Final Report: The Hudson Riverscope Prototype: Real-Time Monitoring of Rivers and Estuaries for Research, Education and Science-Based Decision SupportEPA Grant Number: CR830976
Title: The Hudson Riverscope Prototype: Real-Time Monitoring of Rivers and Estuaries for Research, Education and Science-Based Decision Support
Investigators: Cronin, John , Bell, R. , Nierzwicki-Bauer, S.
Institution: Pace University - New York , Columbia University in the City of New York , Rensselaer Polytechnic Institute
EPA Project Officer: Hahn, Intaek
Project Period: August 14, 2003 through July 31, 2005
Project Amount: $487,500
RFA: Targeted Research Grant (2002) Recipients Lists
Research Category: Targeted Research
Riverscope seeks to bring science to the service of natural resource management. Functionally, it seeks to understand, for the first time in real-time, the flow of water and its transport of organic and inorganic substances along the approximately 300 miles of the Hudson Estuary. Ultimately, it seeks to communicate that understanding to resource managers, policy makers, educators, and interested citizens as well as to the international research community. A pilot framework of research monitoring sites is established. This is the foundation that will enable a national, multidisciplinary understanding of rivers, estuaries and their associated watersheds. Riverscope—the real-time monitoring system for research, education and policy—is integral to achieving New York Governor Pataki’s plan for the Rivers and Estuaries Center. Riverscope’s pilot projects, undertaken under this grant, demonstrated the design and implementation of a river-wide monitoring system. The objectives of this research project were to: (1) monitor sediment transport using acoustic techniques, and (2) monitor the movement of invasive species using cutting-edge oceanographic methodology with sampling of the river’s biota.
Sediment Transport Mechanisms—Acoustic Monitoring
This part of the project demonstrates the application of Acoustic Doppler Current Profilers (ADCPs) to monitor throughput and sediment transport in the Hudson River Estuary. The ADCP experiments monitored the response of the river system to different events, allowing the study of short-term behavior relative to long-term changes.
System Setup. System setup achieved compatible instrumentation in the lower tidal (Lamont Doherty Earth Observatory [LDEO]), middle tidal (U.S. Geological Survey [USGS]), and upper (Rensselaer Polytechnic Institute [RPI]) Hudson. LDEO purchased an R&D Instruments (RDI) ADCP (Workhorse Monitor, 1200 kHz) and a Wetlabs FLNTU-B flourometer and optical backscatter sensor. Together with a second ADCP already owned by LDEO, two identical tripods were equipped. An additional CTD was typically attached to each tripod. Both systems were powered by battery packs that provided sufficient power for approximately 6 weeks of operation. RPI purchased a RDI Channel Master 1200 kHz H-ADCP and a D&A Instrument Company OBS-3A optical turbidity sensor identical to that used by USGS. Equipment for Lock 2 was purchased from YSI/Sontek, which provided an integrated 1500 kHz Argonaut-SL H-ADCP and vertical profiling package to provide information on water chemistry parameters. The vertical profiling package includes a 6600-Series Multiparameter Extended Deployment Sonde, a 6200 data logger, a fixed-mount profiling winch, a solar cell trickle charger, and a cellular data transmission system. In 2005, an additional Sontek 1500 kHz Argonaut-SL H-ADCP was purchased for installation on the Mohawk River just upstream of the confluence with the Hudson.
Deployment. In the lower Hudson two sites were chosen to complement ongoing monitoring programs by USGS in Poughkeepsie and Lock 2. Site 1 was located in the main channel of Haverstraw Bay near Stony Point. Site 2 was located in the Tappan Zee near Piermont Pier. Both sites represent the top of the salt wedge and embrace the widest section of the Hudson River Estuary, which is important for the sediment budget of the system. The tripods were deployed for a total period of approximately 3 months between March and July 2004. After about 6 weeks, both systems were retrieved and redeployed after successful data download and battery change.
Sites for ADCP deployment were chosen on the upper Hudson at Champlain Canal Lock 2 near Mechanicville and in the tidal Hudson at Albany, approximately 10 miles downstream from the confluence with the Mohawk River. In 2005, a third site was chosen at Erie Canal Guard Gate 2 on the Mohawk River just above the confluence with the Hudson. The Lock 2 site is most indicative of suspended sediment transport from the upper Hudson and tributaries, whereas the Albany site is sensitive to upper Hudson inputs as well as the addition of sediment from the predominantly agricultural Mohawk drainage basin. The ADCP site on the Erie Canal directly indicates the contribution of suspended sediment from the Mohawk basin and provides a check on the observations at the Albany site. Like the Haverstraw Bay sites, the Albany site is within the tidal reach of the river but is located well above the influence of salt water. The ADCPs at the Albany and Lock 2 sites became active in July 2004. Although ADCP equipment for these sites arrived too late to catch the main spring freshet in 2004, these sites were active throughout the spring 2005 meltwater event. RPI researchers collected discrete water samples from all three sites throughout the 2004 and 2005 deployment seasons. These water samples provide a daily time series of total suspended sediment (TSS) concentrations prior to ADCP activation. Samples collected since ADCP deployment provide TSS data for calibration of the ADCP-derived acoustic backscatter intensity measurements. Furthermore, the samples of filtered material have been archived for possible future analysis (e.g., percent organic matter and particulate inorganic carbon).
Results. In the lower Hudson, both systems were retrieved and they successfully recorded data. Data from the deployment have been processed and preliminary analysis accomplished. Two tripods were deployed on the bed of the Hudson River estuary for approximately 100 days spanning the 2004 spring freshet. They were placed at the deepest part of the river channel near Piermont (3/24/04 - 7/12/04) and Northern Haverstraw Bay (3/24/04 - 7/3/04). Each held an RDI ADCP (Workhorse Monitor, 1200 kHz) facing upward to monitor currents and acoustic backscatter through the water column. The Piermont tripod held a Seabird SBE-19 CTD, a Wetlabs FLNTU-B flourometer, and an optical backscatter sensor for monitoring chlorophyll concentrations and turbidity. The lower Hudson data have been analyzed and compared with tidal range and phase, river flow, and winds. Tides dominate variability in vertical mixing and sediment transport and this tidal mixing is driven by the interaction of the tidal current with the bottom boundary, a relatively simple task for numerical modeling to reproduce. However, at the Haverstraw mooring, there was high turbulent kinetic energy production (and eddy viscosity) in the middle and upper water column, suggesting that there is an internal water column source of turbulence near the head of the salinity intrusion (e.g. internal waves or density interface drag). One strong (30 knots) and sustained wind event on April 5th led to the full mixing of the water column a few days prior to spring tide, in spite of high river (buoyancy) input to the system. Internally driven and wind-driven vertical mixing can be challenging tasks for numerical modeling, so we plan to analyze these periods in greater detail.
The Upper Hudson data suggest that at least two major flow events occurred during the deployment period. The TSS data from the Lock 2, Cohoes (Mohawk) and Albany sites similarly show that major transport events occurred in early April and mid- to late June. Whereas the April event was driven by the influx of spring meltwater, later events were storm- and rainfall-driven. Acoustic backscatter intensity measurements from Albany follow the tidal cycle, with maximum intensity (maximum transport) occurring between tides when the current is strongest. Both the Albany and the Lock 2 intensity data show obvious variability tied to rainfall events. The high-resolution data collected over the 2005 spring freshet demonstrate the effective linear range of ADCP monitoring as an indicator of sediment transport in the upper Hudson. Preliminary calibrations and time trends of backscatter/TSS from the 2004 and spring 2005 deployments were presented at a meeting of the Northeastern Section of the Geological Society of America in March 2005.
Data from the Upper Hudson and the Lower Hudson provide valuable insights into the design and extension of these efforts system-wide. Specifically, targets for future efforts include the issues of biofouling and wind-driven events in the lower Hudson and the issues of high-flow events and the resolution of the coarse fraction in the upper Hudson.
Zebra Mussel Dispersion and Assessment of Future Trends
This part of the project demonstrates elucidation of the distribution and abundance of the exotic species Dreissena polymorpha, zebra mussels, along the Hudson River. Before this project, there had not been any systematic monitoring of larval zebra mussels, known as veligers, along the stretch of the Hudson from the Troy Dam to the south, and thus there was no capacity to map distribution for management purposes. The project’s objective was to correlate the veliger monitoring information with data acquired through ADCP and SF6 tracer studies to attempt to map the location of adult zebra mussel colonies, the spawning beds of the infestation. The outcome of this objective is discussed in the Tracer Release Experiment below.
System Setup. Based on the preliminary results that were provided in the November 12, 2004, report, a new experimental design was adopted. It was decided that rather than complete the future work, described in the previous report, it was more important and useful to initiate another sampling season. The reason for the change was a strong preference for having three seasons of data on the original sites rather than two and to sample intermediate locations. This would provide a higher resolution than had been provided in the previous 2 years. Planning and preparation for this supplemental sampling season was therefore given the highest priority.
In this additional 2005 season, sampling was completed along two stretches of the River: Stuyvesant to Kingston and from Albany to Castleton. Different sampling regimes were utilized for each stretch. First, during the 2004 experiment, SF6 was injected just north of Stuyvesant, immediately south of Houghtaling Island. This injection site was chosen as the northerly starting point for the first part of the 2005 project. Sampling was completed on both the left and the right bank at every nautical mile downstream to Sturgeon Point, just south of Kingston. Samples at these 35 sites were taken at three discrete depths and an integrated sample of the entire water column. The location pattern was chosen to supplement the 2003 and 2004 sampling data by replicating the sites that had been sampled in the previous 2 years. Sites also were added, halfway between the original sites. Thus, samples now have been collected from the same sites for 3 successive years, during the same time period (third week of June). Locations sampled in 2004 and 2005, interspersed between the original 26 sites, provide a more detailed picture of veliger distribution throughout the sampled sections of the River.
The second sample period employed three methods:
- A single, medium-depth location (5.5 M, “Glenmont: 2003”) was sampled (both depth-integrated and discrete depths) continuously, taking 12 samples over 6 hours.
- Samples were collected, on the right bank, at one-quarter-mile intervals starting at the Port of Albany (Cuyler Dike: 2003) for 29 locations downstream to Castleton-on-Hudson, replicating the original 2003 sites and amplifying the site resolution by inserting 7 additional sites between each of the 2003 sites.
- Depth-integrated samples were taken at a single deep location (10 M; Buoy 216, Van Wies Point) every 10 minutes for 3 hours totaling 18 samples.
- In 2003 the original sites were spaced at 2 nautical miles throughout the 56-mile study area, from Troy Dam to Kingston.
- In 2004, the 35-mile section from Stuyvesant to Kingston was sampled at 166 locations.
- In 2005, each mile of the Stuyvesant-to-Kingston section was sampled and every one-quarter mile was sampled from the Port of Albany to Castleton-on-Hudson.
Results. Microscopy analysis, using cross-polarized light techniques, is currently in progress to identify and enumerate veligers to construct a high-resolution map of veliger density along the sampled stretches of the Hudson River. The 2005 enumeration results will be used to compare densities for each of the three sample years. Data also are being collected on the larval morphological forms of D. polymorpha, as well as on the sampled range of size and age.
SF6 Tracer Release Experiment
Successful modeling of water flow and contaminant transport requires knowledge and understanding of physical processes such as mean advection, dispersive mixing, and air-water gas exchange. Knowledge of these transport processes aid in the detection, remediation, and treatment of substances discharged into the river and in the evaluation of the role of sediments in the transport of toxic substances through the system. Tracer release data provide valuable complementary information to Eulerian measurements discussed in the Sediment Transport and Zebra Mussel Dispersion sections of Riverscope.
System Setup. A highlight of this particular tracer experiment is the application of an automated, high-resolution SF6 measurement system developed by LDEO in 2003, which dramatically improves the spatial (and temporal) resolution of the measurements and increases the size of a water body that can be surveyed.
Deployment. On June 17, 2004, ca. 4.3 mols of SF6 were injected from a boat into the Hudson River off Stuyvesant, New York. For 10 days after the injection, as the boat traversed the tracer patch, surface water SF6 was measured at one minute resolution by using a continuous SF6 analysis system, described in detailed by Ho, et al. (2002). The system consists of a submersible pump, a gas extraction system, and a gas separation and measurement system. SF6 depth profiles were obtained by lowering a Niskin bottle to different depth, and transferring the samples to glass syringes and analyzing them on a gas chromatograph on board the boat.
Results. The SF6 tracer patch moved down river and spread out as a result of advection and longitudinal dispersion. Preliminary results indicate that, at the end of the experiment, SF6 covered a stretch of the river from New Baltimore (kmp 213) to Kingston (kmp 153). The longitudinal dispersion coefficient calculated from the tidally corrected tracer distribution (26.1 ± 2.3 m2 s-1) was similar to that of a stretch of the river immediately south of our tracer patch (24.6 ± 6.6 m2 s-1; from Kingston to Poughkeepsie) measured in the summer of 2003, and net advection was 3.0 ± 0.1 km day-1.
Complementary Activities. The tracer made it possible to follow a patch of water as it traveled downstream to further study veligers. During the same 10-day period, 166 plankton samples were collected to determine veliger density and size. On 4 days (6/24-6/27) veliger densities at the peak of the tracer patch were determined. Size measurements of veligers in the integrated water sample were made for each of the four sample dates (100 veligers measured for each date). As would be expected if the same “patch” of water was being followed the average size of veligers in samples would increase each day. This is exactly what was observed, with the average veliger sizes as follows: 6/24 (85.6 μm), 6/25 (89.2 μm), 6/26 (93.2 μm), and 6/27 (94.7 μm).
To simultaneously determine possible spatial (vertical) diversity of the veligers in the water column, plankton samples of three types were collected (integrated sample, discrete sample from 1 M, and discrete sample from mid-depth). Also, a spatial analysis of size distribution (vertical profile through the water column) was carried out for samples collected on June 26. This demonstrated that veligers of all sizes were well mixed/distributed in the water column, with similar size distributions in the integrated, 1 M below the surface (D1) and middle of the water column at 4 M (D2) samples. The results, to date, have served to examine the temporal and spatial distribution of veligers throughout a stretch of the Hudson River Estuary and to create a map of veliger distribution and abundance.
The ADCP deployments clearly demonstrate the potential to capture and monitor runoff events and associated sediment plumes in the Hudson River Estuary. The movement along the system and travel speed can be monitored. Longer time series and systematic monitoring of sediment plumes associated, for example, with rainstorms in different parts of the Hudson River watershed, would enable analysis of the contributions of different tributaries and their effects. The New York State Canal Corporation (NYSCC) has granted a permanent occupancy permit for the Lock 2 monitoring equipment. We are awaiting NYSCC approval for a permanent permit for the Mohawk site. We anticipate submitting manuscripts detailing the results in early 2006.
Analysis of all samples collected is presently the focus of the project. This involves the enumeration of veliger density at each location and at each depth and the measurement of a sampling of these veligers to provide a size distribution. When completed, the integration of veliger size data with the flow data from the SF6 experiment will be possible. The end product of this integration will be to trace the origin of individual veligers back to their spawning beds.
Publication of the results is anticipated for the winter of 2005.
Education and Outreach
Education and outreach are planned for the next phase of Riverscope.