Results

Effects of Season, Reef, and Depth on Phycoperiphyton Productivity

1. All reefs over three seasons.

The main effects of season, reef, and depth were significant for all four variables, as well as significant interaction effects on all variables (Table 1). A significant (p<0.000) season*reef interaction effect was found for Pn, R, chl-a, and Borg indicating that seasonal responses in all four variables are different across reefs (Table 1). Respiration was also significantly (p<0.027) affected by the interaction of season*depth indicating that trends in R with depth are not constant from one season to the next. Although significant season*reef and reef*depth interaction effects were observed for Pn, variance due to these interactions is better explained by the significant (p<0.000) season*reef*depth interaction (Table 1). This three-way interaction suggests that, not only are the temporal and vertical distribution of Pn different among reefs, but also that seasonal patterns in the vertical distribution of Pn are different across reefs. In addition to the season*reef effect, a significant interaction (p<0.020) between reef*depth was observed for Borg (Table 1), implying that the vertical distribution of organic biomass differed among reefs.

In the subsequent one-way analyses, season and reef treatments had significant effects on all four variables, while depth had significant effects on Pn and chl-a only. Respiration was significantly affected by season (F2,311 = 26.81; p<0.000), and by reef (F3,310 = 17.49; p<0.000), but not depth. Net productivity was significantly affected by season (F2,311 = 38.6; p<0.000), by reef (F3,310 = 13.6; p<0.000), and by depth (F2,311 = 5.9; p<0.003). Chlorophyll-a was significantly affected by season (F2,162 = 6.8; p<0.001), by reef (F3,161 = 3.6; p<0.015), and by depth (F2,162 = 3.1; p<0.046). Organic biomass was significantly affected by season (F2,144 = 7.7; p<0.001) and by reef (F3,143 = 10.1; p<0.000), but not depth.

Post-hoc comparisons of the seasonal effect on R show that spring had significantly higher respiration rates (476 mgO2 hr-1 m-2) than summer and fall, which did not differ significantly from each other. However, this difference is being driven by the extremely high respiration rates the occurred at Kat during the spring and does not accurately represent R for all other reefs during this season. Net productivity varied significantly among all three seasons with spring having the most negative (-157 mgO2 hr-1 m-2) rate. The highly negative Pn values during the spring are caused by the values measured at the LRRs (Table 3). The significantly higher chl-a values observed in the fall (9.2 mg m-2), compared to summer and spring that were not significantly different from each other, can be attributed to the high concentration of chl-a measured at Leg during the fall. Organic biomass was also highest in the fall (52.3 g m-2), although not significantly greater than the spring sample, and lowest (24.4 g m-2) during the summer. In summary, spring had the highest R, lowest Pn, lowest chl-a, and intermediate amounts of Borg, whereas the fall had the lowest R, highest Pn, highest chl-a, and highest Borg.

Post-hoc comparisons of reef effects on respiration show that Kat had significantly higher R (501 mgO2 hr-1 m-2) than the other three reefs, although this is likely caused by the spring sample. Net productivity was significantly lower on the two reefs located closer to shore (Leg and USM) than on those further from shore (Kat and Hand) but this result is largely driven by the highly negative responses measured in the spring sample. Chlorophyll-a concentrations were significantly higher at Kat and Leg (central reefs) than at USM (western reef), but chl-a at Hand was not significantly different from any of the other reefs. Although the intermediate levels of chl-a (5.9 mg m-2) at Hand seem inconsistent with its high Pn values, this could be explained by the higher chlorophyll to biomass ratio (0.32) at this reef. Kat, which had the second highest Pn rates, also had high chlorophyll concentrations (8.1 mg m-2) as well as a high chlorophyll to biomass ratio (0.23). Legacy and USM, which had the lowest Pn values, have correspondingly low chlorophyll to biomass ratios (0.12 and 0.11, respectively). Organic biomass was significantly higher at Leg and similar among the other three reefs.

A consistently declining trend in all four variables with increasing depth was observed. Depth had a significant effect on Pn and was only borderline significant (p<0.046) on chl-a, but was not significant for R or Borg. Net productivity and chl-a were both significantly higher on surface plates than bottom plates, and mid plates were not different from either surface or bottom plates.

2. Low relief reefs over four seasons.

Low relief reefs (LRRs) were examined separately as data was collected from these reefs in all four sampling seasons, enabling the examination of seasonal patterns in more detail . In the three-way ANOVA, season had a significant effect on all four variables (R, Pn, chl-a, and Borg), reef had a significant effect on R, chl-a, and Borg but not Pn, and depth significantly affected Pn and Borg only (Table 2).  A significant (p<0.001) season*reef interaction effect was found for all four variables (Table 2), suggesting that the two LRRs are behaving differently within and across seasons. Respiration was also significantly (p<0.002) affected by the interaction of season*depth (Table 2), indicating that trends in R with depth are not constant from one season to the next. Although significant two-way interaction effects were observed for all combinations of main effects on Pn, variance due to these interactions is better explained by the significant (p<0.000) season*reef*depth interaction (Table 2). This three-way interaction suggests that, not only are the temporal and vertical distribution of Pn different among the two reefs, but also that seasonal patterns in the vertical distribution of Pn are different across the two reefs.

Based on the subsequent one-way analyses, season had a significant effect on Pn, chl-a, and Borg, but not on R (Table 3). Post-hoc comparisons of the significant seasonal effect on Pn showed a significant increase from summer (-79.1 mgO2 hr-1 m-2) to fall, statistically similar rates between fall and winter (mean = 17.2 mgO2 hr-1 m-2), followed by a decrease in spring (-320.6 mgO2 hr-1 m-2). This is consistent with the results from the three-way ANOVA, which showed that both reefs experienced this same seasonal pattern. Post-hoc comparisons on the significant chlorophyll-a results revealed that chl-a concentrations were highest in the fall (Table 3), but did not vary significantly during the other three seasons (mean = 3.5 mg m-2). However, results from the three-way ANOVA (Table 2) indicate that the high chl-a concentrations during the fall are being driven by the values observed at Leg during that season and that USM did not exhibit similarly high values. Biomass was significantly higher in the fall and spring compared to the summer and winter on both reefs (Table 3), but Leg had especially high amounts in the fall.

The results for reefs indicate that R, chl-a, and Borg were all significantly greater at Leg compared to USM, but the two reefs did not differ significantly in their Pn values (Table 3). Depth only had a significant effect on Pn (Table 3) with the surface and bottom plates significantly different from each other, but mid-depth plates were not significantly different from either surface or mid plates (Table 3).

Seasonal and Spatial Patterns in Phytoplankton

            The results of the two-way ANOVAs on water column samples revealed that the main effects of season (summer, fall, winter, spring) and reef (Kat, Hand, Leg, and USM) were both significant for all four variables (R, Pn, chl-a, and POM). The season*reef interaction was also significant for all four variables, indicating that reefs exhibited different seasonal patterns. The subsequent one-way ANOVAs showed that the main effect of season had a significant effect on all four variables (Table 4) as previously indicated by the two-way analysis but, unlike the two-way results, reef was only significant for respiration (Table 4).

Landscape Production Estimates

            The colonizable surface area of the two HRR was calculated as 21,239.6 m2 for Katrina and that of Square Handkerchief was 14,184.2 m2 . The volume of the water column over Kat was 31,374 m3 and 20,432 m3 over Hand. At Kat, annual benthic production was around -1,338 kg C while water column production was estimated at 24,325 kg year-1 (Table 5). The sum of benthic and water column production estimates results in a net production estimate of approximately 23 t C year-1 at Katrina reef (Table 5). Annual benthic production at Hand was -74 kg C and water column production was 18,242 kg C year-1. Net annual habitat production was approximately 18 t C year-1 for this reef (Table 9).

            The surface areas of the two LRR (Legacy and USM) were 3,199.1 and 4,034.8 m2, respectively. The volume of the water column overlying Leg was 5,854 m3 and the volume over USM was 7,384 m3. Annual benthic production on Legacy was -915.6 kg C and annual phytoplankton production in the water column overlying the reef was 4,867 kg C. The net habitat production for Legacy was calculated by adding the water column production estimates to the negative benthic production estimates and was found to be approximately 3,952 kg C year-1 (Table 5). Annual benthic production on USM was -843 kg C year-1 and 8,051 kg C year-1 for phytoplankton in the overlying water. Annual net habitat production for USM reef was estimated at 7,208 kg C (Table 5).

            The estimated mean annual benthic and water column production was standardized to a one square meter in order to facilitate a direct comparison among the annual reef production estimates. The two LRRs had the lowest mean annual net benthic production values; – 286.2 and -208.9 g C m-2 year-1 for Leg and USM, respectively (Table 5). The two HRRs had substantially greater mean annual net benthic production compared to LRRs. However, Hand had considerably greater net benthic production rates compared to Kat, with estimates of -5.2 and -63.0 g C m-2 year-1 respectively (Table 5). Mean annual water column production estimates were fairly similar among reefs, ranging from a low of 775.3 g C m-3 year-1 at Kat to a high of 1090. g C m-3 year-1 at USM (Table 5), with a mean of 897 g C m-3 year-1 across all four sites.