Oxygen and irradiance constraints on visual habitat in a changing ocean: The luminoxyscape

Abstract

Changing oxygen conditions are altering the distribution of many marine animals. Zooplankton vertical distributions are primarily attributed to physiological tolerance and/or avoidance of visual predation. Recent findings reveal that visual function in marine larvae is highly sensitive to oxygen availability, but it is unknown how oxygen, which affects light sensitivity and generates limits for vision, may affect the distribution of animals that rely heavily on this sensory modality. This study introduces the concept of a “visual luminoxyscape” to demonstrate how combinations of limiting oxygen and light could constrain the habitat of marine larvae with oxygen-demanding vision. This concept reveals the impact of sublethal climate change vulnerabilities in visual marine animals and provides an additional hypothesis for habitat compression under ocean deoxygenation, which we argue deserves attention. Manifestations of climate change in the ocean, such as warming, acidification, and deoxygenation, alter the physiological and behavioral responses of marine organisms, and shift their distributions (Somero 2012). Species vulnerability to changing environments is routinely assessed using extreme physiological tolerance limits. “Hypoxia,” often defined as 2 mg O2 L−1, ~ 60 μmol kg−1, or 5 kPa, for example, is a commonly used oxygen threshold in marine life, though it may not accurately reflect an organism's oxygen limits (Vaquer-Sunyer and Duarte 2008). Studies have used a metabolic oxygen limit, or Pcrit (oxygen at which an animal's metabolic rate changes), as a threshold for physiological tolerance to oxygen (Seibel et al. 2021). This limit has inspired several indices (e.g., Metabolic Index, Aerobic Growth Index) that integrate metabolic and biogeographic data to predict future changes in species' distributions by examining combinations of oxygen and temperature conditions that are suitable for metabolic function (Penn et al. 2018; Deutsch et al. 2020; Clarke et al. 2021). Here we introduce a novel concept that relates species-specific experimental visual limits to potential changes in distributions under different irradiance and oxygen conditions to examine the sublethal effects of climate change. Physiological tolerance limits can help to anticipate species-specific consequences of climate change, however, they exclude ecosystem complexity and do not account for important sublethal effects that can decrease survival (Howard et al. 2020). Sublethal effects may impact individual reproduction, locomotion, sensory function, growth, and/or behavior, and can cause population and community changes, including altering organism interactions with other species and the environment. Sensory processing (e.g., vision, olfaction, hearing) can be affected by climate change through direct physiological effects (e.g., Williams et al. 2019) or a change to the predicted sensory environment (Caves and Johnsen 2021). Vision has very high oxygen demands (Niven and Laughlin 2008), particularly in marine organisms with sophisticated eyes supporting high spatial and temporal resolution. Highly visual marine animals include cephalopods, arthropods, and fish (McCormick and Levin 2017); their early life stages (larvae) also rely on vision for diel vertical migration (Forward 1988), prey capture (Chen et al. 1996), and/or predator avoidance (Forward 1977). For these organisms, vision can be inhibited at oxygen levels higher than extreme physiological limits (McCormick and Levin 2017), and therefore is an underappreciated sublethal metric. Sufficient oxygen and irradiance are required for vision, and both can change behavior and distribution of marine organisms (Netburn and Koslow 2015). Irradiance can act as a migration cue for species that undergo diel vertical migration (Forward 1988), and both oxygen and irradiance can alter the daytime maximum depth of this migration (Bertrand et al. 2011; Aksnes et al. 2017; Hobbs et al. 2021). Small changes in oxygen and irradiance (such as from solar eclipses or smoke) can alter distribution and community structure in pelagic organisms (Bright et al. 1972; Kampa 1975; Last et al. 2016; Urmy et al. 2016; Wishner et al. 2018). On longer time scales, global deoxygenation has reduced the average ocean oxygen content by 2% (Schmidtko et al. 2017; Breitburg et al. 2018), and caused habitat compression and changes in abundance for many species, including invertebrate larvae (Weinstock and Collin 2021), euphausiids (Seibel et al. 2016), and fish (Koslow et al. 2011; Stramma et al. 2012; Netburn and Koslow 2015; Mislan et al. 2017). However, few studies have examined how visual impairment from changing ocean conditions affect species' distribution. We propose the visual luminoxyscape: a novel concept linking visual sensitivity to reduced oxygen and irradiance to changes in marine organism habitat. We also identify information needed to advance understanding of this concept. Anthropogenic climate change exacerbates hypoxia in regions with natural variability including, but not limited to, coastal areas, fjords, estuaries, regions with shallow oxygen minimum zones, and eastern boundary upwelling systems (Pitcher et al. 2021). The California Current System, including the Southern California Bight, is an eastern boundary upwelling region with a rich history of environmental monitoring (Ohman and Hobbie 2008; McClatchie 2014). This region hosts a range of economically and ecologically important species with different oxygen tolerances that are subject to fluctuations in aerobic habitat (Howard et al. 2020) caused by changes in source waters or increased remineralization (Evans et al. 2020). Existing environmental gradients and variability over time make the Southern California Bight an excellent case study location to illustrate the visual luminoxyscape concept. We focus on early life stages of four, co-occurring species common to the Southern California Bight: California market squid Doryteuthis opalescens paralarvae, two-spot octopus Octopus bimaculatus paralarvae, graceful rock crab Metacarcinus gracilis megalopae, and tuna crab Pleuroncodes planipes zoea II larvae. These early life stages (hereafter all called larvae) reside in nearshore regions of the Southern California Bight, vertically migrate in the upper water column, and require vision for survival (Boyd 1960; Fields 1965; Ambrose 1988; Gotshall 2005). However, adults of each species live in habitats with different oxygen conditions and may have different physiological tolerances to reduced oxygen (Seibel et al. 2018). In all four species, retinal function declined by 60–100% under reduced partial pressure of oxygen (3.8–5.2 kPa) (McCormick et al. 2019). From these experiments, we developed a metric for oxygen sensitivity to vision, where V90, V50, and V10 are the partial pressures of oxygen that cause 90%, 50%, and 10% retinal function, respectively. In D. opalescens and O. bimaculatus larvae, V10–V50 oxygen conditions affected visual behavior in response to light (photobehavior) (McCormick et al. 2022). Here, oxygen limits for vision are compared to the more common Pcrit metric (Table 1; McCormick 2019). We assess the ecological implications of ocean change for larval organisms by interpreting experimental results in the context of actual environmental fluctuations. We adopt the “luminoxyscape,” which relates species-specific experimental visual limits to potential changes in distributions under different irradiance and oxygen conditions, and demonstrate the benefits of quantifying the effects impaired vision may have on both low-oxygen sensitive species (D. opalescens and M. gracilis) and low-oxygen tolerant species (O. bimaculatus and P. planipes). We hypothesize that the vertical distributions (here termed habitat) of larvae with oxygen-demanding vision are constrained by oxygen and irradiance that permit a sufficient level of visual function for survival. We call this habitat the “visual luminoxyscape,” which includes all depths at which an animal is predicted to have ≥ 50% visual function given the oxygen availability, and assuming a necessary minimum level of irradiance of 0.0311 μmol photons m−2 s−1 (see Supporting Information for threshold details). The maximum depth of this habitat is the “visual luminoxyscape depth” (VLD). Visual function is defined here as the electrical response of the retina to light stimulation, called an electroretinogram; 50% visual function implies a retinal response that is half the maximum visual response under normal (near-surface water) oxygen conditions, and the irradiance limit is the lowest light level tested to elicit a visual response (see Supporting Information for details). To calculate the visual luminoxyscape, we quantify the oxygen and irradiance conditions of larval habitat (0–300 m depth), using visual and metabolic limits for larvae of D. opalescens, O. bimaculatus, M. gracilis, and P. planipes (Table 1) and environmental data from the Southern California Bight. We used oxygen and irradiance data from daytime hydrographic casts collected by the California Cooperative Ocean Fisheries Investigations (CalCOFI) monitoring program between 1984 and 2019 (Appendix S1). We analyzed data from La Jolla to Point Conception, California, a region where the habitats of all four species overlap. Data were categorized into nearshore (within 30 km from shore) and offshore (31–215 km from shore) regions, and organized into winter (December–February), spring (March–May), summer (June–August), and fall (September–November) seasons. The interannual El Niño-Southern Oscillation (ENSO) phase was determined by classifying years as El Niño, La Niña, or “non-ENSO” using the Oceanic Niño Index (climate.gov). We then determined if, when, and where conditions could be limiting for vision and evaluated whether the VLD varied with distance from shore, seasonally, during ENSO phases, and over decadal time scales in the Southern California Bight. In addition, we compared oxygen limits for vision to metabolic limitation, and assessed whether changes in the visual luminoxyscape could affect the spatial and temporal distributions of the marine invertebrate larvae studied. In the Southern California Bight, the visual luminoxyscape determined from our prior empirical relationships varied with distance from shore and across multiple time scales for larvae of D. opalescens, O. bimaculatus, M. gracilis, and P. planipes. For all species, visual limitation from insufficient oxygen alone was more likely than visual limitation by irradiance or limitation by both oxygen and irradiance (Fig. 1A). For the two more sensitive species, D. opalescens and M. gracilis, respectively, 67% and 59% of all hydrographic observations (to 300-m depth) analyzed had oxygen levels that could limit vision. In contrast, larvae of O. bimaculatus and P. planipes would be oxygen limited for vision in 47% and 45% of these observations, respectively. Visual limitation from insufficient irradiance alone would occur infrequently (< 1.3% for all species). However, when irradiance limitation occurred, it occurred at shallower depths than oxygen limitation for the more hypoxia-tolerant species (O. bimaculatus and P. planipes) (Fig. 1B). The depth of oxygen limitation shoaled closer to the coast (nearshore), in comparison to offshore, and was shallower in larvae of D. opalescens and M. gracilis than in larvae of O. bimaculatus and P. planipes (Fig. 1B). The depth of irradiance limitation for all species also shoaled nearshore in comparison to offshore. The depth of oxygen limitation shoaled from offshore to nearshore by more than 20 m for all species, from 75 to 54 m for D. opalescens, 175 to 148 m for O. bimaculatus, 102 to 82 m for M. gracilis, and 186 to 163 m for P. planipes larvae (Kruskall–Wallis, all p < 0.01). The VLD varied seasonally for all species in nearshore regions (Kruskal–Wallis, all p < 0.001) (Fig. 2A,C,E,G). In addition, the VLD differed between all seasons but fall and winter (Dunn test, p < 0.05) in nearshore regions, except for spring–summer comparisons for O. bimaculatus and P. planipes (Table S1). The VLD was shallowest to deepest in spring, summer, winter, and fall, respectively. This trend follows general oxygen dynamics in southern California, where the upwelling of low-oxygen water occurs in spring and summer and the oxycline deepens in fall and winter. Interannual ENSO variability in the nearshore VLD was significant only for larvae of O. bimaculatus and M. gracilis (Kruskal–Wallis, both p = 0.03; Table S2). In general, the VLD was deeper during El Niño years than in La Niña and non-ENSO years; this difference was most apparent in fall and winter and almost reversed in summer (Fig. 2). Between 1984 and 2019, the nearshore, springtime VLD decreased for all species at a rate of 1–3.8 m yr−1 (Fig. 2; generalized linear model, all p < 0.001). The mean, nearshore VLD shoaled by 35 m for D. opalescens, 30 m for M. gracilis, 108 m for O. bimaculatus, and 107 m for P. planipes larvae (Fig. 2). The linear decreases in nearshore VLD echo the decreases in oxygen observed in the coastal California Current (Meyer-Gutbrod et al. 2021), emphasizing that changes to suitable visual habitat were primarily driven by changes in oxygen content. Offshore, the VLD also shoaled with time, and showed interannual variability around the mean, following the multidecadal trend in the oxycline (Gallo et al. 2019). Oxygen and irradiance conditions that could impair vision occurred in the upper 300-m depth in the Southern California Bight during 1984–2019. These changes in the species-specific visual luminoxyscape could impact the behavior and distributions of D. opalescens, O. bimaculatus, M. gracilis, and P. planipes larvae, as well as of other taxa not yet studied. Only two studies quantified the vertical distributions of these species (Wing et al. 1998; Zeidberg and Hamner 2002), but both indicate that larvae are found within their visual luminoxyscape and not in visually unsuitable areas. In the Channel Islands, an archipelago 12–115 km off the coast of southern California, daytime abundance of D. opalescens larvae was greatest at 30-m depth (Zeidberg and Hamner 2002). At the time of the study (1999–2001), that depth was shallower than the mean springtime, nearshore VLD (~ 50 m). Similarly, the daytime depth range previously reported for larval M. gracilis in Point Reyes, CA (30–60 m; Wing et al. 1998) is shallower than the VLD we calculated for summer and fall when larvae are present. Larvae of P. planipes and O. bimaculatus had the shallowest VLD during their spring and summer larval periods, respectively (Fig. 2). The Southern California Bight experienced severe declines in oxygen concentrations, at a rate much faster than global averages (Bograd et al. 2008, 2015). Between 1984 and 2019, the nearshore VLD of all species shoaled (Fig. 2), following the decreasing depth of the oxycline in the region (Evans et al. 2020). For larvae of D. opalescens and M. gracilis, the shoaling nearshore VLD intersects with known larval vertical distributions, suggesting that these larvae could already experience visual stress. Larval distributions for the species examined, however, were obtained from net samples and reported over coarse (10–30 m) depth bins (Wing et al. 1998; Zeidberg and Hamner 2002), which does not permit for a true test of the luminoxyscape concept. Because larval abundances are typically low, even using high-resolution acoustic data (e.g., Urmy and Benoit-Bird 2021) paired with nets for species identification can be challenging to get robust abundance estimates. Emerging zooplankton-imaging technology such as the Zooglider (Whitmore and Ohman 2021) or In Situ Ichthyoplankton Imaging System (Cowen and Guigand 2008) equipped with irradiance and oxygen sensors, or combined with in situ genomic capabilities such as the Mesobot (Yoerger et al. 2018), are most promising to permit the exploration of the luminoxyscape concept. With this essay, we hope to solicit and inspire such studies. Using the luminoxyscape concept, we found vision to be limited at higher oxygen conditions than the critical oxygen for metabolism (Pcrit; Table 1), and the depth associated with Pcrit in D. opalescens and O. bimaculatus larvae was generally deeper than the VLD (> 225 and > 457 m, respectively, data not shown). The impacts of ocean deoxygenation in these species are, therefore, likely to be underestimated if only metabolic oxygen limits are considered. When matching species-specific empirical physiological thresholds with abundance data, it is important to consider both sublethal effects and the role of behavior (Pörtner 2012; Keil et al. 2021). The concept of the luminoxyscape places laboratory-derived physiological and behavioral oxygen limits for vision in the context of past and current environmental conditions. The shoaling of the luminoxyscape over time highlights the potential sublethal impacts of changing ocean conditions for a variety of species (coastal, pelagic, and benthic) and life stages. Examining visual impairment as a sublethal impact of climate change is especially crucial in regions where low-oxygen conditions reach the photic zone. The impact of anthropogenic change (including deoxygenation) on vertical distributions is not well established in larvae and translating population-based distribution data to studies of species-specific vulnerability is crucial (Bandara et al. 2021). For adaptive management of species in the face of climate change, the combined effects of sublethal limits to multiple stressors across life stages should be considered. For example, changing temperature and exposure to reduced carbon dioxide has also been shown to affect visual physiology and behavior (Fritsches et al. 2005; Cohen and Frank 2006; Forsgren et al. 2013). While the luminoxyscape concept focuses on the effects of reduced oxygen and irradiance on marine larvae, additional studies should conduct comparative tests to parse out the synergistic thresholds to holistic environmental changes in both larvae and adults of other species. Here we propose that variability in the VLD could affect larval distributions, and argue that the visual luminoxyscape is a valuable concept to assess species-specific vulnerability and potential for changing habitats in a future of ocean deoxygenation. Appendix S1. Supporting Information Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. We acknowledge and thank the continuous efforts of everyone involved in all aspects of the California Cooperative Oceanic Fisheries Investigations program for providing the hydrographic data used in this analysis. We thank the SURF REU program for supporting SG during the initial analysis of these data (National Science Foundation REU GEO/OCE-1659793). We thank the four anonymous reviewers for constructive comments on the manuscript versions. We are grateful to a Limnology and Oceanography Letters Early Career Publication Honor to LRM for providing support and enabling this research to be open access. This research was supported by a National Science Foundation Graduate Research Fellowship grant DGE-1144086 to LRM, a PEO. Scholar Award to LRM, and a National Science Foundation grant OCE-1829623 to LAL and NWO. JCG was supported by a National Science Foundation grant OCE-1459393 to Peter J. Franks and Andrew J. Lucas.