2. OBSERVING SYSTEM
3. MODELING SYSTEM
4. OPERATIONAL FORECAST SYSTEM
5. INFORMATION DSISTRIBUTION
6. ACOUSTIC TRANSMISSION LOSS
The New York Harbor Observing and Prediction System (NYHOPS, viewable at http://www.stevens.edu/maritimeforecast ) was designed and installed by the Center for Maritime Systems to provide operations-critical information to the Port of New York and New Jersey. The system was developed with support from the Office of Naval Research and the State of New Jersey Department of Transportation, and through continuing partnerships with NOAA, the US Coast Guard, the Port Authority of New York and New Jersey, the States of New York and New Jersey, and several maritime industry participants. NYHOPS combines maritime environment and vessel observing and forecasting systems to provide:
NOAA currently operates a Physical Oceanography Real-Time System (PORTS) ocean monitoring system in the Harbor. This system, viewable at http://www.co-ops.nos.noaa.gov/nyports/nyports.html provides ocean data and ocean model forecasts in the approaches to the berthing areas to support commercial vessel traffic. The Stevens NYHOPS system significantly enhances and expands the NOAA PORTS system by the addition of new sensors in the Hudson River and in the lower Harbor and Raritan Bay - areas that are heavily trafficked by ferry vessels, recreational craft, fishing vessels, and sightseeing/cruise ships. This expanded system will support the region's MTS, as well as the port security and first responder communities. During the continuing development of the NYHOPS system, Stevens and NOAA are strongly collaborating in order to maximize benefit to the maritime community.
NYHOPS allows for high-resolution coverage of the region while also permitting the scientific analysis of several phenomena unique to this urban harbor area, including:
The New York Harbor Observing and Prediction System is an open-access network of distributed sensors and linked computer estuarine and coastal ocean forecasting models. The system is designed to allow for the real-time assessment of ocean, weather, and environmental conditions throughout the New York Harbor region, and forecast of conditions in the near and long-term. As such, the system functions as both an operational product and a research tool to foster the economic and environmental health of the region. NYHOPS includes three main components: observations, numerical modeling, and information distribution. The overall system architecture that shows the interaction among these three components is shown in Figure 1. The architecture is designed so that the modeling system interacts with the sensor network by turning on a particular sensor, by changing its sampling rate or by dispatching a service crew to effect repairs when necessary.
Figure 1. The operational system architecture is designed to have an interaction among the observing system, numerical model, and information distribution system. These three components must all function together for the system to be effective.
Real-time oceanographic information within New York Harbor is obtained using various sensors placed at strategic locations to monitor the current state of the estuarine environment. The challenge of providing this information is particularly difficult in urban environments. The urban environments are typically shallow coastal areas and estuaries, where oceanographic and atmospheric conditions exhibit high spatial and temporal variability due to the influences of freshwater inflows, combined sewer overflows (CSO), tides, micro-climate, and bottom and land topography which is subject to change because of dredging/deepening programs. The design and installation of observing systems in these environments is further complicated by the presence of high turbidity, strong stratification, strong currents, fog, vessel traffic and limited shoreline access. The sensor design and network configuration are patterned after an existing Stevens system on the New Jersey coast (Bruno, et. al. 2001). The network sensors, all of which sprovide their data in real-time, consist of:
The shore-based and moored data are sampled every 10 min and are transmitted on an hourly basis via a radio link. The ferry data are sampled and transmitted every 30 seconds via a cellular connection. This observation network provides very large spatial coverage of the region of interest at sufficiently high resolution to provide a real-time "mosaic" of the oceanographic and weather conditions (see Figure 2 for sensor locations).
Figure 2. General setting of the Port of New York and New Jersey and the location of sensors in the NYHOPS sensor network. "Partner" sensors are owned and maintained by NOAA/NOS, Rutgers University and The River Project. Supplementing this network are observations made from ferries that travel throughout these waters on a routine basis.
Supplementing this mosaic are observations from the Stevens Coastal Monitoring Network (Bruno, et al, 2001). The network provides real-time observations and archived records at four locations along the coast of New Jersey of shallow water (5m) wave characteristics, water temperature, water level, and meteorological conditions. Each station is fitted with a digital camera that provides a picture of the beach every 5 minutes. The recorded data is disseminated via the Internet at: http://cmn.dl.stevens-tech.edu/.
The land, moored and ferry-based sensors transmit data via secure wireless networks. Data QA/QC procedures are in place to ensure the reliability of the observations. The data is also transferred to the NOAA National Weather Service's National Data Buoy Center for distribution via the NDBC website, http://www.ndbc.noaa.gov/Maps/Northeast.shtml. In partnership with Rutgers University's Institute of Marine and Coastal Sciences (IMCS), a HF CODAR current measurement system is under development to provide real-time measurements of the surface currents at the entrance of the Harbor. An example of the data processing protocol from collection to distribution is shown in Figure 3, which illustrates the surface temperature, surface salinity and surface dissolved oxygen derived from the Pioneer-based sensor. The Pioneer sails out of the South Street Museum several times a day during the summer. Interestingly, as a result of this very detailed information, we are able for the first time to learn about the circulation at the confluence of the East River and the Hudson River. At the end of the ebb tide along Manhattan, water moves to the south carrying fresher Hudson River water with it. At the same time in the East River water is moving towards the Battery from Long Island Sound carrying saltier water from the sound via Hell's Gate. These two water masses meet where the Pioneer typically sails and the data permits a close look at the water masses interact.
Figure 3. Surface temperature, salinity and dissolved oxygen on July 13, 2005 as measured from a sensor onboard the Pioneer as it sails in the Hudson River in the vicinity of the Battery, NY. There is a marked change in water mass properties in this highly active region.
The modeling system consists of a three-dimensional circulation model and a wave model. It is designed to provide reliable and timely meteorological and oceanographic nowcasts/forecasts for the maritime waters of New York and New Jersey including the waters eastward to the continental shelf break of the New York Bight in one contiguous fashion.
The modeling system is based on the Estuarine Coastal and Ocean Model (ECOMSED), a derivative of the Princeton Ocean Model (POM - http://www.aos.princeton.edu/WWWPUBLIC/htdocs.pom/)) developed by Blumberg and Mellor (1987).. The ECOMSED/POM combination is in use by almost 1900 research groups around the world and over 600 papers have been published with them as the modeling engine. The spatial extent of the NYHOPS domain incorporates the core area of NY/NJ Harbor and extends beyond to include the Hudson River Estuary up to the Troy Dam, all of Long Island Sound and the New York Bight out to the continental shelf (Figure 4).
Figure 4. The domain of NYHOPS and the attendant grid mesh.
The model was calibrated by comparing water levels at 14 locations, current velocities at 6 locations, salinity and temperature at 35 locations over a period of one year from October 1988 to September 1989. A hydrologically different simulation period, October 1994 to September 1995, was used for model validation. Compared with 1988-89, the latter period is relatively colder, with 10 W/m2 lower yearly mean shortwave solar radiation, and significantly drier, with only 66 percent of inflows from rivers. The model was validated against elevation data at 13 locations, current data at five locations and salinity and temperature data at 27 locations. Detailed calibration and validation efforts have been described by Blumberg et al. (1999). The model has been run continuously beginning with the year 1999 to the present time without any loss of predictive skill.Back to top
The prediction system is designed to provide accurate and comprehensive realizations of meteorological and oceanographic conditions in the past (hindcast), present (nowcasts) and future (forecasts). It is based on the operational protocol designed by Kelley, et al (1997) for the NOAA/NWS Regional Ocean Forecast System (ROFS) which produces daily forecasts of ocean properties for the coastal ocean of the United States. The NYHOPS simulations are performed and archived on the Stevens Hydrodynamic Computational Laboratory's high-performance computer cluster resident in the Davidson Laboratory. Forcing is based on NOAA/NWS/NCEP's Mesoscale Eta 12km atmospheric forecast model, USGS river flow gauges, NOAA/NOS water level stations and effluents from 110 wastewater treatment plants obtained from the Interstate Sanitation Commission. Stormwater runoffs are being incorporated at the present time. The system is designed to automatically transfer forcing data and perform forecasts once daily, at 0000 hours. The NOAA Operational Model Archive Distribution System (N.O.M.A.D.S.) is used to FTP the Eta nowcast and forecast fields. wwwgrab, a PERL5 script, is used to take river discharge data from the USGS and water level and associated ancillary data from NOAA/NOS.
Figure 5. Schematics of Nowcast/Forecast Cycles
The prediction system is scheduled to simulate 72 hours - 24 hours in the past (the hindcast mode) and then 48 hours into the future (the forecast mode). A schematic of the flow of data in the prediction system is shown in Figure 5. The hindcast part of the cycle uses observed forcing functions and a file is written out that forms the basis for the next cycle. The forecast simulations are performed using forecasted forcing functions except for the river flows, which are based on daily persistence. The modeling system saves the proper hydrodynamic information for a restart. A smooth and seamless execution occurs to start the next cycle, which is scheduled to start 24 hours later. Three-dimensional fields of salinity, temperature, and currents and two dimensional fields of water level, significant wave height, period and direction are archived every 30 minutes. The forecast for a period 6 hours into the future is shown in Figure 6. This is for the NY/NJ Estuary subdomain and depicts surface current speed and direction. Note that at this time the currents are quite intense at several locations in the Harbor and knowledge of the current patterns is particularly useful for safe navigation. Similar maps are available for the full suite of predicted parameters and for many subdomains of the entire forecast region.
Figure 6. The forecast map for surface current speed and direction in the NY/NJ Estuary subdomain a period 6 hours into the future. The arrows denote the direction of the currents. Speed in knots is denoted on the color bar. Only a subset of the model grid points is shown for clarity.
The most accurate part of the system cycle is the hindcast. It ends at midnight when a file is written out which forms the basis for the next hindcast cycle. For the following 48 hours of the forecast cycle, model accuracy is enhanced through the assimilation of the high-resolution observations of temperature, salinity and water level that are automatically retrieved from the data repository. Every 30 minutes a map of "Present Conditions" is prepared by assimilating all the available data for that 30 minute period with the appropriate forecast. An objective optimal interpolation technique is used to create the assimilated distribution. These distributions are saved for use with the next cycle of the prediction system. Figure 7 illustrates "Present Conditions" of surface salinity in the NY/NJ Estuary for 3 May, 2004, 15:30 local time.
Figure 7. The present surface salinity in the NY/NJ Estuary subdomain. The spatial distribution is created using data from the sensor network available in real-time, the forecasted field available every 30 minutes, and an objective interpolation technique. The dots represent the locations of the sensors.
The observational network and modeling system are parts of the integrated, sustained ocean observing system envisioned by the National Oceanographic Partnership Program (NOPP), under the OCEAN.US office, and the NOAA/ONR Coastal-IOOS (Integrated Ocean Observing System).Back to top
The NYHOPS information distribution system is designed to sustain high-volume data availability and storage, providing rapid access to information by a broad range of users. It is also designed to present the modeling and observational data in creative ways that can be used effectively by governmental, educational and environmental groups for a variety of purposes. The architecture allows the modeling system to interact with the sensor network by turning on a particular sensor, by changing its sampling rate or by dispatching a service crew to effect repairs when necessary.
The in-situ observational data and model forecasts from NYHOPS will be used by the MTS community in the Port of New York and New Jersey, as well as many other institutions and agencies for a wide variety of purposes. It is an important goal of the program to design a system to sustain high-volume data availability and provide the government, commercial, and academic communities rapid access to quality information. A broader goal is to distribute real-time data and product information to the public in multiple formats. The NYHOPS web site mentioned previously has been established to provide for easy distribution of the information. The web site displays are targeted to the MTS community, port security officials, emergency management personnel, teachers, fishermen, sail and power pleasure boaters and swimmers. They include horizontal maps of surface and bottom water temperature, salinity and currents, water levels, waves and winds. Time series of these quantities for selected locations are also presented.Back to top
NYHOPS based Transmission Loss (TL) Forecast System predicts the temporal and spatial properties of the acoustic field by using a high-resolution hydrodynamic model in conjunction with an acoustic wave propagation model. Predictions of TL are made for 48 hours corresponding to the forecast mode of NYHOPS. This system is viewable by selecting the Hudson Longitudinal Section under the region drop-down list.
The range-dependent version of BELLHOP Gaussian Beam Tracing program is implemented (Porter and Bucker, 1987) is used to forecast TL. The program makes use of NYHOPS computed sound velocity profiles at four locations along the propagation path separated by a distance of 300 meters. The locations correspond to 0, 300, 600 and 900 meters from the acoustic source located near Castle Point, New Jersey along the direction of propagation (Figure 8). The acoustic field is represented by a 2-dimensional slice plot of TL as a function of depth and range of propagation. TL is expressed in decibels relative to the source level at 1 meter from the source. The propagation range is fixed at 1000 meters and represents the longitudinal section of Hudson River between Castle Point, New Jersey and the Battery, New York. The frequency of acoustic source is 50 kHz. The acoustic field is generated for three source locations, namely, 3, 6, 9 meters from the water surface in water column of depth 12 meters.Back to top
Blumberg, A. F., L. A. Khan, and J. P. St. John, 1999. (Three-Dimensional Hydrodynamic Model of New York Harbor Region,(J. Hydr. Engrg., 125(8), 799-816.
Blumberg, A. F. and G. L. Mellor, 1987, A Description of a Three-Dimensional Coastal Ocean Circulation Model, In Three-Dimensional Coastal Ocean Models, Coastal and Estuarine Sciences, Vol. 4, N. Heaps ( Ed.) , American Geophysical Union, Washington, D.C., 1-16.
Bruno, M.S., R. Chant, K. Rankin, G. Korfiatis, N. Dimou, S. Glenn, B. Fullerton, P. Burke, A. Pence, R. Hires, E. Creed, C. Haldeman, and E. Hunter, 2002. Hydrodynamic and Water and Sediment Quality Measurements in the Hudson-Raritan Estuary. Presented at the Fall AGU Conference, San Francisco, American Geophysical Union.
Herrington, T.O., M.S. Bruno and K.L. Rankin, 2000. The New Jersey Coastal Monitoring Network: A Real-Time Coastal Observation System, J. Marine Env. Engg. Vol. 6, pp. 69-82.
Kelley, J. G. W., F. Aikman, L.C. Breaker, G.L. Mellor, 1997, Coastal Ocean Forecasts, Sea Technology, May.
Porter, M. B., and Bucker, H. P., 1987. Gaussian Beam Tracing for Computing Ocean Acoustic Fields. J. Acoust. Soc. Am. 82, 1349-1359.Back to top