Biology lab experiment report ( due in 20 hours)

1

RESULTS FROM PRIOR NSF SUPPORT

First time primary investigator, no prior NSF support.

PROJECT DESCRIPTION

RUI: OCEAN ACIDIFICATION: EFFECTS OF CONSTANT AND VARIABLE CO2-INDUCED OCEAN ACIDIFICATION ON FISH PHYSIOLOGY AND BEHAVIOR

STATEMENT OF THE PROBLEM

Recent investigation of ocean acidification (OA) impacts on larval and juvenile fishes has indicated a wide range of sublethal effects, including commonly reported impacts on behavior (Dixson et al. 2010; Munday et al. 2010) and otolith formation (Checkley et al. 2009; Bignami et al. 2013a). Behavioral and otolith effects have both been linked to a basic physiological response of fishes to increased environmental pCO2: the sustained increase in blood bicarbonate ion concentration ([HCO3

-]; Checkley et al. 2009; Esbaugh et al. 2012; Nilsson et al. 2012). At this point in time, investigations of OA effects on fishes have begun to provide a foundation of understanding that will help guide future research. However, nearly all studies have relied upon constant acidification treatments throughout the duration of experimental exposure. This does not accurately simulate naturally variable pCO2 that many species experience on a diurnal, episodic, and seasonal timescale (Hofmann et al. 2011; Frieder et al. 2012). The biological impact of naturally variable OA could be greatly different than what is now understood based upon studies using constant OA (Shaw et al. 2013). For example, OA-induced changes in blood chemistry may undergo similar diurnal cycles in response to variable pCO2, possibly reducing or eliminating effects on fish behavior or otolith growth. The characteristics of such a physiological response to variable pCO2 have not been described in the literature and could greatly affect the current understanding of OA impacts on fishes.

The overarching goal of this proposed research is to characterize the physiological response of fishes to variable-pCO2 OA and determine if such variability mitigates or eliminates the previously observed behavioral and otolith growth impacts of OA. This proposed research will be transformative to the field of OA research on fishes: it will address one of the most important knowledge gaps facing the field and produce data that could fundamentally change the current understanding of OA effects on fishes. The results of this research will guide the direction for future OA research, improve the ability to understand OA impacts on fishes in general, and ultimately influence future strategies to mitigate the impacts of OA on fish populations. Additionally, a description of the physiological response of fishes to variable pCO2 will provide the broader scientific community with information that is relevant to research on the physiology, biology, or ecology of fishes and their communities in present-day variable-pCO2 habitats.

BACKGROUND

Early literature on the effects of elevated pCO2 on marine fishes reported few impacts on adults and juveniles until treatments reached extreme levels (often 50,000 µatm pCO2 or higher; reviewed by (Ishimatsu et al. 2008). Comparatively, anthropogenic OA is projected to reach nearly 1000 µatm pCO2 by the year 2100 according to the Representative Concentration Pathway 8.5 model (RCP 8.5), corresponding to a 0.3 unit decline in average ocean pH (IPCC 2013). Concern over the susceptibility of the early life stages of fishes to OA has led to many recent studies that have re-focused attention on this taxonomic group under projected anthropogenic OA conditions. To date, these studies have reported mostly sublethal effects including changes in metabolism (Miller et al. 2012), growth, survival (Baumann et al. 2011), reproductive output (Miller et al. 2013), neural receptor function (Nilsson et al. 2012; Hamilton et al. 2013), otolith formation (Checkley et al. 2009; Bignami et al. 2013a), and behavioral characteristics (Dixson et al. 2010; Munday et al. 2010; Hamilton et al. 2013, among others). These

2

effects are not always negative and typically vary among species and regions, making it difficult to draw general conclusions about the potential impact that future OA will have on fishes.

Two consistently reported impacts of OA on fishes are altered otolith formation and behavioral changes. Otoliths are paired, calcium carbonate, inner ear stones which contribute to auditory and vestibular sensation in teleost fishes. Calcium carbonate is passively deposited in annual and daily increments, with widths that are proportional to fish somatic growth, therefore they are often used as a proxy for age and growth by fisheries biologists and ecologists (Campana 2005). Enhanced otolith growth has been commonly observed under OA conditions of 800 µatm pCO2 and above (Checkley et al. 2009; Munday et al. 2011b; Bignami et al. 2013a; 2013b, among others), although a lack of effect on otolith growth (Munday et al. 2011a; Frommel et al. 2013) and one report of reduced otolith growth (Maneja 2012), have been reported over a range of pCO2 treatments (600 to 3200 µatm). Changes in otolith formation have implications for the sensory function of otoliths (Bignami et al. 2013a) as well as the interpretation of otolith growth history data, due to a decoupling of the relationship between otolith increment width and somatic growth (Bignami et al. 2013b). Both implications are relevant under future OA conditions or in present day acidified environments such as fjords (Thomsen et al. 2010) or upwelling zones (Feely et al. 2008).

OA-induced behavioral effects on fishes range from disruption of olfactory homing ability (Munday et al. 2009), auditory behavior (Simpson et al. 2011), and behavioral lateralization (Domenici et al. 2012), to altered learning ability, risk assessment (Ferrari et al. 2012a; 2012b), and increased anxiety (Hamilton et al. 2013). It has also been demonstrated that the altered behavior of juvenile fishes raised under OA conditions can produce up to 9-fold higher predation mortality (Munday et al. 2010) and alter fish habitat preference (Devine & Munday 2012) in the wild. Under future OA conditions, such effects could ultimately influence important ecological processes such as recruitment of juvenile fishes, post- settlement ecological interactions, and population dynamics. The occurrence of behavioral effects in multiple study species, across trophic levels (Dixson et al. 2010; Munday et al. 2012), and between geographic regions (i.e., temperate Forsgren et al. 2013; Hamilton et al. 2013, as well as tropical) implies that these effects may be a pervasive impact of OA on fishes throughout the ocean.

It has been proposed that the otolith and behavioral effects of OA have a common driving mechanism, tied to the basic physiological response of fishes to elevated environmental pCO2. When exposed to elevated pCO2, teleost fishes experience a corresponding decline in blood pH, which is compensated for via the retention and active uptake of bicarbonate ions (Claiborne & Heisler 1986; Esbaugh et al. 2012). Full compensation of blood pH is achieved in as little as 2 to 4 hrs (Esbaugh et al. 2012) and is maintained for the duration of elevated pCO2 exposure (Claiborne & Heisler 1986; Michaelidis et al. 2007; Esbaugh et al. 2012). Increased extracellular bicarbonate ion concentration ([HCO3

-]) likely produces an increased aragonite saturation state in the endolymph fluid that surrounds otoliths, therefore contributing to accelerated calcification (Checkley et al. 2009; Bignami et al. 2013a). Additionally, as HCO3

- is retained in the blood, negatively charged chloride ions (Cl-) are excreted from the blood to balance electric charge (Claiborne & Heisler 1986). However, this process also alters the electrochemical gradient across neuron cell membranes. In adult vertebrates, GABAA receptors are the main neuronal inhibitory mechanism, which hyperpolarize the neuron by allowing the flow of Cl- from extracellular to intracellular space (Bormann et al. 1987; Hamilton et al. 2013). Nilsson et al. (2012) suggest that under OA conditions, decreased extracellular [Cl-] results in a reversal of Cl- conductance (i.e. outflow from the neuron), and thus an excitatory depolarization of the neuron when GABAA is activated. Data supporting this mechanism was provided when administration of a GABAA receptor antagonist (gabazine) was shown to reverse OA-induced behavioral effects by preventing the activation of GABAA receptors (Nilsson et al. 2012). More recently, Hamilton et al. (2013) verified this mechanism with the use of both a GABAA agonist (muscimol) and a GABAA

antagonist (gabazine) in splitnose rockfish (Sebastes diplopora). Consistent with a reversed gradient for Cl- ions in OA exposed rockfish, Hamilton et al. (2013) observed an increase in OA-treatment fish anxiety that was potentiated by

3

muscimol, whereas muscimol administration produced a decrease in anxiety in control rockfish, consistent with the proposed OA-induced changes in electrochemical gradient (Fig. 1).

The behavioral effects of OA can be induced via exposure to OA for a minimum period of 2 to 4 d (Munday et al. 2010; Nilsson et al. 2012), which implies that a sustained period of high [HCO3

-]/low [Cl-] may be required to elicit an effect (given the ability of fishes to acclimatize their blood chemistry within hours; Esbaugh et al. 2012). To date, exposure to constant OA treatments has been the status quo in all OA-related behavioral studies of fishes and in all physiological studies of fish blood chemistry response to elevated pCO2. However, constant ocean pH/pCO2 is not an accurate assumption for the natural environments in which most study species live (McElhany & Shallin Busch 2012; Shaw et al. 2013), such as coral reefs, seagrass meadows, and kelp forests. These habitats can experience large diurnal pH fluctuations of up to 0.7 units (Ohde & van Woesik 1999; Perez-Dominguez et al. 2006; Hofmann et al. 2011; Frieder et al. 2012). In particular, kelp forests off southern California have been shown to exhibit diurnal pH variability of up to 0.36 units, with an average of 0.11 units, largely driven by differences in nocturnal respiration (high pCO2) and daytime photosynthesis (low pCO2; Frieder et al. 2012). Put into context with projected future OA, a pH change of 0.30 units is equivalent to the total magnitude of change in average ocean pH projected to occur by year 2100 under RCP 8.5 (936 µatm atmospheric pCO2; IPCC 2013). Therefore, a diurnal swing of 0.30 pH units (equivalent to approximately 600 µatm pCO2) in kelp forest ecosystems could drive pCO2 levels 300 µatm above and below projected atmospheric values. For example, under RCP 6.0 projected conditions for the year 2100 (670 µatm pCO2; IPCC 2013), organisms in a kelp forest could experience pCO2 values ranging from 370 µatm during the day to 970 µatm at night. Likewise, if conditions meet the RCP 8.5 projections of 936 µatm pCO2 (IPCC 2013), values could range from 636 (i.e., day) to 1236 µatm (i.e., night). It has recently been suggested that variable environmental conditions could greatly influence the impact of OA on fishes, as diurnal cycling may shift pH/pCO2 above and below an organism’s “threshold” for impact (Bignami 2013c; Shaw et al. 2013). However, these arguments have been made in theory alone, and the empirical data

Figure 1. Model of ocean acidification-induced changes in GABAA receptor ionic gradients. Under normal conditions the concentration of Cl- ions is slightly higher in the cerebrospinal fluid (extracellular) relative to the cytoplasm (intracellular) and the equilibrium potential for Cl- (ECl) is near the resting membrane potential. When GABAA receptors open, Cl- flows into the neuron and counters depolarization, keeping the membrane potential more negative, and reducing neuronal activity. Ocean acidification (OA) induces hypercapnic acidosis in plasma, which fish counteract by excreting excess H+ and accumulating HCO3

-. This leads to a decrease in plasma [Cl-] (to maintain charge balance), thus leading to an alteration in ECl. In this condition, opening of GABAA receptors results in net Cl- movement out of the neuron, causing membrane depolarization and increasing excitation of neural pathways. Gabazine is an allosteric antagonist of GABAA receptors that prevents the channel from opening, whereas muscimol is a selective GABAA agonist that binds to the same site as GABA and increases channel opening (independent of presynaptic GABA release). Muscimol has opposite effects in control and OA-acclimated fish when the membrane potential is close to resting because ECl is altered. Figure and legend provided courtesy of T. Hamilton (Hamilton et al. 2013).

4

necessary to address this important and potentially paradigm-shifting concept is currently lacking. This proposed research aims to produce the data necessary to address this crucial knowledge gap.

When blood pH changes more than what can be compensated for with non-bicarbonate blood buffering capacity, then retention/excretion of HCO3

- is necessary (Melzner et al. 2009). The result of exceeding this non-bicarbonate blood buffer capacity has been demonstrated in gulf toadfish (Opsanus beta), which increase blood [HCO3

-] when exposed to environmental pCO2 levels at or above 750 µatm (Esbaugh et al. 2012). Conversely, toadfish exhibit no significant change in blood chemistry at 560 µatm, suggesting that 560 µatm pCO2 does not induce a change in blood pH beyond the non-bicarbonate blood buffer capacity (Esbaugh et al. 2012). While some studies have indirectly addressed the consequences of removing OA conditions by evaluating the persistence of behavioral effects (which can last hours to days; Munday et al. 2010; Devine & Munday 2012; Hamilton et al. 2013), the influence of variable pCO2 on the blood chemistry of fishes is not known. However, if diurnal shifts in pCO2 cross the blood buffer threshold, then similar variability in blood [HCO3

-] can be expected because removal of high-pCO2 conditions will likely result in a shift from HCO3

- retention to excretion. Likewise, diurnal cycles in pCO2 may drive a similar diurnal variation in blood chemistry. Until the entire range of diurnal pCO2 variability occurs above the non-bicarbonate blood buffer capacity, daily pCO2 minimums could result in a “release” of the physiological need to retain HCO3

-. This would prevent the occurrence of the sustained 2-4 d exposure necessary to cause behavioral effects in fishes, and could effectively reduce or remove the basic physiological mechanism driving OA-induced behavioral and otolith effects in fishes. This has the potential to fundamentally alter the current understanding of projected OA impacts on fishes and is the primary justification for pursuing this proposed research.

To address the target research area of predicting the consequences of OA on ecosystem health and function, it is necessary to gain a better understanding of the physiological mechanisms of individual organisms. Until the response to natural OA variability is better understood for fishes, it may not be possible to answer the question of To what extent will OA affect an organism’s performance? The proposed research will improve the ability to answer this question by providing definitive insight into the response of fishes to variable OA, influencing the interpretation of past data, and guiding research in the future. If variable pCO2 has an effect on the response of fishes to OA, then it will be necessary to determine if the results of past research are relevant to OA in the natural environment. More importantly, an improved understanding of the response to variable-pCO2 OA will greatly impact the direction and interpretation of future OA research, resulting in data that are more representative of the natural environment. This will create an opportunity to formulate more confident scientific conclusions, that will be necessary to consider and apply during the development of OA mitigation strategies or management efforts.

The potential impact of the proposed research also applies to disciplines other than OA. For example, ocean geochemistry is affected by marine fish because they contribute a significant portion of the oceanic carbonate budget (3-15 %) via precipitation of carbonate as a byproduct of maintaining osmotic balance (Wilson et al. 2009). This contribution greatly influences the oceanic carbon cycle and Wilson et al. (2009) predicted that it will be exaggerated under future high-pCO2 conditions. However, conclusions that are rooted in the current understanding of physiological mechanisms under constant- pCO2 exposure may need to be reinterpreted with consideration for how present day or future variability influences these processes. There is also a need to understand the physiological and behavioral response of fishes to variable pCO2 because fish that inhabit high-pCO2 environments such as upwelling zones (Feely et al. 2008) and fjords (Thomsen et al. 2010) may already exhibit the effects of dramatic natural variability in pCO2. For example, rockfish (Sebastes spp.) and many other species that inhabit coastal waters naturally experience pCO2 levels that exceed 1000 µatm (Feely et al. 2008). In an experimental setting, similar constant-pCO2 treatments (1125 µatm pCO2) produced increased anxiety and altered the sheltering behavior of juvenile rockfish (Hamilton et al. 2013). Although Hamilton et al. (2013) presented this study from an OA perspective, the authors also made the important distinction that these effects could

5

be experienced during present day upwelling events and could ultimately influence juvenile mortality, community ecological interactions, and population dynamics. Such behavioral effects may also occur in other species and could be currently impacting fish population dynamics in regions with great temporal or spatial pCO2 variability. Therefore, it is critical to gain a better understanding of how fishes respond to these environmental factors regardless of projected OA scenarios, especially when working with ecologically and economically important species.

STUDY SPECIES

The study species proposed for this research is the kelp rockfish (Sebastes atrovirens). Sebastes spp. are native predatory fish in temperate Eastern Pacific waters and are a popular sport fishing and commercial fishery target, with many populations experiencing decline in recent decades (Love et al. 1998; Lea et al. 1999). The kelp rockfish grows to a maximum size of approximately 40 cm over a maximum lifespan of 25 years (Lea et al. 1999). Mating occurs between February and June, with viviparous females releasing planktonic larvae that settle into kelp forests after a 1-2 month pelagic larval duration (Nelson 2001). Late-stage larvae and juveniles recruit to the kelp canopy before migrating to midwater and demersal zones of kelp forests as they mature (Schiel & Foster 1985; Nelson 2001). Trials with adult fish will utilize this species, but depending on availability of recruits, trials with juveniles may utilize another species from the genus Sebastes.

HYPOTHESES, GOALS, AND OBJECTIVES

The physiological response of blood chemistry to variable OA could differ in a number of ways from the currently understood sustained increase in blood [HCO3

-] due to constant high-pCO2. For simplicity, the three potential scenarios addressed here will be referred to as Scenario A, B, and C. In Scenario A, pCO2 is diurnally elevated above and decreased below the blood buffer capacity. In response, blood [HCO3

-] reflects this change and undergoes full compensation and “recovery” with each increase and decrease in pCO2 (Fig. 2). Scenario A does not result in a sustained increase in [HCO3

-] (or corresponding decrease in [Cl-]) and may therefore prevent the development of behavioral and otolith formation effects. In Scenario B, variable pCO2 diurnally crosses the blood buffer threshold as described in Scenario A, except the physiological result is [HCO3

-] oscillation that gradually increases average [HCO3

-] with more restricted diurnal shifts than in Scenario A. Scenario B does not provide full “recovery” of normal blood [HCO3

-] each day, and because it represents a sustained increase in average [HCO3

-], behavior and otolith formation may be impacted similarly to a constant elevated-pCO2 treatment. In Scenario C, the entire range of environmental pCO2 variability occurs above the blood

Figure 2. Diurnal variability of pCO2 (top panel) and response of blood [HCO3

-] (bottom panel). RCP 6.0 (top panel, blue curve) crosses a threshold for non-bicarbonate blood buffer capacity (black dashed line) each night (grey shaded areas), causing an increase in blood [HCO3

-]. The opposite occurs when pCO2 drops below the threshold during the day. Resulting [HCO3

-] could be variable with full diurnal recovery (Scenario A, dark blue dashed line) or an overall increase in [HCO3

-] and more limited diurnal variability (Scenario B, light blue dotted curve). RCP 8.5 (top panel, red line) consistently exceeds the buffer threshold, thus [HCO3

-] will remain elevated while undergoing some diurnal variation (Scenario C, dashed red curve). The well-described sustained response to constant elevated-pCO2 is displayed for comparison (black curve).

6

buffer threshold. Therefore, recovery of normal blood chemistry never occurs despite diurnal variation in [HCO3

-], resulting in behavioral and otolith effects equivalent to a constant elevated-pCO2 treatment. In this proposed research, Scenarios A and B are represented by the RCP 6.0 variable-pH treatment, and Scenario C is represented by the RCP 8.5 variable-pH treatment (described below).

The overarching goal of the proposed research is to characterize the physiological response of fishes to variable-pCO2 and determine if such variability ultimately mitigates or eliminates previously observed impacts of OA. The specific objectives correspond to three subsets of alternative hypotheses that will be tested by exposing rockfish to constant-pCO2 and variable-pCO2 OA scenarios. The first two objectives focus on treatment impacts on adults, while the third extends the study to include the juvenile life stage. Juveniles are included because OA impacts could influence many processes that are of critical ecological importance during this stage (i.e., recruitment, sheltering, and habitat transition). Subsequent reference to “observed” effects refers to previously described OA effects on fishes, specifically increased anxiety in rockfish (Hamilton et al. 2013) and increased otolith growth in other species (Checkley et al. 2009; Bignami et al. 2013a).

Objective 1: To measure and compare the changes in blood chemistry of adult kelp rockfish in response to constant-pCO2 and diurnally-variable-pCO2 OA scenarios that reflect RCP 6.0 and 8.5 projections for the year 2100.

HA1a: Diurnal variability of environmental pCO2 under RCP 6.0 conditions will cause adult kelp rockfish to exhibit full diurnal compensation and recovery of blood [HCO3

-] (i.e., Scenario A).

HA1b: Diurnal variability of environmental pCO2 under RCP 6.0 conditions will cause adult kelp rockfish to exhibit a diurnal increase and decrease of blood [HCO3

-], but without full recovery of normal blood chemistry during periods of low-pCO2 (i.e., Scenario B).

HA1c: Diurnal variability of environmental pCO2 under RCP 8.5 OA conditions will cause an increase in average blood [HCO3

-] with variability but no diurnal recovery (i.e., Scenario C).

Objective 2: To determine and compare the behavioral and otolith growth effects of constant-pCO2 and diurnally-variable-pCO2 OA on adult rockfish, under RCP 6.0 and 8.5 projected scenarios for the year 2100.

HA2a: Diurnal variability of pCO2 under RCP 6.0 OA conditions will result in the reduction or absence of behavioral and otolith growth effects otherwise observed under constant-pCO2 RCP 6.0 OA conditions (i.e., Scenario A).

HA2b: Diurnal variability of pCO2 under RCP 8.5 conditions will produce similar effects compared to those observed under constant-pCO2 RCP 8.5 OA conditions (i.e., Scenario C).

Objective 3: To determine and compare the behavioral and otolith growth impacts on juvenile rockfish exposed to constant-pCO2 and diurnally-variable-pCO2 OA, under RCP 6.0 and 8.5 projected scenarios for the year 2100.

HA3a: Diurnal variability of pCO2 under RCP 6.0 OA conditions will result in the reduction or absence of behavioral and otolith growth effects otherwise observed under constant-pCO2 RCP 6.0 OA conditions (i.e., Scenario A).

HA3b: Diurnal variability of pCO2 under RCP 8.5 conditions will produce similar effects compared to those observed under constant-pCO2 RCP 8.5 OA conditions (i.e., Scenario C)

EXPERIMENTAL TIMELINE

The proposed research will take place over a 3-yr period, split into a development phase and subsequent experimental phases (Table 1). The development phase will involve construction and testing of a seawater system, OA treatment system, and laboratory equipment. The development phase will last approximately 10 months, during the 2014-2015 academic year. The subsequent experimental phases will

7

occur throughout the remaining 26 months. Most research will be conducted during the summer months, when classes are not in session, although preliminary work and some experiments may be conducted during the academic year where time allows.

During the first summer, Objective 1 will be addressed using experiments designed to evaluate the impact of constant vs variable OA on blood chemistry. Experiments will last approximately 2 weeks each and will utilize adult rockfish. To accommodate four treatments per experiment (constant control and OA treatment, variable control and OA treatment), each experiment will be conducted at least once with control and RCP 6.0 treatments, and at least once with control and RCP 8.5 treatments. Objective 2 will be addressed during the second summer of research. These experiments will last at least three weeks, and will test the effects of constant vs variable OA on adult kelp rockfish behavior and otolith growth. To minimize the limitation of experimental duration due to mortality, no surgery or blood chemistry analyses will be attempted during these experiments. During the third summer of research, Objective 3 will be addressed by testing the impact of constant versus variable OA on juvenile rockfish behavior and otolith growth. These experiments will last at least three weeks and will not involve surgery or blood analysis.

Table 1. Objective overview & timeline with summarized treatments, hypotheses, samples sizes, & methods.

YEAR OBJECTIVE TREATMENTS ALTERNATIVE HYPOTHESES SAMPLE SIZE

METHODS

Development N/A N/A N/A Construction RAS, CO2 control system, & lab equipment

Present day constant & variable, Yr 2100 RCP 6.0 constant & variable

HA2a: Variable RCP 6.0 OA allows full [HCO3

-] recovery HA2b: Variable RCP 6.0 OA does not allow full [HCO3

-] recovery

1

Objective 1: Adult rockfish blood chemistry (2 exp. minimum)

Present day constant & variable, Yr 2100 RCP 8.5 constant & variable

HA2c: Variable and constant RCP 8.5 OA do not allow full [HCO3

-] recovery

8 to 10 per treatment, per exper.

- Adult rockfish collected & acclimated to RAS - Fish surgically catheterized - Diurnal variation in pH vs. constant pH - Blood samples (high & low time resolution) for min. of 3 d - Blood pH & Total CO2 measured - Body size & weight compared pre/post experiment

Present day constant & variable, Yr 2100 RCP 6.0 constant & variable

HA3a: Behavior/otoliths will be impacted by constant but not variable RCP 6.0 OA

2

Objective 2: Adult rockfish behavior & otoliths (2 exp. minimum)

Present day constant & variable, Yr 2100 RCP 8.5 constant & variable

H3b: Behavior/otoliths will be impacted by both constant and variable RCP 8.5 pH

8 to 10 per treatment, per exper.

- Adult rockfish collected & acclimated to RAS - Otoliths stained with Alizarin Red - Exposed to treatments; 3 wk minimum - Light/dark preference (anxiety) & shelter behavior assessed at 1 and 3 wk - Otoliths extracted for size, shape, increment analysis - Pre/post body length & weight compared

Present day constant & variable, Yr 2100 RCP 6.0 constant & variable

HA4a: Behavior/otoliths will be impacted by constant but not variable RCP 6.0 OA

3

Objective 3: Juvenile rockfish behavior & otoliths (2 exp. minimum)

Present day constant & variable, Yr 2100 RCP 8.5 constant & variable

HA4b: Behavior/otoliths will be impacted by both constant and variable RCP 8.5 OA

15 to 25 per treatment, per exper.

- Rockfish recruits collected & acclimated - Otoliths stain (alizarin red) - Exposed to treatments; 3 wk minimum - Light/dark preference (anxiety) & shelter behavior assessed at 1 & 3 wk - Otoliths extracted for size, shape, increment analysis - Pre/post body length & weight compared

8

METHODS

EXPERIMENTAL SEAWATER SYSTEM

All fish will be held and exposed to treatments in a recirculating aquarium system (RAS) with approximately 12000 L (~3000 gal) capacity, constructed on the undergraduate campus of Concordia University, Irvine (CUI). The RAS will serve a dual purpose, providing a platform for this proposed research, but also providing a resource for undergraduate marine science education. To ensure proper RAS filtration and capacity, the final design and installation will be completed under direct consultation of an aquaculture and RAS design expert from the University of Miami Experimental Hatchery (Benetti et al. 2008, Stieglitz et al. 2011). The basic components of the RAS can be categorized into three main areas, described below: seawater filtration, carbonate chemistry manipulation, and experimental tanks.

Seawater filtration will include a series of mechanical and biological filtration techniques, as well as water sterilization and temperature control. Seawater will receive initial mechanical filtration in a bead filter before being directed through a protein skimmer to remove smaller organic compounds. Biological filtration will be applied with a fluidized bed filter, which will provide microbial breakdown of toxic compounds (i.e. ammonia and nitrite) into less toxic compounds (i.e. nitrate). Final mechanical filtration through 1-µm canister filters will be applied prior to UV light and ozone sterilization. Then, seawater will flow through an 18,000 BTU heat exchanger (chiller) and in-line heaters to maintain RAS water temperature at approximately 15 °C. During experimental use, RAS water will be replaced at a rate of 5- 10 % d-1, from a large reservoir tank that will be replenished with commercially supplied filtered natural seawater (Catalina Water Co., Long Beach).

Following filtration, sterilization, and temperature adjustment, seawater will flow to one of four 1000 L pCO2 treatment chambers. Each chamber will utilize a pH stat system, controlled by an AquaMedic AT Control System, to adjust seawater pH to the desired level via direct addition of pure CO2 gas using a solenoid valve (Miller et al. 2013). The AT Control System is capable of maintaining constant pH or a pre-programmed variable pH cycle. Seawater pH and temperature will be continuously monitored using AquaMedic probes installed directly into the seawater plumbing prior to delivery of seawater to experimental tanks. Additional monitoring of experimental tank pH, temperature, dissolved oxygen, and conductivity will be conducted twice d-1 using a portable meter (Hach HQ40d).

Experimental tanks will be composed of 16 replicated 150 L (~40 gal) aquaria, each capable of holding 2 adult rockfish, or 10-15 juvenile rockfish, depending on size. Removable opaque dividers will be placed in each tank prior to and during experimentation, to separate individual adult fish and prevent any aggressive interaction. Dividers will be perforated to allow even water flow throughout both tank sections. This RAS design will allow 4 treatments with 4 full replicates to be applied during any one experiment. Individual aquaria will function on a flow-through process, receiving water that has already been filtered and adjusted to the desired pCO2 treatment level. Tanks will receive at least 300 % water exchange h-1 and outflow from all experimental aquaria will be combined into a common drainage sump, from which it will be pumped through a degassing chamber to re-establish ambient pCO2 prior to filtration, CO2 treatment, and reintroduction into experimental tanks.

CARBONATE CHEMISTRY PARAMETERS

Experimental treatments will be tested and calibrated prior to the start of any individual experiment. During experiments, weekly water samples will be collected from each individual tank and poisoned with saturated mercuric chloride for later analysis. Total alkalinity (TA) will be determined from water samples via gran titration using a Metrohm 870 Titrino Plus Titrator and pH will be measured using a Ross electrode (Langdon 2005). CO2SYS will be used to calculate carbonate chemistry parameters that are not directly measured (i.e., pCO2) with input data from weekly TA measurements, daily tank pH measurements, and continuous AT Control System pH data (Lewis & Wallace 1998).

9

Each experiment will include four treatment conditions: constant present day control, variable present day control, constant elevated-pCO2, and variable elevated-pCO2. To allow for maximum replication within the logistic and spatial limitations of the RAS system, elevated-pCO2 treatments (RCP 6.0 and 8.5) will be applied during separate experimental trails. Elevated-pCO2 treatments were chosen according to RCP 6.0 and 8.5 projected conditions for the year 2100 (IPCC 2013). The first scenario (RCP 6.0) represents a stabilization pathway with atmospheric CO2 concentration reaching 670 ppm and ocean pH declining 0.20 units, while the RCP 8.5 scenario represents a high-CO2 “business-as-usual” pathway with 936 ppm CO2 and an ocean pH decline of 0.30 units by year 2100 (IPCC 2013). Control treatments will represent present day average surface ocean pH of 8.10. The magnitude of the applied diurnal pH variability will be 0.30 units (approximately 600 µatm pCO2), corresponding to field data from southern California kelp forests, where pH changes as much as 0.36 units in a day and more than 0.50 units over multi-day time scales (Hofmann et al. 2011; Frieder et al. 2012). Diurnal pH variability will be applied as a 0.15 unit pH shift above and below the mean treatment pH in each variable-pCO2 treatment, including present day controls. Maximum and minimum pH will be programmed to occur at 1500 hr (maximum pH/minimum pCO2) and 0300 hr (minimum pH/maximum pCO2) to reflect the approximate timing of natural pH variability due to photosynthesis and respiration (Frieder et al. 2012). One key aspect of this particular combination of treatments is that diurnal variation of the RCP 6.0 treatment will likely cross the fish non-bicarbonate blood buffer capacity (based upon data from another species; Esbaugh et al. 2012), while the RCP 8.5 treatment will likely remain above the blood buffer capacity despite diurnal variation. This will allow for a comparison of physiological and behavioral responses to elevated pCO2 when treatments are occasionally or chronically surpassing this threshold. Treatment levels, variability, and corresponding pH/pCO2 minimums and maximums are outlined in Table 2.

Table 2. Proposed treatments to simulate present day and future projected scenarios of ocean acidification. Present day (control) and projected acidification scenarios (RCP 6.0 and 8.5) will be applied as constant (-C) and diurnally variable (-V) treatments. Approximate mean, minimum, maximum, and diurnal range (Δ) of pH and pCO2 are provided. Max. pH/min. pCO2 will occur at 1500 h, and min. pH/max. pCO2 will occur at 0300 h.

Treatment Mean pH pCO2 (ppm) Min-Max pH Min-Max pCO2 (ppm) Δ pH Δ pCO2 (ppm) Control-C 8.10 400 8.10 400 RCP 6.0-C 7.90 670 7.90 670 RCP 8.5-C 7.70 936 7.70 936

0.00 0

Control-V 8.10 400 7.95-8.25 100-700 RCP 6.0-V 7.90 670 7.75-8.05 370-970 RCP 8.5-V 7.70 936 7.55-7.85 636-1236

0.30 600

STUDY SPECIES COLLECTION AND CARE

All methods will be approved by the Concordia University Animal Care and Use Committee prior to any animal collection or experimentation. Rockfish will be collected under a California Department of Fish and Game Scientific Collecting Permit. Adult fish will be collected using hook and line, or by SCUBA divers using nets, from kelp forests along the southern California coast 1-2 weeks prior to experimentation. Location and environmental parameters (e.g., temperature, salinity, DO) will be recorded upon collection, and water samples will be taken as previously described. Juvenile kelp rockfish will be collected using Standard Monitor Units for the Recruitment of Fishes (SMURFs) moored for 24- 48 h at a depth of 1 m, approximately 100 m offshore of kelp forests (Ammann 2004). Alternatively, juvenile fishes may be collected in the kelp canopy by SCUBA divers using hand nets. Prior to introduction into the RAS, fish will be treated with Malachite Green in formalin (AquaVet) to prevent the introduction of ectoparasites (McDonald et al. 2003), and then allowed to recover from capture in the RAS for a minimum period of one week. Baseline measurements of standard length and wet weight will be obtained upon initial collection and 2 d prior to experimentation, using manual measurement and

10

digital measurement of images taken of each fish. Any fish displaying signs of poor health or excessive stress (e.g., refusing food) will be removed prior to beginning the study.

During the capture recovery period, adult and juvenile fishes will be provided a variety of fresh and prepared foods (e.g. market squid, anchovies, prepared flake food) once d-1. Adult fish will be fasted beginning 3 d prior to experiments and during any period when blood samples will be collected (Michaelidis et al. 2007). Fish will be fed a normal ration at all other times, except adult and juvenile fish will also be fasted for a minimum of 18 h prior to behavioral assessment. Following completion of each experiment, fish will be anesthetized and euthanized with an overdose of MS-222 (3 g/L; Wood et al. 1997), standard length and mass will be measured, and each fish will be stored in 95 % ethanol for later analysis.

BLOOD PHYSIOLOGY

Survival surgery procedures and measurement of blood chemistry parameters will be conducted on adult rockfish using previously established methods (Wood et al. 1997; Esbaugh et al. 2012), described briefly here. In addition to literature as a source of methodology, in-depth training will be provided to PI Bignami by an expert, Dr. Martin Grosell, at the University of Miami, FL (see letter of support). Rockfish will be anesthetized using a 0.5 g/L dose of buffered MS-222 prior to surgical implantation of a heparin-saline filled caudal artery catheter according to Wood et al. (1997). Fish will be allowed to recover in tanks individually for a period of at least 48 h. Subsequently, each blood sample will consist of an initial 200 µL withdrawal to clear the catheter, followed by a 250 µL analytical sample. The initial 200 µL sample will be re-infused immediately, and each analytical sample will undergo low- speed centrifugation for 1 min to separate plasma and red blood cells (RBCs). Plasma will be stored for later analysis and RBCs will be re-suspended in 250 µL saline and infused into the fish through the caudal catheter to prevent the occurrence of anemia due to frequent sampling. Blood plasma will be kept out of contact with air at all times to prevent changes in carbonate chemistry. Plasma pH will be measured by injecting a 50 µL subsample of plasma into a temperature-controlled gastight sleeve equipped with a Hach micro-sample pH electrode. An additional 50 µL of plasma will be used to directly measure [Cl-] with an ion-specific micro-sample electrode (Lazar Research Laboratories). Total dissolved inorganic carbon of the remaining sample will be measured using an Olympic total CO2 analyzer. [HCO3

-] will be calculated using the Hendersen-Hasselbach equation.

One hour prior to introduction of treatments, a blood sample will be collected from each fish to establish baseline blood chemistry parameters. Treatments will be brought to full strength over a 1 h period prior to the point of “maximum pH/minimum pCO2” treatment conditions (1500 h) and will thereafter be allowed to follow the appropriate pre-programmed schedule of pH change for each treatment. Therefore, constant-pH treatments will reach their respective pH conditions within one hour, but variable-pH treatments will initially reach the daytime “maximum pH/minimum pCO2” with subsequent decline in pH throughout the evening and nighttime hours until “minimum pH/maximum pCO2” is attained at 0300 h (Table 2). Once treatment conditions reach full strength (designated as 0 h) blood samples will be taken from each fish at 0, 1, 2, 5, 9, 13, 17, 21, and 25 h sampling points, followed by subsequent samples at least every 12 h until fish have been exposed to treatments for 3 d (Fig. 3). The goal of this sampling schedule is to collect a high-resolution profile of blood chemistry on the first day of exposure (Esbaugh et al. 2012), followed by subsequent samples that target the expected maximum and minimum blood [HCO3

-]. More frequent sampling may be possible depending on the resilience of the study species. The timing of sampling is designed to allow for an estimated 1 h of physiological lag following max./min. pCO2 conditions (i.e. sampling blood at 0400 h although minimum water pH occurs at 0300 h). Following the initial 3 d of sampling, collection of samples will be suspended for 6 d to allow fish to feed normally for 3 d and fast for 3 d prior to additional sampling, given adequate survival. These long-term sampling points will be taken at least every 6 h beginning at 0400 h (1 h after minimum diurnal pH conditions) for 2 d, to capture any changes in long term acclimation to treatment conditions.

11

BEHAVIORAL ASSESSMENT

Rockfish behavioral response to each treatment will be assessed using methods adapted from Hamilton et al. (2013). Dr. Trevor Hamilton and co- author, Dr. Martin Tresguerres, have agreed to provide intellectual and personal assistance with the following behavioral tests and analyses (see letters of support). Briefly, behavioral assessment will be conducted after approximately one and three weeks of treatment exposure for adult and juvenile rockfish. The first behavioral assessment will evaluate individual fish scototaxis, the preference for light or dark surroundings, which has been described as an appropriate indicator of anxiety and risk assessment in fishes (Maximino et al. 2010; 2013; Hamilton et al. 2013). Scototaxis, hereafter referred to as light/dark preference, will be evaluated using a light/dark arena consisting of a rectangular enclosure that has white walls at one end and black walls at the other, with an open top for observation (Maximino et al. 2010; Hamilton et al. 2013). Arena dimensions will differ for adult and juvenile rockfish, with length, width, and depth dimensions that are approximately 8, 2, and 2 times the fish body length. Rockfish will be individually introduced to the arena and will be allowed to explore the apparatus freely for a 15 min period while being recorded from above using a digital low-light video camera. No acclimation period will be provided prior to data collection because this test evaluates anxiety, thus the period of initial acclimation to the apparatus may provide the most relevant data (Hamilton et al. 2013). The entire arena will be rotated 180 degrees between each trial, to control for any subtle difference in conditions that might impact side preference. Ethovision XT motion tracking software will be used to analyze digital videos and collect measurements including total time spent in either light or dark compartments, number of entries into the light compartment, average velocity, meandering, and duration of freezing (motionless) events (Maximino et al. 2013). Each trial will be conducted during daylight hours under consistent light intensity and with fresh seawater from the respective average treatment conditions (i.e. 400, 670, and 936 ppm CO2). Repeated testing will be conducted during both daylight and nighttime hours (with infrared illumination), with treatment conditions matching the respective diurnal minimum and maximum pCO2 for each treatment. No fish will be tested more than once in any 24 h period, but individual fish will experience multiple behavioral trials over several days because it has been demonstrated that daily repeated trials with fish do not result in habituation (Maximino et al. 2010).

The second behavioral assessment will be a “shelter test”, in which an individual fish is placed in a circular arena with an unfamiliar object in the center to serve as a shelter (Hamilton et al. 2013). This test will be conducted using the same duration, treatment, and recording techniques described for the light/dark preference test, except using a circular arena with a diameter at least 6 times the fish body length, grey walls/bottom, and a central object with of approximately 0.5 to 1 body lengths. Digital videos will be analyzed to determine the amount of time each fish spends in association with the shelter object (defined as any time spent within 1.5 body lengths of the object), time associated with the container walls (thigmotaxis), and general routine swimming activity such as velocity, meandering, immobility, and net-

Figure 3. Proposed minimum blood sampling times during initial 3 d treatment exposure. Treatments will reach full strength at 0 h (vertical dashed line), after being gradually increased beginning 1 h prior (vertical dotted line). Collection times for blood samples (black circles) will capture the expected fluctuation of blood [HCO3

-] (red dashed curve) in response to changing water pCO2 (blue curve). An initial baseline sample will be collected 2 h prior to full treatment strength followed by high temporal resolution sampling at 0, 1, 2, 5, 9, 13, 17, 21, and 25 h. Subsequent samples will be collected at 37, 49, 61, and 73 h, with the goal of capturing diurnal [HCO3

-] maximums and minimums estimated to occur approximately 1 h after max./min. water pCO2.

12

to-gross displacement (Maximino et al. 2013). Fish will be returned to their respective experimental tanks following each behavioral test.

ASSESSMENT OF OTOLITH GROWTH

One day prior to beginning an experiment, adult and juvenile rockfish otoliths will be stained by exposing fish to 200 mg/L alizarin red solution for 24 h, followed by transfer into alizarin-free water (Liu et al. 2009). Upon termination of the experiment, digital images of all sacrificed fish will be captured and standard length will be measured digitally for comparison to pre-experiment length and weight data. All otoliths (3 pairs, 6 total) will be extracted, with one set (e.g., left side) mounted on microscope slides using thermoplastic glue and the other set stored dry. Digital images of mounted otoliths will be captured with the sulcus side down under light microscopy using a 3.3 megapixel color digital microscope camera (Lumenera, Infinity 2-3). Images will be analyzed using ImagePro Plus (Media Cybernetics) to measure otolith size and shape metrics, including area, length, circularity, rectangularity, and symmetry between corresponding left and right otoliths (Bignami et al. 2013b). To more clearly reveal daily growth increments, a transverse section of each otolith will be made by sanding and polishing the otolith on both sides of the core (D'Alessandro et al. 2010). Images of sectioned otoliths will be captured under light microscopy and daily increment widths measured digitally (Bignami et al. 2013b). Daily increment widths and total otolith growth throughout the experiment will be compared between treatments in conjunction with change in fish standard length. Measurement and comparison of daily increments may only be possible with juvenile fish otoliths because it is often difficult to discern daily increments in adult fish otoliths. However, total adult otolith growth during treatment exposure will be determined by using the alizarin red reference mark.

STATISTICAL ANALYSIS

Statistical comparisons will be made between all combination of treatments within an individual experiment, and between elevated-pCO2 treatments in separate replicated experiments when there is no significant difference between the control treatments of those experiments. The normality of all data will be tested prior to statistical analysis using parametric or non-parametric methods. Tank effects will be tested among replicate treatment tanks and data will be pooled if no tank effect is detected (Hamilton et al. 2013). Blood chemistry data will be tested using standard analysis of variance (ANOVA), unless data are not normally distributed, in which case the non-parametric equivalent (Kruskal-Wallis test) will be used (Esbaugh et al. 2012). Behavioral preference for light/dark and use of the shelter object surroundings, calculated as the difference between time spent in each light/dark or near-object/away- from-object zone, will be tested using ANOVA or Kruskal-Wallis (Hamilton et al. 2013). Analysis of covariance (ANCOVA) will be used to test differences in otolith size or shape with standard length as the covariate, and to test total change in otolith size over the duration of the experiment (Bignami et al. 2013b). ANOVA will be used to test the difference between specific daily increment widths chosen a priori at 5 d intervals (Bignami et al. 2013b).

SIGNIFICANCE AND BROADER IMPACTS

INTELLECTUAL MERIT

The intellectual merit of this proposed research is that it will provide definitive insight into the response of fishes to variable OA. This may transform the current understanding of OA impacts on fishes and influence the direction and interpretation of future research. Although highly variable pCO2 has been reported in many environments (Hofmann et al. 2011), OA literature does not empirically address the implications of this variability on fishes. This proposal is premised by the argument that the current understanding of behavioral and otolith growth impacts in fishes may be misguided due to the use of constant OA conditions. This research will provide two important contributions to the collective body of research on OA and fishes: first, a novel description of changes in blood chemistry due to variable environmental pH/pCO2, which will fill a crucial knowledge gap for both physiology and OA disciplines.

13

Second, this study will investigate the impact of variable-pCO2 OA on previously reported behavioral and otolith growth impacts, relating the newly-described physiological characteristics to adult and juvenile response to OA. If the proposed alternative hypotheses are supported, and impacts are mitigated or eliminated under variable-pCO2 conditions, then the understanding of two of the most commonly reported impacts of OA on fishes will be greatly altered and the focus of future research will necessarily shift. If the alternative hypotheses are not supported, then the proposed research will provide additional insight into the current understanding of OA effects on fishes. Either scenario will ultimately help to increase the limited foundation of knowledge that now exists.

The data produced by this research will also inform research in other disciplines. By characterizing the physiological, behavioral, and otolith growth response to variable pCO2, this research will enable other scientists to incorporate this novel information into their research design and interpretation of results. This will be especially applicable in circumstances where fishes reside in environmentally variable habitats. In these habitats, fishes may already be impacted by present day high-pCO2, whether through behavior or the many inseparable physiological and energetic connections within an organism. Furthermore, the results of the proposed research will also be applicable to research on the biology, ecology, population dynamics, and management of fishes, all of which could be influenced by variable-pCO2 even under present-day conditions. This combination of influence on research in OA and non-OA disciplines will allow the proposed research to directly and indirectly contribute crucial new information to the current understanding of the biological function of fishes. It will also inform the understanding of broader ecosystem dynamics involving fishes in present-day and future marine environments, and provide novel and critical insight into the consequences of ocean acidification on ecosystem health and function.

BROADER IMPACTS ON SOCIETY

As part of the broader impacts on society, this Research in Undergraduate Institutions (RUI) proposal aims to address two NSF core strategies, as described in the NSF Strategic Plan (FY 2011-2016): the seamless integration of research and education and broadening opportunities and expanding participation of groups and institutions that are underrepresented in science, technology, engineering, and mathematics (STEM) disciplines. This will be achieved through five avenues: (1) development of a marine science curriculum for undergraduates at CUI, (2) expansion of CUI undergraduate participation in original research, (3) provision of in-depth marine science education for local primary and secondary school teachers, (4) incorporation of an educational outreach science camp to children from a local homeless shelter, and (5) continued active participation by PI Bignami in effective communication and broader marine science educational outreach. As a RUI proposal, the first and second avenues are outlined in greater detail in the RUI Impact Statement attached as supplementary material, but are described here briefly.

PI Bignami has been tasked by CUI to develop and implement a marine science educational curriculum within the School of Arts and Sciences, which CUI administration has committed to support logistically and financially. The infrastructure that will be produced as part of this proposed research will be essential to achieving this undergraduate educational goal, because much of it will be used for educational purposes when experiments are not actively occurring. In the small class setting at CUI, close interaction between PI Bignami and undergraduate students will ensure that each student receives the individualized attention necessary to develop an in-depth understanding of course material and the ability to use critical thinking to solve problems. Also, to better access students who are not actively enrolled in marine science courses, public marine science seminars will be periodically offered to provide all students and faculty with exposure to visiting researchers and faculty. The new CUI marine science curriculum will directly expand the participation of “small liberal arts institutions” in STEM research and education.

As a RUI proposal, one goal of this proposal is to produce graduates who have the ability to conduct hypothesis-driven research and are experienced with the entire scientific process, from proposal preparation to publication. PI Bignmi will actively recruit undergraduate students to participate in the

14

proposed research throughout each academic year, with one or two 10-week undergraduate research internships offered during each summer (depending on the year). Summer interns will conduct the proposed research and individual research that will be presented at the President’s Academic Showcase of Undergraduate Research, an annual research showcase and competition at CUI. This showcase requires students to produce a publication-quality research paper, a poster presentation of their research, and for finalists, an oral presentation. Undergraduate interns will also present their research to the broader scientific community at national symposia and through publication in peer-reviewed journals.

To expand the impact of this research to a broader student audience, local primary and secondary school teachers will be provided an opportunity to participate in a “teach-the-teachers” workshop. Additionally, marine science topics will be incorporated into the educational curriculum at the CUI School of Education, to better prepare teachers to be science literate prior to becoming certified. These projects will be achieved through the collaborative efforts of PI Bignami and Dr. Michael Schulteis, Director of Undergraduate Teacher Education Programs and Professor of Science Education at the CUI School of Education. The teach-the-teachers workshop will invite 10 local primary and secondary school science teachers to CUI for a 3-d program. Teachers will receive a daily stipend for attendance and professional development credit will be arranged, as applicable, through their respective school districts. The workshop will have two primary goals; to improve teacher understanding of ocean acidification and general marine science, and to develop specific inquiry-based lesson plans for implementation in their classes. Similar goals apply to the incorporation of marine science education into the curriculum at the CUI School of Education. By improving the ability of local teachers to include marine science education in their curricula, many more students will be influenced than would be possible through any individual effort of PI Bignami.

Educational outreach serving youth that are underrepresented in the STEM disciplines will be accomplished through collaborative work with a local long-term family homeless shelter, Village of Hope (VOH) (see letter of support). VOH has existing outreach programs coordinated with CUI and the majority of residents are of Latino heritage. Approximately 15-20 elementary and middle school students will be offered the opportunity to participate in a 5-d marine science summer camp at CUI, staffed by PI Bignami, undergraduate summer research interns, and other undergraduate and faculty volunteers. The science camp will focus on three main areas; introduction to general concepts of marine biology and oceanography, field experience with local marine ecosystems, and participation in an inquiry-based experiment using the RAS, laboratory equipment, or field resources. The students and PI Bignami will give a presentation of the science camp experience to the parents and other families at VOH during a follow-up event.

Additional broader impact of this proposed research will be accomplished through the existing outreach activity and established science communication network of PI Bignami. PI Bignami has been continuously involved in education and communication outreach since before he served as an NSF GK-12 fellow during graduate school. Past outreach activities have ranged from small group lessons to large- scale events, such as an educational presentation to over 10,000 middle school students at the Miami Marlins baseball stadium in 2013. Recently, PI Bignami participated at a National Oceanic and Atmospheric Administration (NOAA) OA communication workshop (2012), organized with the goal of developing a NOAA Guide to Best Practices of Ocean Acidification Communication. As a follow-up to the workshop, PI Bignami currently serves as a working group leader for continued NOAA-coordinated development of effective communication to citizens and policy makers. PI Bignami has also developed OA education kits for NOAA offices along the west coast and served as an expert in charge of marine biology question development for a national trivia competition for high school students, National Ocean Sciences Bowl. Overall, PI Bignami’s prior experience and established outreach networks will provide ample opportunity to broadly disseminate data to a diverse audience.

Each of these broader impact activities will be assessed to determine their success and the amount of short term and long term influence on each target group. Undergraduate educational success will be

15

assessed by evaluating enrollment in marine science courses and participation of non-marine science students at public seminars. The successful improvement of undergraduate research experiences at CUI will be assessed by measuring the increase in marine science entries in the President’s Academic Showcase, as well as the frequency of undergraduate research presentations at symposia and publication in peer-reviewed journals. Effectiveness of the teach-the-teachers workshop will be assessed by conducting before and after surveys of the participating teachers’ understanding of marine science and utilization of marine science class activities. Similarly, education graduate students who receive marine science training will be surveyed following graduation and employment, to assess the influence of the enhanced marine science training on their ability to incorporate marine science into their curricula. VOH residents will receive before and after surveys to assess the short-term improvement in marine science knowledge and familiarity with local ecosystems. Long-term follow-up will also be conducted to track participation and success in science courses and careers during high school and college.