Estuaries and Coasts

, Volume 37, Issue 2, pp 243–258

Evolving Paradigms and Challenges in Estuarine and Coastal Eutrophication Dynamics in a Culturally and Climatically Stressed World

  • Hans W. Paerl
  • Nathan S. Hall
  • Benjamin L. Peierls
  • Karen L. Rossignol
The H.T. Odum Synthesis Essay

DOI: 10.1007/s12237-014-9773-x

Cite this article as:
Paerl, H.W., Hall, N.S., Peierls, B.L. et al. Estuaries and Coasts (2014) 37: 243. doi:10.1007/s12237-014-9773-x

Abstract

Coastal watersheds support more than one half of the world’s human population and are experiencing unprecedented urban, agricultural, and industrial expansion. The freshwater–marine continua draining these watersheds are impacted increasingly by nutrient inputs and resultant eutrophication, including symptomatic harmful algal blooms, hypoxia, finfish and shellfish kills, and loss of higher plant and animal habitat. In addressing nutrient input reductions to stem and reverse eutrophication, phosphorus (P) has received priority traditionally in upstream freshwater regions, while controlling nitrogen (N) inputs has been the focus of management strategies in estuarine and coastal waters. However, freshwater, brackish, and full-salinity components of this continuum are connected structurally and functionally. Intensification of human activities has caused imbalances in N and P loading, altering nutrient limitation characteristics and complicating successful eutrophication control along the continuum. Several recent examples indicate the need for dual N and P input constraints as the only nutrient management option effective for long-term eutrophication control. Climatic changes increase variability in freshwater discharge with more severe storms and intense droughts and interact closely with nutrient inputs to modulate the magnitude and relative proportions of N and P loading. The effects of these interactions on phytoplankton production and composition were examined in two neighboring North Carolina lagoonal estuaries, the New River and Neuse River Estuaries, which are experiencing concurrent eutrophication and climatically driven hydrologic variability. Efforts aimed at stemming estuarine and coastal eutrophication in these and other similarly impacted estuarine systems should focus on establishing N and P input thresholds that take into account effects of hydrologic variability, so that eutrophication and harmful algal blooms can be controlled over a range of current and predicted climate change scenarios.

Keywords

NitrogenPhosphorusHydrodynamicsPhytoplanktonCoastal eutrophicationNutrient limitationClimate change

Introduction

More than one half of the Earth’s human population resides in coastal water- and airsheds (Vitousek et al. 1997; NOAA 2012; Ache et al. 2013). Accelerating agricultural, urban, and industrial development in these “sheds” has put unprecedented pressure on the ecological condition and sustainability of downstream riverine, estuarine, and coastal waters (Bricker et al. 1999; Boesch et al. 2001; Conley et al. 2009). Nutrient over-enrichment has been identified as a prime cause for water quality and habitat degradation (Nixon 1995; Paerl 1997; Boesch et al. 2001; Elmgren and Larsson 2001; Rabalais and Turner 2001; Diaz and Rosenberg 2008). In addition, hydrologic modifications including upstream reservoir construction, and agricultural and urban surface and groundwater water withdrawal, have altered water flow rates and paths, sedimentation rates, and optical properties of receiving waters. These modifications affect estuarine and coastal water and habitat quality (Cloern 2001; Boesch et al. 2001; Rabalais and Turner 2001; Humborg et al. 2007). While human modification of coastal water and airsheds directly, and often negatively, impacts water and habitat quality of these systems (Bricker et al. 1999; National Research Council 2000), climatic factors such as warming, more extreme storms, floods, and droughts modulate these impacts (Paerl and Huisman 2008, 2009; Jeppesen et al. 2010).

Human and climatically induced stresses interact strongly in estuarine watersheds. These interactions affect the delivery and effects of freshwater discharge and its nutrient load on the activity and composition of microalgal primary producers in geographically diverse estuarine and coastal ecosystems. Here, we evaluate the effects of human alterations of watershed nutrients and hydrology on coastal eutrophication and the measures needed to control it and associated habitat degradation in a more crowded and climatically extreme world. This issue is addressed in the context of managing the freshwater to coastal continuum, and the physiographic and biogeochemical gradient coupling water- and airsheds to coastal ecosystems (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig1_HTML.gif
Fig. 1

Conceptual diagram, illustrating the complex interactions of human nutrient inputs, nutrient cycling processes, nutrient limitation (of primary production) characteristics, and hydrologic forcing along the freshwater to marine continuum representing estuarine and coastal ecosystems

Nutrient Enrichment and Limitation in Estuarine and Coastal Ecosystems: Historical and Current Perspectives

The dominant nutrient limitation (of primary production) paradigms applied to this continuum for more than half a century were that phosphorus (P) availability controls primary production in freshwaters, while nitrogen (N) is the dominant nutrient limiting production in the more saline downstream estuarine and coastal waters (Ryther and Dunstan 1971; Schindler 1975; Nixon 1995; Boesch et al. 2001; Smith and Schindler 2009). Brackish waters often exhibit sensitivity to both N and P inputs (Fisher et al. 1999; Rudek et al. 1991; Elmgren and Larsson 2001; Paerl and Piehler 2008). Recent analyses of diverse nutrient limitation studies in both freshwater and marine ecosystems indicate that these paradigms may be “eroding” (Elser et al. 2007; Lewis and Wurtsbaugh 2008; Sterner 2008; Conley et al. 2009; Lewis et al. 2011). Increasingly, incidences of N and P “co-limitation”, i.e., the stimulation of primary production by the addition of N and P in combination, where either N or P alone stimulate production far less, are observed (Elser et al. 2007; Lewis et al. 2011). Also, exclusive N limitation in freshwater ecosystems (as opposed to exclusive P limitation) is more common than believed previously (Elser et al. 2007; Lewis et al. 2011). Concurrently, recent estuarine and coastal studies indicate that N and P and/or P limitation are widespread geographically (Peeters and Peperzak 1990; Elmgren and Larsson 2001; Sylvan et al. 2006; Paerl and Justić 2011; Laurent et al. 2012).

Liebig’s “Law of the Minimum”: Theory Versus Practice

In examining nutrient limitation of primary producers, Liebig’s Law of the Minimum, also referred to as Liebig’s Law or the Law of the Minimum, is an operational principle first developed in agricultural science by Carl Sprengel (1839) and later popularized by Justus von Liebig (cf., Brown 1942). It states that plant growth is controlled not by the total amount of resources available but by the scarcest resource (limiting factor). From an aquatic plant production perspective, the yield is proportional to the amount of the most limiting resource (i.e., nutrient, light, etc.). If the limiting resource is a nutrient, it follows that yields may be improved by supplying the limiting nutrient to the point that some other nutrient is needed in greater quantity than the water body can provide. In the case of N and P co-limitation, the balance between N and P supply may approach the demand so that the addition of both nutrients stimulates primary production, but separate additions are not effective.

The application of Liebig’s Law was demonstrated in chemostat cultures with single phytoplankton species, where nutrient supplies could be controlled tightly and growth yield regulated by an essential nutrient supplied in amounts below those needed to maintain optimal cellular growth. Nutrients supplied to these cultures can be manipulated at ratios and rates very close to those needed to maintain balanced growth (i.e., “Redfield ratio”, Redfield 1958; Redfield et al. 1963). Only a slight increase in the supply of one nutrient will shift the control of yield to the other nutrient. In natural systems supporting complex phytoplankton communities, such nutrient supply shifts can occur as a result of variability in external nutrient inputs (loads), internal nutrient cycling, and sediment–water column nutrient exchange. Furthermore, changes in plant community composition, due to death, grazing, and plant–microbe and higher trophic level interactions (e.g., nutrient regeneration from zooplankton grazers up to fish), also affect nutrient availability and hence nutrient limitation. Thus, even subtle shifts in nutrient supply, community composition, and biogeochemical cycling can affect the nature and complexity of nutrient limitation.

Nutrient Limitation Paradigms

Historically, a much longer and geographically diverse line of nutrient limitation data were available for freshwater than marine ecosystems, likely because the symptoms of nutrient over-enrichment and eutrophication were more evident and problematic in freshwater ecosystems, dating back several centuries to the establishment and expansion of agriculture (i.e., rapid increase in chemical fertilizer use), the industrial revolution, and urbanization (Thienemann 1915; Naumann 1921; Parma 1980).

Freshwater and estuarine ecosystems often have larger water- and airshed areas relative to their surfaces than do more open coastal marine ecosystems or large, deep lakes. Water replacement rate, relative to the volume of that water body (i.e., flushing rate), is often slow in shallow enclosed systems. Therefore, from a nutrient input and enrichment perspective, these systems are influenced heavily by their water- and airsheds. They tend to be dominated by N inputs since N compounds are often soluble and associated with a wide variety of organic and inorganic sources (e.g., plants, soils, atmospheric emissions and combustion products, and microbial transformations), whereas P is associated with rocks and soils, where it is often insoluble and therefore unavailable. Furthermore, N, unlike P, exists in several oxidation states, including significant gaseous forms, and is more mobile and easily transported and transformed in the geospheres, biospheres, and atmospheres. As such, freshwater to oligohaline estuarine systems are often enriched in N relative to P and exhibit P limitation.

Exceptions to this paradigm often relate to specific watershed geochemical characteristics. Silicon (Si) may be deficient in watersheds that are dominated by non-silicious rocks and soils (e.g., carbonates). This pattern can lead to Si limitation, especially for diatoms, in downstream N- and P-enriched waters (Justić et al. 1995; Dortch and Whitledge 1992). Furthermore, construction of upstream dams and reservoirs can promote “trapping” of Si-containing soils and sediments, causing Si deficiency downstream (Humborg et al. 2007; Chai et al. 2009).

Phosphorus limitation occurs commonly in relatively undisturbed watersheds (cf. Wetzel 2001), supporting the early supposition that P is the limiting nutrient in most freshwater ecosystems (cf., Likens 1972). Unfortunately, only a few “undisturbed” air- and watersheds remain to evaluate this paradigm. Today, agricultural, urban, and industrial expansions have altered the landscape, and amounts and patterns of nutrient loading to freshwater ecosystems. These activities have increased both N and P loading, with wastewater inputs and runoff from land clearing and the establishment of farmland and urban centers as dominant nutrient sources. The early recognition of P as a primary limiting nutrient in these systems (Likens 1972; Schindler 1975), and the linkage of P loading to freshwater eutrophication (Vollenweider 1968), provided the impetus for focusing on P input reductions (Schindler and Vallentyne 2008). Indeed, such reductions were effective in stemming and reversing problematic symptoms of eutrophication, nuisance algal blooms, food web disruption, bottom water hypoxia, and degradation of planktonic and benthic habitats (Wetzel 2001; Schindler and Vallentyne 2008).

Nutrient loading dynamics have changed dramatically over the past several decades. While P reductions were pursued, human population growth and agricultural and urban expansion in watersheds were paralleled by increased rates of N loading (Peierls et al. 1991; Howarth 1998), often exceeding those for P (Rabalais 2002; Galloway and Cowling 2002). In the Baltic Sea region, subjected to several centuries of human nutrient enrichment, effective control of eutrophication requires considering the total amounts and ratios of N and P discharged to a nutrient-sensitive, river–fjord–sea continuum (Elmgren and Larsson 2001; Conley et al. 2009). Similarly, single nutrient input reductions, including a P-detergent ban and improved wastewater treatment for P during the 1980s in North Carolina’s (USA) Neuse River System, helped arrest freshwater blooms, but failure to reduce N inputs simultaneously exacerbated blooms in downstream N-sensitive estuarine waters (Paerl et al. 2004). In both cases, parallel N and P input reductions were required to stem eutrophication (Elmgren and Larsson 2001; Paerl 2009). In Florida’s (USA) extensive lake–river–estuary systems, excessive N loading, mainly from expanding wastewater and agricultural discharges, was identified (in addition to P) as supporting eutrophication (Kratzer and Brezonik 1981). N2-fixing cyanobacteria often dominate in Lake Okeechobee, Florida’s largest lake. However, non-N2-fixing genera (e.g., Microcystis) and facultative N2-fixing genera (e.g., Cylindrospermopsis sp., Lyngbya sp.) compete effectively for reactive N when it is available and only fix N2 when other available N is depleted (Moisander et al. 2012). In all cases, both N and P reductions are needed to control eutrophication and harmful (toxic, hypoxia-generating) cyanobacterial bloom genera (Howarth et al. 2000; Havens et al. 2001).

Lake Erie, USA–Canada, seemed to have “recovered” from eutrophication due to P (but not N) reduction programs (Schindler 2012), but eutrophication resurged with phytoplankton dominated by non-N2-fixing cyanobacteria (Microcystis sp.; Lyngbya sp.) (cf., Steffen et al. 2014). Oligo- to mesohaline regions of large estuaries and coastal bays and seas (e.g., Chesapeake Bay, Albemarle–Pamlico Sound, NC, Florida Bay, FL, Coastal North Sea and Baltic Sea; Moreton Bay, Australia) also reveal N and P co-limitation (Peeters and Peperzak 1990; Rudek et al. 1991; Fisher et al. 1999; Elmgren and Larsson 2001; Watkinson et al. 2005; Ahern et al. 2007). This co-limitation occurs because previously loaded P and N are retained and recycled. This observation is especially true for P, which exists only in a limited number of soluble (orthophosphate, dissolved organic P) and particulate forms, which are cycled between the water column and bottom sediments (Vollenweider 1968; Boynton and Kemp 1985; Wetzel 2001).

In contrast, N exists in multiple forms, including dissolved (nitrate, nitrite, ammonium, organic N), particulate, and gaseous; P has no significant gaseous forms. Gaseous forms of N (N2, N2O, NO, NO2, NH3, volatile organic N compounds) are produced by microbial transformations, including ammonification, anammox, nitrification, and denitrification (Capone et al. 2008), which can escape into the atmosphere. In particular, denitrification is a major N loss mechanism. However, this process does not keep up with externally supplied “new” N inputs, especially in systems impacted by N over-enrichment (Seitzinger 1988; Nixon et al. 1996).

Agricultural and domestic synthetic fertilizers, fossil fuel combustion, wastewater treatment, and a wide range of industrial chemical processes are major sources of biologically reactive N to estuarine and coastal waters, and have increased dramatically over the past half-century (Galloway and Cowling 2002; US EPA 2011). In Northern Gulf of Mexico waters receiving discharge from the Mississippi River Basin, agricultural and urban N inputs in the Basin have increased so rapidly that the receiving marine waters now exhibit P limitation during spring with elevated runoff, while N limitation prevails during the drier summer months (Sylvan et al. 2006). Anthropogenically generated P has also increased, but in many instances not nearly as rapidly as N (Justić et al. 1995), especially in places where P detergent bans and improved wastewater treatment for P have been implemented. In many intensively farmed, urbanized, and industrialized regions, however, historic and current P loads are still quite high because of P-saturated soils and continued P fertilizer applications. As a result, residual P supplies in sediments have remained high and available.

While a N input “glut” is occurring due to expanding anthropogenic inputs, a fraction of the N supplied to water bodies is “lost” as N2 via denitrification (Seitzinger 1988) or converted to other gaseous forms (e.g., NH3, N2O, and NO emissions) (US EPA 2011). N2 fixation rates are not sufficient to offset N losses via denitrification and anammox in these systems, perpetuating N-limited conditions (Paerl and Scott 2010). Thus, despite receiving ever-increasing anthropogenic N inputs, these systems can still assimilate such inputs and become more eutrophic, without becoming exclusively P limited, due to N losses via denitrification and other gas-generating processes, while much of the P is internally cycled. This phenomenon appears widespread in meso- to eutrophic freshwater and marine ecosystems, which exhibit N limitation or N and P co-limitation (Granéli et al. 1999; Elmgren and Larsson 2001; Elser et al. 2007; Sterner 2008; Paerl and Piehler 2008; Finlay et al. 2010; Lewis et al. 2011).

Interestingly, nutrient management efforts in freshwater components of the estuarine continua continue to focus largely on “P only” reduction strategies (cf., Schindler and Vallentyne 2008; Schindler et al. 2008), despite pioneering studies in the 1960s showing a role for N in freshwater eutrophication (cf., Goldman 1981; Wetzel 2001), and more recent studies demonstrating sensitivity of a range of lakes and reservoirs to N inputs (Elser et al. 2007; North et al. 2007; Lewis and Wurtsbaugh 2008; Finlay et al. 2010; Xu et al. 2010; Paerl et al. 2011b; Spivak et al. 2011; Lewis et al. 2011). This approach is based on the assumption that cyanobacteria can supply N via nitrogen (N2) fixation (Schindler et al. 2008). However, at the ecosystem level, only a fraction, usually far less than 50 %, of primary production demands are met by N2 fixation, even when P supplies are sufficient (Scott et al. 2008; Paerl and Scott 2010). As a result, “perpetual N limitation” can occur in many freshwaters due to seasonal inorganic N drawdown (Scott et al. 2009). This chronic N deficit appears to be even more pronounced for estuarine and coastal waters (Howarth et al. 1988). Indications are that N2 fixation is controlled by factors in addition to just P availability (Paerl 1990; Scott and McCarthy 2010). N-limitation may persist in aquatic ecosystems, even in the presence of N2 fixers. Therefore, external N inputs play a key role in controlling primary production along the continuum (cf., McCarthy et al. 2007; Conley et al. 2009; Paerl 2009; Paerl et al. 2011b).

From a nutrient management perspective, important questions and research needs include (1) How do patterns of reactive N drawdown lead to seasonal N limitation or N + P co-limitation? (2) How is this drawdown partitioned between phytoplankton N demand and denitrification? (3) What is the critical balance (i.e., threshold) between N/P and loading rates and removal processes (i.e., uptake, denitrification) over episodic, seasonal, and multi-annual time scales?

The Interacting Roles of Climate Change: Warming and Increased Hydrologic Variability

Climatic changes, specifically warming and altered rainfall amounts and patterns, interact with nutrient enrichment in modulating eutrophication dynamics. Temperature controls microbial metabolism and hence carbon and nutrient cycling. In addition, it controls microalgal growth rates, and the controls appear taxa-specific (cf., Paerl et al. 2011a) (Fig. 2). Most notable is the stimulatory effect of warming on growth rates of cyanobacteria, the only prokaryotic phytoplankton group (Paerl and Huisman 2008). Being “bacterial”, cyanobacterial species tend to show growth optima that are in the 25–30 °C range, in contrast to eukaryotic taxa, which typically exhibit growth maxima at lower temperatures (Paerl et al. 2011a). Therefore, longer, warmer growing seasons favor cyanobacterial species (Jōhnk et al. 2008; Paerl et al. 2011a). Also, there are taxa-specific effects of warming on eukaryotic phytoplankton groups. For example in the Baltic Sea, shifts in dominance from diatoms to dinoflagellates were linked to long-term warming trends (Kraberg et al. 2011). It follows that differential effects of warming can alter phytoplankton community composition and hence the roles of phytoplankton in nutrient and carbon cycling, food web dynamics, and water quality (Scheffer et al. 2001; Elliott et al. 2005; Feuchtmayr et al. 2009; Jeppesen et al. 2010; Moss et al. 2011; Paerl and Paul 2011).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig2_HTML.gif
Fig. 2

Temperature dependence of the specific growth rates of representative species from three eukaryotic phytoplankton classes and of bloom-forming cyanobacterial species common in temperate freshwater and brackish environments. Data points represent a 5 °C moving average of from laboratory growth experiments under light- and nutrient-saturated conditions. References for the individual species can be found in Paerl et al. (2011b). Solid lines are a best-fit temperature optimum function (Chen and Millero 1986)

While these studies have largely focused on freshwater systems, estuarine and coastal systems are also affected (Paerl and Paul 2011; Kraberg et al. 2011) since freshwater and marine phytoplankton taxa overlap in these systems. In addition, surface warming enhances vertical density stratification, especially in oligohaline waters. Changes in the strength, distribution, and duration of stratification affect phytoplankton community structure by favoring motile taxa such as dinoflagellates, other flagellated species, and buoyant cyanobacteria over passive sinking taxa like diatoms (Reynolds 2006; Hall and Paerl 2011). Interestingly, phytoplankton groups containing harmful (i.e., toxic, food web disrupting) species, namely cyanobacteria and dinoflagellates, are favored selectively by warming effects, and warming appears responsible for the geographic expansion of toxic cyanobacterial bloom genera, including Anabaena, Cylindrospermopsis, Microcystis, and Lyngbya. Examples include lakes in Northern Europe (Padisak 1997; Stüken et al. 2006; Wiedner et al. 2007) and the Baltic Sea (Suikkanen et al. 2007). Temperature regimes and relative cyanobacterial dominance were related positively for 146 lakes along a latitudinal gradient ranging from the sub-Arctic to southern South America (Kosten et al. 2012).

In estuarine and coastal benthic environments (seagrass beds, reefs, subtidal shelf, and intertidal mudflats), filamentous attached cyanobacteria (Lyngbya spp., Oscillatoria spp.) are proliferating in systems that are impacted simultaneously by warming and nutrient enrichment (Paul 2008; Paerl and Paul 2011). Examples include Moreton Bay, Queensland, Australia (Watkinson et al. 2005; Ahern et al. 2007), coastal Florida (Paerl et al. 2008), and Guam (Kuffner and Paul 2001). Detrimental effects include smothering of seagrass and coral communities, hypoxia, an increase in coral diseases (e.g., “black band disease”, caused by cyanobacteria), and declining finfish and shellfish habitats.

Climatic changes also affect magnitudes, geographic distributions, and temporal patterns of precipitation (Trenberth 2005; Webster et al. 2005; O’Goman 2012. In some regions, both the amounts and extremes of precipitation are altered, as evidenced by record droughts and floods and changes in the frequency and intensity of tropical cyclones (Webster et al. 2005; Emanuel et al. 2008). These hydrologic changes impact phytoplankton production and bloom dynamics by (1) altering, and as a result of extreme precipitation events, enhancing, nutrient loading by increasing erosion potentials and mobilizing land-based nutrients; (2) in the case of protracted droughts, increasing water residence time, which helps promote algal blooms, especially among slower-growing species (e.g., cyanobacteria, some dinoflagellates); (3) increasing water column stratification, which benefits motile/buoyant bloom-forming species (e.g., dinoflagellates, cyanobacteria); and (4) influencing the location and magnitude of phytoplankton production.

Hydrologic variability influences both the amounts and proportions of nutrients delivered to estuarine and coastal waters. The interactive effects of nutrient loading and hydrologic variability on estuarine phytoplankton production and community dynamics are particularly evident in poorly flushed (i.e., long water residence time) estuaries, which provide the opportunity to observe the fate of freshwater and nutrient inputs over spatial and time scales that overlap with those required for growth and bloom responses (i.e., relatively free of tidal flushing). Two neighboring semi-lagoonal estuaries located in Eastern North Carolina, the New and Neuse River Estuaries (Fig. 3), provide an opportunity to examine these interactions. Both rivers are monitored routinely for water quality parameters, and phytoplankton biomass and composition (Paerl et al. 2010; Peierls et al. 2012; Hall et al. 2013).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig3_HTML.gif
Fig. 3

Maps of the a New River Estuary and b Neuse River Estuary, NC, showing the locations of sampling stations and US Geological Service (USGS) river gaging stations. The USGS gaging station on the Neuse River is located at Fort Barnwell approximately 26 km upstream from New Bern and out of the area covered by (b)

Both systems have a history of anthropogenic (urban, agricultural, and industrial) nutrient loading (Paerl et al. 1995, 2007; Mallin et al. 2005). They have also been influenced by increasing hydrologic variability, including a recent increase in tropical cyclone activity interspersed with record droughts (Paerl et al. 2001, 2006, 2010; Peierls et al. 2012). The Neuse River Estuary has had an intensive water quality and phytoplankton dynamics monitoring program in place since the early 1990s and hence is well suited for examining these interactions over a relatively long time frame (18 years). High rainfall tropical cyclones and extratropical nor’easters have played a key role in modulating N and P loads to the estuary (Fig. 4). Not surprisingly, storms that deposited high amounts of rainfall in the estuarine watersheds were associated with high nutrient loads. Far lower seasonal and annual nutrient loads occurred during years without such events (e.g., 1994).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig4_HTML.gif
Fig. 4

Dissolved inorganic nitrogen (a) and dissolved inorganic phosphorus (b) loading, as metric tons per day, entering the Neuse River Estuary during years having no major tropical cyclones impacting its watershed (1994) versus years in which storms (named) impacted the watershed. Note the variability in nutrient loads associated with specific storms, which was largely attributed to the storm tracks relative to the watershed and the amount of rainfall deposited by each storm

Total external N and P loads to both estuaries are controlled by freshwater discharge, and the ratios of these nutrients are influenced by the magnitude of discharge. A strong positive relationship between dissolved inorganic nitrogen (DIN) and total N concentrations and discharge under low to moderate discharge conditions occurs in both the Neuse and New River Estuaries (Fig. 5). However, DIN and TN concentrations drop dramatically under elevated discharge conditions (Fig. 5), presumably from dilution (Borsuk et al. 2004). DIP and TP concentrations are also strongly influenced by discharge over the entire range of discharge conditions (Fig. 5). In fact, TP concentrations actually increase under very high discharge conditions (Fig. 5). This suggests that the mobilization and transport dynamics of N and P fractions by freshwater discharge differ substantially under different flow regimes. Soluble forms of N and P are most dominant under low discharge conditions, while particulate fractions, including suspended sediments, play a larger role under high discharge conditions (data not shown). Total P concentrations tended to increase under very high discharge conditions, presumably due to the elevated suspended sediment load (from upstream soil and riverbed erosion) under these conditions. As a result, molar ratios of DIN to DIP and TN to TP are relatively small under low discharge, increase to maximum values during moderate discharge, and then fall under high discharge conditions (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig5_HTML.gif
Fig. 5

Concentrations and molar ratios of nitrogen and phosphorus versus river flow in the Neuse and New Rivers just above the limit of salt intrusion. River flow and nutrient concentrations were measured at the USGS Gum Branch gaging station for the New River. Neuse River flow was measured at Fort Barnwell and nutrient concentrations were measured at Streets Ferry Bridge

These shifts in N/P supply ratios likely affect relative nutrient availabilities, limitation, and microalgal utilization/growth dynamics, which can influence inter-taxa competition and community structure in downstream waters (Tomas et al. 2007; Altman and Paerl 2012). These interactions are potentially important yet poorly understood examples of how climatic changes may impact estuarine and coastal eutrophication and phytoplankton/benthic microalgal community structure (including harmful algal blooms, HABs).

The effects of rainfall events on coastal ecosystems are complex and interactive. For example, elevated nutrient loads associated with freshwater discharge provide resources for resident microalgal communities; however, high flow can advect phytoplankton cells downstream or even out of the estuary. Hence, the relative magnitudes of rainfall in these storms as well as their trajectory across the watershed and estuary are important determinants of the magnitudes and locations of phytoplankton biomass and bloom responses. Examples of several storm events are shown for the Neuse River Estuary, NC (Fig. 6).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig6_HTML.gif
Fig. 6

Impacts of four tropical cyclones on river flow and the downstream distribution of chlorophyll a in the Neuse River Estuary. Dates for each storm represent landfall on the North Carolina coast. Letters above the flow panel correspond with the letters on the chlorophyll a distribution panels. Off-scale chlorophyll a values are written below the peaks. Extremely high flow events, such as the one following Hurricane Floyd (1999), while delivering large amounts of nutrients, also led to “washout” of resident phytoplankton communities, thereby negating the potential stimulatory effect of this event

In a most extreme hydrologic scenario, rainfall and flooding from sequential Hurricanes Dennis (10 days prior to Floyd), Floyd, and Irene (30 days after Floyd) impacted the Neuse River watershed over a period from late August through October 1999. The Neuse River Estuary was flushed completely for more than 2 weeks following Floyd, leading to very low levels of resident phytoplankton throughout the estuary (Fig. 6). After about 3 weeks, phytoplankton biomass began to build in the lower estuary, but increased flows due to Irene prevented bloom development. Throughout the 2-month period following Dennis and Floyd, high flushing losses did not permit biomass accumulation in the upper half of the estuary. Phytoplankton growth rates did not catch up with flushing losses until late November when a phytoplankton bloom developed in the mid-estuary region (Fig. 6). Decreasing light availability from November to January prevented development of significant additional blooms in the estuary (data not shown).

Hydrologic impacts of Hurricane Isabel (mid-September, 2003) were less severe. While Isabel was a powerful storm (Cat. 2), it contributed much smaller amounts of rainfall to the Neuse River watershed (<20 cm) than the massive deluge (∼1 m) that resulted from Floyd (1999) (Paerl et al. 2001). Prior to Isabel, phytoplankton blooms were present at upstream and mid-estuary stations (Fig. 6). Passage of this storm had little effect on phytoplankton biomass except for a slight downstream shift in peak biomass (Fig. 6). Lack of a significant growth stimulation of biomass may have resulted from the high nutrient concentrations existing prior to the storm (Wetz and Paerl 2008).

Heavy rainfall from tropical storm Ernesto produced strong freshwater discharge to the Neuse River Estuary in late August 2006 (Figs. 6 and 7a). Phytoplankton populations were largely flushed out of the upper reaches of the estuary as evidenced by the ∼50 % reduction of phytoplankton biomass at station 60 (Fig. 7d). The freshwater inputs from the storm led to rapid increases in DIN in both surface and bottom waters of the upper estuary. A month-long period of intense, salinity-based vertical stratification (Fig. 7b) allowed high levels of ammonium to accumulate in the bottom waters (Fig. 7c). As the storm waters receded and residence time increased, the combination of high nutrient availability and favorable residence times (Fig. 7d) permitted bloom development by the toxic dinoflagellate, Karlodinium venificum. Cell concentrations reached 219,000 cells mL−1 and the highly toxic cells were implicated in several fish kills in the area as the bloom crashed (Hall et al. 2008).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig7_HTML.gif
Fig. 7

Environmental conditions leading to the development of a toxic Karlodinium veneficum bloom in the Neuse River Estuary. a Neuse River flow at Fort Barnwell, NC showing the runoff pulse from Tropical Storm Ernesto during the fall of 2006. b Time series of surface (0 m) and bottom (3 m) salinity from automated USGS instrumentation at station CM11, ∼4 km downstream of station 60 where peak bloom concentrations were observed on 19 October. c Time series of surface and bottom dissolved inorganic nitrogen at station 60. Inset pie graphs below time series lines show the relative proportion of nitrate (black fill) versus ammonium (white fill) in the DIN pool of surface and bottom waters on 3 October. d Time series of surface water chlorophyll a, K. veneficum cell abundance, and flushing time at station 60 (see Fig. 3 for its location)

A further contrast was provided by Hurricane Irene (late August, 2011), whose eye also passed directly over the Pamlico Sound. The rain bands from this massive storm moved sufficiently inland to deliver large amounts of rainfall to the watershed. Prior to Irene, maximum phytoplankton biomass occurred in the upper regions (10–30 km downstream). After Irene, a well-defined peak in freshwater discharge increased flushing (Fig. 6), which pushed the phytoplankton biomass peak downstream, but not out of the estuary as observed after Floyd. Once freshwater discharge subsided and flushing rates decreased, phytoplankton biomass peaks resumed further upstream, where nutrient inputs were high.

Additional examples of how anthropogenic nutrient-driven eutrophication and algal bloom dynamics were modulated by climatically driven hydrologic variability were shown for the New River Estuary (NRE) (Hall et al. 2013; Peierls et al. 2012) (Fig. 8). This N-sensitive, microtidal system has exhibited a history of nutrient-enhanced primary production (Ensign et al. 2004; Mallin et al. 2005), HABs (Tomas et al. 2007), and bottom water hypoxia (Mallin et al. 2005). Phytoplankton biomass and composition are modulated by variations in flow of the New River, the main freshwater input to the estuary, due to its influence on nutrient delivery and flushing time of the NRE. Phytoplankton biomass increased rapidly up to a threshold flushing time of ∼10 days and then declined slowly at longer flushing times (Fig. 8).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig8_HTML.gif
Fig. 8

Relationship between phytoplankton biomass and flushing time in the New River Estuary, NC from October 2007 to September 2013

This unimodal relationship indicated a balance between advective losses due to flushing and nutrient stimulation of biomass by riverine loading. Significant differences among group-specific optimal flushing times and rates of decline during longer flushing times suggest that hydrologic forcing is important to determining phytoplankton composition and size structure (Hall et al. 2013; Paerl et al. 2013). Temperature also controlled phytoplankton composition. The diagnostic photopigment zeaxanthin, primarily associated with picoplanktonic forms such as cyanobacteria, showed a strong, positive relationship with water temperature. The potentially toxic, bloom-forming raphidophytes also occurred seasonally as temperatures warmed in late spring (Fig. 9).
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig9_HTML.gif
Fig. 9

Seasonality of picocyanobacteria and raphidophyte blooms in the New River Estuary. Contour plot of zeaxanthin concentration versus time and distance upstream from the New River Inlet with average temperatures from all stations (solid black line, error bars are standard deviation). White circles show the time and spatial locations of raphidophyte blooms defined as chlorophyll a concentrations greater than 40 μg L−1 and raphidophytes comprising more than half of total phytoplankton biovolume

Despite efforts to ameliorate eutrophication by reducing point source nutrient inputs through sewage treatment upgrades (Mallin et al. 2005), the New River is still impacted regularly by algal blooms, especially in the microtidal, upper estuarine region. Blooms are generally linked to elevated nutrient inputs in response to high flow periods. However, some blooms occur during droughts. This trend suggests that internal nutrient loading from the sediments may also play a critical role in bloom development. A positive feedback of phytoplankton biomass and sediment nutrient flux exists whereby blooms decrease light availability to the microphytobenthic community, decreasing benthic N demand, and increasing sediment N fluxes to the water column (Anderson et al. 2013).

Bloom-forming flagellate species are sensitive to changes in riverine nutrient inputs (Tomas et al. 2007; Altman and Paerl 2012). Sensitivity of the phytoplankton community was documented when sewage treatment upgrades reduced nutrient loading to the New River Estuary by ∼200,000 kg N year−1 and PP biomass fell by ∼70 % (Mallin et al. 2005). Prior to sewage treatment upgrades, silica, at times, potentially limited (∼0.5 μmol L−1) the growth of diatoms (Mallin et al. 1997), possibly explaining flagellate dominance of the NRE. Current silica concentrations (3–92 μmol L−1) are unlikely to limit diatom growth (Dortch and Whitledge 1992), yet blooms are still dominated by flagellates, including some HAB species.

While nutrient reduction strategies may help reduce the magnitude of these blooms, density-driven stratification, which is largely attributable to precipitation and freshwater runoff conditions, also likely plays an important role in determining phytoplankton community structure. Selective advantages gained by motility during stratified periods may explain why blooms in this estuary are dominated by flagellates.

Closing Remarks

Synergistic anthropogenic nutrient loading and climatic changes, including warming, enhanced vertical stratification, and periods of increased flushing time, provide the mechanisms for accelerating eutrophication and promoting an increase in HAB frequency and intensity in estuarine and coastal waters (Fig. 10). Nutrient input reductions are the most obvious “knobs” that can be “tweaked” to reduce estuarine and coastal eutrophication, and as such should be the key component of most management strategies. We have long been aware of the role P inputs play in such a strategy. However, increasing evidence suggests that N input reductions are also needed. A dual N and P reduction strategy offers the most realistic and effective long-term approach to eutrophication management along the estuarine continuum. A key management priority is to establish N and P input thresholds [e.g., the USEPA’s Total Maximum Daily Loads or TMDLs; the European Union Water Framework Directive—(http://ec.europa.eu/environment/water/index_en.htm)], below which eutrophication and HABs can be controlled in terms of magnitude, and temporal and spatial coverage. These thresholds should incorporate the effects of hydrologic and temperature conditions, predicted to vary in response to climate change, as these factors affect the ratio of N/P loads and modulate estuarine and coastal responses to nutrient inputs.
https://static-content.springer.com/image/art%3A10.1007%2Fs12237-014-9773-x/MediaObjects/12237_2014_9773_Fig10_HTML.gif
Fig. 10

Potential interactions among anthropogenic drivers, changes, and effects on estuarine and coastal ecology and resource use, which may benefit from controlling both nitrogen and phosphorus

Both the total amounts and ratios of N to P inputs should be considered when developing these thresholds. Ideal input ratios should not favor specific HAB taxa over others, but unfortunately a universal ratio—above or below—to control HABs consistently and reliably is not available. Thus, system-specific input ratios are needed for appropriate and effective control. Total molar N/P ratios above ∼15 may discourage CyanoHAB dominance (cf., Smith and Schindler 2009). However, if the nutrient concentrations in receiving waters of N or P exceed uptake saturation values, a ratio approach for reducing eutrophication and HAB formation may not be effective.

There are many ways to reduce nutrient inputs on ecosystem-specific scales. Nutrient inputs are classified as point source and non-point source. Point sources are associated with well-defined and identifiable discharge sites; therefore, these nutrient inputs are often considered “low hanging fruit”, i.e., relatively easy to control. Most short-term successes in nutrient input control were accomplished with point source reductions of wastewater, industrial effluents, and other distinct input sources. The major remaining challenge in many watersheds is targeting and controlling non-point sources, which frequently are the largest sources of nutrients discharged to coastal waters (National Research Council 2000; US EPA 2011); their controls are critical in mitigating eutrophication and HABs.

Manipulating physical factors that play key roles in controlling the composition and function of phytoplankton (and benthic microalgal) communities will benefit systems sensitive to eutrophication and HABs. Vertical mixing devices, bubblers, and other means of destratification are effective in controlling outbreaks and persistence of some HABs (e.g., cyanobacteria), but only in relatively small systems (cf., Hudnell 2008). These devices have limitations in estuarine and coastal waters because they are not applicable to large areas and volumes. Furthermore, artificial mixing does not mitigate the underlying problem of nutrient over-enrichment. In fact, this approach may be counterproductive in vertically stratified waters, by transporting bottom water nutrients across the pycnocline up to the surface, potentially exacerbating surface-dwelling blooms.

Increasing the flushing rates can reduce or control HABs in these systems (cf., Paerl et al. 2011b). However, the flushing water must have low nutrient content to prevent further enrichment. Algal community structuring effects of changing N/P ratios, which can take place as freshwater discharge is altered, must be considered. Furthermore, few communities can afford to use precious water resources normally reserved for drinking or irrigation water for flushing purposes, especially in regions with limited or drought-impacted freshwater runoff. Lastly, flushing can alter the circulation regimes of estuarine and coastal waters. Care must be taken to prevent trapping of the HABs in the system by altering the physical environment (e.g., increasing thermal or chemical density stratification, entrainment bays, and arms of water bodies), rather than flushing them out of the system.

Nutrient input reductions are in general the most direct, simple, ecologically rational, and economically feasible eutrophication and HAB management strategy. Nutrient input reductions aimed specifically at reducing HAB competitive abilities, and possibly combined with physical controls (in systems amenable to those controls), are often the most effective strategies. An obvious strategy which is gaining traction is applying fertilizers at “agronomic rates”, i.e., satisfying crop needs, while avoiding excesses and modifying drainage ditches and tile drains to increase their efficiency in minimizing nutrient losses (David et al. 2010). Nutrient (specifically N) removal from wastewater can be prohibitively expensive, so that alternative nutrient removal strategies are needed. Alternate strategies may include construction of wetlands, cultivation and stimulation of macrophytes, and stocking of herbivorous (specifically cyanobacteria consumers) fish and shellfish species.

In addition to nutrient input reductions, future water management strategies will need to accommodate the hydrological and physical–chemical effects of climatic change. Without curbing greenhouse gas emissions, future warming trends and hydrological extremeness will degrade estuarine and coastal ecosystem water and habitat quality further, increasing the potential for expansion by opportunistic nuisance microalgae and cyanobacteria.

Acknowledgments

We thank co-workers who assisted with field and laboratory work and manuscript preparation, including B. Abare, J, Braddy, A. Joyner, L. Kelly, and R. Sloup. The editorial input from I. Anderson and W. Gardner and review provided by K. McGlathery are appreciated. This research was funded by the Strategic Environmental Research and Developmental Program (SERDP)–Defense Coastal/Estuarine Research Program, Project SI-1413, The Lower Neuse Basin Association/Neuse River Compliance Association, the North Carolina Department of Environment and Natural Resources (ModMon Program), and National Science Foundation Projects OCE 0825466, OCE 0812913, ENG/CBET0826819 and 1230543, and DEB 1119704 and 1240851.

Copyright information

© Coastal and Estuarine Research Federation 2014

Authors and Affiliations

  • Hans W. Paerl
    • 1
  • Nathan S. Hall
    • 1
  • Benjamin L. Peierls
    • 1
  • Karen L. Rossignol
    • 1
  1. 1.UNC–CH Institute of Marine SciencesMorehead CityUSA