1. Introduction
Marine bio-optical modeling, the validation of ocean color products, and the vicarious calibration of optical space sensors all require accurate in-water radiance and irradiance measurements. Light perturbation by the instrument itself, or by the large structures associated with the deployment of optical instruments (e.g., ships, offshore towers, moored buoys, etc.), can significantly reduce the accuracy of in-water radiometric measurements. Instrument self-shading, for example, can increase measurement uncertainty from a few percent to several tens percent as a function of wavelength, instrument radius, and illumination conditions (Gordon and Ding 1992). For such an effect, a correction scheme based on the absorption of seawater, instrument radius, sun zenith angle, and the ratio between diffuse and direct solar irradiances was proposed by Gordon and Ding (1992) and experimentally validated by Zibordi and Ferrari (1995).
In addition to instrument self-shading, various studies were performed to evaluate uncertainties induced in underwater optical measurements by large structures, such as ships. Investigations carried out by different authors (Gordon 1985; Voss et al. 1986; Helliwell et al. 1990; Weir et al. 1994; Saruya et al. 1996) indicate that radiance and irradiance measurement uncertainties increase substantially when the deployment distance of the instrument from the stern of a ship is reduced. To minimize the ship-induced perturbation, a simple method was suggested by Mueller and Austin (1995) to estimate the minimum ship distance from the instrument deployment point as a function of the diffuse attenuation coefficient during clear-sky conditions.
The relatively infrequent use of offshore towers for in-water radiometric measurements is the reason for the dearth of studies on tower perturbation effects (Kearns et al. 1996). Consequently, minimization schemes for tower shading effects have not yet been developed, nor has it been assessed that the ship-based methods are applicable to a particular tower—the superstructure of a tower can be much different than a research vessel.
The need for quantifying the tower shading effects on optical measurements taken in the northern Adriatic Sea near the Acqua Alta Oceanographic Tower (AAOT) in support of calibration and validation activities (Zibordi et al. 1995) is the rationale for the present study. The major objectives of the work are the evaluation of (a) the experimental shading effects induced in in-water radiance and irradiance measurements by the tower, and (b) the capability of theoretically modeling the tower shading effects as a first step toward the development of an operational scheme for the correction of in-water optical measurements collected at the tower site.
2. Description of the experiment
An AAOT field experiment was designed to estimate shading effects on in-water radiometric data collected in close proximity to the tower. Sequential optical profiles taken with the Low Cost NASA Environmental Sampling System (LoCNESS) were used to determine the shading effects on downwelling irradiance, Ed(χ, z, λ, t), and upwelling radiance, Lu(χ, z, λ, t), where χ is the distance of the profiler from the tower, z is the water depth (or height above water), λ is the wavelength, and t is time. Concurrent with LoCNESS data, incident surface (z = 0+ m) solar irradiance data (at χ = 0 m), Es(0, 0+, λ, t), were collected with a Multichannel Visible Detector System (MVDS). These incident irradiance measurements were used to normalize the underwater radiometric data to account for changes in the ambient light field during data collection. Underwater optical data were also taken with the Wire-Stabilized Profiling Environmental Radiometer (WiSPER) system, which is regularly used on the tower to collect radiance and irradiance profiles during measurement campaigns supporting ocean color calibration and validation activities.
a. The oceanographic tower
The AAOT is located in the northern Adriatic Sea (45.31°N, 12.51°E) approximately 15 km southeast of the city of Venice. The average water depth immediately below the tower is 17 m, and the composition of the nearby seafloor is primarily sand and silt. The tower was built in 1975 and is owned and operated by the Institute for the Study of Large Masses of the Italian Consiglio Nazionale delle Ricerche (National Research Council) in Venice. The tower is composed of four levels supported by four large pillars. Each level is approximately 7.2 m × 5.2 m in size with the exception of the lowest level, which is 5.2 m × 5.2 m.
The lowest tower level, about 4.5 m above the water, has an open grid deck and no facilities. The second level is approximately 7 m above the water and contains a workshop, two 12-kW diesel-powered electrical generators, a portable scientific laboratory, and storage spaces for a large complement of lead–acid batteries, fuel tanks, etc. At this level, a special open grid platform, 3.5 m wide, extends 6.5 m out over the sea toward the southeast. The platform provides mounting points for instruments to be deployed above, or into, the sea. The third deck contains the primary accommodation space. The fourth (uppermost) deck, at about 13 m above the water, contains a wind generator, solar panels, a variety of meteorological instruments, communications antennas, plus freshwater and seawater storage tanks.
b. The optical instruments
LoCNESS is a free-falling instrument based on the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) Optical Profiling System (SeaOPS) composed of an above-water and in-water set of sensors, and several subsystems (Hooker and McClain 1999). The in-water optical sensors are a downward-looking radiance sensor, which measures Lu(χ, z, λ, t), and an upward-looking irradiance sensor, which measures Ed(χ, z, λ, t). The sensor used for upwelling radiance measurements is a Satlantic OCR-200 sensor (S/N 021), and the sensor used for downwelling irradiance measurements is an OCI-200 sensor (S/N 029). All of the sensors send their analog signals to an underwater data unit, a Satlantic DATA-100 (S/N 004), which digitizes the analog signals (16 bits) and converts the counts to RS-485 serial communications.
In the LoCNESS configuration, the DATA-100 and the two light sensors are connected in line using extension brackets with the OCR-200 at the nose (pointing downward to measure Lu) and the OCI-200 at the tail (pointing upward to measure Ed). The addition of weight to the nose bracket and buoyant (foam) fins to the tail bracket produces a balanced package that falls through the water column with minimum tilt (less than 2°). The power and telemetry cable to the profiler extends through the field of view (FOV) of the irradiance sensor, but the small diameter of the cable (7 mm) minimizes any negative effect on the measured light field.
The above-water part of SeaOPS, a Satlantic MVDS, measures Es(0, 0+, λ, t). The MVDS unit (S/N 009) is composed of an OCI-200 irradiance sensor (S/N 030) packaged with a separate DATA-100 module that converts the analog output of the OCI-200 radiometer to RS-485 serial communications. For the shading experiments, the DATA-100 and OCI-200 units were mounted on a pole that was sited on the uppermost deck of the AAOT. The pole was long enough to ensure the tower superstructure did not shadow the irradiance sensor under almost all illumination conditions.
The RS-485 signals from the two DATA-100 units are combined in a Satlantic deck box, the PRO-DCU (S/N 023), and converted to RS-232 communications for computer logging. The deck box also provides the (computer controlled) power for all the sensors and is designed to avoid instrument damage due to improper power-up sequences over varying cable lengths. The RS-232 data were logged on two separate Macintosh PowerBook computers using software developed at the University of Miami Rosenstiel School for Marine and Atmospheric Science (RSMAS) and the SeaWiFS Project of the National Aeronautics and Space Administration’s Goddard Space Flight Center (NASA/GSFC). The software controls the logging and display of the data streams as a function of the data collection activity being undertaken: dark data (caps on the radiometers), down cast, constant depth soak, up cast, etc. All of the telemetry channels are displayed in real time, and the operator can select from a variety of plotting options to visualize the data being collected.
The WiSPER system is permanently installed on the tower and is operated from the 6.5-m platform extension on the second level. WiSPER uses a custom-built profiling rig, and the positioning of the equipment on the rig was developed with a geometry that ensures all of the radiometers do not view any part of the mechanical supports. The radiometers are mounted on a 1-m extension boom, which places them approximately 7.5 m from the nearest tower leg. Two taut wires, anchored between the tower and the sea bottom, prevent the movement of the rig out of the vertical plane defined by the wires. The narrow geometry of the rig was designed to provide a minimal optical cross section. The FOV of the irradiance sensor is obstructed by the power and telemetry cable, as well as the stabilization wires, but all of these have very small cross sections that minimize the impact on measurements.
Careful attention was paid to the rigidity and stability of the rig, so tilt or roll sensors are not needed. WiSPER uses the same kind of optical sensors as LoCNESS: one OCI-200 (S/N 71) to measure Ed(7.5, z, λ, t) and one OCR-200 (S/N 46) to measure Lu(7.5, z, λ, t). The center wavelengths for the radiometers used with all of the optical systems are given in Table 1. The commonality in spectral properties for the optical devices used in the experiment was an important part of the equipment selection criteria. A special kind of DATA-100 (S/N 5) provides the analog-to-digital (A/D) and telemetry capability for the WiSPER instruments. The equipment is powered directly from 12-V lead–acid batteries, which are stored and kept charged on the tower. WiSPER is raised and lowered from the southeastern side of the tower. The typical lowering and raising speed of the winch used is approximately 0.1 m s−1.
c. The experimental methodology
The experimental setup began with siting a black buoy approximately 30 m from the southern tower leg;the buoy was aligned perpendicularly to the southeastern side of the tower and displaced approximately 2 m to the side of the WiSPER instrument. Two pulleys were then attached—one to the buoy and the other to the tower—and a closed loop of line (60 m long) with marks on it every 2.5 m was run through the pulleys. A cable ring was linked to one of the cable marks that defined the current position of the profiler (the power and telemetry cable for the profiler passed through the ring). The ring and, thus, the profiler, was moved to a selected distance from the tower leg by pulling on the closed loop of line until the desired number of cable marks between the ring’s position and the tower leg was achieved. A schematic of the experimental setup is given in Fig. 1.
LoCNESS is sufficiently easy to handle that one person can deploy it. Under normal circumstances, the handler keeps a few coils of the power and telemetry cable in the water, so the profiler can fall freely through the water column; once the desired depth has been reached, the cast is terminated and the profiler is pulled back to the surface (data are only logged during the down cast). For the tower deployments, the profiler was slowly lowered by hand to control the descent rate (approximately 0.4 m s−1). A cable block, which could not pass through the cable ring, was used to prevent the profiler from going deeper than 15 m and accidentally impacting the seafloor (which might damage the radiometer in the nose of the instrument).
An experiment was defined as a sequence of profile deployments going away or toward the platform during a relatively short period of time (typically 20 min). For the tower deployments, 10 different experiments were conducted during almost clear-sky conditions: the first 9 with variable LoCNESS deployment distances (ranging from 2.5 to 22.5 m) with respect to the tower, and the last 1 with the deployment distance fixed at 7.5 m (the same distance the WiSPER measurements were made with respect to the tower).
At the end of the experiments, the LoCNESS and WiSPER optical sensors were intercompared by mounting them on the WiSPER frame in pairs: first, both OCR-200 sensors were used to collect coincident Lu(7.5, z, λ, t) data, and then both OCI-200 sensors were used to collect Ed(7.5, z, λ, t) coincident data.
3. Numerical modeling
The presence of a deployment structure at the measurement site introduces abrupt medium changes within a predominantly plane-parallel system, causing light field inhomogeneities in all three spatial dimensions. Consequently, the simulation of tower shading effects on in-water radiometric measurements requires a fully three-dimensional (3D) geometry description and radiative transfer modeling of the ocean–atmosphere system. In this fully 3D problem, the radiative transfer processes can be modeled with stochastic simulation methods to achieve a solution of the generalized radiative transfer equation (RTE). In particular, backward Monte Carlo (MC) techniques should be applied. These techniques, compared to forward MC techniques, produce much lower variances and concentrate only on significant contributions to the simulated light field.
a. Principles
The Photon-Transport (PHO-TRAN) MC code—jointly developed at the University of London, Imperial College of Science, Technology and Medicine (IC) together with the Joint Research Centre (JRC) of the Commission of the European Communities (CEC)—has been previously validated for the ocean–atmosphere system (Doyle and Rief 1998) and expanded to include backward MC techniques. This code has been employed in this work to assess the tower shading perturbations on in-water radiometric measurements. Algorithms, based on Case’s reciprocity relationship (Case 1957) and Gordon’s implementations (Gordon 1985), were introduced in the PHO-TRAN backward MC code to compute radiances and irradiances at a specific point in the modeled system.
Within the simulations, the ocean–atmosphere system is modeled on a 3D grid. The grid delimits the largest macroscopic volumes (cells) containing medium of uniform optical properties. In each cell of the grid, the optically active components (air or water molecules, hydrosols, aerosols, etc.) are specified, and their spectral inherent optical properties (IOPs) are assigned (e.g., the beam attenuation coefficient c, the single scattering albedo ω0, and the scattering phase function P). Cell boundaries are spectrally characterized by transmittance and reflectance, and by the associated transmission and reflection angular distribution functions (ADFs). Specific ADFs are also defined both for the source angular emission and for the detector angular collection. The air and water refractive indices are assigned to the atmosphere and ocean, respectively.
Photons detected by a radiometer are a fraction of those emitted by the sun and reach the sensor after absorption, scattering, reflection, and refraction processes in the ocean–atmosphere system. The properties of physical processes characterizing the propagation of photons are time reversal invariant; hence the inverse course of events can be applied in simulations. This principle is the basis for the backward MC technique. Random-walk photon trajectory sampling is performed the same in backward MC simulations, as in forward MC simulations, but contributions to the detected signal are deterministically computed ensuring high computational efficiency.
In the backward MC scheme, photons, initially having unitary statistical weight, are released from the detector within its FOV according to the predefined ADF. A free-flight optical distance to the next collision point is sampled (Lux and Koblinger 1991), possible cell-boundary crossing processes are considered to define flight direction modifications, and finally, photon trajectory is computed taking into account possible changes in IOP values along the trajectory. At the collision point (defined as the point where the sampled optical distance is exhausted), a scatterer is sampled; photon survival probability, given by ω0, is computed (accounting for absorption a); and the flight direction of the surviving photon is determined by retrieving the scattering angle from a random sampling of P. The latter is adequately modeled into an equal probability interval table (Lux and Koblinger 1991).
The probability of a backtracked photon propagating toward the sun is virtually zero; therefore, at collision and boundary crossing points, photon contributions to the detected signal are deterministically computed. This is done by calculating the probability that a photon scatters in directions leading to the sun and evaluating the associated weight reduction, given both by attenuation along the traveled path and by possible absorption at interface transmission. If this virtual photon encounters a purely absorbing medium (i.e., the tower structure), its contribution to the detected signal is zero.
By tracking so-called twin photons, one interacting with the tower and the other not, a correlated sampling scheme (Spanier and Gelbard 1969) is produced, which minimizes the score variance of differences between tower-perturbed and unperturbed signals. The number of initiated twin photons defines the estimated statistical relative error on simulated data.
b. Simulations
The simulation frame consisted of one large 3D box that enclosed the grid defining the geometrical features of the problem. The atmosphere, the ocean, and the boundaries (top of the atmosphere, sea surface, and seafloor) were modeled as horizontally plane-parallel. This plane-parallel symmetry was broken by introducing, at a specific location within the reference frame, a geometrical object schematically representing the tower with completely absorbing surfaces. The reference system for the simulation frame used Cartesian orthonormal coordinates (x, y, z), centered at a tower leg. Figure 2 shows the schematic of the tower structure and the relevant 3D features introduced in the geometry description of the PHO-TRAN code. It is important to note the presence of the platform, which is used for routine deployment of the optical instruments and stretches out from the main structure of the tower. Moreover, in this section, the modeling is described from the forward MC perspective, even though backward MC is the actual technique adopted in tracking and tallying photons through the modeled ocean–atmosphere system.
The direct source (i.e., the source in a forward MC scheme perspective—the sun) was represented by a parallel beam of photons uniformly incident at the top of the atmosphere. The ADF for the source emission was formulated by a δ Dirac centered on the sun zenith θ0 and sun azimuth ϕ0 angles. Radiometers, located at specific points in the reference frame, were described by their FOV and the associated photon-collection ADF. Radiances were simulated assuming an in-water 20° FOV (as per the OCR-200 specifications) with a unitary collection ADF; irradiances were simulated assuming a 2π sr FOV with a cosine collection ADF. Downwelling irradiances and upwelling radiances were simulated for intercomparison with field data taken at the tower site. Simulations were performed at different wavelengths (coinciding with the center wavelengths of the radiometers), at different depths, and at different distances from the tower legs.
The atmosphere was modeled by 50 plane-parallel and homogeneous layers 1 km thick, and was stratified according to the Elterman (1968) atmospheric model. The individual ozone, molecular, and aerosol relative concentration profiles were weighted by the respective optical thickness as derived climatologically for ozone (Vigroux 1953; Robinson 1966; and Leckner 1978), theoretically for molecules (Frölich and Shaw 1980; Young 1980), and experimentally for aerosols. In particular, the aerosol optical thicknesses τA at the wavelengths of interest were obtained by applying the Ångström law (Ångström 1961) to measurements taken during the experiment at 440, 500, and 670 nm with a sun photometer (CIMEL CE-318). The optical thicknesses of the atmospheric components were then used to compute c and ω0 for each atmospheric layer. The Rayleigh molecular scattering phase function and the Gordon and Castaño (1987) marine aerosol scattering phase function were adopted.
The sea surface was assumed flat, and the incident light was reflected and refracted according to the Fresnel and Snell laws, respectively. The extraterrestrial solar irradiance was assumed to be a constant in the computations (1.0 W m−2 nm−1). An absolute quantity is not needed, because the field data and simulated data intercomparison relies on radiance or irradiance ratios to derive tower shading perturbations. The ADF for seafloor reflection was assumed to be Lambertian. The seafloor reflectance Rb was derived by performing the ratio of upwelling to downwelling irradiance, as provided by ad hoc measurements taken close to the seafloor.
The water column was modeled by N plane-parallel, homogeneous layers (where N is the number of significant stratification layers chosen for the simulations). Profiles of c and a, taken with an attenuation absorption meter (WET-LABS AC9) operated during the experiment, were used to define ω0 = (c − a)/c for each water layer. The Rayleigh and Gordon KA (Gordon et al. 1975) scattering phase functions were used for water molecules and particulate matter, respectively.
4. Comparison of the in situ and numerical results
The depth coregistration of optical data was achieved by data binning. The nEd(χ, z, λ, t) and nLu(χ, z, λ, t) profile data were binned at depths zi using depth intervals of 0.5 m (ranging from zi + 0.25 m to zi − 0.25 m) resulting in nEd(χ, zi, λ, ti) and nLu(χ, zi, λ, ti) values, where ti is the average time for each binning interval associated with depth zi.
The analysis of nEd(χ, zi, λ, ti) and nLu(χ, zi, λ, ti) has shown that for several experiments the shading effects were masked out by the presence of mucilage (Degobbis et al. 1995; Molin et al. 1992), which was clearly observed during many of the optical deployments. The mucilage, in the form of highly scattering, inhomogeneous, thick layers, of gelatinous aggregates (Berthon et al. 1999, submitted to Limnol. Oceanogr.), induces spatial and temporal variability in the in-water light field. This is clearly shown in Fig. 3 for the LoCNESS nEd(χ, zi, λ, ti) and nLu(χ, zi, λ, ti) profiles, which were obtained in almost ideal conditions with no evidence of mucilage (Fig. 3a) and in the presence of mucilage (Fig. 3b). The data analysis is restricted to 443, 555, and 665 nm, assumed as representative wavelengths of the visible spectrum.
The multiple profiles taken in the absence of mucilage and at a variety of distances from the tower (Fig. 3a) show almost regular and correlated variations; in comparison, the multiple profiles taken in the presence of mucilage (Fig. 3b) show irregular variations with respect to both distance and depth. A careful examination of the nEd(χ, zi, λ, ti) values indicates higher noise in the surface data between 0- and 4-m depth for the former than for the latter. This is a consequence of the more pronounced wave effects induced in Fig. 3a profiles by a sea state of about 2 (code “S” units), which was higher than the sea state of approximately 1 observed during the collection of the Fig. 3b profiles. Further examination of nLu(χ, zi, λ, ti) in Fig. 3a clearly shows the bottom effects that, becoming more pronounced on depth increase, induce a change in the slope of profiles. This change is not appreciable at 443 nm but appears significant at 555 nm and 665 nm in agreement with the relatively low Rb at 443 nm, and the much higher Rb values at 555 nm and 665 nm (see Table 2). Different from Fig. 3a, the bottom effects on nLu(χ, zi, λ, ti) profiles shown in Fig. 3b are masked out by the scattering of mucilage layers.
For the shading analysis presented below, only the data in Fig. 3a (i.e., not contaminated by mucilage) were used.
To minimize wave effects—which can induce noise of the order of several percent in surface values even after fitting of profile data (Mueller and Austin 1995)—as well as to avoid intervals close to the bottom where data were not available, the analysis of
A plot of
The
A remarkable agreement between simulated and experimental data is observed for
5. Discussion
In the following sections, accounting for the satisfactory results obtained from the intercomparison between simulated and experimental data, a theoretical analysis of the tower shading effects is presented as a function of typical environmental conditions at the AAOT measurement site. In addition, an intercomparison of LoCNESS and WiSPER data is briefly presented to demonstrate the equivalence of the tower shading effects on measurements taken with the two devices working in identical geometric and environmental conditions.
a. Theoretical analysis of the tower shading effects
The intercomparison between experimental and PHO-TRAN simulated data has shown mean percentage differences α of 1.0% and 0.8% for irradiance and of 1.7% and 1.0% for radiance at 443 and 555 nm, respectively. Although code assessment with field data is restricted to a single experiment (because of the difficulty in performing the needed measurements under the required environmental conditions), the result of the intercomparison suggests that PHO-TRAN is a viable code for the evaluation of the AAOT shading errors on underwater radiance and irradiance collected at various geometric and physical conditions.
Simulated downwelling irradiance errors,
Analysis of the data in Table 4 shows values almost independent of wavelength for both
Values of
b. Intercomparison of WiSPER and LoCNESS data
In order to demonstrate that the tower shading effects estimated using the LoCNESS profiler were consistent with WiSPER data (i.e., that radiometric measurements performed with LoCNESS and WiSPER were equally influenced by tower shading), sample profiles obtained simultaneously with the two systems operating at 7.5 m from the tower legs were intercompared.
The first step in the radiometric intercomparison was the intercalibration of the OCI-200 and OCR-200 instruments used with LoCNESS and WiSPER. Intercalibration factors were obtained using the profiles collected with the two radiance sensors, and then with the two irradiance sensors, installed at the same time on the WiSPER system. By assuming that the active radiometers were measuring the same light field, the intercalibration factors for the two OCR-200 sensors were obtained by forming the ratio between the two radiance profiles on a depth-binned basis; the intercalibration factors for the two OCI-200 sensors were obtained in an analogous fashion using the irradiance profile data. The derived calibration factors were applied to LoCNESS and WiSPER data for intercomparing Ed and Lu obtained with the two profiling systems operating at the same time at 7.5 m from the tower. Average percentage differences obtained from binned LoCNESS and WiSPER profile data are within 2% for irradiances and within 3% for radiances. The differences are explained by the presence of mucilage during the intercomparison casts. A better agreement is expected under normal measurement conditions, that is, in the absence of inhomogeneous scattering layers in the water column. Thus, it is assumed that any conclusion on tower shading effects achieved from the analysis of LoCNESS data is also correct for WiSPER data.
6. Conclusions
The sequences of irradiance and radiance profiles taken at varying distances from the AAOT to quantify the tower shading perturbations on underwater optical measurements have shown very appreciable shading effects at 443 and 555 nm, while reflections from the tower surfaces and noise induced by surface waves have largely masked out the shading effects at 665 nm. At the 7.5-m deployment distance regularly used on the AAOT for the collection of underwater optical measurements, the shading effect at a depth of 7 m and at a wavelength of 443 nm with a relatively low sun zenith angle of 22° has been quantified within 3% for downwelling irradiance and within 8% for upwelling radiance. For the specific field experiment, the shading effects have been shown to be negligible for both Ed and Lu at deployment distances greater than 15 and 20 m, respectively.
The PHO-TRAN code, assessed with the experimental data, was used to extend the analysis to AAOT shading effects for measurements taken at a depth of 0− m and at a deployment distance of 7.5 m during different illumination conditions. Simulations, performed assuming seawater IOPs representative of the AAOT site, showed that the measurement conditions minimizing the tower perturbations are those performed on very clear sky conditions with θ0 between 20° and 60°. In such a case, for the simulated conditions, the error is lower than approximately 4.0%, 1.5%, and 0.5% for irradiance and lower than approximately 4.5%, 2.5%, and 0.6% for radiance at 443, 555, and 665 nm, respectively.
Simulations have also shown, however, that the shading error can be as high as approximately 20% and almost independent of wavelength for both irradiance and radiance measurements taken during overcast sky conditions. Then, actual AAOT shading errors at a depth of 0− m and 7.5 m from the tower (for the average seawater IOPs characterizing the site) will result in a combination of the case studies presented above. Because of this, given that the SeaWiFS Project requires radiometric field data with a measurement uncertainty below 5% (Hooker and Esaias 1993), a correction scheme to minimize the tower shading effects in optical measurements taken near the tower is needed if the data are to be used for ocean color validation activities.
This study suggests the development of an operational correction scheme based on lookup tables of AAOT shading correction factors, computed with the PHO-TRAN code for a wide range of measurement conditions identified by a minimum set of parameters collected simultaneously as the in-water optical profiles. The adoption of lookup tables, making the correction of AAOT shading effects independent from the relatively high MC computational time required by PHO-TRAN, would make possible real-time correction of the optical in-water measurements.
Acknowledgments
The authors acknowledge Dirk van der Linde for his support in preparing the tower shading experiment and during its execution, Jean-François Berthon for providing the AC-9 processed data and contributing to helpful discussions, Stephane Maritorena for providing comments on an early draft of the paper, and Elaine Firestone for the editorial revision of the manuscript. The authors would also like to thank Luigi Alberotanza for ensuring the use of the tower and Pierluigi Cova, Armando Penzo, and Narciso Zennaro for the logistic support provided during the field work.
REFERENCES
Ångström, A., 1961: Techniques of determining the turbidity of the atmosphere. Tellus,13, 214–223.
Case, K. M., 1957: Transfer problems and the reciprocity principle. Rev. Mod. Phys.,29, 651–663.
Degobbis, D., S. Fonda-Umani, P. Franco, A. Malej, R. Precali, and N. Smodlaka, 1995: Changes in the northern Adriatic ecosystem and the hypertrophic appearance of gelatinous aggregates. Sci. Total Environ.,165, 43–58.
Doyle, J. P., and H. Rief, 1998: Photon transport in 3D structures treated by random walk techniques: Monte Carlo benchmark of ocean color simulations. Math. Comput. Simul.,47, 215–241.
Elterman, L., 1968: UV, visible, and IR attenuation for altitudes to 50 km. Rep. AFCRL-68-0153, U.S. Air Force Cambridge Research Laboratory, Bedford, MA, 56 pp.
Frölich, C., and G.E. Shaw, 1980: New determination of Rayleigh scattering in the terrestrial atmosphere. Appl. Opt.,19, 1773–1775.
Gordon, H. R., 1985: Ship perturbation of irradiance measurements at sea. Part 1: Monte Carlo simulations. Appl. Opt.,24, 4172–4182.
——, and D. J. Castaño, 1987: The Coastal Zone Color Scanner atmospheric correction algorithm: Multiple scattering effects. Appl. Opt.,26, 2111–2122.
——, and K. Ding, 1992: Self-shading of in-water optical instruments. Limnol. Oceanogr.,37, 491–500.
——, O. B. Brown, and M. M. Jacobs, 1975: Computed relationships between the inherent and apparent optical properties of a flat homogeneous ocean. Appl. Opt.,14, 417–427.
Helliwell, W. S., G. N. Sullivan, B. Macdonald, and K. J. Voss, 1990:Ship shadowing: Model and data comparison. Proc. Ocean Optics X, Orlando, FL, SPIE, 55–71.
Hooker, S. B., and W. E. Esaias, 1993: An overview of the SeaWiFS project. Eos, Trans. Amer. Geophys. Union,74, 241–246.
——, and C. R. McClain, 1999: A comprehensive plan for the calibration and validation of SeaWiFS data. Progress in Oceanography, Pergamon Press, in press.
Kearns, E., R. Riley, and C. Woody, 1996: A bio-optical time series collected in coastal waters for SeaWiFS calibration and validation: Large structure shadowing considerations. Proc. Ocean Optics XIII, Halifax, NS, Canada, SPIE 2963, 697–702.
Kirk, J. T. O., 1994: Estimation of the absorption and the scattering coefficients of natural waters by use of underwater irradiance measurements. Appl. Opt.,33, 3276–3278.
Leckner, B., 1978: The spectral distribution of solar radiation at the Earth’s surface—Elements of a model. Sol. Energy,20, 143–150.
Lux, I., and L. Koblinger, 1991: Monte Carlo Transport Methods; Neutron and Photon Calculations. CRC Press, 517 pp.
Molin, D., E. Guidoboni, and A. Lodovisi, 1992: Mucilage and the phenomena of algae in the history of the Adriatic: Periodization and the antrophic context (17th–20th centuries). Proc. of the International Conf. Marine Coastal Eutrophication, R. Avollenweider, R. Marchetti, and R. Viviani, Eds., Elsevier, 511–524.
Mueller, J. L., and R. W. Austin, 1995: Ocean optics protocols for SeaWiFS validation, Revision 1. NASA Tech. Memo. 104566, Vol. 25, S. B. Hooker, E. R. Firestone, and J. G. Acker, Eds., NASA/Goddard Space Flight Center, Greenbelt, MD, 66 pp.
Robinson, N., 1966: Solar Radiation. Elsevier, 347 pp.
Saruya, Y., T. Oishi, M. Kishino, Y. Jodai, K. Kadokura, and A. Tanaka, 1996: Influence of ship shadow on underwater iradiance fields. Proc. Ocean Optics XIII, Halifax, NS, Canada, SPIE 2963, 760–765.
Spanier, J., and E. M. Gelbard, 1969: Monte Carlo Principles and Neutron Transport Problems. Addison-Wesley, 234 pp.
Vigroux, E., 1953: Contribution à l’étude expérimentale de l’absorption de l’ozone. Ann. Phys.,8, 709–762.
Voss, K. J., J. W. Nolten, and G. D. Edwards, 1986: Ship shadow effects on apparent optical properties. Proc. Ocean Optics VIII, Orlando, FL, SPIE, 186–190.
Weir, C. T., D. A. Siegel, A. F. Michaels, and D. W. Menzies, 1994:In situ evaluation of a ship’s shadow. Proc. Ocean Optics XII, Bergen, Norway, SPIE 2258, 815–821.
Young, A. T., 1980: Revised depolarization corrections for atmospheric extinction. Appl. Opt.,19, 3427–3428.
Zibordi, G., and M. Ferrari, 1995: Instrument self-shading in underwater optical measurements: Experimental data. Appl. Opt.,34, 2750–2754.
——, V. Barale, G. M. Ferrari, N. Hoepffner, L. Alberotanza, P. Cova, and C. Ramasco, 1995: Coastal Atmosphere and Sea Time-Series project (CoASTS): An ocean color remote sensing calibration–validation project. Proc. Third Thematic Conf. Remote Sensing for Marine and Coastal Environments, Vol. II, 96–100.
APPENDIX A
Symbols
APPENDIX B
Acronyms
AAOT Acqua Alta Oceanographic Tower
A/D Analog-to-Digital
ADF Angular Distribution Function
CEC Commission of the European Communities
FOV Field of View
GSFC Goddard Space Flight Center
IC Imperial College of Science, Technology and Medicine
IOP Inherent Optical Property
JRC Joint Research Centre
LoCNESS Low Cost NASA Environmental Sampling System
MC Monte Carlo
MVDS Multichannel Visible Detector System
NASA National Aeronautics and Space Administration
PHO-TRAN Photon-Transport code
RSMAS Rosenstiel School for Marine and Atmospheric Science
RTE Radiative Transfer Equation
SeaOPS SeaWiFS Optical Profiling System
SeaWiFS Sea-viewing Wide Field-of-view Sensor
S/N Serial Number
3D Three-dimensional
WiSPER Wire-Stabilized Profiling Environmental Radiometer
A schematic of the technical setup at the AAOT showing the equipment and its relative positioning with respect to one another. The WiSPER instrumentation is shown in a slightly distorted view so all components are visible; in reality, the two radiometers and the two stabilizing wires are aligned in a plane perpendicular to the page, which means the light data are collected approximately 7.5 m away from the tower leg. Details of the LoCNESS instrument are shown in the inset panel.
Citation: Journal of Atmospheric and Oceanic Technology 16, 11; 10.1175/1520-0426(1999)016<1767:OTSEOI>2.0.CO;2
Schematic of the tower as defined in the PHO-TRAN Monte Carlo code. The south–north direction is rotated 135° with respect to the positive x axis. The surfaces obstructing photons are shown in gray.
Citation: Journal of Atmospheric and Oceanic Technology 16, 11; 10.1175/1520-0426(1999)016<1767:OTSEOI>2.0.CO;2
LoCNESS profiles of nEd and nLu data in the first six panels (a) ideal conditions, and in the last six panels (b) in the presence of mucilage. Profiles are given at 443-, 555-, and 665-nm center wavelengths and for different distances from the tower. The profiles in (a) were taken with θ0 between 22° and 23° and ϕ0 between 182° and 194°; the profiles in (b) were taken with θ0 between 43° and 47° and ϕ0 between 101° and 105°.
Citation: Journal of Atmospheric and Oceanic Technology 16, 11; 10.1175/1520-0426(1999)016<1767:OTSEOI>2.0.CO;2
Irradiance and radiance relative variations
Citation: Journal of Atmospheric and Oceanic Technology 16, 11; 10.1175/1520-0426(1999)016<1767:OTSEOI>2.0.CO;2
Computed shading errors
Citation: Journal of Atmospheric and Oceanic Technology 16, 11; 10.1175/1520-0426(1999)016<1767:OTSEOI>2.0.CO;2
Computed shading errors
Citation: Journal of Atmospheric and Oceanic Technology 16, 11; 10.1175/1520-0426(1999)016<1767:OTSEOI>2.0.CO;2
Nominal and actual center wavelengths [nm] of the optical radiometers used in the tower shading experiments. All instrument channels have 10-nm bandwidths.
Truth data used for theoretical computations given in Fig. 4 at different nominal center wavelengths [nm].
Typical IOPs used for the computations given in Fig. 5 at different wavelengths [nm].
Computed shading errors on subsurface downwelling irradiance Ed (7.5, 0−, λ) and upwelling radiance Lu (7.5, 0−, λ) (the dependence on the time parameter t is not necessary here) for a diffuse light source assuming typical values of inherent optical properties, at different wavelengths [nm]. Confidence intervals are given in parentheses.
