Abstract

In 1991 and 1992, cyanobacterial blooms depleted sponges, the primary refuge of juvenile Caribbean spiny lobsters, in ~20% of the nursery in the Florida Keys, USA. Long-term data from the affected middle Keys were used to study the impact of sponge loss, juvenile abundance, recruitment and shelter use. A manipulative experiment (1991–93) involved artificial shelters on 27 ~0.5-h sites. Conditions on 19 sites over the affected ~ 500 km 2 area were compared before and after the blooms. The entire nursery (~10,000 km²) was surveyed to estimate the impact of the disturbance on total juvenile recruitment.

Refuge and lobster abundances declined and the pattern of shelter use changed on previously sponge- rich sites without alternative shelter (solution holes, coral heads, cement blocks, etc.). Although sponge loss often locally reduced juvenile abundance, the juvenile lobster population overall declined by ~5%. The availability of alternative, previously underused shelter (solution holes, coral heads, etc.) in the affected region, continued production in the larger unaffected nursery region, and high postlarval supply that fortuitously coincided with sponge loss all offset a stronger effect. However, postlarval supply is unlikely to remain high until the sponges repopulate the middle Keys (10+ years), so a major factor ameliorating the effect of sponge loss on lobster recruitment may disappear.