CALIFORNIA STATE UNIVERSITY, NORTHRIDGE

Foundation species loss alters ecosystem functioning within temperate tidepool communities

A thesis submitted in partial fulfillment of the requirements

For the degree of Master of Science in Biology

By

Jennifer Fields

December 2020

The thesis of Jennifer B. Fields is approved:

______Dr. Kerry J. Nickols, Ph.D. Date

______Dr. Steve R. Dudgeon, Ph.D. Date

______Dr. Nyssa J. Silbiger, Ph.D. (Chair) Date

California State University, Northridge

ii ACKNOWLEDGEMENTS

Science never happens in isolation. This Master’s is the product of endless hard work from a team of amazing and passionate people. My thesis advisor, Dr. Nyssa Silbiger, has been the mentor I could only dream of being myself one day. Without your connections with Oregon

Department of Fish and Wildlife (ODFW) Marine Reserve team, I would not be able to do the master’s project that I have been dreaming about since studying barnacles in undergrad. Thank you for your endless edits, letters of recommendation, and support as I pushed my computational abilities and gained confidence in R stats. You have encouraged all of us to be not only better scientists and science communicators, but also better humans that advocate to make this world a more inclusive and equitable space. The theme of #everythingisconnected and memories from the Silbiger Lab (e.g. impromptu dance parties) are something I will carry fondly forever.

Thank you to my thesis committee members, Dr. Kerry Nickols and Dr. Steve Dudgeon for their expertise and feedback on study design and statistical analyses. I have learned so much from you both and I always felt supported by you with my project and my scientific abilities. Dr.

Dudgeon, thank you for your insight into grazing experiments and for helping with the omnivore’s dilemma. Your knowledge of intertidal intricacies is always incredibly appreciated.

Dr. Nickols, thank you for always bringing me back to the bigger oceanographic and physical picture. You have challenged me to dive deeper into my data to understand more of “how”, and

“why” questions within tide pool systems. Thank you for being my inspiration to pursue coastal policy as my future career. I literally would not be a future CA Sea Grant State Fellow without your support.

I am extremely grateful for my labmates, Amanda Chiachi, Danielle Barnas, Danielle

Becker, Katie Hannibal, and Jamie Kerlin, for their advice, support, and, in non-COVID times,

iii hugs. I am grateful for my shared (Silbiger/Nickols) labmates, Deme Panos, Brian Cohn,

Taylorann Smith, and Emily Wilson, for always inspiring me, providing tea, and endless laughter. For all the undergraduate students in our lab present and past, Maria Martone, Emily

Rukstales, Natalie Dahan, Miranda Gilhuys, and Julio Rosales, it has been amazing to see you grow as rockstar scientists and thank for inspiring me with your passion. Thank you to my broader California State University, Northridge (CSUN) cohort, especially the CSUN trail run club (Sarah Wenner, Robert Hogg, Kiran Reed, Jack Girard, and Chelsea Brisson) and my housemates Jennica Moffat (feat. Hoodini), Chris Polinski, and Danielle Becker, whom I would not have gotten through the last 2.5 years without. Thank you to the Biology Department and stock room staff for always providing generous support and assistance to my laundry list of tasks.

Thanks to my incredible team of intertidal rockstar volunteers: A. Chiachi, M. Barrios, C.

Polinski, J. Moffat, S. McClarence, C. Garvey, S. Vogel, D. Zuk, A. Marino, O. Martin, M.

Hauer-Jensen, K. Bauer, C. Fields, D. Becker, M. Nguyen, A. Chin, M. Yamane, T. Miles, C.

Weiler, and H. Fulton-Bennett, especially during night water sampling. Special thank you to my core assistants M. Gilhuys, J. Rosales, and D. Barnas, whom learned along with me how to be flexible and manage a multifaceted field experiment. I could have never accomplished this without your enthusiasm and hard work. This project would not have come to life without the research, lab, and housing support from ODFW Marine Reserve staff, Oregon State Parks, and

Oregon State University’s Hatfield Marine Science Center.

Funding for this project was provided by CSUN Research and Graduate Studies, CSUN

Department of Biology, CSU Council on Ocean Affairs, Science & Technology Graduate

Research Grant, Association of Retired Faculty, Sigma Xi Grants-in-Aid of Research, and Dr.

iv Julie Gorchynski research grant. This project was conducted under ODFW Scientific

Taking Permit #22930 and Oregon Parks and Recreation Department Scientific

Take Permit #017-19 as well as California DFW permit #13047 and CA State Parks permit #18-

820-50 for preliminary testing.

Lastly, to my family and friends that responsible for the person I am today, thank you for your endless support of my dreams and for always inspiring me to pursue my passions.

v DEDICATION

I dedicate this thesis to literal metric ton of mussels I removed from the intertidal. I thank you for the endless wisdom I have gained from destroying thousands of critters’ home. I hope you made some hungry Pisaster very happy.

vi Table of Contents

Signature Page ii

Acknowledgements iii

Dedication vi

List of Tables viii

List of Figures ix

Abstract x

Introduction 1 Methods 5 Results 17 In-text tables and figures 19 Discussion 30

References 36

Appendix A: Tables 43

Appendix B: Figures 55

vii LIST OF TABLES

TABLE 1 43

TABLE 2 45

TABLE 3 46

TABLE 4 47

TABLE 5 48

TABLE 6 49

TABLE 7 50

TABLE 8 51

TABLE 9 52

TABLE 10 53

viii

LIST OF FIGURES

FIGURE 1 19

FIGURE 2 20

FIGURE 3 21

FIGURE 4 23

FIGURE 5 25

FIGURE 6 26

FIGURE 7 29

Appendix B FIGURE 1 55

Appendix B FIGURE 2 56

Appendix B FIGURE 3 57

Appendix B FIGURE 4 58

ix ABSTRACT

Foundation species loss alters ecosystem functioning within temperate tidepool communities

By

Jennifer Fields

Master of Science in Biology

Foundation species are organisms that create shelter, enhance biodiversity, and maintain ecosystem functioning within their environment. Within the rocky intertidal ecosystem, a coastal ecosystem that can be dominated by mussels and surfgrass, foundation species are expected to decline with climate change, extreme climatic events, and increased human impact. Although studies have investigated the impact of intertidal foundation species on community structure and some resource fluxes, there is a gap in knowledge to understand the magnitude of effects that intertidal foundation species loss will have on multiple facets of ecosystem functioning in concert using a mechanistic approach. I studied the direct and indirect effects of foundation species loss of mussels (Mytilus californianus) and surfgrass (Phyllospadix spp.) on community structure and resource fluxes via in situ tide pool manipulations. I measured multiple facets of ecosystem function including stocks (community structure), fluxes (light, temperature, dissolved oxygen, dissolved inorganic nutrients, pH), and ecosystem metabolism (net ecosystem calcification [NEC] and net ecosystem production [NEP]) in mussel- and surfgrass-dominated tide pools in central Oregon before and after removal of foundation species. I quantified the interdependency between changes in stocks, fluxes, and ecosystem metabolism along a gradient of foundation species loss using a structural equation model. Surfgrass and mussel loss caused significant changes in the sessile communities, increased the light and temperature environment, shifted biogeochemistry, and directly and indirectly increased pH and ecosystem metabolism. Shifts in both surfgrass and mussel sessile communities with foundation species loss were driven mostly by increasing diatom cover. Surfgrass loss increased daily maximum temperature by 7.7°C and percent light by 30%, whereas mussel loss increased temperature by 3.8°C, but not the light environment over the month-long study. Tide pools with foundation species present shifted in accordance with changing ocean biogeochemistry, while pools with foundation species loss had higher in dissolved oxygen, pH, and temperature variance values. Surfgrass loss had strong positive indirect effects of NEP, pH, and NEC mediated by the direct increase in light environment with percent surfgrass loss. Mussel loss indirectly increased NEP mediated by direct increases in micro/macroalgal cover with mussel loss and directly increased pH through the removal of the dominant consumer within the system. This study highlights the importance of measuring the ecosystem as a whole with a mechanistic approach to understand the magnitude of changes that will occur with intertidal foundation species loss with increasing anthropogenic impacts.

x Introduction

Foundation species are a major focus of ecological studies due to their importance to ecosystem processes and conservation management (Byers et al. 2006, Halpern et al. 2007,

Angelini et al. 2011, Ellison 2019). As the dominant, habitat-building organism within biological communities, foundation species facilitate interactions that promote biodiversity, stabilize community structure, and maintain fluxes (e.g. nutrient cycling, carbon cycling, or production rates) of ecosystem functioning (Bruno and Bertness 2001, Kinzig et al. 2001, Angelini et al.

2011, Ellison 2019). Recent studies have emphasized the connection between foundation species and ecological services, such as shoreline protection, nutrient cycling, and food security, as the direct consequences of foundation species loss (Smale et al. 2019). It is essential to understand both the direct (loss of habitat) and indirect (altered fluxes and ecosystem functioning) impacts of foundation species loss for the conservation management of nearshore marine habitats that provide over 43% of the global total of ecosystem goods and services per year (Costanza et al.

1997, Ellison 2019).

In coastal marine ecosystems worldwide, anthropogenic climate change and acute temperature anomalies have caused latitudinal shifts (Sagarin et al. 1999, Angelini et al. 2011) and losses (Alongi 2002, Orth et al. 2006, Smith et al. 2006, Dudgeon et al. 2010, Byrnes et al.

2011, Sorte et al. 2017) of foundation species in marine coastal ecosystems worldwide.

Foundation species such as seagrasses, mangroves, corals, kelps, and mussels have decreased dramatically in abundance over the last several decades (Alongi 2002, Orth et al. 2006, Smith et al. 2006, Dudgeon et al. 2010, Byrnes et al. 2011, Krumhansl et al. 2016, Sorte et al. 2017). For example, kelp beds along the West coast of Australia decreased ~ 2,300 km2 and experienced a

100-kilmometer range contraction due to ocean warming and a marine heat wave event

1 (Wernberg et al. 2016). Further, coral cover on the Great Barrier Reef declined ~40-90% due to a single marine heat wave event in 2016 (Hughes et al. 2018). Foundation species declines have already led to significant and documented losses in ecosystem functioning, (Pandolfi et al. 2009,

Waycott et al. 2009, Carranza et al. 2011, Krumhansl et al. 2016, Hughes et al. 2018, Edwards et al. 2020), and foundation species will be continue to be disproportionately removed from systems via anthropogenic disturbance (Dayton 1972, Ellison et al. 2005, Castorani et al. 2018).

One ecosystem with significant and recently documented depletion of multiple foundation species is the rocky intertidal, a coastal ecosystem that can be dominated by the foundation species of mussels (e.g., Mytilus spp.) and seagrasses (e.g., Phyllospadix spp.).

Mytilus californianus cover and biomass decreased by 40.2% and 51.3%, respectively, within southern California since the mid-1970s to 1980s (Smith et al. 2006), and localized massive die- offs have occurred during recent terrestrial heat wave events (Stark 2019). Seagrasses are one of the most rapidly declining marine foundation species in the world and have decreased within near-shore and intertidal environments by 58% worldwide due to habitat loss, eutrophication, and climate change (Orth et al. 2006, Waycott et al. 2009, Honig et al. 2017). Temperate seagrass habitats have already experienced substantial (>90%) localized diebacks related to marine heat wave events (Thomson et al. 2015), which are increasing with anthropogenic climate change (Laufkötter et al. 2020). In the northeastern Pacific, a dominant rocky intertidal seagrass called surfgrass (Phyllospadix spp.) is particularly susceptible to desiccation and heat stress during low midday tides, anthropogenic development, and nutrient addition (Littler and Murray

1975, Raimondi et al. 1999). Both mussels and surfgrasses will continue to disappear as a result of increasing anthropogenic impacts (Smith and Murray 2005, Orth et al. 2006, Werfhorst and

Pearse 2007, Wootton et al. 2008, Waycott et al. 2009, Honig et al. 2017, Sorte et al. 2017),

2 which will cause extensive changes in ecosystem functioning and threaten the ecological services that these foundation species help provide (Christensen et al. 1996, Angelini et al. 2011, Smale et al. 2019).

Mussels and surfgrasses not only maintain ecosystem functioning by creating habitat for mobile and sessile flora and fauna (Stewart and Myers 1980, Crouch 1991, Suchanek 1992,

Bracken and Nielsen 2004, Shelton 2010), but also by regulating fluxes of the ecosystem.

Specifically, M. californianus recycles nutrients and promotes primary productivity through increased nutrient availability (Bracken and Nielsen 2004, Pfister 2007, Pfister et al. 2019), decreases pH and dissolved oxygen (DO) through respiration (Silbiger and Sorte 2018,

Ninokawa et al. 2019), decreases temperature within mussel beds (Stephens and Bertness 1991), and maintains net ecosystem calcification (NEC) as a dominant calcifier in the intertidal

(Silbiger and Sorte 2018). Phyllospadix spp., a primary producer, takes up nutrients (Terrados and Williams 1997), modulates pH and DO over diel cycles (Duarte and Chiscano 1999, Silbiger and Sorte 2018), decreases temperature (Shelton 2010), and increases net ecosystem production

(NEP), which increases pH, and makes an environment more favorable for calcification to occur

(Bracken et al. 2018, Silbiger and Sorte 2018). The loss in Mytilus or Phyllospadix species could have both direct (loss of food and habitat) and indirect (altered biogeochemistry and thermal environment) effects on ecosystem functioning (Shelton 2010, Sorte and Bracken 2015, Bracken et al. 2018, Silbiger and Sorte 2018).

While removal of Mytilus californianus and Phyllospadix spp. (from here on as mussel and surfgrass) has been investigated on individual fluxes or community metrics (Pfister 2007,

Shelton 2010), the relative importance of foundation species loss on multiple fluxes of ecosystem function in concert remains unknown. To address this gap in knowledge, I used in situ tide pool

3 manipulations to investigate how intertidal foundation species removal affects ecosystem function via changes in community structure and resource fluxes one month after foundation species removal. Specifically, I examined the direct and indirect effects of mussel and surfgrass foundation species removal from tide pools on (1) community composition of sessile and mobile species, (2) thermal and light environments, (3) local biogeochemistry, and (4) ecosystem metabolism (NEC and NEP) using general linear models and a structural equation model (SEM) one month after foundation species removal. I hypothesize that (1) community structure will shift towards more inferior space competitors, (2) temperature and light will increase, (3) biogeochemistry will shift in accordance with community structure changes, and (4) indirect effects of foundation species loss on ecosystem metabolism (NEP and NEC) via altered local biogeochemistry as a result of shifted community composition will be greater than direct effects. By addressing both direct and indirect effects using a causal approach, I broaden the mechanistic understanding of how tide pool communities will respond to mussel and surfgrass loss, including the importance of these foundation species’ role in ecosystem functioning.

4 Methods

Study Site

I selected 32 tide pools within the Otter Rock Marine Reserve (ORMR), Oregon USA

(44°45'09.1"N 124°03'58.0"W) during the months of June–August 2019. At the site, one mussel species (M. califorinanus) and two surfgrass species (P. scouleri and P. torreyi) were present. Of the 32 tide pools, 16 were dominated by the California mussel M. califorinanus) and 16 were dominated by surfgrass (Phyllospadix spp.; 10 pools contained Phyllospadix scouleri and six pools had both Phyllospadix scouleri and Phyllospadix torreyi). Percent cover of foundation species ranged from 45.3–98.9% in the mussel pools and 49.5–100% in the surfgrass tide pools.

Some tide pools had both mussels and surfgrasses; however, the presence of second foundation species did not exceed 7.2% of the remaining tide pool cover. Tide pools were located in the mid to low intertidal zone ranging from 0.71 m to 1.77 m above mean lower-low water. At this tide height, tide pools were isolated for 4-6 hours during summer low tides.

Experimental Design and Tide Pool Manipulation

A paired Before-After-Control-Impact (BACI) design was used to account for changes in ocean chemistry, timing of low tide, and variability within tide pools throughout the experimental period. The BACI design consisted of two time periods (before and after removal) with control and removal tide pools (Stewart-Oaten et al. 1986), where foundation species were removed from removal tide pools (n = 8 per foundation species) between the before and after removal periods. The before-removal period occurred June to mid-July 2019 and the after- removal period occurred mid-July to August 2019. Tide pools were selected for control or removal groups by using a random number generator in R (v3.2.0) statistical program software

(R Core Team 2018). Tide pools within removal and control groups were kept to ± 5.8% of the

5 ratio of surface area to percent cover of foundation species (Appendix A Table 1). Removal of foundation species occurred mid-July 2019 and tide pools had a one-week recovery period before any surveys or samples were taken. Foundation species were removed from tide pools using a garden hoe (Boese et al. 2009). Within surfgrass removal tide pools, rhizomes were also removed due to their ability to alter nutrient cycling (Terrados and Williams 1997). During each time period, I characterized physical parameters (e.g. tide height or pool volume), community composition, and biogeochemical fluxes (e.g. dissolved oxygen, pH, nutrients) in each tide pool.

Tide Pool Characterization

Because tide pools vary in physical properties that can affect ecosystem function, I measured multiple physical characteristics of each tide pool and included them as covariates in statistical analyses. Physical parameters included tidal height (location within intertidal) and ratio of bottom surface area to volume (size of pool). Tide heights for each pool were surveyed with a laser level and stadia rod (DeWalt, Towson, MD, USA). Surface area (SA) of the pool was measured by counting the number of 10 x 10 cm squares within the pool’s perimeter, which was marked with demarcated metal chain along the edge of the water. Tide pool volume (V) was determined using a dye method (Pfister 1995) and measured with a SmartSpec3000 spectrophotometer (Bio-Rad Lab, Hercules, CA, USA). Since removal of foundation species altered the initial surface area and volume measurements, these parameters were re-measured post-removal. The average values in surface area to volume ratio (SA:V) and tide height between both time periods were used in statistical analyses.

Community changes as a function of foundation species loss

I conducted two rounds of community composition surveys for sessile percent cover and mobile organism counts: before removal and one-month post-removal. I temporarily removed

6 seawater from the tide pool and placed a flexible mesh quadrat with 10 x 10 cm squares into each pool. I measured percent cover for sessile organisms and counted mobile organisms, identifying down to the lowest possible taxonomic unit in the field (usually genus level; Appendix A Table

2.1 & 2.2). I normalized non-foundation species sessile cover to be relative to 100% cover per tide pool. All tide pool characterization and community composition methods were completed at least 24 hours before any water sampling event to allow the pools to be flushed at least twice by the ocean before measurements.

Direct effect of foundation species loss on temperature and light environment

Temperature (°C) and light intensity were recorded every 15 minutes for one month during each time period using HOBO® Pendant loggers bolted facing up on the flattest part of the tide pool on a level platform in the interstitial spaces of the foundation species (Onset

HOBO Pendent Light Intensity Data Logger MX2202 , Bourne, MA, USA). Light intensity

(lumens m-2) was converted to photon flux density (PFD; µmol photons m-2 s-1) following Long et al. (2012). Change in maximum temperature and percent change in maximum light were summarized for each time period (n = 30 days each) for statistical analysis. For the causal model, average PFD and maximum temperature were extracted from the logger data for the specific dates and times of water collection for comparison with biogeochemistry and ecosystem metabolism measurements.

Effects of foundation species loss on biogeochemistry and ecosystem metabolism

To determine biogeochemistry fluxes and ecosystem metabolism (NEC and NEP) before and after removal of foundation species, I collected daytime and nighttime water samples during low tide. I used a block design for water sampling with two daytime and two nighttime sampling events, where n= 16 pools (n=8 per foundation species) were on separate day and night sampling

7 events (Appendix A Table 3−4). Each sampling event had equal number of pools per foundation species and treatment group (removal or control: n = 8 pools). In situ temperature, DO, salinity,

+ − − 3 − pH, and discrete samples for dissolved inorganic nutrients (NH4 , NO2 + NO3 , PO4 ) were collected hourly over a four-hour period in each pool and the adjacent ocean following methods by Silbiger & Sorte (2018). Temperature, DO, and salinity were measured with a calibrated multi-parameter pro meter directly in each pool (YSI Pro 2030 Lot #18B100763, Yellowsprings,

OH, USA). For pH, nutrients, and TA, I collected 400 mL discrete water samples into a sealed

Erlenmeyer flask using a vacuum hand pump (Mityvac, St. Louis, MO, USA) from the deepest part of the pool. Discrete samples (~250mL) for total alkalinity (TA) were taken at the first and last time point (hour 1 and hour 4) to calculate an integrated change in TA over the low tide period (similar to Bracken et al. 2018). To compare tide pool conditions to the open ocean, ocean measurements of all parameters were taken from the same general location at the surface adjacent to the site.

pHT was measured within one hour of water collection from the sealed Erlenmeyer flask using an Orion Star Multiparameter Meter with a ROSS Ultra glass electrode (Thermo Scientific,

USA, accuracy = ±0.2 mV, resolution = ±0.1, drift < 0.005 pH units per day) and a traceable digital thermometer (5-077-8, accuracy = 0.05 °C, resolution = 0.001 °C; Control Company,

Friendswood, TX, USA) following Dickson SOP 6 (Dickson et al. 2007). The glass electrode measured millivolts (mV) and was calibrated within 48 h of each sampling event using a multipoint calibration to a tris standard solution from the Dickson Lab at Scripps Institution of

Oceanography following Dickson SOP 6a (Dickson et al. 2007). TA seawater samples were placed in 250 mL Nalgene bottles with 100 µl of 50% saturated HgCl2 to preserve the water within five hours of collection. Seawater samples for nutrient analysis were filtered through

8 GF/F filters (0.7µm) with a syringe into designated 50 ml centrifuge tubes and the samples were frozen within five hours of collection. All sampling and storage containers were soaked in 10%

HCl for 24 hours, rinsed with MilliQ water, and rinsed three times with sample water before sampling events.

Sample Processing

In situ pH was calculated using the seacarb package in R (Gattuso et al. 2018) by correcting for the in situ temperature in each tide pool from a multi-parameter pro meter. Total alkalinity seawater samples were processed using open-cell potentiometric titrations on a

Mettler-Toledo T5 auto-titrator (Columbus, OH, USA) following Dickson SOP 3b (Dickson et al. 2007). A certified reference material (CRM) from the Dickson Lab at the Scripps Institution of Oceanography was used at the beginning of each sample group run. The accuracy of the CRM never exceeded ± 0.79% (precision = 5 µmol kg-1) from the standard value and TA samples were corrected for deviations from the standard value. TA was normalized to a constant salinity of 34 units to account for changes in evaporation during sample processing. Dissolved inorganic

+ − − 3 − nutrients (NH4 , NO2 + NO3 , PO4 ) were analyzed at Moss Landing Marine Laboratory using

+ − − a Lachat Quickchem 8000 Flow Injection Analyzer (± 1.21% NH4 , ± 0.26% NO2 + NO3 , ±

3 − 3.57% PO4 instrument precision; Hach, Loveland, CO, USA). After processing nutrients, one mussel control pool (Pool ID 30) was removed from all resource flux analyses due to abnormally high values of ammonium compared to other pools (4290 ± 188.1 µmol g-1), likely due to contamination, on one sampling day leaving n = 7 control pools for mussels from analyses, excluding community composition.

Calculations of NEC and NEP

9 I used the total alkalinity anomaly technique (Chisholm and Gattuso 1991) to calculate

−2 −1 net ecosystem calcification (mmol CaCO3 m hr ) with the following equation:

∆푇퐴 ∗ 휌 ∗ 푉 푁퐸퐶 = 2 ∗ 푆퐴 ∗ 푡

Total alkalinity values were divided by 1000 to convert from µmol kg-1 to mmol kg−1. ΔTA/2 is the salinity-normalized TA (mmol kg−1) between the first and last time points (1 and 4) divided by 2, where one mole of CaCO3 is formed per two moles of TA; ρ is the density of seawater

(1023 kg m−3); V is the volume of water in the pool at each time point (m3); SA is the bottom surface area of the tide pool (m2); and t is the time between sampling points (h). Samples were corrected for the effect of nutrients on carbonate chemistry with the following equation (Wolf-

Gladrow et al. 2007):

− − 3 − + Nutrient-corrected TA = TA – [NO2 + NO3 ] − (2 × [PO4 ]) + [NH4 ].

Positive NEC values indicate net calcification while negative values indicate net dissolution.

Net ecosystem production rates (mmol C m-2 hr-1) were calculated from differences in dissolved inorganic carbon (DIC), calculated from pHT and TA (error propagation = 7.09 ± 0.06 mmol kg -1) with the seacarb package in R (Gattuso et al. 2019), using the following equation

(Gattuso et al. 1999):

∆퐷퐼퐶∗ 휌∗푉 푁퐸푃 = − 푁퐸퐶 − 퐹퐶푂 . 푆퐴∗푡 2

ΔDIC is the difference in salinity-normalized DIC (mmol kg−1) between the first and last time points. NEC is subtracted to account for changes in DIC by the precipitation or dissolution of

−2 −1 CaCO3, and FCO2 (mmol m hr ) is the air-sea flux of CO2, which was subtracted to account for the flux in CO2 from the air-sea exchange. FCO2 was calculated as:

퐹퐶푂2 = 푘 ∗ 푠(퐶푂2−푤푎푡푒푟 − 퐶푂2−푎푖푟 ),

10 where k is the gas transfer velocity (m h−1) calculated from wind speed (Ho et al. 2006) using the closest weather station around 10 miles south of Otter Rock (NOAA Station NWPO3;

44°36'46.8"N 124°04'01.2"W); s is the solubility of CO2 in seawater calculated from in situ temperature and salinity (Weiss 1974); CO2 (µatm) in water is calculated from pH and TA values; CO2 in air was 410 µatm based on concurrent measurements at the Mauna Loa

Observatory (Tans & Keeling 2019). Positive NEP values indicate net photosynthesis and negative NEP values indicate net respiration.

Statistical Analysis

I conducted separate statistical analyses for surfgrass and mussel tide pools for all response variables due to the difference in effects that each foundation species has on community composition, biogeochemistry, and ecosystem metabolism. Change in foundation species cover between the before and after removal periods varied continuously across treatment groups with

49.5–100% cover in the surfgrass tide pools and 45.3-98.9% cover in the mussel pools

(Appendix B Figure 1); therefore, I investigated foundation species loss as a gradient rather than using an ANOVA design typical for BACI studies (Smith 2002).

Multivariate plots and analyses were used assess how community composition and biogeochemistry parameters changed between time periods. I used non-metric multidimensional scale plots with function “metaMDS” in the vegan package (Oksanen et al. 2007) to visualize how the gradient of foundation species loss altered sessile organism cover and mobile count communities. Both sessile and mobile community data was fourth-root transformed to reduce the effect of rare species. A PERMANOVA with the “adonis” function (Oksanen et al. 2007) was used to test if foundation species loss altered community composition, with tide height and tide pool size included as covariates. For the biogeochemistry data, principal component analyses

11 (PCA) were used to reduce the dimensionality within the data. Specifically, shifts between the before and after-removal period in biogeochemistry (DO, nutrients, and pH) and temperature mean, maximum, and variance of n = 8 (four day and four night) time points were visualized to observe differences between local tide pool conditions and the adjacent ocean.

I ran general linear models to test the effect of foundation species loss on change (after – before removal periods) in species richness, light, and temperature in the tide pools, with tide height and tide pool size (SA:V) as covariates. For light and temperature, I used the change in maximum temperature and percent change in light between before and after removal periods for the analyses. One outlier in the light model (light value + 3 S.D. away from the mean) within mussel pools was removed to meet model assumptions. I used the ggeffects package (Lüdecke

2018) to obtain the marginal effects of foundation species loss while accounting for the covariates (tide height and tide pool size) of the model. Model assumptions for all general linear models were verified by investigating plots of residuals of the model for homogeneity of variance and normality using the car package (Fox et al. 2007) and data that did not meet assumptions were transformed. Multi-collinearity was tested with Pearson correlations between each predictor using the GGally package (Schloerke et al. 2018). All data was processed and analyzed using R statistical program software (v 4.0.2; R Core Team 2018).

I conducted a piecewise structural equation model (SEM) using the R package

PiecewiseSEM 2.1.0 (Lefcheck 2016) to understand how mussel and surfgrass removal and physical characteristics of tide pools affected community changes, physical environmental conditions, biogeochemistry, and ecosystem metabolism of tide pool communities. Piecewise

SEMs combine multiple linear models into a single causal framework (Shipley 2009, Lefcheck

2016). Because the model is pieced together with local estimation, it is more flexible than

12 traditional SEMs and allows for inclusion of non-normal distributions, random effects, nested models, and smaller sample size within the causal model network (Lefcheck and Duffy 2015).

Specifically, I tested the indirect and direct pathways of surfgrass and mussel removal on ecosystem metabolism (NEC and NEP) and pH mediated by changes in micro/macroalgae cover, light, temperature, nutrients (dissolved nitrogen to phosphate ratio), and physical parameters of the tide pool. The fit of each model was determined with a Shipley’s test of d-separation, which determines there are no missing inferences from the model using a Fisher’s C statistic, where p>0.05 indicates that the model is a good fit for the data (Shipley 2009). Path coefficients were standardized and centered, allowing for comparisons of the magnitude of effect among groups.

For all model components, I averaged the daytime and nighttime measurements to obtain the net effect of the system, except for light measurements which were only daytime values, and then calculated the difference between the before and after removal time periods (n = 16 per parameter for surfgrass model; n =15 per parameter for mussel model), such that positive values indicate an increase, zero indicates no change, and negative indicates a decrease in that parameter after the removal period. For the community composition components of the model, I used the sessile community metrics that changed the most between the before and after removal period, including, surfgrass, mussel, and non-surfgrass fleshy producer (all micro and macroalgae not including coralline algal species) cover. Surfgrass loss, mussel loss, and producer cover were determined by the change in percent cover between before and after period. Tide pool size, represented as SA to V ratio (SA:V), and tide pool height were used as a covariate in all models because physical parameters affect temperature, light, and biogeochemistry (Legrand et al. 2018, Wolfe et al. 2020).

13 Due to the collinearity between nutrient species, the dissolved nitrogen to phosphate ratio

([nitrite + nitrate + ammonium]:phosphate) was used as the measurement for nutrients in SEM model. Because temperature and light were highly correlated due to underlying physical factors and because this relationship is not causal, light and temperature were included as correlated errors within both SEMs (Appendix B Figure 2). I used existing knowledge of intertidal systems to create hypothesized paths within each of the models (Grace 2008, Grace et al. 2012). The series of equations and their hypotheses are listed below with surfgrass model on the top and mussel model on the bottom, if there were different models for each foundation species:

(1) Micro/macroalgal cover ~ Foundation species loss + SA:V + Tide height

Model 1 represents the hypothesis that loss of surfgrass and mussel cover would increase producer cover by increasing free space for fast colonizers, with the size of the pool and tide height as a covariates (Dethier 1981, 1984).

(2) Light ~ Foundation species loss + SA:V + Tide height;

Model 2 represents the hypothesis that the light environment (measured as daytime average photon flux density; µmol photons m-2 s-1) will increase with foundation species loss, with size of the pool and tide height as covariates (Suchanek 1979, Stewart and Myers 1980).

(3) Maximum Temperature ~ Foundation species loss + SA:V + Tide height

Model 3 tests the understanding that temperature will increase with surfgrass and mussel loss

(Stephens and Bertness 1991, Shelton 2010), with size of the pool and tide height as a covariates.

(4) N:P ~ Foundation species loss + SA:V + Tide height

Model 4 represents the hypothesis that the N:P ratio will increase with surfgrass loss because the loss of dominant producers will lead to less nitrogen uptake (Terrados and Williams 1997,

Ramirez-Garcia et al. 2002). I hypothesize that the N:P ratio will decrease with mussel loss due

14 to the mussels’ ability to recycle nitrogen (Nielsen 2003, Bracken and Nielsen 2004, Pfister

2007, Pfister and Altabet 2019). N:P ratio may be affected by the size of tide pool due to diffusive properties of tide pools (Hurd 2000) and tide height due to emersion time.

(5) NEP~ Light + Micro/Macroalgal Cover + N:P+ SA:V+ Tide Height;

NEP ~ Maximum Temperature + Micro/Macroalgal Cover + N:P + SA:V+ Tide Height

Model 5 represents the different mechanisms that influence NEP. NEP is increased by light, producer cover, temperature, and nutrients (N:P) and affected by the size of the pool and tide height (Pfister 2007, Kwiatkowski et al. 2016, Takeshita et al. 2016, Silbiger and Sorte 2018,

Pfister and Altabet 2019, Wolfe et al. 2020). Surfgrass and mussel loss were both correlated with most of the components of the NEP model and were therefore excluded from the models.

Because light and temperature were highly correlated, only one could be included in the NEP models. I determined whether light or temperature had the lower AIC value in each model. Light was included in the NEP model for surfgrass and temperature in the NEP model for mussel

(Appendix A Table 5), although both models had a ΔAIC of < 1 meaning the two models are indistinguishable from each other.

(6) pH ~ NEP + Foundation species loss + SA:V + Tide height

Model 6 represents our understanding that there is a positive relationship between NEP and pH

(Silbiger and Sorte 2018). Production increases pH, while respiration decreases pH due to the uptake and release of CO2 respectively. Foundation species loss, size of the pool, and tide height may also affect pH.

(7) NEC ~ pH + Maximum temperature + Foundation species loss + SA:V + Tide Height

15 Model 7 represents our understanding that NEC increases with pH and temperature

(Kwiatkowski et al. 2016, Silbiger and Sorte 2018, Wolfe et al. 2020) and may be affected by the size of the pool, tide height, and foundation species loss.

(8) Correlated errors

Max Temperature %~~% Light

Light and temperature are correlated, but not causally linked through underlying physical parameters.

Data Availability Statement

All data and code used for this study are available at https://github.com/jenniferfields/EcoFunORTidepools

16 Results

Foundation species loss and physical parameters alter tide pool communities

Within surfgrass pools, surfgrass loss significantly affected sessile community composition (Fig. 1.1; PERMANOVA: F1,31=15.35, p=0.001) but not mobile communities (Fig.

1.3; PERMANOVA: F1,31=1.10, p=0.36). Shifts in sessile community with surfgrass loss were driven particularly by diatoms, algae film, Ptilota spp., bleached crustose coralline algae,

Analipus japonica, and Chaetomorpha linum cover (Fig. 1.3). In addition, size of the tide pool

(PERMANOVA: F1,31=5.34, p=0.004) and tide height (PERMANOVA: F1,31=2.71, p=0.03) significantly affected surfgrass sessile communities. Mobile communities were significantly affected by both size of tide pool (PERMANOVA: F1,31=4.93, p=0.001) and tide height

(PERMANOVA: F1,31=4.73, p=0.001), but not surfgrass loss. Species richness for sessile or mobile communities was not affected by surfgrass loss, size of tide pool, nor tide height (Fig. 3;

Appendix A Table 6).

Mussel loss significantly affected sessile community composition (PERMANOVA:

F1,31=8.84, p=0.001; Figure 2.1), driven mostly by changes in diatom cover. Mobile communities did not significantly change as a function of mussel loss (PERMANOVA: F1,31=1.73, p=0.07).

Size of tide pool also significantly affected sessile community composition in mussel pools

(PERMANOVA: F1,31=3.81, p=0.003). In addition, sessile species richness decreased 0.96 species (± 1.01 species SE) along the gradient of mussel percent loss (Fig. 3.2 ; t3,12=-3.97, p=0.002), but was not affected by size of pool (t3,12=1.29, p=0.22) or tide height (t3,12=-1.84, p=0.09). Mussel mobile communities were not significantly affected by mussel loss, tide pool, or tide pool size (Fig. 2.3; Appendix A Table 7). Mobile species richness significantly increased

17 with tide height (t3,12=4.36, p<0.001), but was not affected by mussel loss (t3,12=-0.82, p=0.43) or size of the tide pool (t3,12=1.32, p=0.21).

18 FIGURE. 1. nMDS plot (1.1& 1.3) and species biplot (1.2 & 1.4) of (1.1 & 1.2) on quad-rooted sessile and (1.3 & 1.4) mobile organism community composition for surfgrass pools before and one-month after removal of surfgrass removal across n = 16 tide pools at Otter Rock, OR. Green points are colors and sized proportional to surfgrass percent loss (0 in before period, varying in after period). Enlarged triangles represent the centroids (median) of each grouping (control and foundation species removal pools) with arrows between before (filled shape) and after (empty shape) period to show the magnitude of change. Grey triangles indicate removal pools and blue triangles indicate control pools.

19 FIGURE. 2. nMDS (2.1& 2.3) and species biplot (2.2 & 2.4) of (2.1 & 2.2) on quad-rooted sessile and (2.3 & 2.4) mobile organism community composition for mussel pools before and one- month after removal of surfgrass removal across n = 16 tide pools at Otter Rock, OR. Blue points are colors and sized proportional to mussel percent loss (0 in before period, varying in after period). Enlarged triangles represent the centroids (median) of each grouping (control and foundation species removal pools) with arrows between before (filled shape) and after (empty shape) period to show the magnitude of change. Grey triangles indicate removal pools and blue triangles indicate control pools.

20

FIGURE. 3. Relationships between foundation species loss and change in (3.1 & 3.2) surfgrass sessile and mobile species richness and (3.3 & 3.4) mussel sessile and mobile species richness. Solid dots represent control pools, open dots represent removal pools. Solid lines with gray shaded 95% confidence intervals represent significant relationships.

21 Foundation species loss increased physical light and temperature environment

Over the two month-long study, temperature increased as a function of both surfgrass and mussel loss in the tide pools (Appendix A Table 8). Surfgrass loss had a strong positive effect on maximum temperature (Fig. 4.1), increasing by 0.13°C (± 0.03 °C SE) per percent loss of surfgrass (t3,12 = 4.46; p<0.001). Overall, daily maximum temperature was on average 7.7°C (±

1.5°C SE) warmer in the pools that had surfgrass removed relative to the control pools. Mussel loss also significantly increased max temperature in the tide pools by 0.05°C (± 0.02°C SE) per percent loss (Fig. 4.2; t3,12=2.64, p=0.002), with temperatures increasing by an average 3.8°C (±

1.5°C SE) compared to control pools. There was no effect of the tide pool physical parameters on maximum temperature over the month-long period in surfgrass (size of pool: t3,12 =0.87; p=0.40 and tide height: t3,12= -0.25; p=0.81) or mussel (size of pool: t3,12 =-0.916; p=0.38 and tide height: t3,12= 0.571; p=0.58) pool, reinforcing that the effect of biological removal of foundation species exceeded the effect of physical parameters of the tide pool one-month post removal.

There was a similar trend with percent change in light, where surfgrass loss increased percent max light by 0.78% (± 0.21% SE) per percent loss of surfgrass (Fig. 4.3; t3,12=3.73, p=0.003). Overall, average percent max light increased by almost 30% (± 15.75% SE) in removal compared to control pools. Mussel removal did not significantly increase percent

2 maximum light over the study period (Fig. 4.4; multiple regression: F3,10=2.33, r = 0.23, p=0.14;

Mussel loss: t3,10: 1.75, p=0.11).

22

FIGURE. 4. Relationships between foundation species loss and change in (4.1 & 4.2) max temperature and (4.3 & 4.4) percent change in max light per tide pool for (4.1, 4.3) surfgrass and (4.2,4.4) mussel pools. Solid dots represent control pools, open dots represent removal pools. Solid lines with gray shaded 95% confidence intervals represent significant relationships.

23 Shifts in tide pool and ocean biogeochemistry and temperature

Surfgrass control pools shifted in accordance to changing ocean biogeochemistry, whereas removal pools deviated from the shift in ocean biogeochemistry (Fig. 5). During the after-removal period, coastal upwelling led to both ocean samples and control surfgrass pools becoming higher in nitrate/nitrite maximum, mean, and variance and phosphate mean and maximum values (Fig. 5B) as well as overall decreases in temperature (Appendix A Table 4).

Pools with surfgrass loss in the after period became higher in dissolved oxygen and pH maximum, mean, and variance as well as temperature variance. Temperature mean and maximum values were greater in surfgrass pools in the before period.

In mussel pools, control pools also shifted in similar magnitude and direction as ocean samples in the after-removal period (Fig. 6). Removal pools with increased mussel loss increased in both dissolved oxygen and pH maximum, mean, and variance as well as temperature variance

(Fig. 6.2).

24

FIGURE. 5. (5.1) Principal component analysis of the biogeochemistry (pH, DO, and nutrients + − 3 − (NH4 , NO3 , PO4 ) and temperature mean, max, and variance of n=16 pools before and three- weeks after surfgrass removal. Green points represent PC scores of each pool, where the color and size are proportional to surfgrass percent loss (0 in before period, varying in after period). Enlarged triangles represent the centroids (mean) of each grouping (foundation species control and removal pools and ocean) with arrows between before (filled shape) and after (empty shape) period to show the magnitude of change. Grey triangles indicate removal pools and blue triangles indicate control pools. (5.2) Loadings of the PCA with strength of driving for each biogeochemical parameter. PC axes 1 and 2 explain 60.78% of the variation of the data.

25

FIGURE. 6. (6.1) Principal component analysis of the biogeochemistry (pH, DO, and nutrients + − 3 − (NH4 , NO3 , PO4 ) and temperature mean, max, and variance of n=16 pools before and three- weeks after surfgrass removal. Blue points represent PC scores of each pool, where the color and size are proportional to mussel percent loss (0 in before period, varying in after period). Enlarged triangles represent the centroids (mean) of each grouping(foundation species control and removal pools and ocean) with arrows between before (filled shape) and after (empty shape) period to show the magnitude of change. Grey triangles indicate removal pools and blue triangles indicate control pools. (6.2) Loadings of the PCA with strength of driving for each biogeochemical parameter. PC axes 1 and 2 explain 61.91% of the variation of the data.

26 Foundation species loss and tide pool physical parameters alter ecosystem fluxes

Using a SEM model, I demonstrate the relative magnitude of tide pool physical parameters (SA:V and tide height) and foundation species loss on the ecosystem metabolism

(NEP and NEC) mediated by changes in community structure (% cover producers), biogeochemistry (nutrients and pH), and physical environment (light and temperature) of tide pool ecosystems. Structure of the overall model was a good fit using Fisher’s C (p>0.05) for both surfgrass (C30=24.25 p=0.76, Fig. 7.1) and mussel (C30 =41.10, p = 0.085, Fig 7.2) SEM models.

In the surfgrass model, pH and NEC parameters explained the most variance within the model

(83% for pH and 81% for NEC). For the mussel model, pH (83%) and NEP (81%) explained the most variance (Appendix A Table 9).

Surfgrass loss directly increased average daytime light (PFD) values by 18.80 (± 7.43

SE) µmol photons m-2 s-1 per percent loss (p=0.02) and increased micro/macroalgal cover 0.69%

(± 0.28 SE) per percent loss (p=0.03). The change in light significantly increased NEP by 0.002

(± 0.001 SE) mmol C m-2 hr-1 per unit change in PFD (p=0.05), which then caused a significant

0.02 (± 0.01 SE) increase in pH (p=0.02) per unit increase in NEP. The strongest causal relationship in the model was the relationship between pH and NEC, where there was a 10.51 (±

−2 −1 2.53 SE) mmol CaCO3 m hr increase in NEC per pH unit (p=0.002). Maximum temperature

−2 −1 (°C) significantly increased NEC by 0.31 (± 0.10 SE) mmol CaCO3 m hr per change in 1°C (p

= 0.01). Size of the tide pool (SA:V) significantly affected pH and NEC, increasing pH by 0.007

−2 −1 (± 0.002 SE; p=0.01) and decreasing NEC by 0.09 (± 0.03 SE) mmol CaCO3 m hr per unit increase in SA:V (p=0.01). Tide height (m) only significantly negatively affected NEC (p=0.01),

−2 where a 1 m increase in tide height resulted in a decrease in 2.61 (± 0.87 SE) mmol CaCO3 m

27 hr−1 in NEC. Nutrients (N:P ratio) were not significant predictors of NEP (p=0.71). Light and maximum temperature were positively correlated in surfgrass pools (p<0.001).

Mussel loss significantly increased light (PFD) values by 7.82 (± 3.35 SE) µmol photons m-2 s-1 per percent loss (p=0.04) and maximum temperature by 0.02°C (± 0.01 SE) per percent loss (p=0.02). With removal, the N:P ratio of the pool was not significantly affected by percent loss of mussels (p=0.07). Change in micro/macroalgae cover increased by 0.73% (± 0.21 SE) per percent of mussel loss (p=0.02), which led to a 0.04 (± 0.01 SE) mmol C m-2 hr-1 increase in NEP rate per percent change of producer cover (p<0.001). The largest direct effect of mussel loss was on pH, where pH increased by 0.003 (± 0.001 SE) units per percent change in mussel loss

(p=0.007). Size of the tide pool had less of an effect in mussel than surfgrass pools, causing a decrease of 0.40 (± 0.11 SE) per unit of N:P per unit change in SA:V (p=0.01). Similar to the community composition results, tide height had a significant effect on micro/macroalgae cover, where micro/macroalgae cover decreased 65.20% (± 30.12 SE) per meter increase in tide height

(p=0.05). Maximum temperature did not have a significant effect on NEC (p=0.36) and nutrients

(N:P ratio) were not significant predictors of NEP (p=0.44). Light and maximum temperature were significantly positively correlated in mussel pools (p<0.001). Full model output can be found in Appendix A Table 10 and regressions of significant relationships can be found in

Appendix B Fig. 3 and 4.

28

FIGURE. 7. Structural equation models to determine how (7.1) surfgrass and (7.2) mussel loss and physical parameters directly and indirectly affect community structure, biogeochemistry, and ecosystem metabolism. Blue lines represent positive paths, red lines represent negative paths, grey lines represent nonsignificant paths (P > 0.05), and black dotted lines are correlated errors. Arrow width is proportional to standardized path coefficients and the coefficient values are written on each significant pathway.

29 Discussion

While studies on foundation species loss have been linked to changes in community structure, physical environment, and biodiversity (Ellison et al. 2005, Shelton 2010, Castorani et al. 2018, Smale et al. 2019), few have focused on a causal approach with multiple facets of ecosystem functioning. This study helps to fill a gap by measuring the magnitude of effects of intertidal foundation species loss on ecosystem functioning through changes in community structure and resource fluxes over a two-month experimental period. My results show that removal of foundation species from tide pool ecosystems had significant cascading and causal impacts on multiple components of ecosystem function. Surfgrass loss led to significant changes in the sessile community, light, and temperature environment. Changes in surfgrass pools’ community structure and physical environment indirectly impacted pH, and ecosystem metabolism (NEC and NEP). Mussel loss caused significant direct changes in sessile communities, light, temperature, and pH, while indirectly influencing ecosystem metabolism.

My findings generally affirm similar intertidal foundation species removal studies (Dethier 1982,

1984, Pfister 2007, Shelton 2010) as well as more recent studies on effects of intertidal communities on ecosystem function (Legrand et al. 2018, Silbiger and Sorte 2018, Wolfe et al.

2020).

Sessile community shifts with foundation species loss increased the presence of other inferior space competitors one month after foundation species loss. After disturbance, both surfgrass and mussel pools saw an increase in early colonizers like diatoms, similar to foundational disturbance studies of the benthic intertidal community (Emerson and Zelder 1978,

Dethier 1981, Murray and Littler 1981, Dethier 1982, 1984, Dethier and Duggins 1984).

Surprisal, only mussel sessile species richness decreased along the gradient of mussel loss as

30 seen in similar mussel (Perumytilus purpuratus) removal studies (Valdivia and Thiel 2006).

Increasing disturbance usually decreases species richness due to the facilitation of secondary sessile and mobile species with the presence of foundation species (Kimbro and Grosholz 2006).

However, it is possible species richness was not affected one-month after foundation species removal. Because recovery and community stability after disturbance of tide pool communities can take more than three years (Dethier 1984, Shelton 2010), it is likely the pools will continue to be dominated by a suite of early successional species along with seasonal increases in ephemeral diatom algae (Dethier 1982, 1984, Shelton 2010) and corticated algal species during upwelling along the Oregon coast (Nielsen and Navarrete 2004, Wieters 2005). There was no significant change in mobile communities in surfgrass or mussel pools, perhaps due the variability between sampling time points (Shelton 2010) or because there may have not been enough time for the community to respond to the disturbance. Mobile invertebrates and fishes can move across the intertidal environment during high tide and may have found environments that had more suitable habitat and resources. My results give a snapshot in time of these tide pool communities that may continue to change with the increased stress of the physical environment with foundation species loss.

Foundation species help maintain lower stress environments through shading and decreased temperature (Bruno and Bertness 2001), with their removal, there are wide-reaching organism, community, and ecosystem-level effects that occur with the change in physical environment. Maximum light and temperature environments increased with removal of mussel and surfgrass within the SEM and most month-long analyses in accordance with previous studies

(Stewart and Myers 1980, Stephens and Bertness 1991, Shelton 2010). For example, increased maximum light can bleach light-sensitive algae, like coralline algae, reducing the production of

31 the functional group like was seen in the surfgrass pools sessile community (Fig. 1.2; Shelton

2010). Increased temperature from foundation species loss can increase success of less temperature-sensitive invasive species within marine (Sorte et al. 2010, Olabarria et al. 2013), freshwater (Rahel and Olden 2008), and terrestrial (Prevéy et al. 2010) systems. Temperature extremes as a result of foundation species loss may also cause mortality if temperatures exceed thermal tolerances of tide pool species (Shelton 2010; Somero 2002). My study found the maximum temperature of pools exceeded on average + 7.7°C in removal surfgrass pools and +

3.8°C in removal mussel pools compared to pools with foundation species. Since select removal pools in this study had maximum temperatures exceeding 30.0°C, which is a maximum threshold for many Northeast Pacific temperate tide pools species including Tegula spp. (Tomanek and

Somero 1999, Tomanek and Helmuth 2002), tidepool sculpin (Oligocottus spp.; Nakano and

Iwama 2002), and >40 macrophyte species (Lüning and Freshwater 1988). The increases in temperature seen in the current study could indicate that pools without foundation species act as ecological traps (Vinagre et al. 2018), where organisms may not survive with >30.0°C maximum temperatures over summer low tide periods (4–6 hours).

Both physical parameters (pool size and tide height) influenced community structure and ecosystem fluxes. These results were contrary to Silbiger and Sorte (2018) that did not find an effect of pool size or tide height on pH and ecosystem metabolism over a larger spatial scale

(Southern CA to Oregon sites). Tide pools in the current study had a wider range in surface area

(0.58–7.54 m2), volume (27.4–868.8L), and tide height (0.71–1.77m) than in Silbiger & Sorte

(2018) and most other tide pool studies (Shelton 2010, Sorte and Bracken 2015, Bracken et al.

2018, Legrand et al. 2018, Wolfe et al. 2020), likely leading to these differences (Appendix A

Table 1.2). Over this range, surfgrass sessile and mobile species composition was affected by

32 both tide height and pool size and mussel sessile species composition was affected by pool size.

This effect may be due to the influence of both tide height and tide pool size on the stability of environmental conditions provided by lower tidal height and larger tide pools (Martins et al.

2007, Bracken et al. 2011). Mobile species richness increased with tide height, possible due to the addition of higher intertidal species, like Littorina spp. (Bracken et al. 2011). Nutrients (N:P ratio) decreased along the SA:V gradient within mussel pools, where nutrients were higher in smaller pools. This result supports diffusive properties of tide pools indicating that smaller pools would be more affected by changes in nutrients output from the mussel community (Hurd 2000).

Within surfgrass pools, both physical parameters, tide height and tide pool size, negatively impacted NEC rates possibly due to emersion time and diffusion rates (Wolfe et al. 2020). To assess ecosystem functioning of tide pools, studies must take into account physical parameters in a complex and dynamic system as they interact with biological and chemical processes (Legrand et al. 2018).

Using a causal model, I determined the magnitude of both surfgrass and mussel loss on the physical and biogeochemical processes, and ecosystem metabolism. Contrary to my predictions, there were not as many direct effects of surfgrass loss as with mussel loss. Mussel pools shifted from consumer dominated to producer dominated with mussel loss, which may have led to more direct effects in the model. In fact, micro/macroalgal cover became a significant predictor of NEP in the mussel model, whereas light was the strongest predictor in the surfgrass model. Both light and producer cover are linked to increasing NEP within other intertidal community structure studies (Takeshita et al. 2016, Silbiger and Sorte 2018). Contrary to hypothesized relationships between N:P and nutrient concentrations on ecosystem production

(Bracken 2004, Bracken and Nielsen 2004, Pfister 2007, Aquilino et al. 2009, Peng et al. 2017),

33 the N:P ratio did not have a significant effect on NEP. It is likely the nutrient environment of tide pools was altered by nutrient cycling ability of mussels (Pfister and Altabet 2019) and surfgrass

(Terrados and Williams 1997, Honig et al. 2017) that was not captured with this study’s instantaneous sampling and analysis design.

Surfgrass pools has a significant biophysical causal pathway from NEP pH NEC, which was expected based on prior studies (Silbiger and Sorte 2018). Mussel pools did not exhibit this biophysical casual pathway possibly due to biological differences in the type of dominant calcifiers between surfgrass and mussel pools or smaller sample size (n = 15 versus 57 in Silbiger and Sorte 2018). Within mussel pools, mussels were the dominant calcifier, whereas surfgrass pools were dominated by crustose coral algae (CCA) calcifiers. With mussel loss, there was a direct removal of that dominant calcifier, which could account for the lack of casual pathway between pH and NEC. In addition, CCA calcifies nearly twice as fast as mussels: studies in temperate species, Lithothamnion corallioides, showed that CCA can calcify 9.7 ±

0.5 µmol CaCO3 g-1 dry wt d-1 (Martin et al. 2006), whereas mussel (M. edulis) calcification rates did not exceed 4 µmol CaCO3 g-1 dry wt d-1 (Wahl et al. 2018). Therefore, changes in TA

(an instantaneous measure of calcification) may have detected stronger calcification signals from surfgrass pools than mussel pools.

This study demonstrates a snapshot of the ecosystem functioning post-removal of foundation species. It is likely that as community structure and the physical environment shift, ecosystem function will shift along with it (Takeshita et al. 2016, Silbiger and Sorte 2018). Even with the short-term nature of the study, I saw strong causal relationships that have not been previously described. However, to make informed conservation management decisions, longer- term and seasonal impacts of mussel and surfgrass loss should be addressed (Ellison et al. 2005).

34 Although tide pools can be used as a natural laboratory to better understand the intertidal system, the relative effect of physics during high tide in contrast to the tide pool biological system is unknown. The strong mechanistic signaling in this study could be either amplified or dulled during high tide, when larger-scale physical processes (e.g., waves and currents) are at play.

Nevertheless, this study underscores that intertidal foundation species loss alters multiple facets of ecosystem function and is likely to have longer-lasting effects on ecosystem functioning in the future as the community recovers from disturbance.

Focusing on the interplay of both community and ecosystem ecology, this study connects foundational studies of disturbance dynamics with contemporary studies focused on the influence of community structure on the physical, biogeochemical, and ecosystem metabolism landscape of tide pools. Dominant organisms (e.g., surfgrass, mussels, diatoms) within the system drive changes in ecosystem functioning. As foundation species do not have a linear effect on the ecosystem (Bruno and Bertness 2001), it is important test the effects of foundation species with a continuous gradient rather than species presence or absence (e.g. Smale et al. 2019). It will be of increasing importance to measure the magnitude of foundation species loss not only to biodiversity and community structure, but also to fluxes of biogeochemical cycling and ecosystem metabolism (Ellison 2019). These flux processes feedback to influence the broader ecological community (Silbiger and Sorte 2018) and the ecological services (Angelini et al.

2011) that we depend on from the ecosystem.

35 REFERENCES

Alongi, D. M. 2002. Present state and future of the world’s mangrove forests. Environmental Conservation 29:331–349. Angelini, C., A. H. Altieri, B. R. Silliman, and M. D. Bertness. 2011. Interactions among foundation species and their consequences for community organization, biodiversity, and conservation. BioScience 61:782–789. Aquilino, K. M., M. E. S. Bracken, M. N. Faubel, and J. J. Stachowicz. 2009. Local-scale nutrient regeneration facilitates seaweed growth on wave-exposed rocky shores in an upwelling system. Limnology and Oceanography 54:309–317. Boese, B. L., J. E. Kaldy, P. J. Clinton, P. M. Eldridge, and C. L. Folger. 2009. Recolonization of intertidal Zostera marina L. (eelgrass) following experimental shoot removal. Journal of Experimental Marine Biology and Ecology 374:69–77. Bracken, M. E. S. 2004. Invertebrate-mediated nutrient loading increases growth of an intertidal macroalga. Journal of Phycology 40:1032–1041. Bracken, M. E. S., E. Jones, and S. L. Williams. 2011. Herbivores, tidal elevation, and species richness simultaneously mediate nitrate uptake by seaweed assemblages. Ecology 92:1083– 1093. Bracken, M. E. S., and K. J. Nielsen. 2004. Diversity of intertidal macroalgae increases with nitrogen loading by invertebrates. Ecology 85:2828–2836. Bracken, M. E. S., N. J. Silbiger, G. Bernatchez, and C. J. B. Sorte. 2018. Primary producers may ameliorate impacts of daytime CO2 addition in a coastal marine ecosystem. PeerJ 6:e4739. Bruno, J. F., and M. D. Bertness. 2001. Habitat modification and facilitation in benthic marine communities. Pages 201–218 Marine Community Ecology. Byers, J. E., K. Cuddington, C. G. Jones, T. S. Talley, A. Hastings, J. G. Lambrinos, J. A. Crooks, and W. G. Wilson. 2006. Using ecosystem engineers to restore ecological systems. Trends in Ecology and Evolution 21:493–500. Byrnes, J. E., D. C. Reed, B. J. Cardinale, K. C. Cavanaugh, S. J. Holbrook, and R. J. Schmitt. 2011. Climate-driven increases in storm frequency simplify kelp forest food webs. Global Change Biology 17:2513–2524. Carranza, A., L. Airoldi, M. C. Kay, B. Hancock, M. W. Beck, X. Guo, C. Crawford, R. D. Brumbaugh, H. S. Lenihan, L. D. Coen, C. L. Toropova, G. J. Edgar, O. Defeo, G. Zhang, and M. W. Luckenbach. 2011. Oyster reefs at risk and recommendations for conservation, restoration, and management. BioScience 61:107–116. Castorani, M. C. N., D. C. Reed, and R. J. Miller. 2018. Loss of foundation species: disturbance frequency outweighs severity in structuring kelp forest communities. Ecology 99:2442– 2454. Chisholm, J. R. M., and J. Gattuso. 1991. Validation of the alkalinity anomaly technique for investigating calcification and photosynthesis in coral reef communities. Limnology and Oceanography 36:1232–1239. Christensen, N. L., A. M. Bartuska, J. H. Brown, S. Carpenter, C. D. Antonio, R. Francis, J. F. Franklin, J. A. Macmahon, R. F. Noss, J. Parsons, C. H. Peterson, M. G. Turner, and R. G. Woodmansee. 1996. The report of the Ecological Society of America Committee on the scientific basis for ecosystem management. Ecological Applications 6:665–691. Costanza, R., R. D’Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem,

36 R. V O’Neill, J. Paruelo, R. G. Raskin, P. Sutton, and M. van den Belt. 1997. The value of the world’s ecosystem services and natural capital. Nature 387:253. Crouch, C. A. 1991. Infaunal polychaetes of a rocky intertidal surfgrass bed in southern California. Bulletin of Marine Science 48:386–394. Dayton, P. K. 1972. Toward an understanding of community resilience and the potential effects of enrichments to the benthos at McMurdo Sound, Antarctica. Pages 81–96 Proceedings of the Colloquium on Conservation Problems. Allen Press, Lawrence, Kansas, USA. Dethier, M. N. 1981. Heteromorphic algal life histories: The seasonal pattern and response to herbivory of the brown crust, Ralfsia californica. Oecologia 49:333–339. Dethier, M. N. 1982. Pattern and process in tidepool algae: factors influencing seasonality and distribution. Botanica Marina 25:55–66. Dethier, M. N. 1984. Disturbance and recovery in intertidal pools: Maintenance of mosaic patterns. Ecological Monographs 54:99–118. Dethier, M. N., and D. O. Duggins. 1984. An “indirect commensalism” between marine herbivores and the importance of competitive hierarchies. The American Naturalist 124:205–219. Dickson, A. G., C. L. Sabine, and J. R. Christian. 2007. Guide to best practices for ocean CO2 measurements. PICES Special Publication 3:1–191. Duarte, C. M., and C. L. Chiscano. 1999. Seagrass biomass and production: A reassessment. Aquatic Botany 65:159–174. Dudgeon, S. R., R. B. Aronson, J. F. Bruno, and W. F. Precht. 2010. Phase shifts and stable states on coral reefs. Marine Ecology Progress Series 413:201–216. Edwards, M., B. Konar, J. K. Id, S. Gabara, G. Sullaway, T. Mchugh, M. Spector, and S. Small. 2020. Marine deforestation leads to widespread loss of ecosystem function. PLOS ONE 15(3):1–21. Ellison, A. M. 2019. Foundation species, non-trophic interactions, and the value of being common. iScience 13:254–268. Ellison, A. M., M. S. Bank, B. D. Clinton, E. A. Colburn, C. R. Ford, D. R. Foster, B. D. Kloeppel, J. D. Knoepp, G. M. Lovett, J. Mohan, D. A. Orwig, N. L. Rodenhouse, W. V Sobczak, A. Kristina, J. K. Stone, C. M. Swan, J. Thompson, B. Von Holle, R. Jackson, A. M. Ellisonl, M. S. Bank, B. D. Clinton, E. A. Colburnm, K. Elliott, C. R. Ford, D. R. Foster, B. D. Kloeppel, J. D. Knoepp, G. M. Lovett, J. Mohan, D. A. Orwig, N. L. Rodenhouse, W. V Sobczak, K. A. Stinson, J. K. Stone, C. M. Swan, J. Thompson, B. Von Holle, and J. R. Webster. 2005. Loss of foundation species: Consequences for the structure and dynamics of forested ecosystems. Frontiers in Ecology and the Environment 3:479–486. Emerson, S. E., and J. B. Zelder. 1978. Recolonization of intertidal algae: An experimental study. Marine Biology:315–324. Fox, J., G. G. Friendly, S. Graves, R. Heiberger, G. Monette, H. Nilsson, B. Ripley, S. Weisberg, M. J. Fox, and M. Suggests. 2007. The car package. R Foundation for Statistical Computing. Gattuso, J.-P., J.-M. Epitalon, H. Lavigne, and J. Orr. 2018. seacarb: seawater carbonate chemistry. Gattuso, J.-P., M. Frankignoulle, and S. V. Smith. 1999. Measurement of community metabolism and significance in the coral reef CO2 source-sink debate. Proceedings of the National Academy of Sciences 96:13017–13022. Grace, J. B. 2008. Structural equation modeling for observational studies. Journal of Wildlife

37 Management 72:14–22. Grace, J. B., D. R. Schoolmaster, G. R. Guntenspergen, A. M. Little, B. R. Mitchell, K. M. Miller, and E. W. Schweiger. 2012. Guidelines for a graph-theoretic implementation of structural equation modeling. Ecosphere 3:art73. Halpern, B. S., K. A. Selkoe, F. Micheli, and C. V. Kappel. 2007. Evaluating and ranking the vulnerability of global marine ecosystems to anthropogenic threats. Conservation Biology 21:1301–1315. Honig, S. E., B. Mahoney, J. S. Glanz, and B. B. Hughes. 2017. Are seagrass beds indicators of anthropogenic nutrient stress in the rocky intertidal? Marine Pollution Bulletin 114:539– 546. Hughes, T. P., J. T. Kerry, A. H. Baird, S. R. Connolly, A. Dietzel, C. M. Eakin, S. F. Heron, A. S. Hoey, M. O. Hoogenboom, G. Liu, M. J. McWilliam, R. J. Pears, M. S. Pratchett, W. J. Skirving, J. S. Stella, and G. Torda. 2018. Global warming transforms coral reef assemblages. Nature 556:492–496. Hurd, C. L. 2000. Water motion, marine macroalgal physiology, and production. Journal of Phycology 36:453–472. Kimbro, D. L., and E. D. Grosholz. 2006. Disturbance influences oyster community richness and evenness, but not diversity. Ecology 87:2378–2388. Kinzig, A. P., S. W. Pacala, and D. Tilman. 2001. The functional consequences of biodiversity: empirical progress and theoretical extensions. Princeton University Press. Krumhansl, K. A., D. K. Okamoto, A. Rassweiler, M. Novak, J. J. Bolton, K. C. Cavanaugh, S. D. Connell, C. R. Johnson, B. Konar, S. D. Ling, F. Micheli, K. M. Norderhaug, A. Pérez- Matus, I. Sousa-Pinto, D. C. Reed, A. K. Salomon, N. T. Shears, T. Wernberg, R. J. Anderson, N. S. Barrett, A. H. Buschmann, M. H. Carr, J. E. Caselle, S. Derrien-Courtel, G. J. Edgar, M. Edwards, J. A. Estes, C. Goodwin, M. C. Kenner, D. J. Kushner, F. E. Moy, J. Nunn, J. Vásquez, R. S. Steneck, J. Watson, J. D. Witman, and J. E. K. Byrnes. 2016. Global patterns of kelp forest change over the past half-century. Proceedings of the National Academy of Sciences 113:13785–13790. Kwiatkowski, L., B. Gaylord, T. Hill, J. Hosfelt, K. J. Kroeker, Y. Nebuchina, A. Ninokawa, A. D. Russell, E. B. Rivest, M. Sesboue, and K. Caldeira. 2016. Nighttime dissolution in a temperate coastal ocean ecosystem increases under acidification. Scientific Reports 6. Laufkötter, C., J. Zscheischler, and T. L. Frölicher. 2020. High-impact marine heatwaves attributable to human-induced global warming. Science (New York, N.Y.) 369:1621–1625. Lefcheck, J. S. 2016. piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods in Ecology and Evolution 7:573–579. Lefcheck, J. S., and J. E. Duffy. 2015. Multitrophic functional diversity predicts ecosystem functioning in experimental assemblages of estuarine consumers. Ecology 96:2973–2983. Legrand, E., P. Riera, L. Pouliquen, O. Bohner, T. Cariou, and S. Martin. 2018. Ecological characterization of intertidal rockpools: Seasonal and diurnal monitoring of physico- chemical parameters. Regional Studies in Marine Science 17:1–10. Littler, M. M., and S. N. Murray. 1975. Impact of sewage on the distribution, abundance and community structure of rocky intertidal macro-organisms. Marine Biology 30:277–291. Long, M. H., J. E. Rheuban, P. Berg, and J. C. Zieman. 2012. A comparison and correction of light intensity loggers to photosynthetically active radiation sensors. Limnology and Oceanography: Methods 10:41–424. Lüdecke, D. 2018. ggeffects: Tidy data frames of marginal effects from regression models.

38 Journal of Open Source Software 3:772. Lüning, K., and W. Freshwater. 1988. Temperature tolerance of Northeast Pacific marine algae. Journal of Phycology 24:310–315. Martin, S., M. D. Castets, and J. Clavier. 2006. Primary production, respiration and calcification of the temperate free-living coralline alga Lithothamnion corallioides. Aquatic Botany 85:121–128. Martins, G. M., S. J. Hawkins, R. C. Thompson, and S. R. Jenkins. 2007. Community structure and functioning in intertidal rock pools: Effects of pool size and shore height at different successional stages. Marine Ecology Progress Series 329:43–55. Murray, S. N., and M. M. Littler. 1981. Biogeographical analysis of intertidal macrophyte floras of southern California. Journal of Biogeography 8:339. Nakano, K., and G. K. Iwama. 2002. The 70-kDa heat shock protein response in two intertidal sculpins, Oligocottus maculosus and O. snyderi: Relationship of hsp70 and thermal tolerance. Comparative Biochemistry and Physiology - A Molecular and Integrative Physiology 133:79–94. Nielsen, K. J. 2003. Nutrient loading and consumers: Agents of change in open-coast macrophyte assemblages. Proceedings of the National Academy of Sciences 100:7660– 7665. Nielsen, K. J., and S. A. Navarrete. 2004. Mesoscale regulation comes from the bottom-up: Intertidal interactions between consumers and upwelling. Ecology Letters 7:31–41. Ninokawa, A., Y. Takeshita, B. M. Jellison, L. J. Jurgens, and B. Gaylord. 2019. Biological modification of seawater chemistry by an ecosystem engineer, the California mussel, Mytilus californianus. Limnology and Oceanography 65:157–172. Oksanen, J., R. Kindt, P. Legendre, B. O’Hara, M. H. H. Stevens, M. J. Oksanen, and M. Suggests. 2007. The vegan package. Community ecology package 10:631–637. Olabarria, C., F. Arenas, R. M. Viejo, I. Gestoso, F. Vaz-Pinto, M. Incera, M. Rubal, E. Cacabelos, P. Veiga, and C. Sobrino. 2013. Response of macroalgal assemblages from rockpools to climate change: Effects of persistent increase in temperature and CO2. Oikos 122:1065–1079. Orth, R. J., T. J. B. Carrunthers, W. C. Dennison, C. M. Duarte, J. W. Fourqurean, K. L. Heck Jr., A. Randall Hughes, G. A. Kendrick, W. Judson Kenworthy, S. Olyarnik, F. T. Short, M. Waycott, and S. L. WIlliams. 2006. A global crisis for seagrass ecosystems. BioScience 56:987–996. Paine, R. T., and S. A. Levin. 1981. Intertidal landscapes: Disturbance and the dynamics of pattern. Ecological Monographs 51:145–178. Pandolfi, J. M., R. H. Bradbury, E. Sala, T. P. Hughes, K. A. Bjorndal, R. G. Cooke, D. McArdle, L. McClenachan, M. J. H. Newman, G. Paredes, R. R. Warner, and J. B. C. Jackson. 2009. Global trajectories of the long-term decline of coral reef ecosystems. Science 955:10–14. Peng, Y., F. Li, G. Zhou, K. Fang, D. Zhang, C. Li, G. Yang, G. Wang, J. Wang, and Y. Yang. 2017. Linkages of plant stoichiometry to ecosystem production and carbon fluxes with increasing nitrogen inputs in an alpine steppe. Global Change Biology 23:5249–5259. Petraitis, P. S., and S. R. Dudgeon. 2020. Declines over the last two decades of five intertidal invertebrate species in the western North Atlantic. Communications Biology 3:1–7. Pfister, C. A. 1995. Estimating competition coefficients from census data: A test with field manipulations of tidepool fishes. The American Naturalist 146:271–291.

39 Pfister, C. A. 2007. Intertidal invertebrates locally enhance primary production. Ecology 88:1647–1653. Pfister, C. A., and M. A. Altabet. 2019. Enhanced microbial nitrogen transformations in association with macrobiota from the rocky intertidal:193–206. Prevéy, J. S., M. J. Germino, N. J. Huntly, and R. S. Inouye. 2010. Exotic plants increase and native plants decrease with loss of foundation species in sagebrush steppe. Plant Ecology 207:39–51. R Core Team. 2018. A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rahel, F. J., and J. D. Olden. 2008. Assessing the effects of climate change on aquatic invasive species. Conservation Biology 22:521–533. Raimondi, P., R. Ambrose, J. Engle, S. Murray, and M. Wilson. 1999. Monitoring of rocky intertidal resources along the central and southern California mainland. 3-Year Report for San Luis Obispo, Santa Barbara, and Orange Counties (Fall 1995-Spring 1998). OCS Study,. Ramirez-Garcia, P., J. Terrados, F. Ramos, A. Lot, D. Ocaa, and C. M. Duarte. 2002. Distribution and nutrient limitation of surfgrass, Phyllospadix scouleri and Phyllospadix torreyi, along the Pacific coast of Baja California (México). Aquatic Botany 74:121–131. Sagarin, R. D., J. P. Barry, S. E. Gilman, and C. H. Baxter. 1999. Climate-related change in an intertidal community over short and long time scales. Ecological Monographs 69:465–490. Schloerke, B. C. J., D. Cook, F. Briatte, M. Marbach, E. Thoen, A. Elberg, and J. Larmarange. 2018. GGally: Extension to ’ggplot2’. CRAN. R-project. Shelton, A. O. 2010. Temperature and community consequences of the loss of foundation species: Surfgrass (Phyllospadix spp., Hooker) in tidepools. Journal of Experimental Marine Biology and Ecology 391:35–42. Shipley, B. 2009. Confirmatory path analysis in a generalized multilevel context. Ecology 90:363–368. Silbiger, N. J., and C. J. B. Sorte. 2018. Biophysical feedbacks mediate carbonate chemistry in coastal ecosystems across spatiotemporal gradients. Scientific Reports 8:1–11. Smale, D. A., T. Wernberg, E. C. J. Oliver, M. S. Thomsen, B. P. Harvey, S. Straub, M. T. Burrows, L. V. Alexander, J. A. Benthuysen, M. G. Donat, M. Feng, A. J. Hobday, N. J. Holbrook, S. E. Perkins-Kirkpatrick, H. A. Scannell, A. Sen Gupta, B. L. Payne, and P. J. Moore. 2019. Marine heatwaves threaten global biodiversity and the provision of ecosystem services. Nature Climate Change 9:306–312. Smith, E. P. 2002. BACI design. Encyclopedia of Environmetrics 1:141–148. Smith, J. R., P. Fong, and R. F. Ambrose. 2006. Long-term change in mussel (Mytilus californianus Conrad) populations along the wave-exposed coast of southern California. Marine Biology 149:537–545. Smith, J. R., and S. N. Murray. 2005. The effects of experimental bait collection and trampling on a Mytilus californianus mussel bed in southern California. Marine Biology 147:699–706. Somero, G. N. 2002. Thermal physiology and vertical zonation of intertidal animals: Optima, limits, and costs of living. Integrative and Comparative Biology 42:780–789. Sorte, C. J. B., and M. E. S. Bracken. 2015. Warming and elevated CO2 interact to drive rapid shifts in marine community production. PLoS ONE 10:1–12. Sorte, C. J. B., V. E. Davidson, M. C. Franklin, K. M. Benes, M. M. Doellman, R. J. Etter, R. E. Hannigan, J. Lubchenco, and B. A. Menge. 2017. Long-term declines in an intertidal

40 foundation species parallel shifts in community composition. Global Change Biology 23:341–352. Sorte, C. J. B., S. L. Williams, and R. A. Zerebecki. 2010. Ocean warming increases threat of invasive species in a marine fouling community. Ecology 91:2198–2204. Stark, K. 2019. Interview: Researcher on Bodega Bay mussel die-off. KQED Science. Stephens, E. G., and M. D. Bertness. 1991. Mussel facilitation of barnacle survival in a sheltered bay habitat. Journal of Experimental Marine Biology and Ecology 145:33–48. Stewart-Oaten, A., W. W. Murdoch, and K. R. Parker. 1986. Environmental impact assessment: “Pseudoreplication” in time? Ecology 67:929–940. Stewart, J. G., and B. Myers. 1980. Assemblages of algae and invertebrates in southern California Phyllospadix-dominated intertidal habitats. Aquatic Botany 9:73–94. Straub, S. C., T. Wernberg, M. S. Thomsen, P. J. Moore, M. T. Burrows, B. P. Harvey, D. A. Smale, and C. E. Cornwall. 2019. Resistance, extinction, and everything in between – The diverse responses of seaweeds to marine heatwaves. Frontiers in Marine Science 6:1–13. Suchanek, T. H. 1979. The Mytilus californianus community: studies on the composition, structure, organization, and dynamics of a mussel bed. PhD Thesis, University of Washington. Suchanek, T. H. 1992. Extreme biodiversity in the marine environment: mussel and communities of Mytilus californianus. Northwest Environment Journal 8:150–152. Takeshita, Y., W. Mcgillis, E. M. Briggs, A. L. Carter, E. M. Donham, T. R. Martz, N. N. Price, and J. E. Smith. 2016. Assessment of net community production and calcification of a coral reef using a boundary layer approach. Journal of Geophysical Research: Oceans 121:5655– 5671. Terrados, J., and S. L. Williams. 1997. Leaf versus root nitrogen uptake by the surfgrass Phyllospadix torreyi. Marine Ecology Progress Series 149:267–277. Thomson, J. A., D. A. Burkholder, M. R. Heithaus, J. W. Fourqurean, M. W. Fraser, J. Statton, and G. A. Kendrick. 2015. Extreme temperatures, foundation species, and abrupt ecosystem change: an example from an iconic seagrass ecosystem. Global Change Biology 21:1463– 1474. Tomanek, L., and B. Helmuth. 2002. Physiological ecology of rocky intertidal organisms: A synergy of concepts. Integrative and Comparative Biology 42:771–775. Tomanek, L., and G. N. Somero. 1999. Evolutionary and acclimation-induced variation in the heat-shock responses of congeneric marine snails (genus Tegula) from different thermal habitats: Implications for limits of thermotolerance and biogeography. Journal of Experimental Biology 202:2925–2936. Valdivia, N., and M. Thiel. 2006. Effects of point-source nutrient addition and mussel removal on epibiotic assemblages in Perumytilus purpuratus beds. Journal of Sea Research 56:271– 283. Vinagre, C., V. Mendonça, R. Cereja, F. Abreu-Afonso, M. Dias, D. Mizrahi, and A. A. V. Flores. 2018. Ecological traps in shallow coastal waters-Potential effect of heat-waves in tropical and temperate organisms. PLoS ONE 13:1–17. Wahl, M., S. Schneider Covachã, V. Saderne, C. Hiebenthal, J. D. Müller, C. Pansch, and Y. Sawall. 2018. Macroalgae may mitigate ocean acidification effects on mussel calcification by increasing pH and its fluctuations. Limnology and Oceanography 63:3–21. Waycott, M., C. M. Duarte, T. J. B. Caruthers, R. J. Orth, W. C. Dennison, S. Olyarnik, A. Calladine, J. W. Fourqurean, K. L. Heck Jr, R. A. Hughes, G. A. Kendrick, W. J.

41 Kenworthy, F. T. Short, and S. L. Williams. 2009. Accelerating loss of seagrasses across the globe threatens coastal ecosystems. Proceedings of the National Academy of Sciences 106:12377–12380. Werfhorst, L. C. Van De, and J. S. Pearse. 2007. Trampling in the rocky intertidal of Central California: a follow-up study. Bulletin Of Marine Science 81:245–254. Wernberg, T., S. Bennett, R. C. Babcock, T. De Bettignies, K. Cure, M. Depczynski, F. Dufois, J. Fromont, C. J. Fulton, R. K. Hovey, E. S. Harvey, T. H. Holmes, G. A. Kendrick, B. Radford, J. Santana-garcon, B. J. Saunders, D. A. Smale, M. S. Thomsen, C. A. Tuckett, and F. Tuya. 2016. Climate-driven regime shift of a temperate marine ecosystem. Science 353:169–172. Wieters, E. A. 2005. Upwelling control of positive interactions over mesoscales: A new link between bottom-up and top-down processes on rocky shores. Marine Ecology Progress Series 301:43–54. Wolf-Gladrow, D. A., R. E. Zeebe, C. Klaas, A. Körtzinger, and A. G. Dickson. 2007. Total alkalinity: The explicit conservative expression and its application to biogeochemical processes. Marine Chemistry 106:287–300. Wolfe, K., H. D. Nguyen, M. Davey, and M. Byrne. 2020. Characterizing biogeochemical fluctuations in a world of extremes: A synthesis for temperate intertidal habitats in the face of global change. GLOBAL CHANGE BIOLOGY 26:3858–3879. Wootton, J. T., C. A. Pfister, and J. D. Forester. 2008. Dynamic patterns and ecological impacts of declining ocean pH in a high-resolution multi-year dataset. Proceedings of the National Academy of Sciences 105:18848–18853.

42 APPENDIX A: TABLES TABLE 1. (1.1) The ratio between surface area to % cover foundation species was +/- 5.8% for control or removal pools as well as designation for sampling day. Tide pools 1-16 were on sampling day 1 and night 2 and tide pools 17-32 were on sample day 2 and night 1. (1.2) Tide pool physical parameters for all pools. 1.1 % cover surfgrass or SA to % Pool ID Type Removal mussel cover 1 Surfgrass 65.6 18.17 N 2 Surfgrass 51.6 25.99 Y 3 Surfgrass 76.2 14.35 Y 4 Surfgrass 72.6 59.07 N 5 Surfgrass 83.99 54.79 N 6 Surfgrass 77 112.74 Y 7 Surfgrass 58.9 19.7 Y 8 Surfgrass 55.7 62.9 N 9 Mussel 88.3 13.68 N 10 Mussel 78.8 38.6 Y 11 Mussel 98.9 93.13 Y 12 Mussel 45.3 17.75 N 13 Mussel 80.9 19.45 N 14 Mussel 92.8 159.18 N 15 Mussel 61.1 57.1 Y 16 Mussel 77.3 11.24 Y 17 Mussel 49.3 44.53 Y 18 Surfgrass 49.5 9.83 N 19 Mussel 82.1 15.09 N 20 Surfgrass 68 14.63 Y 21 Surfgrass 100 45.88 Y 22 Surfgrass 96.4 23.91 N 23 Mussel 57 35.85 Y 24 Mussel 51.7 52.76 Y 25 Mussel 71.6 62.64 N 26 Surfgrass 58.9 48.96 Y 27 Surfgrass 53.9 46.31 N 28 Surfgrass 72.8 31.15 Y 29 Surfgrass 76.8 102.4 N 30 Mussel 81.5 51.42 N 31 Mussel 96.3 28.93 N 32 Mussel 95.3 29.75 Y

43 1.2

44 TABLE 2. Community composition for both surfgrass and mussel tide pools at Otter Rock (n=16 surfgrass; n = 16 mussel). (2.1) sessile species list, (2.2) mobile species list. 2.1 Sessile species list

Diatoms Analipus japonicus Mytilus californianus Algae film Cryptosiphonia woodii Chthamalus spp. Acrosiphonia coalita Cumagloia andersonii Semibalanus cariosus Cladophora columbiana Erythrophyllum Balanus nibulis Centroceras or Ceramium delesserioides Balanus glandula Chaetomorpha linum Halosaccion glandiforme Pollicipes polymerus Savoiea robusta Cryptopleura spp. Tubeworm Polysiphonia spp. Odonthalia floccosa Clathria pennata Turf Algae O. washingtoniensis Halichondria spp. Callithamnion pikeanum Neorhodomela larix Haliclona permollis Pyropia spp. Plocamium spp. Stylantheca spp. Ulva spp. Osmundea spectabilis Anthopluera elegantissima Smithora naiadum Endocladia muricata A. xanthrogrammica Scytosiphon lomentaria Microcladia borealis A. artemisia Halosaccion glandiforme Fucus gardneri Urticina coriacea Leathesia marina Laminaria setchellii Epiactis prolifera Mazzaella splendens Costaria costata Mazzaella oregona Phyllospadix spp. Mazzaella flaccida Non-coralline crusts Palmaria hecatensis Crustose corallines Schizymenia pacifica Bossiella spp. Ptilota spp. Calliarthron tuberculosum Mastocarpus spp. Corallina vancouveriensis Farlowia mollis Corallina spp.

2.2 Mobile species list Lottorina spp. Tegula funebralis Paciocinebrina lurida Lottia spp. Petrolisthes spp. Ceratostoma foliatum Acmaea mitra S. purpuratus Nucella lamellosa Fissurella volcano Brittle Star Lirabuccinum dirum Amphissa spp. Hemigrapsus nudus Alia carinata Bittium eschrichtii Pachygrapsus crassipes Peltodoris nobilis Callistoma canaliculata Heptacarpus stichensis Aeolidia papillosa Cyanoplax hartwegii Pagurus spp. Acanthodoris nanaaimoensis Idotea spp. (Isopod) Pugettia producta Hermissenda opalescens Katharina tunicata Diodora aspera Diaulula sandiegensis Lepidochiton spp. Gunnel Cancer spp. Mopalia spp. Sculpin Pisaster ochraceus Nuttalina spp. Nucella ostrina Leptasterias hexactis Tonicella spp Nucella canaliculata Henricia spp.

45 TABLE 3. Range of chemical parameters from all tide pools (n =31) during day and night sampling in both time periods (before and after) from both removal and control pools of surfgrass (Phyllospadix spp.) and mussels (M. californianus).

46 TABLE 4. Ocean chemistry from all water sampling days before and after foundation species removal periods.

47 TABLE 5. AIC, BIC, Fisher’s C and p-value for light and temperature comparisons in the NEP models both surfgrass and mussel SEMs.

Foundation AIC BIC Fisher's Species Model value value C P-value NEP ~ Light + MicroMacroAlgaeCover Surfgrass 104.25 135.15 24.25 0.76 +NtoPRatio + SAtoVRatio + TideHeight NEP ~ MaxTemp + MicroMacroAlgaeCover Surfgrass 105.15 136.05 25.15 0.72 +NtoPRatio + SAtoVRatio + TideHeight NEP ~ MaxTemp + MicroMacroAlgaeCover Mussel 121.10 149.42 41.10 0.09 +NtoPRatio + SAtoVRatio + TideHeight NEP ~ Light + MicroMacroAlgaeCover Mussel 121.9 150.2 41.97 0.07 +NtoPRatio + SAtoVRatio + TideHeight

48 TABLE 6. Multiple regression summary output for species richness analysis. (6.1) Full model summary; (6.2) individual model responses with coefficients, SE, t-statistic, and p-value. Bolded terms indicate significant values. 6.1

6.2

49 TABLE 7. Output of PERMANOVA models on surfgrass and mussel sessile and mobile community data.

50 TABLE 8. Multiple regression summary output for light and temperature analysis. (8.1) Full model summary; (8.2) individual model responses with coefficients, SE, t-statistic, and p-value. Bolded terms indicate significant values. 8.1

8.2 Surfgrass and maximum temperature Term Coefficient SE T-statistic P-value (Intercept) -2.84 1.72 -1.65 0.12 Surfgrass loss (%) 0.06 0.01 4.25 0.001 SA to V Ratio 0.03 0.05 0.62 0.55 Tide height (m) 1.70 1.81 .936 0.37 Surfgrass and light Term Coefficient SE T-statistic P-value (Intercept) -5.94 24.17 -0.25 0.81 Surfgrass loss (%) 0.78 0.21 3.73 0.003 SA to V Ratio -1.37 0.76 -1.79 0.098 Tide height (m) 3.07 25.48 0.12 0.91 Mussel and maximum temperature Term Coefficient SE T-statistic P-value (Intercept) -3.71 2.25 -1.65 0.13 Mussel loss (%) 0.02 0.01 2.14 0.06 SA to V Ratio -0.05 0.08 -0.69 0.51 Tide height (m) 3.15 1.53 2.07 0.06 Mussel and light Term Coefficient SE T-statistic P-value (Intercept) -2.48 4.28 -0.58 0.57 Mussel loss (%) 0.03 0.02 1.75 0.11 SA to V Ratio -0.14 0.15 -0.96 0.36 Tide height (m) 2.38 2.85 0.83 0.42

51 TABLE 9. R2 values of response variables for both surfgrass and mussel structural equation models.

Surfgrass SEM Mussel SEM Response model R2 model R2 Micro and macroalgae cover 0.39 0.58 Light PFD 0.41 0.40 µmol photons m-2 s-1 Maximum temperature (°C) 0.39 0.43 Nitrate to phosphate ratio 0.13 0.66 Net ecosystem production 0.41 0.81 (mmol C m-2 hr-1) pH 0.83 0.83 Net ecosystem calcification −2 −1 0.81 0.41 (mmol CaCO3 m hr )

52 TABLE 10. Full output of structural equation models. (10.1) surfgrass model and (10.2) mussel model responses and predictors with both unstandardized estimates and standardized estimates. Bolded values indicate significant values. 10.1 Surfgrass model

53 10.2 Mussel model

54 APPENDIX B: FIGURES

FIGURE 1. Distribution of foundation species cover after removal for (1.1) surfgrass pools and (1.2) mussel pools. Blue shading represents control pools and grey shading represents removal pools.

55 FIGURE 2. Relationship between average light and daily maximum temperature values on water sampling days in (2.1) Phyllospadix spp. SEM model and (2.2) M. californianus SEM model. Filled dots are control pools, and empty dots are removal pools and gray shading represents 95% confidence intervals. Light and temperature were positively correlated in (2.1) (pearson’s correlation: t14=4.21, r=0.748, p<0.001) and (2.2) (pearson’s correlation: t13=5.84, r=0.851, p<0.001).

56 FIGURE 3. Estimated marginal effects for significant pathways within the surfgrass structural dequation model. Solid dots represent control pools, open dots represent removal pools, and gray areas represent the 95% confidence interval of these relationships.

57 FIGURE 4. Estimated marginal effects for significant pathways within the mussel structural equation model. Solid dots represent control pools, open dots represent removal pools, and gray areas represent the 95% confidence interval of these relationships.

58