Seagrass monitoring by underwater videography: Disturbance regimes, sampling design, and statistical power

Surface-based underwater videography with submeter-accuracy differential GPS is a method with several advantages, including high spatial and visual resolution, effectiveness at all depths at which seagrasses occur, non-destructive sampling, and rapid data collection in the field. Here I investigate the statistical power of this method applied to a natural meadow and to virtual meadows created by a spatially explicit model of seagrass disturbance, regrowth, and colonization. The approach is found to detect a 5-10% loss in short timespans at 80% probability, with a sampling design emphasizing long transects (≅400 m or greater) and a paired analysis in which the same permanent transects are followed each monitoring event. A field effort function shows that this precision is possible within a single working field day for sampling regions ≅1 km2. These conclusions hold regardless of the presence of positional error in remonitoring previous transects by submeter DGPS navigation.