Introduction

In terms of food production, marine aquaculture is now the fastest growing sector globally, making it one of the most promising industries for future food security (Duarte et al. 2022). More specifically, seaweed aquaculture, which takes place mostly in Asian countries such as China and Indonesia, is a lucrative industry with a projected growth of up to US$11.8 billion by 2030 (World Bank 2023). Harvesting seaweed from aquaculture not only provides a valuable food source for a growing world population, but also supplies a range of industries and their development of essential materials such as cosmetics, fertilisers, food additives, and pharmaceuticals (Hasselström et al. 2018). In addition, seaweed aquaculture could have a range of environmental benefits, such as providing protection against coastal erosion by reducing wave energy (Løvås and Tørum 2001). Also, by acting as a ‘biological filter’ for the global oceans, seaweed can absorb large quantities of nutrients during growth and produce oxygen, which could help to improve coastal waters where eutrophication is an issue (Hu et al. 2022). Importantly, when produced on an industrial scale, seaweeds may help to mitigate against the impacts of global climate change, as they sequester large quantities of CO2 (Duarte et al. 2013; Zheng et al. 2019).

Despite the environmental and economic benefits of the expanding seaweed aquaculture industry, it remains only a marginally economically viable activity throughout Europe (Zheng et al. 2019). Several European countries including Ireland, Norway, and Sweden are supporting research that focuses on seaweed cultivation techniques to aid accessibility for the European market (Handå et al. 2013; Forbord et al. 2020a, b; Kerrison et al. 2020). Seaweed is generally grown using long-line techniques which can be highly labour intensive (Peteiro et al. 2016; Campbell et al. 2019). Therefore, technological modifications are needed to ensure economic viability for the seaweed aquaculture industry to grow throughout Europe (Edwards and Watson 2011). These modifications are not only needed to reduce labour costs in the harvesting of seaweed biomass, but also to reduce the hatchery phase where zoospore or gametophyte cultures are seeded onto twine and grown in tanks for a minimum of four weeks prior to their introduction to coastal seas (see Edwards and Watson 2011; Campbell et al. 2019; Kerrison et al. 2020).

Consequently, the scope to expand seaweed aquaculture throughout Europe, and indeed globally, will be limited by the high production costs associated with labour and the processes involved during the hatchery growth phases (Forbord et al. 2020a, b). For reference, a detailed description of the different hatchery processes is outlined in the best practise report for seaweed cultivation by Mooney-McAuley et al. (2016). However, mechanising parts of the aquaculture industry is being seen as a way of increasing the viability of the seaweed market (Werner et al. 2004). While current research is investigating the efficacy of mechanised processes including the deployment of seedlings from twine to ropes, biomass harvesting and crop handling logistics (Stévant et al. 2017), the impact of mechanised and pressurised spraying of seaweed cultures on growth are yet to be established. Mechanised spraying of substrata with gametophyte cultures during the initial hatchery growth phase (Mooney-McAuley et al. 2016) could reduce the time and labour costs of preparing large quantities of spools for cultivation purposes, however an initial investigation into the effects of pressure on kelp zoospores is needed. Hence, in this study we aimed to determine the efficacy of a future mechanised seeding procedure by investigating how differing pressure treatments impact upon the growth of Saccharina latissima zoospores when they are seeded onto twine in the hatchery.

Methods

Sample collection and spore release

Ten fertile drift sporophytes of Saccharina latissima were collected from Kilclief beach, Northern Ireland (54.3359° N, 5.5435° W) on 24 October 2021. The S. latissima sporophytes were wiped clean and placed on paper wetted with seawater before being enclosed in plastic bags and refrigerated overnight (4 °C).

Following a 24-h period, sporangial areas were excised from each individual blade and randomly divided between five containers to a total mass of 75 g per container. At 10 min intervals, to allow for time to carry out experimental treatments (see below), each batch of 75 g was added to a 1250-mL beaker containing autoclaved seawater and placed on an orbital shaker (Stuart SSL1, Davidson & Hardy Ltd, UK) at 150 rpm to release spores.

Experiment

After 30 min of spore release on the orbital shaker, 200 mL of the resulting spore solution was filtered through 45-µm nylon mesh and funnelled into a clean 1-L pressure cylinder (North Sea Winches Ltd, Scarborough, UK). The solution was then subjected to a pressure treatment for 5 min before being poured into a 250-mL conical flask. This procedure was then repeated six times for each of five pressure treatments (N = 30): 1 (control), 2, 3, 4 and 5 bar (or 100, 200, 300, 400, and 500 kPa). The pressure cylinder, funnel and filter were rinsed with deionised water between each replicate.

Simultaneously, spores were visually checked for motility under a light microscope (Leica DMIRB) and 9 mL of each of the five spore solutions was added to a falcon tube with 1 mL of 7% Lugol’s iodine (Sigma-Aldrich; Product no. L6146) for analysis of spore density mL–1 in the resulting 0.1% (v/v) solution. Subsequent spore counts using a compound microscope (Olympus BX41) and haemocytometer (Hawksley, England, UK) (Guillard and Sieracki 2005) revealed mean (± SE) spore densities of 31,666 ± 5068 mL–1 (1 bar), 54,333 ± 11,181 mL–1 (2 bar), 33,333 ± 10,833 mL–1 (3 bar), 56,000 ± 10,981 mL–1 (4 bar) and 11,666 ± 3333 mL–1 (5 bar). While the exceptionally low spore density in the solution used for the 5 bar treatment requires careful attention in the interpretation of our results for that treatment, the other densities are fairly consistent and randomly distributed among treatments.

Each flask was equipped with cotton twine (10 cm lengths) and placed under a white 12:12 light cycle (20 µmol photons m–2 s–1) in a temperature-controlled room (12 °C). Spores were left to settle on the twine for 2 days with no airflow. Following this, the flasks were emptied, and the twine was suspended in new flasks containing 200 mL of continuously aerated autoclaved seawater, supplemented with 0.1 mL germanium dioxide (104.61 g mol−1; Sigma-Aldrich; CAS no. 1310–53-8) to reduce diatom growth, and 0.2 mL of 10% F/2 medium (Sigma Aldrich; Product no. G0154) for nutrients, and left to grow for 7 weeks. Water changes and replenishment of nutrients was carried out twice per week.

Quantification

Following completion of the experiment after 7 weeks, three 1-cm sections (middle and ends) of each 10-cm twine (N = 30) were photographed using an Olympus SZX16 stereomicroscope with an Olympus E600 DSLR camera attachment. All photographs were taken at 10 × magnification with a 10-mm object for scale. Using Coral Point Count v4.1 with Excel extensions software (Kohler and Gill 2006), a 10 mm quadrat was added to each photograph and the percentage cover of S. latissima was calculated by assigning 10 random points (< 1 mm in size) within the quadrat and recording presence or absence. These steps resulted in a binomial response variable for sporophyte cover with 30 × 3 × 10 = 900 technical replicates (ones and zeros).

To corroborate estimated sporophyte cover, up to five of the longest sporophytes from each replicate were measured using ImageJ v1.53 k (Rasband 2012). Nine replicates did not have measurable sporophytes leading to a reduced sample size (N = 21). These steps resulted in a continuous response variable for sporophyte length with 101 technical replicates.

Data analysis and visualisation

Data were analysed and visualised in R v4.2.1 (R Core Team 2022) within the integrated development environment RStudio v2022.12.0 + 353 (RStudio Team 2022). We modelled the effects of pressure on sporophyte cover and length using generalised linear mixed models (GLMMs) written with the glmer function of lme4 v1.1–31 (Bates et al. 2015). Multilevel modelling was strictly necessary to counteract artificial inflation of sample size and concurrently confidence through our extensive technical replication. Using the hierarchical structure of GLMMs, we nested technical replicates within independent statistical ones by defining flask as a random factor, allowing the model intercept to vary among flasks and adding this additional source of uncertainty to our estimation of confidence. We did not choose to factor variation in spore density of the inoculating solutions into our models since this would have convoluted them and no potential confounding effect was evident (see Sect. "Experiment").

The binomial (presence, absence) nature of our cover variable dictated a binomial GLMM for which we selected the logistic link function. Since binomial data are not symmetrically distributed around the mean and normal approximation (Wald) is consequently suboptimal, asymmetric Wilson score 95% confidence intervals and s.e.m. were calculated with binom v1.1–1 (Dorai-Raj 2015). Sporophyte length was right-skewed like most biological response variables but only takes positive values so was perfectly suited for a gamma GLMM with a logarithmic link function. Model optimality was determined based on graphical scrutiny of residuals for normality and homogeneity (Zuur et al. 2009).

Type II Wald chi-square (χ2) tests of the effect of pressure – testing the null hypothesis that the slope is not different from zero – were performed with the Anova function of car v3.1–1 (Fox & Weisberg 2019). Data were visualised in ggplot2 v3.3.3 (Wickham 2016), plots aligned in cowplot v1.1.1 (Wilke 2020) and illustrations added in Affinity Designer v1.10.6 (Serif Ltd, Nottingham, UK). All analysis and visualisation steps can be scrutinised and replicated using the files provided at github.com/lukaseamus/spores-under-pressure.

Results

Sporophyte growth was visually confirmed in all treatments over the course of the 7-week experiment (Fig. 1). Our model suggests that the percentage cover (c) of S. latissima (%) decreases with increasing pressure (p) (N = 30, χ21,30 = 13.36, p < 0.001, Fig. 2a), a trend which is best described by \({\textit{c}}= \frac{100}{{1}\mathrm{+}{\mathrm{e}}^{\mathrm{0.33}{\mathrm{p}}\mathrm{ + }\mathrm{0.12}}}\). Across our range of pressure treatments, c decreased by 22.22% from 41.67 [24.19, 61.53] % (mean [– s.e.m., + s.e.m.], n = 6) at 1 bar (ambient atmospheric pressure) to 19.44 [8.23, 39.39] % at 5 bar, although it reached its minimum with 17.78 [7.21, 37.55] % at 4 bar. The largest stepwise decrease between pressure treatments (10%) was observed from ambient to 31.67 [16.51, 52.06] % at a doubling of atmospheric pressure (2 bar).

Fig. 1
figure 1

Sporophytes of Saccharina latissima grown over a 7-wk period from juvenile zoospores subjected to increasing pressure treatments (1–5 bar)

Full size image
Fig. 2
figure 2

The percentage cover (a) and lengths (b) of Saccharina latissima sporophytes under increasing pressure. N = total technical (statistical) sample size. Points are technical replicates (measurements) and point-ranges are statistical replicates (flasks) and their corresponding measurement error (technical 95% confidence intervals). Lines are model fits and shaded areas are 95% confidence intervals, augmented with intercept variance between flasks as an additional source of uncertainty

Full size image

Length (l) of individual S. latissima sporophytes (mm) also decreases with increasing pressure according to our model (N = 21, χ21,21 = 7.60, p = 0.006, Fig. 2b), a trend which is best described by \({\mathrm{l}} \, \mathrm{=}{ \, {\mathrm{e}}}^{-\mathrm{0.23}{\mathrm{p}}\mathrm{+}\mathrm{0.12}}\). These findings corroborate the pressure effect on c. Across our range of pressure treatments, l decreased by 0.53 mm from 0.87 ± 0.11 mm (mean ± s.e.m., n = 6) at ambient pressure to its minimum 0.34 ± 0.04 mm (n = 5) at 5 bar, although it reached its maximum with 0.93 ± 0.4 mm (n = 3) at 3 bar. The largest stepwise decrease between pressure treatments (0.52 mm) was observed from 3 bar to 0.41 ± 0.11 (n = 2) at 4 bar.

Discussion

Here we demonstrate that both S. latissima percentage cover and sporophyte lengths are significantly reduced when zoospores are subjected to pressure at the start of the hatchery growing stage. Our results highlight that S. latissima grew more effectively and in higher numbers over 7 weeks when zoospores were only exposed to normal atmospheric pressure, and therefore our models indicate that minimal pressure is ideal for future mechanised hatchery techniques.

For this study we adopted a direct seeding approach, whereby viable zoospores are directly attached to the substrata upon release from the fertile blade material (Mooney-McAuley et al. 2016) to determine the percentage growth coverage at increasing pressures. This form of direct seeding (not to be confused with binder-seeding techniques; Forbord et al. 2020b) has been highlighted as the most cost effective, favourable and simplest seeding approach used in the aquaculture of S. latissima (Mooney-McAuley et al. 2016). Also, while natural fertile material can only be used at certain times of the year (Mooney-McAuley et al. 2016), year-round sorus induction of S. latissima is possible by manipulating light levels (Forbord et al. 2012). Despite only obtaining 40% sporophyte coverage in our study at 1 bar, compared to the ~ 80% identified in Forbord et al. (2020b) over the same period, different methods were used to calculate percentage cover which may explain the discrepancy. Further, zoospore densities of ~ 250,000 mL−1 were reported in Forbord et al. (2020b), whereas in our study, much lower densities (~ 11,000—~ 56,000 mL−1) were observed in our pressure treatment samples. Although zoospore densities fluctuated from high to low across our five pressure treatments, a decreasing correlation of percentage sporophyte cover and lengths was still found at the end of the experiment.

While we have determined the effects of increasing pressure on S. latissima zoospores in this study, investigation into the impacts on the gametophyte stage would be beneficial. This would help in determining the broader effects of pressure, especially in terms of spraying cultures onto cultivation materials. This is particularly important as the gametophyte stage of the kelp lifecycle can be used to seed lines all year round by storing them under red light where they continue to increase in density through cellular division (Lüning 1979). It also would be beneficial to monitor the zoospore germination and onward development of the kelp after pressure treatments to determine if the negative effects found here occur immediately or as development of the spores continues over time. To further understand the cost benefit analysis of using a mechanised gametophyte spraying approach in comparison to manual spraying, a follow up assessment of growth in the field and an investigation into optimal seeding densities for successful growth is also needed. Particularly, as studies suggest that lab-based results do not necessarily translate to the field (Kerrison et al. 2019) and that the optimal density of gametophytes or sporophytes on substrata for any seaweed species remains unknown (Kerrison et al. 2015). This would help to determine if the pressure treatments impede growth after the hatchery stage and could be tested using both direct seeding of zoospores and seeding of gametophytes. Finally, an investigation into the effect of mechanical and physical stress of mechanised spraying on zoospores and gametophytes may also be beneficial.

In conclusion, our models highlight that sporophyte cover and length declines with increasing pressure applied during seeding, indicating that minimal pressure is optimal for sporophyte growth in future mechanised hatchery techniques. Further research is also needed to determine how increasing pressure impacts upon the growth of the mature gametophyte life stage in a hatchery setting. A follow-up validation study is also needed to determine how hatchery pressure treatments impact upon the subsequent growth of S. latissima sporophytes in the field.