The Shinnecock Canal Study
Aerial photograph of the Shinnecock Canal, Long Island, NY
The Shinnecock Canal Study was conducted by Karl-Heinz Szekielda, Samir Ahmed, Fred Moshary, Barry Gross, Jorge Peche and Yiping Zhang. Dr. Szekielda is in the The Graduate Center, Earth and Environmental Sciences, of the City of University of New York (CUNY). His colleagues are with CUNY's Optical Remote Sensing Laboratory in the Center for Water Resources and Environmental Research. Dr. Szekielda contributed the following description of this study.
Background: In near-coastal waters with high chlorophyll concentrations and where there is a significant presence of inorganic particulate matter, the algorithms used to calculate chlorophyll concentration and other properties may no longer be valid. The data must be interpreted in qualitative terms, e.g., for patch recognition and processes related to tidal and current changes. The failure of existing algorithms over turbid water can be further attributed to invalid assumptions, such as the assumption that there is no radiation from the water surface in the near-infrared bands at 765-865 nm (Ruddick et al., 2000). Although spectral analysis of reflected light from the open ocean is well understood, the problem and need of interpreting optical properties of near-shore water have been addressed, amongst others, by Bagheri, Zetlin, and Dios (1990).
Field Experiments: The field experiments were designed to determine time frames for the settling of suspended particles under a variety of current velocity and turbulence conditions, to distinguish between settleable and non-settleable fractions in the water column, and observe the effect of settling on reflected spectra.
To conduct these experiments, the Shinnecock Canal of eastern Long Island (New York) was selected as a natural tank in which to observe optical parameters and to estimate the residence time of suspended matter in the water column under the different current speeds and turbulence arising from tidal changes. Inlets with tidal ranges demonstrate large ranges of particulate and dissolved organic and inorganic constituents. They are therefore excellent test areas to relate spectral properties with the varying concentrations of material in suspension and in solution that occur over a tidal cycle.
Landsat image of the Shinnecock Canal
The principal function of the Shinnecock Canal and its operating lock system is to control the flow of water in one direction, from the Peconic Bay to the Atlantic Ocean through Shinnecock Inlet. This prevents the flow of water with low salinity from Shinnecock Bay to Peconic Bay. As a result, the flow of saline water from Peconic Bay to Shinnecock Bay flushes through the Shinnecock gate with a particle load originating primarily from one major source area.
In the field experiments, spectral reflectance measurements were carried out using a spectroradiometer (GER 1500) covering the UV, visible, and near IR at wavelengths from 0.35 to 1.05 µm. For final analysis, however, only the 400-850 nm spectral region was selected for further processing. The spectroradiometer uses a diffraction grating with a silicon diode array with 512 discrete detectors. It includes a memory for stand-alone operation as well as capability for computer-assisted operation. A total of 483 spectral readings can be stored for subsequent downloading and analysis using a personal computer with a standard serial port and GER operating software.
Computer-based operation allows for real-time display and data analysis. In the experiments, upwelling radiance was monitored through a calibrated fiber cable positioned in a down-looking micro-buoy. Results are presented as a percent of the incident solar irradiance. Salinity, temperature, pH and turbidity were collected using a Hydrolab H20 multi-sensor simultaneously with the spectroradiometer readings. Instead of using an organic dye (formazine) for calibrating the turbidity measurements, however, calibration was carried out with a montmorillonite suspension, since the optical properties of clays are closer to the spectral behavior of Total Suspended Sediments (TSS) in coastal regions. Correlation of varying concentrations of suspended matter with turbidity (in NTU) and spectral reflectance data, showed good reproducibility and are in excellent agreement with data published more recently.
Turbidity Measurements: Figure 1 shows the turbidity variations measured over a lunar month in the tidal gates of Shinnecock Canal. These measurements cover the period 22 September—21 October, 2000. Time is given in Julian days, and the vertical axis represents the concentration of suspended sediment in milligrams measured against a standard montmorillonite suspensions. These field experiments demonstrate the occurrence of rapid turbidity changes, and also indicate that the settling of TSS is highly dependent on the tidal stage and the residence time (reduced turbulence) of entrapped water after the tidal gate is closed.
Figure 1: Suspended matter concentration expressed in terms of mg montmorillonite per liter
Spectral Measurements: Spectral measurements of upwelling radiance were taken at different time intervals (before, during and after closing of the tidal gate). These measurements confirm the fast settling of particles during reduced turbulence conditions which occur while the sampled water parcel is trapped during closed gates. Figure 2 shows reflected spectra measured for different settling times in the tidal lock. The data represent averaged spectra for different days with a standard deviation of less than 1% reflectance.
Figure 2: Reflectance spectra of suspended matter in the Shinnecock Canal. Times shown refer to canal lock events: AC=After Closing, BC=Before Closing, AO=After Opening.
For each spectral reflectance measurement corresponding to a specific water condition, 32 complete spectra were sampled and averaged. To check reproducibility of data, this step was repeated and the standard deviation calculated throughout the spectrum. Figure 3 shows the high degree of reproducibility attained.
Figure 3: Standard deviation of reflectance spectra measurements
Examination of the reflected spectra show that the visible reflectance is typically reduced by about 50% after particulate matter in an entrapped water parcel settles for 4 hours. This reduction is seen in the reflectance spectra for 550-600 nm wavelengths. In general, the data show that reflectance decreases with decreasing turbidity throughout the monitored spectral regions. In fact, reflectance, turbidity, and concentration of suspended material are correlated throughout the spectral range of 400 to about 850 nm. This is in good agreement with laboratory results reported by Bhargava and Mariam (1990), who showed that for the spectral region 700-900 nm, high correlation and low standard errors existed between these variables for the clay material used in suspension in their experiments. Our laboratory measurements over a continuous spectral range from 400-860 nm using a miniature fiber-optic spectrometer equipped with a high sensitivity CCD detector also confirm the correlation of reflectance over the entire wavelength range with changes in the water constituents and their concentration.
Conclusions: The results of the field experiments discussed above demonstrate the importance of taking into account the separation of settleable and non-settleable TSS for the interpretation of ocean color satellite imagery over coastal regions. This analysis is required in order to resolve the time-and-space relationships of the TSS distribution patterns. For future work, revisits by satellites spaced as closely as one hour may be required in order to better understand the coastal dynamics of suspended sediments.
Bagheri, S., C. Zetlin, and R. Dios (1999). Estimation of optical properties of near-shore water. Int. J. Remote Sensing, 20, 3393-3397.
Bhargava, D.S., and D.W. Mariam (1990). Spectral reflectance relationship to turbidity generated by different clay materials. Photogrammetric Engineering and Remote Sensing, 56, 225-229.
Ruddick, K.G., F. Ovido, and M. Rijkeboer (2000). Atmospheric correction of SeaWiFS imagery for turbid coastal and inland waters. Applied Optics, 39, 897-912.
<< Back | Next >>
Return to Index