Direct effects and prey-mediated effects of global change in projections of early life stages of pelagic predators

Global change will impact the distribution and abundance of predators through a combination of abiotic variables, such as temperature, and biotic variables, such as prey availability. However, there is a poor understanding of how distribution projections with biotic variables differ from those with abiotic variables, particularly in resource limited and marine systems. We address this knowledge gap using the planktonic larvae of iconic and economically important pelagic fish predators. We leverage a multidecadal, long-term sampling program from the western Atlantic Ocean to assess the efficacy of using zooplankton prey (copepods, larvaceans and cladocerans) and climate variables to predict the distribution of larvae of seven pelagic fish species, including tunas, billfishes and mahi-mahi. Zooplankton prey, particularly larvaceans, showed high importance for predicting the distribution of smaller tunas. Temperature showed high importance for true tuna (Thunnus spp.), billfish and mahi-mahi. Statistical models linking predator, prey and abiotic variables were forced with climate projections from an ensemble of earth system models to assess zooplankton and fish larvae distribution changes. Redistributions and declines of zooplankton prey led to minimal changes in abundance and distribution for most larval taxa. However, direct climate change effects, driven partially by ocean warming, led to increases in abundance and northward distribution shifts for multiple larval taxa. These climate change-zooplankton–fish larvae relationships highlight that future distribution and abundance changes of predators can be dampened when assessing impacts of prey availability changes. We also show that in a resource-limited system, key pelagic predators, many of which produce lucrative fisheries, are spatiotemporally linked with their preferred zooplankton prey. Direct effects and prey-mediated effects of global change in projections of early life stages of pelagic predators

Global change will impact the distribution and abundance of predators through a combination of abiotic variables, such as temperature, and biotic variables, such as prey availability. However, there is a poor understanding of how distribution projections with biotic variables differ from those with abiotic variables, particularly in resource limited and marine systems. We address this knowledge gap using the planktonic larvae of iconic and economically important pelagic fish predators. We leverage a multidecadal, long-term sampling program from the western Atlantic Ocean to assess the efficacy of using zooplankton prey (copepods, larvaceans and cladocerans) and climate variables to predict the distribution of larvae of seven pelagic fish species, including tunas, billfishes and mahi-mahi. Zooplankton prey, particularly larvaceans, showed high importance for predicting the distribution of smaller tunas. Temperature showed high importance for true tuna ( Thunnus spp.), billfish and mahi-mahi. Statistical models linking predator, prey and abiotic variables were forced with climate projections from an ensemble of earth system models to assess zooplankton and fish larvae distribution changes. Redistributions and declines of zooplankton prey led to minimal changes in abundance and distribution for most larval taxa. However, direct climate change effects, driven partially by ocean warming, led to increases in abundance and northward distribution shifts for multiple larval taxa. These climate change-zooplankton–fish larvae relationships highlight that future distribution and abundance changes of predators can be dampened when assessing impacts of prey availability changes. We also show that in a resource-limited system, key pelagic predators, many of which produce lucrative fisheries, are spatiotemporally linked with their preferred zooplankton prey.

Trophic, Species distribution model, Highly migratory species, Zooplankton, Fisheries, Climate change

Projecting population responses to climate change is vital for ecosystem and resource management. One key concern is climate-driven spatial distribution shifts, which can lead to reorganized food webs, reduced effectiveness of spatial and resource management measures, and loss of ecosystem services (Peci et al. 2017, Lawlor et al. 2024). Correlative species distribution models (SDMs) are frequently used to assess distribution shifts, as they can be forced with past, present and future climate conditions (Beaumont et al. 2007). Abiotic effects (e.g., warming) are often assessed, with biotic effects including prey availability changes receiving far less attention in SDM projections (Godsoe et al. 2015, Dormann et al. 2018, Khosravi et al. 2021). More work is needed to assess the role of prey availability changes in distribution shifts under climate change, particularly in resource limited systems where prey availability may play dominant roles in influencing predator distributions (Van der Putten et al. 2010, Patel et al. 2024). In the marine realm, climate change impacts on resource limited, subtropical systems are proportionally understudied (Poloczanska et al. 2016). These important systems function as spawning grounds for pelagic, predatory fish of high economic and ecological importance (Trenkel et al. 2014, Collette and Cole 2010) including “true” tuna (Scombridae, Thunnus spp.), billfish (Istiophoridae), mahi-mahi ( Coryphaena spp.) and therefore host their environmentally sensitive planktonic larvae (i.e., ichthyoplankton). These taxa in part comprise what are known as highly migratory species (HMS) and support large recreational and commercial fisheries worldwide, including coveted but threatened sportfish such as sailfish and marlin (Dell’Apa et al. 2023). Despite their wide distributions, spawning and larval development of HMS is often spatiotemporally constrained to individual seasons and international basins including the Gulf of Mexico (hereafter, Gulf) and Mediterranean Sea (Block et al. 2005, Collette and Cole 2010). Previous studies have projected climate change impacts on larval HMS in these basins, revealing species-specific shifts in distribution and habitat suitability (Muhling et al. 2011, 2015,) but trophic dynamics are rarely incorporated in projection models (but see Lehodey et al. 2008). This represents a major knowledge gap, as spatiotemporal overlap of predator and prey can affect persistence of predator populations, and is particularly important for ichthyoplankton survival, i.e., the match-mismatch hypothesis (Hjort 1914, Cushing 1990, Hinrichsen et al. 2002). Extensive work has documented selective planktivory of larval tuna and billfish on zooplankton including larvaceans, cladocerans, and copepods (e.g., Llopiz and Cowen 2008, Llopiz and Hobday 2015, Shiroza et al. 2022, and references therein) and has highlighted the importance of bottom-up processes supporting tuna larvae (Landry et al. 2019, Malca et al. 2022, Stukel et al. 2022, Quintanilla et al. 2024). This foundational trophic ecology knowledge highlights the strength in using larvae of pelagic predators as an opportunity to investigate broad questions regarding the importance of prey availability in SDM projections. Our ultimate objective was to compare projected abundance and distributional shifts in larvae of select pelagic predators via direct effects (climate variables) and indirect, or prey-mediated effects (biotic variables including zooplankton prey). We first use SDMs to project climate change impacts on specific zooplankton taxa (“Zooplankton models”: Fig. 1). We then assessed the ability of individual zooplankton taxa and climate variables to predict the distribution of ichthyoplankton (“Biotic Models” and “Abiotic Models”: Fig. 1). Lastly, we used SDMs to project climate change impacts on ichthyoplankton distributions through direct and prey-mediated climate change effects. We hypothesized that biotic models would produce similar prediction skill of larvae of pelagic predators relative to the abiotic models. We also hypothesized that projected changes in abundance and distribution for each taxon would differ between biotic and abiotic models.

Data Collection: We leveraged long-term monitoring data from annual spring Southeast Area Monitoring and Assessment Program (SEAMAP) plankton surveys in the Gulf. These surveys occur annually from April to May in the northern Gulf and were designed to characterize spawning dynamics and larval abundances of Atlantic bluefin tuna ( Thunnus thynnus ) in efforts to improve adult stock assessments (Scott et al. 1993, Ingram et al. 2010). However, >300 species of larval fish and ~50 zooplankton taxa are also identified from samples. Surveys occur at a fixed grid of stations in offshore waters, of depths typically > 200m (Supporting Information). At each station, oblique bongo tows (nets of 61cm net diameter, and 333 μm mesh size) capture zoo- and ichthyo-plankton simultaneously. Tows are accompanied with CTD casts to obtain temperature, chlorophyll a and salinity measurements at three depths: surface, the depth of the chlorophyll maximum and 200m (or within 5m of the bottom). These data are made available upon request from the Southeast Fisheries Science Center. We obtained abundance values for seven ichthyoplankton taxa: frigate and bullet tuna ( Auxis spp., hereafter just frigate tuna), mahi-mahi ( Coryphaena spp.), little tunny ( Euthynnus alletteratus ), skipjack tuna ( Katsuwonus pelamis), billfish (Istiophoridae), Atlantic bluefin tuna ( T. thynnus ) and other “true” tuna ( Thunnus spp.). We chose these seven taxa largely based on economic and ecological significance, and use of offshore Gulf habitat for spawning (Habtes et al. 2014). We also obtained abundance values for four zooplankton prey taxa: calanoid copepods (Order Calanoida, hereafter, calanoids), cladocerans (Class Diplostraca), cyclopoid copepods (Order Cyclopoida, hereafter, cyclopoids) and larvaceans (Class Appendicularia). We chose these four taxa based on a literature review that identified zooplankton taxa that represent key prey items for our target ichthyoplankton taxa (Supporting Information, also see Llopiz and Hobday 2015). See Supporting Information for further survey and data collection details. Species Distribution Modeling: All analyses were conducted using R Statistical Software (v4.3.0; R Core Team 2023) and all plots were created using the “ggplot2” R package (v3.5.1; Wickham 2016). Our data set included 700 observations (i.e., tows) from 2006-2019 for which ichthyoplankton, zooplankton and climate data were available and sampled at a minimum depth of 200m (Supporting Information). One outlier was identified and removed (see Supporting Information for further details). We used generalized additive models (GAMs) as our SDMs, built in the “mgcv” R package (v1.9.1; Wood 2011; Marra and Wood 2011). We chose GAMs as they allow for customizing parameters that reduce error due to ecologically unrealistic relationships and extrapolation. Specifically, we adjusted the number of knots (‘k’ parameter) for each predictor to ensure relationships consistent with a priori knowledge of environmental niches and previous studies (e.g., see Teo et al. 2007, Muhling et al. 2010) and adjusted the order of the first derivative (‘m’ parameter) to 1.0 (Barnes et al. 2022) to prevent extrapolating increasing values outside of training ranges (Brodie et al. 2022). We used a Tweedie distribution, as the data displayed skew, moderate zero inflation and heterogeneity of variance. Residual plots confirmed that Tweedie distributions produced appropriate models (Supporting Information). GAMs predicting zooplankton abundance (“Zooplankton Models”: Fig. 1) included six total predictor variables: time of day (hour and minute of day) with a cyclic cubic spline to account for its cyclic nature, s(lon, lat) to account for uncaptured environmental variability and four climate variables of sea surface temperature (SST) (°C), temperature at 200m depth ( °C), surface chlorophyll (μg/L), and surface salinity. We acknowledge chlorophyll is technically not an abiotic factor, but refer to it as so in this group of predictors for the sake of simplicity. Pearson correlation coefficients among climate variables were all weaker than 0.6 (Supporting Information). We used a multi-phase model/variable selection method to choose the model, for each taxa, to be used for projections. This process is described for zooplankton models (Fig. 1), as follows. First, we set ‘select = TRUE’ when using the ‘gam’ function, allowing the estimated degrees of freedom, and therefore effect, of unimportant predictor variables to be shrunk to zero. Any such predictor variables were then removed from further consideration. We then used the ‘dredge’ function in the MumIn R package (v1.47.5: Bartoń 2023) to create models with all possible combinations of predictor variables. This allowed us to create a subset of candidate models, which included all models with an Akaike Information Criterion (AIC) within 2.0 of the lowest AIC and/or with an AIC weight greater than 0.05. Lastly, each candidate model underwent a leave one (year) out (LOO) analysis. Here, a year of data was removed, the model was fit to the remaining years and that model was then used to predict values for the omitted year. This procedure was repeated for each year for each model. The mean LOO R 2 across withheld years was then calculated (hereafter, predictive skill). The model with the highest predictive skill was used for projections. GAMs predicting ichthyoplankton abundance underwent the same procedure, but in two iterations: one for Abiotic Models (Fig. 1) and a second for Biotic Models (Fig. 1). Biotic models supplemented the four climate variables with the four zooplankton prey taxa as predictors. For each ichthyoplankton taxa, the abiotic and biotic model with the highest predictive skill were used for projections, giving two sets of projections for each taxon. Projections and Distribution Maps: We used NOAA’s CMIP6 (Coupled Model Intercomparison Project) Climate Change Web Portal (https://psl.noaa.gov/ipcc/cmip6/ccwp6.html) to download ESM output of our four environmental variables at 1° spatial resolution. We used an ensemble mean of four CMIP6 ESMs previously used for Gulf projections (Supporting Information) and presenting diverging scenarios to aid quantifying SDM projection uncertainty (Brodie et al. 2022). A moderate shared socioeconomic pathway with 2.5-4.5 W/m 2 of radiative forcing by 2100 (SSP2-4.5) scenario was used. ESM output was averaged for two periods: historical (1985-2014) and future (2070-2099) for spring months (April, May and June). Values for each climate variable from ESMs were bias corrected using the delta method following Ho et al. (2012), using observed values from spring SEAMAP plankton surveys from the same period (1985-2014) as observations. Bias corrected ESM output from historical and future periods were then used as input for SDMs. Distribution maps were created for each model for each plankton taxon for each time-period. Percent changes in abundance were calculated by subtracting the future abundance of a given taxon at a given grid cell from that of the historical period, and dividing by that of the historical period. The geographic center of abundance (GCOA) was calculated for each model for each taxon for each period as the weighted mean latitude and longitude, with taxa abundances at each grid cell representing the weights (Pinsky et al. 2013). To quantify displacement for each taxon, the distance (haversine, km) and direction (degrees, and corresponding cardinal direction) between historical and future GCOAs were calculated using the ‘distHaversine’ and ‘bearing’ functions, respectively, in the geosphere R package (v1.5-18: Hijmans 2022).

Model Performance: Models for calanoids and cladocerans explained more variance (DE of 35.8 and 43.7%, receptively) and displayed superior predictive skill (0.31 and 0.20, respectively) relative to those for larvaceans and cyclopoids (Table 1). Temperature was an important predictor for all four zooplankton taxa, with temperature at depth producing partial DE values as high at 15%, for calanoids (Supporting Information). SST generally produced U-shaped relationships with zooplankton abundance and temperature at depth generally produced negative relationships with zooplankton abundance (Supporting Information). For Atlantic bluefin tuna, other true tuna, frigate tuna and mahi-mahi, abiotic and biotic models produced similar predictive skill (Table 1). For little tunny, the biotic model produced higher predictive skill, but for billfish and skipjack tuna, abiotic models produced higher predictive skill (Table 1). Differences in predictive skill between biotic and abiotic models were the most drastic for billfish, for which the abiotic model displayed the highest predictive skill (0.43) and DE (42.4%) of all ichthyoplankton models (Table 1). Abiotic models explained more variance, relative to biotic models, for all ichthyoplankton taxa, except little tunny and mahi-mahi (Table 1). SST was retained in the best abiotic model for all seven ichthyoplankton taxa, and produced partial DE values up to 20%, for billfish (Supporting Information). SST produced positive or dome shaped relationships with all seven ichthyoplankton taxa (Supporting Information). Temperature at depth was also highly prevalent in abiotic models and produced partial DE values up to 23%, for frigate tuna (Supporting Information). Larvaceans were retained in the best biotic model for all seven ichthyoplankton taxa, and produced partial DE values up to 9.1%, for little tunny, and explained more variability than latitude and longitude for skipjack tuna (Supporting Information). Larvaceans produced dome shaped or positive relationships with all seven ichthyoplankton taxa (Supporting Information). Cladocerans were also highly prevalent in biotic models and produced partial DE values up to 9.2%, for billfish (Supporting Information). Climate Projections: Projected SST increases ranged from 2.0-2.7°C, with observed, historical values primarily 23.5-28.0°C (Supporting Information). Greater SST increases occurred in northeast waters, although maximum projected SSTs occurred in southeast waters, exceeding 29°C (Supporting Information). Projected temperature at depth increases were higher than that of SST (up to 3.0°C), with observed, historical values primarily 10.0-25.0°C (Supporting Information). Similar to SST, greater temperature at depth increases occurred in northeast waters (Supporting Information). Projected surface salinity increases were greatest in northeastern waters and reached maximum values of 37.4, with observed, historical surface values primarily >35.5 (Supporting Information). Projected chlorophyll concentrations generally increased, with observed, historical values of 0.06-0.21 μg/L (Supporting Information). Zooplankton SDMs and Projections: Historical predictions of zooplankton abundances were generally higher in eastern, than western, waters (Fig. 2). Future projections showed larvaceans and cyclopoids becoming more widespread, and north- and/or east-ward distribution shifts for all four taxa (Fig. 2). Projected changes in zooplankton GCOA between periods were relatively minor (Fig. 3). Cladocerans yielded the largest GCOA change of 55 km to the southeast (Fig. 3). Changes in zooplankton abundance were more pronounced than GCOAs, with projected declines in abundance for calanoids, cyclopoids and cladocerans (Fig. 4). Cladocerans showed the largest mean decline in abundance (76%), whereas larvaceans showed the smallest change of a 1% mean increase in abundance (Fig. 4). See Supporting Information for observed zooplankton abundance and frequency of occurrence values. Ichthyoplankton SDMs and Projections: Broadly, biotic and abiotic SDMs aligned for historical predictions of individual ichthyoplankton taxa (Fig. 5). For example, both model types predicted high abundances of Atlantic bluefin tuna in northern waters, high billfish abundances in southern waters and high frigate tuna abundances in eastern waters (Fig. 5). Although there was some misalignment, as in western waters, mahi-mahi biotic models produced lower abundances compared to abiotic models and other true tuna abiotic models produced lower abundances compared to biotic models (Fig. 5). Historical predictions of abundance value ranges aligned well between biotic and abiotic models for other true tuna, skipjack tuna, little tunny and mahi-mahi (Fig. 5). However, for Atlantic bluefin and frigate tuna, abiotic models predicted higher historical abundances and for billfish, biotic models predicted higher historical abundances (Fig. 5). Abiotic models generally yielded more drastic projected abundances increases and distribution shifts that biotic models (Fig. 4). Biotic models produced minimal changes in abundance, except for little tunny and frigate tuna, which yielded an 18% increase and 29% decrease in mean abundance, respectively (Fig. 4). Abiotic models produced more severe changes in abundance, with five taxa (mahi-mahi, little tunny, billfish, skipjack tuna and other true tuna) yielding >45% increases in mean abundance (Fig. 4). There was some alignment between model types for direction (rather than magnitude) of abundance changes, as both biotic and abiotic models projected decreases in frigate tuna abundance, increases in little tunny abundance and minimal changes in Atlantic bluefin tuna abundance (Fig. 4). Changes in GCOA were less pronounced than abundance changes, with only three models (abiotic other true tuna, abiotic billfish and abiotic little tunny) projecting >50 km changes in GCOA, all of which showed GCOA shifts to the northwest (Fig. 4). While changes in GCOA were not pronounced for most ichthyoplankton taxa, SDMs still revealed notable projected distribution shifts, with the magnitude and direction of the shift differing between model type (Fig. 5). For example, for other true tuna, little tunny, skipjack tuna and mahi-mahi, biotic models produced minor distribution shifts but abiotic models produced notable northeastward distribution shifts (Fig. 5). For the other three taxa, SDMs revealed minor distribution shifts for both abiotic and biotic models (Fig. 5). See Supporting Information for observed, historical ichthyoplankton abundance and frequency of occurrence values.

Our finding that changes in prey availability (i.e., biotic models) may dampen climate change impacts on distribution and abundance shifts of pelagic predators fills a broad knowledge gap in the overlapping field of trophic and climate change ecology. The importance of prey availability, particularly in resource limited systems (Johnson and Wallace 2005, Lai et al. 2024) is well recognized, as predator distributions may be tied to prey availability (Woodworth-Jefcoats et al. 2017, Patel et al. 2024). While predator-prey overlap may occur due to shared responses to environmental factors, shared habitat preferences or direct trophic interactions (e.g., bottom-up effects), cooccurrence can increase survival and/or growth of predators due to increased resource availability (Dormann et al. 2018). These potential positive effects of predator-prey overlap for predators (although negative effects are possible through disease transmission, intraspecific competition, etc.) have in part sparked recent efforts to include trophic interactions in SDM projections (e.g., Zhang et al. 2022, Liu et al. 2023, Zabihi‐Seissan et al. 2024, Liu et al. 2025). Despite the influx of studies considering biotic interactions in SDMs, few studies have assessed how climate change projections including prey availability differ from those that do not (but see Khosravi et al. 2021). Such information is needed because if climate change is projected to decrease prey availability and/or create spatiotemporal mismatches between predator and prey, predators that rely on spatiotemporal overlap with prey may also endure negative changes (Carroll et al. 2024, Speakman et al. 2024). We show that for early life stages of pelagic predators, abiotic factor alone led to notable abundance increases and distribution shifts. However, when accounting for distribution shifts and abundance declines in zooplankton prey, abundance increases and distribution shifts of predators were far less substantial. One strength of our study is inclusion of taxon-specific predator-prey relationships. In the marine realm, bulk estimates of prey availability such as mesozooplankton biomass are often used for future projections of predators (e.g., Dueri et al. 2014, Woodworth-Jefcoats et al. 2017). However, these variables are often found to have less predictive power than abiotic variables, as predators may consume only a small portion of the taxa in such prey estimates (Siddon et al. 2011, Dueri et al. 2014). We found that prey availability can have similar, if not greater predictive skill than abiotic factors. This unique finding could be due to our efforts in matching individual predator and prey taxa based on the extensive larval tuna and billfish diet literature. Terrestrial studies have also found that taxon-specific predator-prey relationships can alter, if not improve model predictions, as was demonstrated with harpy eagles and three-toed sloths (Sutton et al. 2023) as well as ladybugs and aphids (Ge et al. 2024). We therefore recommend that when allowed by data and known diets, future SDM projections consider taxon-specific predator-prey relationships. While our results support the idea that prey availability aids projections of larval pelagic predators, the importance of prey availability varied among species. For example, prey availability was highly important for little tunny but relatively unimportant for billfish and other true tuna. Historical distributions showed that little tunny occupy more inshore waters at eastern and western Gulf edges, whereas billfish and other true tuna occupy more offshore waters, including the highly oligotrophic waters of the Loop Current. Therefore, inshore larvae may behave more like classic predators, associated with productive, inshore waters that contain higher prey concentrations, whereas offshore larvae may be more constrained by other factors (Boehlert et al. 1994, Davoren 2013). Mechanisms for this difference may relate to the trade-off between predation risk and food availability (Houston et al. 1993). While inshore waters may produce more food, they also may produce higher predator encounter rates (Maciej Gliwicz et al. 2006, Shropshire et al. 2022). Therefore, the risk of predation may not be worth the reward of food. Predation has indeed been shown as a primary factor limiting survival of late-stage tuna larvae (Shropshire et al. 2022). The relative weight of predation and starvation risks may drive our findings of some taxa having weaker spatiotemporal links with zooplankton prey. Although other factors including species specific energy requirements, onsets of piscivory, and adaptation to utilizing microbial food webs may also be prescriptive (Llopiz et al. 2010, Llopiz and Hobday 2015, Muhling et al. 2017). Our study is the first to document spatiotemporal links between larval tuna, billfish, and mahi-mahi and their preferred zooplankton prey. Of all zooplankton prey, larvaceans produced the highest partial DE values for all taxa other than billfish, for which cladocerans were the most important. Larvaceans were particularly important for skipjack tuna and little tunny. These results align with previous studies documenting larval skipjack tuna and little tunny diet being nearly exclusively comprised of larvaceans (Llopiz et al. 2010) and strong preferences by larval sailfish ( Istiophorus platypterus ) and blue marlin ( Makaira nigricans ) for cladocerans, including Evadne spp. (Llopiz et al 2008). The high relative importance of larvacean prey could be due to larvaceans having inferior escape capabilities compared to the other zooplankton prey taxa (Kodama et al. 2017, Landry et al. 2019). The relatively strong spatiotemporal links with larvacean prey could also be due to the ability of larvaceans to utilize microbial trophic pathways (Llopiz et al. 2010, Stukel et al. 2022). Larger phytoplankton such as diatoms may be relatively scarce compared to picophytoplankton in the offshore, oligotrophic waters of the Gulf (Selph et al. 2021, Stukel et al. 2022). Therefore, it may be a trophic advantage for larval fish to rely on zooplankton that utilize picophytoplankton, including larvaceans, rather than zooplankton such as calanoid and cyclopoid copepod that utilize nano and microphytoplankton (e.g., diatoms), and subsequently, food chains with more trophic steps (Landry et al. 2019, Stukel et al. 2022). This theory aligns with the Trophic Efficiency in Early Life (TEEL) hypothesis, which states that shorter and/or more efficient food chains increase larval survival and therefore recruitment (Swalethorp et al. 2023, Kwan et al. 2024). More broadly, our results support the notion that billfish and tuna larvae may be cladoceran and larvacean specialists, respectively (Kodama et al. 2017). Marine specialists may be particularly vulnerable to climate change relative to generalists, particularly those with specialized spawning migrations and areas, such as HMS (Bakun 2014). This highlights the need to better understand global change impacts on these species. Recent work has shown that climate change may affect habitat availability and distributions of adult HMS. For example, Braun et al. 2023 projected habitat loss and northward distribution shifts for 12 HMS in the western Atlantic, with billfish being an exception and encountering habitat gains. Additionally, Lezama-Ochoa et al. 2024 projected divergent responses for 10 HMS in the eastern Pacific, with some taxa including swordfish encountering habitat gains and northward distribution shifts. However, both studies did not include, yet identified the importance of, assessing changes in prey availability for projections. Zheng et al. 2023 projected distribution changes of Japanese Spanish mackerel and found that prey availability (anchovy) improved prediction skill. Additionally, Gleiber et al. 2024 highlighted the importance of prey use by albacore tuna undergoing climate-driven range shifts, and the role of prey traits. However, to our knowledge, few, if any, other studies have included prey availability in HMS projections, particularly at earlier life stages, although spatial ecosystem and population dynamics model (SEAPODYM) studies include a modeled food requirement index (e.g., Lehodey et al. 2008). In general, our results suggest that through direct climate change effects only, warming and increased salinities will increase abundances of larval tuna, billfish, and mahi-mahi, suggesting these taxa may be climate change ‘winners’ in the Gulf. Therefore, our results and those of Braun et al. 2023 suggest that both larval and adult billfish may be climate change ‘winners; in the western Atlantic. Furthermore, the results of our study, Braun et al. 2023 and Lezama-Ochoa et al. 2024 together suggest that most HMS on both sides of North America will encounter northward distributions shifts. Such shifts should be considered by fisheries management of these species. See Supporting Information for species-specific discussions of spatiotemporal trends, comparisons with previous studies and their implications for conservation and management. Our objectives and hypotheses were based on a bottom-up system where prey availability drives predator abundance and distribution, rather than predation driving prey abundance and distribution. This assumption comes from 1) knowledge of strong bottom-up influences in oligotrophic systems (Benoit-Bird and McManus 2012, Yoneda et al. 2022); 2) selective planktivory of larval tuna and billfish (Kodama et al. 2017, Landry et al. 2019); and 3) previous hypotheses emphasizing important roles of prey availability for larval tuna and billfish (Llopiz et al. 2015, Landry et al. 2019, Gleiber et al. 2020, and references therein). Although our correlative models cannot provide causal inference, the positive relationships between predator and prey distribution may be indicative of bottom-up effects (Worm and Myers 2003, Frederiksen et al. 2006) and at the very least indicate cooccurrence with suitable prey, which may lead to positive effects for growth and survival of predators, particularly under climate change (Carroll et al. 2024). We also recognize our predator-prey relationships may not be driven by trophic interactions. Shared habitat preferences, shared responses to omitted abiotic variables, and common movement behavior may also drive cooccurrence (Dormann et al. 2018). We particularly emphasize the roles of hydrodynamic convergence leading to cooccurrence of zooplankton and ichthyoplankton (Schmid et al. 2020) and active selection by adults to spawn in resource limited waters to minimize larval predation mortality (Muhling et al. 2017, Shropshire et al. 2022, and references therein). Future studies may consider empirical dynamic modeling and/or mechanistic models to better support causal predator-prey relationships. Our models also omit hydrodynamic processes (e.g., fronts, currents, eddies) known to affect Gulf epipelagic ichthyoplankton abundance and distribution (Rooker et al. 2012, Cornic and Rooker 2021). We note the importance of these features, however, global climate models are of insufficient spatial resolution to capture these processes, particularly in the Gulf (Kemp et al. 2016, Dee et al. 2019, Alexander et al. 2020). We used GAMs for our study, as they are easy to interpret, produce univariate taxa-taxa relationships, and offer ability to constrain biologically unrealistic responses and known environmental niches. For example, the thermal niche of larval Atlantic bluefin tuna in Gulf waters, based on larval occurrence and spawning activity, is roughly 24-28°C (Teo et al. 2007, Muhling et al. 2010). This allowed us to tune GAM parameters, creating relationships that somewhat aligned with this thermal niche. While GAMs were appropriate for our particular objectives, other statistical model types may have shown different responses (e.g., Brodie et al. 2022), or have parameterized species interactions in different ways (Norberg et al. 2019). Our approach of focusing on one statistical model type thus underestimates the uncertainty attributable to SDM type. Future work in the fast-paced SDM field may consider adapting our framework of including taxon-specific, predator-prey relationships within projections for other SDM approaches, including SDMs based on Bayesian structural equation models (Poggiato et al. 2025). We also acknowledge that the low variance explained by our models may lead to highly uncertain projections for some taxa. However, we emphasize that our goal was to assess the role of prey availability in distribution projections of larval pelagic predators and to assess uncertainty contributed by including prey fields in SDMs.

Despite their known ecological importance, taxon-specific measures of prey availability are not often used in SDMs, particularly when projecting redistributions due to climate change. Here, we hypothesized that biotic models including prey availability would produce similar predictive skill of larvae of key pelagic predators relative to abiotic models that include traditional, direct climate change impacts. We found support for this hypothesis for most investigated larval taxa. We also hypothesized that while biotic and abiotic models may yield similar predictive skill, they would produce different projected shifts in distribution and abundance. We also found support for this hypothesis across most taxa, as changes in prey availability dampened projected distribution and abundance shifts relative to projections from climate variables. Our models showed that abundances of zooplankton prey were generally projected to decrease, although prey decreases did not lead to decreases of their larval tuna, billfish and mahi-mahi predators. Rather, direct climate change effects including ocean warming, led to substantial projected increases in abundance and northward distribution shifts for most larval taxa. We highlight the importance and role of prey availability in subtropical, resource limited systems, particularly for early life stages of economically and ecologically important species. As such, our results may inform strategic fisheries management of these species and warrant consideration for incorporation in ecosystem models. We acknowledge that data sets containing simultaneously measured abundances and/or frequencies of occurrence of individual predator and prey taxa are not common. However, when they are available, or when ecologists can appropriately synthesize multiple data sets, we encourage future work to consider suites of abiotic and biotic predictors when projecting redistributions under climate change, as our results clearly show that the inclusion and/or exclusion of such predictors affects the direction and magnitude of distribution and abundance shifts. Table 1: For each ichthyoplankton and zooplankton taxa, the biotic (top) and abiotic (bottom) models with the best predictive skill are shown (abiotic only for zooplankton). Predictive skill is based on Leave-One-Out (LOO) R 2 values (see methods for further details). Included predictor variables and deviance explained (DE) values are reported for each model. See Supporting Information for all candidate models. | Calanoids | Abiotic | Surface Temp, Temp at Depth, Salinity, Chlorophyll, Time of Day, Lat Lon | 35.8 | 0.35 | | Cladocerans | Abiotic | Surface Temp, Salinity, Time of Day, Lat Lon | 43.7 | 0.20 | | Cyclopoids | Abiotic | Surface Temp, Temp at Depth, Salinity, Chlorophyll, Lat Lon | 21.6 | 0.17 | | Larvaceans | Abiotic | Surface Temp, Temp at Depth, Salinity, Time of Day, Lat Lon | 22.0 | 0.20 | | Other True Tuna | Biotic | Cyclopoids, Larvaceans | 5.4 | 0.08 | | Abiotic | Surface Temp, Salinity, Temp at Depth, Chlorophyll, Lat Lon | 16.2 | 0.11 | | | Atlantic Bluefin Tuna | Biotic | Cladocerans, Larvaceans, Lat Lon | 28.6 | 0.10 | | Abiotic | Surface Temp, Salinity, Lat Lon | 30.9 | 0.08 | | | Skipjack Tuna | Biotic | Larvaceans, Lat Lon | 11.0 | 0.12 | | Abiotic | Surface Temp, Temp at Depth, Salinity, Chlorophyll, Lat Lon | 31.2 | 0.19 | | | Billfish | Biotic | Cladocerans, Larvaceans, Lat Lon | 22.8 | 0.13 | | Abiotic | Surface Temp, Temp at Depth, Lat Lon | 42.4 | 0.43 | | | Little Tunny | Biotic | Calanoids, Cyclopoids, Larvaceans, Lat Lon | 32.2 | 0.19 | | Abiotic | Surface Temp, Temp at Depth, Lat Lon | 24.8 | 0.14 | | | Mahi-Mahi | Biotic | Cladocerans, Larvaceans, Lat Lon | 10.7 | 0.06 | | Abiotic | Surface Temp, Temp at Depth, Salinity, Lat Lon | 10.7 | 0.08 | | | Frigate Tuna | Biotic | Calanoids, Cladocerans, Larvaceans, Lat Lon | 23.1 | 0.20 | | Abiotic | Surface Temp, Temp at Depth, Salinity, Chlorophyll, Lat Lon | 31.2 | 0.19 | Figure 1: Conceptual diagram displaying models linking climate variables, zooplankton and ichthyoplankton. Zooplankton taxa include Calanoids, Cyclopoids, Cladocerans and Larvaceans. Ichthyoplankton taxa include Atlantic Bluefin Tuna, Other True Tuna, Skipjack Tuna, Frigate Tuna, Little Tunny, Billfish and Mahi-Mahi. Different colored circles denote different data groups. ‘Zooplankton Models’ refer to predicting zooplankton abundance from climate variable data. ‘Abiotic Models’ refer to predicting ichthyoplankton abundance from climate variable data. ‘Biotic Models’ refer to predicting ichthyoplankton from zooplankton abundance data. When forced with Earth System Model output (i.e., climate variable projections), zooplankton models and abiotic models include direct effects of climate change. However, biotic models include indirect effects of climate change, as with biotic models, zooplankton projections are used as input. Figure 2: Projected spring abundances for Larvaceans (A, B), Cyclopoids (C,D), Calanoids (E,F) and Cladocerans (G,H) for historical (1985-2014, left) and future (2070-2099, right) periods. Projections include generalized additive model abundances (individuals under 1m 2 of sea surface). Figure 3: Vector plot displaying displacement distance (km) and cardinal direction changes in geographic center for abundance for each plankton taxa and model type for ichthyoplankton (biotic or abiotic). Blue shades denote ichthyoplankton predicted directly from climate variables (i.e., abiotic models), yellow shades denote ichthyoplankton predicted from zooplankton projections (i.e., biotic models), and green shades denote zooplankton predicted directly from climate variables. Only taxa with displacement distances greater than 20 km are shown for increased visual interpretability. See Supporting Information for exact displacement distances and directions for each taxa and model. *Atl. = Atlantic. Figure 4: Percent change in abundance (individuals under 1m 2 of sea surface) of zooplankton and ichthyoplankton predicted from generalized additive models. Changes are based on differences in historical (1985-2014) and future (2070-2099) climate conditions. Larger, colored circles denote mean values. Blue denotes ichthyoplankton predicted directly from climate variables (i.e., abiotic models), yellow denotes ichthyoplankton predicted from zooplankton projections (i.e., biotic models), and green denotes zooplankton predicted directly from climate variables. Mean values for four models (e.g., Billfish (Abiotic) are inflated due to very low predicted abundances values and are therefore not visible on this scale. See Supporting Information for values of such models. *Atl. = Atlantic. Figure 5: Spring projections from biotic models (left) and abiotic models (right) for two 30-year time periods (historical and future) for larval Other True Tuna (A), Atlantic Bluefin Tuna (B), Skipjack Tuna (C), Billfish (D), Little Tunny (E), Mahi-Mahi (F) and Frigate Tuna (G). Projections include generalized additive model abundances (individuals under 1m 2 of sea surface). *Atl. = Atlantic. Figure 5 (continued): Spring projections from biotic models (left) and abiotic models (right) for two 30-year time periods (historical and future) or larval Other True Tuna (A), Atlantic Bluefin Tuna (B), Skipjack Tuna (C), Billfish (D), Little Tunny (E), Mahi-Mahi (F) and Frigate Tuna (G). Projections include generalized additive model abundances (individuals under 1m 2 of sea surface).

Bakun A. Active opportunist species as potential diagnostic markers for comparative tracking of complex marine ecosystem responses to global trends. ICES Journal of Marine Science. 2014 Oct 1;71(8):2281-92. Barnes CL, Essington TE, Pirtle JL, Rooper CN, Laman EA, Holsman KK, Aydin KY, Thorson JT. Climate‐informed models benefit hindcasting but present challenges when forecasting species–habitat associations. Ecography. 2022 Oct;2022(10):e06189. Bartoń K. MuMIn: Multi-Model Inference. 2023, R package. Beaumont LJ, Pitman AJ, Poulsen M, Hughes L. Where will species go? Incorporating new advances in climate modelling into projections of species distributions. Global Change Biology. 2007 Jul;13(7):1368-85. Benoit-Bird KJ, McManus MA. Bottom-up regulation of a pelagic community through spatial aggregations. Biology Letters. 2012 Oct 23;8(5):813-6. Block BA, Teo SL, Walli A, Boustany A, Stokesbury MJ, Farwell CJ, Weng KC, Dewar H, Williams TD. Electronic tagging and population structure of Atlantic bluefin tuna. Nature. 2005 Apr 28;434(7037):1121-7. Boehlert GW, Mundy BC. Vertical and onshore-offshore distributional patterns of tuna larvae in relation to physical habitat features. Marine Ecology Progress Series. 1994 Apr 21:1-3. Braun CD, Lezama-Ochoa N, Farchadi N, Arostegui MC, Alexander M, Allyn A, Bograd SJ, Brodie S, Crear DP, Curtis TH, Hazen EL. Widespread habitat loss and redistribution of marine top predators in a changing ocean. Science Advances. 2023 Aug 9;9(32):eadi2718. Brodie S, Smith JA, Muhling BA, Barnett LA, Carroll G, Fiedler P, Bograd SJ, Hazen EL, Jacox MG, Andrews KS, Barnes CL. Recommendations for quantifying and reducing uncertainty in climate projections of species distributions. Global Change Biology. 2022 Nov;28(22):6586-601. Carroll G, Abrahms B, Brodie S, Cimino MA. Spatial match–mismatch between predators and prey under climate change. Nature Ecology & Evolution. 2024 Sep;8(9):1593-601. Collette BB, Cole K. Reproduction and development in epipelagic fishes. Reproduction and sexuality in marine fishes: Patterns and processes. 2010:21-63. Cornic M, Rooker JR. Temporal shifts in the abundance and preferred habitats of yellowfin and bigeye tuna larvae in the Gulf of Mexico. Journal of Marine Systems. 2021 May 1;217:103524. Cushing DH. Plankton production and year-class strength in fish populations: an update of the match/mismatch hypothesis. InAdvances in marine biology 1990 Jan 1 (Vol. 26, pp. 249-293). Academic Press. Davoren GK. Distribution of marine predator hotspots explained by persistent areas of prey. Marine Biology. 2013 Dec;160:3043-58. Dell’Apa A, Boenish R, Fujita R, Kleisner K. Effects of climate change and variability on large pelagic fish in the Northwest Atlantic Ocean: implications for improving climate resilient management for pelagic longline fisheries. Frontiers in Marine Science. 2023 Sep 20;10:1206911. Dormann CF, Bobrowski M, Dehling DM, Harris DJ, Hartig F, Lischke H, Moretti MD, Pagel J, Pinkert S, Schleuning M, Schmidt SI. Biotic interactions in species distribution modelling: 10 questions to guide interpretation and avoid false conclusions. Global Ecology and Biogeography. 2018 Sep;27(9):1004-16. Dueri S, Bopp L, Maury O. Projecting the impacts of climate change on skipjack tuna abundance and spatial distribution. Global Change Biology. 2014 Mar;20(3):742-53. Frederiksen M, Edwards M, Richardson AJ, Halliday NC, Wanless S. From plankton to top predators: bottom‐up control of a marine food web across four trophic levels. Journal of Animal Ecology. 2006 Nov;75(6):1259-68. Ge X, Griswold CK, Newman JA. Robust species distribution predictions of predator and prey responses to climate change. Journal of Biogeography. 2024. Gleiber MR, Sponaugle S, Cowen RK. Some like it hot, hungry tunas do not! Implications of temperature and plankton food web dynamics on growth and diet of tropical tuna larvae. ICES Journal of Marine Science. 2020 Dec;77(7-8):3058-73. Gleiber MR, Hardy NA, Morganson CJ, Nickels CF, Muhling BA, Portner EJ, Wells BK, Brodeur RD, Auth TD, Santora JA, Glaser SM. Trait-based indicators of resource selection by albacore tuna in the California Current Large Marine Ecosystem. Ecological Indicators. 2024 Jan 1;158:111473. Godsoe W, Murray R, Plank MJ. Information on biotic interactions improves transferability of distribution models. The American Naturalist. 2015 Feb 1;185(2):281-90. Habtes S, Muller‐Karger FE, Roffer MA, Lamkin JT, Muhling BA. A comparison of sampling methods for larvae of medium and large epipelagic fish species during spring SEAMAP ichthyoplankton surveys in the Gulf of Mexico. Limnology and Oceanography: Methods. 2014 Feb;12(2):86-101. Hijmans R. geosphere: Spherical Trigonometry. 2022, R package. Hinrichsen HH, Möllmann C, Voss R, Köster FW, Kornilovs G. Biophysical modeling of larval Baltic cod ( Gadus morhua ) growth and survival. Canadian Journal of Fisheries and Aquatic Sciences. 2002 Dec 1;59(12):1858-73. Hjort J. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Conseil Permanent International pour l’Exploration de la Mer. 2014; 20:1-228. Ho CK, Stephenson DB, Collins M, Ferro CA, Brown SJ. Calibration strategies: a source of additional uncertainty in climate change projections. Bulletin of the American Meteorological Society. 2012 Jan 1;93(1):21-6. Houston AI, McNamara JM, Hutchinson JM. General results concerning the trade-off between gaining energy and avoiding predation. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences. 1993 Sep 29;341(1298):375-97. Ingram Jr GW, Richards WJ, Lamkin JT, Muhling B. Annual indices of Atlantic bluefin tuna ( Thunnus thynnus ) larvae in the Gulf of Mexico developed using delta-lognormal and multivariate models. Aquatic Living Resources. 2010 Jan 1;23(1):35-47. Johnson BR, Wallace JB. Bottom-up limitation of a stream salamander in a detritus-based food web. Canadian Journal of Fisheries and Aquatic Sciences. 2005 Feb 1;62(2):301-11. Kemp GP, Day Jr JW, Yáñez-Arancibia A, Peyronnin NS. Can continental shelf river plumes in the northern and southern Gulf of Mexico promote ecological resilience in a time of climate change?. Water. 2016 Mar 4;8(3):83. Khosravi R, Hemami MR, Malakoutikhah S, Ashrafzadeh MR, Cushman SA. Prey availability modulates predicted range contraction of two large felids in response to changing climate. Biological Conservation. 2021 Mar 1;255:109018. Kodama T, Hirai J, Tamura S, Takahashi T, Tanaka Y, Ishihara T, Tawa A, Morimoto H, Ohshimo S. Diet composition and feeding habits of larval Pacific bluefin tuna Thunnus orientalis in the Sea of Japan: integrated morphological and metagenetic analysis. Marine Ecology Progress Series. 2017 Nov 16;583:211-26. Kwan GT, Walsh KA, Thompson AR, Ben‐Aderet NJ, Fennie HW, Semmens BX, Swalethorp R. Trophic level influences larval Shortbelly Rockfish development. Marine and Coastal Fisheries. 2024 Dec 1;16(6):e10319. Lai SY, Pálsson A, Guðbergsson G, Jónsson IR, Ólafsson JS, Bárðarson H. The prey availability and diet of juvenile Atlantic salmon ( Salmo salar L.) in low‐productivity rivers in northern Europe. Journal of Fish Biology. 2024 Apr 17. Landry MR, Beckley LE, Muhling BA. Climate sensitivities and uncertainties in food-web pathways supporting larval bluefin tuna in subtropical oligotrophic oceans. ICES Journal of Marine Science. 2019 Mar;76(2):359-69. Lawlor JA, Comte L, Grenouillet G, Lenoir J, Baecher JA, Bandara RM, Bertrand R, Chen IC, Diamond SE, Lancaster LT, Moore N. Mechanisms, detection and impacts of species redistributions under climate change. Nature Reviews Earth & Environment. 2024 Apr 18:1-8. Lehodey P, Senina I, Murtugudde R. A spatial ecosystem and populations dynamics model (SEAPODYM)–Modeling of tuna and tuna-like populations. Progress in Oceanography. 2008 Sep 1;78(4):304-18. Lezama‐Ochoa N, Brodie S, Welch H, Jacox MG, Pozo Buil M, Fiechter J, Cimino M, Muhling B, Dewar H, Becker EA, Forney KA. Divergent responses of highly migratory species to climate change in the California Current. Diversity and Distributions. 2024 Feb;30(2):e13800. Liu S, Tian Y, Liu Y, Alabia ID, Cheng J, Ito SI. Development of a prey-predator species distribution model for a large piscivorous fish: a case study for Japanese Spanish mackerel Scomberomorus niphonius and Japanese anchovy Engraulis japonicus . Deep Sea Research Part II: Topical Studies in Oceanography. 2023 Feb 1;207:105227. Liu OR, Kaplan IC, Hernvann PY, Fulton EA, Haltuch MA, Harvey CJ, Marshall KN, Muhling B, Norman K, Pozo Buil M, Rovellini A. Climate change influences via species distribution shifts and century‐scale warming in an end‐to‐end California current ecosystem model. Global Change Biology. 2025 Jan;31(1):e70021. Llopiz JK, Cowen RK. Precocious, selective and successful feeding of larval billfishes in the oceanic Straits of Florida. Marine Ecology Progress Series. 2008 Apr 21;358:231-44. Llopiz JK, Richardson DE, Shiroza A, Smith SL, Cowen RK. Distinctions in the diets and distributions of larval tunas and the important role of appendicularians. Limnology and Oceanography. 2010 May;55(3):983-96. Llopiz JK, Hobday AJ. A global comparative analysis of the feeding dynamics and environmental conditions of larval tunas, mackerels, and billfishes. Deep Sea Research Part II: Topical Studies in Oceanography. 2015 Mar 1;113:113-24. Maciej Gliwicz Z, Slon J, Szynkarczyk I. Trading safety for food: evidence from gut contents in roach and bleak captured at different distances offshore from their daytime littoral refuge. Freshwater Biology. 2006 May;51(5):823-39. Malca E, Shropshire T, Landry MR, Quintanilla JM, Laiz-CarriÓn R, Shiroza A, Stukel MR, Lamkin J, Gerard T, Swalethorp R. Influence of food quality on larval growth of Atlantic bluefin tuna ( Thunnus thynnus ) in the Gulf of Mexico. Journal of Plankton Research. 2022 Sep;44(5):747-62. Marra G, Wood SN. Practical variable selection for generalized additive models. Computational Statistics & Data Analysis. 2011 Jul 1;55(7):2372-87. Muhling BA, Lamkin JT, Roffer MA. Predicting the occurrence of Atlantic bluefin tuna ( Thunnus thynnus ) larvae in the northern Gulf of Mexico: building a classification model from archival data. Fisheries Oceanography. 2010 Nov;19(6):526-39. Muhling BA, Liu Y, Lee SK, Lamkin JT, Roffer MA, Muller-Karger F, Walter III JF. Potential impact of climate change on the Intra-Americas Sea: Part 2. Implications for Atlantic bluefin tuna and skipjack tuna adult and larval habitats. Journal of Marine Systems. 2015 Aug 1;148:1-3. Muhling BA, Lamkin JT, Alemany F, García A, Farley J, Ingram GW, Berastegui DA, Reglero P, Carrion RL. Reproduction and larval biology in tunas, and the importance of restricted area spawning grounds. Reviews in Fish Biology and Fisheries. 2017 Dec;27:697-732. Norberg A, Abrego N, Blanchet FG, Adler FR, Anderson BJ, Anttila J, Araújo MB, Dallas T, Dunson D, Elith J, Foster SD. A comprehensive evaluation of predictive performance of 33 species distribution models at species and community levels. Ecological Monographs. 2019 Aug;89(3):e01370. Patel J, Sharma M, Khanyari M, Bijoor A, Mishra C, Harihar A, Suryawanshi K. Influence of predator suppression and prey availability on carnivore occurrence in western Himalaya. Journal of Zoology. 2024 Jan;322(1):3-11. Poggiato G, Andréoletti J, Pollock LJ, Thuiller W. Integrating food webs in species distribution models can improve ecological niche estimation and predictions. Ecography. 2025 Jan:e07546. Pinsky ML, Worm B, Fogarty MJ, Sarmiento JL, Levin SA. Marine taxa track local climate velocities. Science. 2013 Sep 13;341(6151):1239-42. Poloczanska ES, Burrows MT, Brown CJ, García Molinos J, Halpern BS, Hoegh-Guldberg O, Kappel CV, Moore PJ, Richardson AJ, Schoeman DS, Sydeman WJ. Responses of marine organisms to climate change across oceans. Frontiers in Marine Science. 2016 May 4;3:62. Quintanilla JM, Borrego-Santos R, Malca E, Swalethorp R, Landry MR, Gerard T, Lamkin J, García A, Laiz-Carrión R. Maternal effects and trophodynamics drive interannual larval growth variability of Atlantic bluefin tuna ( Thunnus thynnus ) from the Gulf of Mexico. Animals. 2024 Apr 28;14(9):1319. R Core Team. R: A Language and Environment for Statistical Computing. 2023, R Foundation for Statistical Computing, Vienna, Austria Rooker JR, Simms JR, Wells RD, Holt SA, Holt GJ, Graves JE, Furey NB. Distribution and habitat associations of billfish and swordfish larvae across mesoscale features in the Gulf of Mexico. PloS one. 2012 Apr 11;7(4):e34180. Scott GP, Turner SC, Churchill GB, Richards WJ, Brothers EB. Indices of larval bluefin tuna, Thunnus thynnus, abundance in the Gulf of Mexico; modelling variability in growth, mortality, and gear selectivity. Bulletin of Marine Science. 1993 Aug 1;53(2):912-29. Selph K, Swalethorp R, R Stukel M, B Kelly T, N Knapp A, Fleming K, Hernandez T, R Landry M. Phytoplankton community composition and biomass in the oligotrophic Gulf of Mexico. Journal of Plankton Research. 2022 Sep;44(5):618-37. Shiroza A, Malca E, Lamkin JT, Gerard T, Landry MR, Stukel MR, Laiz-Carrión R, Swalethorp R. Active prey selection in developing larvae of Atlantic bluefin tuna ( Thunnus thynnus ) in spawning grounds of the Gulf of Mexico. Journal of Plankton Research. 2022 Sep;44(5):728-46. Shropshire TA, Morey SL, Chassignet EP, Karnauskas M, Coles VJ, Malca E, Laiz-Carrión R, Fiksen Ø, Reglero P, Shiroza A, Quintanilla Hervas JM. Trade-offs between risks of predation and starvation in larvae make the shelf break an optimal spawning location for Atlantic Bluefin tuna. Journal of Plankton Research. 2022 Sep;44(5):782-98. Siddon EC, Duffy-Anderson JT, Mueter FJ. Community-level response of fish larvae to environmental variability in the southeastern Bering Sea. Marine Ecology Progress Series. 2011 Mar 28;426:225-39. Speakman CN, McHuron EA, Arnould JP. Vulnerability to fluctuations in prey and predation landscape in a central place foraging marine predator. Frontiers in Marine Science. 2024 Mar 27;11:1367655. Stukel MR, Gerard T, Kelly TB, Knapp AN, Laiz-Carrión R, Lamkin JT, Landry MR, Malca E, Selph KE, Shiroza A, Shropshire TA. Plankton food webs in the oligotrophic Gulf of Mexico spawning grounds of Atlantic bluefin tuna. Journal of Plankton Research. 2022 Sep;44(5):763-81. Sutton LJ, Anderson DL, Franco M, McClure CJ, Miranda EB, Vargas FH, Vargas González JD, Puschendorf R. Prey resources are equally important as climatic conditions for predicting the distribution of a broad‐ranged apex predator. Diversity and Distributions. 2023 May;29(5):613-28. Swalethorp R, Landry MR, Semmens BX, Ohman MD, Aluwihare L, Chargualaf D, Thompson AR. Anchovy boom and bust linked to trophic shifts in larval diet. Nature Communications. 2023 Dec 5;14(1):7412. Teo SL, Boustany AM, Block BA. Oceanographic preferences of Atlantic bluefin tuna, Thunnus thynnus, on their Gulf of Mexico breeding grounds. Marine Biology. 2007 Oct;152:1105-19. Trenkel VM, Huse G, MacKenzie BR, Alvarez P, Arrizabalaga H, Castonguay M, Goñi N, Grégoire F, Hátún H, Jansen T, Jacobsen JA. Comparative ecology of widely distributed pelagic fish species in the North Atlantic: implications for modelling climate and fisheries impacts. Progress in Oceanography. 2014 Dec 1;129:219-43. Van der Putten WH, Macel M, Visser ME. Predicting species distribution and abundance responses to climate change: why it is essential to include biotic interactions across trophic levels. Philosophical Transactions of the Royal Society B: Biological Sciences. 2010 Jul 12;365(1549):2025-34. Wickham H. ggplot2: Elegant Graphics for Data Analysis. 2016 Springer-Verlag New York. R package. Wood SN. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. 2011. Journal of the Royal Statistical Society (B), 73(1), 3-36. Woodworth‐Jefcoats PA, Polovina JJ, Drazen JC. Climate change is projected to reduce carrying capacity and redistribute species richness in North Pacific pelagic marine ecosystems. Global Change Biology. 2017 Mar;23(3):1000-8. Worm B, Myers RA. Meta‐analysis of cod–shrimp interactions reveals top‐down control in oceanic food webs. Ecology. 2003 Jan;84(1):162-73. Yoneda M, Fujita T, Yamamoto M, Tadokoro K, Okazaki Y, Nakamura M, Takahashi M, Kono N, Matsubara T, Abo K, Xinyu G. Bottom-up processes drive reproductive success of Japanese anchovy in an oligotrophic sea: A case study in the central Seto Inland Sea, Japan. Progress in Oceanography. 2022 Aug 1;206:102860. Zabihi‐Seissan S, Baker KD, Stanley RR, Tunney TD, Beauchamp B, Benoit HP, Brickman D, Chabot D, Cook A, Deslauriers D, Koen‐Alonso M. Interactive effects of predation and climate on the distributions of marine shellfish in the Northwest Atlantic. Oikos. 2024 May:e10524. Zhang Y, Zhang C, Xu B, Ji Y, Ren Y, Xue Y. Impacts of trophic interactions on the prediction of spatio-temporal distribution of mid-trophic level fishes. Ecological Indicators. 2022 May 1;138:10882 Information & Authors Information Version history Peer review timeline Published Ecography Version of Record8 Sep 2025Published Copyright This work is licensed under a Non Exclusive No Reuse License. Collection

Authors Metrics & Citations Metrics Article Usage 376views 275downloads Citations Download citation Raymond Czaja, Barbara Muhling, ESTRELLA MALCA, et al. Direct effects and prey-mediated effects of global change in projections of early life stages of pelagic predators. Authorea. 11 July 2025. DOI: https://doi.org/10.22541/au.175222759.98154210/v1 DOI: https://doi.org/10.22541/au.175222759.98154210/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.