MEPS 567:57-78 (2017)  -  DOI: https://doi.org/10.3354/meps12054

ABSTRACT: Estuarine environments support dynamic phytoplankton blooms, especially in low-latitude regions, where the effects of local drivers dominate. Identifying key bloom drivers from entangled ecological and anthropogenic influences is particularly challenging in stressed systems where several disturbances interact. Additionally, processes controlling bloom and non-bloom phytoplankton biomass dynamics can differ spatially, further confounding characterization of disturbance regimes that create bloom-favorable conditions. This study aims to explore the question of whether the shift from non-bloom to bloom conditions is matched by a shift in the relative importance of water quality drivers. Florida Bay (USA), a shallow subtropical inner shelf lagoon, was chosen as the study site due to its unique bloom dynamics and low-latitude location, as well as for the availability of long-term (16 yr) water quality data consisting of monthly measurements from 28 locations across the 2200 km² bay. At each of the locations, we applied a novel threshold-based quantile regression analysis to chlorophyll a data to define bloom conditions, separate data from non-bloom conditions, and evaluate phytoplankton biomass dynamics of each of the 2 states. The final suite of explanatory covariates revealed spatial trends and differences in the relative importance of water quality descriptors of phytoplankton between the 2 conditions. The effects of turbidity and salinity on phytoplankton biomass became pronounced during blooms, whereas non-bloom conditions were primarily explained by autoregressive phytoplankton biomass trends and nutrient dynamics. The proposed analytical approach is not limited to any particular aquatic system type, and can be used to produce practical spatiotemporal information to guide management, restoration, and conservation efforts.

KEY WORDS: Quantile regression · Time series modeling · Spatial patterns · Cyanobacteria · Florida Bay