Polymetallic nodules are concretions of manganese and iron oxides that lie on the sediment surface of abyssal plains, most densely in the Clarion-Clipperton Zone of the eastern Pacific at depths of four to five kilometres. They grow at rates of millimetres per million years and hold nickel, copper, cobalt and manganese in quantities that compare with land reserves (Hein et al., 2013). Under the United Nations Convention on the Law of the Sea the seabed beyond national jurisdiction and its minerals are the common heritage of mankind, administered by the International Seabed Authority, which is obliged to ensure effective protection of the marine environment from the harmful effects of mining (United Nations, 1982). The Authority has issued exploration contracts and is drafting rules for exploitation.

Whether protection can be effective depends on a factual question: what happens to the benthos after the nodules have been taken, and over what time. The abyssal seafloor is cold, dark and poor in food, and its communities are diverse, sparse and slow growing (Smith et al., 2008). Glover and Smith (2003) judged nodule mining to be among the largest prospective human disturbances of the deep sea, because a single operation would work over hundreds of square kilometres of seabed each year. Since nobody has mined at commercial scale, the evidence on recovery comes from experiments designed to imitate part of the disturbance. This review examines that evidence.

The article is a narrative review of the published literature on experimental and test-mining disturbance of nodule-bearing abyssal sediments. It concentrates on studies that returned to a disturbed site after a known interval and compared it with undisturbed ground nearby, and on the syntheses that have drawn those studies together. It was not compiled through a formal systematic search and it does not pool effect sizes. Mining of seafloor massive sulphides at hydrothermal vents and of cobalt-rich crusts on seamounts raises related questions but involves different habitats and is mentioned only in passing. General overviews of the industry and its environmental issues are available elsewhere (Miller et al., 2018).

The record begins with the pilot mining tests that industrial consortia ran in the Clarion-Clipperton Zone in 1978, which left collector tracks on the seabed that scientists later relocated. The first experiment designed for ecological study was DISCOL, the Disturbance and Recolonization experiment, begun in 1989 in the Peru Basin. A plough-harrow eight metres wide was towed repeatedly across a circular area some two nautical miles in diameter, burying the nodules and turning over the upper sediment, and the site was sampled before, immediately after, and at intervals of months to years (Thiel et al., 2001). Figure 1 places DISCOL among the experiments that followed it.

The experiments of the 1990s took a different approach. Instead of ploughing, they towed a device that fluidised the surface sediment and discharged it a few metres above the bottom, so as to study the blanketing of nearby seabed by resettled material. The Benthic Impact Experiment of the United States National Oceanic and Atmospheric Administration used this design in the Clarion-Clipperton Zone, and the same disturber was then employed by Japanese, eastern European and Indian programmes, the last in the Central Indian Ocean Basin. Results for meiofauna from these resuspension experiments were compared by Radziejewska (2002). Table 1 sets out the main groups of studies, the disturbance each applied and what was later reported.

The findings are best taken by size class, since the classes differ in mobility, in generation time and in how they were sampled.

All the authors cited above caution against reading the experimental record as a forecast, and their reasons point in one direction. The first is scale. DISCOL disturbed a few square kilometres, of which only a fraction was touched by the plough, so that unploughed strips remained within metres of every track to supply colonists. A mine site would be worked contiguously over an area orders of magnitude larger, and recolonisation from the edges would have much further to travel. The second is the nature of the disturbance. The plough buried nodules but left them in the sediment, and the resuspension experiments left them on the surface. A collector removes them, together with the top few centimetres of sediment in which most of the fauna and most of the labile organic matter are found.

The third is the plume. Collectors will raise fine sediment that settles over ground beyond the mined strips, and the processing vessel will return water and fines to the ocean at some depth. Measurements on experimentally generated plumes indicate that abyssal clay particles aggregate and settle faster than their individual size would suggest, which would concentrate deposition nearer the source (Gillard et al., 2019), though how far this holds at industrial discharge rates is not known. Metals released from crushed nodules and pore water may add a toxic component for which deep-sea dose-response data are almost absent (Hauton et al., 2017). Nor is the seabed the only habitat exposed: Drazen et al. (2020) argued that discharge plumes and noise put midwater communities at risk and that these have been left out of most assessments.

The experimental evidence bears on three regulatory questions. One is the meaning of serious harm, the threshold at which the Authority must act. Levin et al. (2016) proposed that harm be judged against the spatial extent, duration and reversibility of effects, and the revisits supply the duration: decades at least for sediment communities, and geological time for nodule fauna. A second is whether losses can be offset. Van Dover et al. (2017) held that biodiversity loss from mining is unavoidable and that restoration at abyssal depths is not feasible at the necessary scale, and Niner et al. (2018) concluded that a goal of no net loss cannot be met. The third is spatial planning. If mined areas will not recover, protection has to come from areas left unmined, which is the logic of the network of Areas of Particular Environmental Interest designed for the Clarion-Clipperton Zone (Wedding et al., 2013).

What the old experiments cannot supply defines the agenda for new ones. No published study has yet followed a site after disturbance by a full-scale collector for more than a few years, and trials of prototype collectors are only now producing baseline and early post-impact data. The area over which plume deposition changes the fauna, and the thickness of deposit that does so, are poorly constrained. Connectivity is largely unknown: most abyssal species have been collected only once or twice, so that ranges, larval dispersal and the capacity of distant populations to reseed a mined area cannot be estimated with confidence. Baselines themselves vary from year to year with the supply of sinking food, and separating a mining effect from that variation needs longer reference series than most contractors hold. Finally, almost all the evidence concerns the Peru Basin and the eastern Clarion-Clipperton Zone, and its transfer to other basins is an assumption.

Four decades of experiments give a coherent account. Disturbance of nodule-bearing sediment causes immediate loss across the benthos; mobile and small organisms come back in years, in altered assemblages; and sessile fauna, microbial activity and carbon cycling remain impaired for at least a quarter of a century where the surface layer and its nodules were disturbed. Because the experiments were smaller and gentler than mining would be, these results mark the optimistic end of the range. The sound working assumption for regulators is that mined seabed is lost as habitat for nodule fauna, and that conservation will depend on what is set aside.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.