Introduction

Climate change is recognised as a growing threat to commercial fisheries around the world (Cheung et al. 2010; Cochrane et al. 2009; Free et al. 2019; Worm and Lotze 2021). In particular, rising ocean temperatures (Hoegh-Guldberg and Bruno 2010), ocean acidification (Doney et al. 2009), reduced oxygen levels (Breitburg et al. 2018), and shifts in oceanic currents (Hays 2017) pose a direct threat to many commercially important species. These risks are most pronounced in ocean warming hotspots, such as south-eastern Australia (Hobday and Lough 2011), a region prone to marine heatwaves (Oliver et al. 2017, 2018; Roberts et al. 2019), and where sea surface temperatures (SSTs) are rising at four times the global average (Frusher et al. 2014; Hobday and Pecl 2014). Here, fisheries and aquaculture are already showing signs of climate stress (Holbrook and Johnson 2014; Pecl et al. 2014; Wernberg et al. 2011), with declines in productivity linked to direct physiological stress (Martino et al. 2019; McLeay et al. 2019; Wade et al. 2019), distributional shifts (Champion et al. 2018; Gervais et al. 2021a; Ramos et al. 2018; Robinson et al. 2015), habitat modifications (Johnson et al. 2011; Ling 2008) and disruption of trophic interactions (Holland et al. 2021; Ling et al. 2009; Vergés et al. 2016). Impacts of climate change are expected to intensify over the coming decades (IPCC 2021; Pereira et al. 2010), causing significant concern for industry viability, threatening food security, human health, and socio-economic values (Rice and Garcia 2011; Sumaila et al. 2011; Vianna et al. 2020). Future adaptive management of commercial fisheries and aquaculture in south-eastern Australia, and other regions of the world impacted by climate change hinges on our ability to predict how these industries will respond to future climate challenges and interventions capable of enhancing fisheries resilience (Bryndum-Buchholz et al. 2021).

South-eastern Australia is home to the world’s largest wild-capture abalone fishery, primarily targeting blacklip abalone (Haliotis rubra), as well as a rapidly expanding aquaculture industry which predominantly farms greenlip (H. laevigata) and hybrid abalone (H. rubra × H. laevigata; Cook 2019; Mayfield et al. 2012). Combined annual harvests exceed 3000 tonnes valued at approximately AUS$120 million (Tuynman and Dylewski 2022). Some wild abalone fisheries in the region have suffered significant declines in recent decades, due to overexploitation (Gorfine et al. 2001), disease (Mayfield et al. 2011) and invasive urchin species (Ling 2008). However, recent evidence now points to declining wild abalone stocks independent of fishing pressure, disease exposure and urchin activity, with ocean warming implicated as an additional driver of biomass reductions (Young et al. 2020). A recent study demonstrated that ocean warming is likely to be decoupling critical trophic interactions supporting the wild H. rubra fisheries, causing reductions in abalone biomass (Holland et al. 2021). However, evidence also suggests that ocean warming is likely to be directly impacting the health and development of H. rubra and contributing to declines in some fisheries (Dang et al. 2012; Morash and Alter 2016; Young et al. 2020). Farm fisheries are also expected to become increasingly threatened by ocean warming, due to their dependency on coastal water inputs, leading to reduced growth rates and elevated risks of mortality and disease outbreaks (Hooper et al. 2014; Morash and Alter 2016; Stone et al. 2014).

Abalone are poikilothermic and thermally sensitive marine organisms (Morash and Alter 2016). Elevated SSTs have been shown to compromise all life stages, supressing larval development (McCormick et al. 2016), delaying gonad maturation (Sukhan et al. 2021), reducing fecundity (Rogers-Bennett et al. 2010), stunting growth (Vilchis et al. 2005), and increasing susceptibility to disease (Dang et al. 2012; Raimondi et al. 2002). Extreme marine heat wave events have decimated abalone stocks in northern California (Rogers-Bennett and Catton 2019), and in Western Australia (Caputi et al. 2019; Kajtar et al. 2021; Pearce and Feng 2013), and at least one marine heatwave event has been linked to declines in H. rubra stocks in Tasmania (Oliver et al. 2017). Parts of south-eastern Australia are projected to experience an increase in warming of 1.5–3 °C by the year 2070, depending on the severity of ongoing carbon emissions (Hobday and Lough 2011). Understandably, fisheries managers and industry stakeholders in south-eastern Australia are concerned about these projections and future threats to the abalone industry, as well as knowledge gaps that currently hinder adaptive management.

Across the animal kingdom, early life-history stages (gametes and larvae), and larger, mature individuals are thought to be more vulnerable to thermal stress (Clark et al. 2012; Klockmann et al. 2017; Leiva et al. 2019; Peck et al. 2009). The mechanisms underpinning these ontogenetic patterns are not clear, but have previously been attributed to size-dependent oxygen transport limits (Clark et al. 2008; Peck et al. 2004, 2013). Evidence suggests that these patterns are strongest in aquatic species (Horne et al. 2015; Pinsky et al. 2019), with several studies demonstrating strong linear relationships between body size and thermal stress tolerance in a range of marine vertebrates and invertebrates (Clark et al. 2017; Di Santo and Lobel 2017; Illing et al. 2020; Messmer et al. 2017; Peck et al. 2013). Body size also generally correlates with reproductive output in fishes, with smaller individuals often producing fewer or smaller offspring (Barneche et al. 2018; Clark et al. 2013a, b). Consequently, it has been suggested that ocean warming may impact the future productivity of some fisheries around the world by selecting for smaller and less fecund individuals (Audzijonyte et al. 2013; Baudron et al. 2014; van Rijn et al. 2017). However, it is not well understood if such allometric patterns apply to marine invertebrates such as abalone, where studies of this nature remain limited (Dı́az et al. 2000; Gilroy and Edwards 1998; Hecht 1994). This highlights the need for further research geared toward improving our understanding of potential climate change effects on marine invertebrates, specifically those of commercial importance in climate change ‘hotspots’ such as south-eastern Australia.

Whilst studies suggest that patterns of vulnerability across developmental life stages persist across species distributions (Daufresne et al. 2009; Leiva et al. 2019), critical thermal limits among populations from contrasting thermal environments can differ significantly due to adaptive genetic differences (Eliason et al. 2011; Howells et al. 2012; Newton et al. 2010). Species with wide latitudinal ranges often exhibit genetically determined clines across thermal gradients (Aitken and Bemmels 2016; Halbritter et al. 2018; Jeffery et al. 2017; Miller et al. 2020). Abalone species from south-eastern Australia are widely distributed spanning strong thermal gradients, particularly H.rubra which has a latitudinal range of over 1500 km where annual average SST range from 14 to 22 °C (Huang et al. 2017), suggesting that adaptive differences among abalone populations may exist. Previous population genomic studies have demonstrated adaptive genetic structure relating to temperature in H. rubra (Miller et al. 2019) and H. laevigata (Sandoval-Castillo et al. 2018), as well as haliotids from the northern hemisphere (De Wit and Palumbi 2013; Yu et al. 2021a). However, functional validation of adaptive genetic differences among populations from contrasting thermal environments has only been demonstrated in a single haliotid from China, which has since been further characterised using transcriptomic and metabolomic data (H. discus hannai; Yu et al. 2021a; b, 2023). Similar quantitative experiments are needed to provide functional validation of adaptive genetic differences among Australian abalone relating to temperature. Such studies are critical for identifying vulnerable wild and farm fishing stocks (Gervais et al. 2021a, b), understanding the potential for abalone stocks to evolve and adapt to future environments (Miller et al. 2020; Waldvogel et al. 2020), and identifying temperature tolerant genotypes that can be used to enhance the resilience of both wild and farm stocks through strategic translocation activities and breeding programs (Turko et al. 2021).

In this study, controlled trials were used to test the effects of body size and provenance (two farm sites) on maximal thermal limits of Australian farmed hybrid abalone. Specifically, critical thermal maximum (CTmax) was measured in abalone from different life stages and thermal environments, using both acute (1 °C per h) and chronic (1 °C per 12 h) heating rates. The inclusion of two heating rates is uncommon in studies of this nature (Dı́az et al. 2000; Gilroy and Edwards 1998; Searle et al. 2006), but was deemed important for understanding differential impacts of acute versus chronic thermal stress on abalone resilience. This study was limited to the use of farmed hybrid abalone, which may not accurately reflect the thermal maxima of wild, pure-bred animals. However, this approach provides a necessary first step into investigating the effects of size and provenance on the critical thermal maxima of Australian abalone broadly, as well as a baseline for how farmed animals may respond to ongoing ocean warming. Based on the findings of previous correlative (De Wit and Palumbi 2013; Miller et al. 2019; Sandoval-Castillo et al. 2018; Yu et al. 2021a) and quantitative (Yu et al. 2021b) genetic research, we hypothesise that significant differences in CTmax will be observed between abalone from different thermal environments (farm sites). Furthermore, given the existing knowledge from the literature, and anecdotal reports from abalone farms reporting higher mortalities in larger size classes following marine heatwave events, we hypothesise that a negative linear relationship between body size and CTmax will be observed at both heating rates.

Methods

Animal selection and collection

Experimental trials were performed on farmed hybrid abalone (female H. rubra × male H. laevigata) from south-eastern Australia. Hybrid abalone exhibit heterosis (also known as ‘hybrid vigour’) providing growth and behavioural advantages and possibly enhanced resilience to thermal stress (Alter et al. 2017; Li 2008; Tripp-Valdez et al. 2021; Xiao et al. 2022), although they are rarely encountered in the wild (Brown 1995). While conducting experimental trials on pure line animals sourced from the wild would help to maximise the translation of research findings to natural settings, replicated sampling of wild specimens across developmental life stages is logistically difficult due to the cryptic nature of juveniles (Prince et al. 1988), and time constraints involved with the project.

A total of 324 hybrid abalone were sourced from two aquaculture farms from south-eastern Australia, with 162 animals collected on October 25th 2021 for an acute thermal stress trial and an additional 162 animals collected on November 25th 2021 for a chronic thermal stress trial. The two farms were Southern Ocean Mariculture (SOM) from Port Fairy (38° 23′ 25.5″ S, 142° 10′ 43.5″ E) and Jade Tiger Abalone (JTA) from Avalon (38° 4′ 50.4″ S, 144° 27′ 18.2″ E), and were selected based on their contrasting thermal environments. The average summer sea surface temperature (SST) and average annual range recorded at SOM (summer mean = 19.11 °C, annual range = 11.53 °C) are considerably lower than those from JTA (summer mean = 22.51 °C, annual range = 17.371 °C), providing a strong thermal gradient for experimental tests of provenance effects on maximum thermal limits. Each experimental trial (acute and chronic thermal stress conditions) used a gradient of animal size classes to investigate the effect of body size and ontogeny on thermal tolerance. Eighteen individuals were selected from each of four size/age classes from each farm (wet weight and age: 10–20 g, ~ 1.0–1.5 years; 40–50 g, ~ 2.0 years; 70–80 g, ~ 2.5–3.0 years; 100–110 g, ~ 3.5 years), while a fifth size/age class was included from SOM (200–210 g, ~ 4 years), providing a total of 72 animals from JTA and 90 animals from SOM (n = 162) for each of the acute and chronic experiments. Experimental abalone from SOM were progeny of F4 broodstock of mixed genetic backgrounds originally sourced in 2006, with male H. rubra broodstock sourced from Port Phillip Bay and female H. laevigata sourced from Tasmania, as well as Port Lincoln, South Australia. Animals from JTA were also progeny of broodstock initially sourced from local and non-local sources and bred in captivity for several generations. However, exact details could not be shared due to intellectual property constraints.

Abalone were removed from farm grow-out tanks and transported to Deakin University’s Warrnambool campus according to standard industry practices and translocation requirements (Victorian Fisheries Authority, Permit Number RP1445). To prevent physical harm to abalone, tanks were initially drained of seawater, and abalone were gently pried off the tank surface using a blunt spatula. Animals were subsequently weighed and transferred into live export baskets, before being placed into polystyrene foam boxes with absorbent pads, ice and oxygen for transport to Deakin University. The total time from abalone collection to arrival was 3 h for each farm.

Acclimation and animal husbandry

Upon arrival at Deakin University’s research facilities, the shell of each abalone was briefly dried and labelled for diagnostic purposes with Glue-on Shellfish Tags (Hallprint, Hindmarsh Valley, South Australia, Australia), according to the manufacturer’s instructions. Labelled abalone were returned to export baskets and submerged in a 900 L recirculating aquaculture system (RAS) at a temperature of 14 °C ± 1 °C, consisting of two experimental aquaria and one sump tank, filled with seawater collected locally from Lady Bay, Warrnambool. After 24 h, 12 animals per size class, per farm, were transferred and distributed evenly across two 450 L RASs, also at a temperature of 14 °C ± 1 °C, each including two experimental aquaria and one sump tank. Even distributions of size classes and farms were employed intentionally to control for potential tank effects during the experimental trials. Additionally, 5 animals per size class per farm were transferred to a single aquarium in the 900 L RAS at a temperature of 14 °C ± 1 °C to be used as experimental control animals.

In each RAS, flow was regulated at 750 L h−1 by AquaPro® AP3000 Water Feature Pumps (Aquatec Equipment Pty Ltd., Perth, Western Australia, Australia), and constant aeration was provided by UniStar® Tornado 150-LM Air Pumps (UniStar Australia Pty Ltd., Brisbane, Queensland, Australia). All RASs contained biofiltration media, which was established eight weeks prior with Seachem Stability® (Seachem Laboratories Inc., Madison, Georgia, USA). Animals were acclimated for two weeks at 14 °C ± 1 °C as this approximated the water temperature at each farm at the time of animal collection (H. Ebery, SOM & J. McIntyre, JTA, pers. comm., Oct 2021), with water temperatures maintained using a Hailea® HC-1000 A 1HP Chiller (Guangdong Hailea Group Co., Ltd., Chaozhou, Guangdong, China). Abalone were kept in constant darkness as they are photophobic (Shepherd and Turner 1985), except during animal husbandry procedures and experimental observation which utilised red coloured lighting to minimise disturbance.

All animals were starved for a period of five days to facilitate purging of feed consumed prior to collection. Abalone are able to endure weeks of starvation before depletion of body reserves occurs (Carefoot et al. 1993), and a loss of appetite and reduced feed intake are common in the days following movement and physical handling (Hooper et al. 2011). Additionally, the majority of faecal matter is voided between 12 and 60 h for adults (Shipton and Britz 2001), and total gut content accounts for 6% of body mass in juveniles (Britz et al. 1996). This starvation period controlled for any dietary discrepancies between individual animal responses from each farm (Stone et al. 2014).

Following the five-day starvation period, abalone were fed an artificial diet of Aquafeeds® 3 mm Abalone Food (Aquafeeds Pty Ltd., Mount Barker, South Australia, Australia) to satiety every 24 h. For each day of the acclimation period, nitrogenous waste levels and pH were measured with an API® Saltwater Master Test Kit (Mars Inc., Chalfont, Pennsylvania, USA), and dissolved oxygen and salinity were monitored with a YSI® Pro2030 (YSI Inc., Yellow Springs, Ohio, USA). Waste was siphoned from aquaria and a third of the total water volume in each RAS was replaced with fresh seawater daily.

Acute and chronic thermal limit trials

Twenty-four hours prior to the commencement of the experiments, aquaria were drained, and the abalone were removed from their resting position using a blunt spatula and temporarily held in separate, aerated aquaria. PVC pipe frames lined with 100 mm nylon gutter-brush (Sello Products, Oakleigh, Victoria, Australia) were fitted around all edges of each aquarium, including the control aquaria, to prevent animal movement from vertical to horizontal surfaces during the experiment (Fig. 1). Aquaria were subsequently refilled with seawater and animals were returned to their aquaria, allocating an even number of animals of each size class from each farm to the front and back vertical walls of each tank. Abalone were left to habituate for 12 h overnight in a vertical position prior to the thermal ramping trials commencing.

Fig. 1
figure 1

Single experimental aquarium from experimental system (2 × 150 L aquaria and 1 × 150 L sump), with nylon gutter brush attached to PVC pipe frames, restricting animal movement to faces of the aquaria. Abalone are adhered to the front (ventral view) and back (dorsal view) aquaria faces

Full size image

Two separate experimental trials were performed to test the effects of provenance and body size on CTmax achieved under fast (herein ‘acute’) and slower (herein ‘chronic’) thermal ramping. Both experiments started from a base temperature of 14 °C, with acute and chronic thermal ramping trials consisting of temperature ramping rates of 1 °C per hour and 1 °C per 12 h, respectively. Chronic heating rates of this nature reflect observed rates of summer temperature change at JTA and SOM, where water temperatures can shift by 1–2 °C in a single day, with changes in the nursery and during heatwave events being more intense (H. Ebery, SOM and J. McIntyre, JTA, pers. comm., Feb 2021). Heating was controlled using Schego® 2400 W Titanium Heaters (Schego Schemel & Goetz Electrical Appliance Co., Offenbach, Hessen, Germany) programmed by Aqua Logic® Digital Thermostat Controllers (Aqua Logic Inc., San Diego, California, USA). In contrast, control animals were maintained at a constant temperature of 14 °C. Abalone responses and tank conditions were monitored every 30 min for the acute trial and every 6 h for the chronic trial. The temperature at which an abalone lost pedal adherence to the vertical surface was deemed its critical thermal maximum (CTmax), an endpoint used in previous abalone thermal stress studies (Dı́az et al. 2000; Gilroy and Edwards 1998; Hecht 1994; Searle et al. 2006). The measure of loss of pedal adherence has previously been demonstrated to be a reliable physiological indicator of CTmax in abalone. Chen et al. (2020) observed sharp declines in the heart rates in H. gigantea, H. discus hannai, and the hybrid H. d. hannai × H. fulgens during controlled thermal challenges, correlating strongly with the loss of adhesion. Additionally, exposure to high temperatures is known to cause reduced function in mucous production and physical degradation in H. discus hannai pedal tissue, leading to loss of adherence (Park et al. 2015). The CTmax temperature was recorded to 0.1 °C with a YSI® Pro2030 (YSI Inc., Yellow Springs, Ohio, USA) and was cross-validated using digital thermometer probes. For a schematic of this experimental process, refer to Figure S1.

For both the acute and chronic thermal ramping trials, total ammonia (NH4+/NH3), nitrite (NO2−), and nitrate (NO3−) did not exceed 0.5 mg L−1, 1.0 mg L−1 and 20.0 mg L−1, respectively, and pH remained stable at 8.0. Salinity was maintained in the range of and 33.7–35.6 ppt. Dissolved oxygen concentration steadily decreased for the duration of the experiment as a result of the increasing temperature, with a maximum saturation of 8.2 mg L−1 and a minimum saturation of 7.5 mg L−1.

Sample processing and histological analysis

Individual abalone were removed from experimental aquaria upon reaching their CTmax, with the water temperature and time of detachment recorded. Abalone were subsequently weighed (g) to one decimal place using digital scales (Westlab Pty. Ltd., Ballarat, Victoria, AUS), and farm of origin noted (based on tag ID). Individual abalone were also sexed by lifting the mantle and observing the gonad colour, where pale beige and deep purple/green colourations were used to distinguish males and females, respectively (Heasman and Savva 2007). Sexing using this method could not be reliably performed for all of the smallest animals (10–20 g) due to a lack of gonad development, and subsequent analyses investigating the effect of sex excluded some of these individuals. A semi-quantitative estimate of physical condition of each animal was also recorded, based on visual inspection for abrasion of the epithelial layer of the pedal sole, based on the method by Ellis et al. (1993). A scale ranging from 0 to 4 was assigned to individuals, denoting immaculate, good, fair, poor and extremely poor condition, respectively. Behavioural data were also collected during the trial to identify individual animals where performance could have been compromised due to interactions with other individuals or with the nylon brush used in the experimental system. Animals maintained under constant temperature conditions were monitored in parallel with animals from the treatment groups as experimental controls.

Histological analyses were performed to assess the physiological consequences of thermal stress and to gain insights into the relationship between CTmax and gonad development. Upon animals reaching their thermal maxima, a cross section of gonad and digestive gland (one-third of the way along the conical appendage), pedal sole and side wall of the foot was excised from each animal using a sterile scalpel and fixed in 10% buffered formalin. Samples were subsequently processed through dehydration across an increasing ethanol concentration series of 30–100%, cleared in Histolene® and impregnated with paraffin wax at 60 °C under vacuum (LEICA ASP 3005 automated tissue processor). Processed tissues were embedded in paraffin at 60 °C (LECIA EG1150H embedding station), then hand-sectioned to 4 μm (HM325 Microm microtome) onto glass slides. Slides were stained using standard haematoxylin and eosin (H&E) protocol and cover-slipped using D.P.X mounting agent. Each slide was analysed blind using a Zeiss Axiovert Universal Microscope and images captured using Ziess AxioCam HRc microimaging AxioVision digital capture system software. Male and female gonads were observed to confirm or determine sex, and gonad maturation was categorised against Gastropoda: Haliotidae gonad maturation stages (Shin et al. 2020; Table 1). Digestive gland tubules were observed for evidence of atrophy and scored against a semi-quantitative digestive tubule atrophy scale modified from Weis et al. (1993; Table 2), and categorised against images of digestive gland metaplasia (Couch 1985), and juvenile green abalone digestive gland (Serviere-Zaragoza et al. 2016). The central pedal sole and side wall sections of the foot were examined for prevalence of histological alteration linked to potential loss of adhesion, including laceration injury, oedema, abscess, ulceration, epithelial erosion and cavitation of the pedal sole, lack of mucous cell secretion, vacuolation and necrosis of muscle fibres (Grenon & Walker 1978; Di et al. 2012; Hooper et al. 2014; Kang et al. 2020; Loeher and Moore 2020; Li et al. 2021).

Table 1 Gonad development stages in abalone modified from Shin et al. (2020)
Full size table
Table 2 Semi-quantitative scale for digestive tubule atrophy modified from Weis et al. (1993)
Full size table

Statistical analysis

Animals were removed from subsequent analyses if CTmax values could not be reliably recorded due to mortalities during acclimation, nylon brush preventing loss of adhesion, and animals ‘piggy-backing’ at CTmax endpoints. This resulted in the exclusion of 14% of animals from each trial. A test for significant effects were conducted using a mixed effects ANOVA to compare provenance, sex, body mass and condition as fixed factors and tank number as a random factor. Boxplots were generated for each of the acute and chronic trials using the ggpubr package in R (Kassambara 2020) to visualise the effect of provenance on abalone CTmax. Linear regressions were performed in R to investigate the relationships between abalone CTmax and wet weight, with Pearson’s correlation coefficients fitted to trendlines to measure the strength and significance of the linear relationship.

Male and female gonad development stages were combined to investigate the effect of maturity as estimates of gonad development, based on the metric provided by Shin et al. (2020), are relative to common sexual maturity stages of both sexes. A boxplot were generated using the ggpubr package in R (Kassambara 2020) to visualise the effect of gonad development stage on abalone CTmax, with significance testing performed using a mixed effects ANOVA with gonad development stage as a fixed factor and tank number as a random effect. Post-hoc testing was conducted using Tukey’s HSD test. A Wilcoxon non-parametric test was used to identify differences in semi-quantitative digestive gland tubule atrophy scores between controls and CTmax from experimental animals. A one-way ANOVA was used to test for significant differences between digestive gland tubule atrophy scores and animal size classes. Probit analysis was used to predict the effective CTmax temperature (ECTmax) at which a 10% (ECTmax 10), 25% (ECTmax 25) and 50% (ECTmax 50) reduction in digestive tubule wall thickness (atrophy) is predicted to occur. The lower the ET point estimate value, the higher the relative effect from temperature. Effect temperature estimate values were calculated by means of a logistic (sigmoidal) curve by minimising unweighted squared residuals sum (maximum likelihood) using log 10 transformed rates (adapted from Effective Concentration estimate calculation).

Results

Acute thermal stress trial

CTmax values were recorded for a total of 93 hybrid abalone exposed to acute thermal ramping conditions. The experiment lasted a total of 9 h, at which point the final animal reached its CTmax. The mean CTmax was 23.9 °C (range 19.4–28.3 °C) for SOM abalone and 24.5 °C (range 19.5–28.3 °C) for JTA abalone (Fig. 3). The mean CTmax across all experimental animals was 24.2 °C. A total of 49 control animals were monitored, none of which lost pedal adhesion during the trial, where the aquarium remained at 14 ± 1 °C.

Regression analyses indicated no significant effect of abalone wet weight on CTmax (P > 0.05; Fig. 2a). The mixed effect ANOVA revealed no significant effect of body mass, sex, provenance, condition or tank number on CTmax (P > 0.05; Table 3).

Fig. 2
figure 2

Linear regressions comparing body mass and CTmax of abalone subjected to a acute thermal ramping (1 °C per hour; n = 81) and b chronic thermal ramping (1 °C per 12 h; n = 92). Pearson’s correlation coefficient significance values and linear equations, and mixed model ANOVA significance values are depicted in the bottom right of each plot

Full size image
Table 3 ANOVA results comparing the effect of animal body mass, provenance and sex as fixed effects, and tank as a random effect, on CTmax under acute thermal ramping conditions
Full size table

Chronic thermal stress trial

CTmax values were recorded for a total of 93 hybrid abalone exposed to chronic thermal ramping conditions. The experiment lasted a total of 180 h, at which point the final animal reached it’s CTmax. The mean CTmax was 26.1 °C (range 21.3–28.9 °C) for SOM abalone and 25.7 °C (range 21.3–28.9 °C) for JTA abalone (Fig. 3). The mean CTmax across all experimental animals was 25.9 °C. A total of 51 control animals were recorded, of which a total of six individuals lost pedal adhesion during the trial, occurring within 24 h of each other on the penultimate and final time points. The control aquarium remained at 14 ± 1 °C for the duration of the trial.

Fig. 3
figure 3

Boxplots comparing farm of origin and CTmax of abalone subjected to acute thermal ramping (purple; 1 °C per hour; n = 81) and chronic thermal ramping (orange; 1 °C per 12 h; n = 92). Mixed model ANOVA significance values for provenance effect on CTmax are depicted in the top right of the plot. Asterisks represent outliers and black dots represent the mean

Full size image

In contrast to the acute thermal ramping trials, regression analysis indicated a significant effect of abalone wet weight (P = 0.018; Fig. 2b) on CTmax measured during the chronic trials. Despite the significant effect, the Pearson correlation coefficient was weak (R = − 0.25), with the regression indicating that significance is driven by the smallest and largest animals only (with high variability among intermediate size classes). However, the mixed model ANOVA revealed a significant effect of body mass on CTmax (P = 0.0047; Table 4). Contrary to our hypothesis, provenance did not have a significant effect on CTmax, nor did sex or condition (P > 0.05; Table 4). Interestingly, tank number as a random effect had a significant effect on CTmax (P < 0.001; Table 4), which appears to be driven by a disparity between the two systems (system 1 = tanks 1 and 2, system 2 = tanks 3 and 4; Figure S8). While the exact cause of this phenomenon is not known, it may be driven by minor differences in ramping between temperatures progressively impacting CTmax over time, resulting in a “lagging” effect (Figure S8).

Table 4 ANOVA results comparing the effect of animal body mass, provenance and sex as fixed effects, and tank as a random effect, on CTmax under chronic thermal ramping conditions
Full size table

Histological analyses

Tubule atrophy baseline condition was estimated as the mean pooled value across all control animals (1.03 ± 0.46 SE), and suggested a mean tubule wall atrophy up to 25% digestive tissue loss in control animals. Analysis of animals subject to the acute temperature stress treatment revealed no change in digestive tubule atrophy relative to control animals. However, a moderate positive relationship between CTmax and digestive gland atrophy (Rho = 0.6800, P < 0.0001) was observed in animals subject to chronic thermal stress conditions. A significant increase of digestive gland tubule atrophy above control condition was observed in SOM abalone at CTmax 23 °C and above (χ2 = 37.31, DF = 7, P < 0.0001) and in JTA abalone at CTmax 25 °C and above (χ2 = 25.00, DF = 7, P < 0.001; Figure S2), however, these differences between animals from the two farms were not significant (P > 0.05). No correlation was identified between size class and atrophy index within each CTmax (Table S2). Probit analysis of the pooled digestive gland atrophy index data from the chronic thermal stress trial indicates abalone are predicted to exhibit a non-linear 10% reduction in tubule wall thickness at CTmax 21.50 °C, up to 25% reduction at CTmax 23.67 °C, and up to 50% reduction at CTmax 27.98 °C (Figure S3). Atrophy index values of 3–3.5 (indicating up to 75−87.5% digestive tissue reduction) were recorded in 35.7% of the pooled abalone across the CTmax range of 25–28 °C.

Analysis of animals subject to temperature stress treatments and control animals revealed no histological alteration in the pedal sole and side wall of the foot (Figure S5). There was no evidence of laceration, haemorrhage or oedema, indicating the animals had not sustained a significant physical injury to the pedal sole during transfer to and between tanks. There was no cavitation or epithelial erosion of the pedal sole that would be indicative of ecdysis where extensive loss of epithelial cells would inhibit pedal sole adhesion to the substrate. Similarly, mucous secretions indicated mucous cells in the epidermal layer are active in the crucial adhesion area of the pedal sole. Further, there was no evidence of vacuolation or necrosis of muscle fibre bundles indicative of potential inhibition of pedal muscle contraction. These findings contradict the previous works of Park et al. (2015), however, energetic demands associated with gonad development and/or cardiovascular stress are possible influences on loss of adhesion (Litaay and De Silva 2003; Lopez and Tyler 2006; Chen et al. 2020).

Histological gonad development assessments were not performed for animals from acute treatment trials given body mass had no significant effect on CTmax. In contrast, analyses indicated a significant effect of gonad development stage on CTmax in animals exposed to chronic thermal stress conditions (P = 0.04104; Fig. 4; Table 5), however, Tukey’s HSD tests could not identify which group(s) were responsible for driving this effect (P > 0.05), likely due to low replication. Early gonad development stages (1 and 2) were represented entirely by size class 1, with the exception of two individuals (size classes 2 and 4; Figure S7). Again, we detected a significant tank effect on CTmax (P = 0.002153; Table 5).

Fig. 4
figure 4

Boxplots comparing gonad development stages of all abalone and CTmax of abalone subjected to chronic thermal ramping (1 °C per 12 h; n = 93). One-way ANOVA significance values are depicted in the top right of the plot. Asterisks represent outliers and black dots represent the mean

Full size image
Table 5 ANOVA results comparing the effect of gonad development stage as a fixed effect, and tank as a random effect, on CTmax under chronic thermal ramping conditions
Full size table

Discussion

Worldwide, rising ocean temperatures pose a significant risk to many marine species supporting commercial fisheries and aquaculture (Brander 2010; Cochrane et al. 2009; Free et al. 2019), particularly thermally sensitive organisms such as abalone (Morash and Alter 2016). Predicting the future responses of commercial fisheries to ocean warming depends on understanding which life stages are most vulnerable (Kaustuv et al. 2001), the potential for evolutionary adaptive responses to changing thermal environments (Munday et al. 2013), and the availability of temperature tolerant genotypes which can be used to enhance resilience in wild fisheries and aquaculture through strategic interventions (Foo and Byrne 2016; Liu et al. 2022). This study aimed to address current knowledge gaps surrounding the vulnerability of Australian abalone to thermal stress by investigating the influence of body size and provenance on CTmax in farmed hybrid abalone (H. rubra × H. laevigata) under controlled acute and chronic thermal ramping conditions. These trials showed body mass-and gonad maturation of hybrid abalone had a significant effect on CTmax under chronic thermal ramping conditions, suggesting that larger, mature animals are most vulnerable at elevated temperatures. These findings are consistent with anecdotal evidence from farm settings in the region, and previous studies from other aquatic organisms (Daufresne et al. 2009; Messmer et al. 2017; Peck et al. 2013). The results of these trials support the notion that marine heatwaves and ongoing ocean warming may favour smaller, less fecund animals in farm and potentially natural settings.In contrast, provenance had no detectable effect on critical thermal limits despite our experimental animals coming from contrasting thermal environments. These findings are inconsistent with those from similar trials (Yu et al. 2021b) and population genomic studies in other haliotids (De Wit and Palumbi 2013; Miller et al. 2019; Sandoval-Castillo et al. 2018; Yu et al. 2021a), which indicate adaptive genetic structure relating to temperature across haliotid distributions. However, our findings are inconclusive due to a number of potential limiting factors (discussed below), highlighting the need for further research, ideally using animals from natural populations. Future research of this kind will be key to informing adaptive management of abalone fisheries and understanding the possible effects of ocean warming on the future productivity of global fisheries.

The influence of body size on CTmax

Previous studies investigating the relationship between body size and CTmax in a range of aquatic ectotherms have provided valuable insights into intraspecific patterns of thermal sensitivity. Several studies on fishes have shown that juveniles and young adults often exhibit fitness advantages at elevated temperatures in contrast to earlier life stages or larger and more mature individuals (Clark et al. 2012; Daufresne et al. 2009; Di Santo and Lobel 2017). For example, studies on tropical clownfish (Amphiprion melanopus) and barramundi (Lates calcarifer) have found positive correlations between body mass and CTmax from larval to early juvenile developmental stages (Illing et al. 2020), whereas studies on coral trout (Plectropomus leopardus) have found inverse correlations between life stages spanning juveniles to late adulthood (Messmer et al. 2017). Increasing evidence suggests that these patterns also apply to invertebrates, with Peck et al. (2013) finding greater thermal tolerance in juvenile soft-shelled clams (Laternula elliptica), sea cucumbers (Cucumaria georgiana), sea urchins (Sterechinus neumayeri), and seastars (Odontaster validus) than their mature counterparts. However, thermal stress studies of this nature on haliotids have been limited to either single life stages or a narrow range of size classes (Diaz et al. 2015; Gilroy and Edwards 1998; Hecht 1994; Searle et al. 2006) with limited statistical power for detecting allometric relationships (Nyboer et al. 2020). For the first time, this study demonstrated elevated thermotolerance in small (< 20 g), reproductively immature farmed Australian hybrid abalone under chronic thermal ramping conditions, suggesting that larger, mature animals may be more thermally sensitive.

Despite the influence of body size and maturation on CTmax under chronic thermal ramping, we found no such relationship under acute thermal ramping conditions. Contrasting patterns between controlled acute and chronic thermal ramping conditions have been shown in a range of aquatic taxa (Galbreath et al. 2004; Illing et al. 2020; Middlebrook et al. 2010). Higher overall CTmax values are seen in faster heating rates due, at least in part, to a lag between the environmental temperature and the internal temperature of the organism (reference). In contrast, a slower heating rate can allow some acclimation but also typically results in a lower CTmax, perhaps due to an accumulation of thermal stress (Illing et al. 2020; Kingsolver and Umbanhowar 2018). Faster rates can also overwhelm the organism, leading to increased physiological stress, ultimately masking the ecological relevance of the thermal maxima recorded, which may explain the lack of any observable differences in CTmax during the acute trial in this study (Christen et al. 2018; Ekstrom et al. 2017).

Nonetheless, findings here corroborate anecdotal evidence from both the abalone farms involved in this study, which observe higher summer mortalities in larger, mature individuals (H. Ebery, SOM and J. McIntyre, JTA, pers. comm., Feb 2021). While the drivers of the difference in CTmax observed across animal life stages remain uncertain, some previous studies propose that oxygen limitation may be responsible (Pörtner 2010; Pörtner and Knust 2007). Elevated temperature typically increases metabolic oxygen demand but decreases water oxygen availability, with larger animals having greater absolute oxygen demands and typically succumbing to hypoxia earlier than smaller conspecifics (Forster et al. 2012; Rubalcaba et al. 2020). Such observations have been recorded in abalone, with Vosloo et al. (2013) finding adult South African abalone (H. midae) to be more susceptible to the debilitating effects of hypoxia than juveniles. Increased susceptibility to hypoxia may be attributed to reductions in the diffusion surface area of the gills relative to body mass, leading to a progressive limitation in oxygen availability (Ragg and Taylor 2006). However, mounting evidence contradicts ‘oxygen limitation’ hypotheses (Clark et al. 2013b; Jutfelt et al. 2018; Lefevre et al. 2017; Scheuffele et al. 2021) and as such the ideas should be interpreted with caution within the context of this study. Additionally, we found that maturity inferred from gonadal development stage also had a significant effect on CTmax. These findings are highly consistent with the literature pointing to the high energy demands of reproduction and the suspectiptibility of more mature animals to environmental stress (Brokordt et al. 2015; Li et al. 2007; Sokolova et al. 2012). However, low levels of replication restricted the separation of males and females into their respective gonad development stages. While there was no apparent effect of sex on CTmax, the effect of gonadal development physiology specific to each sex may have an effect which was not observed here, possibly due to low replication. Future studies should therefore prioritise improved replication of gonad development stages across each sex to better investigate this.

The influence of provenance on CTmax

Several population genomic studies have demonstrated correlative patterns of adaptive genetic variation relating to temperature in a range of wild haliotids, including H. rubra and H. laevigata (Miller et al. 2019; Sandoval-Castillo et al. 2018; Yu et al. 2021a). These findings have been further supported with experimental trials on farmed disk abalone (Haliotis discus hannai) that provided functional validation of adaptive genetic differences among farmed abalone stocks from contrasting thermal environments (Yu et al. 2021b). However, the present study found that provenance had no significant influence on CTmax. It is possible our findings are influenced by the mixed genetic backgrounds of the experimental animals, the strength of selection, heterosis driven by hybridisation, and the time needed to adapt to respective farm environments. Animals used in these trials are progeny of broodstock sourced from several thermally variable locations and reared in captivity across multiple generations (H. Ebery, SOM, pers. comm., Feb 2021; J. McIntyre, JTA, pers. comm., Feb 2021). Previous studies have shown that adaptation to culture environments can occur in as little as a single generation in fishes, crustaceans and molluscs (Araki et al. 2007; Araki and Schmid 2010; Christie et al. 2012). Additionally, Holland et al. (2022) recently demonstrated evidence of rapid selection in H. rubra following a single generation of exposure to disease. Considering the potential for rapid evolutionary responses, it might be expected that multiple generations of captive rearing would be sufficient for local adaptation. Repeat trials, ideally using animals from natural populations spanning a greater thermal gradient, will help to confirm the presence of adaptive genetic differentiation relating to temperature in Australian haliotids.

Despite a lack of evidence for adaptive genetic differences among animals from JTA and SOM farms, farm observations indicate differences in thermal tolerance. Specifically, summer mortality at JTA is minimal at temperatures up to 25 °C, while moderate-high mortalities at SOM have been observed at temperatures approaching 21 °C (H. Ebery, SOM and J. McIntyre, JTA, pers. comm., Feb 2021). It is possible that these observations on farm could reflect plastic responses related to seasonal acclimation. JTA is located in Port Philip Bay and experiences considerable seasonal variation in SST (~ 7 °C difference between summer and winter averages), and greater exposure to heatwave events (Harris et al. 1996; Hobday and Lough 2011). In contrast, SOM is situated on the open coast where the upwelling of cold nutrient-rich waters from the continental shelf buffer summer SSTs and reduces the variability between seasons (~ 3 °C difference between seasonal averages; Kämpf et al. 2004; Nieblas et al. 2009). Therefore, the low mortality observed on-farm in animals from JTA may suggest that they are potentially accustomed to greater temperature fluctuations and acclimatise to the conditions throughout spring in the lead up to summer. Reciprocal transplant studies, involving the movement of animals between farms, would be beneficial for assessing the relative performance of genotypes under different environmental conditions, and the role of seasonal acclimation and plasticity in thermal stress response (Aitken and Bemmels 2016; Halbritter et al. 2018).

Future directions and implications for industry

Abalone fisheries in south-eastern Australia are facing numerous challenges, including disease, habitat modifications, and the potential decoupling of trophic interactions that support the region’s fisheries (Gorfine et al. 2001; Holland et al. 2021; Hooper et al. 2007). Findings from this study indicate that ocean warming may compound these threats by favouring and potentially selecting for smaller immature animals, possibly leading to reductions in harvestable biomass and disrupting population dynamics in wild-capture fisheries, and hindering production of larger animals of high value in farm settings (Audzijonyte et al. 2013; Baudron et al. 2014; van Rijn et al. 2017). Similar concerns have been raised for a range of fish and invertebrate species that also show allometrically determined maximum thermal limits under controlled experimental conditions (Clark et al. 2012, 2017; Daufresne et al. 2009; Messmer et al. 2017; Peck et al. 2013). These concerns have been reinforced by several studies which have demonstrated body size reductions in several wild fisheries due to ocean warming independent of size-selective fishing pressure (Daufresne et al. 2009; van Rijn et al. 2017; Wootton et al. 2021). Here, we also demonstrate that elevated temperatures can lead to significant reductions in digestive tubule thickness, potentially impacting nutrient uptake, aerobic metabolism and overall nutritional condition due to reduced numbers of digestive cells in the tubule wall (Garmendia et al. 2011; Kang et al. 2010). Reductions in nutrient uptake in conjunction with temperature stress are likely to impact growth (Vilchis et al. 2005) and potentially reduce harvestable biomass in both farm and wild settings. Nevertheless, we are cautious with our interpretations as these effects are unlikely to be spatially uniform due to the potential for genetically determined clinal variation among stocks across temperature gradients that could contribute to differences in thermal stress limits between wild and farm stocks from different thermal origins.

While our findings provide evidence of non-uniform warming threats to abalone of varying size and maturity, repeat trials using a greater number of animals aimed at testing for provenance, body size, sex and maturity effects on thermal stress responses in wild pure-bred H. rubra and H. laevigata populations are needed to improve the translation of these findings to natural settings. Several studies have demonstrated interspecific hybrid abalone to exhibit heterosis relating to thermal tolerance (Tripp-Valdez et al. 2021; Xiao et al. 2022). Therefore it is possible that heterosis is potentially masking, and even underestimating, the actual risks of thermal stress to pure-bred wild species. Such studies will be essential for providing validation of allometrically determined differences in thermal limits in pure-bred wild species. Similarly, these studies will provide more reliable assessments of heritable genetic variation relating to temperature across species distributions. Such information is needed for assessing the likelihood of standing genetic variation being available for adaptation to rising SSTs, and also for identifying ‘climate ready’ genotypes that could be used to enhance resilience of vulnerable wild stocks through targeted translocations (Hoffmann et al. 2021) and aquaculture through selective breeding programs (Li 2008).

Finally, our study highlights the importance of slow and ecologically realistic experimental heating regimes when testing for thermal tolerance limits in haliotids. Previous studies on haliotids have relied primarily on controlled heating rates of 1 °C per hour or faster (Gilroy and Edwards 1998; Hecht 1994; Searle et al. 2006), none of which have reported significant relationships between body size and CTmax. Experiments on other aquatic taxa have also demonstrated contrasting patterns of thermal tolerance between contrasting heating rates (Galbreath et al. 2004; Illing et al. 2020; Middlebrook et al. 2010). Briefly, consideration should be made regarding the effect of the experimental system used for future studies of this nature, as a significant tank effect was observed in the chronic experiment for both the multi factor and gonad interrogations. Despite the suitable water quality, differences were seen between systems. While we still saw a significant effect of body mass, it is recommended that future studies of this nature prioritise using larger, more uniform study systems to limit the effect of minor system differences that may impact the outcomes observed. Nevertheless, our comparative assessment of animal performances under both acute and chronic thermal ramping conditions adds to the literature utilising constrasting heating rates, highlighting the importance of slow temperature ramping for detecting allometric and ontogenetic effects on thermal stress responses, revealing patterns that would be otherwise masked under acute stress conditions.