2018 Projects

We propose to continue for a second year a project to develop a mechanistic eutrophication model of the Neuse River Estuary. That original one-year project (WRRI Project No. 16-02-W, “Comparing the Impact of Organic vs. Inorganic Nitrogen Loading to the Neuse Estuary with a Mechanistic Eutrophication Model”) was narrowly focused and designed to leverage our earlier model development work in the estuary. We now have successfully updated the model and used it to investigate the system’s sensitivity to nitrogen load reduction (Harrigan and Bowen 2016, Bowen 2017). An additional year of funding would enable us to make significant additional contributions and improvements to the model and would allow us to work with the potential users of the model such as the North Carolina Division of Water Resources (DWR) who are interested in running it for basin planning and water quality assessment.We plan model development in three areas. First, after May 2009 our data set lacks water surface elevation at the downstream boundary. Adding this single time history would enable us to extend the model period by more than five years. We plan to use the ADCIRC (Luettich Jr, Westerink et al. 1992) model to simulate elevations. Second, we will develop a simple second order sediment diagenesis sub-model to predict denitrification and sediment related oxygen demand to replace the approximate approaches that we currently use. Third we have leveraged our earlier work on pathogen modeling (Froelich, Bowen et al. 2013) to develop an automated calibration system that can run the model in parallel over 2-50 processors. This system allows us to test thousands of candidate parameter vectors during calibration and then run the calibrated model with many suitable parameter vectors. Because of time restrictions we have the system sparingly, but in the coming year we would use it to systematically test and calibrate alternate approaches to modeling the algal constituents. We plan also to have several briefings during the coming year with DWR staff to update them on our progress and to query them on how best to tailor the model and create the front and back-end systems needed for their use of the model.   

Nitrogen (N) loading to our streams and rivers has improved since the mid-1990s through management practices that have focused on pollution from advanced wastewater treatment plants, urban stormwater, and agricultural facilities. However, load reductions to surface waters like the Neuse River have not reached targeted goals and eutrophication remains a major concern, particularly in the face of projected population increases and shifts in precipitation patterns that will only increase N pollution. A strategic shift in focus to smaller wastewater treatment systems in rural communities offers another solution for reducing N pollution. Often overlooked, the discharge limits for smaller systems for ammonia-nitrogen (NH4-N) are often high (10 mg/L) or even non-existent. Facilities that use aerobic processes to treat wastewater in smaller, rural communities often successfully treat NH4-N to low levels through the process of nitrification, but the effluent contains the byproduct nitrate-nitrogen (NO3-N) which in-turn also contributes to eutrophication. Installation of constructed wetlands, known for high N removal potential, placed strategically in the landscape to intercept water from smaller rural wastewater treatment facilities, could be a solution to help North Carolina get closer to its N reduction goals. These systems have been successful in NC; however, few are in operation, performance data are limited, and some older systems are showing signs of decreased N removal. The objectives of this research and outreach project are to study and improve removal performance of an older existing constructed wetlands in Walnut Cove, NC and evaluate and demonstrate the impact constructed wetlands could have on overall watershed N reduction when used as tertiary treatment for smaller wastewater package plants.  

There is increasing evidence that dissolved organic nitrogen (DON) can stimulate phytoplankton growth and play an important role in structuring phytoplankton community composition. Specifically, high ratios of DON to dissolved inorganic nitrogen (DIN) may select for potentially toxic harmful algal blooms (HABs) which continue to plague lakes, rivers and freshwater reservoirs. Recent attention has been focused on the composition and quality of DON discharged in waste water treatment plant (WWTP) effluent, particularly as WWTPs switch from conventional activated sludge (CAS) to biological nutrient removal (BNR) processes in response to increasing pressure to reduce total nitrogen (TN) loading in WWTP effluent. BNR systems significantly reduce TN discharged in effluent, but increases the concentration of DON that may be bio-reactive and capable of stimulating phytoplankton growth. The goals of this project were to determine whether BNR processes produce bioreactive DON and determine whether these compounds stimulate phytoplankton growth, particularly HAB species. A combination of bulk DON analyses, fluorescent spectroscopy, humic, and protein DON analyses were used to characterize the DON in the influent and effluent waters of a representative BNR WWTP to assess whether DON is produced during BNR treatment or is inherent to the influent water. Two experimental nutrient additions of influent and effluent waters were made during summer and fall 2018 to natural phytoplankton and microbial assemblages from a eutrophic reservoir (Jordan Lake, North Carolina) to determine the potential stimulatory role of effluent DON and to assess whether bio-stimulatory DON is produced by the BNR process or is inherent to the received influent water. Influent and effluent characterization revealed that influent waters contain high concentrations of potentially bioreactive protein. However, about 80% of the protein is removed within the WWTP process, and protein contained in the effluent appeared largely refractory. During the summer experiment, effluent additions stimulated total phytoplankton biomass beyond the stimulation that could be attributed to the effluent inorganic nutrient content. Accessory pigment analysis, indicated that diatoms/chrysophytes, chlorophytes, and dinoflagellates were the primary taxa that were stimulated. During the fall experiment, only the diatom/chrysophyte group was stimulated by effluent additions. Potentially harmful cyanobacteria taxa were either slightly inhibited (summer) or showed no response to effluent organic matter additions (fall). Based on canonical phytoplankton stoichiometry and lack of decline in the DON pool, phytoplankton growth during the experiments could be best explained solely by observed uptake of DIN. Therefore, it seems possible that the observed stimulation of phytoplankton growth is due to some other substance, e.g. a vitamin, other co-factor, or micronutrient, rather than direct incorporation of effluent DON. Results from these experiments indicate that although BNR effluent may contain a growth stimulatory substance, it does not contain significant quantities of highly labile DON that could fuel high biomass attainment beyond the carrying capacity set by its DIN concentration. Thus, in nature, BNR effluent would not be expected to strongly stimulate phytoplankton biomass increases in highly N limited reservoirs like Jordan Lake.

Sedimentation is a major pollution problem in North Carolina according to the North Carolina Sediment Pollution Control Act. Gullies formed by road drainage are a potential source of sediment, especially in the mountainous western portion of the state (WNC). However, due to steep topography and dense vegetation, these features are difficult to locate. This study aimed to leverage high resolution LiDAR-derived digital elevation models (DEMs) from the state of North Carolina to locate road-draining gullies remotely using a geospatial model. Specifically, the goals of the project were to identify road-draining erosional features automatically, examine whether they are linked with particular environmental conditions, and determine how they may impact the landscape in terms of being connected sources of sediment. We used an empirical GIS approach to identify gully candidates within the two-county area which included slope as part of a multivariate logistic model. This approach highlighted areas that may be gullies but also identified broad areas of DEM cells that were not gullies. In order to answer the project questions, we visually examined the areas highlighted by the model and then determined areas that were gullies adjacent to roads using a field-based approach. We identified gullies in both Jackson and Haywood counties, and our initial conclusion is that they play a small role in the overall sediment budgets of the landscapes. However, this process is likely important on a hillslope or small watershed scale, especially because it is likely that more gullies may be located with an improved algorithm or more searching time. The promising aspect of this work is that it does appear we can more clearly answer questions about water and sediment connectivity on high resolution DEMs (0.5 m grid cells). Several environmental variables (e.g., landcover, geology and soils) were examined to determine any links to gully formation, and our initial conclusion is that these variables had little impact on the system’s ability to resist gully formation. The key variables that seem to emerge to explain gully locations are very local site conditions: slope and ditch length, the latter of which we estimated in the field. We hope to go back to determine if we can either improve our empirical approach or adopt another approach that improves our ability to automatically locate gullies relative to non-gully locations. We hope the results of this study have enhanced understanding of human-influence over drainage networks and erosion/sedimentation issues in WNC, and be useful to local governments and water quality advocacy organizations.

The goal of this project was to develop a comprehensive, deterministic, distributed and physically-based hydrologic model (MIKE-SHE) to provide the City of Jacksonville (NC) with historical and current visualizations of how their current municipal wastewater treatment system, a forest land application site (FWR), functions hydrologically among the seasons in response to weather, forest age, and forest management. Project objectives were to use the model to (1) forecast FWR response under different scenarios of weather extremes and (2) to forecast FWR response under different regimens of water reuse or forest composition using a water balance approach. The City of Jacksonville is exploring strategies to increase FWR capacity for future demand. Simulated evapotranspiration (ET) and water table depth (WTD), using MIKE-SHE and twenty years of measured precipitation and irrigation data from the land treatment facility, were used to calculate drainage (runoff and lateral flow) across the site. Irrigation impacted ET and WTD to the greatest extent for forest areas surrounded by irrigation fields but caused little change in annual ET. Forest water use was relatively unchanged by irrigation, and annual watershed drainage increased proportionally to irrigation input. The drivers of on-site drainage were rainfall and the amount of irrigation. In wet years, ET and groundwater levels remained constant while drainage increased in response to rainfall and irrigation. WTD in wells surrounded by wastewater irrigation remained consistently closer to the surface than wells with only partial irrigation nearby. The model under-predicted WTD for wells on the site’s exterior during below average rainfall periods. Groundwater withdrawal for agricultural use by adjacent landowners may explain this discrepancy. Extreme rainfall events, such as Hurricane Florence, resulted in high volumes of drainage but rapid recovery of groundwater storage in the FWR. The model provided insight to management practices that could increase FWR efficiency and flexibility for managing variable weather with climate change. One observation of the model was that current increases in irrigation volumes from winter application volumes to higher summer application volumes lag behind ET. Hence, one management option to increase irrigation capacity is to increase irrigation volume earlier in March rather than May. A sensitivity analysis of rooting depth and leaf area index to water use showed that rooting depth mattered more for water use than LAI. Forest management practices such as bedding for replanting would improve rooting depth of trees. Nancy Gibson presented project results to the City of Jacksonville in March 2019. COJ was very excited that project results support current operation perspectives that the FWR could treat more wastewater if allowed more flexibility to land apply when conditions are optimal rather than prescribed volumes per week. Model observations that irrigation does not limit FWR ET and that rainfall drives FWR export of water were key outcomes that resonated with city officials and operators. Project results will be provided as an executive summary for the City of Jacksonville personnel to use in discussions with NCDEQ as both organizations discuss revised permits to avoid emergency spraying for extreme storms. This study has shown that these unique forest systems offer insights to water balance dynamics in irrigated forests and forest resiliency to extreme hydraulic loading that can be of use to regional wastewater land treatment systems for North Carolina.

Many potable source-waters in the U.S. are sustaining cultural eutrophication from elevated nutrient supplies concomitant with major shifts in nutrient ratios away from healthy conditions as indicated by Redfield proportions (16:1, molar). Falls Lake is a representative eutrophic impoundment (reservoir; length ~38 km, width ~0.2-2.9 km) in a rapidly urbanizing watershed in the southeastern U.S. The shallow upper reservoir is especially prone to algal blooms in comparison to the deeper lower reservoir (mean depth 5 m and 14 m, respectively). Using a long-term water quality dataset provided by the Center for Applied Aquatic Ecology (CAAE; 2011-2019, biweekly to monthly samples), we compared the upper and lower reservoir over time for nitrogen (N) and phosphorus (P) nutrient regimes, phytoplankton biomass (as chlorophyll a [chla] concentrations), and phytoplankton assemblage composition during bloom conditions defined as > 40 µg chla/L, the state water quality standard. In the upper reservoir, TP and inorganic N (Ni) were chronically elevated and chla commonly exceeded the state standard, with noxious cyanobacteria dominant in cell number. In contrast, the lower reservoir was generally characterized by moderate nutrient regimes except for elevated NH4 + (up to ~400 µg/L) in surface waters during fall-winter. Short-term experiments were conducted to gain further insights about assemblage responses to changing nutrient regimes during two summers that were planned to be replicate seasons; however, one summer had average precipitation (considering the past decade; 127 mm) whereas the other had much higher precipitation (170 mm). Microcosm experiments in situ (duration, 5 days) were used to assess reservoir phytoplankton assemblage responses to inorganic N form (Ni) + inorganic P (Pi) enrichment, and to Ni:Pi ratios. The abundance and composition of phytoplankton functional groups were similar reservoir-wide in both summers, except for higher relative abundance of cyanobacteria in the upper region under average precipitation. In both summers, there was a positive relationship between chla and Ni concentrations (both forms), with or without Pi enrichment, and a stronger relationship between chla and NH4+ than between chla and NO3. Also in both summers, the eutrophic lower region assemblage responded more strongly to nutrient enrichment. Maximal final phytoplankton biomass as chla was attained, for the upper region assemblage, with Ni + Pi enrichment as NO3. For the lower region assemblage, maximum chla occurred with Ni + Pi enrichment as NH4+. The toxigenic cyanobacterium Cylindrospermopsis raciborskii was the most abundant taxon initially reservoir-wide; it was stimulated by enrichment with either Ni form, especially along with Pi enrichment. Similar responses in the two summers likely occurred because the precipitation/dilution/washout regimes would only have affected the phytoplankton before they were placed into closed microcosms in the experiments. The precipitation differential during the two summers, although substantial, probably was not enough to cause major changes in these resilient reservoir. Overall, this study contributed species-level insights about seasonal influences of chronic cultural eutrophication on reservoir phytoplankton blooms. The findings indicate that reservoir assemblages are well-adapted to variable precipitation/hydrologic changes, and that cyanobacteria will continue to be favored under warming temperatures and high N/P enrichment. This study also supports co-managed reductions of both N and P supplies, as in the Falls Lake Rules, to minimize noxious algal blooms in this important, representative potable source-water reservoir of the North Carolina Piedmont.