Detect coastal disturbances and climate change effects in coralligenous community through sentinel stations

(coordinates 36˚ 4,770 N - 05˚ 25,184 W; GPS, DATUM WGS 84). These quadrats were placed on surfaces containing target species. The covers were monitored over ten consecutive years (2005–2014) in spring and summer seasons. A four video footage of 50 x 50 cms were filmed sequentially, covering a total of 1 x 1 m quadrat, in order to gain higher resolution of the subsequent photos taken from the video. A total of 320 photographs were digitally manipulated using the Adobe Photoshop 6.0 (2000) program as follows: four 50x50-cm areas were created for each quadrat (Fig 5A); the contrast and colour saturation were adjusted (Fig 5B); a complete digital frame was created (Fig 5C); and a digital grid that was adjusted to the perimeter of the monitoring area was superimposed (Fig 5D). To correct the angular deformation. To assess the degree of the cover of the target species, a system involving the determination of the presence/absence of cover using the digital grid was applied. This system aided estima- tion accuracy [89,90] and optimised the working time [64,91]. Nevertheless, it should be noted that, in multistratified communities, the system tends to underestimate the cover [64,69,89– 92] because larger species “mask” those that developed underneath them. Statistical analyses Repeated measures analysis of variance (RM-ANOVA) was applied to test whether the mean coverage of the target species significantly varied either through time (intra-subject factor ‘time’; ten levels: 2005 to 2014) or according to orientation (inter-subject factor ‘orientation’;

two levels: horizontal and vertical). Mauchly’s test of sphericity [93] was used to test the assumption that the variances in the differences between all possible pairs of groups were equal. When the sphericity condition was not verified, the F test value was corrected with the Greenhouse-Geisser epsilon index [94,95]. The factors of time and orientation were consid- ered orthogonal and fixed. The same two-factor design was considered to test for any signifi- cant differences in the mean coverages via RM-permutational analysis of variance Fig 4. The updated method with the objective of minimizing the degree of intrusion. A: Hand drill, B: Expansion anchor bolt, C: Drilling and bolt fixing sequence, D: Metal crosshead fixation, E: Sequential allocation of the detachable quadrat, F: Representation of a complete monitoring point. https://doi.org/10.1371/journal.pone.0231641.g004 PLOS ONE Detect coastal disturbances in coralligenous PLOS ONE | https://doi.org/10.1371/journal.pone.0231641 May 5, 2020 7 / 23