Scott H. Ensign

Demographic, landscape, and meteorological factors controlling the microbial pollution of coastal waters

Coastal areas in the United States and many other countries are considered to be desirable regions to live and recreate. However, as human use of coastal land and water increases, so does the incidence of aquatic-borne disease from contact with contaminated water and eating contaminated shellfish. Movement of humans into coastal areas both greatly increases the number of sources of microbial pathogens and radically alters the landscape through increased construction activity and paving of former natural areas. On a regional scale, increases in human population over a 14-year period in coastal North Carolina were strongly correlated with increases in shellfish bed closures due to high fecal coliform bacterial counts. On a watershed scale, an analysis of several tidal creeks found strong correlations between mean estuarine fecal coliform bacterial counts and watershed population, percent developed area and especially with percent impervious surface coverage. Conversion of natural landscapes to impervious surfaces (roads, drives, sidewalks, parking lots and roofs) removes the lands natural filtration capability, allows for increased concentration of pollutants at the lands surface and provides a means of rapid conveyance of pollutants to downstream waterways. An analysis of rural watersheds in the Coastal Plain found that stream fecal coliform counts and turbidity were both strongly correlated with rainfall in the previous 24 h in watersheds containing extensive industrial swine and poultry operations, as well as watersheds containing more traditional agriculture and cattle husbandry. In contrast, in watersheds rich in swamp wetlands these relationships were not significant, even in watersheds containing extensive animal production. Based on these findings, we suggest that waterborne microbial pathogen abundance can be minimized in urbanizing coastal areas through reduced use of impervious surfaces and maximal use of natural or constructed wetlands for passive stormwater runoff treatment. In animal husbandry areas, retention of natural wetlands and management practices designed to minimize sediment runoff can likely reduce inputs of pathogenic microbes into streams.

HURRICANE EFFECTS ON WATER QUALITY AND BENTHOS IN THE CAPE FEAR WATERSHED: NATURAL AND ANTHROPOGENIC IMPACTS

In the summer of 1996, southeastern North Carolina, United States, was struck by two hurricanes, with the second (Hurricane Fran) doing considerably more damage than the first (Hurricane Bertha). The Cape Fear watershed, largest in North Carolina, suffered from severe water quality problems for weeks following Fran, including a massive fish kill in the Northeast Cape Fear River. Post-hurricane flooding caused inputs of riparian swamp water to river channels, and sewage treatment plant and pump station power failures caused diversions of millions of liters of raw and partially treated human waste into rivers. Additionally, several swine waste lagoons were breached, overtopped, or inundated, discharging large quantities of concentrated organic waste into the system, particularly into the Northeast Cape Fear River. Dissolved oxygen (DO) decreased to 2 mg/L in the mainstem Cape Fear River, and fell to zero in the Northeast Cape Fear River for >3 wk. Biochemical oxygen demand in the Northeast Cape Fear River w...

Comparative impacts of stormwater runoff on water quality of an urban, a suburban, and a rural stream

Water quality data at 12 sites within an urban, a suburban, and a rural stream were collected contemporaneously during four wet and eight dry periods. The urban stream yielded the highest biochemical oxygen demand (BOD), orthophosphate, total suspended sediment (TSS), and surfactant concentrations, while the most rural stream yielded the highest total organic carbon concentrations. Percent watershed development and percent impervious surface coverage were strongly correlated with BOD (biochemical oxygen demand), orthophosphate, and surfactant concentrations but negatively with total organic carbon. Excessive fecal coliform abundance most frequently occurred in the most urbanized catchments. Fecal coliform bacteria, TSS, turbidity, orthophosphate, total phosphorus, and BOD were significantly higher during rain events compared to nonrain periods. Total rainfall preceding sampling was positively correlated with turbidity, TSS, BOD, total phosphorus, and fecal coliform bacteria concentrations. Turbidity and TSS were positively correlated with phosphorus, fecal coliform bacteria, BOD, and chlorophyll a, which argues for better sedimentation controls under all landscape types.

Stream water quality changes following timber harvest in a coastal plain swamp forest

The Goshen Swamp, a fourth order blackwater creek in southeastern North Carolina, was clearcut of 130 acres of riparian and seasonally flooded forest in late May through September 1998. Downstream water quality had been monitored monthly for 2 1/2 years before the clearcut, during the clearcut, and for two years following the clearcut. The objective of this paper was to test the hypothesis that clearcutting in the Goshen Swamp watershed negatively impacted downstream water quality. To do so, data from the Goshen Swamp were compared with data collected from a neighboring control creek (Six Runs Creek) of similar size, land use, and hydrologic characteristics. Compared with the control creek, the post-clearcut Goshen Swamp displayed significantly higher suspended solids, total nitrogen, total phosphorus, total Kjeldahl nitrogen and fecal coliform bacteria, and significantly lower dissolved oxygen over a 15 month period. Longer-term deleterious effects included recurrent nuisance algal blooms that had not been present during the 2 1/2 years before the clearcut. Although a 10 m uncut buffer zone was left streamside, this was insufficient to prevent the above impacts to stream water quality.

PHOTOSYNTHETIC AND HETEROTROPHIC IMPACTS OF NUTRIENT LOADING TO BLACKWATER STREAMS

Blackwater stream systems are the most abundant type of freshwater lotic system on the Coastal Plain of the eastern United States. Many of these ecosystems drain watersheds that receive large anthropogenic nutrient inputs, whereas some blackwater systems remain relatively pristine. A series of nutrient addition experiments was carried out over a four-year period to assess the roles of nitrogen and phosphorus loading on the phytoplankton, bacterioplankton, and respiration of two third-order and two fifth-order blackwater streams in southeastern North Carolina. Stream water was distributed into 4-L containers, amended with various nutrient addition treatments, and incubated in gently agitated outdoor pools over a six-day period. Chlorophyll a production, direct bacterial counts, ATP, and biochemical oxygen demand (BOD) were measured as response variables. Significant phytoplankton production over control occurred in most experiments involving nitrogen additions, regardless of whether it was in the form of a...

Application of digital bathymetry data in an analysis of flushing times of two large estuaries

The objectives of this study were: (1) define the best method of using digital bathymetric data to compute estuarine flushing time using the fraction of freshwater method and (2) use this method to compare flushing times of two neighboring estuaries of different trophic state. We examined the sensitivity of the fraction of freshwater method to various methods of calculating estuarine volume using digital bathymetric data. Raster and vector bathymetry data are available from the National Geophysical Data Center (NGDC), and can be used to calculate estuarine volume using a geographic information system (GIS). The vector data was of higher spatial resolution than the raster data (NGDC Coastal Relief Model) and produced a higher estuarine volume, but did not produce significantly different flushing times than the raster data. Water column salinity data can be used to quantify segmented vertical freshwater volumes for integration along the estuary, thereby providing a two-dimensional freshwater distribution profile of the estuary. The vertical representation of water column salinity did not produce flushing times significantly different from a vertically averaged salinity method. Processing and analysis of the Coastal Relief Model raster data is faster and less complex than processing the vector data available from the NGDC. We conclude that the Coastal Relief Model raster data is the preferred bathymetric data source, and that representation of vertical salinity distribution is unnecessary for the analysis of estuaries with morphology similar to the Cape Fear Rivers. After using the Cape Fear River estuary as a test site for the above comparisons, we applied the preferred method to the New River estuary. In addition to having a direct connection with the ocean, the Cape Fear River has much higher freshwater inflow than the New River, and therefore has a much faster mean flushing time. The Cape Fear River estuary flushing time ranged from 1 to 22 days, while the New River estuary ranged from 8 to 187 days. Similar seasonal patterns were observed in both estuaries: short flushing times occurred during the high-flow winter months and long flushing times occurred during the low-flow summer months.

Alternative Reference Frames in River System Science

Understanding environmental processes begins with mental conceptualizations of system components and interactions. Conceptualizing rivers begins with adopting one of two reference frames for observing movement: Eulerian, wherein the flux of objects is observed in a spatially bounded area, or Lagrangian, wherein specific objects are tracked through time. Mechanistic studies include Eulerian and Lagrangian data, with some negotiation of how much Eulerian and Lagrangian information may be needed to maximize the accuracy of understanding processes and the efficiency of data collection. Most studies rely on a presupposed reference frame, yet we suspect breakthroughs lurk in explicit alterations of presupposed reference frames. We analyze the importance of reference frames by contrasting the extent to which alternative reference frames have been used and combined in studies of sediment transport, fish migration, and river biogeochemistry. We show how adopting alternative or nonintuitive reference frames can facilitate novel research questions and observations, potentially triggering new research trajectories.

Spatial and temporal variability in excessive soil phosphorus levels in eastern North Carolina

Numerous studies have shown that accumulation of excessive soil phosphorus raises the potential for phosphorus export and eutrophication of adjacent surface waters. Soil phosphorus data from the North Carolina Agronomy Divisions database were analyzed for two-year periods spanning the decades of the 1980s and 1990s for 39 eastern North Carolina counties. Eastern North Carolina supported extensive row crop agriculture, rapidly growing intensive livestock industries, and a growing human population during these decades. Excessive soil phosphorus levels, defined as having a soil phosphorus index (P-I, based on Mehlich III testing) > 100, occurred in over 40% of almost a million samples reported for the three two-year periods analyzed. Excessive soil P-I levels were most frequent in central eastern North Carolina, declined in the 1980s and rose again in the 1990s. The distribution of row crop area with excessive soil P-I levels was very similar in time and space. Increases in the area harvested for cotton (+635%) and pasture (+523%) with excessive soil P-I levels were particularly large during the 1990s, when crop areas harvested associated with excessive soil P-I levels for other major crops (corn, tobacco, peanuts) declined. Residential and recreational land uses were associated with similarly high frequencies of excessive soil P-I levels, but these land uses were relatively unimportant (<5% area) compared to agricultural land use (~34%) in the region. Recent increases in fertilizer shipments (approximately twofold in the late 1990s) likely reflected increased cotton production. Rapid growth in concentrated animal production (almost twofold increase in total animal units (AU) between 1992 and 2001), with accompanying land application of wastes, accounted for increases in soil P-I values in pasturelands in the 1990s, particularly in central eastern North Carolina, where these activities were concentrated. The potential threat to water quality from export of excessive soil phosphorus is therefore greatest in this region. North Carolina is currently developing a Phosphorus Loss Assessment Tool (PLAT) in an attempt to manage the challenge posed by excessive soil phosphorus levels.

Hydrologic Versus Biogeochemical Controls of Denitrification in Tidal Freshwater Wetlands

Tidal freshwater wetlands (TFW) alter nitrogen concentrations in river water, but the role of these processes on a river’s downstream nitrogen delivery is poorly understood. We examined spatial and temporal patterns in denitrification in TFW of four rivers in North Carolina, USA and evaluated the relative importance of denitrification rate and inundation on ecosystem-scale N2 efflux. An empirical model of TFW denitrification was developed to predict N2 efflux using a digital topographic model of the TFW, a time series of water level measurements, and a range of denitrification rates. Additionally, a magnitude-frequency analysis was performed to investigate the relative importance of storm events on decadal patterns in N2 efflux. Spatially, inundation patterns exerted more influence on N2 efflux than did the range of denitrification rate used. Temporal variability in N2 efflux was greatest in the lower half of the tidal rivers (near the saline estuary) where inundation dynamics exerted more influence on N2 efflux than denitrification rate. N2 efflux was highest in the upper half of the rivers following storm runoff, and under these conditions variation in denitrification rate had a larger effect on N2 efflux than variability in inundation. The frequency-magnitude analysis predicted that most N2 efflux occurred during low flow periods when tidal dynamics, not storm events, affected TFW inundation. Tidal hydrology and riparian topography are as important as denitrification rate in calculating nitrogen loss in TFW; we present a simple empirical model that links nitrogen transport in rivers with loss due to denitrification in TFW.

Geomorphic Implications of Particle Movement by Water Surface Tension in a Salt Marsh

Measurements and modeling of salt marsh morphodynamics typically focus on the sediment fraction suspended in water, but another fraction moves at the air-water interface on water surface tension. We compared the geomorphic significance of this surface microlayer sediment fraction to the suspended fraction in a mesohaline, microtidal salt marsh in North Carolina. Conventional methods of suspended sediment and surface microlayer sampling were used and a method was developed to measure the surface microlayer material generated per area of benthic surface. The mass of material held in the surface microlayer was 4% to 16% of the suspended sediment mass in the water column overlying the marsh. This surface microlayer material mass was lower than the mass generated from benthic surfaces and it declined over time, indicating that the surface microlayer released sediment to the water column. Loss on ignition, carbon:nitrogen, and particle size were similar in the surface microlayer and suspended sediment, although the surface microlayer material was higher in chlorophyll a. We estimate that 19% to 100% of the suspended sediment that settled onto the marsh was lifted by surface tension during the next rising tide. This process helps explain the topography of tidal marshes and patterns of sediment accretion.