Introduction

Since ancient times seaweeds have been grown for food and a variety of chemical constituents. The seaweed sector is currently projected to be worth US$ 6 billion per year (FAO 2022). In 2020, 35.01 million tonnes of seaweed were generated through cultivation, representing 97.45 percent of the total 35.93 million tonnes of seaweed produced worldwide (FAO 2022). Tropical seaweed production in the Coral Triangle has increased from 0.39 to 11 million tonnes of fresh weight from 1990 to 2020. This includes commercially significant algae, including Eucheuma spp., Gracilaria spp., and Kappaphycus alvarezii. The Coral Triangle is one of the world's most biodiverse regions in terms of marine life, spanning six countries in Southeast Asia and the Pacific, namely Indonesia, the Philippines, Malaysia, Papua New Guinea, the Solomon Islands, and Timor-Leste (Asia Development Bank 2016).

While seaweeds such as Gracilaria are grown largely for human consumption, e.g., agar–agar, Eucheuma and Kappaphycus are mostly cultivated for the manufacturing of carrageenan, a popular ingredient in the cosmetic and food processing industries. Seaweed production in Malaysia increased from 111,298 t in 2008 to 182,061 t in 2020, with a market value of US$ 1.4 billion (FAO 2022). Both the hanging basket and longline methods are cultivation methods commonly adopted by the seaweed industry in Malaysia, especially in the cultivation of K. alvarezii (Sade et al. 2006; Phang et al. 2019). Rising demand for cultivated seaweeds in the magnitude of orders greater than the current levels has been anticipated, mainly due to its dietary benefits and capability to generate multiple ecosystem services (Kang et al. 2011; Rizki et al. 2023) in response to the United Nations Sustainability Development Goals (Duarte et al. 2021).

Even though widespread cultivation of seaweeds has been essential to improving the socioeconomic position of coastal communities, the growth in seaweed production in the tropical Coral Triangle might lead to an increase in the region's halocarbon burden (Phang et al. 2015). Coastal seaweeds are recognised as significant sources of volatile halocarbons such as CHBr3 (Keng et al. 2020). Short-lived halocarbons account for 30% of total stratospheric bromine (WMO 2018). Previous research indicated that K. alvareziii, Gracilaria changii, and Gracilaria edulis are among the farmed tropical seaweeds that are capable of emitting significant quantities of such halocarbons (Leedham et al. 2013; Mithoo-Singh et al. 2017). Much of the short-lived halocarbon load emitted from the surface is removed by photolysis and oxidation in the troposphere, but some fraction can enter the stratosphere, notably through rapid convective uplift in the tropics, where photolysis releases reactive inorganic bromine species that destroy ozone (Hossaini et al. 2012a; Liu and Zipser 2015; Pan et al. 2017). These brominated short-lived halocarbons are responsible for significant loss of lower stratospheric ozone and may be 3.6 times more powerful than long-lived halocarbons in exerting a radiative impact through ozone loss (Hossaini et al. 2015).

The emission of volatile halocarbons by seaweeds has been linked to environmental factors such as variations in irradiance, temperature, saltwater nutrients, and salinity (Laturnus et al. 2000; Abrahamsson et al. 2003; Bondu et al. 2008). Along the electron transport chain, harmful reactive oxygen species such as hydrogen peroxide and hydroxyl compounds are often generated in chloroplasts and mitochondria (Manley and Barbero 2001; Dummermuth et al. 2003; Gill and Tuteja 2010). Bromoperoxidases in the seaweeds catalyse the reaction between H2O2 and halides to produce hypohalide, which then reacts with either the ketones in the seaweeds or dissolved organic matter (DOM) in the environment to produce CHBr3 (Wever et al. 1991; Opsahl and Benner 1997; Lin and Manley 2012; Wever and van der Horst 2013). Disturbances to the H2O2 equilibrium, such as changes in environmental conditions, may have an impact on the emission of halocarbons by seaweed, since the accumulation of reactive oxygen species (Li et al. 2010; Dietz 2016) may result in an increase in the generation of volatile halocarbons (Küpper et al. 2013).

Our previous works investigating the emission of halocarbons by tropical seaweeds, including the effect of environmental factors such as irradiance, pH, and temperature on halocarbon emissions through a series of laboratory-based studies, showed that tropical seaweeds are important coastal sources of halocarbons and could contribute to the regional, if not global, halocarbon pool (Keng et al. 2013; Leedham et al. 2013; Mithoo-Singh et al. 2017; Keng et al. 2021). However, gaps remained in this area as emissions data from the tropics remained limited, and much remained unknown about how environmental changes could affect the emission of halocarbons by the seaweeds during cultivation. The lack of data from oceanic input and the under-representation of coastal and extreme emissions often cause huge variability and uncertainties in the modelling and estimation of global oceanic halogen load (Hossaini et al. 2012b; Ziska et al. 2013, 2017; Fuhlbrügge et al. 2016; WMO 2018). In this study, the emissions of CHBr3, CHBr2Cl, and CHBrCl2 by G. manilaensis and the economically important K. alvarezii harvested from the various cultivation systems were quantified. The effects of temperature, irradiance, seawater salinity and nutrient levels are also investigated. The emission rate of CHBr3 by K. alvarezii from the offshore platform was then used to estimate the release of Br by farmed K. alvarezii during harvesting in Malaysia. This study will provide in-depth information on the halocarbon emission by farmed seaweeds in the tropics, which could be useful for future estimation work, particularly with the potential expansion of the seaweed farming industry, considering the vast beneficial properties and the newly explored potential of exploiting the high CHBr3 content in red seaweeds like Asparagopsis taxiformis to reduce methane production by ruminants (Glasson et al. 2022).

Materials and methods

Seaweed cultivation

Gracilaria manilaensis Yamamoto and Trono (Gracilariaceae, Rhodophyta; PSM13041) was collected at Tanjung Kupang, Johor, Malaysia (1°20′26″ N 103°36′16″ E), while Kappaphycus alvarezii (Doty) L.M.Liao (Solieriaceae, Rhodophyta; PSM13827) was purchased from a seaweed farm off Tawau, Sabah, East Malaysia (4°18′58″ N 117°46′26″ E). The seaweeds (Fig. 1) were then transported to the Bachok Marine Research Station (BMRS) in Bachok, Kelantan. Gracilaria manilaensis and K. alvarezii were both grown in the BMRS hatchery's onshore overflow tank cultivation system and on an offshore platform. Gracilaria manilaensis was also cultivated in a cage culture system at a river mouth close to the BMRS. The cultivation of G. manilaensis was tested in the river, as this seaweed is commonly found in sheltered bays, estuaries, and river mouths with sandy or sandy-mud bottoms.

Fig. 1
figure 1

The cultivated seaweed species are, from left to right, Gracilaria manilaensis and Kappaphycus alvarezii. Further descriptions of these seaweeds are provided in the main text

Full size image

Onshore tank cultivation system

The onshore tank cultivation was made up of several overflow systems. Each of the overflow systems consists of three polypropylene tanks, each measuring 1.3 × 0.9 × 0.5 m, attached to a reservoir tank with a top filter (Fig. 2a). Approximately 1000 L of seawater was used to fill the reservoir and the three tanks, and a continuous flow of seawater within the system was maintained by submersible pumps. About 30% of the seawater in the system was refreshed with fresh natural seawater every three days, and the cultures were maintained at a 28 ± 3 °C seawater temperature and under natural sunlight at the BMRS outdoor hatchery. Two onshore overflow systems were used for G. manilaensis and K. alvarezii cultures, respectively. The designated tanks were each filled with 1 kg of G. manilaensis and 500 g of K. alvarezii and these culture studies were carried out for 4 weeks. An initial study was carried out in April 2018 and successive studies in October 2018.

Fig. 2
figure 2

The three systems from which seaweed was cultivated were: onshore tanks (a), offshore platform (b), and cage culture at Sungai Melawi (c, d)

Full size image

Offshore platform

An offshore platform measuring 5 × 5 m was constructed from wooden planks and 120 L polyethylene drums as floats. The platform was anchored about 600 m offshore in 4.5 – 6 m depth of water, depending on the tide. Fishing net of 2 cm mesh was attached to the underside to protect the seaweed from fish grazing. The seaweeds were cultivated for four weeks; slightly shorter than the typical cycle of 45 days for the cultivation of G. manilaensis and K. alvarezii. This was because our preliminary investigations had indicated that both G. manilaensis and K. alvarezii have higher biomass productivities and growth rates in weeks 1–3, with a decline beginning in week 4. The seaweeds were generally healthy with no visible signs of senescence or bleaching throughout the chosen cultivation period.

In these studies, G. manilaensis was cultivated in cuboid baskets, while K. alvarezii was tied directly to a monoline system. Cuboid baskets of 0.3 × 0.3 × 0.1 m were made from 1 cm mesh fishing net wound around PVC tubing. Each basket contained 500 g of G. manilaensis and three baskets were spaced out along a 50-m nylon monoline and kept floating about 10 cm under the sea surface (Fig. 2b). Bunches of K. alvarezii, each weighing 150 g, were tied along a monoline at intervals of 20 cm, with a 15 cm gap between the seaweed bunches and the monoline. All monolines were kept 0.5 m apart from each other. Loggers to collect irradiance and temperature data were tied to the lines. There was a total of three rows of G. manilaensis in the hanging baskets and five lines of K. alvarezii bunches cultivated at the platform between October 4 and 31, 2018.

River mouth cage culture

A cage culture system measuring 1.5 × 1.0 × 0.5 m with three compartments as replicates was constructed out of the same material as the hanging basket used in the offshore platform cultivation system. The three compartments were each filled with 500 g of G. manilaensis (Fig. 2c). The system was then deployed at the river mouth of Sungai Melawi (6º1′23.24″ N, 102º25′3.47″ E), close to the BMRS (Fig. 2d). The water depth at the river mouth ranged from 0.2 to 1.5 m, influenced by tidal fluctuations. The cage culture system, larger than the offshore hanging cuboid basket, allows the seaweed to float at varying tidal levels. This design ensures the seaweeds remained mostly immersed in water while being capable of floating to the surface during high tide.

Sampling and data collection

Four sampling trips were made (Table 1). Seaweeds sampled during trips 2, 3, and 4 were from the same starting batch, resembling a cycle of seaweed cultivation.

Table 1 The sampling dates, the culture age of the seaweed from which air samples were collected (via flux chamber and air canisters), the sampling conditions, i.e., daylight and darkness, the number of air samples (n) collected, and the net emission of CHBr3, CHBr2Cl, and CHBrCl2 (average ± standard deviation of biological replicates; minimum to maximum; ppbv) by G. manilaensis and K. alvarezii from the three cultivation systems
Full size table

Seaweed from the offshore platform was collected by boat: a random rope line with hanging baskets of G. manilaensis, and another with bunches of K. alvarezii, were brought back to the onshore hatchery. Each trip took between 5 and 10 min. The seaweeds were immersed in a basin containing seawater during transport. All seaweed samples used for incubation in the flux chamber from the onshore tanks and offshore platform were pre-weighed and immersed in an isolated compartment in the onshore tanks for no longer than two hours prior to incubation. At the onshore tanks, approximately 500 g of seaweed biomass were harvested from randomly selected tanks during each sampling event. For the offshore and river culture samples, seaweed from a single bunch or compartment was randomly chosen for sampling. In cases where the biomass was low, seaweed from the next closest bunch or compartment was also harvested to ensure an adequate sample of biomass.

Data loggers (Onset, HOBO) were attached to the three cultivation systems and within the flux chambers to record irradiance and temperature, with the latter being used to correlate with halocarbon emission data during incubation. The temperature and irradiance levels in the flux chamber were within the range experienced by the seaweeds at their various cultivation locations, and is therefore representative of those experienced during growth. Seawater salinity (Master-S38M, ATAGO) was determined during seaweed sampling. Seawater samples were collected and transported back to Kuala Lumpur in a chilled condition for analysis of nutrient content. The seawater nutrient levels were determined using the HACH DR 3900 spectrophotometer and the salicylate (Reardon et al. 1966), diazotization (USEPA 1979), cadmium reduction (APHA 1998), and ascorbic acid (APHA 1998) methods (Hach 2001).

Flux chamber air sample collection

A custom-made flux chamber (Fig. 3) was used to sample the halocarbon emissions from the seaweed. The 40-L flux chamber was designed and modified according to the EPA-recommended flux chamber (Eklund 1992) and that described by Sartin et al. (2002). A small axial fan unit powered by a 12 V battery was built in to help move air around in the chamber.

Fig. 3
figure 3

The flux chamber connected to a canister via an air pump: (a) the collection of control air; (b) the collection of air sample containing halocarbon emissions from seaweeds

Full size image

Flux chamber measurements on seaweeds cultivated from the onshore tank and the offshore platform were made at the onshore hatchery, whilst measurements on samples from the cage culture system were made on shore near the river mouth to minimise changes in emissions during transport. All emission measurements on seaweeds from the offshore platform and river mouth were carried out during daylight, whilst seaweeds from the onshore tanks were also sampled after sunset to determine their emissions in the dark. The seaweeds were only removed from the isolated compartments of the tanks, or from the river mouth, immediately before incubation in the flux chamber.

Whole seaweeds were blotted dry for weighing and around 500 g were placed in the flux chamber in which the seaweeds were re-moistened by sprinkling 100 mL of seawater or river water, depending on the locations where the seaweeds were sampled. This was done to prevent complete drying-out of the seaweed during the 30 min incubation period and to better mimic that from freshly harvested seaweed. The same steps were repeated for the control sample, with the same amount of river or seawater used, but no seaweed. The seaweeds and controls were left for 30 min in the flux chamber prior to the collection of the air samples. The temperature within the flux chamber was recorded at the start and at the end of the incubation, while two to three readings of irradiance levels were also recorded.

Air samples from the flux chamber were collected using SilcoCan canisters (Entech Instruments). Canisters with volumes of 2.7 L and 3.2 L were cleaned prior to sampling through repeated baking (at 90 °C) and flushing with high-purity nitrogen gas (Linde Malaysia) for six cycles before evacuating (Canister Cleaning System, Entech).

Air from the flux chambers was drawn into the canisters through PFA transfer tubing connected to a 12 V battery-powered pump (Air Dimensions Inc.) to a final pressure of 10 psig after flushing once with the same air to the same pressure. Samples were transported back to the laboratory for analysis within three weeks.

Air samples, with or without seaweeds, collected after sunrise and before sunset were classified as ‘daylight’ samples, while air samples collected after sunset were classified as ‘dark’. Control air samples in daylight (n = 5) and dark (n = 4) without seaweed were collected for comparison to the air samples containing seaweed. Net emissions by seaweeds were then determined by subtracting the mixing ratios of compounds in the flux chamber containing seaweeds from the controls, which were taken prior to the sampling of the seaweeds on the same day, either during daylight or darkness. All halocarbon mixing ratios (ppbv) measured from the flux chambers represent the emissions from 500 g of seaweed biomass in 30 min, standardised from the actual measured weight of biomass in each.

Sample analysis

Filled canisters were attached to the inlet of an Entech pre-concentrator (7200 Sample Preconcentrator, Entech Instruments) coupled to an Agilent GCMS (7890B/5977B, Agilent Technologies). The pre-concentrator uses a three-stage trapping process known as ‘Extended Cold Trap Dehydration’ to improve the sensitivity of the analyses. This includes passing 100 cm3 air samples through an empty Silonite-D treated trap (M1) pre-cooled to − 40 °C, eliminating water from the samples. The analytes were then transferred to a second-stage Tenax trap (M2) at − 50 °C for trapping. This was followed by back-desorption of the Tenax trap to the final focusing trap (M3) prior to injection into the GCMS system (Markle et al. 2017). Further details on the pre-concentration parameters are summarised in Supplementary Table S1.

The GC system was fitted with a 60-m capillary column (J&W DB-VRX; film thickness 1.40 μm, internal diameter 0.25 mm). The GC oven was programmed to hold the temperature at 35 °C for 10 min, ramp up to 220 °C at a rate of 8 °C min−1 and hold for 3 min (Markle et al. 2017). The analyte concentrations were then determined by comparing the peak areas or target ion masses against the peak areas of the same in a standard gas. A standard gas containing volatile organic compounds was diluted to a concentration of 10 ppbv (courtesy of Entech Instruments) with a calibration volume of 10 cm3. Of these, three compounds, CHBr3, CHBr2Cl, and CHBrCl2, were chosen for monitoring.

A total of three technical replicates were analysed from each canister containing air samples, and the mean value was represented as a single data entry for each of the control and biological replicates (n), therefore, all standard deviations mentioned in the text represent values from the biological replicates. The detection limit of < 1 ppbv for all three compounds was determined from the sum of three times the standard deviation and the blank mean of the system (Kaiser 1970).

Statistical analysis

Data collection, figure constructions and initial analyses were performed in Microsoft Excel. Following initial data processing, statistical analyses were performed on the dataset using SPSS Statistics Version 22 (IBM).

The following statistical tests were performed:

  • To compare halocarbon emissions during daylight and darkness at the onshore tank cultivation system, the dataset (n = 7) for daylight and dark emissions by G. manilaensis; and n = 6, 3 for daylight and dark emissions by K. alvarezii was analysed using the Mann–Whitney U test.

  • One-way ANOVA was performed to compare the halocarbon emissions from G. manilaensis (n = 7, 9, 2 for onshore tank, offshore platform, and river mouth culture respectively) and the means between seawater nutrient and salinity levels from different cultivation sites (n = 15, 16, 2, and n = 24, 16, 2 for the readings of nutrients and salinity levels at the onshore, offshore, and cage cultivation systems).

  • An independent sample t-test was used to test the difference in the transformed emission data by K. alvarezii between the onshore tank (n = 6) and offshore platform (n = 7).

  • To compare G. manilaensis and K. alvarezii halocarbon emissions at various temperatures and irradiance during incubation in the flux chamber, Pearson’s bivariate (n = 25, 14) and Pearson's partial correlation (n = 22, 13) were run on temperature and irradiance data collected during each flux chamber incubation. The latter was done to eliminate the effect of irradiance. The linear relationships between these factors were assessed by scatterplots and partial regression plots. The normality of transformed data was assessed by Shapiro–Wilk's test (p > 0.05), and there were no univariate or multivariate outliers, as assessed by boxplots and Mahalanobis Distance, respectively.

  • The Pearson Product-Moment Correlation was used to look at the relationships between the transformed daylight emission data of G. manilaensis (n = 15, 17) and K. alvarezii (n = 12, 13), which was pooled from all daylight measurements from the three cultivation systems, and the measured levels of nutrients and salinity of the seawater data collected during the sampling of the seaweeds. The significance levels (2-tailed) were determined. Preliminary analyses showed the relationship to be linear with both variables normally distributed, as assessed by Shapiro–Wilk's test (p > 0.05), and there were no outliers.

Estimation of emissions from the harvesting of K. alvarezii in Malaysia and the Coral Triangle

To estimate the bromine emission from the harvesting of K. alvarezii in Malaysia and the Coral Triangle, the emission range of halocarbon per kg fresh weight of K. alvarezii per hour (ppbv kg FW–1 h–1) was estimated from the emission range of 500 g of K. alvarezii incubated in the flux chamber for 30 min (ppbv). The estimated range (ppbv kg FW–1 h–1) was then extrapolated to the respective countries' total biomass production in 2020 using the following formula, assuming a one-hour average harvesting period and field conditions similar to those in the flux chamber:

$$\mathrm{Estimated}\;\mathrm{emission}\;\mathrm{in}\;2020\;\left(\mathrm{mol}\;\mathrm{Br}\right)=\;\mathrm X\times\mathrm{mixing}\;\mathrm{ratio}\;\mathrm{of}\;\mathrm{halocarbon}\;\left(\mathrm{ppbv}\;\mathrm{per}\;\mathrm{kg}\;\mathrm{FW}\;\mathrm{per h}\right)\times1.5956\times10^{-9}\times\mathrm{biomass}\;\left(\mathrm{kg}\;\mathrm{FW}\right)\times1\mathrm h$$

where 1.5956 is the mol of air, derived from the ideal gas law in the 40 L flux chamber at an average temperature of 32.35 °C; X = 3, 2, and 1 for the number of bromine atoms in CHBr3, CHBr2Cl, and CHBrCl2, respectively.

Results

Halocarbon emissions by seaweeds from the three cultivation systems

Throughout the sampling period, net halocarbon emissions by G. manilaensis and K. alvarezii collected from the cultivation systems showed that the seaweeds released a higher amount of CHBr3, followed by CHBr2Cl and CHBrCl2 (Fig. 4). The emissions of CHBr3 were influenced by the cultivation of G. manilaensis and K. alvarezii in different systems, possibly due to the varying environmental conditions present in each system. The comparison of halocarbon mixing ratios by G. manilaensis among the three cultivation systems showed that the emission of CHBr3 was enhanced in the river mouth cage culture system (267 ± 13 ppbv) compared to the onshore tanks and offshore platform, which exhibited mixing ratios of 69 ± 69 ppbv and 78 ± 47 ppbv, respectively (Table 1). An independent t-test showed that both onshore and offshore cultivation systems did not affect the emissions of either CHBr2Cl or CHBrCl2, but CHBr3 emitted by K. alvarezii was statistically higher at the onshore tanks (t(11) = 2.59, p = 0.025; Supplementary Table S2).

Fig. 4
figure 4

The net mixing ratios of CHBr3, CHBr2Cl, and CHBrCl2 released during incubation by 500 g of G. manilaensis (GM) and K. alvarezii (KA) cultivated at the onshore cultivation tank, offshore platform, and river mouth cage culture, respectively. Daylight (L) and dark (D) measurements were taken for the onshore cultivated seaweed. In this, the horizontal bars of the onshore and offshore box plots represent the median values, the boxes give the upper and lower quartile ranges, and the error bars show the spread of the data. Outlier data ( between 1.5 and 3 box lengths from the box edges) and extreme cases (> 3 box lengths from the box edges) are denoted by circles. Sample numbers, n, are included at the top right corner of the CHBr3 box plot and remained the same for CHBr2Cl and CHBrCl2. Since there were only 2 replicates for daylight G. manilaensis samples cultivated at the river mouth cage culture, these are represented as individual points in the graph

Full size image

Between daylight data from the two seaweed species at the onshore tanks, higher mixing ratios were observed in G. manilaensis, e.g., CHBr3 (128 ± 119 ppbv), CHBr2Cl (8 ± 6 ppbv), and CHBrCl2 (0.9 ± 0.8 ppbv) compared to K. alvarezii, which showed CHBr3 (88 ± 13 ppbv), CHBr2Cl (3.0 ± 0.0 ppbv), and CHBrCl2 (0.3 ± 0.09 ppbv; Table 1).

At the offshore platform, higher mixing ratios of CHBr3 (115 ± 60 ppbv) were observed during the second week of cultivation G. manilaensis compared to the first (76 ± 34 ppbv) and fell to 43 ± 16 ppbv during the fourth week of cultivation (Fig. 5). The emissions of CHBr2Cl, remained rather low and decreased from the first to the fourth week of cultivation. Emissions of CHBr3 from K. alvarezii increased throughout the cultivation period from week one, from 23 ± 3 ppbv to 29 ± 10 ppbv and reached 78 ± 21 ppbv in the fourth week of cultivation (Fig. 5). The recorded emissions of CHBr2Cl and CHBrCl2 were around one and three magnitudes lower than CHBr3, which, similar to G. manilaensis, showed a general decreasing trend throughout the cultivation period (Fig. 5). Also, throughout the cultivation period, we observed mechanical grazing of K. alvarezii by rabbit fish.

Fig. 5
figure 5

Net mixing ratios (ppbv) of CHBr3, CHBr2Cl, and CHBrCl2 released by 500 g of G. manilaensis and K. alvarezii cultivated at the offshore platform throughout the cultivation cycle. The horizontal bar represents the median value, the box gives the upper and lower quartile ranges, and the error bar shows the spread of the data; n = 2 or 3 (denoted by circles at the median)

Full size image

Difference in emissions during daylight and dark at the onshore tanks

A Mann–Whitney U test result from the onshore cultivation system showed that the median net mixing ratios of both CHBr3 and CHBr2Cl by G. manilaensis, and CHBr3, CHBr2Cl, and CHBrCl2 released by K. alvarezii were significantly higher during daylight than in the dark (Supplementary Table S3), while insignificant changes (p = 1.0) were observed in the controls for all three compounds between daylight and dark. The changes in emissions between daylight and darkness could only be attributed to the presence of seaweeds since no significant difference was noted between the daylight and dark reading of control samples (Supplementary Table S3).

The mean daylight to dark (L:D) ratios showed that halocarbon emissions by G. manilaensis in the day could be around a magnitude (13–16 times) higher than in the dark and could reach an upper bound ratio of 82–100. K. alvarezii from this study reported an average increase of three to six times the emissions of the investigated halocarbons during daylight relative to the dark, with a maximum ninefold increase observed in the emission of CHBr3 (Supplementary Table S4).

Correlations between halocarbon emissions and environmental parameters

During cultivation, G. manilaensis at the river mouth cage culture experienced the greatest variation in temperature, ranging from 24 to 43 °C (Table 2). The onshore system experienced the least fluctuation in temperature compared to the other two systems. Irradiance levels were between 0–3466 μmol photons m−2 s−1 at the cultivation sites. Seawater phosphate level was particularly elevated at the river mouth, with a significantly lower salinity level of 4 ± 0 ppt compared to the other two cultivation systems (Table 3).

Table 2 Temperature and irradiance data collected from the various cultivation systems and within the flux chamber
Full size table
Table 3 Nutrient contents and salinity readings (mean ± standard deviation) at the various seaweed cultivation systems at Bachok, Kelantan, tested using one-way ANOVA
Full size table

To further determine the relationship between environmental factors and halocarbon emissions by G. manilaensis, the emissions data from G. manilaensis grown at all three cultivation systems was pooled and plotted against the irradiance and temperature readings at the time of measurement in the flux chamber (Table 4), and all daylight emission data against the seawater nutrients and salinity where the seaweeds were grown.

Table 4 Bivariate Pearson’s correlations (r) and Pearson’s Partial correlations (rpartial) when irradiance was controlled for between the pooled halocarbon emissions by G. manilaensis with the temperature and irradiance at the time of incubation in the flux chamber
Full size table

Positive correlations were recorded between temperature and the emissions (r(23) = 0.7–0.8; p < 0.01; Table 4), and between irradiance and the emissions of all halocarbons (r(23) = 0.5–0.6; p < 0.01). The correlations between temperature and emission became weaker (rpartial(22) = 0.5–0.6; p < 0.01; Table 4) when the influence of irradiance was eliminated statistically, while correlations between temperature and the emission of CHBr3 by K. alvarezii became stronger when the influence of irradiance was eliminated. Of all the correlations between the daylight emission data and the seawater nutrients and salinity, only the emission of CHBr3 by G. manilaensis was found to be positively correlated to seawater phosphate levels (r(15) = 0.54; p < 0.05).

Discussion

Halocarbon emissions from seaweeds at the three cultivation systems

In this study data from the onshore tank arises from two different batches of seaweeds due to the separate surveys conducted, while data from the offshore and river mouth cultures were generated from the same (second) batch of seaweeds. The first survey was conducted in April 2018 as a preliminary trial, while the second survey was conducted in October 2018 to compare the different cultivation systems. The inclusion of the April data provides a better overview of the emissions from the tank culture in general, where the range of emissions is comparable to those recorded from the other two cultivation systems (Fig. 4). The cultivation of seaweeds for both batches lasted for a month.

Among the different cultivation systems, the highest data point of CHBr3 came from the river mouth where G. manilaensis was cultivated (Fig. 4). It was believed that the production of mixed halocarbons such as CHBr2Cl and CHBrCl2 resulted from the nucleophilic substitution of CHBr3 (Tokarczyk and Moore 1994; Leedham et al. 2013). As such, there is a possibility that the effect of environmental changes could be more evident in the emission of CHBr3 relative to CHBr2Cl and CHBrCl2. The higher CHBr3 could be due to the significantly higher level of phosphorus (0.39 ± 0.01 mg L−1, F(2,30) = 20.71, p < 0.001) and a rather low salinity level (4 ± 0 ppt, F(2,39) = 19.36, p < 0.001) at the river where the cage cultures located (Table 3). There were also greater temperature fluctuations (Table 2) of up to 19 °C (24.2–43.2 °C) at the river mouth compared with the offshore system (24.7–40.3 °C) and the onshore tank system (24.6–35.7 °C) and could have been due to tidal changes in the river. Although the bromide load in seawater is higher and therefore might lead to higher in situ production of CHBr3, factors such as the spatial differences in the distribution of dissolved organic matter could also affect the formation of CHBr3 (Liu et al. 2015; Shah et al. 2015). These, together with the physiological changes exerted by the different environmental parameters on the seaweeds throughout the cultivation period, may be interactive and thereby affect the ability of the seaweeds to produce bromoform. There is, however, a low sample number for the river mouth study, this is a potential area for further exploration.

At the offshore platform, where K. alvarezii was grown using the most common (monoline) cultivation method used in the Coral Triangle, mechanical grazing by rabbit fish may have contributed to the increased CHBr3 emissions observed throughout the cultivation period. The effect of mechanical stresses such as grazing and wounding on enhancing halocarbon emissions in the seaweeds had been previously reported by Nightingale et al. (1995). As K. alvarezii was tethered to the rope at the offshore platform, as opposed to G. manilaensis, which was cultivated in woven poly net baskets, rabbit fish were able to freely access and feed on the cultivated K. alvarezii. In addition, K. alvarezii has a softer texture than G. manilaensis, which may make it a more desirable diet for the fish. Lower emissions by G. manilaensis during Week 4 of culture may have been attributable to reduced stress responses to the environment as a result of adaptation.

Difference in emissions during daylight and dark at the onshore tanks

The significantly higher amounts of halocarbons released by G. manilaensis and K. alvarezii during daylight than darkness at the onshore tank systems (except CHBrCl2 by G. manilaensis) suggests a strong link between light and light-related activities. Reactive oxygen species (ROS) are commonly produced during physiological activities such as photosynthesis and respiratory electron transport (Rutherford et al. 2012; Wever and van der Horst 2013) and could be enhanced by environmental stresses (Nightingale et al. 1995). With more ROS present, more halocarbons are produced (Wever and van der Horst 2013; Abrahamsson et al. 2018). Lower temperatures in the dark could also affect the emissions by reducing enzymatic activities or slow down fluxes (Punitha et al. 2018; Macdonald et al. 2020).

The higher emissions of CHBr2Cl, and CHBrCl2 by K. alvarezii during daylight could be due to the higher amount of CHBr3 produced by the seaweed (Fig. 4). Strong correlations had been observed between the emission of CHBr3 with CHBr2Cl, and to a lesser extent, CHBr2Cl with CHBrCl2 in seaweeds (Leedham et al. 2013). During the production of halocarbons by haloperoxidase, CHBr3 is produced as a result of halide oxidation by H2O2, which increases in stressful conditions. This is followed by the reaction with ketones or dissolved organic matter (DOM) in the environment (Wever et al. 1991; Opsahl and Benner 1997; Lin and Manley 2012; Wever and van der Horst 2013; Liu et al. 2015). Brominated compounds such as CHBr2Cl and CHBrCl2 could possibly be formed through the same enzymatic pathway or through subsequent nucleophilic substitution of CHBr3 (Abrahamsson et al. 2018).

Carpenter et al. (2000) reported a tenfold increase in CHBr3 during daylight compared to dark emissions by the temperate brown seaweed Laminaria digitata. This was the largest increase of all halocarbons investigated in that study, with most of the other compounds, including CHBr2Cl, reporting at least a two-fold increase during the day. The quantification was derived from seawater measurements of the compounds (Carpenter et al. 2000). Although calculations from the previous study were based on seawater contents of halocarbons, the values were still within the range of L:D ratios observed in G. manilaensis and K. alvarezii. These ratios could provide details on the changes in emissions between dark and illuminated conditions by farmed seaweeds in the tropics, reducing the uncertainty in macroalgal halocarbon emissions arising from diurnal changes.

Correlations between halocarbon emissions and environmental parameters

There is evidence that both temperature and irradiance have an impact on halocarbon emissions. This study indicates a stronger correlation between temperature change in the flux chamber and halocarbon emissions (r(23) = 0.7–0.8; p < 0.01) compared to irradiance (r(23) = 0.5–0.6; p < 0.01). A direct correlation between seawater salinity and halocarbon emissions could not be established, even though higher CHBr3 emissions were observed at the river mouth, where phosphate levels were higher and salinity lower (Fig. 4; Supplementary Table S1). The positive correlation between seawater phosphate levels (r(15) = 0.54; p < 0.05) and the CHBr3 emission by G. manilaensis could indicate that a higher phosphate level at the river mouth could be exerting a dominant effect in driving higher emissions of CHBr3 instead of the salinity change, since G. manilaensis is tolerant to a wide range of salinity, e.g., being able to grow in fishponds at lower salinities between 23–28 ppt and was found to grow well at a salinity of 15 ppt (Mohamad Hidayat et al. 2021). However, there is an underrepresentation of low salinity river samples (n = 2) in this study.

Estimated emissions from the harvesting of K. alvarezii in Malaysia and the Coral Triangle

While data on G. manilaensis production is limited, a recent statistic from FAO (2022) showed a total production of 182,061 t (FW) of K. alvarezii in Malaysia in 2020. A simple extrapolation based on this biomass data could give an estimate of the potential contribution of K. alvarezii farming in Malaysia in terms of its halocarbon emissions. The estimated range of bromine contribution based on the release of the dominant compound, CHBr3, during the harvesting of K. alvarezii alone is between 63 and 322 mol Br for each hour of harvesting activity (Table 5). This was based on the mixing ratios of CHBr3 emitted from K. alvarezii collected from the offshore platform (Fig. 5), as this method of cultivation closely resembles the mass cultivation of K. alvarezii in Malaysia and the Coral Triangle (Lim et al. 2021). As the air samples from this study were collected from fresh and moistened K. alvarezii, this estimate may provide relevance to the potential increase in halocarbon load into the atmosphere during harvesting of the seaweeds at the K. alvarezii farms in Malaysia. However, this estimation deriving from the abundance in a closed flux chamber may not directly align with the open layer of the surface atmosphere due to various factors including wind, temperature and humidity, which influence the dispersal and distribution of gases in the open environment.

Table 5 Estimation of halocarbon emissions by farmed K. alvarezii in Malaysia, the Philippines, Indonesia, and the Coral Triangle based on total production in the year 2020
Full size table

To further explore the contribution of Kappaphycus farming to the regional halogen pool, the biomass production data from Indonesia and the Philippines were used to extrapolate the rate of halogen released by K. alvarezii during harvesting. With this, a combined emission of 3378–17,122 mol Br for an hour of harvesting activity was estimated from the harvesting of Kappaphycus in the Coral Triangle in 2020 alone (Table 5). This estimate was based on an hour of harvesting activity and the condition of the field being similar to the flux chamber within an hour. In addition to the cultivation and harvesting of seaweeds, further downstream processing, such as the post-harvest drying necessary to produce carrageenan, also contributes to the emission of volatile halocarbons such as CHBr3 (Leedham Elvidge et al. 2015).

The estimated harvesting emissions from this study could be the first estimate of halocarbon emissions of farmed seaweed in the tropics, which is an important region for the convective transport of reactive bromines into the upper troposphere and lower stratosphere. In the tropics, both G. manilaensis and K. alvarezii are red seaweeds with high CHBr3 emission rates (Keng et al. 2020). In light of the recently debated proposal to mass cultivate seaweeds with high CHBr3 emissions as livestock feed to counter the production of enteric methane, this could provide an insight into the need to balance the environmental impact of CHBr3 emissions while addressing the methane crisis (Glasson et al. 2022).

Conclusion

Measurements of cultivated seaweeds showed higher emissions of CHBr3, CHBr2Cl, and CHBrCl2 during daylight than in the dark. Daylight emissions from the farmed G. manilaensis were found to be 13 to 16 times higher than dark emissions on average, while emissions from K. alvarezii were between three and six times higher during the day. The emission of halocarbons by G. manilaensis in the different cultivation systems was affected by changes in temperature, irradiance, and possibly seawater phosphate levels as seen in the river culture, while the emission of CHBr3 by K. alvarezii was affected by temperature. Predicted values from this study showed that the harvesting activity of K. alvarezii in Malaysia in 2020 could have released 63–322 mol Br for each hour of harvesting activity, and in the Coral Triangle, 3378–17,122 mol Br.

This study provides an estimate of the halocarbon emission from farmed seaweed in the tropics, shedding light on its environmental impact. The findings underscore the need for further exploration of factors like salinity, dissolved organic matter, and emissions during cultivation, especially in mass-cultivated regions, to better assess the potential of seaweed farming-related policy as climate solutions, and enhance the prediction of climate change’s impact on halocarbon emissions by seaweeds.