Sargassum species composition
Fifty-one seaweed species were successfully identified in the three monitoring studies conducted in May 2018 (the first monitoring study), July 2020 (the second monitoring study), and October 2020 (the third monitoring study) (see Table S1 in the online Supplementary Material). Regarding species diversity, nine species of large brown seaweed of Sargassum and Myagropsis were confirmed as S. confusum, S. fulvellum, S. hemiphyllum, S. horneri, S. macrocarpum, S. patens, S. piluliferum, S. siliquastrum, and M. myagroides (Table S1). Based on Table S1, the ranges of large brown seaweed observed at the six artificial reef sites (Stns. A-1 to A-6) were 1 to 3 species (S. horneri, S. patens, and S. piluliferum), 5 to 7 species (S. confusum, S. hemiphyllum, S. macrocarpum, S. patens, S. piluliferum, S. siliquastrum, and M. myagroides), and 5 to 9 species (S. confusum, S. fulvellum, S. hemiphyllum, S. horneri, S. macrocarpum, S. patens, S. piluliferum, S. siliquastrum, and M. myagroides), with an average of 2, 6, and 6 species for the first, second, and third monitoring studies, respectively. At the three natural reef sites (Stns. B-1 to B-3), the ranges were 1 to 3 species (S. horneri, S. macrocarpum, and S. patens), 2 to 3 species (S. macrocarpum, S. patens, and S. siliquastrum), and 2 to 4 species (S. horneri, S. macrocarpum, S. patens, and S. piluliferum), with an average of 2, 2, and 3 species for the first, second, and third monitoring studies, respectively.
We performed cluster analysis by nonmetric multidimensional scaling (NMDS) for the nine Sargassum and Myagropsis species in the first, second, and third monitoring studies both on the artificial and natural reefs (Fig. 3). The span distance value was 0.153. The maximum average silhouette value was 0.458 when the number of seaweed community groups was seven. It was thus appropriate to divide the monitoring activities into seven groups (Table 3). In the seven groups, we classified roughly as follows. Group 1: the artificial reef in the first monitoring study; Group 3: the artificial reef in the second and third monitoring studies; and Groups 2, 4, 5, 6, and 7: the natural reef in the first, second, and third monitoring studies. In short, when the 26 survey sites (9 survey sites x 3 observation minus 1 site (Stn. A-5)) were classified based on seaweed species composition and their coverage, those on the artificial and natural reefs were different.
The indicator species of each group obtained using the indicator value (IndVal) method were significantly large: S. horneri in Group 1, S. hemiphyllum/S. confusum in Group 3, and S. siliquastrum in Group 4 (Table 4). No indicator species were extracted in Groups 2, 5, 6, and 7. The species S. patens was not identified as an indicator species because it appeared at all 26 survey sites. The analysis of all 26 survey sites by permutational multivariate analysis of variance (PERMANOVA) confirmed a significant difference in similarity between the artificial and natural reefs (p=0.003) (Table 5).
Seaweed coverage
Regarding seaweed coverage of all seaweed species, on the artificial reef, no species with seaweed coverage of 5% or more were identified in the first monitoring study. S. horneri was confirmed in the first monitoring study, but S. horneri decreased and S. hemiphyllum/S. patens increased in the second and third monitoring studies. On the natural reef, S. patens dominated and S. macrocarpum increased (Fig. 4).
In the first monitoring study (May 2018), the average seaweed coverage of all seaweed species on the natural reef was significantly larger (55%) than that on the artificial reef (7%), with a difference of 48%, while in the second monitoring study (July 2020), the coverage on the natural reef was 67% compared to the artificial reef of 45%, with a difference of 22%. In the third monitoring study (October 2020), the average brown seaweed
coverage on the artificial reef was 77%, which nearly approached that on the natural reef of 83%, with a difference of only 7% (Fig. 5).
Statistical analysis (ANOVA) showed that there were significant differences in the average seaweed coverage for all seaweed species, including green algae, brown algae, and red algae. A significant difference was confirmed for reef types (Artificial VS Natural reefs) and observation times (First, Second, and Third monitoring times), but no significant difference was confirmed for the interaction of these two factors (Table 6).
Notably, the seaweed coverage of the family Corallinaceae in the red algae on the natural reef was significantly larger than that on the artificial reef (paired two-way ANOVA, p=0.0276). The significant difference was confirmed for observation times and the interaction of these two factors (paired two-way ANOVA, p=0.0001 and 0.0246, respectively).
Growing environment for all seaweeds
A Wilcoxon's rank sum test for water depth, water temperature, pH, DO concentration, salinity, and turbidity revealed no significant differences between the artificial and natural reefs in all items. However, the water depth tended to increase from landward to offshore from Stns. A-1 to A-3, from Stns. A-4 to A-6, and from Stns. B-1 to B-3 (Table 7). The maximum water depth value among these nine survey sites was 5.4 m at Stn. A-6 on the artificial reef, and the minimum value was 2.5 m at Stn. B-1 on the natural reef.
The multiple linear regression analysis indicated that the predictive equation regarding the relationship between all seaweed coverage at each survey site and water quality parameters was significant (p=0.029). The formula is as follows: all seaweed coverage = 200.510 – 15.815 water depth – 16.076 water temperature – 71.105 pH + 7.100 DO concentration + 26.679 salinity – 0.048 turbidity (Table 8). However, among the six water quality parameters, statistically, all parameters had a weak relationship with all seaweed coverage, except for water depth (Pr(>|t|) = 0.008, Table 8). Based on Fig. 1 and Table S1, the average seaweed coverages at shallow stations A-1, A-4, and B-1 for the three monitoring campaigns were higher of 55.3, 50.3, and 78.3%, respectively. On the other hand, those at deep stations A-3, A-6, and B-3 were lower of 41.7, 18.3, and 53.3%, respectively. Thus, these facts show that the stations in shallow waters have higher seaweed coverage than those in deep waters. It is because sunlight reaches the seabed better in shallow waters than in deep waters, which makes it suitable for seaweeds to photosynthesize.
Additionally, in line with the analysis results of the influence of depth on seaweed coverage, these results also show an inverse pattern of depth on seaweed diversity, where shallow depths (Stns. A-1, A-4, and B-1) have a lower number of species of 20, 24, and 21 species than deep depths (Stns. A-3, A-6, and B-3) of 32, 28, and 27 species, respectively. It is because the transition in shallow waters is more advanced than in deep waters, which leads to an emergence of dominant species and a decrease in the number of species.
Figure 6 depicts the seasonal changes in water temperature in the survey area. Regarding the water temperature over one year, the maximum value was 30.9°C (July 31, 2021), the minimum was 9.9°C (February 19, 2021), and the average was 18.9°C with a standard deviation of 5.9°C. According to the multiple linear regression analysis, the water temperature had a weak correlation with all seaweed coverage (see Table 8).