Exploring estuarine eutrophication sensitivity to nutrient loading

The sensitivity of surface chlorophyll (Chl) and bottom water dissolved oxygen (DO) to total nitrogen (TN) load was investigated using a Bayesian-based process model fit to data from a range of estuaries. The model was used to test if the sensitivity of DO depletion to TN loads is dependent only on factors controlling the sensitivity of surface Chl, or if additional factors are important. Results indicate that separate processes control Chl and DO sensitivity, and that these sensitivities vary among estuaries. Analysis of fitted parameters across estuaries showed that Chl sensitivity to TN loading was positively correlated with water residence time, and DO sensitivity to Chl was positively correlated with relative mixing depth. Ecosystem sensitivity to external stressors and the biophysical control of this sensitivity have long been of interest to ecologists. Recent work has focused on the combined effects of multiple stressors and on the relative sensitivity of ecosystem functions (Blake and Duffy 2010; Wiley et al. 2010). Many open questions remain in both of these areas. In this study, we explored factors controlling the relative sensitivity of chlorophyll (Chl) concentration and bottom water dissolved oxygen (DO) depletion to nutrient stress across estuaries. Because of a convergence of freshwater and saltwater habitats, their association with high human population densities, and their semi-enclosed nature, estuaries are highly sensitive to loads of nutrients, sediment, and toxins and are excellent study systems for exploring questions of sensitivity. Nutrient overfertilization has led to significant eutrophication in many systems (Bricker et al. 2007), and that is the subject of this study. Nutrient fertilization of coastal waters causes both primary and secondary symptoms of eutrophication. Primary symptoms include increases in Chl and primary production, changes in nutrient ratios, increased sedimentation of organic carbon, and changes in phytoplankton species composition. These changes can cascade into secondary symptoms, including changes in water transparency, decreases in bottom water DO, animal mortality or displacement, shading or toxic inhibition of submerged aquatic vegetation, and many others (Cloern 2001). Here we focus on two key symptoms, increased Chl concentration and decreased bottom water DO, because of their ecological importance and management attention. Increases in phytoplankton biomass, often represented as Chl concentration, are the most commonly reported symptoms of eutrophication in estuaries and are a key precursor to many of the secondary symptoms (Cloern 2001). Sinking and decomposition of phytoplankton biomass are major drivers of DO depletion. DO depletion can lead to widespread ecosystem changes, including fish kills (Diaz and Rosenberg 2008), decreased or displaced fish production (Rabalais and Turner 2001), altered biogeochemical cycles (Turner et al. 2008), and decreased societal value through reduced recreational opportunities and fisheries harvest losses (Renaud 1986). The issue has become widespread worldwide (Diaz and Rosenberg 2008; Zhang et al. 2010). Clear links exist between nutrient loads, especially total nitrogen (TN), and bottom water DO, for example, in the Gulf of Mexico (Turner et al. 2008; Greene et al. 2009) and in Chesapeake Bay (Kemp et al. 2005; Liu and Scavia 2010). However, the sensitivity of bottom water DO to TN loading has also been shown to vary through time (Liu and Scavia 2010; Liu et al. 2010; Scully 2010) and among estuaries (Zhang et al. 2010). Here, we expand on these analyses to explore the relative simultaneous sensitivity of Chl concentration and DO depletion to TN loading in a cross-estuary context. Estuarine eutrophication sensitivity to TN loading is modulated through ecological filters, system-specific processes that modulate responses to enrichment (Cloern 2001). These filters include the degree of tidal flushing and turbulent mixing that can dilute TN loads and reduce eutrophication symptoms (Monbet 1992; Cloern 2001); the degree of freshwater flushing and residence time, which determine the time during which nutrients are available for estuary processing (Swaney et al. 2008; Steward and Lowe 2010); and biological factors such as top-down control of phytoplankton biomass (Cloern 2001). As part of a toolbox for eutrophication filter evaluation, a mechanistic estuarine model of TN-driven surface-layer Chl concentration has been tested and is available (Scavia and Liu 2009). We expand that model to a two-layer version that includes surface Chl and bottom-layer organic carbon and DO. The model was constructed and fit using Bayesian inference based on seasonally and spatially averaged, satellite-derived Chl estimates used in Scavia and Liu (2009) and summer average, bottom water DO data from the National Estuarine Eutrophication Assessment (NEEA; Bricker et al. 1999, 2007). We used the model to test if the sensitivity of DO depletion to TN loads is * Corresponding author: [email protected] 1 Present address: U.S. Geological Survey, Great Lakes Science Center, Ann Arbor, Michigan Limnol. Oceanogr., 58(2), 2013, 569–578 E 2013, by the Association for the Sciences of Limnology and Oceanography, Inc. doi:10.4319/lo.2013.58.2.0569