Settlement plates can be deployed to assess whether the colonising community has the same species composition as the previous community and/or

set aside area. Genetic analysis comparing the fauna colonising artificial or newly-generated natural substrate to the original populations could enable the source of colonisers to be identified drug discovery and the suitability of set aside areas to be assessed. The monitoring program needs to be implemented at suitable spatial and temporal scales (IMMS, 2011), although the appropriate length of long-term study required is at present unclear. Levels of natural variation need to be evaluated before any appreciable operations begin, in order to establish fluctuations that could, for example, be seasonal or related to changing chemical conditions. Also, following disturbance, succession of species composition and abundance is to be expected, and so any monitoring must span sufficient time. Recovery from natural disturbance at sites along the EPR (Lutz et al.,

1994 and Mullineaux et al., 2010) and Juan de Fuca Ridge (Tunnicliffe et al., 1997) and the rapid re-growth of deposits at Solwara 1 (Gwyther, 2008a) indicate that monitoring for a few years following the cessation of mining activities may be sufficient. However, experimental polymetallic nodule mining resulted in Tariquidar disturbance to the benthic community assemblage for at least 26 years following mining activity (Miljutin et al., 2011), suggesting that in keeping with the precautionary principle, suitable long-term monitoring could be on the scale of decades rather than years. Monitoring programmes by themselves are all very well, but they need to be evaluated against pre-determined decision rules. The latter will be derived from management objectives, and involve a management response when a monitored parameter value exceeds a certain level. For example, mining may have to stop in an area if sediment plume deposition thicknesses exceed a certain ZD1839 cell line depth. The design of baseline, impact and long-term monitoring studies also needs to consider the importance of replication to address the natural

environmental variability at SMS sites at both temporal and spatial scales. Ideally, this should utilise a design similar to BACI (before-after-control-impact, Green (1979)) or Beyond BACI (Underwood, 1991 and Underwood, 1992), with multiple unimpacted (control or set aside) and impacted (mined) sites (Collins et al., 2013a). However, BACI design at SMS sites will probably be asymmetrical with the potential for multiple unimpacted sites but only one impacted site (Underwood, 1991 and Underwood, 1992), as mining is likely to be concentrated at one site. There is also the question of cost. Coastal or shallow water impact studies may be able to investigate multiple sites but the logistics (time and cost) of investigating multiple sites in deep-sea SMS mining impact studies may be prohibitive.