussatus. 3.4. Burrowing activity 3.4.1. Responses under ebb-tide conditions The four species were unable to burrow after four days of stress at the lowest salinities of each treatment, i.e. 5, 10 and 15 compared to the control (χ2 values all p < 0.0001; Fig. 8-A,B,C,D, Table 3). The models explained between 40% and 70% of deviance. Only a few individuals of C. edule burrowed at salinities of 10 and 15; none burrowed at 5. 3.3. The importance of variables in SFG The relative importance analysis confirmed clearance rate as the main relevant variable, with adjusted R2 values from 0.6 to 0.95 across all experiments and species. Respiration was the second variable in importance ranging from 0.25 to < 0.1 followed by absorption efficiency and ammonia, which were never above 0.1 (see Supplementary material, S.1). The highest values of relative importance for clearance rate were for C. edule and R. philippinarum in all seasons (> 0.8). The maximum relative importance of respiration for any of the four species was 0.25, found in V. corrugata in Dec15 and in March16 for R. 3.4.2. Responses under flood-tide conditions Individuals of V. corrugata showed significant differences in burrowing probability among treatments on day 6 in Dec15 and March16 (Fig. 9A, Table 3) due to a marked increase in the probability of burrowing with increasing salinity in the reduced salinity treatments and 8 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 5. Scope for growth, clearance rate and respiration rate (Jh-1g−1) of Ruditapes decussatus in A: Dec15, B: March16 and C: May16 experiments, measured under flood tide conditions. Lines are the predicted lines of the salinity model with 95% confidence bands. The symbols are the raw data for each salinity treatment. Yellow solid circles: day 4, red rhombus: day 6, blue crosses: day 8/9. salinities. In March16, on day 6 highly significant differences were found with the model explaining 32% of deviance. The lowest probability of individuals to burrow was at S5-20 and S10-25 and then it increased markedly with higher salinity treatments. In May16 marginal differences were detected only on day 8 and the model explaining 11% of deviance although the pattern was the opposite to the previous days, with highest probability of burrowing at S5-20, and then decreased markedly with salinity treatments, reaching the lowest probability at control. Cerastoderma edule in Dec15 had a similar pattern all days, with significant differences on day 9 due to lower probability of individuals to burrow in all reduced salinity treatments compared to the control and 16% of deviance explained (Fig. 9D, Table 3). In March and May16 only marginally significant differences among salinity treatments were found. the model on day 4 in March16 explained 24% of the deviance showed higher probability to burrow with increasing salinity. In May16 the model on day 6 with very little deviance explained (10%) and lower models explaining 19% and 39% of deviance, respectively. After recovery on day 8/9, no significant differences were found among treatments. In May16, marginally significant differences were found on day 4 (Fig. 9A, Table 3) with the model explaining 13% of the deviance and showing the same pattern of higher probability of burrowing for individuals at salinities above S5-20. On day 6 in Dec15 no differences were found for R. decussatus (Fig. 9B, Table 3), as on day 9, when a clear recovery was observed. Both models explained little of the deviance (< 10%). In March16 the models explained very little of the deviance (< 5%) and did not show significant differences among treatments. In May16 significant differences were found only on day 4, explaining 14% of deviance. The lowest probability of individuals to burrow was at S5-20 and it increased markedly from S10-25 to higher salinities. Ruditapes philippinarum showed differences on day 6 in Dec15 and the model explained 21% of deviance (Fig. 9C, Table 3). The lowest probability was at S5-20 and then it increased smoothly with increasing 9 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 6. Scope for growth, clearance rate and respiration rate (Jh-1g−1) of Ruditapes philippinarum in A: Dec15, B: March16 and C: May16 experiments, measured under flood tide conditions. Lines are the predicted lines of the salinity model with 95% confidence bands. The symbols are the raw data for each salinity treatment. Yellow solid circles: day 4, red rhombus: day 6, blue crosses: day 8/9. treatments, but on day 6 neither exhibited higher SFG in the controls (Table 2). In terms of burrowing ability, the venerids V. corrugata and R. philippinarum, were the most affected species at lower salinity treatments although they recovered by day 8. In spring (May16), there were two types of compensation response of SFG that led to an inability of individuals to recover: the venerids V. corrugata, R. philippinarum and to lower extent R. decussatus showed parabolic curves under stress on Day 6 with higher values for treatments S10-25 and S15-30. The cockle showed a pronounced sinusoidal response with higher values at treatment S10-25 on day 6. In terms of burrowing ability, V. corrugata and R. decussatus were the most impacted species at the lowest salinity treatments although they recovered after the stress (Table 3). Valve closure was the strategy used by all species under the higher stress (salinity 15 and below), as expected from previous findings (Carregosa et al., 2014; Gharbi et al., 2016), though with remarkable differences (Table 1). One third of V. corrugata individuals opened in all treatments, which could explain its higher abundance in the low burrowing probabilities at lower salinities were found. 4. Discussion The overall response to stress was the same for all species at the lower salinities (5, 10, 15), with almost no activity (valves closed and no burrowing) and SFG near 0, supporting the general hypothesis of salinity drops tapering physiological activity (Fig. 3A-D). Under the higher salinities (20, 25, 30), species reacted diversely. In autumn (Dec15) all species resisted generally well in terms of SFG, except for C. edule which showed no recovery with lower values for the reduced salinity treatments compared to controls (Table 2: Dec15: Day 9). All species recovered burrowing ability although the responses differed. Whereas the effects of low salinity treatments were still evident in V. corrugata and C. edule, in the other two species they were not (Fig. 9, Day 9). In winter (March16), both V. corrugata and R. decussatus exhibited reduced SFG on day 4 as expected in the reduced salinity 10 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 7. Scope for growth, clearance rate and respiration rate (Jh-1g−1) of Cerastoderma edule in A: Dec15, B: March16 and C: May16 experiments, measured under flood tide conditions. Lines are the predicted lines of the salinity model with 95% confidence bands. The symbols are the raw data for each salinity treatment. Yellow solid circles: day 4, red rhombus: day 6, blue crosses: day 8/9. Widdows (1997) imposed temperature stress on R. decussatus and obtained similar values for respiration or ammonia but the clearance rates were higher. Similar values to ours have been reported for SFG following stress due to ambient ammonia after low tide in estuaries (Sobral and Fernandes, 2004). Albentosa et al. (2007) reported lower respiration rates given food limitation for V. corrugata and R. decussatus. The values of SFG found in this study were comparable to those found in the literature considering that the experiments reported here were designed to be sublethal with intermittent exposure to stress, while the majority of those in the literature were designed for a continuous exposure to low salinity and often animals suffer significant mortality (Kim et al., 2001). Mean negative values of SFG were found in the controls of V. corrugata and R. philippinarum at the end of the experiment in Dec15 and in March16. Moreover, these species had negative values of SFG in the basal measurements before the imposition of stress in May16. The highly energy-demanding process of reproduction (Griffiths and intertidal and subtidal, where it is less exposed to long periods of low salinity and presumed higher mortality associated with valve opening at low salinities (Akberali and Trueman, 1985; McFarland et al., 2013; Carregosa et al., 2014). Valve closure under lower salinities was more extended in R. decussatus and R. philippinarum (Gharbi et al., 2016; Kim et al., 2001), as in other intertidal species (Shumway, 1977). Similarly to present results, adults of C. edule have low activity below 10 (Verdelhos et al., 2015), or recruits at 15 (Peteiro et al., 2018). However, a variable percentage remained open, as expected in a species with intermitent gaping under periodic salinity fluctuations (Nossier, 1986; Malham et al., 2012). Once salinity rose above 20, all four species opened valves to resume respiration, feeding and excretion of waste products, in agreement with previous findings (Kim et al., 2001; Verdelhos et al., 2015). The values of the components of SFG reported here are similar to other studies of salinity stress when clams were exposed continuously to salinities of 20 or below (Kim et al., 2001; Nie et al., 2016). Sobral and 11 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 8. A,B,C,D. Burrowing probabilities for A: Venerupis corrugata, B: Ruditapes decussatus, C: Ruditapes philippinarum and D: Cerastoderma edule on day 4 of the Dec15 experiment at the lowest salinity of each ramp. Lines are the predicted lines of the salinity model with 95% confidence bands. The dots are the raw data for each salinity treatment. Montaudouin and Bachelet, 1996; Solidoro et al., 2000; Garcia et al., 2001; Bidegain et al., 2013). Failure of burrowing may be an indicator of high physiological stress and compromise individual survival (Haider et al., 2018). With salinities ≤15 on day 4 of Dec15 no individuals buried at salinity 5 and only a few C. edule at 10 and 15 (Fig. 8, Table 3). Under less extreme stresses such as salinities ≥20, behaviour can show different patterns due to energy trade-offs with basal maintenance and reproduction (Ansell and Trueman, 1973). Generally, in the three experiments during the stress the ability to burrow decreased under lower salinity treatments, increasing during the recovery period as expected (Verdelhos et al., 2015). Our results indicated that clearance rate was a good proxy to assess salinity stress of adult bivalves in the laboratory when a complex experimental design to measure many variables is not feasible. Nevertheless, respiration rate gives information that is useful to understand underlying processes such as physiological compensation (Kim et al., 2001) and it is an easy variable to measure directly without much manipulation. Finally, ammonia excretion is already avoided in many studies since its weight in the calculation of SFG is small (Widdows and Staff, 2006). For two of the species reported here, R. philippinarum and C. edule, clearance rate had a relative importance of ≥0.8 in all experiments and ≥0.9 in both March16 and May16. In contrast, for V. corrugata and R. decussatus, clearance rate was ≥0.6 in all experiments, but respiration was ≥0.15 in all but one case. For the first two species therefore, one might be able to measure just clearance rate to estimate SFG, but not in the other two (see Supplementary material, Table S.1). According to these results, ≤15 was the salinity threshold for high frequencies of valve closure and sharp decrease of activity for most individuals. The data on SFG and valve closure activity resulting from these experiments together with field data from a salinity monitoring system in the field (own data not published) allowed us to make some Griffiths, 1987) is often associated with an increase in respiration rates in bivalves (Widdows, 1978; Rueda and Smaal, 2004), as seen on V. corrugata, R. philippinarum and C. edule, which are reproductive in May. During late autumn and winter V. corrugata and R. philippinarum are mostly in the gametogenic stage, progressing to a ripening period before spawning in spring (Joaquim et al., 2010; Rodríguez-Moscoso et al., 1992) while gametogenesis in C. edule occurs somewhat earlier (Martínez-Castro and Vázquez, 2012). Conservative species such as those in this study (Pérez-Camacho et al., 2003; Karray et al., 2015) mainly rely on energy obtained from exogenous feeding for reproduction. In the current study, clams received 1% of the mean dry weight clams−1 day−1 of algae, a diet chosen to allow detection of stress. The clams were not fed continuously, instead having their entire daily ration delivered over an approximate time span of 4–6 h. The decrease of SFG of animals in controls in Dec15 and March16 may indicate that clams received insufficient food to afford energetic costs associated with gametogenesis and they were relying on internal energy reserves. Negative basal values in May16 may be related to the exhaustion of species after a spawning event. SFG values can be classified into three categories (Widdows et al., 2002): low growth potential/high stress (< 5 Jh−1g−1), moderate growth potential/ moderate stress (5–15 Jh−1g−1), and high growth potential/low stress (> 15 Jh−1g−1). Low growth potential was found for all species at salinities ≤15, when many animals stayed closed or siphons were mostly retracted temporarily reducing pumping activity (Table 1: Dec15, D4; Woodin et al., 2020). For salinities > 15, R. decussatus had the lowest growth potential, regardless of season, while V. corrugata ranged between low in autumn and moderate in winter and spring, R. philippinarum between low in autumn and moderate to high in winter and spring, and C. edule often had the highest growth potential, particularly in spring (Figs. 3–6). Those results are in agreement with field growth data of these species (Stenton-Dozey et al., 1994; 12 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Table 3 Summary of results of the GLM analyses to test the effect of salinity treatments on the burrowing activity. Values in bold are statistically significant (p < 0.05) and for those graphical patterns are presented by: + when the expected pattern of lower SFG for lower salinities was apparent in the graphical analysis, − when the iverse pattern was apparent and 0 when neither of these patterns was apparent. a: Note that measurements on day 4 of the Dec15 experiment were done under different conditions, i.e. at the lowest salinity of each ramp. Species V. corrugata Exp Dec 15 Mar 16 May 16 R. decussatus Dec 15 Mar 16 May 16 R. philippinarum Dec 15 Mar 16 May 16 C. edule Dec 15 Mar 16 May 16 Day Df Chisq, pvalue % DE Patterns in Burrowing 5–20 lower 30–30 higher 4a 6 9 4 6 8 4 6 8 3 3 3 3 3 3 3 3 3 27.977, p = 3.672e−06*** 10.162, p = 0.01724* 3.3425, p = 0.3418 0.94212, p = 0.8153 23.968, p = 2.537e−05*** 1.512, p = 0.6795 7.807, p = 0.05017. 5.0851, p = 0.1657 3.721, p = 0.2932 64.7 18.8 5.4 1.4 38.9 2.5 12.9 7.6 6.4 0 + + + + + + + 4a 6 9 4 6 8 4 6 8 3 3 3 3 3 3 3 3 3 23.891, p = 2.633e−05*** 5.6528, p = 0.1298 0.50802, p = 0.9171 2.2805, p = 0.5163 2.8539, p = 0.4147 3.0698, p = 0.381 8.8016, p = 0.03205* 2.8683, p = 0.4124 2.4924, p = 0.4767 59.4 9.09 1.03 3.4 4.4 4.7 14.3 5.1 4.1 0 + + + 4a 6 9 4 6 8 4 6 8 3 3 3 3 3 3 3 3 3 32.831, p = 3.495e−07*** 13.612, p = 0.003483** 1.0455, p = 0.7902 3.0545, p = 0.3833 20.199, p = 0.0001544*** 0.94212, p = 0.8153 4.612, p = 0.2025 0.98551, p = 0.8048 7.2211, p = 0.06517 70.7 20.9 7.9 4.6 32.3 1.4 7.2 1.5 10.9 0 + + + + + – – 4a 6 9 4 6 8 4 6 8 3 3 3 3 3 3 3 3 3 22.196, p = 5.938e−05*** 3.5895, p = 0.3093 10.322, p = 0.01602* 7.7682, p = 0.05105. 0.88275, p = 0.8296 4.1795, p = 0.2427 2.3875, p = 0.496 6.7767, p = 0.07937. 4.1442, p = 0.2463 39.6 5.7 15.6 24.2 2.1 8.1 3.6 10.5 6.2 + + + 0 + 0 + + projections of seasonal fitness for each species to define management strategies. During high tide, seawater salinity in one of the most productive shellfish beds (Carril, Ria de Arousa) during autumn 2015, was ≤15 during 9.5% of the time, while in winter and spring salinities ≤15 occurred 21% and 23% of the time, respectively. The SFG values multiplied by the proportion of active animals at reduced salinities (Table 1) and the amount of time when salinity was ≤15 yield SFG in the field (f-SFG) for each species and season (see Supplementary material, Table S.2). These data allow us to address the question about the impact of low salinities in the field on energetics and behavioral responses. A striking result is the gross reduction in SFG by applying the field salinity adjustment. The loss of > 20% of the original SFG value due to short-term exposures to salinities ≤15 is impressive. For time periods when SFG without stress is already low such as March16 for R. decussatus and May16 for V. corrugata, reducing SFG due to low salinities in the field with the behavioral response of valve closure increases the probability of organisms in energy deficit. The incidence of extreme events such as heavy rainfall is expected to increase according to future climate change scenarios. As these results indicate, consequences for growth of tidally fluctuating salinities can be significant even though mass mortality does not occur, and managers need to consider the consequences in their management plans. In this sense we present a conceptual model (Fig. 10) that can serve as a planning tool for managers and it can be used to develop understanding among researchers, managers and policy makers. CRediT authorship contribution statement Rula Domínguez: Conceptualization, Methodology, Investigation, Formal analysis, Writing - original draft, Visualization. Elsa Vázquez: Conceptualization, Methodology, Investigation, Formal analysis, Resources, Writing - review & editing, Supervision, Project administration, Funding acquisition. Sarah A. Woodin: Conceptualization, Methodology, Investigation, Formal analysis, Resources, Writing - review & editing, Funding acquisition. David S. Wethey: Conceptualization, Methodology, Investigation, Formal analysis, Resources, Funding acquisition. Laura G. Peteiro: Methodology, Investigation, Writing - review & editing. Gonzalo Macho: Methodology, Investigation, Writing - review & editing. Celia Olabarria: Conceptualization, Methodology, Investigation, Formal analysis, Resources, Writing - review & editing, Supervision, Project administration, Funding acquisition. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. 13 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 9. A,B,C,D. Burrowing probabilities for A: Venerupis corrugata, B: Ruditapes decussatus, C: Ruditapes philippinarum and D: Cerastoderma edule of the three experiments at the highest salinity of each ramp. From left to right column: Dec15, March16 and May16. Yellow solid circles: day 4, red rhombus: day 6, blue crosses: days 8/9. 14 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Fig. 10. Conceptual model applied to the fishery management. The main effects and consequences of salinity changes on SFG and burrowing activity on the four species over time. and Noia respectively, for providing the clams and cockles and valuable comments for the experiment. Acknowledgments This research was supported by grants CTM2014-51935-R from the Spanish Ministerio de Economıa y Competitividad to the project MARISCO, a pre-doctoral grant for RD (CTM2014-51935-R), and the Autonomous government Xunta de Galicia-FEDER (project GRC2013004) and grants NNX11AP77G from the US National Aeronautics and Space Administration (NASA) and OCE1129401 from the US National Science Foundation to DSW and SAW. Facilities were kindly provided by the Estacion de Ciencias Mariñas de Toralla (ECIMAT) of the University of Vigo. We want to thank Esther Pérez and all the staff at ECIMAT for their technical support. We also thank J.C. Mariño and L. Solís, technical Assistants of the Cofradía de Pescadores of Cambados Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ecolind.2019.106031. References Akberali, H.B., Trueman, E.R., 1985. Effects of environmental stress on marine bivalve molluscs. Adv. Mar. Biol. 22, 101–198. Albentosa, M., Fernández-Reiriz, M.J., Labarta, U., Pérez-Camacho, A., 2007. Response of two species of clams, Ruditapes decussatus and Venerupis pullastra, to starvation: 15 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. management. Ocean Coast. Manage. 69, 316–326. https://doi.org/10.1016/j. ocecoaman.2012.08.007. Karray, S., Smaoui-damak, W., Rebai, T., Hamza-chaffai, A., 2015. The reproductive cycle, condition index, and glycogen reserves of the cockles Cerastoderma glaucum from the Gulf of Gabès (Tunisia). Environ. Sci. Pollut. Res. 22, 17317–17329. https:// doi.org/10.1007/s11356-015-4337-6. Kim, W.S., Huh, H.T., Huh, S.H., Lee, T.W., 2001. Effects of salinity on endogenous rhythm of the Manila clam, Ruditapes philippinarum (Bivalvia: Veneridae). Mar. Biol. 138, 157–162. https://doi.org/10.1007/s002270000430. Lockwood, A.P.M., 1976. Physiological adaptation to life in estuaries. In: Newell, R.C. (Ed.), Adaptation to Environment. Butterworths, London, pp. 315–392. Macho, G., Woodin, S.A., Wethey, D.S., Vázquez, E., 2016. Impacts of sublethal and lethal high temperatures on clams exploited in European fisheries. J. Shellfish Res. 35, 405–419. https://doi.org/10.2983/035.035.0215. Malham, S.K., Hutchinson, T.H., Longshaw, M., 2012. A review of the biology of European cockles (Cerastoderma spp.). J. Mar. Biol. Assoc. UK 92, 1563–1577. https://doi.org/10.1017/S0025315412000355. Martínez-Castro, C., Vázquez, E., 2012. Reproductive cycle of the Cockle Cerastoderma edule (Linnaeus 1758) in the Ría De Vigo (Galicia, Northwest Spain). J. Shellfish Res. 31, 757–767. https://doi.org/10.2983/035.031.0320. McFarland, K., Donaghy, L., Volety, A.K., 2013. Effect of acute salinity changes on hemolymph osmolality and clearance rate of the non-native mussel, Perna viridis, and the native oyster, Crassostrea virginica, in Southwest Florida. Aquat. Invasions 8, 299–310. https://doi.org/10.3391/ai.2013.8.3.06. McLachlan, A., Defeo, O., Chapter 14 – Fisheries, The Ecology of Sandy Shores (Third Edition), Academic Press, 2018, Pages 331-374, ISBN 9780128094679, doi: 10.1016/ B978-0-12-809467-9.00014-X. Méndez, G., Vilas, F., 2005. Geological antecedents of the Rias Baixas (Galicia, northwest Iberian Peninsula). J. Mar. Syst. 54, 195–207. https://doi.org/10.1016/j.jmarsys. 2004.07.012. Molares, J., Parada, J.M., Navarro-Perez, E., Fernandez, A., 2008. Variabilidad interanual de las ventas de los principales recursos marisqueros de Galicia y su relación con las condiciones ambientales. Rev. Galega Recursos Mariños 2, 1–42. Morgan, E., O’Riordan, R.M., Culloty, S.C., 2013. Climate change impacts on potential recruitment in an ecosystem engineer. Ecol. Evol. 3, 581–594. https://doi.org/10. 1002/ece3.419. Montaudouin, X., Bachelet, G., 1996. Experimental evidence of complex interactions between biotic and abiotic factors in the dynamics of an intertidal population of the bivalve Cerastoderma edule. Oceanol. Acta 19, 449–463. Nie, H., Chen, P., Huo, Z., Chen, Y., Hou, X., Yang, F., Yan, X., 2016. Effects of temperature and salinity on oxygen consumption and ammonia excretion in different colour strains of the Manila clam, Ruditapes philippinarum. Aquac Res. 48, 2778–2786. https://doi.org/10.1111/are.13111. Nogueira, E., Pérez, F.F., Ríos, A.F., 1997. Seasonal patterns and long-term trends in an estuarine upwelling ecosystem (Ría de Vigo, NW Spain) 285–300. Nossier, M.A., 1986. Ecophysiological responses of Cerastoderma edule (L.) and Cerastoderma glaucum (Bruguiere) to different salinity regimes and exposure to air. J. Molluscan Stud. 52 (2), 110–119. Parada, J.M., Molares, J., 2008. Natural mortality of the cockle Cerastoderma edule (L.) from the Ria of Arousa (NW Spain) intertidal zone. Rev. Biol. Mar. Oceanogr. 43, 501–511. https://doi.org/10.4067/S0718-19572008000300009. Parada, J.M., Molares, J., Otero, X., 2012. Multispecies mortality patterns of commercial bivalves in relation to estuarine salinity fluctuation. Estuar. Coasts 35, 132–142. https://doi.org/10.1007/s12237-011-9426-2. Peteiro, L.G., Woodin, S.A., Wethey, D.S., Costas-Costas, D., Martínez-Casal, A., Olabarria, C., Vázquez, E., 2018. Responses to salinity stress in bivalves: Evidence of ontogenetic changes in energetic physiology on Cerastoderma edule. Sci. Rep. 8, 1–9. https://doi. org/10.1038/s41598-018-26706-9. R Core Team, 2014. R: a Language and Environment for Statistical Computing. The R Foundation for Statistical Computing, Vienna, Austria, ISBN 3–900051-07-0. Available online at: http://www.Rproject.org/. von Richthofen, F., 1886. Fqhrer fqr Forschungsreisende. Oppen-heim, Berlin. Resgalla, C., Elisângela, J., Brasil, D.S., Salomão, L.C., 2007. The effect of temperature and salinity on the physiological rates of the mussel Perna perna (Linnaeus 1758). Brazilian Arch. Biol. Technol. 50, 543–556. https://doi.org/10.1590/S151689132007000300019. Rodríguez-moscoso, E., Pazó, J., García, A., Fernández, Cortés F., 1992. Reproductive cycle of Manila clam, Ruditapes philippinarum (Adams & Reeve 1850) in Ria of Vigo (NW Spain). Sci. Mar. 56 (1), 61–67. Rueda, J.L., Smaal, A.C., 2004. Variation of the physiological energetics of the bivalve Spisula subtruncata (da Costa, 1778) within an annual cycle. J. Exp. Mar. Bio. Ecol. 301 (2), 141–157. https://doi.org/10.1016/j.jembe.2003.09.018. Sará, G., Romano, C., Widdows, J., Staff, F.J., 2008. Effect of salinity and temperature on feeding physiology and scope for growth of an invasive species (Brachidontes pharaonis – MOLLUSCA: BIVALVIA) within the Mediterranean Sea. J. Exp. Mar. Bio. Ecol. 363, 130–136. https://doi.org/10.1016/j.jembe.2008.06.030. Shumway, S.E., 1977. Effect of salinity fluctuation on the osmotic pressure and Na+, Ca2+ and Mg2+ ion concentrations in the hemolymph of bivalve molluscs. Mar. Biol. 41, 153–177. https://doi.org/10.1007/BF00394023. Sobral, P., Fernandes, S., 2004. Physiological responses and scope for growth of Ruditapes decussatus from Ria Formosa, southern Portugal, exposed to increased ambient ammonia. Sci. Mar. 68 (2), 219–225. https://doi.org/10.3989/scimar.2004.68n2219. Sobral, P., Widdows, J., 1997. Effects of elevated temperatures on the scope for growth and resistance to air exposure of the clam Ruditapes decussatus (L.), from southern Portugal. Sci Mar. 61, 163–171. Solidoro, C., Pastres, R., Canu, D.M., Pellizzato, M., Rossi, R., 2000. Modelling the growth Physiological and biochemical parameters. Comp. Biochem. Physiol. – B Biochem. Mol. Biol. 146, 241–249. https://doi.org/10.1016/j.cbpb.2006.10.109. Alvarez, I., deCastro, M., Gomez-Gesteira, M., Prego, R., 2005. Inter- and intra-annual analysis of the salinity and temperature evolution in the Galician Rías Baixas-ocean boundary (northwest Spain). J. Geophys. Res. C: Oceans 110 (4), 1–14. https://doi. org/10.1029/2004JC002504. Aníbal, J., Esteves, E., Rocha, C., 2011. Seasonal variations in gross biochemical composition, percent edibility, and condition index of the clam Ruditapes decussatus cultivated in the Ria Formosa (South Portugal). J. Shellfish Res. 30, 17–23. https:// doi.org/10.2983/035.030.0104. Ansell, A.D., Trueman, E.R., 1973. The energy cost of migration of the bivalve Donax on tropical sandy beaches. Mar. Behav. Physiol. 2, 21–32. Aranguren, R., Gomez-León, J., Balseiro, P., Costa, M.M., Novoa, B., Figueras, A., 2014. Abnormal mortalities of the carpet shell clam Ruditapes decussatus (Linnaeus 1756) in natural bed populations: a practical approach. Aquac. Res. 45, 1303–1310. https:// doi.org/10.1111/are.12074. Barbosa, A.M., Real, R., Muñoz, A.R., Brown, J.A., 2013. New measures for assessing model equilibrium and prediction mismatch in species distribution models. Divers. Distrib. 19, 1333–1338. https://doi.org/10.1111/ddi.12100. Bayne, B.L., 1998. The physiology of suspension feeding by bivalve molluscs: an introduction to the Plymouth “Trophee” workshop. J. Exp. Mar. Biol. Ecol. 219, 1–19. https://doi.org/10.1016/S0022-0981(97) 00172-X. Beninger, P., Lucas, A., 1984. Seasonal variations in condition, reproduct. activity, and gross biochemical comp. of Tapes decussatus and Tapes philipinarum. J. Exp. Mar. Biol. Ecol. 44, 51–56. Berger, V.J., Kharazova, A.D., 1997. Mechanisms of salinity adaptations in marine molluscs. Hydrobiol. 355, 115–126. Bidegain, G., Sestelo, M., Roca-Pardiñas, J., 2013. Estimating a new suitable catch size for two clam species: implications for shellfishery management. Ocean Coast. Manage. 71, 52–63. https://doi.org/10.1016/j.ocecoaman.2012.09.009. Burton, R.F., 1983. Ionic regulation and water balance. In: The Mollusca: Physiology, Vol. 5. (A. S. M. Saleuddin and K. M. Wilbur, eds.), 292–341. Academic Press, New York). Carpenter, E., Capone, D., 1983. Nitrogen in the Marine Environment. Academic Press, New York. Carregosa, V., Velez, C., Soares, A.M.V.M., Figueira, E., Freitas, R., 2014. Physiological and biochemical responses of three Veneridae clams exposed to salinity changes. Comp. Biochem. Physiol. Part B, Biochem. Mol. Biol. 177–178, 1–9. https://doi.org/ 10.1016/j.cbpb.2014.08.001. Elliott, J.M., Davidson, W., 1975. Energy equivalents of oxygen consumption in animal energetics. Oecologia 19, 195–201. FAO. 2018. The State of World Fisheries and Aquaculture 2018 – Meeting the sustainable development goals. Rome. Licence: CC BY-NC-SA 3.0 IGO. Fox J., Weisberg S., 2011. An {R} Companion to Applied Regression, Second Edition. Thousand Oaks CA: Sage. URL: http://socserv.socsci.mcmaster.ca/jfox/Books/ Companion. Garcia, A., Alcalde, A., Coo, A., Martinez, M., Parada, J.M., Sebe, M.P., 2001. Population dynamics of carpet-shell Ruditapes decussatus (Linnaeus, 1758) from natural fixation in intertidal zones of the Ria de Arousa, Galicia. Direccion General de Universidades E Investigacion Consejeria de Edicacion Cultura y Deportes Gobierno de Canarias, Instituto Canario de Ciencias Marinas ICCM. Gharbi, A., Farcy, E., Wormhoudt, A. Van., 2016. Response of the carpet shell clam (Ruditapes decussatus) and the Manila clam (Ruditapes philippinarum) to salinity stress. Biologia (Bratisl) 71, 551–562. https://doi.org/10.1515/biolog-2016-0072. Gnaiger, E., 1983. Heat dissipation and energetic efficiency in animal anoxibiosis: Economy contra power. J. Exp. Zool. 228, 471–490. Gosling, E., 2015. Bivalve Molluscs: Biology, Ecology and Culture, CEUR Workshop Proceedings. doi: 10.1017/CBO9781107415324.004. Griffiths, C.L., Griffiths, R.J., 1987. Bivalvia. In: Pandian, T.J., Vernberg, F.J., (ed.), Animal Energetics, Bivalvia through Reptilia. Academic Press, California. 2, 1–87. Grömping, U., 2006. Relative importance for linear regression in R: the package relaimpo. J. Stat. Soft. 17 (1), 1–27. Guzmán-Agüero, J.E., Nieves-Soto, M., Hurtado, M.A., Piña-Valdez, P., Garza-Aguirre, M. del C., 2013. Feeding physiology and scope for growth of the oyster Crassostrea corteziensis (Hertlein, 1951) acclimated to different conditions of temperature and salinity. Aquac. Int. 21, 283–297. https://doi.org/10.1007/s10499-012-9550-4. Haider, F., Sokolov, E.P., Sokolova, I.M., 2018. Effects of mechanical disturbance and salinity stress on bioenergetics and burrowing behavior of the soft shell clam Mya arenaria. J. Exp. Biol. 221, jeb.172643. https://doi.org/10.1242/jeb.172643. Hoegh-Guldberg O., Jacob D., Taylor M., Bindi M., Brown S., Camilloni I., et al., 2018. Impacts of 1.5 °C global warming on natural and human systems. In: MassonDelmotte, V., Zhai, P., Pörtner, H.O., Roberts, D., Skea, J., Shukla, P.R., et al. (Eds.), Global Warming of 1.5°C. An IPCC Special Report on the Impacts of Global Warming of 1.5 °C above Pre-Industrial Levels and Related Global Greenhouse Gas Emission Pathways, in the Context of Strengthening the Global Response to the Threat of Climate Change, Sustainable Development, and Efforts to Eradicate Poverty. Hothorn, T., Bretz, F., Westfall, P., 2008. Simultaneous inference in general parametric models. Biometric J. 50 (3), 346–363. Joaquim, S., Matias, D., Matias, A.M., Moura, P., Arnold, W.S., Chícharo, L., Baptista Gaspar, M., 2010. Reproductive activity and biochemical composition of the pullet carpet Shell Venerupis senegalensis (Gmelin, 1791) (Mollusca: Bivalvia) from Ria de Aveiro (northwestern coast of Portugal). Sci. Mar. 75, 217–226. https://doi.org/10. 3989/scimar.2011.75n2217. Juanes, J.A., Bidegain, G., Echavarri-Erasun, B., Puente, A., García, A., García, A., Bárcena, J.F., Álvarez, C., García-Castillo, G., 2012. Differential distribution pattern of native Ruditapes decussatus and introduced Ruditapes phillippinarum clam populations in the Bay of Santander (Gulf of Biscay): considerations for fisheries 16 Ecological Indicators 111 (2020) 106031 R. Domínguez, et al. Warren, C.E., Davis, G.E., 1967. Laboratory studies on the feeding, bioenergetics, and growth of fish. In: Gerking, S.D. (Ed.), The Biological Basis of Freshwater Fish Production. Blackwell Scientific Publications, Oxford, pp. 175–214. Whyte, J.N.C., 1987. Biochemical composition and energy content of six species of phytoplankton used in mariculture of bivalves. Aquaculture 60, 231–241. Wickham, H., 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag, New York. Widdows, J., 1978. Physiological indices of stress in Mytilus edulis. J. Mar. Biol. Assoc. UK 58, 125–142. https://doi.org/10.1017/S0025315400024450. Widdows, J., Staff, F., 2006. Biological effects of contaminants: measurement of scope for growth in mussels. ICES Tech. Mar. Environ. Sci. 40, 1–30. Wood, S.N., 2011. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. Royal Stat. Soc. (B) 73 (1), 3–36. https://doi.org/10.1111/j.1467-9868.2010.00749.x. Woodin, S.A., Wethey, D.S., Olabarria, C., Vázquez, E., Domínguez, R., Macho, G., Peteiro, L., 2019. Behavioral responses of three venerid bivalves to fluctuating salinity stress. J. Exp. Mar. Biol. Ecol., 522, January 2020, 151256. doi: 10.1016/j. jembe.2019.151256. of Tapes philippinarum in Northern Adriatic lagoons. Mar. Ecol. Progr. Ser. 199, 137–148. Stenton-Dozey, J.M.E., Brown, A.C., 1994. Short-term changes in the energy balance of Venerupis corrugatus (Bivalvia) in relation to tidal availability of natural suspended particles. Mar. Ecol. Ser. 103, 57–64. Sunoko, H.R.A., 1997. The effect of abrupt changes in salinity on the SFG of the mussels. J. Coast Dev. 1, 59–69. Verdelhos, T., Marques, J.C., Anastácio, P., 2015. The impact of estuarine salinity changes on the bivalves Scrobicularia plana and Cerastoderma edule, illustrated by behavioral and mortality responses on a laboratory assay. Ecol. Ind. 52, 96–104. https://doi.org/ 10.1016/j.ecolind.2014.11.022. Villalba, A., Iglesias, D., Ramilo, A., Darriba, S., Parada, J.M., No, E., Abollo, E., Molares, J., Carballal, M.J., 2014. Cockle Cerastoderma edule fishery collapse in the ría de Arousa (Galicia, nw Spain) associated with the protistan parasite Marteilia cochilli. Dis. Aquat. Organ. 109, 55–80. https://doi.org/10.3354/dao02723. Wang, Y., Hu, M., Wong, W.H., Shin, P.K.S., Cheung, S.G., 2011. The combined effects of oxygen availability and salinity on physiological responses and scope for growth in the green-lipped mussel Perna viridis. Mar. Pollut. Bull. 63, 255–261. https://doi.org/ 10.1016/j.marpolbul.2011.02.004. 17