This data set consists of a scored time-series of Autonomous Underwater Vehicle (AUV) images from the Bicheno region on the east coast of Tasmania. Surveys were conducted between 2011 and 2016 within the Governor Island Marine Reserve and nearby sites outside the reserve. Governor Island was surveyed in 2011, 2013, 2014 and 2016. The outside sites of Trap Reef, Cape Lodi and Butlers Point were surveyed in 2011, 2013 and 2016. Imagery across all surveys was scored for the presence of Centrostephanus rodgersii urchin barrens across rocky reef at each site. Prior to analysis the data was subsetted to every fifth image to avoid overlapping images. The data set also contains depth information for each image and a measure of rugosity (Vector Rugosity Measure) computed in ArcGIS software from a one metre resolution bathymetric map covering the survey sites. Analysis was conducted to examine the trend in the presence of barrens through time and to compare the occurrence of barrens inside the Governor Island Marine Reserve with sites outside the reserve. A spatio-temporal model incorporating both spatial and temporal correlation in the time-series of data was used. This data set contains the scored data used in the analysis. Further details of the methods used and results are contained in the following article. Please cite any use of the data or code by citing this article: Perkins NR, Hosack GR, Foster SD, Monk J, Barrett NS (2020) Monitoring the resilience of a no-take marine reserve to a range extending species using benthic imagery. PLOS ONE 15(8): e0237257. https://doi.org/10.1371/journal.pone.0237257

Shifts from productive kelp beds to impoverished sea urchin barrens occur globally and represent a wholesale change to the ecology of sub-tidal temperate reefs. Although the theory of shifts between alternative stable states is well advanced, there are few field studies detailing the dynamics of these kinds of transitions. In this study, sea urchin herbivory (a ‘top-down’ driver of ecosystems) was manipulated over 12 months to estimate (1) the sea urchin density at which kelp beds collapse to sea urchin barrens, and (2) the minimum sea urchin density required to maintain urchin barrens on experimental reefs in the urbanised Port Phillip Bay, Australia. In parallel, the role of one of the ‘bottom-up’ drivers of ecosystem structure was examined by (3) manipulating local nutrient levels and thus attempting to alter primary production on the experimental reefs. It was found that densities of 8 or more urchins m-2 (≥ 427 g m-2 biomass) lead to complete overgrazing of kelp beds while kelp bed recovery occurred when densities were reduced to ≤ 4 urchins m-2 (≤ 213 g m-2 biomass). This experiment provided further insight into the dynamics of transition between urchin barrens and kelp beds by exploring possible tipping-points which in this system can be found between 4 and 8 urchins m-2 (213 and 427 g m-2 respectively). Local enhancement of nutrient loading did not change the urchin density required for overgrazing or kelp bed recovery, as algal growth was not affected by nutrient enhancement.