Three jellyfish species Aurelia aurita (AA), Cyanea capillata (CC), and Periphylla periphylla (PP) were obtained from different locations around the Norwegian- and Baltic Sea. Jellyfish were captured by vessels during cruises or in the harbor and along the coastline of fjords. They were frozen at –20°C directly after harvesting and shipped to the facilities of the Christian-Albrechts-University in Kiel, Institute for Crop Science and Plant Breeding.

Prior to further processing, jellyfish samples were dried until constant weight. For drying, two different processes were used: (1) samples were oven-dried at 50°C for 48–72 h, depending on the species and the time needed to reach constant dry weight. Dried materials were weighed, put in laboratory glass bottles and kept in a desiccator because of their sensitivity to humidity. (2) The alcohol-dried method of Pedersen et al. (2017) was applied in which the fresh or frozen jellyfish were submerged in ethanol (70%) for 24 h. Jellyfish were then removed and placed in distilled water for 30 min, after which they were dried at room temperature (around 21°C).

The jellyfish dry matter (DM) was homogenized in a porcelain mortar before chemical analyses. The C and N content was measured via dry combustion (Vario Max CN, Elementar Analysensysteme GmbH, Hanau, Germany). For further elemental analysis, 200 mg DM of ground jellyfish material was digested with 10 mL 15.6 M HNO₃ at 190°C for 45 min in an 1,800 W microwave oven (MARS 6, Xpress, CEM, Matthews, MC, United States). After digestion, the concentrations of Ca, K, Mg, and Na were quantified with an atomic absorption spectrometer (AAS 5EA Thermo Electron S, Carl Zeiss, Jena, Germany). P concentrations in all extracts were determined photometrically with a continuous flow analyzer (Skalar Analytical B.V., Breda, Netherlands) by using the modified molybdenum–ascorbic acid blue method (Murphy and Riley, 1962). The nutrient content of different jellyfish species used for plant growth experiments are given in Table 1. With regards to species PP, only oven-dried material was used for plant growth experiments. The coefficient of variation across the different parameters measured were 1.1% C, 0.03% N, 4.3% P, 1.5% K, 2.1% Na, 1.6% Ca and 1.6% for Mg.

Plant experiments were conducted under controlled conditions in a greenhouse at the facility of Kiel University. Cylindrical pots (12 cm diameter × 10 cm height) were filled with sieved and dried soil (86% sand, 10% silt, 4% clay) to a bulk density of 1.6 g/cm³ collected from fields in the Baltic Sea Region [Eckernförde Bay (54°27N, 9°57E)]. The nutrient content per 100 g dry-soil was 20 mg P₂O₅, 7 mg K₂O and 7 mg Mg. The pH-value was 6.9. Seeds of Lolium multiflorum Lam. Var. westerwoldicum were sown at 0.5 cm soil depth at a sowing density equivalent to 8 g per m². Good growing conditions were maintained due to the constant water filled pore space of 50–60% using capillary tubes (Ortmann, Votloh, Germany). Temperature was kept constant at 22 ± 1°C. Illumination was provided by vapor-lamps (Hortilux Schreder, HPS 400W) for 10 h per day. After germination, the plants were grown until flowering stage and cut with scissors at a height of 5 cm above the soil surface. This was repeated 5 times during 2 months to obtain a fully established grass sward in each pot and scarcity of nutrients in the soil. Dried jellyfish and mineral fertilizer were applied after the fifth cut at rates equivalent to 5 g N per m². The fertilizer treatments comprised dried material from the species AA and CC, obtained by both oven- and alcohol-dried methods, and that of PP, oven-dried only. These were compared with a mineral fertilizer [Calcium-Ammonium-Nitrate (CAN)] and a non-fertilized control (CT). Thus, there were seven treatments including the non-fertilized control, with five replicates of each, and the pots were arranged in a fully randomized design (Figure 1).

There was no additional supplementation of other trace elements. Thus, the amount of other nutrients applied per treatment differed according to the chemical composition of the different sources of fertilizer. Total amounts of the major nutrients supplied to all treatments is given in Table 2.

For application to the pots, the dried and ground jellyfish material was mixed with soil to a volume of 9 mL and inserted into three small holes (0.6 cm × 0.6 cm) below the soil surface. The same procedure was used for the CAN treatment. Afterward, two more cycles of grass regrowth after cutting were performed over a 56-day period. After the second cut, following fertilizer application, all the aboveground biomass (AGB) was removed at soil-surface level and divided into leaves, stalks and stubble. The soil inside the pots was processed in a hydropneumatic elutriation system (Smucker et al., 1982) to separate roots with a 2-mm sieve, which is referred to as belowground biomass (BGB). The AGB (leaves and stalks from the first and second cut after fertilization + stubble) and BGB were oven-dried at 54°C and the samples were then milled to pass a 2 mm sieve. For the C:N-ratio, ABG and BGB were analyzed by dry combustion (Elementar, Vario Cube, Hanau, Germany). In order to calculate the apparent nitrogen recovery (ANR) of the tested treatments the N-yield of the non-fertilized control was deduced from the N-yield (g N) of the respective fertilized treatment and divided by the total N applied.

After application of the different fertilizers, the pots were sealed inside a gastight chamber (d:12.5 cm, h:19.0 cm) twice a week for 60 min. Each chamber was equipped with a vent (diameter: 0.5 mm, length: 110 mm) to allow balanced air pressure between the inside and outside chamber. Gas sampling measurements were taken over a period of 2 months. To allow gas sampling, the chambers were equipped with a rubber septum on the top. Gas samples were taken at 20-min intervals (0, 20, 40, and 60 min) and stored in 12 ml pre-evacuated septum-capped vials (Labco, High Wycombe, United Kingdom) using a 30 ml syringe. The gas samples were analyzed for N₂O through a gas chromatograph (SCION 456-GC, Bruker, Leiderdorp, Netherlands), equipped with a ⁶³Ni electron-capture detector using helium as carrier gas and argon-methane as make-up gas. Samples were injected using an autosampler (model 271 LH, Gilson Inc., Middleton, WI, United States). Data were processed using the software Compass CDS (Version 3.0.1). Fluxes were calculated for each treatment and replicate by linear regression between measured N₂O concentrations and time. The accumulated N₂O emissions were calculated by linear interpolation between the measured daily fluxes for each replication.

Soil salinity was measured conductometrically during four consecutive days after AGB removal at the end of the experiment using an electrode (Decagon Devices Inc., WA, United States). Prior to BGB analyses, soil sub-samples were taken from each pot. Soil samples were oven-dried (100°C for 48 h) and ground to mesh size of 1 mm (Modell RMO 100, Retsch GmbH & Co., KG, Haan, Germany). Chemical soil properties (C, N, P, K, Mg) were measured using ICP-OES.

For statistical computing and graphical support the statistical Software R (2020) was used to evaluate the data. Generalized least squares were used to evaluate the AGB, BGB and accumulated N₂O emissions depending on the treatment (Laird and Ware, 1982; Verbeke and Molenberghs, 2000). The data were normally distributed and heteroscedastic due to the different fertilizer treatments. All pair comparisons according to Tukey (1953) and Hothorn et al. (2008) were conducted. Significance of the tested factor and comparisons of means were declared when p < 0.05. The coefficient of variation (CV) was used as a parameter for yield stability (Sanderson, 2010).

The contents of macro- and micronutrients in the DM of the different jellyfish species are shown in Table 1. The C content in the DM varied from 3.1 to 45.1% and N content from 0.7 to 7.8% among the tested jellyfish treatments. Differences in nutrient values were mainly affected by different salt contents. The presence of salts was predominately NaCl (>90%) (data not shown). This is reflected in the high Na contents in the jellyfish DM. Thus, Na contents correlated negatively with C due to a dilution effect of organic matter by NaCl. However, Na contents showed wide variation (55–334 mg Na g^–1 DM) among the different treatments and the highest values were found for oven-dried AA. Accordingly, the oven-dried material from AA used as fertilizer led to significantly lower grass biomass yields, and this treatment had the lowest biomass accumulation for both AGB and BGB (p < 0.05). All other jellyfish treatments showed higher DM yields of AGB, compared with the control, with no differences compared with the CAN treatment (p > 0.05). The highest DM yields, on average, were obtained in the alcohol-dried AA treatment. Thus, the range of total DM yields in all treatments were in the order: AA oven-dried [0.59 g DM (SE 0.07)] < control [1.47 g DM (SE 0.21)] < PP oven-dried [2.13 g DM (SE 0.50)] < CC alcohol-dried [2.40 g DM (SE 0.40)] < CC oven-dried [2.8 g DM (SE 0.31)] < CAN [2.98 g DM (SE 0.24)] < AA alcohol-dried [3.22 g DM (SE0.26)] (Figure 2A). The highest yield stability was observed for alcohol-dried AA and CAN (CV 0.18). Oven-dried AA (CV 0.24) and CC (CV 0.25) were comparable to the non-fertilized control (CV 0.32). The highest yield variability was observed for alcohol-dried CC (CV 0.40) and oven-dried PP (CV 0.54).

The N-yields followed the same patterns as the DM yields (Figure 2B). The ANR between the treatments was negative in the oven-dried AA treatment and lowest, yet positive, in the oven-dried PP (0.15) treatment. The highest ANR values were observed for alcohol-dried (0.23) and oven-dried (0.29) CC, CAN (0.38) and alcohol-dried AA (0.41).

After application of fertilizers, N₂O-fluxes reached their maximum within 2 weeks after application, with lowest fluxes for the non-fertilized control and CAN. The highest fluxes were measured in CC and AA alcohol-dried (Figure 3). However, for oven-dried AA and CC the fluxes were maintained over a longer period than all other treatments. As a result, the accumulated N₂O emissions during the study period were lowest in the control and highest in the AA and CC treatments irrespective of drying method used. With the exception of oven-dried PP and alcohol-dried AA, all jellyfish species treatments showed significantly higher (p < 0.05) accumulated N₂O-emissions compared with that of CAN (Figure 4).

Soil chemical properties after the experiment showed lowest pH values (7.2) in the PP oven-dried treatment (p < 0.05); however, this difference was of a low extent. The pH value of all other treatments did not differ and showed an average pH of 7.4 (Table 3). Soil C:N ratio did not show any differences between treatments at the end of the experiments. Soil nutrient contents showed no differences for P; however, for K, the amounts were lowest for AA alcohol-dried and highest for Mg and K in the AA oven-dried among all other treatments. Salinity was highest in the oven-dried AA and CC treatments; however, with more than threefold higher amounts of Na in AA oven-dried in comparison to all other treatments.

The analysis of nutrient contents in the jellyfish DM showed comparable amounts of N to that of other organic fertilizers derived from animal husbandry (Table 4), but P contents were slightly lower and Ca contents were 2-times lower, whereas K content was 6–10-fold lower and Mg contents up to 2-fold higher (LWK-SH, 2019). The C:N ratio was narrower compared with that of most organic manures leading to a potentially higher turnover rate in the soil after application and thus a high plant availability of applied nutrients in the very short-term. Microelements such as Mn, Cu, and Zn were present in variable amounts (data not shown) and generally sufficient for plant nutrition, but mainly below the amounts commonly expected in other types of organic manure, such as digestates from biogas plants (Möller and Schultheiß, 2014; Emadodin et al., 2020a). The results for heavy metals such as Al and Fe were low (data not shown), however, other toxic elements were not analyzed in this study even though marine biomass can show a high potential for accumulating anthropogenic emissions from water, which was e.g., shown for pelagic Sargassum along the Mexican Caribbean (see Rodríguez-Martínez et al., 2020).

The different drying methods affected the nutrient contents in the jellyfish DM. This was mainly due to differing salt levels after drying, and means that high NaCl levels result in a lower content of organic matter and thus lead to lower measured nutrient contents in the dry matter. During the alcohol-drying, NaCl became dissolved and hence its content in the jellyfish DM was reduced. This assumption was supported by the different measured Na levels and C contents of the dried materials, with lower Na contents and higher C contents in the alcohol-dried treatment (Table 1). However, the desalination effect of this drying method was different between the different jellyfish species, with a lower success for CC. This leads to the conclusion that alcohol-drying should not be universally applied for desalination among different jellyfish species, and other techniques, such as dialysis, might be more appropriate (Kogovšek et al., 2014). However, it should be mentioned that the fresh matter of CC already showed 10-times lower initial Na values compared to AA, indicating a lower success of desalination during the alcohol-drying as the Na content were already on a lower level.

The plant growth in our study was negatively affected by high Na levels only in the treatment with oven-dried AA. However, if Na can be successfully reduced, as was shown for AA, jellyfish as an organic fertilizer showed similar plant growth responses to that of CAN when supplied at the same level of N input. Moreover, after reducing salinity, the AA treatment gave higher biomass yields, on average, than CAN, which may be explained by additional nutritional effect from P, K and other microelements (Table 2). The plant growth is dependent mainly on the availability of the macro-nutrients such as N, P, and K. However, micro-nutrients might also inhibit or promote plant growth, if their supply is below/higher the necessary minimum/maximum nutrient supply. Organic biomass is usually rich in many nutrients. Thus, even though in this experiment we accounted for N, we did not control for other elements, probably causing differences in the N-yield response of the different treatments. In contradiction to that, the calculated ANR in this study was below the expected values given in the literature for mineral and other organic fertilizers, which usually exceeds 50% (Nannen et al., 2011). This can be explained by the high frequency of cutting intervals without excessive fertilizer amendments prior to the application of jellyfish. The scarcity of nitrogen, which is indicated by a high C:N ratio (15:1) at the end of the experiment in the soil of all treatments, potentially provoked an intensive N-immobilization (Kuzyakov et al., 2000). Moreover, nutrient deficits led to a shift in plant functional traits of grass, allocating more energy to roots rather than the shoot (Chen et al., 2016), suppressing the AGB allocation of N during the weeks after fertilization. However, a comparison of the different treatments shows that the AA alcohol-dried treatment revealed a similar ANR to the mineral fertilizer (CAN), indicating a fast decomposition of material due to the narrow C:N ratios. This is indeed an advantage relative to other types of organic fertilizer, where a high ratio of N is bound in the organic matter (N_org) and only available in mid-term (Gutser et al., 2005). These facts might potentially open possibilities in extensive land use systems, where soil nutrient deficits are present and plant available nutrients are needed in the short-term after application.

The results presented here are the first of several consecutive experiments. The amount of jellyfish, which can be dried at lab scale is limited by the size of equipment available. In this study, after drying and desalination, the processed material was only sufficient for five replicated pots per treatment. As there are large differences among different jellyfish species, a higher number of replicates should be considered in any further experiments. However, the effects of the two pre-processing methods were significant relative to the control, thereby presenting evidence of the potential of jellyfish as a source of fertilizer and of its positive effect on plant growth.

Several other studies have indicated elevated N₂O fluxes after application of C- and N-rich organic matter in comparison with mineral fertilizers, when water-filled pore space exceeds 50% (Senbayram et al., 2009). This elevation is mainly attributed to a higher activity of soil denitrifiers using C-rich material as an electron source and N as an electron acceptor, when oxygen is limiting (Davidson et al., 1986). However, this is also heavily dependent on the plant-N demand, environmental circumstances and experimental design. The experimental comparisons in plant experiments on N₂O emissions are often related to either the mineral phase of N (NH₄ + NO₃) in the organic matter or the total N amount (NH₄ + NO₃ + N_org). For the latter, other experiments have also found lower N₂O emissions from organic fertilizers in comparison with mineral fertilizers (Hansen et al., 2014) as N_org is not available for soil N₂O-producing bacteria in the short-term. However, increased N₂O-fluxes from soil can be triggered additionally when the C:N ratio is narrow (Biernat et al., 2020). The C:N ratio of jellyfish was below 5, which is a lower value than reported for other marine debris (Emadodin et al., 2020a), explaining the relatively high measured N₂O-emissions even if the total amount of N applied is considered. Consequently, the hypothesized reduction of N₂O-emissions from jellyfish-based fertilizers due to other metabolites (Yu et al., 2005, 2015) was not confirmed for AA and CC, although we observed significantly lower emissions from PP. However, recent observations have found that integrated fertilizer strategies, which combine mineral and organic fertilizer sources, could increase the production efficiency with only a minimum amount of N₂O released, and this applies in tropical and temperate climates (Nyamadzawo et al., 2017). Thus, approaches using jellyfish biomass with a narrow C:N ratio together with other sources of nutrients that have a higher C:N ratio (e.g., seagrass or seaweed) could be considered for coastal areas in the context of increasing overall fertilization and nutrient-use efficiency, while additional N₂O emissions from applied organic matter can be kept to a minimum.

Nevertheless, when considering marine biomass as an alternative resource for supplying plant nutrients with low environmental impact in terms of GHG emissions, the effect of N₂O emissions from soils has to be considered in a full life-cycle-assessment (LCA), as on-farm emissions in agricultural production are responsible for the majority of emissions associated with agricultural commodities (Poore and Nemecek, 2018). From the perspective of LCA, marine biomass as a source of fertilizers may be beneficial in comparison with mineral fertilizer due to its relatively low energy use during production of raw material (e.g., harvest) (Seghetta et al., 2017; Halfdanarson et al., 2019). However, this advantage only applies when GHG emissions that result from land application (e.g., N₂O), in addition to the emissions from the process chain, do not exceed those of mineral fertilizers (Christensen et al., 2014). There are additional expected GHG emissions associated with using jellyfish as a fertilizer arising from its transport, drying and desalination. Therefore, the results from measured N₂O-emission can be crucial in evaluating the suitability of this raw material as a fertilizer in view of the very high global warming potential of N₂O, at 265 times that of carbon dioxide (IPCC, 2014). Finally, an effective evaluation of the full potential and benefits of valorizing jellyfish into fertilizers and thus its contribution to the circular bioeconomy, requires three additional aspects to be taken into consideration. First, any of the pre-treatment methods, such as desalination, should be cost-efficient. Second, two levels of sustainability must be guaranteed: (i) sustainable supply of biomass regardless of its seasonal and geographical variability and (ii) sustainable transporting and storage methods which will not impact the quality and efficiency of the final product. Third, this study has shown a trend for jellyfish inter-species variability in terms of their nutrient contents, subsequent grass biomass yields after fertilization and N₂O emissions. Therefore, to advance to a higher technological readiness in terms of sustainability goals, further studies will be obligatory. Unfortunately, our study was not designed to cover all these specifications; however, to provide a more holistic view on the jellyfish drying treatments used during this study, as well as lyophilization as an alternative to oven-drying, we constructed a decision matrix (Table 5). We selected six overall categories to help determine the two best drying methods in comparison with using CAN as the source of N fertilizer. The overall cost was selected as an important criterion, as it affects the competitiveness of fertilizers in terms of market value. The ranking included the known prices for CAN, ethanol and the estimated need for infrastructure and materials for other drying methods (oven-drying, lyophilization, dialysis, salt + aluminum). Energy demand was selected as the second selection criterion, as techniques that are highly demanding in terms of energy are not competitive in the market as they inflate the final fertilizer price. As salinity might be a plant growth-inhibiting factor, the potential salinity of the fertilizers (CAN in comparison to other jellyfish drying methods) was theoretically ranked based on the expected desalination capacity. The nutrient content, N₂O emission potential and plant growth effect were ranked according to the findings in this study and on the assumed soil plant interaction controlled by inhibited/increased N₂O emission due to increased/inhibited plant-N-uptake. In summary, for this matrix we found the lowest score and thus the highest efficiency for CAN. However, oven- and alcohol-drying turned out to show only three- and twofold higher values compared to CAN and lower values compared to more energy intensive procedures such like lyophilization. The salt + aluminum drying adapted from the food industry acted in this context as a negative example as it is completely unsuitable for fertilizer preparation due to elevated salt contents as confirmed by the decision matrix. The matrix also makes clear that a higher readiness for commercial use can be generated by the reduction of energy use and costs. For instance, the market price for 1 t calcium-ammonium-nitrate (27%N) is 173–225 € (0.64–0.83 € per kg N). We estimated operation costs of 20 € for drying one 1 kg jellyfish biomass to dry weight. Taking an average N-content of 2.6–7.2%N into account (Emadodin et al., 2020a, b) the price for 1 kg N accounts 769–278 €. Thus, at high N contents, the costs are comparable yet still higher compared to mineral fertilizers. For alcohol-drying we assume that costs are about threefold compared to oven-drying, however, but offering other advantages with regards to reduced salinity and consequently higher yields and potentially lower N-losses compared to oven-drying. Therefore, technical methods for cost efficient processing need further exploration.