Assessing the use of artificial substrates to monitor Gambierdiscus populations in the Florida Keys.

Four distinct coastal locations were sampled on a monthly basis near Long Key (Florida Keys, USA) over a 13-month period to study Gambierdiscus population dynamics on different substrates, including four macrophyte species (Dictyota spp., Halimeda spp., Laurencia spp., and Thalassia testudinum) and three artificial substrates (polyvinyl chloride (PVC) tiles, burlap, and fiberglass window screen). Cell densities of Gambierdiscus were generally lower on Dictyota versus Halimeda and Laurencia. Cell densities of Gambierdiscus were significantly correlated among macrophyte hosts in 54% of the comparisons, and between macrophyte hosts and artificial substrates in 72% of the comparisons. Predictive slopes determined from regression analyses between cell densities on artificial substrates and macrophyte hosts indicated that, on an areal basis, fewer cells were present on macrophytes versus artificial substrates (cells cm-2 ) and that slope variation (error) among the different macrophytes and sites ranged from 5% to 200%, averaging 61% overall. As the data required log-transformation prior to analyses, this level of error translates into two-orders of magnitude in range of estimation of the overall average abundance of Gambierdiscus cells on macrophytes (135 cells g-1 wet weight); 20-2690 cells g-1 ww. The lack of consistent correlation among Gambierdiscus cell densities on macrophytes versus artificial substrates, coupled with the high level of error associated with the predictive slope estimations, indicates that extreme caution should be taken when interpreting the data garnered from artificial substrate deployments, and that such deployments should be thoroughly vetted prior to routine use for monitoring purposes.