MEPS 517:143-157 (2014)  -  DOI: https://doi.org/10.3354/meps11068

ABSTRACT: We studied 2 seagrass beds in the Mexican Caribbean that were grazed by green turtles. Grazing impact was moderate at Puerto Morelos (<20% of the area was grazed), whereas at Akumal, 45 to 55% of the bed was grazed. The turtles practiced rotational (cultivation) grazing and thereby increased the nitrogen leaf content in the dominant seagrass Thalassia testudinum by 30 to 33%. Average seagrass leaf productivity decreased under grazing at Puerto Morelos from 3.09 to 0.93 g dry wt m^-2 d^-1, whereas it did not change significantly at Akumal (0.88 to 1.17 g dry wt m^-2 d^-1). At Puerto Morelos, the turtles maintained grazing plots for 13 mo to >2 yr, creating a mosaic of patches that were grazed, ungrazed or recovering from grazing. Cover of the faster-growing Halodule wrightii and rhizophytic algae increased in the grazing plots, whereas that of the dominant but slow-growing T. testudinum decreased. After the turtles stopped grazing the plots, the cover of T. testudinum gradually increased again, but recovery to pre-grazing conditions lasted >1 yr and was not observed in this study. Cover of Syringodium filiforme decreased when the turtles opened up new grazing areas but remained stable afterwards because the turtles usually avoided consumption of this seagrass. At Akumal, the system approached carrying capacity for grazing, and the turtles returned to grazing plots that had not fully recovered from past grazing. Here, the dominant climax seagrass T. testudinum became less abundant from 2008 until 2012, resulting in a less patchy landscape with low seagrass biomass and higher prevalence of the early seral species H. wrightii.

KEY WORDS: Chelonia mydas · Halodule wrightii · Plant-herbivore interactions · Thalassia testudinum