Identifying the emergence of SCTLD in situ is based on epidemiological characteristics, including which species are affected and when they become affected, since no specific pathogen has been identified that would allow for a confirmed diagnosis. Therefore, although it is possible to determine when SCTLD appears at a site, it is difficult to say for certain that acute tissue loss disease signs on an individual coral represent true SCTLD without following that coral through time. This is because recent mortality indicative of disease can be attributable to other acute tissue loss diseases, such as white plague. Whereas white plague lesions emerge at the base or margin of a colony, expand rapidly, but often do not coalesce and halt before full colony mortality (Brandt et al., 2013; Smith et al., 2013), SCTLD lesions appear as multi-focal lesions across a colony’s surface, expand rapidly, and often coalesce resulting in total colony mortality (Meiling et al., 2020). To be conservative, we combine and present data for all acute tissue loss (ATL) disease signs affecting corals in the survey data.

The Territorial Coral Reef Monitoring Program (TCRMP) has been conducting annual coral and fish community assessments at locations across the USVI since 2002 (Ennis et al., 2019). Of the 34 total monitoring locations, 19 are within the northern USVI (St. Thomas and St. John), while 15 locations are distributed around St. Croix; however, SCTLD was not observed on St. Croix during the time period of this study and so St. Croix monitoring locations are not included here. The deepest (63 m) northern USVI location, Ginsburg’s Fringe, was also excluded since it is a mesophotic lettuce coral (Agaricia undata) reef making it ecologically distinct from the other TCRMP monitoring locations. Annual monitoring data at the other 18 St. Thomas and St. John TCRMP locations are presented here going back to 2005, with the exception of the Brewers Bay and Coral Bay locations, which begin in 2008 and 2011, respectively. In addition to these annual data, several monitoring locations were sampled outside of annual monitoring when SCTLD was first reported (Table 1).

The TCRMP methodology collects data on the benthic community composition at each location by surveying six marked 10 m transects annually using video methods detailed in Ennis et al. (2019) and summarized in Smith et al. (2013). A diver on SCUBA takes continuous video along each transect keeping the lens approximately 0.4 m above the substrate (a wand guide is used to maintain height). Twenty to fifty non-overlapping images are then captured from each transect’s video. Captured images represent an area of approximately 0.31 m² (0.64 m × 0.48 m). Coral Point Count with Excel Extension software (Kohler and Gill, 2006; prior to 2019) or the “imager” (Barthelme, 2020) and “spatstat” (Baddeley et al., 2015) packages in R is then used to superimpose randomly located points on each image (20 points/image) and the substrate type under each point is identified to the most descriptive level possible.

The TCRMP methodology collects coral demographic and health metrics. Individual coral colonies along the same transects are assessed in situ for a number of coral health metrics focused on mortality, bleaching, and disease following modified AGGRA methodology (Ennis et al., 2019). As part of a project intended to understand coral juvenile abundances, three of the six transects at some sites (indicated by^b in Table 1), were expanded in 2016 to include any coral that was within or intercepted a 10 m × 0.5 m belt transect on the left side of each transect. For each coral assessed, multiple characteristics are recorded. Those presented here include species and the presence of disease lesions and/or bleaching. Bleaching was defined as stark white tissue.

When SCTLD was first observed at Flat Cay (FC) in January of 2019, periodic roving surveys by trained observers were instituted on SCUBA or snorkeling at 59 sentinel monitoring locations surrounding St. Thomas ranging in depth from 3 to 40 m (Figure 1). These locations included the 18 long-term monitoring locations (indicated by stars in Figure 1 and methods described above) as well as 42 strategically placed locations along the potential path of disease spread. Identification of SCTLD was determined after evaluating the species affected, severity and description of lesions (supported with photographic evidence, see below), and in some cases locations were revisited within 1 week to further confirm disease progression and species susceptibility. The presence of SCTLD was confirmed if multifocal to coalescing lesions were found on highly susceptible species that included D. cylindrus, M. meandrites, P. strigosa, and D. labyrinthiformis. Roving surveys were completed at the sentinel locations approximately monthly to bi-monthly and were used to determine the first calendar month when the disease was reported at each location.

To compare species susceptibility with studies from other regions, prevalence (% affected) of ATL diseases within species was calculated using corals assessed in TCRMP transects. Corals of each species were pooled across all transects and sites for surveys that occurred immediately after the first report of ATL indicative of SCTLD at a site (as indicated in Table 1). Differences in disease prevalence among species were tested using a likelihood ratio test. For FC, where there was repeated sampling of transects between February 2019 and February 2020, prevalence of ATL was calculated for all affected species for each sample time point in 2019 and 2020 (Table 1). ATL prevalence for these species at FC was also calculated for all transects pooled among all annual surveys between 2005 and 2018 for comparison. Differences in ATL prevalence through time for each species were analyzed using a likelihood ratio test. For two species, C. natans and D. cylindrus, a Fisher’s Exact test was used instead due to low sample sizes.

Metrics calculated at the overall coral community level included percent coral cover from video transects. Percent coral cover values were calculated for each transect at each time point as the number of randomly assigned points (described above) identified as living coral divided by the total number of identifiable points that landed on reef substrate. Relative change in coral cover was calculated as the difference between coral cover values in subsequent years divided by the previous year’s value.

Species richness, diversity (H’), and disease prevalence were also calculated for each transect at each time point from line intercept and belt transects. Disease prevalence (%) was calculated as the number of corals exhibiting ATL divided by the total number of corals assessed on that transect. Similarly, to Walton et al. (2018) and Muller et al. (2020), ATL was used instead of specifically SCTLD due to the difficulty in distinguishing between SCTLD and other acute tissue loss diseases such as white plague disease. However, it is likely that the majority of ATL lesions represented SCTLD, given the large difference in prevalence of ATL at sites considered SCTLD affected versus unaffected (detailed in results below). ATL lesions were distinguished from bleached tissue by the absence of living tissue.

To determine the impact of SCTLD on St. Thomas and St. John coral communities, percent coral cover, coral species richness, coral species diversity, and disease prevalence were each analyzed through time at monitoring locations using Friedman’s rank tests. Relative coral cover change was calculated as the change in coral cover from 1 year relative to the value recorded in the year immediately prior. For example, the relative change in coral cover for 2006 was calculated as coral cover in 2005 minus coral cover in 2006 divided by coral cover in 2005. Relative coral cover change was also analyzed through time for each location using Friedman’s rank tests. All statistical analyses were performed in R (v 3.5.1).

No SCTLD-like disease was found at long-term TCRMP monitoring locations surrounding St. Thomas and St. John in late 2018. However, a disease with characteristics described for SCTLD (i.e., multifocal large lesions) affecting species identified as highly and intermediately susceptible in the SCTLD Case Definition (2018) was found at the FC TCRMP monitoring site (depth 14 m) and on the shallow reefs of FC (depth 3–5 m) on 28 January 2019 (Figure 1). No ATL diseases were observed on monitored corals at the FC monitoring site one-month prior during normal annual TCRMP monitoring activities (see following data section), nor were any ATL disease signs noted on corals in observations outside of the transects in December 2018 (Brandt personal observation). In subsequent surveys of strategically placed sentinel sites around the island immediately following the first identification of SCTLD, SCTLD-affected reefs were found to be concentrated to the southwest of St. Thomas (Figure 1). SCTLD was not observed at sites to the southeast, east, or north of the island, or on reefs surrounding St. John.

Between the first identification of SCTLD in late January 2019 and June 2020, the disease spread west to the island of Culebra, Puerto Rico (first identified November 2019), and east to the British Virgin Islands (first identified May 2020, noted on vicoraldisease.org). Results from monitoring of the 59 sentinel sites located around St. Thomas and St. John indicated that the disease was concentrated on the southwest side of St. Thomas until April 2019 when it appeared on reefs on the northwest side of the island (Figure 1). The disease then remained concentrated on the west side of the island until the fall of 2019 when it began appearing on mesophotic (30–40 m depth) coral reefs sites (Supplementary Figure 1) at the shelf-edge as well as sites southeast of St. Thomas. In January 2020, the disease emerged at multiple locations on the east and northeast end of St. Thomas and appeared at locations on the west end of St. John. In the 6 months following that, the disease emerged at locations across St. John, though some embayments and locations were not yet affected by June 2020.

When all surveys from first observations of SCTLD at long-term monitoring locations were combined, prevalence of ATL due to disease showed differences among species (LR = 130.4, p < 0.0001; Figure 2). Highly susceptible species (i.e., species exhibiting >10% prevalence across all affected sites, as in Alvarez-Filip et al., 2019) showed overlap with species considered highly or intermediately susceptible in Florida and Mexico (Figure 2). These included P. strigosa, Montastraea cavernosa, D. cylindrus, C. natans, Eusmilia fastigiata, and Orbicella faveolata/franksi (Figures 3A–F). Species showing low prevalence (<10%) included the previously classified intermediately susceptible Orbicella annularis (Figure 3G) and Siderastrea spp. (Figure 3I). In addition, some species previously classified as unknown or not susceptible were found to have low prevalence of ATL, including Madracis spp. (Figure 3H), Agaricia spp. (Figure 3J), and Porites astreoides (Figure 3K). Species that did not show signs of disease during these initial surveys included Mycetophyllia spp., Stephanocoenia intersepta, branching Porites spp., and M. meandrites.

When examining the time series of species affected at FC, species showed different levels of disease prevalence as well as changes in disease prevalence through time (Figure 4). Prior to 2019, ATL was very rare and found on only Orbicella spp. (3 of 684 corals assessed, 0.4%; Figure 4H) and C. natans (1 of 23 assessed, 4.3%; Figure 4B). All species except C. natans and D. cylindrus showed significant differences in ATL disease prevalence through time (Table 2).

At FC prior to the onset of SCTLD, the most abundant types of corals found in transects included Orbicella spp., Siderastrea spp., Agaricia spp., and M. cavernosa (Figures 4A,E,H,J, respectively). Of these, Orbicella spp. and M. cavernosa showed increasing or steady prevalence of ATL after SCTLD emerged and declining abundances until the last sampling in February 2020, when only 12% of combined O. franksi and O. faveolata, 30% of O. annularis, and 11% of M. cavernosa colonies counted 1 year prior remained. Agaricia spp. and Siderastrea spp. also declined to 13 and 21% of their abundances 1 year prior, respectively, but generally had low (<10%) ATL prevalence. Low prevalence among these corals may have been due to their small size; once affected by SCTLD these corals may die quickly, not allowing for active disease to have been picked up in the periodic surveys.

Other species that were in low abundance at FC (n < 4) yet were affected by disease included C. natans (Figure 4B), D. cylindrus (Figure 4C), E. fastigiata (Figure 4D), and P. strigosa (Figure 4I). At least one of colony of each of these species was affected early on, but then then no disease was recorded after March 2019, likely because individuals of these species died from disease and no longer appeared in the transects, save for one E. fastigiata. M. meandrites, a highly susceptible species and often used as an indicator of the onset of SCTLD at a site, was also low in abundance (n = 8) and showed no disease prevalence within transects at FC when first sampled after SCTLD emerged (Figure 4F). However, subsequent samplings showed disease prevalence within this species, and then the loss of the species from transects by October of 2019 (Figure 4F).

Prevalence of ATL disease was significantly different through time at 6 of the 8 SCTLD-affected monitoring sites (Table 3). In 2019, ATL prevalence was statistically higher than previous sample year at five of these sites (Black Point, College Shoal, FC, Savana, Seahorse; Figure 5A and Supplementary Table 1 connecting letters) and was statistically equal to mass bleaching-related ATL disease outbreaks at one site (Botany Bay; Figure 5A, and Supplementary Table 1). Two SCTLD-affected sites, Botany Bay and South Water, showed elevated disease in 2019 but did not show statistically significant changes in disease prevalence through time (Table 3). Two additional sites that were unaffected by SCTLD (Hind Bank, Meri Shoal) showed significant changes in ATL disease prevalence through time (Table 3). These changes, however, were related to the mass bleaching events in 2005 and 2010 which immediately preceded white plague disease outbreaks the following year (Figure 5B and Supplementary Table 1, reported in Smith et al., 2013). All other sites unaffected by SCTLD showed little to no ATL prevalence through time (Figure 5B).

Over the 2005–2019 monitoring period, 5 of the 8 SCTLD-affected sites showed significant changes in coral cover through time, but only FC, where SCTLD was first detected, saw a significant decline related to the emergence of SCTLD (Figure 6A, Table 3, and Supplementary Table 2 connecting letters). Coral cover did not show recent declines at sites unaffected by SCTLD (Figure 6B). Coral species richness and diversity declined at many sites around the 2005 bleaching event, but then slowly rebounded or surpassed their 2005 levels prior to the appearance of SCTLD in the USVI (Figure 7, Table 3, Supplementary Figure 2, and Supplementary Tables 3, 4). Species richness showed significant declines related to SCTLD at two sites, FC and Botany Bay (Figure 7A and Supplementary Table 3) and species diversity (as measured by the Shannon index) also declined significantly in relationship to the emergence of SCTLD for one site, Botany Bay (Supplementary Figure 2A and Supplementary Table 4).

Relative change in coral cover was significantly different through time at 5 of the 8 SCTLD-affected sites (Table 4). At those sites, the relative drop in coral cover in 2019 was as large or larger than that which occurred in 2006 due to the 2005 mass bleaching event. Flat Cay showed the most dramatic loss with an average relative drop of 53.3% from the previous year (Table 4). Three additional sites that were not affected by SCTLD showed significant changes in relative coral cover through time (Coral Bay, Meri Shoal, South Capella). For sites Meri Shoal and South Capella, these changes were due to losses from the 2005 mass bleaching event. At Coral Bay, significant change was seen in 2018, almost certainly related to the passage of Hurricanes Irma and Maria in 2017 (Irma tracked directly over St. John).

The higher frequency sampling at FC showed that percent cover of both highly and intermediately susceptible species dropped between the last SCTLD-free monitoring time point and the first survey when SCTLD was present (Figure 8). Highly susceptible species percent cover dropped again, but only slightly by the next sampling, and then remained approximately the same. Highly susceptible species at FC included primarily M. meandrites, E. fastigiata, D. labyrinthiformis, and P. clivosa (Supplementary Figure 3A). While the former three species showed large losses, P. clivosa remained approximately the same, suggesting that this species may be less susceptible at this location. However, percent cover of intermediately susceptible species dropped continuously across all samplings in 2019 and 2020. Intermediate species included Orbicella spp., M. cavernosa, Myceptophyllia spp., S. intersepta, and S. siderea. Of these the largest drops were observed in Orbicella spp. and M. cavernosa (Supplementary Figure 3B). Low susceptible species showed no large changes during the time period, and included Porites spp., Madracis spp. (Supplementary Figure 3D). ATL prevalence increased in early 2019 and remained high until mass bleaching was observed in October when no ATL was recorded. ATL was then once again observed in February 2020 corresponding with a drop in bleaching prevalence (Figure 8).