Abstract
Seabed substrates have the capacity to support a variety of marine communities. However, habitats provided by natural substrates are increasingly modified and supplemented by man-made structures. These provide hard surfaces suitable for colonisation by sedentary and/or non-migratory organisms, and may contribute to an interconnected system of benefit to diverse marine populations. Robust assessment of the influence of such structures is, therefore, a necessary consideration for their long-term management. The challenge of compiling and manipulating data for input to two North Sea models is described. Source data were processed and gridded at three different spatial resolutions to investigate the effect of scale on spatial relationships. Choice of grid size was found to exacerbate existing uncertainty in location and extent of features, influencing interpretation of their spatial distributions at the different scales examined. The small spatial footprint of man-made structures, compared with natural substrates, may lead to underestimation of the influence of the former at coarser model scales. Choices must be made between data availability, spatial resolution and accuracy, modelling and analysis requirements, to identify robust approaches to reliable outcomes. Model sensitivity and uncertainty analyses are recommended for application in data-limited situations. Greater openness and cooperation in data-sharing is required for robust scientific modelling to underpin decision-making in the marine environment.