Abstract. Ecological connectivity can influence the distributions of diversity and productivity among ecosystems, but relationships among multiple marine ecosystems remain relatively uncharacterized. Sandy beaches are recipient ecosystems that support coastal food webs through deposits of drift macrophytes (wrack), and serve as test cases for exploring within-seascape connectivity. We present results from the first comprehensive survey of geographic and temporal patterns of wrack cover and composition on beaches along the North Central Coast of California and test the role of local donor ecosystems and physical factors in predicting wrack distribution. We surveyed wrack at 17 beaches in August 2010, and monthly at a subset of 10 beaches for 13 months. We estimated explanatory variables of (1) local donor ecosystem cover (kelp forests, rocky intertidal, and bays and estuaries), (2) biomass transport, and (3) beach morphology. Regression analyses were used to evaluate relationships among the cover of six key wrack categories and the explanatory variables above, for two time periods. We found persistent geographic variation in wrack composition and detected significant relationships between wrack cover and cover of local donor ecosystems for five of the six wrack categories (Nereocystis, Zostera, Postelsia, mixed red algae, and mixed brown algae). Transport mechanisms (wind exposure, swell exposure) or attributes of the recipient ecosystem (beach width, beach slope) explained additional spatial variation for three of the six wrack categories (Zostera, Phyllospadix, and mixed red algae). Our results support the concept of considering ecological connectivity (particularly the role of donor ecosystems upon which recipient ecosystems rely) in the design and management of protected areas.

Abstract. Contradictions between system-specific evidence and broader paradigms to explain ecosystem behavior present a challenge for natural resource management. In Florida (USA) springs, increasing nitrate (NO 3 À ) concentrations have been implicated as the cause of algal overgrowth via alleviation of N-limitation. As such, policy and management efforts have centered heavily on reduction of nitrogen (N) loads. While the N-limitation hypothesis appears well founded on broadly supported aquatic eutrophication models, several observations from Florida springs are inconsistent with this hypothesis in its present simplified form. First, NO 3 À concentration is not correlated with algal abundance across the broad population of springs and is weakly negatively correlated with primary productivity. Second, within individual spring runs, algal mats are largely confined to the headwater reaches within 250 m of spring vents, while elevated NO 3 À concentrations persist for several kilometers or more. Third, historic observations suggest that establishment of macroalgal mats often lags behind observed increases in NO 3À by more than a decade. Fourth, although microcosm experiments indicate high thresholds for N-limitation of algae, experiments in situ have demonstrated only minimal response to N enrichment. These muted responses may reflect large nutrient fluxes in springs, which were sufficient to satisfy present demand even at historic concentrations. New analyses of existing data indicate that dissolved oxygen (DO) has declined dramatically in many Florida springs over the past 30 years, and that DO and grazer abundance are better predictors of algal abundance in springs than are nutrient concentrations. Although a precautionary N-reduction strategy for Florida springs is warranted given demonstrable effects of nutrient enrichment in a broad suite of aquatic systems worldwide, the DO-grazer hypothesis and other potential mechanisms merit increased scientific scrutiny. This case study illustrates the importance of an adaptive approach that explicitly evaluates paradigms as hypotheses and actively seeks alternative explanations.

1. Reduced grazing can lead to increases in autotroph biomass and changes in taxonomic composition similar to those associated with nutrient enrichment. In Florida's iconic spring ecosystems, algal proliferation has become widespread, yet the causes remain unclear. 2. We tested three linked hypotheses: (i) loss of top-down grazer control explains algae proliferation, a change often attributed to nitrate enrichment; (ii) grazer control is mediated by dissolved oxygen (DO) concentration; and (iii) an algal-dominated state may persist if biomass exceeds a critical level beyond which grazers can no longer constrain accumulation. 3. We tested these hypotheses using hierarchically nested benthic surveys of algal, vascular plant and gastropod biomasses along with physicochemical measurements in eleven springs spanning gradients in nitrogen enrichment, algal cover and DO. 4. We observed a significant and temporally consistent negative association between algal and gastropod biomasses (R 2 = 0.38), with gastropods displaying the strongest explanatory power in multivariate prediction models that explained 45% of algal variation in the best fitted model. 5. A modest but significant positive bivariate association was observed between gastropod biomass and DO (R 2 = 0.23); a multivariate model including temperature, velocity, canopy cover, submerged aquatic vegetation and conductivity explained 56% of gastropod variation. 6. Residuals from a linear relationship between gastropod and algal biomasses were strongly bimodal above a threshold grazer biomass of 20 g dry weight m À2 (c. 235 snails m À2 ) suggesting alternative states of high and low algal abundance. 7. These observations support the hypothesis that gastropods have the potential to control benthic algae in Florida's springs, identify DO as a partial explanation for variation in grazer abundance and imply potentially important hysteretic behaviour in top-down algal control.

No abstract