AUTHOR=Ardron Jeff A. , Simon-Lledó Erik , Jones Daniel O. B. , Ruhl Henry A. TITLE=Detecting the Effects of Deep-Seabed Nodule Mining: Simulations Using Megafaunal Data From the Clarion-Clipperton Zone JOURNAL=Frontiers in Marine Science VOLUME=6 YEAR=2019 URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2019.00604 DOI=10.3389/fmars.2019.00604 ISSN=2296-7745 ABSTRACT=
The International Seabed Authority (ISA) is in the process of preparing exploitation regulations for deep-seabed mining (DSM). DSM has the potential to disturb the seabed over wide areas, yet there is little information on the ecological consequences, both at the site of mining and surrounding areas where disturbance such as sediment smothering could occur. Of critical regulatory concern is whether the impacts cause “serious harm” to the environment. Using metazoan megafaunal data from the Clarion-Clipperton Zone (northern equatorial Pacific), we simulate a range of disturbances from very low to severe, to determine the effect on community-level metrics. Two kinds of stressors were simulated: one that impacts organisms based on their affinity to nodules, and another that applies spatially stochastic stress to all organisms. These simulations are then assessed using power analysis to determine the amount of sampling required to distinguish the disturbances. This analysis is limited to modelling lethal impacts on megafauna. It provides a first indication of the effect sizes and ecological nature of mining impacts that might be expected across a broader range of taxa. To detect our simulated “tipping point,” power analyses suggest impact monitoring samples should each have at least 500–750 individual megafauna; and at least five such samples, as well as control samples should be assessed. In the region studied, this translates to approximately 1500–2300 m2 seabed per impact monitoring sample, i.e., 7500–11,500 m2 in total for a given location and/or habitat. Detecting less severe disturbances requires more sampling. The numerical density of individuals and Pielou’s evenness of communities appear most sensitive to simulated disturbances and may provide suitable “early warning” metrics for monitoring. To determine the sampling details for detecting the desired threshold(s) for harm, statistical effect sizes will need to be determined and validated. The determination of what constitutes