<p><strong>Seamounts are important marine habitats that commonly occur in deep waters, with global estimates ranging from tens of thousands to several million. Interactions between seamount topographies and surrounding currents increase local productivity at some features, supporting high abundances and biomasses of both pelagic and benthic fauna. For this reason, many seamounts host fish aggregations, including several commercially-important species. Early observations of seamount fish aggregations led to their initial targeting in the 1960s by bottom trawlers. Since then, this practice has expanded globally. In New Zealand, seamounts were initially targeted by fisheries in the late 1970s, and by the early 1990s a majority of the annual orange roughy (Hoplostethus atlanticus) catch was obtained from several seamounts. Bottom trawling has significant impacts not only on the target fish population, but also on the non-target benthic fauna caught as bycatch, as they can be comprised of fragile, emergent, sessile animals (e.g., deep-sea corals and sponges) that are easily damaged or removed by trawling gear. However, the extent to which these communities can recover from trawling impacts is uncertain, and this information is necessary to design effective spatial management measures, particularly in New Zealand where many seamounts are located within the Exclusive Economic Zone and actively fished. Designing studies to address this knowledge gap and provide management recommendations first requires an assessment of current understanding regarding resilience of seamount benthic communities to fishing. In this thesis, such an assessment was done through a global literature review and meta-analysis of seamount benthos, fisheries, known trawling impacts, and a three-fold evaluation of community resilience (Chapter 1). This review indicated that recovery will likely be initially patchy, and information on community spatial patterns is a necessary starting point to determine recovery potential. In order to provide such baseline spatial context to subsequent studies of previously fished seamounts, the structure and distribution of benthic communities at two unfished seamounts was investigated to characterise fine-scale community structural patterns and identify environmental drivers of these patterns (Chapter 2). Megabenthic (seafloor fauna with body sizes >2 cm) fauna were identified and described from 1,233 total seafloor images collected with camera transects on both seamounts in 2015 to identify communities. The predicted spatial distributions of these communities were mapped and characteristics of the patches they were predicted to occupy were described. Habitat drivers were found to differ between the two features at different spatial scales, and these differences were reflected in the predicted spatial distributions and patch characteristics of communities characterising each seamount. Post-fishing responses of benthic communities were determined by examining fine-scale spatio-temporal changes in their structure and function over 20 years on a seamount that was intensively trawled from the early 1990s until its protection in 2001 (Chapter 3). As part of a long-term research programme, camera surveys were conducted on this seamount (and others) in 2001, 2006, 2009, 2015, and 2020. In total, 2,343 images were analysed across years to obtain megabenthic faunal counts of 148 taxa. Results showed there was a temporal shift in community structure driven by small relative changes in abundances of several taxa, indicative of early recovery. However, community function showed little sign of ongoing recovery. Juvenile colonies of a reef-forming stony coral, Solenosmilia variabilis, were observed, demonstrating that the fisheries closure may have facilitated the early recovery process of this species and its associates. Data obtained from these analyses was used to develop joint-species distribution models to predict changes in benthic faunal occurrence and abundance up to 200 years post-fisheries closure (Chapter 4). Spatial distributions of 36 aggregated megabenthic taxa were predicted for five post-closure time-steps in 2001, 2020, 2040, 2100, and 2200. Nine communities were described from the spatial predictions using cluster analysis and spatial patterns of the patches they occupied were characterised. Results of this study suggested that the fisheries closure may enable the recovery of most, but not all, taxa. Additionally, several communities on the seamount summit may undergo succession towards a reef climax community, but are unlikely to fully reverse fishing impacts even on this long timescale.</strong></p><p>Key findings from the preceding chapters were synthesised, and their implications for seamount management and future research discussed (Chapter 5). The latter includes the need for studies determining the physiology and life history traits of seamount benthic species to better predict their resilience to fishing impacts. Major contributions of this thesis study to the field of seamount ecology were also presented, including a conceptual model of seamount recovery, hypothesised changes in community patch characteristics after trawling disturbance, and a decision framework for seamount management prioritisation.</p>