Advertisement

Biogeochemistry

, Volume 141, Issue 3, pp 307–332 | Cite as

Two decades of tropical cyclone impacts on North Carolina’s estuarine carbon, nutrient and phytoplankton dynamics: implications for biogeochemical cycling and water quality in a stormier world

  • Hans W. Paerl
  • Joseph R. Crosswell
  • Bryce Van Dam
  • Nathan S. Hall
  • Karen L. Rossignol
  • Christopher L. Osburn
  • Alexandria G. Hounshell
  • Randolph S. Sloup
  • Lawrence W. HardingJr.
Article

Abstract

Coastal North Carolina (USA) has experienced 35 tropical cyclones over the past 2 decades; the frequency of these events is expected to continue in the foreseeable future. Individual storms had unique and, at times, significant hydrologic, nutrient-, and carbon (C)-loading impacts on biogeochemical cycling and phytoplankton responses in a large estuarine complex, the Pamlico Sound (PS) and Neuse River Estuary (NRE). Major storms caused up to a doubling of annual nitrogen and tripling of phosphorus loading compared to non-storm years; magnitudes of loading depended on storm tracks, forward speed, and precipitation in NRE-PS watersheds. With regard to C cycling, NRE-PS was a sink for atmospheric CO2 during dry, storm-free years and a significant source of CO2 in years with at least one storm, although responses were storm-specific. Hurricane Irene (2011) mobilized large amounts of previously-accumulated terrigenous C in the watershed, mainly as dissolved organic carbon, and extreme winds rapidly released CO2 to the atmosphere. Historic flooding after Hurricanes Joaquin (2015) and Matthew (2016) provided large inputs of C from the watershed, modifying the annual C balance of NRE-PS and leading to sustained CO2 efflux for months. Storm type affected biogeochemical responses as C-enriched floodwaters enhanced air–water CO2 exchange during ‘wet’ storms, while CO2 fluxes during ‘windy’ storms were largely supported by previously-accumulated C. Nutrient loading and flushing jointly influenced spatio-temporal patterns of phytoplankton biomass and composition. These findings suggest the importance of incorporating freshwater discharge and C dynamics in nutrient management strategies for coastal ecosystems likely to experience a stormier future.

Keywords

Tropical cyclones Carbon/nutrient cycling Estuarine Coastal Phytoplankton North Carolina 

Introduction

Since the mid-1990s, increased activity of tropical cyclones has been observed in both the Atlantic and Pacific Basins (Goldenberg et al. 2001; Emanuel 2005; Webster et al. 2005; Holland and Webster 2007; Bender et al. 2010; IPCC 2014; Kang and Elsner 2015); this active period has surpassed previous cycles in both frequency and intensity of cyclones (NOAA Hurricane Center; Bender et al. 2010; Seneviratne et al. 2012; IPCC 2014). Physical perturbations of estuarine and coastal ecosystems, such as erosion, sediment re-suspension, regional flooding, increased freshwater inputs, and over-wash of wetlands and marshes, are often accompanied by massive inputs of nutrients and organic matter from affected watersheds (Conner et al. 1989; Paerl et al. 2000, 2006a, b, 2010a, b; Christian et al. 2004; Roman et al. 2005; Crosswell et al. 2014). These inputs may stimulate primary production, promote harmful algal blooms, and culminate in bottom water hypoxia (Paerl et al. 2006a, b, c; Harding et al. 2016).

Negative consequences for higher trophic levels are commonly observed, including habitat degradation, increased disease and mortality of fish and shellfish, abrupt changes of salinity, dissolved oxygen, and pH, and inputs of contaminants from affected watersheds (Ruzecki et al. 1976; Conner et al. 1989; Eby and Crowder 2002; Adams et al. 2003; Roman et al. 2005; Miller et al. 2006; Paerl et al. 2006a, b, 2010a, b; Phlips et al. 2011; Hagy et al. 2004, 2006). Prior studies have documented the impact of tropical cyclones on major biogeochemical processes in estuarine and coastal systems, including air–water CO2 exchange, nutrient loading and regeneration, and re-mineralization of soil organic carbon (C) (Bauer et al. 2013; Bianchi et al. 2013; Wetz and Yoskowitz 2013). However, most observational studies have focused on a single event or limited time period, which cannot be extrapolated to the broader impacts of changing storm activity. In order to quantify whether storms have a significant impact relative to baseline conditions requires long-term datasets that resolve a range of storms with diverse characteristics in a tractable study sytem.

Previous work documented profound impacts of tropical cyclones on lagoonal and semi-enclosed water bodies separated from oceans and seas by sandbars, barrier islands, and reefs because of restricted water exchange and relatively long water-residence times (Kennish and Paerl 2010; Paerl et al. 2010b; Laruelle et al. 2017). Lagoonal estuaries are low-energy ecosystems with small tidal amplitudes and low current velocities, promoting storage and recycling of materials (Kennish and Paerl 2010). These features prolong the recovery period following episodic perturbations, making them susceptible to cumulative impacts of closely spaced events (Paerl et al. 2006b). North Carolina’s Pamlico Sound (PS) is the largest lagoonal ecosystem and second largest estuarine complex in North America. This ecosystem and its largest microtidal sub-estuary, the Neuse River Estuary (NRE) that feeds into PS, exemplify these conditions and represent the focus of this study.

The NRE-PS is an ecosystem subject to increasing human activity, resulting in nutrient enrichment, sedimentation, accumulation of pollutants, and habitat disturbances from rapidly expanding agriculture, industry, and urbanization in the airshed and watershed (Stanley 1988; Copeland and Gray 1991; Stow et al. 2001; Paerl et al. 2006a, 2007, Burkholder et al. 2006). The relatively long residence time of NRE-PS (~ 1 year), combined with efficient use of nutrients and organic matter, make this system highly susceptible to nutrient over-enrichment and eutrophication, resulting in nuisance algal blooms, hypoxia, anoxia, finfish and shellfish diseases, and mass mortalities (Copeland and Gray 1991; Christian et al. 1991; Paerl et al. 1995, 2006a, 2010a, 2014; Buzzelli et al. 2002; Eby 2001; Eby and Crowder 2002). This ecosystem is also strongly influenced by climatic perturbations, particularly the passages of tropical cyclones and “nor’easters”, as it is located on the Atlantic “hurricane track” with a very active history of storm landfalls (NOAA Hurricane Center: http://www.nhc.noaa.gov). Climatologists have noted a “new climatic era” in the southeastern and mid-Atlantic coastal region of the United States, characterized by increased intensity and frequency of storms since the mid-1990s (Goldenberg et al. 2001; Webster et al. 2005; Holland and Webster 2007; Bender et al. 2010; Seneviratne et al. 2012; IPCC 2014; Wuebbles et al. 2014). The coastal region of North Carolina has experienced 35 tropical storm or hurricane-strength landfalls from 1996 through 2016, notably Hurricanes Bertha and Fran in 1996, Bonnie in 1998, three successive cyclones, Hurricanes Dennis, Floyd and Irene, in a 6-week period in September and October 1999, Isabel in 2003, Charlie and Alex in 2004, Ophelia in 2005, Tropical Storm Ernesto in 2006, Hurricanes Irene in 2011, Arthur in 2014, and most recently, Matthew in 2016 (Fig. 1). Three of these storms, Hurricanes Fran, Floyd and Matthew, caused “100–1000-year flood events” in the watershed of Bales (2003), Bales et al. (2000), https://www.colorlines.com/articles/noaa-declares-hurricane-matthew-1-1000-year-flood-event.
Fig. 1

Named storms that have impacted the North Carolina coastline since 1996. Storm track data are from the NOAA National Hurricane Center: Storm tracks with solid lines directly impacted the coastline, while tracks with open lines (e.g., Nicole, Joaquin) did not directly cross the study area, but impacted it with heavy rains and winds

Effects of recent tropical cyclones on water quality, habitat, and fisheries in NRE-PS vary substantially in type, magnitude, and duration (Paerl et al. 2001a, b, 2006a, b; Peierls et al. 2003; Paerl and Peierls 2008; Adams et al. 2003; Tester et al. 2003), with large storm events exerting biogeochemical impacts lasting from several months into subsequent years (Peierls et al. 2003; Burkholder et al. 2004; Paerl et al. 2006a, b). As such, these storms pose a recurring threat to NRE-PS and other low-lying estuarine-coastal ecosystems along the US East Coast and Gulf of Mexico that have a number of large lagoonal or semi-lagoonal estuaries, bays, and sounds draining the coastal plains (i.e., Chesapeake Bay, Pamlico Sound, Florida Bay, Lake Ponchartrain, Texas lagoonal estuaries and bays). Fully evaluating the biogeochemical and ecological effects of hurricanes and other intense storms affords an opportunity to understand the linkages between coastal watershed geology, hydrology, biogeochemistry, water quality, and fisheries to resiliency of lagoonal ecosystems worldwide (e.g., S.E. Asia, Australia).

In this contribution, we focus on the magnitudes of nutrient, inorganic and organic C loading and cycling, and responses of phytoplankton production and composition to tropical storm events. Phytoplankton account for ~ 80% of “new” primary production to sustain planktonic and benthic food webs in NRE-PS (Paerl et al. 1998); thus quantifying phytoplankton responses is important to develop predictive capabilities for trophodynamic responses to tropical cyclones. Large-scale responses to tropical cyclones are superimposed on seasonal and inter-annual variability associated with human impacts and natural processes, requiring that we distinguish these effects to resolve impacts. To this end, we draw on extensive data from long-term monitoring programs to define relationships between nutrient and C inputs, impacts on water quality, habitat conditions, and biological responses from storm-related perturbations in comparison to more chronic, seasonal patterns. We classify three different types of storms based on rainfall and wind speed, and quantify estuarine responses to these different storm types relative to non-storm conditions.

Methods and materials

Neuse River Estuary-Pamlico Sound (NRE-PS) study site

NRE is the second largest tributary of PS in terms of freshwater discharge. Its watershed is comprised of managed forests and rapidly expanding agricultural (animal and row-crop operations), rural, urban (Raleigh-Durham-Research Triangle) and industrial land uses. Anthropogenic N and P loadings to NRE approximately doubled in the past four decades (Stow et al. 2001). Urban regions have sprawled in the upper NRE watershed, while industrial-scale animal (swine, poultry) farms in the coastal plain region proliferated, with an increase of swine from less than 1 million in 1989 to more than 12 million today. Excessive nutrient inputs dominated by non-point sources (80%) associated with these expanding human activities led to expressions of eutrophication, including nuisance algal blooms, hypoxia, and food web alterations (Copeland and Gray 1991; Rudek et al. 1991; Mallin et al. 1993; Paerl et al. 1995, 1998, 2006c; NC Dept of Environment and Natural Resources, Division of Water Qulaity 2001; Burkholder et al. 2006). Changes in the landscape fueling eutrophication are also reflected in organic matter loading to NRE (Osburn et al. 2016a).

PS has a surface area of 5335 km2 and drains five major watersheds (Neuse, Tar-Pamlico, Roanoke, Chowan, and Pasquotank Rivers). These watersheds cover an area ~ 80,000 km2, total freshwater discharge of ~21 km3 y−1, and drain about 40% of North Carolina’s and 10% of Virginia’s surface area (Table 1).
Table 1

General characteristics of the Neuse River Estuary and Pamlico Sound, North Carolina

Characteristic

Neuse River Estuary

Pamlico Sound

Watershed size (km2)

16,108

80,400

Surface area (km2)

455

5335

Average depth (m)

2.7

5

Discharge (m3 s−1)

50–1000

600–1500

Freshwater flushing time (d)

7–200

200– > 400

Hypoxic area (% of bottom)

20– > 70

Not known

Tidal exchange with the coastal Atlantic Ocean and nearby Gulf Stream is restricted to three narrow inlets, resulting in a ~ 1 year residence time during average hydrologic years (Pietrafesa et al. 1996). The long residence time allows resident phytoplankton and vascular plants ample time to assimilate nutrients, resulting in high productivity per unit nutrient input. These characteristics are key to the PS serving as a highly productive nursery, supporting ~ 80% of US mid-Atlantic commercial and recreationally caught finfish and shellfish species (Copeland and Gray 1991; NC Dept. of Environmental Quality, http://portal.ncdenr.org/web/apnep/fastfacts).

Sampling sites and methodologies

Data analyzed here were obtained from two long-term monitoring programs in NRE, ModMon and FerryMon, designed to cover NRE and western PS (Fig. 2; Table 2). ModMon is a collaborative University—State of North Carolina (NC-Dept. of Environmental Quality-DEQ), and UNC-Chapel Hill Institute of Marine Sciences (IMS) program (www.unc.edu/ims/neuse/modmon/index.htm), that began in 1994. Sampling consisted of twice monthly visits to 11 mid-river stations along the estuarine portion of NRE (Fig. 2), including vertical profiles with collection of near-surface and near-bottom water for water quality (Paerl et al. 1995, 2006a, b). Monthly samples were also collected at nine stations in the western PS as part of the ModMon program starting in 2000 (Fig. 2). Profiles of temperature, salinity, and dissolved oxygen were made at 0.5 m depth intervals using a YSI 6600 multi-parameter water quality sonde (Yellow Springs Inc., Yellow Springs, OH). Sondes were calibrated prior to each sampling trip according to the YSI User’s Manual. Analytical uncertainties of all reported measurements are provided in Appendix I within the Supplementary Information.
Fig. 2

Map of the Neuse River Estuary (NRE) and Pamlico Sound (PS) system, showing ModMon sampling stations (by numbers referred to in text) and the FerryMon crossings

Table 2

Comparison of sampling programs employed in this study

Program

Location

Resolution

Key measurements

Spatial

Temporal

ModMon

Along river axis

73 km coverage;

1–14 km (distance between sites)

Bi-weekly

YSI vertical sonde profiles (temperature, salinity, pH, O2, chlorophyll a., turbidity); Discrete: DOM & POM, CDOM, inorganic nutrients, HPLC-pigments, preserved phytoplankton and microzooplankton, primary productivity

FerryMon

Across river channel and sound

~ 20 m horizontal (sampling resolution)

NRE = 34 crossings d−1

PS = 4 crossings d−1

YSI sonde near surface (as above)

Discrete: inorganic nutrients, DOM and POM, CHN, HPLC-pigments

Samples for nutrient concentrations, phytoplankton biomass, and community composition were collected along the ModMon transect in NRE from 1994 through 2016, and in the western PS from 2000 through 2016. Water samples were collected at the surface (0.2 m depth) and bottom (0.5 m above bottom) using a non-destructive diaphragm pump, dispensed into 4 L polyethylene bottles, and returned to the laboratory at the UNC-CH Institute of Marine Sciences, Morehead City (IMS) for processing within 4 h of collection.

FerryMon (www.ferrymon.org) consists of two N.C. Dept. of Transportation ferries equipped with automated, water-quality monitoring systems (Buzzelli et al. 2003; Ensign and Paerl 2006; Paerl et al. 2009). The NRE ferry makes 34 crossings per day between Cherry Point and Minnesott Beach, a mid-estuarine location in the mesohaline salinity zone of NRE characterized by high primary productivity and seasonal hypoxia (Fig. 2). Water is drawn from a through-hull intake located at ~ 1.5 m depth where it passes through a debubbler and a flow-through chamber that houses a YSI 6600 multi-parameter data sonde. Temperature, salinity, dissolved oxygen, turbidity, pH, and chlorophyll-a (Chl-a) via in vivo fluorescence are measured at a sampling rate of 0.25 Hz, and data are transmitted to a shore-based data server (IMS) via a cellular modem. Continuous data were augmented by discrete samples for nutrients and phytoplankton photopigments using automated ISCO samplers at programmed intervals. Physical–chemical parameters (Table 2) at NRE stations were also sampled monthly by NC-DEQ, and data were provided in a common database maintained by UNC-CH IMS and NC-DEQ.

Water Quality monitoring in NRE was expanded in 2009 to include high-resolution surveys of surface-water partial pressure of carbon dioxide (pCO2) to estimate air–water fluxes (Crosswell et al. 2012). These spatial surveys of pCO2 were coordinated with water-quality monitoring in PS, and the NRE ferry was fitted with an autonomous pCO2 system in 2011 to study the impact of Hurricane Irene (Crosswell et al. 2014). For periods where direct pCO2 measurements were unavailable (prior to 2009), pCO2 was calculated from measured DIC and pH (NBS) via CO2Sys (Lewis et al. 1998), using the carbonic acid dissociation constants of Millero (2010). This approach may overestimate pCO2 by upwards of 10% in organic rich, acidic water, due to the relatively large contribution of organic acids to total alkalinity (Abril et al. 2015). However, this error is reasonably small, relative to the large spatial and temporal variability of pCO2 in these estuaries (Crosswell et al. 2012; Van Dam et al. 2018), particularly during storm events (Crosswell et al. 2014).

Wind, freshwater discharge, and material loading

Hourly average windspeed was gathered from the National Data Buoy Center station CLKN7 at Cape Lookout, NC. Daily average Neuse River discharge was measured by the United States Geological Survey (USGS) at Fort Barnwell (USGS 02,091,814), and divided by 0.69 to account for ungaged downstream inputs (Peierls et al. 2012). Daily Neuse River loads of carbon and nutrient forms were estimated using Weighted Regressions on Time Discharge and Season (WRTDS) (Hirsch and DeCicco 2014; Hirsch et al. 2010) based on daily average discharge and concentrations measured by ModMon (or NC DEQ for total N and total P) at the head of the estuary (Fig. 2). Half-window widths of the tricube weight function for seasonality, time, and discharge were set to default values of 6 months, 7 years, and 2 natural log units, respectively (Hirsch and DeCicco 2014).

Nutrient and carbon analyses

Nutrients measured at ModMon stations included: total dissolved nitrogen (TDN), nitrate plus nitrite (NO3 + NO2), ammonium (NH4+), soluble reactive phosphate (SRP), dissolved organic carbon (DOC), dissolved inorganic carbon (DIC), particulate organic carbon (POC), and particulate nitrogen (PN). Dissolved inorganic nitrogen (DIN) was NO3 + NO2 + NH4+. Dissolved organic nitrogen (DON) was computed by difference as TDN–DIN. Nutrient analyses used 100 mL aliquots filtered through pre-combusted (4 h at 525 °C) 25 mm diameter Whatman GF/F filters into acid-washed and sample-rinsed 150 mL polyethylene bottles that were subsequently frozen at − 20 °C. Filtrates were analyzed for dissolved N forms and SRP with a Lachat/Zellweger Analytics QuickChem 8000 flow injection autoanalyzer using standard protocols (Lachat method numbers 31-107-04-1-C, 31-107-06-1-B, and 31-115-01-3-C, respectively) (Peierls et al. 2012). Particulate organic carbon (POC) and nitrogen (PON) were measured on seston collected on pre-combusted GF/F filters, analyzed by high-temperature combustion using a Costech ECS 4010 analyzer (Peierls and Paerl 2011). DIC and DOC were measured on a Shimadzu Total Organic Carbon Analyzer (TOC-5000A) (Crosswell et al. 2012).

Phytoplankton biomass and community composition

Chlorophyll a (Chl-a) was measured for near-surface and near-bottom samples by filtering 50 mL of NRE water onto GF/F filters. Filters were frozen at −20 °C and subsequently extracted using a tissue grinder in 90% acetone (EPA method 445.0, Arar et al. 1997). Chl-a of extracts was measured using the non-acidification method of Welschmeyer (1994) on a Turner Designs Trilogy fluorometer calibrated with pure Chl-a standards (Turner Designs, Sunnyvale, CA). Taxonomic-group diagnostic photopigments were used to estimate spatio-temporal changes in abundance of algal taxonomic groups. Samples were analyzed by high performance liquid chromatography (HPLC) using a Shimadzu model LC-20AB HPLC equipped with a SIL-20AC autoinjector and a Shimadzu SPD-M20AC photodiode array spectrophotometric detector. Pigment extraction and processing techniques were described by Pinckney et al. (1998, 1999, 2001) and Van Heukelem et al. (1994). Absorption spectra were used to identify photopigments using commercial standards (DHI, Denmark). Major phytoplankton taxa and diagnostic pigments included; chlorophytes or green algae (chlorophyll b, lutein, violaxanthin), cyanobacteria (zeaxanthin, myxoxanthophyll, echinenone, canthaxanthin), diatoms, raphidophytes (fucoxanthin, chlorophyll c, diadinoxanthin, diatoxanthin), dinoflagellates (peridinin), and cryptophytes (alloxanthin) (Jeffrey et al. 1999).

Data analysis

Thirty-five named storms that affected the North Carolina coast from 1996 through 2016 (Fig. 1) were designated as wet or dry and windy or calm, based on recorded wind speeds near the NRE-PS and the impact of their precipitation on freshwater delivery to the NRE (Table 3). Wet storms produced a 7 day mean Neuse River flow, (Fort Barnwell, NC) that was greater than 191 m3 s−1, the 90th percentile of weekly average flows for the storm season (May–October). Windy storms were defined as having a 36 h (12 h before landfall and 24 h after landfall) maximum hourly average wind speed that was greater than 14.1 m s−1, the 90th percentile of 36 h maximum hourly average wind speeds during the storm season. Named storms that were not categorized as wet or windy (8 storms) were considered as part of the baseline conditions upon which to compare storm impacts. Using these criteria, storms were assigned to three groupings: windy but not wet = dry and windy (16 storms), wet but not windy = wet and calm (3 storms), and wet and windy (8 storms) (Table 3). Storm flow periods were defined for each “wet” storm as the time period that encompassed the rise in river flow associated with the storm through the falling limb of the hydrograph, and ended at the first increase in flow once baseflow conditions occurred (see Fig. 3). Storm loads were calculated by summing daily loads over each storm period. To determine the relative importance of storm loads on annual and long-term loading to the NRE, the sum of storm loads was compared directly against total loads on an annual basis and for the long-term (21 years) record. The proportion of storm to total loads, however, minimizes the importance of storm loading to some degree because storm loads also increase the total load. To determine the increase in annual loads directly attributable to storms, the total storm load was divided by a theoretical total load where the storm period loadings were substituted with average non-storm period loading values from the months of May through October.
Table 3

Named storms affecting North Carolina from 1996 through 2016, with storm designation and dates

Storm

Date

Classification

Storm

Date

Classification

Arthur

20 Jun 1996

Dry and windy

Gaston

30 Aug 2004

Bertha

12 Jul 1996

Dry and windy

Ophelia

14 Sep 2005

Dry and windy

Fran

6 Sep 1996

Wet and windy

Ernesto

1 Sep 2006

Wet and windy

Josephine

8 Oct 1996

Wet and windy

Barry

3 Jun 2007

Dry and windy

Danny

24 Jul 1997

Gabrielle

9 Sep 2007

Bonnie

27 Aug 1998

Dry and windy

Christobol

20 Jul 2008

Earl

4 Sep 1998

Dry and windy

Hanna

5 Sep 2008

Dennis

4 Sep 1999

Wet and windy

Earl

3 Sep 2010

Dry and windy

Floyd

16 Sep 1999

Wet and windy

Nicole

29 Sep 2010

Wet and calm

Irene

18 Oct 1999

Wet and windy

Irene

27 Aug 2011

Wet and windy

Gordon

19 Sep 2000

Wet and windy

Beryl

30 May 2012

Dry and windy

Helene

23 Sep 2000

Dry and windy

Andrea

7 Jun 2013

Dry and windy

Allison

13 June 2001

Arthur

4 Jul 2014

Dry and windy

Gustav

10 Sep 2002

Dry and windy

Ana

7 May 2015

Wet and calm

Isabel

18 Sep 2003

Dry and windy

Joaquin

3 Oct 2015

Wet and windy

Alex

3 Aug 2004

Hermine

2 Sep 2016

Dry and windy

Bonnie

13 Aug 2004

Matthew

8 Oct 2016

Wet and windy

Charley

14 Aug 2004

Wet and calm

   

“–” The storms were not classified as wet or windy

Fig. 3

Impacts of “wet storms” on river flow and total N (TN) loading to the Neuse River Estuary. a Daily Neuse River discharge. b Daily TN load estimated by Weighted Regressions on Time Discharge and Season. Filled red circles indicate the landfall dates of “wet storms”. Time periods defined as storm flows are shaded red. c Expanded view of TN load resulting from Hurricane Irene. d Relationship between observed TN load based on monthly TN measurements and WRTDS estimates of load for the same set of dates. (Color figure online)

Impacts of tropical cyclones on estuarine carbon, nutrient and algal biomass related parameters were determined for each storm type for the NRE and southern PS. For each parameter and for both the NRE and PS, volume weighted averages were calculated to provide a single metric for examining the response of each water body. For the NRE, volume weighted concentrations were calculated according to Peierls et al. (2012) based on segment volumes assigned to each NRE station. Due to the even spatial distribution and weaker spatial gradients in PS, the volume weighted average was assumed equal to the arithmetic mean of observations from the nine PS sampling stations (Fig. 2). Pigment data were log transformed prior to all analyses. Volume weighted average parameter values from sampling cruises within two weeks following a storm were averaged and the resulting averages were aggregated by storm type. Kruskal–Wallis tests were used to determine significance of differences in each parameter between four storm types: (1) baseline, non-event, (2) dry and windy, (3) wet and calm, and (4) wet and windy storms. Tukey–Kramer post hoc tests on ranks within each storm type grouping were used to determine which groups were different when Kruskall Wallis tests indicated significant differences of medians between groups (α = 0.05). To evaluate the evolution of storm impacts over time, this process was repeated for samples collected 3–4 weeks after the storms (Table 4).
Table 4

Influence of “wet” storms on long-term (1996–2016) material loads to the Neuse River Estuary

Parameter

Percent of long term load during storm flows

Percent increase over baseline due to storms

Water

13.9

15.5

TN

11.6

12.6

DIN

7.2

7.5

DON

16.0

18.3

PN

16.0

18.2

TP

21.5

25.7

SRP

26.0

32.8

DOC

21.2

25.6

POC

17.0

19.6

DIC

14.1

15.7

Following these estuary-wide analyses, phytoplankton biomass as Chl-a was examined with a greater degree of spatial resolution in the NRE, where averaging across the estuary may obscure significant differences in downstream Chl-a spatial gradients due to storm impacts. Volume weighted Chl-a concentrations were calculated for upper (stations 0–30), middle (stations 50–120) and lower estuary (stations 140–180) segments, and Chl-a responses to the different storm types were determined as for the whole estuary.

Results and discussion

Freshwater discharge

Major storms, including Hurricanes Floyd in September 1999 and Matthew in October 2016, resulted in peak daily average discharges more than ten and twenty times higher than the long-term mean (~ 110 m3 s−1) and median (~ 70 m3 s−1), respectively (Fig. 3). These significant storm pulse events are superimposed on a highly dynamic flow regime regulated by seasonal cycles of evapotranspiration, years of very high non-tropical storm related precipitation (e.g. 1998 and 2003), and multi-year droughts (e.g. 2007–2012). Consequently, storm impacts on annual discharge depend on the intensity of storm discharge, relative to background annual flow. For example, 1999 was an otherwise dry year but the cumulative sequential impacts of Hurricanes Dennis, Floyd, and Irene more than doubled the annual discharge, making annual flows in 1999 the highest during the 21 year period of study (Fig. 4A). Although the storm pulse from Hurricane Matthew was a significant event, 2016 was a year marked by high rainfall to the watershed, and Matthew only constituted about 10% of the annual total flow. Hurricane Irene (2011) produced about half the discharge of Matthew but occurred during the prolonged drought when annual flows were the lowest during the 21 year study period. As a result, discharge following Irene was twice the fraction of the annual total compared to Matthew.
Fig. 4

Total annual and storm flow loading of nutrients and carbon forms (left panels). Flow versus concentration relationships during non-storm in black symbols and storm flows in red symbols (left panels). Annual water loads are in units km−3 per year. N, P, and C loads are in units 106 kg of N, C, or P per year. Concentrations at Streets Ferry Bridge are in units log10 (µmol L−1). (Color figure online)

While it is clear that tropical cyclones can contribute significant pulses of freshwater to the NRE-PS system, no significant effects of storms on volume weighted average salinities were determined for either the NRE or PS (Fig. 5A, F). Although the median salinities did decrease for wet and calm and wet and windy storms in both first 2 weeks and following weeks three and four after the storm (Fig. 5A, F), the high degree of variability resulted in these difference being statistically insignificant. River flow exerts a strong influence on the salinity of the NRE-PS. However, the salinity budget of the system is also strongly impacted by changes in exchange with shelf waters driven by differences in water level between PS and nearby Raleigh Bay (Jia and Li 2012), the occasional formation of new inlets (Paerl et al. 2009), and wind-forced salinity intrusion (Jia and Li 2012).
Fig. 5

Box plots of volume weighted nutrient-related parameters within the NRE & PS. during non-storm conditions and two time spans following the passage of dry and windy, wet and calm, and wet and windy tropical cyclones. Boxes represent the interquartile range. Whiskers represent 1.5 times the interquartile range, and red, plus symbols represent outliers. Different letters above boxes identify statistically distinguishable rank values between groups. (Color figure online)

Continuous monitoring of salinity in PS by FerryMon provided an excellent example of how creation of a new inlet can affect exchange and salinity regimes within PS. Prior to Hurricane Isabel in September 2003, the eastern side of PS was growing increasingly fresh compared to the western basin due to higher than usual riverine discharge to the Roanoke and Chowan rivers that eventually drain into eastern PS (Paerl et al. 2009). The inlet created by Hurricane Isabel enhanced exchange of the eastern basin with Atlantic shelf waters causing a relative increase in salinity in the eastern basin compared to the western basin of PS. This trend of increasing relative salinity in the eastern basin continued until later in the fall of 2003 when the new inlet was filled by the N.C. Dept. of Transportation and the east–west gradient in salinity began to once again decrease (Fig. 6).
Fig. 6

Difference in weekly average FerryMon salinity measurements between the east and west basins of Pamlico Sound during the period prior to the creation of a new inlet by Hurricane Isabel through its eventual closure in late fall 2003.

Figure reproduced with permission from Paerl et al. (2009)

Nutrient (N and P) and carbon loadings

Variability in river flow spanned nearly three orders of magnitude while concentrations of nutrient and C forms generally varied only by 1–1.5 orders of magnitude (Fig. 4, K-BB). Consequently, flow rather than concentration was the strongest driver of loading of nutrients and C forms to the estuary. High river flows following “wet” storms consistently led to periods of intense C and nutrient loading to the estuary, and often contributed significantly to annual total loading. For example, in 1999 loading of TP and SRP during the back to back storm flow periods following Dennis, Floyd, and Irene constituted greater than 90% of the annual load (Fig. 4F, G). These same storm flows, however, only constituted about 50% of the annual DIN loading (Fig. 4C). Similarly, SRP loading during Hurricane Matthew constituted about half of the annual load while DIN loading was less than 20%.

These differences in the relative importance of storm loadings to the annual total can be explained by examining the relationship between flow and concentrations of each parameter during storm and non-storm flows. Under non-storm flow condtions, both TP and SRP exhibit weak negative relationships with increasing flow (Fig. 4X, Y), which is indicative of a dominant contribution by point sources in the watershed. In contrast, under storm flows, concentrations of P forms increase as flow increases, which magnifies the impact of elevated water loading to the estuary on P delivery. The mechanism behind the enhancement of P concentration under storm flow conditions may involve sediment P release from inundated wetlands and agricultural fields under the hypoxic river water conditions that often occur following the passage of tropical cyclones. DOC concentrations exhibited a similar positive slope with flow under both storm and non-storm flow conditions (Fig. 4Z). However, under storm flows, DOC concentrations are higher than non-storm flows (Fig. 4Z), with a similar effect of magnifying the relative importance of storm flow loading of terrigenous DOC to the estuary. It seems very likely that high P and DOC loadings following storms are linked. Tropical cyclones occur during summer and fall when waters are still warm and favorable for bacterial growth (Peierls et al. 2012). High loading of labile terrigenous DOC during storm flows likely fuels water column oxygen consumption, creating hypoxic conditions that promotes redox sensitive P release from surficial sediments (Paerl et al. 1998).

Both TN and DIN exhibited a non-monotonic relationship with flow with concentrations increasing from the lowest flows up to about 30 m3 s−1 and then subsequently declining weakly with increasing flows (Fig. 4T, U). This pattern has been previously recognized and ascribed to the initial importance of overland, non-point source delivery of N as flows begin to increase followed by a dilution effect of increasing flows (Alemeddine et al. 2011). Within the range of storm-flows, the relationship between flow and concentrations of TN and DIN is negative (Fig. 4T, U), and therefore mutes the impact of storm flows on loading relative to P forms and DOC. DON and PON showed no clear relation with flow (Fig. 4V, W), and therefore the fraction of storm loading to total loading for these N forms closely mirrored the fraction storm water loading to total water loading.

The strong correlation of N loadings with freshwater discharge (Fig. 4K–N) confirmed the importance of non-point sources, including agricultural and urban storm-water runoff, as major sources of N in the NRE-PS basin. These results were consistent with earlier estimates of annual point vs. non-point sources of N in the NRE-PS basin that showed ~ 80% of total N loading is of non-point origin, much of it organic N (NC DENR 2001; Lebo et al. 2012; Osburn et al. 2016a). Moreover, results presented here illustrate the importance of event-scale loadings, with “wet” storms contributing significant fractions of annual totals for N loading to the NRE and PS. Years with large wet storm events corresponded to increased annual loadings compared to years with no events or that experienced “dry” storms. Fall 1999 (Floyd and Dennis), fall 2006 (Ernesto) and fall 2016 (Matthew) are examples of years whose N loadings were dominated by storm events (Fig. 4B). Several other years had rainy seasons in an absence of major storms, exemplified by summer 2003. In contrast, seasons and years that featured dry and windy storms and droughts showed lowest seasonal and annual N loadings to the NRE and PS. These data documented that the frequency and magnitude of “wet” events strongly affected N loadings to the NRE-PS.

Although storms do increase N loading to the NRE-PS, of the forms of N investigated (PON, DON, and DIN), only DIN showed significantly elevated concentrations compared to the baseline conditions, and only in the NRE during the first 2 weeks following wet and windy storms (Fig. 5B–D).

While phytoplankton in the NRE-PS were N-limited from spring to fall, P was a “co-limiting” factor occasionally (Rudek et al. 1991; Paerl et al. 1995, 2014). Co-limitation by N and P was observed during periods of high freshwater discharge accompanied by exceptionally high N loadings. Compared to baseline, non-storm conditions, volume weighted phosphate concentrations were not significantly different in either the NRE or PS following any storm type (Fig. 5E, J).

Organic matter (OM)

Storms appear to affect pools of DOC and POC differently in NRE-PS. Pulses of terrigenous DOC occurred throughout NRE-PS following wet and calm and wet and windy events, and the spatial extent of these DOC pulses often reached the mid-estuary, depending on the magnitude of freshwater discharge (Figs. 7, 8). Exceptional DOC pulses affected PS, particularly following large, wet tropical storm events in 2006 and 2016 (Figs. 7, 8). This result was consistent among wet and windy storms and resulted in median volume-weighted DOC values that were significantly higher than non-storm conditions in NRE (Fig. 7). DOC concentrations in NRE were approximately 200 and 300 µmol L−1 higher during the first three to 4 weeks following wet and windy storms (Fig. 7). In PS, the same trend was observed but DOC concentrations were only significantly higher than baseline conditions (~ 100 µmol L−1 higher) during the three to 4 week period following wet and windy storms (Fig. 7). Median values of volume-weighted POC concentrations in NRE and PS were remarkably consistent at ~ 150 and 120 µmol L−1, respectively.
Fig. 7

Box plots of volume weighted carbon-related parameters within the NRE & PS. Figure configuration follows Fig. 5

Fig. 8

Carbon loading and air–water CO2 gradients (∆pCO2) in the NRE, where positive ∆pCO2 indicates CO2 exchange from the water to the atmosphere. Hurricane Joaquin (2015) did not make landfall (also shown in Fig. 1), but indirectly impacted NC as part of a large ‘multiple-low’ storm complex

While DOC concentrations are often used to track the influence of terrigenous OM in estuarine systems, C:N serves as an indicator of the origin and quality of DOM and POM, with lower C:N (4–10) attributed to phytoplankton biomass and the higher C:N (> 20) characteristic of terrigenous material (Bianchi and Bauer 2011). In NRE, DOM is largely terrigenous with some modulation of its quality by planktonic processes during low discharge (Osburn et al. 2012; Dixon et al. 2014). For example, just prior to Hurricane Irene, NRE experienced low-flow conditions and DOM quality reflected high Chl-a (Osburn et al. 2012). After passage of that storm, both DOM and POM pools reflected terrigenous sources, including increases in terrigenous DON. Another major source of autochthonous DOM originated in sediments, as evidenced by changes of DOM quality under moderate discharge and wind-driven mixing (Dixon et al. 2014). C:N ratios in DOM for NRE ranged from 12 to 30, indicating the predominantly terrigenous influence.

In NRE, increases of particulate C:N (p = 0.09) were observed during wet and windy storms four weeks following Hurricane Irene (Fig. 7B), suggesting that storms with high freshwater delivery increased the ratios of terrigenous to phytoplankton-derived organic matter. Particulate C:N ratios in PS primarily reflected phytoplankton OM (C:N < 10) as the influence of terrigenous OM was negligible. Tropical cyclones and precipitation extremes are known to produce significant changes in terrigenous OM loading to NRE and PS (Brown et al. 2014; Osburn et al. 2012; Paerl et al. 1998, 2001a). The significant increase of DOC for wet and windy storms (Fig. 7C, I) was consistent with an impact of these storms on DOC loading (Fig. 4H, Q). Large pulses of terrigenous OM into NRE were observed for days to months following storm passages. DOC pulses following high freshwater discharge affected microbial production and respiration, exacerbating eutrophication and habitat degradation, i.e., hypoxic/anoxic zones; fish kills; decreased water clarity and quality (Brown et al. 2014; Paerl et al. 1998; Peierls and Paerl 2010, 2011).

Understanding how the impact of these DOC pulses on the eutrophic NRE-PS is important in the context of their roles in accelerating eutrophication, given a predicted increase of major tropical cyclones (Category 3 or higher) in the western Atlantic in future climate scenarios (Bender et al. 2010). Our findings suggest that storm-driven flushing of terrigenous OM to NRE-PS, confirmed by DOC and particulate C:N signals (Fig. 7), results in a shift to terrigenous-dominated OM from phytoplankton-dominated OM, stretching from riverine sources to the marine end-member (Dixon et al. 2014; Osburn et al. 2016b). Extreme events, particularly events that resulted in historic floods (i.e., Hurricanes Floyd, Matthew) resulted in extreme pulses of terrigenous DOC that extended through NRE and into PS, indicating these extreme discharge events follow the pulse-shunt concept whereby NRE acts as a pipeline for terrigenous DOC that is exported to PS largely unaltered (Osburn et al. 2012; Paerl et al. 2001a, b, 2006a, b, c; Osburn et al. 2016b; Raymond et al. 2016).

CO2 fluxes

Average estuarine pCO2 in NRE was significantly elevated above baseline conditions during the two week period following wet and windy storms (Fig. 7E). This increase in pCO2 was also reflected in significant decreases in pH (Fig. 7F). Although individual storms did produce elevated pCO2 in PS (e.g. Hurricane Irene: Crosswell et al. 2014), there were no consistent, statistically significant differences in PS pCO2 or pH following any storm type compared to non-storm conditions (Fig. 7K, L). Due to the high bicarbonate concentration of seawater and its dominance of the DIC pool at estuarine pH, changes of DIC by storm type resembled those for salinity (compare Figs. 5A, F and 7D, J) and no significant changes from baseline, non-storm conditions were observed.

Extreme winds dramatically increase gas exchange via increased turbulence and physical processes associated with wave breaking. Crosswell et al. (2014) reported increases of pCO2 and piston velocity following Hurricane Irene led to air–water CO2 efflux from NRE and PS of 57–170 and 5–16 mmol C m−2 h−1, respectively, for the 42-h period when NRE-PS experienced hurricane-force winds. By contrast, air–water CO2 efflux over the 17 days following Irene were only 4.2–6.0 and 0.2–0.3 mmol C m−2 h−1 in NRE and PS, respectively. A significant increase in calculated pCO2 after wet and windy storms (Fig. 7E) was consistent with pCO2 directly measured during the case study of Hurricane Irene. Median calculated pCO2 following wet and windy storms was at least three times higher than wet and calm storms, dry and windy storms, or non-storm conditions.

Variation in storm intensity and frequency

Over the past two decades, the largest effects on riverine C-loadings (DIC, DOC, and POC) followed intense (1996, 2016) and back-to-back (1999) hurricanes (Fig. 8). Riverine C-loading following these types of storms was 10× greater than occurred immediately after Hurricane Irene (2011) or Hurricane Joaquin (2015) (Fig. 8). Storm intensity is typically characterized by the Saffir-Simpson scale that is based on wind speed. However, other metrics of storm intensity, such as rainfall, location, and duration may better capture the extent to which C is mobilized from the watershed and transported through rivers to estuaries (Fig. 8). The largest riverine C-loadings were in response to Hurricane Dennis (1999), the longest-duration storm to impact the state of North Carolina during the study period. Hurricane Fran (1996) was a more intense storm in terms of rainfall and wind speed, but Fran was a relatively fast-moving system that followed a northern track, and riverine C-loading was slightly less than after Dennis. A similar trend was observed among the less intense storms. The northern track of Hurricanes Dennis (1999) and Irene (2011) across the NC coastline led to relatively low riverine C loading. By contrast, the curved track of Hurricane Matthew along the South Atlantic Bight extended the time of impact on Piedmont watersheds and produced extreme flooding and C-loadings from rivers to coastal waters (Fig. 8).

Fate of terrigenous C in NRE-PS

The role of rivers as passive conduits or ‘pipelines’ that deliver terrigenous DOC and POC to estuaries and coastal waters has long been recognized as a major component of global biogeochemical cycles (Schlesinger and Melack 1981; Bianchi 2007; Bianchi and Bauer 2011; Bauer et al. 2013). However, the concept of streams and rivers as active, rather than passive “pipelines” for OC (Cole et al. 2007) has received support in the past few decades (Battin et al. 2009; Tranvik et al. 2009). This has led to a recognition of inland waters as globally significant sources of CO2 (Butman and Raymond 2011; Raymond et al. 2013) and CH4 (Cole et al. 2007, Bastviken et al. 2011) to the atmosphere, driven by the re-mineralization of terrigenous OC (Cole and Caraco 2001; Regnier et al. 2013). Hydrologic events significantly alter both the timing and quality of OC flux from streams to receiving bodies such as estuaries (Hilton et al. 2008; Osburn et al. 2012; Bauer et al. 2013; Bianchi et al. 2013; Osburn et al. 2016b; Almeida et al. 2017). DOC and POC concentrations generally increase with freshwater discharge as a larger fraction of runoff occurs in overland flow rather than infiltration and percolation through soils (Schlesinger 2005). Hence, freshwater discharge affects the quality and quantity of terrigenous OC delivered by rivers, and the residence time of this material within receiving waters. The recently described ‘pulse-shunt’ concept (Raymond et al. 2016) is an extension of earlier frameworks describing OC dynamics in drainage networks, including ‘active pipe’ (Cole et al. 2007) and nutrient spiraling (Ensign and Doyle 2006; Hall et al. 2013). This concept argues that hydrologic events alter the delivery of terrigenous OC through drainage networks by mobilizing a ‘pulse’ of labile material, which is rapidly ‘shunted’ downstream.

The fate of this labile OC depends on the capacity of estuarine waters to buffer changes in pCO2 generated during the respiration and photochemical oxidation of organic matter. If waters are poorly buffered, more of the remineralized OC will be lost to the atmosphere as CO2, but if the buffering capacity is high, a greater fraction may be exported to the coastal ocean as HCO3 and CO32−. Because this buffering capacity varies with DIC and alkalinity (Egleston et al. 2010; Hu and Cai 2013), delivery of these materials to the estuary will shape the response of the estuary to OC loading.

The data presented here for NRE-PS support the pulse-shunt concept and extend its applicability to estuarine systems. Storms mobilize large pulses of terrigenous OC and deliver this C to the estuary (Osburn et al. 2012). The eventual fate of this terrigenous OC is determined by the relative importance of three processes: burial, flushing to the coastal ocean, and degradation to CO2, resulting in exchange with the atmosphere. We were able to constrain one of these loss terms, air–water CO2 exchange in this study. ‘Dry’ events, such as Hurricane Irene (2011), were associated with enhanced physical mixing that increased CO2 loss to the atmosphere relative to terrigenous OC flux and storage in PS (Crosswell et al. 2014). ‘Wet’ storms, exemplified by Joaquin and Matthew, were associated with relatively low CO2 fluxes during storm passage. Instead, the main impact of wet storms on CO2 fluxes occurred after the storm due to enhanced net heterotrophy, and possibly photochemical oxidation, in response to large terrigenous DOC loadings generated by extreme rainfall. Figure 9 illustrates this effect as increased DOC loadings associated with the large freshwater pulse from Hurricane Matthew, concurrent flushing of phytoplankton biomass from NRE, and dilution of POC associated with phytoplankton biomass in NRE-PS that saturated these ecosystems with terrigenous DOC. Within 1–2 months, freshwater discharge decreased, allowing phytoplankton-driven production of POC to increase in importance as the large terrigenous DOC in freshwater discharge subsides (Figs. 8, 9).
Fig. 9

An example of the pulse-shunt concept as shown by the organic matter-enriched freshwater pulse that discharged to the NRE-PS after very wet Hurricane Matthew struck coastal North Carolina in the fall of 2016. Note the rapid enrichment of DOC associated with the pulse. This caused a rapid shift from an authochthonous-driven system, as shown by phytoplankton-based Chl-a, to a allochthonous-dominated condition in response to Matthew’s floodwaters, which flushed much of the phytoplankton biomass out of the estuary. Two months after this event, the system returned to authochthonous-dominated conditions, mediated by nutrient-driven phytoplankton production and reduced flushing rates

Responses of phytoplankton biomass (Chl-a) to storm events

Within NRE and PS, volume-weighted phytoplankton pigment concentrations following the different storm types showed no consistent, statistically significant differences from non-storm conditions (Fig. 10). However, this finding did not indicate a lack of response by the phytoplankton community to storm events, rather it suggested there was not a predictable estuary-wide response. At finer degrees of spatial and temporal aggregation, significant changes in Chl-a due to storm type were observed in NRE. In the upper and mid-estuary, wet and windy storms led to decreased Chl-a, with impacts persisting up to four weeks following storm passage (Fig. 11A, B). In the upper estuary, even small increases of freshwater discharge promoted rapid flushing and prevented an accumulation of phytoplankton biomass (Peierls et al. 2012; Paerl et al. 2014). Though less sensitive to flushing than the upper estuary, the middle NRE was susceptible to rapid flushing by major wet and windy storms. Examples included Hurricanes Bertha and Fran in 1996 and Hurricane Floyd in 1999 (Paerl et al. 2001a, b, 2006a, b) that caused the middle NRE to become a freshwater, highly turbid, lotic system. The lack of a significant flushing effect from wet and calm storms is likely due to sample size as only three storms were designated as wet and calm. Although not statistically significant, the lower NRE responded to wet and wet and windy storms with increased phytoplankton biomass for up to 4 weeks following the storms (Fig. 11C). Under baseline conditions, riverine nutrients are assimilated by phytoplankton far upstream, within the upper and middle regions of NRE, leading to chronic nutrient-limitation in the lower NRE-PS (Pinckney et al. 1998; Piehler et al. 2004; Peierls et al. 2012; Paerl et al. 2014). Increased phytoplankton biomass in the lower NRE following wet and wet and windy storms was consistent with an alleviation of chronic nutrient-limitation by pulsed nutrient loadings to NRE-PS and sufficient residence time to allow significant biomass accumulation.
Fig. 10

Box plots of volume weighted log2 tranformed phytoplankton pigment concentrations within the NRE & PS. Figure configuration follows Fig. 6

Fig. 11

Box plots of volume weighted log2 tranformed chlorophyll a concentrations within the upper, middle, and lower reaches of the NRE. Figure configuration follows Fig. 6

Previous studies suggested dry storms stimulate phytoplankton biomass, presumably by promoting sediment re-suspension and release of pore-water nutrients to the water column (Wetz and Paerl 2008). However, results presented here showed that windy storms with little precipitation had no detectable impact on phytoplankton biomass in nutrient-limited regions of NRE-PS. Freshwater discharge associated with tropical cyclones was the strongest modulator of phytoplankton biomass along NRE-PS, with impacts explained by dual influences of enhanced nutrient loading and enhanced flushing. Previous studies showed low to moderate rainfall and freshwater discharge stimulated phytoplankton biomass in the upper to middle regions of NRE where high Chl-a was observed, while high rainfall and freshwater discharge either stimulated biomass in downstream locations, or prevented biomass from accumulating in NRE due to excessive flushing (Peierls et al. 2012). In the latter case, such as we observed following the very wet Hurricane Floyd, significant increases of biomass occurred in the long-residence time PS (Paerl et al. 2001a, b, 2006b). Moderately wet storms, such as Ernesto (2006), led to significant biomass peaks at mid-NRE locations, including blooms of a toxic dinoflagellate (Karlodinium venificum) that were responsible for several fish kills (Hall et al. 2008).

Two recent examples of variable Chl-a responses to interactive effects of residence time and nutrient loadings occurred following Hurricane Joaquin (September 23, 2015) and Hurricane Matthew (October 18, 2016) (Fig. 12). Effects on freshwater discharge were both chronic (i.e., sustained, above-normal precipitation leading to a prolonged, weeks-to-months increase of discharge) and acute (i.e., precipitation not sustained, resulting in a brief, days, increase of discharge) following these storms. Chronic discharge followed multiple precipitation events that occurred concurrently within a time period leading to saturated soil conditions, resulting in increased overland flow, a higher water table, and stream base flow conditions. High chronic discharges occurred following Hurricane Joaquin from October 2015 through early March 2016, with a peak of 566.4 m3 s−1 on February 2, 2016. Increased Chl-a developed in surface and bottom waters of PS following Joaquin and persisted until April 2016, returning to low baseline conditions in summer 2016, until Hurricane Matthew made landfall on October 8, 2016 (Fig. 12). Such high acute discharges also occurred immediately following Matthew from October 8, 2016 to October 26, 2016, with a peak discharge of 1393.2 m3 s−1 on October 15, 2016.
Fig. 12

Chlorophyll a and high performance liquid chromatography determinations of Chlorophyll a and diagnostic pigments for the dominant phytoplankton groups plotted for the NRE-PS as well as mean freshwater discharge and major storm events from 1994 to 2016. Blooms over 40 μg L−1 are seen in red; J June, D December, Perid peridinin, fuco fucoxanthin, allo alloxanthin, zea zeaxanthin, chl b chlorophyll b, and chl a = chlorophyll a

A pair of surface Chl-a maxima occurred in PS before and after Hurricane Matthew. The first maximum closely followed increased rainfall from several tropical storms prior to Matthew, and the second maximum occurred following increased freshwater discharge from Matthew (Fig. 9). There was a slight increase of surface Chl-a for stations closest to the mouth of NRE, but no blooms affected Chl-a in bottom samples following Matthew. Most likely, this was due to salinity stratification that isolated surface from bottom waters as the freshet from Matthew moved downstream over denser, saltier bottom water in the lower NRE. This effect was distinct from the increase of Chl-a throughout the water column that we observed in PS as a response to chronic high discharges following Hurricane Joaquin. Persistent high freshwater discharge in the 2–3 week aftermath of high rainfall and flooding from Hurricane Joaquin led to complete flushing and vertical mixing of NRE, triggering high Chl-a throughout the well-mixed PS. In contrast, acute and extreme freshwater discharge associated with Hurricane Matthew in October 2016 produced a sharp increase of N- and P-loadings, followed by an abrupt return to low levels as the storm subsided. Thus, the distinct hydrologic impacts of these large storms included significantly different responses of phytoplankton biomass (Figs. 9, 10, 11, 12).

Flushing rates and nutrient loadings interacted to determine responses of phytoplankton biomass, production, and composition to storm events in NRE. Extreme freshwater discharge led to high nutrient loadings, but concurrently led to rapid flushing that exceeded phytoplankton growth rates (Peierls et al. 2012; Paerl et al. 2014) (Figs. 9, 11, 12). We suggest there is a tradeoff between nutrient loadings and flushing rates wherein moderately wet storms stimulate higher biomass in NRE than extremely wet storms. Two of the “wettest” storms on record confirmed this suggestion. Hurricanes Floyd and Matthew had extremely high nutrient loadings but did not stimulate blooms as rapid flushing precluded biomass accumulation in NRE. Distinct from storm effects on NRE, phytoplankton biomass was stimulated in PS where the residence time is much longer.

Storm effects on phytoplankton composition

Extensive data on HPLC diagnostic photopigments showed that nutrient loading and flushing interacted to influence phytoplankton community composition. During high discharge (e.g. post Ernesto in 2006, post Joaquin in 2015), phytoplankton were flushed down the estuary; how far depended upon the magnitude and duration of the event (Figs. 11, 12). Blooms often followed farther downstream. During low discharge periods (e.g. summers of 2002, 2007), higher pigment concentrations could be found farther upstream and even near SFB (0 km) with occasional blooms (Fig. 12).

At baseline conditions, total pigment concentrations were often highest ~ 20–50 km downstream in NRE where the river widens and residence time increases (Fig. 12a–c, e–f; Paerl et al. 2014). However, there were exceptions, as discussed below for individual taxonomic groups. Alloxanthin (cryptophytes) and Chl-b (chlorophytes) were responsive to storm events, displaying significant spatial heterogeneity and temporal lags (Fig. 12c, e, respectively). These pigment markers for cryptophytes and chlorophytes responded rapidly to storm events as inherent growth rates of these taxa are higher than other groups (Paerl et al. 2013, 2014), and both are favored by relatively low salinity and high nutrient conditions associated with storm passage.

Peridinin (dinoflagellates) increased during typical winter/spring blooms in NRE since 2011 and long-term data showed high concentrations persisted for weeks to months throughout the estuary (Fig. 12a). Growth rates of dinoflagellates are typically low, limiting responses to hydrologic perturbations and making members of this taxonomic group susceptible to flushing. Nonetheless, we observed storm effects on dinoflagellate biomass including a mixed bloom of perdinin- and fucoxanthin-containing dinoflagellates in 2006 following TS Ernesto (Fig. 12; Hall et al. 2008). Moderately wet storms supported increased peridinin as nutrient loadings and freshwater discharge provided suitable conditions for biomass accumulation with minimal flushing. These storm events also enhanced vertical stratification in NRE, favoring motile taxonomic groups such as dinoflagellates that use vertical migration to access light and nutrient resources (Hall and Paerl 2011).

Of the diagnostic pigments used to identify major taxa, zeaxanthin (cyanobacteria) was the only indicator of significant differences between storm types. Compared to wet and calm storms, cyanobacteria biomass was significantly higher following dry and windy storms in NRE. This stemmed from a slight, though statistically insignificant, increase above baseline conditions following dry and windy storms, and a stronger decrease following wet and calm storms (Fig. 10). Low growth rates characteristic of this taxonomic group and their susceptibility to flushing may explain this observation. Additionally, reduced N forms (e.g. ammonium and DON) that would be prevalent during resuspension events (Giffin and Corbett 2003) favor the growth of picoplanktonic cyanobacteria (Glibert et al. 2004) that are predominant members of the cyanobacteria assemblage in NRE (Gaulke et al. 2010).

Seasonal patterns of cyanobacteria growth and biomass were related to the annual temperature cycle, with biomass maxima typically occurring during warm summer and fall periods (Pinckney et al. 1998, 1999; Gaulke et al. 2010; Paerl et al. 2013, Fig. 12d). Over multiple years, cyanobacteria abundance tracked surface water temperatures, consistent with observations that cyanobacteria “like it hot” (Paerl and Huisman 2008). Furthermore, fucoxanthin (diatom) abundance was relatively unaffected by storm events, although a slight time lag was observed at some locations (Fig. 12b). Growth rates of diatoms are relatively high, but this taxonomic group generally comprised less than 20% of phytoplankton biomass in NRE-PS (Pinckney et al. 1998, 1999). This may explain the “insensitivity” of diatom abundance to storm events.

Major storm events strongly affected biogeochemical cycling and phytoplankton production, with storms exhibiting interactive effects of nutrient loadings and flushing. Some phytoplankton taxa, notably cyanobacteria, are highly seasonal with spatial–temporal distributions primarily associated with the annual temperature cycle and secondarily by flushing. From a predictive perspective, a period of elevated storm activity favors fast-growing chlorophytes and cryptophytes, especially if the ecosystem “freshens up” following high freshwater discharge from storm events. Dinoflagellates, such as Karlodinium venificum, are also susceptible to removal by flushing, but intermediate conditions with modest flushing accompanied by vertical density stratification, permit this taxonomic group to form blooms by virtue of motility that confers an advantage in responding to nutrient pulses (Hall et al. 2008).

Conclusions

Estuarine and coastal systems such as NRE-PS are highly susceptible to hydrologic perturbations from storm events that deliver pulsed inputs of nutrients and OM. In this overview, we report the scales, magnitudes, and extent of biogeochemical and biotic responses for a variety of storm types spanning 2–3 decades. A conceptual diagram (Fig. 13) depicts important ramifications of storm impacts, focusing on time-scales and magnitudes. Storms with high rainfall mobilized terrestrial material from inland watersheds and river channels, whereas storms with extreme winds mobilized previously stored terrigenous nutrients and OM. Perturbations of NRE-PS by storm winds or high freshwater flow were expressed at time scales from hours to weeks. Enhanced DOC, POC, and nutrient fluxes following storms exceeded annual fluxes observed in baseline years lacking major storms. Sequences of intense, frequent storms, exemplified by years with multiple tropical cyclones in a single season, exerted the largest impact on C and nutrient cycling. Each storm showed a unique signature based on physical attributes, such as duration, wind speed, precipitation, and geographic track. Widely used indices of storm intensity, however, such as the Saffir-Simpson Hurricane Wind Scale (SSHWS), were insufficient to predict storm effects on C and nutrient fluxes.
Fig. 13

Conceptual diagram, adapted from Crosswell (2013), showing the idealized response of the NRE to Hurricane Irene, a model ‘windy’ storm (right)*. The track and intensity of Hurricane Irene as it impacted the study area is also shown (left). *Salinity profiles were constructed from direct measurements before and after the storm. Data shown are 15 Aug. 2011 (a), hypothetical salinity values under well-mixed conditions (b), and 30 Aug. 2011 (c)

The upper region of NRE-PS was dominated by terrigenous DOC loading following storm events that delivered high rainfall to the watersheds (Fig. 13). Steep increases of DOC were recorded following storm events of all types. These responses were not limited to tropical cyclones, but occurred following major precipitation events that were sufficient to flush terrigenous material with high DOC from the watersheds (e.g., Dixon et al. 2014). The mid- and lower regions of NRE-PS also showed effects of increased terrigenous DOC, but effects were dampened compared to the upper region. PS experienced a lagged increase of DOC reflecting movement of the freshwater pulse down NRE to lower regions and PS. This lag suggests NRE acts as a “pipeline” during extreme storm events of sufficient strength to flush terrigenous material through NRE and into PS. Residence times are longer in PS, allowing water rich in nutrients and OM to fuel primary productivity. This finding is reflected in Chl-a data that reveal phytoplankton blooms following storm events. While the upper estuary was dominated by terrestrial flushing following storm events, the lower estuary and PS were dominated by in situ processes fueled by deliveries of nutrients and OM (Fig. 13).

Storm events can influence phytoplankton community composition through the interaction of nutrient loading and flushing rates. While very high freshwater discharge is responsible for maximum nutrient loads, they are also responsible for maximum flushing rates that can exceed phytoplankton growth rates. Some taxa (cyanobacteria) are most sensitive to seasonal temperature changes and their spatial–temporal biomass largely reflect the annual temperature cycle, with a secondary influence of flushing. From a predictive perspective, a period of elevated storm activity favors fast-growing chlorophytes and cryptophytes, especially if the system “freshens up” following major discharge events. Dinoflagellates are strongly affected by flushing, but under moderate and intermediate flushing conditions wherein nutrient enrichment from runoff and vertical stratification coincide, they can take advantage of advantageous conditions and form harmful blooms (Valdes-Weaver et al. 2006; Paerl et al. 2013).

The fate and impacts of terrigenous OM on estuarine and coastal ecosystems such as NRE-PS following storm events remains an open question. Bacterial production is responsive to loads of “fresh” terrigenous OM stored in watersheds and exported to coastal waters where increased residence time facilitates degradation of OC to CO2 (Peierls and Paerl 2010). Photochemical oxidation of DOC is also an important process in long residence times. Nutrients that transit the “estuarine filter” of NRE should fuel secondary production in PS, but insufficient observations are available to confirm this conclusion. We suggest that storm events can “prime” coastal waters for intense biogeochemical cycling as a result of nutrient and OM pulses to receiving waters such as PS.

Lastly, we report storm effects on freshwater discharge, nutrient loadings, and C dynamics with commensurate influences on spatial–temporal distributions of phytoplankton biomass and composition in NRE-PS, responses that should be considered in management strategies for coastal ecosystems as we encounter stormier, increasingly extreme climate conditions.

Notes

Acknowledgements

We appreciate the assistance of J. Braddy, A. Joyner, H. Walker, B. Abare, B. Peierls and all students and technicians that participated in the field and laboratory work supporting this publication. This research was funded by NSF Projects DEB 1119704, DEB 1240851, OCE 0825466, OCE 0812913, OCE 1705972, OCE 1706009, and CBET 0932632, the Defense Coastal/Estuarine Research Program (DCERP) Project Number: RC-2245, The North Carolina Dept. of Environmental Quality (ModMon Program), the Lower Neuse Basin Association, NC Sea Grant and the UNC Water Resources Research Institute.

Supplementary material

10533_2018_438_MOESM1_ESM.pdf (17 kb)
Supplementary material 1 (PDF 17 kb)

References

  1. Abril G, Bouillon S, Darchambeau F, Teodoru CR, Marwick TR, Tamooh F et al (2015) Technical note: large overestimation of pCO2 calculated from pH and alkalinity in acidic, organic-rich freshwaters. Biogeosci 12(1):67–78.  https://doi.org/10.5194/bg-12-67-2015 CrossRefGoogle Scholar
  2. Adams SM, Greeley MS, Law JM, Noga EJ, Zelikoff JT (2003) Application of multiple sublethal stress indicators to assess the health of fish in Pamlico Sound following extensive flooding. Estuaries 26:1365–1382CrossRefGoogle Scholar
  3. Alemeddine I, Qian S, Reckhow K (2011) A Bayesian changepoint-threshold model to examine the effect of TMDL implementation on the flow-nitrogen concentration relationship in the Neuse River Basin. Wat Res 45:51–62CrossRefGoogle Scholar
  4. Almeida RM, Pacheco FS, Barros N, Rosi E, Roland F (2017) Extreme floods increase CO2 outgassing from a large Amazonian river. Limnol Oceanogr.  https://doi.org/10.1002/lno.10480 CrossRefGoogle Scholar
  5. Arar EJ, Budde WL, Behymer TD (1997) Methods for the determination of chemical substances in marine and environmental matrices. EPA/600/R-97/072. National Exposure Research Laboratory, U.S. Environmental Protection Agency, CincinnatiGoogle Scholar
  6. Bales JD (2003) Effects of Hurricane Floyd inland flooding, September-October 1999, on tributaries to Pamlico Sound, North Carolina. Estuaries 26:1319–1328CrossRefGoogle Scholar
  7. Bales JD, Oblinger CJ, Sallenger AH (2000) Two months of flooding in eastern North Carolina, September–October 1999: hydrologic, water-quality, and geologic effects of Hurricanes Dennis, Floyd, and Irene. U.S. Geological Survey Water-Resources Investigations Report 00-4093. MI Dissertation Services, Ann ArborGoogle Scholar
  8. Bastviken D, Tranvik LJ, Downing J, Crill PM, Enrich-prast A (2011) Freshwater methane emissions offset the continental carbon sink. Science 331:50.  https://doi.org/10.1126/science.1196808 CrossRefGoogle Scholar
  9. Battin TJ, Luyssaert S, Kaplan L, Aufdenkampe AK, Richter A, Tranvik LJ (2009) The boundless carbon cycle. Nat Geosci 2:598–600.  https://doi.org/10.1038/ngeo618 CrossRefGoogle Scholar
  10. Bauer JE, Cai W-J, Raymond PA, Bianchi TS, Hopkinson CS, Regnier PAG (2013) The changing carbon cycle of the coastal ocean. Nature 504:61–70.  https://doi.org/10.1038/nature12857 CrossRefGoogle Scholar
  11. Bender MA, Knutson TR, Tuleya RE, Sirutis JJ, Vecchi GA, Garner ST, Held IM (2010) Modeled impact of anthropogenic warming on the frequency of intense Atlantic hurricanes. Science 327:454–458CrossRefGoogle Scholar
  12. Bianchi TS (2007) Biogeochemistry of estuaries. Oxford University Press, New YorkGoogle Scholar
  13. Bianchi TS, Bauer JE (2011) Particulate organic carbon cycling and transformation. In: Wolanski E, McLusky DS (eds) Treatise on estuarine and coastal science, vol 5. Academic Press, Waltham, pp 69–117CrossRefGoogle Scholar
  14. Bianchi TS, Garcia-Tigreros F, Yvon-Lewis SA et al (2013) Enhanced transfer of terrestrially derived carbon to the atmosphere in a flooding event. Geophys Res Lett 40:116–122.  https://doi.org/10.1029/2012GL054145 CrossRefGoogle Scholar
  15. Brown MM, Mulligan RP, Miller RL (2014) Modeling the transport of freshwater and dissolved organic carbon in the Neuse River Estuary, NC, USA following Hurricane Irene (2011). Estuar Coast Shelf Sci 139:148–158.  https://doi.org/10.1016/j.ecss.2014.01.005 CrossRefGoogle Scholar
  16. Burkholder J, Eggleston D, Glasgow H, Brownie C et al (2004) Comparative impacts of two major hurricane seasons on the Neuse River and western Pamlico Sound ecosystems. Proc Natl Acad Sci USA 101:9291–9296CrossRefGoogle Scholar
  17. Burkholder JM, Dickey DA, Kinder CA, Reed RE et al (2006) Comprehensive trend analysis of nutrients and related variables in a large eutrophic estuary: a decadal study of anthropogenic and climatic influences. Limnol Oceanogr 51:463–487CrossRefGoogle Scholar
  18. Butman D, Raymond PA (2011) Significant efflux of carbon dioxide from streams and rivers in the United States. Nat Geosci 4:839–842.  https://doi.org/10.1038/ngeo1294 CrossRefGoogle Scholar
  19. Buzzelli CP, Luettich RA, Powers SP, Peterson CH et al (2002) Estimating the spatial extent of bottom-water hypoxia and habitat degradation in a shallow estuary. Mar Ecol Progr Ser 230:103–112CrossRefGoogle Scholar
  20. Buzzelli CP, Ramus JR, Paerl HW (2003) Ferry-based monitoring of surface water quality in North Carolina estuaries. Estuaries 26:975–984CrossRefGoogle Scholar
  21. Christian RR, Boyer JN, Stanley DW (1991) Multi-year distribution patterns of nutrients within the Neuse River Estuary, North Carolina. Mar Ecol Progr Ser 71(3):259–274CrossRefGoogle Scholar
  22. Christian RR, O’Neal JE, Peierls BL, Valdes LM, Paerl HW (2004) Episodic nutrient loading impacts on eutrophication of the southern Pamlico Sound: the effects of the 1999 Hurricanes. UNC Water Resources Research Institute Report No. 349. UNC Water Resources Research Institute, RaleighGoogle Scholar
  23. Cole JJ, Caraco NF (2001) Carbon in catchments: connecting terrestrial carbon losses with aquatic metabolism. Mar Freshw Res 52:101–110CrossRefGoogle Scholar
  24. Cole JJ, Prairie YT, Caraco NF et al (2007) Plumbing the global carbon cycle: integrating inland waters into the terrestrial carbon budget. Ecosystems 10:171–184.  https://doi.org/10.1007/s10021-006-9013-8 CrossRefGoogle Scholar
  25. Conner WH, Day JW, Baumann RH, Randall JM (1989) Influence of hurricanes on coastal ecosystems along the northern Gulf of Mexico. Wetlands Ecol Manag 1:45.  https://doi.org/10.1007/BF00177889 CrossRefGoogle Scholar
  26. Copeland BJ, Gray J (1991) Status and trends report of the Albemarle-Pamlico Estuary. In: Steel J (ed) Albemarle-Pamlico Estuarine Study Report 90-01. North Carolina Department of Environment, Health and Natural Resources, RaleighGoogle Scholar
  27. Crosswell J (2013) Air-water CO2 fluxes in estuaries: sources, sinks and storms. PhD dissertation, University of North Carolina at Chapel HillGoogle Scholar
  28. Crosswell JR, Wetz MS, Hales B, Paerl HW (2012) Air-water CO2 fluxes in the microtidal, seasonally-stratified Neuse River Estuary, North Carolina. J Geophys Res 117:C08017.  https://doi.org/10.1029/2012JC007925 CrossRefGoogle Scholar
  29. Crosswell JR, Wetz MS, Hales B, Paerl HW (2014) Extensive CO2 emissions from shallow coastal waters during passage of Hurricane Irene (August 2011) over the Mid-Atlantic Coast of the U.S.A. Limnol Oceanogr 59:1651–1665CrossRefGoogle Scholar
  30. Dixon JL, Osburn CL, Paerl HW, Peierls BL (2014) Seasonal changes in estuarine dissolved organic matter due to variable flushing time and wind-driven mixing events. Estuar Coast Shelf Sci 151(2014):210–220CrossRefGoogle Scholar
  31. Eby LA (2001) Response of a fish community to frequent and infrequent disturbances in an estuarine ecosystem. PhD dissertation, Duke University, DurhamGoogle Scholar
  32. Eby LA, Crowder LB (2002) Hypoxia-based habitat compression in the Neuse River Estuary: context-dependent shifts in behavioral avoidance thresholds. Can J Fish Aquat Sci 59:952–965CrossRefGoogle Scholar
  33. Egleston ES, Sabine CL, Morel FM (2010) Revelle revisited: buffer factors that quantify the response of ocean chemistry to changes in DIC and alkalinity. Glob Biogeochem Cycles 24:1–9.  https://doi.org/10.1029/2008GB003407 CrossRefGoogle Scholar
  34. Emanuel K (2005) Increasing destructiveness of tropical cyclones over the past 30 years. Nature 436:686–688CrossRefGoogle Scholar
  35. Ensign SH, Doyle MW (2006) Nutrient spiraling in streams and river networks. J Geophys Res Biogeosci 111:1–13.  https://doi.org/10.1029/2005JG000114 CrossRefGoogle Scholar
  36. Ensign SH, Paerl HW (2006) Development of an unattended estuarine monitoring program using ferries as data collecting platforms. Limnol Oceanogr Methods 4:399–405CrossRefGoogle Scholar
  37. Gaulke AK, Wetz MS, Paerl HW (2010) Picophytoplankton: a major contributor to planktonic biomass and primary production in a eutrophic, river-dominated estuary. Estuar Coast Shelf Sci 90:45–54CrossRefGoogle Scholar
  38. Giffin D, Corbett DR (2003) Evaluation of sediment dynamics in coastal systems via short-lived radioisotopes. J Mar Syst 42:83–96CrossRefGoogle Scholar
  39. Glibert PM, Heil CA, Hollander D, Revilla M, Hoare A, Alexander J, Murasko S (2004) Evidence for dissolved organic nitrogen and phosphorus uptake during a cyanobacterial bloom in Florida Bay. Mar Ecol Prog Ser 280:73–83CrossRefGoogle Scholar
  40. Goldenberg SB, Landsea CM, Mestas-Nuñez AM, Gray WM (2001) The recent increase in Atlantic hurricane activity: causes and implications. Science 293:474–478CrossRefGoogle Scholar
  41. Hagy JD, Boynton WR, Keefe CW, Wood KV (2004) Hypoxia in Chesapeake Bay, 1950–2001: long-term change in relation to nutrient loading and river flow. Estuaries 27:634–658CrossRefGoogle Scholar
  42. Hagy JD, Lehrter JC, Murrell MC (2006) Effects of Hurricane Ivan on water quality in Pensacola Bay, Florida. Estuaries Coasts 29:919–925CrossRefGoogle Scholar
  43. Hall NS, Paerl HW (2011) Vertical migration patterns of phytoflagellates in relation to light and nutrient availability in a shallow, microtidal estuary. Mar Ecol Progr Ser 425:1–19CrossRefGoogle Scholar
  44. Hall NS, Litaker RW, Fensin E, Adolf JE, Place AR, Paerl HW (2008) Environmental factors contributing to the development and demise of a toxic dinoflagellate (Karlodinium veneficum) bloom in a shallow, eutrophic, lagoonal estuary. Estuar Coasts 31:402–418CrossRefGoogle Scholar
  45. Hall NS, Paerl HW, Peierls BL, Whipple AC, Rossignol KL (2013) Effects of climatic variability on phytoplankton biomass and community structure in the eutrophic, microtidal, New River Estuary, North Carolina, USA. Estuar Coast Shelf Sci 117:70–82CrossRefGoogle Scholar
  46. Harding LE, Mallonee ME, Perry ES, Miller WD, Adolf JE, Gallegos C, Paerl HW (2016) Variable climatic conditions dominate recent phytoplankton dynamics in Chesapeake Bay. Nat Sci Rep 6:23773.  https://doi.org/10.1038/srep23773 CrossRefGoogle Scholar
  47. Hilton RG, Galy A, Hovius N, Chen M-C, Horng M-J, Chen H (2008) Tropical-cyclone-driven erosion of the terrestrial biosphere from mountains. Nat Geosci 1:759–762.  https://doi.org/10.1038/ngeo333 CrossRefGoogle Scholar
  48. Hirsch RM, DeCicco L (2014) User guide to exploration and graphics for RivEr Trends (EGRET) and data Retrieval: R packages for hydrologic data. Technical report. Techniques and Methods book 4, Chap. A10. US Geological Survey, Reston. http://pubs.usgs.gov/tm/04/a10/
  49. Hirsch RM, Moyer DL, Archfield SA (2010) Weighted regressions on time, discharge, and season (WRTDS), With an application to Chesapeake Bay river inputs. J Am Water Res Assoc 46:857–880CrossRefGoogle Scholar
  50. Holland GJ, Webster PJ (2007) Heightened tropical cyclone activity in the North Atlantic: natural variability of climate trend? Philos Trans R Soc A.  https://doi.org/10.1098/rsta.2007.2083 CrossRefGoogle Scholar
  51. Hu X, Cai W-J (2013) Estuarine acidification and minimum buffer zone—a conceptual study. Geophys Res Lett 40:5176–5181.  https://doi.org/10.1002/grl.51000 CrossRefGoogle Scholar
  52. Intergovernmental Panel on Climate Change (IPCC) (2014) Climate change 2014: synthesis report. In: Core Writing Team, Pachauri RK, Meyer LA (eds) Contribution of working groups I, II and III to the fifth assessment report of the intergovernmental panel on climate change. IPCC, GenevaGoogle Scholar
  53. Jeffrey SW, Wright SW, Zapata M (1999) Recent advances in HPLC pigment analysis of phytoplankton. Mar Freshw Res 50:879–896CrossRefGoogle Scholar
  54. Jia P, Li M (2012) Circulation dynamics and salt balance in a lagoonal estuary. J Geophys Res Ocean 117:C01003.  https://doi.org/10.1029/2011JC007124 CrossRefGoogle Scholar
  55. Kang N-Y, Elsner JB (2015) Trade-off between intensity and frequency of global tropical cyclones. Nat Clim Change 5:661–664CrossRefGoogle Scholar
  56. Kennish M, Paerl HW (2010) Coastal lagoons: critical habitats of environmental change. CRC Marine Science Series. CRC Press, Boca RatonCrossRefGoogle Scholar
  57. Laruelle GG, Goossens N, Arndt S, Cai W-J, Regnier P (2017) Air-water CO2 evasion from U.S. East Coast estuaries. Biogeosciences 14:2441–2468.  https://doi.org/10.5194/bg-2016-278 CrossRefGoogle Scholar
  58. Lebo ME, Paerl HW, Peierls BL (2012) Evaluation of progress in achieving TMDL mandated nitrogen reductions in the Neuse River Basin, North Carolina. Environ Manag 49:253–266CrossRefGoogle Scholar
  59. Lewis E, Wallace D, Allison LJ (1998) Program developed for CO2 system calculations. Carbon Dioxide Information Analysis Center, BerkeleyCrossRefGoogle Scholar
  60. Mallin MA, Paerl HW, Rudek J, Bates PW (1993) Regulation of estuarine primary production by watershed rainfall and river flow. Mar Ecol Progr Ser 93:199–203CrossRefGoogle Scholar
  61. Miller WD, Harding LW Jr, Adolf JE (2006) Hurricane Isabel generated an unusual fall bloom in Chesapeake Bay. Geophys Res Lett 33(L06612):2006.  https://doi.org/10.1029/2005GL025658 CrossRefGoogle Scholar
  62. Millero FJ (2010) Carbonate constants for estuarine waters. Mar Freshw Res 61:139–142.  https://doi.org/10.1071/MF09254 CrossRefGoogle Scholar
  63. NOAA Hurricane Center, Miami, FL. http://www.nhc.noaa.gov
  64. North Carolina Department of Environment and Natural Resources, Division of Water Quality (2001) Phase II of the total maximum daily load for total nitrogen to the Neuse River Estuary, North Carolina, December 2001Google Scholar
  65. Osburn CL, Handsell LT, Mikan MP, Paerl HW, Montgomery MT (2012) Fluorescence tracking of dissolved and particulate organic matter quality in a river-dominated estuary. Environ Sci Technol 46:8628–8636CrossRefGoogle Scholar
  66. Osburn CL, Handsell LT, Peierls BL, Paerl HW (2016a) Predicting sources of dissolved organic nitrogen to an estuary from an Agro-Urban Coastal Watershed. Environ Sci Technol 50:8473–8484CrossRefGoogle Scholar
  67. Osburn CL, Boyd TJ, Montgomery MT, Coffin RB, Bianchi TS, Paerl HW (2016b) Optical proxies for terrestrial dissolved organic matter in estuaries and coastal waters. Front Mar Sci 2:127.  https://doi.org/10.3389/fmars.2015.00127 CrossRefGoogle Scholar
  68. Paerl HW, Huisman J (2008) Blooms like it hot. Science 320:57–58CrossRefGoogle Scholar
  69. Paerl HW, Peierls BL (2008) Ecological responses of the Neuse river-Pamlico Sound estuarine continuum to a period of elevated hurricane activity: impacts of individual storms and longer term trends. Am Fish Soc Symp 64:101–116Google Scholar
  70. Paerl HW, Mallin MA, Donahue CA, Go M, Peierls BL (1995) Nitrogen loading sources and eutrophication of the Neuse River estuary, NC: direct and indirect roles of atmospheric deposition. UNC Water Resources Research Institute Report No. 291. Water Resources Res Institute, RaleighGoogle Scholar
  71. Paerl HW, Pinckney JL, Fear JM, Peierls BL (1998) Ecosystem responses to internal and watershed organic matter loading: consequences for hypoxia in the eutrophying Neuse River Estuary, North Carolina, USA. Mar Ecol Progr Ser 166:17–25CrossRefGoogle Scholar
  72. Paerl HW, Bales JD, Ausley LW, Buzzelli CP, Crowder LB, Eby LA, Go M, Peierls BL, Richardson TL, Ramus JS (2000) Hurricanes’ hydrological, ecological effects linger in major US estuary. EOS 81(40):457–462CrossRefGoogle Scholar
  73. Paerl HW, Bales JD, Ausley LW, Buzzelli CP et al (2001a) Ecosystem impacts of 3 sequential hurricanes (Dennis, Floyd and Irene) on the US’s largest lagoonal estuary, Pamlico Sound, NC. Proc Natl Acad Sci USA 98(10):5655–5660CrossRefGoogle Scholar
  74. Paerl HW, Buzzelli CP, Go M, Peierls BL et al (2001b) Water quality and fisheries habitat changes in the Pamlico Sound after three hurricanes: a short-term and long-term perspective. In: Maiolo JR, Whitehead JC, McGee M, King K, Johnson J, Stone H (eds) Facing our future: hurricane floyd and recovery in the coastal plain. Coastal Carolina Press, Wilmington, pp 255–263Google Scholar
  75. Paerl HW, Valdes LM, Adolf JE, Peierls BL, Harding LW Jr (2006a) Anthropogenic and climatic influences on the eutrophication of large estuarine ecosystems. Limnol Oceanogr 51:448–462CrossRefGoogle Scholar
  76. Paerl HW, Valdes LM, Joyner AR, Peierls BL et al (2006b) ecological response to hurricane events in the Pamlico Sound System, NC and implications for assessment and management in a regime of increased frequency. Estuar Coasts 29:1033–1045CrossRefGoogle Scholar
  77. Paerl HW, Valdes LM, Piehler MF, Stow CA (2006c) Assessing the effects of nutrient management in an estuary experiencing climatic change: the Neuse River Estuary, NC, USA. Environ Manag 37:422–436CrossRefGoogle Scholar
  78. Paerl HW, Valdes LM, Joyner AR, Winkelmann V (2007) Phytoplankton indicators of ecological change in the nutrient and climatically-impacted Neuse River-Pamlico Sound system, North Carolina. Ecol Appl 17(5):88–101CrossRefGoogle Scholar
  79. Paerl HW, Rossignol KL, Hall NS, Guajardo R, Joyner AR, Peierls BL, Ramus J (2009) FerryMon: ferry-based monitoring and assessment of human and climatically-driven environmental change in the Pamlico Sound System, North Carolina, USA. Environ Sci Technol 43:7609–7613CrossRefGoogle Scholar
  80. Paerl HW, Rossignol KL, Hall NS, Peierls BL, Wetz MS (2010a) Phytoplankton community indicators of short- and long-term ecological change in the anthropogenically and climatically impacted Neuse River Estuary, North Carolina, USA. Estuar Coasts 33:485–497CrossRefGoogle Scholar
  81. Paerl HW, Christian RR, Bales JD, Peierls BL et al (2010b) Assessing the response of the Pamlico Sound, North Carolina, USA to human and climatic disturbances: management implications. In: Kennish M, Paerl H (eds) Coastal lagoons: critical habitats of environmental change CRC Marine Science Series. CRC Press, Boca Raton, pp 17–42CrossRefGoogle Scholar
  82. Paerl HW, Hall NS, Peierls BL, Rossignol KL, Joyner AR (2013) Hydrologic variability and its control of phytoplankton community structure and function in two shallow, coastal, lagoonal ecosystems: the Neuse and New River Estuaries. Estuaries and Coasts, North Carolina.  https://doi.org/10.1007/s12237-013-9686-0 CrossRefGoogle Scholar
  83. Paerl HW, Hall NS, Peierls BL, Rossignol KL (2014) Evolving paradigms and challenges in estuarine and coastal eutrophication dynamics in a culturally and climatically stressed world. Estuar Coasts 37(2):243–258CrossRefGoogle Scholar
  84. Peierls BL, Paerl HW (2010) Temperature, organic matter, and the control of bacterioplankton in the Neuse River and Pamlico Sound estuarine system. Aquat Microb Ecol 60:139–149CrossRefGoogle Scholar
  85. Peierls BL, Paerl HW (2011) Longitudinal and depth variation of bacterioplankton productivity and related factors in a temperate estuary. Estuar Coastal Shelf Sci 95:207–215CrossRefGoogle Scholar
  86. Peierls BL, Christian RR, Paerl HW (2003) Water Quality and phytoplankton as indicators of hurricane impacts on a large estuarine ecosystem. Estuaries 26:1329–1343CrossRefGoogle Scholar
  87. Peierls BL, Hall NS, Paerl HW (2012) Non-monotonic responses of phytoplankton biomass accumulation to hydrologic variability: a comparison of two coastal plain North Carolina estuaries. Estuar Coasts 35:1376–1392CrossRefGoogle Scholar
  88. Phlips EJ, Badylak S, Hart J et al (2011) Climatic influences on autochthonous and allochthonous phytoplankton blooms in a subtropical estuary, St. Lucie Estuary, Florida, USA. Estuar Coasts 35:335–352.  https://doi.org/10.1007/s12237-011-9442-2 CrossRefGoogle Scholar
  89. Piehler MF, Twomey LJ, Hall NS, Paerl HW (2004) Impacts of inorganic nutrient enrichment on the phytoplankton community structure and function in Pamlico Sound, NC USA. Estuar Coast Shelf Sci 61:197–207CrossRefGoogle Scholar
  90. Pietrafesa LJ, Janowitz GS, Chao T-Y, Weisberg TH, Askari F, Noble E (1996) The physical oceanography of Pamlico Sound. UNC Sea Grant Publication UNC-WP-86-5. UNC Sea, RaleighGoogle Scholar
  91. Pinckney JL, Paerl HW, Harrington MB, Howe KE (1998) Annual cycles of phytoplankton community structure and bloom dynamics in the Neuse River Estuary, NC (USA). Mar Biol 131:371–382CrossRefGoogle Scholar
  92. Pinckney JL, Paerl HW, Harrington MB (1999) Responses of the phytoplankton community growth rate to nutrient pulses in variable estuarine environments. J Phycol 35:1455–1463CrossRefGoogle Scholar
  93. Pinckney JL, Richardson TL, Millie DF, Paerl HW (2001) Application of photopigment biomarkers for quantifying microalgal community composition and in situ growth rates. Org Geochem 32:585–595CrossRefGoogle Scholar
  94. Raymond PA, Hartmann J, Lauerwald R et al (2013) Global carbon dioxide emissions from inland waters. Nature 503:355–359.  https://doi.org/10.1038/nature12760 CrossRefGoogle Scholar
  95. Raymond PA, Saiers JE, Sobczak WV (2016) Hydrological and biogeochemical controls on watershed dissolved organic matter transport: pulse-shunt concept. Ecology 95:833–844Google Scholar
  96. Regnier P, Friedlingstein P, Clais P, Mackenzie FT et al (2013) Anthropogenic perturbation of the carbon fluxes from land to ocean. Nat Geosci 6(8):597–607.  https://doi.org/10.1038/ngeo1830 CrossRefGoogle Scholar
  97. Roman MR, Adolf JE, Bichy J, Boicourt WC et al (2005) Chesapeake Bay plankton and fish abundance enhanced by Hurricane Isabel. EOS 86:261–265CrossRefGoogle Scholar
  98. Rudek J, Paerl HW, Mallin MA, Bates PW (1991) Seasonal and hydrological control of phytoplankton nutrient limitation in the lower Neuse River Estuary, North Carolina. Mar Ecol Progr Ser 75:133–142CrossRefGoogle Scholar
  99. Ruzecki EP, Schubel JH, Huggett RJ, Anderson AM et al (1976) The effects of tropical storm Agnes on the Chesapeake Bay estuarine system. Chesapeake Bay Research Consortium Publication No 54. Johns Hopkins University Press, BaltimoreGoogle Scholar
  100. Schlesinger WH (2005) Biogeochemistry: an analysis of global change. Academic Press, New YorkGoogle Scholar
  101. Schlesinger WH, Melack JM (1981) Transport of organic carbon in the world’s rivers. Tellus 33:172–187CrossRefGoogle Scholar
  102. Seneviratne SI, Nicholls N, Easterling D, Goodess CM et al (2012) Chapter 3: changes in climate extremes and their impacts on the natural physical environment. In: Field CD, Barros V, Stocker TF, Dahe Q et al (eds) Managing the risks of extreme events and disasters to advance climate change adaptation. A special report of working groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, pp 109–230Google Scholar
  103. Stanley DW (1988) Historical trends in nutrient loading to the Neuse River Estuary, NC. In: Lyke W, Hoban T (eds) Proceedings of the American Water Resources Association symposium on coastal water resources AWRA Technical Publication Series TPS-88-1. AWRA, Bethesda, pp 155–164Google Scholar
  104. Stow CA, Borsuk ME, Stanley DW (2001) Long-term changes in watershed nutrient inputs and riverine exports in the Neuse River, North Carolina. Wat Res 35:1489–1499CrossRefGoogle Scholar
  105. Tester PA, Varnam SM, Culver ME, Eslinger DL et al (2003) Airborne detection of ecosystem responses to an extreme event: phytoplankton displacement and abundance after hurricane induced flooding in the Albemarle-Pamlico Sound system. Estuaries 26:1353–1364CrossRefGoogle Scholar
  106. Tranvik LJ, Downing JA, Cotner JB et al (2009) Lakes and reservoirs as regulators of carbon cycling and climate. Limnol Oceanogr 54:2298–2314.  https://doi.org/10.4319/lo.2009.54.6_part_2.2298 CrossRefGoogle Scholar
  107. Valdes-Weaver LM, Piehler MF, Pinckney JL, Howe KE, Rosignol K, Paerl HW (2006) Long-term temporal and spatial trends in phytoplankton biomass and class-level taxonomic composition in the hydrologically variable Neuse-Pamlico estuarine continuum, NC, USA. Limnol Oceanogr 51(3):1410–1420CrossRefGoogle Scholar
  108. Van Dam BR, Crosswell JR, Anderson IC, Paerl HW (2018) Watershed-scale drivers of air-water CO2 exchanges in two lagoonal North Carolina (USA) estuaries. J Geophys Res Biogeosci.  https://doi.org/10.1002/2017JG004243 CrossRefGoogle Scholar
  109. Van Heukelem L, Lewitus A, Kana T, Craft N (1994) Improved separations of phyto-plankton pigments using temperature-controlled high performance liquid chromatography. Mar Ecol Progr Ser 114:303–313CrossRefGoogle Scholar
  110. Webster PJ, Holland GJ, Curry JA, Chang HR (2005) Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309:1844–1846CrossRefGoogle Scholar
  111. Welschmeyer NA (1994) Fluorometric analysis of chlorophyll a in the presence ofchlorophyll b and pheopigments. Limnol Oceanogr 39:1985–1992CrossRefGoogle Scholar
  112. Wetz MS, Paerl HW (2008) Estuarine phytoplankton responses to hurricanes and tropical storms with different characteristics (trajectory, rainfall, winds). Estuar Coasts 31:419–429CrossRefGoogle Scholar
  113. Wetz MS, Yoskowitz DW (2013) An “extreme” future for estuaries? Effects of extreme climatic events on estuarine water quality and ecology. Mar Pollut Bull 69:7–18.  https://doi.org/10.1016/j.marpolbul.2013.01.020 CrossRefGoogle Scholar
  114. Wuebbles D, Meehl G, Hayhoe K et al (2014) CMIP5 climate model analyses: climate extremes in the United States. Bull Am Meteorol Soc 95:571–583.  https://doi.org/10.1175/BAMS-D-12-00172.1 CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  • Hans W. Paerl
    • 1
  • Joseph R. Crosswell
    • 2
  • Bryce Van Dam
    • 1
  • Nathan S. Hall
    • 1
  • Karen L. Rossignol
    • 1
  • Christopher L. Osburn
    • 3
  • Alexandria G. Hounshell
    • 1
  • Randolph S. Sloup
    • 1
  • Lawrence W. HardingJr.
    • 4
  1. 1.Institute of Marine SciencesUniversity of North Carolina at Chapel HillMorehead CityUSA
  2. 2.CSIRO Oceans and Atmosphere, Ecosciences PrecinctBrisbaneAustralia
  3. 3.Department of Marine, Earth, and Atmospheric SciencesNorth Carolina State UniversityRaleighUSA
  4. 4.Department of Atmospheric and Oceanic SciencesUniversity of California, Los AngelesLos AngelesUSA

Personalised recommendations