1 Introduction

Marine litter is a common concern of humankind with severe environmental, social and economic consequences. One of its biggest and most damaging sources is derelict fishing gear (DFG). DFG is defined as abandoned, lost or otherwise discarded fishing gear in the ocean. It mainly arises from bad weather incidents or gear conflicts, but human errors and illicit disposals also take place (Richardson et al. 2019a, b). While fishing gear losses can be estimated based on gear types (Loubet et al. 2022), the annual global inflow of DFG into the ocean is generally projected at 640 thousand tonnes (Macfadyen et al. 2009, but see Richardson et al. 2021). This is problematic because DFG is made of durable polymers that can continue to catch marine life for years (Tschernij and Larsson 2003). It can also suffocate corals (Kühn et al. 2015), lead to navigational accidents (Cho 2005) and facilitate invasive species to spread (Kiessling et al. 2015). After time, DFG breaks down into smaller pieces and can be mistaken for food (Gall and Thompson 2015). While this itself can be fatal, it can also transport contaminants into the human food chain (Ribeiro et al. 2019). Therefore, and to achieve the Sustainable Development Goal 14.1, to “prevent and significantly reduce marine pollution” by 2025 (United Nations General Assembly 2015), preventive and curative work is increasingly undertaken.

Preventive approaches include fishing gear markings (He and Suuronen 2018), buy-back programmes (Cho 2005) and extended producer responsibility (EPR) schemes (European Commission 2018). Although they are generally less expensive than clean-up efforts (Macfadyen et al. 2009), a complete protection against gear losses or littering is unlikely to be achieved (Schneider et al. 2022). Thus, to reduce the current accumulation of DFG, curative clean-up efforts are essential alongside preventive techniques.

Curative clean-up efforts are conducted in areas of DFG accumulation (Schneider et al. 2018). One of the most prominent examples is the North Pacific Ocean Gyre, where the Ocean Cleanup installed a floating barrier to collect 3640 tonnes of DFG per year (van Giezen and Wiegmans 2020). Large clean-ups are also carried out in other parts of the world, having removed over 15 thousand tonnes of DFG so far (Cho 2011; Richardson et al. 2019a, b; WWF Poland, 2015; Deshpande et al. 2020). The retrieved DFG typically comprises of partly degraded polymers, lead lines and steel as well as contaminants like sand, salt, organic matter and litter that has been picked up in the sea (Stolte and Schneider 2018). This material mix makes a waste treatment difficult so that DFG is frequently landfilled (Schneider et al. 2018). Yet, as a disposal of toxic lead and other resources is of great environmental concern and because improperly managed landfills can be a source of marine litter (Quantis and EA 2020), there is a need to identify more suitable waste treatment options.

Apart from disposal in landfills, DFG waste management options can be divided into a reuse, recycling and energy recovery (Directive 2008/98/EC 2008). Reuse has been reported for some components like pots and ropes (Northwest Straits Foundation 2019; Charter et al. 2018), but only recycling and thermal processing with subsequent energy recovery provide large-scale solutions for this entangled mixed waste. From those, DFG incineration with energy recovery has been established in South Korea and in the USA (McCoy 2010; Jung et al. 2010), whereas industrial recycling applications have yet to emerge. Still, research has shown that mechanical recycling via extrusion and thermal processing via gasification are technically feasible for DFG (Stolte and Schneider 2018; Weißbach et al. 2021). To help countries implement an environmentally sustainable waste management system for DFG, there is a need to assess the potential environmental impacts of available large-scale treatment pathways in a quantitative way.

Life cycle assessment (LCA) is an internationally standardised method (ISO 14040/44) for the evaluation of potential environmental impacts that occur in process chains. Although some areas of its methodology are debated (McManus et al. 2015), its solid methodological foundation coupled with its ability to compare processes, to identify environmental hotspot areas and to track burden shifts (Finnveden et al. 2009) allowed LCA to become one of the most accepted environmental performance tools today. This and its frequent application in the waste management field (Khandelwal et al. 2019) make LCA a suitable tool to evaluate treatment options for DFG.

To the authors’ knowledge, previous LCA studies have not yet been conducted for DFG. However, Kneppers and Laruccia (2021), Storm (2017) and Nofir (2015) carried out LCA studies on waste fishing gear that highlighted the environmental superiority of a mechanical recycling over a virgin polymer production and disposal system. Yet, as the assumptions and inventory data were not clearly presented, it is difficult to meaningfully assess those results. In related fields, more comprehensive LCA studies have been completed for mixed plastic waste (e.g. Faraca et al. 2019; Rigamonti et al. 2014; Shonfield 2008); municipal solid waste (e.g. Ripa et al. 2017; Quirós et al. 2015; Al-Salem et al. 2014) and waste from electrical and electronic equipment (e.g. Wäger and Hischier 2015; Biganzoli et al. 2015). In general, those studies supported the waste hierarchy, meaning that recycling was preferred over energy recovery, which was preferred over disposal, respectively (Directive 2008/98/EC 2008). Yet, as LCA studies depend on distinct conditions such as the waste composition, the energy mix and transport distances, general outcomes should not be assumed (Laurent et al 2014). Especially for a novel waste stream such as DFG, there is a need for a new case-specific LCA.

This study presents the first LCA on waste treatment options for DFG. Based on industrial experiments, mechanical recycling, syngas production and energy recovery are compared with a default landfill disposal. Novel life cycle inventory data is presented for the pre-treatment and gasification steps. A contribution analysis is performed to identify areas of significant environmental impacts and a sensitivity and uncertainty analysis are conducted to investigate the influence of input parameters on the results. The results are discussed and recommendations on possible future DFG waste management systems are given. Overall, this study identifies the least environmentally harmful waste treatment pathways for DFG so that an environmentally sustainable waste management system can be established for this growing waste stream.

2 Methods

Following ISO 14040 (International Organisation for Standardization 2006), this study applies an attributional LCA. In accordance with the standard, this section covers the goal and scope definition, inventory analysis and impact assessment. The interpretation of the results is provided in the results and discussion chapters.

2.1 Goal and scope definition

The aim of this study is to compare the potential environmental impacts of waste treatment pathways for DFG from the Baltic Sea so that decisions on a suitable DFG waste management infrastructure can be made in Europe. DFG arises in small quantities which do not cause large-scale consequences on the background system, like the electricity production mix. According to the ILCD handbook (European Commission 2010), this situation can be described as a micro-level decision support. To solve multifunctionalities, this study implements a system expansion approach based on average data.

2.1.1 Scenarios and functional unit

The investigated DFG treatment pathways are (1) landfill disposal, (2) syngas production, (3) energy recovery and (4) mechanical recycling. To quantitatively compare these scenarios, the waste treatment of 1000 kg of DFG was chosen as the functional unit. The waste composition of DFG varies depending on the type of fishing gear that is retrieved from the ocean. Based on previous experiments from the Baltic Sea that mainly retrieved trawl netting and gillnets (Stolte and Schneider 2018), an average waste composition was established in Table 1.

Table 1 Indicative DFG waste compositions
Full size table

2.1.2 System boundaries

Starting from the retrieval, the system boundaries consider all transport and waste treatment processes until the DFG waste composition is either landfilled or used to substitute similar products or energy. An overview of the scenario specific processing pathways is given in Fig. 1.

Fig. 1
figure 1

Process flow chart for the investigated waste treatment scenarios

Full size image

Apart from the retrieval and transport processes that are modelled in all scenarios,

  • The disposal scenario also includes a landfill step

  • The syngas production a sorting, shredding and gasification step

  • The energy recovery a sorting, shredding, density separation and incineration step

  • The mechanical recycling a sorting, shredding and two-stage density separation as well as a washing, drying and extrusion step (Fig. 1)

Steel and lead recycling are modelled for all scenarios except for the worst-case disposal scenario. Within the mechanical recycling scenario, it is assumed that residual polymers in form of process losses and non-nylon fractions are incinerated. In the LCA model employed here, the extrusion, gasification and incineration processes substitute average nylon, syngas and energy production processes, respectively, whereas the steel and lead recycling replace average steel and lead production processes (Fig. 1).

Fish and possibly other organic materials are separated during the retrieval and returned to the ocean. This was not included into the modelling because the degradation processes would have also occurred naturally. The landfilling of minerals and other residues was not modelled because previous studies showed their low environmental significance over a period of 100 years (Birgisdóttir et al. 2007). Similarly, capital goods like roads and machines only caused minor environmental impacts in previous waste management studies (e.g. Brogaard et al. 2013), which is why capital goods were not included into this work. It was the aim to model the system boundary as complete as possible so that cut-offs were not applied. To ensure the comparability of the waste treatment scenarios, inventory data for large-scale, state-of-the art DFG waste treatment facilities in Europe were sought. However, as such facilities are not yet established, available datasets from similar processes needed to be adapted to form the life cycle inventory.

2.1.3 Impact methodology and impact categories

Due to its wide-spread use (Khandelwal et al. 2019) and robustness across different impact categories (European Commission 2011), ReCiPe is selected as impact methodology (Goedkoop et al. 2013). From ReCiPe’s 18 mid-point level impact categories, the following 12 are chosen for this study: (1) climate change (CC), (2) terrestrial acidification (APt), (3) freshwater eutrophication (EPfw), (4) marine eutrophication (EPm), (5) photochemical oxidant formation (POF), (6) human toxicity (HT), (7) terrestrial ecotoxicity (ETt), (8) freshwater ecotoxicity (ETfw), (9) marine ecotoxicity (ETm), (10) water depletion (WD), (11) metal depletion (MD) and (12) fossil depletion (FD). The main reasons for selecting those are their robust modelling frameworks (European Commission 2011) and their common application in related scientific literature (Khandelwal et al. 2019). As approach to uncertainty, ReCiPe’s hierarchist model with a 100-year time frame is selected to facilitate the comparability with other impact categories.

2.2 Inventory analysis

In collaboration with WWF Germany and the waste management company PreZero Deutschland KG, the life cycle inventory was formed based on retrieval and waste treatment experiments between 2016 and 2019 (Schneider 2020). In case primary data could not be obtained, secondary literature sources were used. While a detailed life cycle inventory is provided in the supplementary materials (Table S1–S18), the most relevant modelling choices are explained below.

2.2.1 Retrieval

DFG retrieval actions in Northern Europe are frequently conducted through dragging operations from fishing vessels. While dragging operations can cause seafloor scouring and sediment suspension with potentially negative effects on the surrounding environment (Sahlin and Tjensvoll 2018), the LCI for the retrieval (Table S1) only considers the vessel operation in relation to its time at sea. Based on recordings from DFG retrievals in the Baltic Sea, it was estimated that it takes approximately 16.3 h to retrieve 1000 kg of DFG (Schneider 2020). A typical lubricant oil, antifouling and boat paint consumption (Villanueva-Rey et al. 2017) and a diesel consumption of 35 L/h (K. Neumann, fishermen, personal communication, 19 July 2017) were assumed. Airborne emissions from burning fuel were adapted from the EMEP/EEA air pollutant emission inventory guidebook (EMEP/EEA 2016a), and two thirds of the antifouling and boat paint were modelled to be released to the water (Hospido and Tyedmers 2005, Table S1). After retrieving DFG, dead and live fish are manually separated from DFG on board of the fishing vessel and returned to the water. While smaller amounts of seaweed and mussels remained in the nets, as a simplification, it is assumed that 100% of DFG’s organic material is removed.

2.2.2 Transport

Road transport is needed to bring DFG to different waste treatment facilities (Fig. 1). It is assumed that pre- and main treatments are carried out in a centralised waste management facility making road transports unnecessary between those steps. The remaining transport journeys are projected to take place in a Euro 6 heavy-duty vehicle with a capacity of 3.5–7.5 t and with an average load of 0.98 t (Spielmann et al. 2007). The LCI is modelled to be distance dependent and as weighted transport distances 20.1 km for the disposal, 546 km for the syngas production scenario, 546 km for the energy recovery and 548 km for the mechanical recycling were determined (Table 2). For the construction of the life cycle inventory (Table S2), information about ancillary materials and emissions to air from combustion as well as from tyre and brake wear was adapted from the EMEP/EEA air pollutant emission inventory guidebook (EMEP/EEA 2016b, c).

Table 2 Calculation of weighted transport distances for each scenario
Full size table

2.2.3 Sorting

The sorting involves the stretching of DFG as well as the detection and removal of large metal scrap and rocks. A steel separation efficiency of 79.2% was determined during industrial experiments (Schneider 2020) and an electricity consumption of 0.184 Wh/kg for a crane and of 0.176 Wh/kg for an angle grinder to cut metal were determined based on processing times and manufacturer datasheets (Table S3).

2.2.4 Shredding

The shredding transfers bulky DFG into smaller and more homogenised parts. It is divided into a rough and fine shredding stage that generate fibres of approximately 120-mm and 30-mm length, respectively. Each stage is coupled with a magnet to separate metal scrap. For this, a steel separation efficiency of 100% is assumed as found in previous research (Unger et al. 2017; Tunesi et al. 2016). An average electricity consumption of 25.0 Wh/kg for the rough shredding and of 64.5 Wh/kg for the fine shredding were determined during the recycling trials carried out at Vecoplan AG, Bad Marienberg, and used for the life cycle inventory (Table S4).

2.2.5 Density separation

The density separation consists of two stages. During the first stage, saline water with a density of approximately 1.15 g/cm3 is used to separate polymers from denser lead and mineral fractions. This is crucial because DFG originating partially from gillnets contains lead weights from sink lines and both gillnets and trawl netting contain significant amounts of residual sediments after retrieval from the seafloor. The second stage uses plain water to separate the less dense polymers polypropylene and polyethylene from the targeted nylon fraction. A focus was placed on nylon due to its high market value and abundance in DFG. Based on industrial experiments and data from a decanter (Andritz Separation n.d.), a separation efficiency of 90% for minerals and of 100% for lead at the first stage and of 100% for polymers in the second stage were assumed. The water, salt and electricity consumption were modelled as 0.3 L/kg, 84.4 g/kg and 68.2 Wh/kg, respectively (Table S5), where the salt consumption was only required during the first density separation stage.

2.2.6 Washing

The modelled washing applies water, friction and radial forces to remove minerals from the targeted polymer fraction. The wastewater is filtered to remove minerals and residual polymers, cleaned and returned to the process. Based on the recycling trials (Schneider 2020), a polymer loss of 15% and a complete removal of minerals is assumed, whereas the electricity and water consumption were estimated based on expert opinion as 100 Wh/kg and 2 L/kg respectively. For the wastewater treatment, approximately 1 ml/kg of an aluminium hydroxide-based coagulation agent and approximately 0.1 g/kg of a cationic resin-based flocculation agent were applied (U. Kramer, Vecoplan AG, personal communication, 27 March 2019) and included into the life cycle inventory (Table S6).

2.2.7 Drying

The drying is necessary to remove water from washed DFG fibres. Based on data for a conductive paddle dryer from Andritz separation, the process consumes approximately 28.4 Wh/kg of electricity and 284 Wh/kg of heat (M. Maingay, Andritz Gouda B.V., personal communication, 02 April 2019, Table S7).

2.2.8 Extrusion

The extrusion of dried DFG fibres is modelled to take place in an Intarema TE 1310 single screw extruder, which is used for similar fishing net recycling schemes (EREMA 2015). The process involves a preparation, extrusion and strand cutting as well as a screening, drying and water cooling. Based on small-scale experiments, it is estimated that 98.2% of the input polymers become recycled pellets (Schneider 2020), which were assumed to substitute an average nylon production at a ratio of 1:1. The remaining 1.8% of the input materials are turned into landfillable filter residues or condensate. The electricity consumption of the whole process was estimated as 0.372 kWh/kg (R. Binder, EREMA Ges.m.b.H., personal communication, 29 March 2019). The water consumption was described as negligible and thus not included in the life cycle inventory (Table S8).

2.2.9 Gasification

In the case that mechanical recycling is not feasible because of material contamination, the next level in the waste hierarchy is waste-to-energy generation. This can lead to the generation of synthetic, hydrogen-rich gas (“syngas”) or to pyrolysis condensate with the potential to be used as fuel. The advantage of thermal waste processing as compared to recycling of DFG is that energy and technology intensive cleaning and drying processes are not required. The gasification of shredded DFG turns polymers into syngas, while completely separating it from minerals and lead (Fig. 1). After quenching, the syngas is directed through an acid and basic scrubber to remove heavy metals and halogens (CleanCarbonConversion 2022). Based on the chemical composition of shredded DFG (Schneider 2020), 0.55% of the input weight are separated in the scrubber, requiring 4.12 g/kg of lime and 1.56 mg/kg of hydrochloric acid for the neutralisation (Table S9). The nitrogen consumption for flushing out oxygen was estimated as 40.8 g/kg, while the water and electricity consumption were determined as 0.35 L/kg and 1.65 kWh/kg (Table S9, Schneider 2020). The experiments showed an average syngas generation of 0.418 m3 per kg DFG (Schneider 2020), which was modelled to replace a biomass-based syngas production at a factor of 1:1 (Table S9).

2.2.10 Incineration

In case of contaminated material or that facilities for thermal waste-to-energy processing are not available, incineration with or without energy recovery remains the next viable waste management option for DFG in the waste hierarchy. Under the European waste framework directive, DFG should only enter incineration plants after lead as a form of hazardous waste was removed. This implies that even before incineration, at least one manual pre-sorting stage is required. In addition, DFG has to be cut or shredded into fragments of less than 1 m in diameter to ensure safety in the transport lines inside the incineration facility. These preparation steps are usually carried out manually and are not quantifiable in terms of energy consumption. However, the energy required for the incineration process itself dominates this processing method. The life cycle inventory for the incineration of DFG is based on an average European incineration plant in which a joint heat and electricity production, a wet scrubbing and a selective catalytic reduction take place (BREF, 2006). The two materials that enter the incineration plant are DFG4 and polymer residues from process losses or non-nylon fractions (Fig. 1). The lower heating values of those materials were used to calculate the steam and electricity generation (Table S10), assuming a thermal and electrical efficiency of 37% and 15% (Laner et al. 2015). In addition to the mineral fraction, 7% of DFG4 and 0.8% of the polymer residues were modelled to become residues, assuming a full removal of heavy metals and halogens (Schneider 2020). As presumably equally common neutralisation agents, the consumption of lime and sodium hydroxide were determined based on the measured chlorine and sulphur content of the input materials (Table S11). Carbon dioxide emissions were calculated from the carbon content in shredded DFG (Table S12), while other emissions and ancillary materials were adapted from the EU Best Available Techniques Reference documents for waste incineration (BREF 2006) and the EMEP/EEA air pollutant emission inventory guidebook (2016d; Table S13).

2.2.11 Landfill

In many areas of the world, and even in some European countries, landfill is the only option for fishing gear waste. Therefore, the impact of landfilling DFG is considered here for comparison with the processing methods along the waste hierarchy. A sanitary landfill with no leakage of DFG debris to the environment is assumed. The landfill inventory (Table S14) considers the landfill closure, the leachate treatment and the generation of heat and electricity from landfill gas (LFG) over a period of 100 years. Based on previous research (Koroneos and Nanaki 2012), machines for the landfill closure were modelled with diesel consumption of 40 MJ/t. The electricity consumption for the leachate treatment was established as 0.694 Wh/kg (Table S14), assuming an estimated leachate generation of 2.5 L/kg waste (Doka 2003a) and a relative electricity consumption of 1 kJ/L (Koroneos and Nanaki 2012). The electricity and heat generation were calculated as 0.705 Wh/kg and 0.347 Wh/kg, respectively (Table S15). This assumes a carbon content of 55.4% in polymers (Othman et al. 2008), a 100-year polymer degradation rate of 1% (Rossi et al. 2015), a carbon to landfill gas ratio of 97.1% (Doka 2003a) and a landfill gas capture rate of 49% (Rossi et al. 2015). It also assumes a methane content of 56% in landfill gas (Doka 2003a), a lower heating value of 50 MJ/kg for methane and a thermal and electrical conversion efficiency of 3.0% and 6.1% (Rossi et al. 2015). The carbon- and lead-related emissions to water and air were linked to the DFG waste composition (Table S16). For this, a leachate capture rate of 85% (Koroneos and Nanaki 2012), a conversion of leachate contained carbon into 24.5% airborne CO2, 65.8% landfillable residues and 9.7% water bound TOC (Doka 2003b), and a 0.1% and 0.033% release of lead into leachate and landfill gas (Koroneos and Nanaki 2012; Doka 2003a) were further assumed. Other emissions to air and water were adapted from a study on municipal solid waste (Department for Environment, Food and Rural Affairs 2004).

2.2.12 Steel recycling

During pre-sorting and manual extraction of metals and lead, steel is commonly removed from DFG before further processing. After extraction, the steel fraction can readily be recycled. The steel recycling is modelled to take place in an electric arc furnace, covering the smelting, refinement, off gas treatment and residues treatment. The steel scrap that was contained in DFG mainly originated from metal chains and anchors that showed clear signs of corrosion. Therefore, a recycling process for low-alloyed steel was assumed. The life cycle inventory (Table S16) was adapted from the EU Best Available Techniques Reference documents for iron and steel production (BREF 2013) and the recycled low-alloyed steel was assumed to replace an average low-alloyed steel production at a ratio of 1:1.

2.2.13 Lead recycling

During pre-sorting, the sink lines containing lead weights need to be removed. As a hazardous heavy metal, lead is not commonly acceptable in thermal processing or polymer recycling facilities. After extraction and separation from the polyester sheathing, lead can readily be recycled. To represent a typical situation in Europe, the lead recycling is modelled as part of a secondary lead production facility with a rotary furnace (BREF 2017). It includes a smelting and refinement as well as an off-gas and water treatment. The lead from DFG is present in its metallic form with little contamination and therefore a direct conversion into recycled lead without material losses is assumed (also see Davidson et al. 2016). Due to its virgin-like material properties, recycled lead was presumed to replace an average lead production at a ratio of 1:1. Information about ancillary material and energy consumption as well as air and water emissions (Table S18) derives from the Best Available Techniques Reference documents for the non-ferrous metals industries (BREF 2017) and the EMEP/EEA air pollutant emission inventory guidebook (2016e).

2.2.14 Background processes

The life cycle inventory for background processes such as the production of Diesel was taken from the Ecoinvent database. One exception is the electricity production, which was modelled separately to represent the 2017 situation for Germany (Frauenhofer 2022). It is based on the following energy mix: hydro (3.7%), biomass (8.2%), wind (19.3%), solar (7.2%), natural gas (9.0%), lignite (24.5%), hard coal (14.9%) and nuclear (13.2%).

2.3 Impact assessment

2.3.1 Classification and characterisation

The LCA software SimaPro was used to perform the mandatory classification and characterisation step. To facilitate a replication of this research, the Ecoinvent process selection choices (Table S19) as well as the characterisation results for the modelled processes (Table S20–S33) were included in the supplementary materials.

2.3.2 Contribution analysis

A contribution analysis compares the potential environmental impacts of the waste treatment scenarios and investigates where hotspot areas occur. For this, the characterisation results were exported into an Excel file (ESM2: T1-2), where basic calculations and graph plotting operations took place. To prevent a cancellation of values with opposing mathematical signs, the environmental impacts from auxiliary materials and process emissions were treated separately from avoided production processes. Inventory data that contributed less than 5% to the impact contribution were aggregated into a rest group (Fig. 2).

Fig. 2
figure 2

Characterisation results for the landfill disposal (LD), syngas production (SP), energy recovery (ER) and mechanical recycling (MR); yellow diamonds show net environmental impacts; and error bars represent 95% confidence levels of the Monte Carlo uncertainty analysis results. a)For better readability, water depletion error bars and net results were scaled down by a factor of 20

Full size image

2.3.3 Uncertainty analysis

To determine the overall uncertainty of the LCA results, an uncertainty analysis through sampling-based Monte Carlo simulations was performed (Bisinella et al. 2016). As probability distribution for the foreground input values, a triangular distribution with minimum, maximum and average values was assumed (also see Faraca et al. 2019). Foreground input values without known min- and maxima were transformed into a lognormal distribution following the Pedigree approach (Weidema et al. 2013). For each scenario, a Monte Carlo simulation was conducted with 10 thousand iterations within SimaPro. The randomly sampled results were brought into an ascending order (ESM2: T3) to determine the 95% confidence ranges of the LCA results (Fig. 2).

2.3.4 Sensitivity analysis

A sensitivity analysis was performed to investigate how changes of selected input parameters affect the overall results (ESM2: T4-6). Changes were made to the (1) waste composition, (2) energy mix, (3) transport distances and (4) avoided production processes. Regarding the waste composition, the mixed DFG in the baseline scenario was altered to represent a 100% trawl net and a 100% gillnet stream (Table 1). The rationale behind this was to investigate whether both types of material result in the same environmentally optimal waste treatment approach. An overview of the system and processes parameters that were adapted to reflect the changed waste composition is given in the supplementary materials 1 (Table S34–S35). To study the implications of an increased share of renewable energy, the baseline 2017 German energy mix was adjusted to a 2030 German energy mix from the following projected energy sources: 3.4% hydro, 5.2% biomass, 43.1% wind, 12.1% solar, 12.1% natural gas, 1.2% lignite and 13.8% hard coal (Agora Energiewende 2017).

The transport distances were changed to a best and worst-case (Table 3). The best case assumes a pre-treatment in the harbour and a 250-km, 20-km and 250-km transport of pre-treated DFG to a further syngas production, energy recovery or mechanical recycling respectively. The worst case assumes that DFG is transported in a wet condition with an additional water content of 25%. Furthermore, the transport distance to the sorting facility was increased from initially 500 km to 1300 km representing the distance between a central German Baltic Sea harbour and a DFG sorting facility in Eastern Europe. The calculation is provided in the supplementary materials (ESM2: T6).

Table 3 Altered weighted transport distances for the sensitivity analysis
Full size table

To investigate the sensitivity of important avoided production processes and its underlying assumptions, best- and worst-case scenarios were constructed for the syngas, nylon and energy production (Table S36). For syngas, the best-case considered an output increase of 25% due to the water content and a 2.5 times higher heating value compared to other market products. A worst-case for syngas was modelled as the baseline scenario. For nylon, the assumed substitution ratio of 100% was reduced to 81% in the worst-case due to differences in market prices between recycled and virgin plastics (Cremiato et al. 2018; Unger et al. 2017; Tunesi et al. 2016). A best-case for nylon was not specified. For the energy production, the thermal conversion efficiency of 37% for the incineration process was modified to 0% in the worst-case scenario to represent a situation in which no thermal energy is generated and to 72% in the best-case (Laner et al. 2015). Similarly, the incineration’s baseline electrical conversion efficiency of 15% was adjusted to 5% and 35% to reflect a worst and best case (Laner et al. 2015).

3 Results

3.1 Impact contributions

Figure 2 illustrates the results of the impact contribution and uncertainty analysis. Considering uncertainties, the mechanical recycling and energy recovery outperformed either or both the syngas production and landfill disposal in climate change, fossil fuel consumption and metal depletion as well as in the eutrophication and toxicity-related impact categories. Large and overlapping uncertainty ranges occurred for terrestrial acidification, photochemical oxidant formation and water depletion. For the latter, the Monte Carlo results significantly exceeded the net environmental impacts, which required that the error bars were scaled down by a factor of 20 to keep the impact contributions visible (Fig. 2). While overlapping uncertainty ranges prevent a clear scenario ranking, the net environmental impacts of the mechanical recycling and energy recovery were significantly lower compared to the syngas production and landfill disposal across all impact categories, giving confidence in the overall ranking.

The mechanical recycling ranked first in climate change, terrestrial acidification, marine eutrophication, photochemical oxidant formation, water depletion and fossil depletion. Its net environmental savings ranged between 24 and 506% compared with the least beneficial options, respectively. While retrieval actions of DFG caused notable environmental burdens (Table S20), the avoided virgin nylon production was able to reduce or offset those impacts (Table S26, Fig. 2). The energy recovery scenario ranked first in the remaining impact categories. Compared to the respective worst option, its net environmental impact savings ranged between 109 and 577%, whereby the highest benefits were achieved in metal depletion. The major impacts of the energy recovery scenario were attributed to the retrieval and incineration emissions, but they were largely compensated by the avoided heat and electricity production (Table S30).

The syngas production and landfill disposal both ranked last in six of twelve impact categories (Fig. 2). The high impacts of the syngas production mainly resulted from the electricity consumption of the gasification process and a limited capacity from avoided biomass syngas to offset those impacts (Table S27). Similarly, the avoided electricity and heat production from landfill gas in the disposal scenario (Table S31) was not enough to offset the retrieval and landfill impacts from diesel or direct emissions to soil, water and air. Landfill disposal also lacked otherwise beneficial steel and lead recycling so that credits from an avoided steel and lead production were not considered in this scenario (Fig. 2). Overall, the single most important impact contributions were generated during DFG retrieval, metal recycling and the main treatment steps (Table 4), whereas transport and pre-treatment processes had less impact on the environmental performances (Fig. 2).

Table 4 Overview of the most significant environmental impact contributions (full contribution list in ESM2)
Full size table

3.2 Parameter sensitivity

The results of the sensitivity analysis are depicted in Fig. 3, where net environmental impacts from the baseline scenarios are compared against the environmental performances that result from input parameter changes to the waste composition, energy mix, transport distances and avoided production.

Fig. 3
figure 3

Sensitivity analysis results for the waste composition, energy mix, transport distances and avoided production processes compared to the baseline scenario results

Full size image

3.2.1 Waste composition

An important part of the waste compositions is the lead content, which amounts to 6.8% in the baseline, to 13.5% in gillnets and to 0% in trawl nets (Table 1). The lead content directly affects the avoided lead production in the energy recovery and recycling scenarios, causing trawl nets to lose credits compared to the baseline, whereas gillnets register impact reductions (e.g. metal depletion). For the disposal scenario, there is an opposite trend, because the lead content directly corresponds to environmental impacts from landfill emissions (see human toxicity). Given those changes, the trawl nets landfill disposal overpassed the performance of a mechanical recycling in freshwater eutrophication and the toxicity-related impact categories, whereas the controlled landfill disposal of gillnets manifested its position as least preferable treatment approach.

3.2.2 Energy mix

The 2017 German electricity mix in the baseline scenario was mainly adjusted to represent a higher share of wind (from 19.3 to 43%) and a lower share of lignite (from 24.5 to 1.2%) in the expected 2030 energy mix. Due to its high electricity demand, the effects of this change can be best observed in the syngas production scenario, for which the highest fluctuations occurred in freshwater eutrophication and the toxicity-related impact categories. Within the syngas production scenario, the environmental burden in the metal depletion, freshwater and marine ecotoxicity increased in the 2030 energy mix due to the higher share of electricity from wind, whereas burdens in the other impact categories decreased. As net electricity producer, the energy recovery was impacted in the opposite way, meaning that performance improvements in a future electricity mix reduced its credits from avoided electricity production. While this allowed the syngas production to gain with respect to the energy recovery in the baseline scenario, the changes were not enough to cause a scenario ranking change.

3.2.3 Transport

The transport process is very robust as changes to the baseline scenario only cause small fluctuations to the environmental impact results. Shorter transport distances improved the scenario results but did not affect their ranking. Longer transport distances mainly increased the impact of the thermal conversion and recycling scenarios, because landfills were assumed to be comparably nearby. This allowed the landfill disposal scenario to move one rank up to the second-best performing scenario in terrestrial ecotoxicity and climate change impact, where initial differences between the second and third rank were small.

3.2.4 Avoided production

Based on technology and quality assumptions, the avoided production processes were adjusted to reflect worst and best cases using the following ranges for the output substitution: 100–312% for syngas, 81–100% for nylon, 33–233% for electricity and 0%–194% for heat in their respective scenarios. Changing the avoided production parameters had large effects on climate change, eutrophication and toxicity-related impact categories. While best-case assumptions did not affect the scenario ranking, the worst-case modelling allowed the landfill disposal scenario to climb from the third to the second-best performing scenario in climate change and terrestrial ecotoxicity impact. However, the scenario differences in those impact categories are very small and do not mark a significant change.

4 Discussion

4.1 Research implications

4.1.1 Waste hierarchy

Most waste treatment LCAs support the waste hierarchy (Laurent et al. 2014), which means that the potential environmental impacts of the recycling pathway are generally lower than of an energy recovery. However, the position of thermal conversion into syngas is subject to an ongoing debate. For example, Demetrious and Crossin (2019) and Zaman (2010, 2013) established an environmental advantage of incineration with energy recovery over gasification, whereas Kreißig et al. (2003) reported opposite results. Evangelisti et al. (2015) gave evidence for both ranking constellations, in which case the environmental impact results mainly depended on the technology choices and their net electrical efficiencies. As thermal conversion covers a wide range of technologies, from pyrolysis at 400 °C up to steam reforming at more than 1000 °C, one reason for these apparent inconsistencies originates from the wide range of energy input required to start the thermal conversion process. The results presented here clearly backed incineration with energy recovery as environmentally advantageous as compared to thermal syngas conversion, but as the results are only representative for the investigated experimental processes, they should not be used to draw conclusions on other technologies. The large impact variability especially among thermal conversion technologies makes it necessary to continue conducting case-specific LCAs before recommendations on technological adoptions can be made.

4.1.2 Waste treatment of DFG

Due to its complex material composition, DFG has been traditionally landfilled (Schneider et al. 2018). However, this study shows that landfill disposal causes high environmental impacts, especially for toxic lead containing gillnets, and that more environmentally friendly options exist in the form of mechanical recycling and incineration with energy recovery. While attempts towards mechanical recycling were made (Weißbach et al. 2021), the material proved too contaminated with lead, fine-grained sediments, and organic matter and technologically challenging to further pursue this waste management pathway for DFG retrieved from the Baltic seafloor. Therefore, Germany, where this study was mainly conducted, has taken first steps to integrate DFG into its existing waste management by dismantling and incineration. In order to enable processing in incineration plants, the described process (Fig. 1) was slightly adapted as it now involves a manual removal of lead lines and lead fragment extraction before pre-cutting and coarse shredding take place. The resulting polymer nets and ropes are free of hazardous waste materials and suited for standard incineration plants. This is time-consuming and labour-intensive, but it eliminates the need for a density separation stage and further reduces the environmental burdens induced if hazardous lead would enter incineration slags and emissions, in addition to conforming with European law. Overall, the approach taken in Germany follows similar systems that were established in South Korea and the USA (McCoy 2010; Jung et al. 2010) and can serve as a blueprint for an environmentally optimised DFG waste treatment system.

4.1.3 Environmental hotspots and process improvements

The retrieval of lost fishing gear from the seas offers significant benefits for the marine ecosystem. Yet, the LCA presented in this study attributed substantial environmental impacts to DFG retrieval campaigns. This is mainly because DFG retrievals are organised as single journeys to suspected DFG locations with low-volume recovery rates (WWF Poland 2015). To decrease the retrieval impact, single journeys could be combined into multi-destination trips and side-scan sonars offer the technology to increase the DFG detection rate. Dedicated and consolidated retrieval journeys to sonar-detected DFG locations would render each journey more efficient and less energy-consuming. Given their success in other countries (Gilardi et al. 2010), such retrieval improvements are now also tested for European Union member states. Under the lead of WWF Germany, WWF Poland, WWF France and Mediterranean and Keep the Estonian Seas Tidy are testing the sonar search methodology to achieve a high retrieval efficiency. The responsibility for establishing healthy marine ecosystems lies, according to European law, with the member country authorities. In Germany, the coastal states are supporting the first state-funded DFG retrieval activities. With the coordination by state authorities, consolidated seafloor-cleanups are easier to realise than un-coordinated actions by individual fisheries or private NGOs. These efforts are expected to result in less energy-intensive, more efficient DFG removal campaigns.

Another major impact hotspot is the gasification process due to its high electricity demand. While incremental process improvements and a technology up-scaling may reduce its electricity consumption, the investigated process requires up to 500 °C higher temperatures than incineration (Schneider 2020). Unless higher credits from the hydrogen-rich syngas can be obtained, there is little room for this technology to compete with direct energy recovery plants at the considered electricity mixes.

The pre-treatment and transport processes caused comparatively low environmental impacts, which is in line with previous research (Faraca et al. 2019; Wäger and Hischier 2015; Biganzoli et al. 2015). Based on this, future LCA studies may exclude those processes from their system boundaries. However, as LCA studies are case-specific, such exclusions should be justified and treated with extreme care.

4.1.4 DFG prevention

The investigated clean-up and waste management scenarios for DFG are addressing an important area of environmental pollution, but they are not stopping the input of new fishing gear into the sea (Schneider et al. 2021). Preventive solutions on the other hand can remove the need for retrieval actions, making them optimal solutions. To prevent DFG, the European Union has proposed an extended producer responsibility scheme that would impose the fishing gear collection and waste treatment cost onto the fishing gear producer (European Commission 2018). The rationale behind this is to create an incentive for the producers to improve fishing gear designs in form of fewer and less harmful materials to facilitate a repair, reuse and recycling. It also provides motivation to setup an effective end-of-life fishing gear collection system, possibly through a deposit return scheme, to make sure that the producer’s collection cost is reduced. While the changes have yet to be implemented, the approach provides a much-needed step into the right direction that may help other countries to implement similar schemes.

4.2 Limitations and future research

4.2.1 Assumptions and limitations

In addition to the specific assumptions described in the life cycle inventory, this study is limited by the adapted modelling framework, system boundaries and data choices. Those limitations mainly include the selection of an attributional LCA, the exclusion of capital goods and uncertainties in the data sources. For the EU Best Available Techniques Reference documents for waste incineration (BREF 2006) and the EMEP/EEA air pollutant emission inventory guidebook (2016a, b, c, d, e), updated newer versions became available throughout this research, which have not been considered. Another limitation resulted from the lack of large-scale DFG treatment plants, which meant that small scale experimental data for DFG or average European data for similar waste streams had to be used as proxies.

4.2.2 Marine litter impact categories

In this study, four common waste treatment pathways for DFG were compared. However, it would be helpful to relate those findings to a prevention or no-action scenario during which either or both retrieval and waste management would be omitted. While the environmental impacts of preventive scenarios may be straightforward to calculate, a no-action scenario would require the consideration of new impact categories such as ghost fishing or the release of microplastics and lead to capture the environmental impact resulting from this approach. Although research in this area is increasingly undertaken (Woods et al. 2019, 2021; Sonnemann and Valdivia 2017), such new impact categories need to be combined with existing impact methodologies to become commonly applied in LCAs.

4.2.3 Sustainability assessment

This research investigated the environmental impacts of DFG management schemes. However, to identify and recommend the most sustainable approach, economic and social factors need to be considered as well (Stamford and Azapagic 2014). Previous research was carried out by numerous researchers and stakeholders in this area (i.e. UNEP-SETAC 2011) which could provide further interesting angles to this study. For example, the social impact of the scenarios could be compared based on their job creation and employment potential as well as through health and safety risks that may result, e.g. from toxic lead handling. Similarly, the scenario-specific funding requirement should be identified in a future economic analysis for the DFG waste management options. Especially for entrepreneurs and waste management companies, it would be helpful to investigate whether the increasing willingness of consumers to pay a premium price for marine litter products would be enough to cover the high retrieval and waste treatment costs of material recycling.

5 Conclusion

The clean-up of derelict fishing gear is very important to restore the health of marine ecosystems. Yet, this study showed that the consequent waste treatment pathways can be linked to high environmental impacts. Especially a default landfill disposal (Scenario 1) can cause severe long-term damages for example due to the potential leakage of hazardous materials such as lead in acidic landfill conditions. Environmentally better options to deal with this novel and increasing waste stream are a mechanical recycling (Scenario 4) and incineration with energy recovery (Scenario 3) which achieve lower potential environmental impacts compared to a landfill disposal or an energy intensive thermal conversion into syngas (Scenario 2). The study results are very robust as input parameter changes to the electricity mix, transport distances and avoided production processes could not significantly alter the scenario ranking results. An uncertainty analysis revealed distinct uncertainty ranges in nine of twelve impact categories between the best and worst performing scenario, which further underlines the credibility of the results. Besides the waste treatment, high environmental impacts could also be attributed to the retrieval. To decrease the environmental impact, better detection and clean-up strategies were proposed. Overall, this study is the first to appraise the environmental impact of marine litter clean-up and waste treatment pathways. As such it establishes important information for policy-makers in Europe to setup an appropriate recovery, reception and recycling structure for derelict fishing gear. The study contains detailed life cycle inventory with valuable primary data for novel processes such as the gasification, adding new evidence to the ongoing debate about the positioning of thermal conversion in the waste hierarchy. This study highlighted the suitability of LCA as a decision support tool, but future research is needed to also examine the influence of economic and social factors as well as marine litter prevention approaches.