This PDF file contains the front matter associated with SPIE Proceedings Volume 6680, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and the Conference Committee listing.

In order to retrieve ocean color from satellite imagery, one must perform atmospheric correction, because when observed from space the ocean signature is weak compared with the strong atmospheric signal. The color of the ocean depends on its optically active constituents: water molecules, dissolved matter, and particulate matter. In the open ocean, the color is mainly due to water molecules and phytoplankton, whereas in the coastal zone, the color also results from the presence of sediments and colored dissolved organic matter. Because coastal waters (Case 2 waters) are much more difficult to decouple from the atmosphere than open ocean (Case 1 waters), operational atmospheric correction algorithms usually separate Case 1 from Case 2 waters processing. The solution proposed in this paper does not separate them. Our algorithm, referred to as Ocean Color Estimation by principal component ANalysis (OCEAN), exploits the fact that ocean is more variable spectrally than the atmosphere, while the atmosphere signal is more variable in magnitude. The satellite reflectance is first decomposed into principal components. The components sensitive to the ocean signal are then combined to retrieve the principal components of the marine reflectance via neural network methodology. The algorithm is described, and results are presented on real and simulated data for POLDER, MERIS, SeaWiFS, and MODIS. Accurate water reflectance estimates are obtained for various aerosol types and contents (including maritime, coastal and urban mixtures), and for the full range of water properties (resulting from realistic combinations of chlorophyll content, sediment content, and colored dissolved matter absorption).

The Ocean Color Estimation by principal component ANalysis (OCEAN) algorithm performs atmospheric correction of satellite ocean-color imagery in the presence of various aerosol contents and types, including absorbing mixtures, and for the full range of water properties (Case 1 and Case 2 waters), retrieving diffuse water reflectance with good theoretical accuracy. It is easy to implement and has several advantages for operational processing lines: (1) It has de-noising abilities, for it is based on principal component analysis and neural networks, (2) it is able to perform atmospheric correction through cirrus and thin clouds, (3) it is able to retrieve water reflectance in the presence of Sun glint until a glint reflectance of 0.2, and more importantly, (4) it is less sensitive to absolute radiometric calibration and directionality than classical ocean-color algorithms. This allows multi-sensor merging (denoted hereafter Level 4 synthesis). These abilities may improve dramatically the daily spatial coverage of ocean color products. In the companion paper (Part I), the theoretical performance of OCEAN in situations of both Case 1 and Case 2 waters is presented for various multispectral radiometers (i.e., POLDER, SeaWiFS, MODIS, MERIS). In this paper (Part II), the focus is made on OCEAN de-noising and merging properties. The ability of the algorithm to work in situations of Sun glint and cirrus/thin clouds is illustrated using MERIS imagery. Multi-directional merging is demonstrated using POLDER imagery (daily and temporal merging), and multi-sensor merging using SeaWiFS and MODIS imagery (daily merging). The resulting products do not show directional artifacts.

An inversion methodology, based on least squares minimization, is proposed to retrieve spectral marine reflectance from top-of-atmosphere reflectance measurements in the visible and near infrared. The problem is first made linear by decomposing into principal components the additive contributions of the water body (the signal of interest) and of the atmosphere and surface (the perturbing signal), after subtraction of molecular effects and proper normalization. For realistic geometric and geophysical conditions, the two contributions can be described adequately by a few eigenvectors. This yields generally (i.e., for current satellite ocean-color sensors) an over-determined system of linear equations, in which the unknown parameters are the coefficients associated with the eigenvectors. The problem is ill conditioned, since the measurements are noisy, the signal of interest is small compared with the top-of-atmosphere reflectance, and some of the eigenvectors of the atmosphere/surface signal are correlated with those of the water-body signal. The system of linear equations is solved in the least squares sense using a regularization scheme, in which a regularization parameter is introduced to stabilize the solution. Once the coefficients are determined, they are used to reconstruct the water-body signal, basically the marine reflectance, and the perturbing signal. Performance is evaluated theoretically for the Sea-viewing Wide Field-of-view Sensor, using a comprehensive non-noisy simulated data set. The inversion scheme yields acceptable root-mean-squared errors of 0.0034, 0.00228, 0.00132, 0.00102, 0.00081, and 0.00021 on the retrieved water-body signal at 412, 443, 490, 510, 555, and 670 nm (Case 1 waters). Performance could be improved by using additional wavelengths in the near infrared, but more eigenvectors might be required to describe the atmospheric/surface signal.

A methodology is presented for obtaining information on aerosol vertical structure, a key variable in studies of aerosol climate forcing and atmospheric correction of satellite ocean-color imagery. The methodology employs ground-based angular measurements of atmospheric radiance, total or polarized, in the oxygen A-band centered on 763 nm. The radiance measured at different zenith angles is sensitive to different atmospheric layers, and the measurements can be inverted to retrieve the vertical profile of aerosol concentration. To solve the inverse problem, in which small errors in the data may yield large errors in the reconstructed profile, an iterative regularization scheme, robust to noise and perturbing effects (e.g., due to multiple scattering and non-null surface reflectance), is developed. Maximum entropy regularized solutions are introduced. The methodology is tested on atmospheric radiance data simulated for typical aerosol profiles and aerosol types. The retrieved aerosol profiles agree with the prescribed ones, indicating that the inversion scheme is efficient in achieving a proper balance between goodness-of-fit to the data and stability of the solution. The methodology has the potential to extend and complement surface observations of aerosol vertical structure made by lidar networks. This perspective is significant, since current information on aerosol vertical structure is insufficient to constrain and verify key assumptions in global aerosol models. The complementary information would contribute, via assimilation, to improving predictions of aerosol radiative forcing and to reducing uncertainties in model simulations of climate change. In addition, the methodology would help to evaluate the retrievals of aerosol vertical structure from space-borne lidars, and would be useful to check the atmospheric correction of satellite ocean-color imagery and develop improved correction algorithms in the presence of absorbing aerosols.

Effect of sun glint reflectance to the measurement of ocean color sensors is generally observed, even some of the sensors are equipped with tilt mechanisms to avoid sun glint. The traditional Cox-Munk model has been widely used to estimate the sun glint reflectance together with objective analysis wind data. To reevaluate the sun glint model at the condition of satellite viewing geometry, ADEOS-II/GLI data are analyzed jointly with SeaWinds microwave scatterometer data which provides the concurrent wind data with the ocean color observation. The probability density of the wave slope is then estimated using GLI data of 865nm band after carefully masking the cloud-contaminated pixels and removing the aerosol effects, the latter being estimated from the SeaWiFS Level 3 daily aerosol data set. The satellite-retrieved probability density functions are then analyzed as a function of wind speed and wave slope angle, by fitting the satellite retrieved probability density of slope to the anisotropic model. Modified model parameters is given and applied into GLI processing. Results are compared with the original anisotropic Cox-Munk model, as well as other published results. Differences in the slope distributions are discussed, which is mostly found at very weak speed or very strong wind speed.

Artificial neural network has been proven a useful technique for deriving water and bottom properties from remote sensing upwelling radiance. Conventionally, a neural network is trained to minimize the overall mean square error of desired products. The approach does not explicitly take into account the change of spectral shapes of upwelling radiance. In this study, we have created four groups of training sets, two groups with ratios of Rrs( λ_i) to Rrs(557), and the others without. Ratios of Rrs( λ_i) to Rrs(557) for λ_i of 409nm, 438nm, 488nm, 507nm, 616nm, 665nm, 683nm, 712nm, 750nm and 779nm have been used as additional inputs in the training of neural networks. Trained neural networks were then applied to an independent testing set which was created for optically different coastal waters. The inclusion of 10 spectral ratios in the training significantly improves the accuracy of derived water depth H, backscattering coefficient b_b(438) and the absorption coefficient a(438). The accuracy of the derived coefficients is 86%, 94% and 92%. Our results clearly show the importance for including spectral ratios in the neural network training process. Remote sensing upwelling radiance over the identified 11 spectral channels provides adequate information for the retrieval of water optical property coefficients when an artificial neural network approach is used.

This work presents the multispectral reconstruction of the vertical profile of Chlorophyll concentration (Chl-a) in Case 1 waters. In such ocean waters, optical properties are driven by the presence of phytoplankton, allowing the use of bio-optical models, where the absorption and scattering coefficients are related with Chl-a. The vertical profile of Chl-a is reconstructed from experimental measurements of water-leaving radiances (nLw) at 10 wavelengths. These radiances are considered for a discrete number of upward polar directions. The inverse problem is formulated as an optimization problem, and iteratively solved by the ACO using the radiative transfer equation as direct model. The objective function is given by the square difference between computed and experimental radiances. For each iteration a population of candidate solutions is generated, pre-selected and evaluated by means of the objective function. Each candidate solution corresponds to a discrete Chl-a profile. The radiative transfer equation is solved for each candidate solution yielding the radiances that are used in the objective function to evaluate it. Since this equation requires the absorption and scattering coefficients, these are calculated using bio-optical models. The radiative transfer equation is solved using the Laplace transform discrete ordinate LTS_N method. A parallel implementation of the ACO is employed and executed in a distributed memory machine. The multispectral approach allows estimating the vertical profile using only nLw, instead of in situ measurements on several depths.

Improved remote sensing retrievals of the chlorophyll fluorescence component in coastal water reflectance can significantly help environmental impact assessments. While retrieval of chlorophyll fluorescence from satellite observations of open ocean reflectance using Fluorescence Line Height (FLH) algorithms is now routine, it is much more complicated in coastal waters where the fluorescence overlaps with a NIR elastic scattering peak arising from the combination of photosynthetic pigment and particulate scattering and absorption, and rapidly increasing water absorption. To examine retrieval accuracies attainable in coastal waters by MODIS and other FLH algorithms, we compared the results of extensive numerical simulations with those of our field measurements in the Chesapeake Bay. The relationship between the contribution of fluorescence in the reflectance spectra and [Chl] and other water constituents was analyzed by simulations of more than 1000 reflectances using the HYDROLIGHT radiative transfer program. For these, IOP were related to parameterized microphysical models, following the same procedures used to generate the IOCCG dataset, but with higher (1 nm) spectral resolution, and wider range of parameters including chlorophyll specific absorption more typical of coastal waters. Results of simulations and field measurements show that the variability of retrieved fluorescence can be attributed largely to its attenuation in the water by algae, CDOM and mineral particles, and much less to the variation of the fluorescence quantum yield. Our systematic parametric study of fluorescence as a function of the other water components is then used to define the range of water parameters where fluorescence contributes significantly to the NIR peak reflectance, and where it is almost undetectable.

A few years ago, Lee and Carder demonstrated that for the quantitative derivation of major properties in an aqua-environment (information of phytoplankton biomass, colored dissolved organic matter, and bottom status, for instance) from remote sensing of its color, a sensor with roughly ~17 spectral bands in the 400 − 800 nm range can provide acceptable results compared to a sensor with 81 consecutive bands (in a 5-nm step). In that study, however, it did not show where the 17 bands should be placed. Here, from nearly 300 hyperspectral measurements of water reflectance taken in both coastal and oceanic waters that covering both optically deep and optically shallow waters, first and second derivatives were calculated after interpolating the measurements into 1-nm resolution. From these hyperspectral derivatives, the occurrence of zero value at each wavelength was accounted for, and a spectrum of the total oc-urrences was obtained, and further the wavelengths that captured most number of zeros were identified. Because these spectral locations indicate extremum (a local maximum or minimum) of the reflectance spectrum or inflections of the spectral curvature, placing the bands of a sensor at these wavelengths maximize the possibility of capturing (and then accurately restoring) the detailed curve of a reflectance spectrum, and thus maximize the potential of detecting the changes of water and/or bottom properties of various aqua environments with a multi-band sensor.

Automated validation methods and a suite of tools have been developed in a Quality Control Center to analyze the stability and uncertainty of satellite ocean products. The automatic procedures analyze match-ups of near real time coastal bio-optical observations from Martha's Vineyard Coastal Observatory (MVCO) with satellite-derived ocean color products from MODIS Aqua and Terra, SeaWIFS, Ocean Color Monitor, and MERIS. These tools will be used to compare MVCO in situ data sets (absorption, backscattering, and attenuation coefficients), co-located SeaPRISM-derived water leaving radiances, and the Aerosol Robotic Network (AeroNet) derived aerosol properties with daily satellite bio-optical products and atmospheric correction parameters (aerosol model types, epsilon, angstrom coefficient), to track the long term stability of the bio-optical products and aerosol patterns. The automated procedures will be used to compare the in situ and satellite-derived values, assess seasonal trends, estimate uncertainty of coastal products, and determine the influence and uncertainty of the atmospheric correction procedures. Additionally we will examine the increased resolution of 250m, 500m, and 1 km satellite data from multiple satellite borne sensors to examine the spatial variability and how this variability affects assessing the product uncertainty of coastal match-ups of both bio-optical algorithms and atmospheric correction methods. This report describes the status of the QCC tool development and potential applications of the QCC tool suite.

Ocean color satellites provide a mechanism for studying the marine biosphere on temporal and spatial scales otherwise unattainable via conventional in situ sampling methods. These satellites measure visible and infrared radiances, which are used to estimate additional geophysical data products, such as the concentration of the phytoplankton pigment chlorophyll a, C_a, via the application of secondary bio-optical algorithms. The operational C_a algorithms for the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and Moderate Resolution Imaging Spectroradiometer (MODIS), for example, perform well in the global open ocean, but often degrade in more optically complex coastal environments where global parameterizations are less applicable. Organizations such as the Chesapeake Bay Program, which have interest in using SeaWiFS and MODIS data products to facilitate regional monitoring activities, must rely on locally parameterized algorithms to achieve requisite accuracies. To facilitate algorithm selection, the NASA Ocean Biology Processing Group recently developed the infrastructure to spatially and temporally evaluate a long-term regional time-series of satellite observations using in situ measurements as ground-truth. Here, we present this approach using a case study in the Chesapeake Bay, where a series of C_a algorithms and atmospheric correction schemes were evaluated for the full SeaWiFS and MODIS-Aqua time-series. We demonstrate how the selection of the best algorithms and processing approaches is driven by trade-offs in coverage needs and relative accuracy requirements. While our case study highlights C_a in the Chesapeake Bay, our methodology is applicable to any geophysical data product and region of interest.

Comparisons between in situ measurements of surface chlorophyll concentration (CHL) and ocean color remote sensing estimates were conducted during an oceanographic cruise in the Brazilian Southeastern continental shelf and slope in November 2004. In situ estimates were based on fluorometry, above-water radiometry and lidar fluorosensor. Three empirical algorithms were used to estimate chlorophyll a concentration from radiometric measurements: Ocean Chlorophyll 3 bands (OC3M), Ocean Chlorophyll 4 bands (OC4v4), and Ocean Chlorophyll 2 bands (OC2v4). The satellite estimates of chlorophyll a were derived from data collected by the Moderate-resolution Imaging Spectroradiometer (MODIS) with a nominal 1.1 km resolution at nadir. Three algorithms were used to estimate chlorophyll concentrations from MODIS data: one empirical - OC3M, and two semi-analytical - Garver, Siegel, Maritorena version 01 (GSM01), and Carder. In this paper, LIDAR, MODIS and in situ above-water radiometry and fluorometry are briefly described and the estimated values of chlorophyll retrieved by these techniques are compared. Chlorophyll concentrations were fairly well estimated by all the methods. In general, the empirical algorithms applied to the satellite and in situ radiometric data showed a tendency for overestimating CHL. The semi-analytical GSM01 algorithm applied to MODIS data performed better than the Carder and the empirical OC3M algorithms.

The water-leaving spectral radiance is a basic ocean color remote sensing parameters required for the vicarious calibration. Determination of water-leaving spectral radiance using in-water radiometry requires measurements of the upwelling spectral radiance at several depths. The Marine Optical System (MOS) Remotely Operated Vehicle (ROV) is a portable, fiber-coupled, high-resolution spectroradiometer system with spectral coverage from 340 nm to 960 nm. MOS was developed at the same time as the Marine Optical Buoy (MOBY) spectrometer system and is optically identical except that it is configured as a profiling instrument. Concerns with instrument self-shadowing because of the large exterior dimensions of the MOS underwater housing led to adapting MOS and ROV technology. This system provides for measurement of the near-surface upwelled spectral radiance while minimizing the effects of shadowing. A major advantage of this configuration is that the ROV provides the capability to acquire measurements 5 cm to 10 cm below the water surface and is capable of very accurate depth control (1 cm) allowing for high vertical resolution observations within the very near-surface. We describe the integrated system and its characterization and calibration. Initial measurements and results from observations of coral reefs in Kaneohe Bay, Oahu, extremely turbid waters in the Chesapeake Bay, Maryland, and in Case 1 waters off Southern Oahu, Hawaii are presented.

Determination of the water-leaving spectral radiance using in-water instrumentation requires measurements of the upwelling spectral radiance (L_u) at several depths. If these measurements are separated in time, changes in the measurement conditions result in increased variance in the results. A prototype simultaneous multi-track system was developed to assess the potential reduction in the Type A uncertainty in single set, normalized water-leaving radiance achievable if the data were acquired simultaneously. The prototype system employed a spectrograph and multi-track fiber-coupled CCD-detector; in situ in-water tests were performed with the prototype system fiber-coupled to a small buoy. The experiments demonstrate the utility of multi-channel simultaneous data acquisition for in-water measurement applications. An example of the potential impact for tracking abrupt responsivity changes in satellite ocean color sensors using these types of instruments as well as for the satellite vicarious calibration is given.

The Columbia River Plume is a highly dynamic water mass that supplies silicate and trace metals, fresh water, and dissolved and particulate organic matter to the Oregon/Washington shelf. The optical and physical properties of the river plume evolve as it travels away from the river mouth and undergoes both aging and dilution by surrounding waters. The objectives of this study were to (1) identify initial optical properties of fresh plume waters at the river mouth, (2) track changes in the optical signature of the water mass as it advects seaward from the mouth, and (3) predict residence time of the water mass on the shelf from changes in the optical signature, using remote sensing data. These results are compared to central California, where river plumes are much more episodic and spatially smaller, to determine the limits of detection using standard (1 km) and high-resolution (250 m) data from the MODIS platform.

The phytoplankton distribution in the East China Sea (ECS) was analyzed by the statistics of the discharge and the sediment load of the Yangtze River at the Datong station from 2000 to 2005 and satellite images of chlorophyll-a concentration which were observed by Sea-viewing Wide Field-of-view Sensor (SeaWiFS) from 2000 to 2005. This study suggests that phytoplankton distribution in the ECS in the spring and fall season could be determined by the amount of discharge, sediment load, and sediment concentration from the Yangtze River. The standing stocks of chlorophyll-a along the water column could be determined by the nutrient concentrations through the winter season. Also, this study suggests that the sediment load may have two functions to increase the primary productivity by an increase of silicate concentration and to decrease the primary productivity by an increase of diffused attenuation coefficient. However, the phytoplankton distribution in the summer season could not be simply determined from the discharge, sediment load, and sediment concentration from the Yangtze River. The associated or inherent parameters like diffused attenuation coefficient or primary productivity may have a significant contribution to the spatial distribution of phytoplankton in the summer.

Variability of particulate beam attenuation coefficient at 532 nm (c_p (532)) and microbial planktonic community (heterotrophic bacteria and phytoplankton) was analyzed in coastal waters of Southern California. The goal of this study was to explore heterotrophic bacteria (HB) response (cell abundance, BA, and carbon production, BCP) with respect to different particle characteristics (concentration, size distribution, and composition) related with c_p(532). We observed a fairly complex pattern of HB response and particle dynamics during seven experiments throughout the summer and winter, which reflected variations in c_p(532). The first experiment showed relatively high values of c_p(532), in conjunction with high chlorophyll a concentration (chl) of about 5.4 mg m^-3. For experiments 2 and 3, a sharp decrease of chl was accompanied by an increased role of detrital particles (non-living matter) as evidenced by increased detrital absorption (ad). The highest values of particle-attached (>1 μm) and free living (<1 μm) BA and BCP were observed in experiment 3. These changes in particle assemblage including HB maintained c_p(532) at relatively high level, comparable to that observed when phytoplankton dominated. A significant decrease of c_p(532) was observed in experiment 4 and 5, which coincided with relatively low BA, BCP, and ad values. In experiment 7, c_p(532) magnitude was comparable to the first experiment and was accompanied by high chl, BA and SPM (suspended particulate matter). Greatest changes in c_p(532) coincided with greatest variations in BA, even though our estimates of the direct contribution of heterotrophic bacteria to c_p(532) for all experiments remained quite low (<10%).

The coastal zone is of enormous importance to the environmental, economic, and social well being of nations. It is subject to increasing pressures from many sources, including industrial development, urban expansion, the exploitation of marine resources, and tourism. In order to understand and address the effects of natural and anthropogenic forces in the Southeastern coastal zone of Brazil, time-series of in-situ and satellite-based environmental observations are being developed to account for the interconnectivity of processes within the system. In this work, data collected during December 2004-January 2006 at the ANTARES time series station near Ubatuba, Southeast Brazil (23°44'S and 45°00'W) are analyzed. The data set includes measurements of near-surface chlorophyll-a concentration (CHL), absorption by particles, detritus, and colored dissolved organic matter (CDOM), and above-water hyperspectral reflectance. A triangular diagram, based on the relative contribution to spectral absorption of the optically active constituents, is used to classify the waters, revealing CDOM-dominated Case 2 waters. Seasonal changes in water composition and optical properties are examined. Applying the OC2v4, OC4v4, and OC3M algorithms to the radiometric data, after proper spectral integration, the CHL estimates are generally too high compared with fluorometric determinations, which might be caused by relatively large CDOM absorption at the coastal site.

Current ocean color sensors, for example SeaWiFS and MODIS, are well suited for sampling the open ocean. However, coastal environments are spatially and optically more complex and require more frequent sampling and higher spatial resolution sensors with additional spectral channels. We have conducted experiments with data from Hyperion and airborne hyperspectral imagers to evaluate these needs for a variety of coastal environments. Here we present results from an analysis of airborne hyperspectral data for a Harmful Algal Bloom in Monterey Bay. Based on these results and earlier studies we recommend increased frequency of sampling, increased spatial sampling and additional spectral channels for ocean color sensors for coastal environments.

Ocean color monitoring on the coastal water is still under study because of an incomplete atmospheric correction over the turbid water like over the coastal water along the China main land. Currently available sensors for science as MODIS on Terra or Aqua will terminate their service in the near future and the NPOESS Preparatory Project (NPP) will be the next satellite to support the satellite oceanography on the coastal water. The Tokyo University of Information Sciences (TUIS) has updated the MODIS receiving system to capture and ingest the Visible/Infrared Imager/Radiometer Suite (VIIRS) data from NPP, which will be launched in 2008. Data processing software from the Direct Readout Laboratory (DRL), such as the Real-time Software Telemetry Processing (RT-STPS), Simulcast, and DB algorithms, will be core programs in our system. VIIRS has seven bands in VIS&NIR, which are for ocean color research. The spatial resolution is 0.742×0.259 meters at nadir. While the MODIS spatial resolution of the nine ocean color bands is 1000m. The higher spatial resolution MODIS data (250 meters) is used to illustrate the advantage of the higher spatial resolution remote sensing data, such as data from VIIRS. In this study, we propose to combine the higher spatial resolution data with the traditional products of chlorophyll-a and sea surface temperature in the low resolution so as to extract further information on the coastal ocean.

Coupling the 3-d ocean optical imagery with 3-d circulation models provides a new capability to understand coastal processes. Particle distribution derived from ocean color optical properties were coupled with numerical circulation models to determine a 24 hour forecast of particle concentrations. A 3-d particle concentration field for the coastal ocean was created by extending the surface satellite bio-optical properties vertically by parameterzing an expediential Gaussian depth profile. The shape of the vertical particle profile was constrained by 1) the depth of the 1% light level 2) the mixed layer depth 3) the intensity of the layer stratification 4) and subsurface current field and the surface bio-optical properties. These properties were obtained from MODIS ocean optical products (phytoplankton absorption and backscattering) and the Intra-America Sea Nowcast Forecast System - Naval Coastal Ocean Model. The 3-d particle distribution was imbedded into a 3-d circulation model and the particles advected hourly using forecast model 3-d current. The particles were diffused, dispersed and differentially settled during the advection processes. Following the 24 hour advection, the resultant particle distribution were accumulated into 1 km spatial grid and vertically to a 1 attenuation length (satellite penetration depth) and the forecast ocean color backscattering image determined. The forecast image was compared with the next day ocean color backscattering image to define the error budget. The ocean color particle tracking, defines fine spatial scales processes such as local upwelling and downwelling, which are essential in understanding the coupling of physical and bio-optical processes. The methods provide new capability for characterizing how subsurface particles layers change in response to cross and along shelf exchange processes. Results show methods to forecast satellite optical properties in coastal areas and examine how sequential MODIS imagery of the particle scattering is related to particle transport and physical processes

The dynamics of rainstorm plumes in the coastal waters of southern California was studied during the Bight'03 Regional Water Quality Program surveys. Measurements of surface salinity and bacterial counts collected from research vessels were compared to MODIS-Aqua satellite imagery. The spectra of normalized water-leaving radiation (nLw) were different in plumes and ambient ocean waters, enabling plumes discrimination and plume area size assessments from remotely-sensed data. The plume/ocean nLw differences (i.e., plume optical signatures) were most evident during first days after the rainstorm over the San Pedro shelf and in the San Diego region and less evident in Santa Monica Bay, where suspended sediments concentration in discharged water was lower than in other regions. In the Ventura area, plumes contained more suspended sediments than in other regions, but the grid of ship-based stations covered only a small part of the freshwater plume and was insufficient to reveal the differences between the plume and ocean optical signatures. The accuracy of plume area assessments from satellite imagery was not high (77% on average), seemingly because of inexactitude in satellite data processing. Nevertheless, satellite imagery is a useful tool for the estimation of the extent of polluted plumes, which is hardly achievable by contact methods.

Water-leaving radiance data obtained from MODIS-Aqua satellite images at spatial resolution of 250 m (band 1 at 645 nm) and 500 m (band 4 at 555 nm) were used to analyze the correlation between plume area and rainfall during strong storm events in coastal waters of Southern California. Our study is focused on the area between Point Loma and the US-Mexican border in San Diego, which is influenced by terrigenous input of particulate and dissolved materials from San Diego and Tijuana watersheds and non-point sources along the shore. For several events of intense rainstorms that occurred in the winter of 2004-2005, we carried out a correlational analysis between the satellite-derived plume area and rainfall parameters. We examined several rainfall parameters and methods for the estimation of plume area. We identified the optimal threshold values of satellite-derived normalized water-leaving radiances at 645 nm and 555 nm for distinguishing the plume from ambient ocean waters. The satellite-derived plume size showed high correlation with the amount of precipitated water accumulated during storm event over the San Diego and Tijuana watersheds. Our results support the potential of ocean color imagery with relatively high spatial resolution for the study of turbid plumes in the coastal ocean.

The multi-parametric data observed by the Synthetic Aperture Radar (SAR) and the Moderate Image Scanning Radiometer (MODIS) were applied to analyze the water distribution in the Philippine Archipelago to build a new radar imaging model. The SAR provides information on the surface structure of the water affected by various phenomena such as currents, internal waves, swells, surface winds, rain falls, natural films, and so on. One difficult parameter among them is the natural film, which exhibits a damping effect to the surface roughness from breaking waves to capillary ones. The chlorophyll-a concentration observed by satellite sensors such as MODIS provide estimates of the standing stock of phytoplankton, which is considered as the parameter estimating the standing stock of zooplankton. The standing stocks of phytoplankton and zooplankton could be a proportional parameter to the amount of the natural film. In this study, a frequency dependent multi-parametric equation was proposed to rebuild the surface roughness with various parameters in the spectrum domain. Possible validation study was conducted with the relationship between the sum of power spectrum for certain frequency range with chlorophyll-a concentration on the Mindanao Sea and the Surigao Strait.

The ballast water exchange at seas has been recognized as one of the operational countermeasures to cope with the invasion of non-indigenous species through the ballast water. The Bay of Bengal is traditionally considered to have low chlorophyll-a concentration thus low phytoplankton counts, which is the reason why the Bay of Bengal (BoB) has been selected as a suitable ballast water exchangeable sea. However an anomalously high K(490) area was found off the coast of Sri Lanka during the northeast monsoon in 2005, which corresponds to higher plankton cell densities than the criterion set by the regulation of International Maritime Organization (IMO). The regression equation between K(490) and corresponding in situ plankton cell densities in the Bay of Bengal is developed to identify suitable ballast water exchangeable area based on the regulations of IMO. According to the results the central and eastern portions in the Bay of Bengal during the northeast monsoon season are found to be suitable, while the unsuitable area broadens during the southwest monsoon season at the western Bay of Bengal. Seasonal and annual variability of K(490) and corresponding cell density is discussed to establish an early routing system for avoiding the high cell density area in advance.

During fall periods in 2002, 2003 and 2004 three major oceanographic expeditions were carried out in Mamala Bay, Hawaii. These were part of the RASP Remote Anthropogenic Sensing Program. Ikonos and Quickbird optical satellite images of sea surface glint revealed ≈100 m spectral anomalies in km² averaging patches in regions leading from the Honolulu Sand Island Municipal Outfall diffuser to distances up to 20 km. To determine the mechanisms behind this phenomenon, the RASP expeditions monitored the waters adjacent to the outfall with an array of hydrographic, optical and turbulence microstructure sensors in anomaly and ambient background regions. Drogue tracks and mean turbulence parameters for 2 × 10⁴ microstructure patches were analyzed to understand complex turbulence, fossil turbulence and zombie turbulence near-vertical internal wave transport processes. The dominant mechanism appears to be generic to stratified natural fluids including planet and star atmospheres and is termed beamed zombie turbulence maser action (BZTMA). Most of the bottom turbulent kinetic energy is converted to ≈ 100 m fossil turbulence waves. These activate secondary (zombie) turbulence in outfall fossil turbulence patches that transmit heat, mass, chemical species, momentum and information vertically to the sea surface for detection in an efficient maser action. The transport is beamed in intermittent mixing chimneys.

Combined use of Advanced Synthetic Aperture Radar (ASAR) onboard ENVISAT and the mesoscale model MM5 is proposed in this study to assess offshore wind energy potential. As far as the validation is concerned, 22 ASAR scenes are processed to estimate wind speed distribution by using CMOD4 and CMOD-IFR2 algorithms over Tanabe Bay, which is the target area of this study and has an offshore wind observation station. These algorithms require inputs of relative wind direction, which is usually defined as the ASAR viewing direction relative to the observed wind direction at the time of ASAR overpass. In this study MM5-simulated wind direction is used as a substitute of the observed wind direction for calculating relative wind direction, and the estimated wind speeds are compared with observed wind speeds at the offshore station. RMS errors and biases of CMOD4 and CMOD-IFR2 with inputs of observed wind direction are 2.05m/s, -0.80m/s and 2.16m/s, -0.54m/s respectively. On the other hand, in the case of using the MM5-simulated wind direction as a substitute of the observed wind direction, these statistics are 2.28m/s, -1.16m/s and 2.39m/s, -0.75m/s, respectively. That is, the accuracies of the estimated wind speed using MM5- simulated wind direction are found to be worse slightly. However, these results indicate that ASAR-based wind estimation in combination with MM5-simulated wind vectors can be a promising approach for offshore wind mapping, especially for coastal waters with complicated onshore terrains and coastlines in Japan.

The Great Islands zone, in the Gulf of California, presents high phytoplankton concentration as a result of the high Turbulent Kinetic Energy (TKE). In this work we looked for dynamics zones based on Empirical Orthogonal Function analysis (EOF). The input data were Sea Surface Temperature (SST) and Chlorophyll-a concentration (Chla) from daily MODIS-AQUA at 1 Km from 2003 to 2006. Time series were generated to define the average conditions for summer and winter spring tides. Results showed that in general and during summer-spring tides, higher Chla concentrations are localized in the west coast, with a displacement to the south. These high Chla were associated with tidal mixing. Zero EOF values in summer showed the boundary between low SST and high Chla. During winter-spring tides there were more spatial variability than during summer time. Zero EOF value in winter time showed low SST and Chla in the west coast due to stronger mixing conditions that stay longer. Results of this work emphasize that a dynamic regionalization can be used in high TKE areas and it helps to define zones with a similar response based on the input parameters chosen.

The SeaWinds on QuikSCAT scatterometer measures near surface ocean winds using radar backscatter values and a geophysical model function. QuikSCAT data is limited in coastal regions due to land contamination of the backscatter measurements. However, wind retrieval in near coastal areas can be successfully accomplished by estimating the amount of land contamination in the backscatter measurements and eliminating measurements which exceed a specified threshold. In order to accurately assess the amount of land contamination in a given measurement, a detailed knowledge of the antenna spatial response is required. The land contribution ratio is used as the contamination metric and is calculated using the spatial response of each QuikSCAT measurement. The land contamination threshold changes during wind retrieval as a function of wind conditions in the local area which allows retrieval closer to the coast as wind speeds increase. To ameliorate processing time the QuikSCAT spatial response is calculated and tabulated prior to wind retrieval. Subjective comparisons to estimated coastal winds show that determining a land contamination threshold using the percent land contribution metric provides accurate wind retrieval up to 25km closer to the coast than current methods. As a result wind speeds can be accurately retrieved as close as 2.5km from the coast.

Comparisons between in situ measurements of surface chlorophyll-a concentration (CHL) and ocean color remote sensing estimates were conducted in the Brazilian Southeast coast, Southwestern South Atlantic. In situ fluorometric data were acquired in four hydrographic cruises carried out during the austral summer and winter of 2001 and 2002. The satellite estimates of CHL were derived from SeaWiFS data recorded in HRPT mode by INPE's station with a nominal 1.1 km resolution. Four algorithms were used to estimate CHL: two empirical - Ocean Chlorophyll 4 bands (OC4v4), and 2 bands (OC2v4); one semi-analytical - Garver, Siegel, Maritorena (GSM01); and one based on neural network (NN). Comparisons of estimated and measured CHL were done within a temporal window of 12 hours from the in situ sampling time. SeaWiFS algorithms values are 5x5 pixel medians centered on the location of in situ sampling station. For the study area CHL was fairly well estimated by all the SeaWiFS algorithms. OC4 performed better (R² = 0.71; rms = 0.22) than the other algorithms (OC2, GSM01, and NN). The OC2 algorithm also showed a good performance with R² = 0.67 and rms = 0.23. The neural network algorithm performed better than the semi-analytical one (R² = 0.62 and 0.55, respectively), but with a higher rms (0.34 and 0.20, respectively). In general, the OC4, OC2, and NN algorithms showed a tendency for overestimating CHL at higher concentrations and underestimating at lower values. The semi-analytic GSM01 algorithm overestimated only the lower CHL, but underestimated most of the other values.

Typical MODIS ocean color products are at 1 kilometer (km) spatial resolution, although two 250 meter (m) and five 500 m bands are also available on the sensor. We couple these higher resolution bands with the 1km bands to produce pseudo-250m resolution MODIS bio-optical properties. Finer resolution bio-optical products from space significantly improve our capability for monitoring coastal ocean and estuarine processes. Additionally, increased resolution is required for validation of ocean color products in coastal regions due to the shorter spatial scales of coastal processes and greater variability compared to open-ocean regions. Using the 250m bands coupled with the 1km and 500m bands (which are bi-linearly interpolated to 250m resolution), we estimate remote sensing reflectances (Rrs) at 250m resolution following atmospheric correction. The aerosol correction makes use of the 1km near infrared (NIR) bands at 748 nanometers (nm) and 869 nm to determine aerosol type and concentration. The water leaving radiances in the NIR bands are modeled from retrieved water leaving radiances in the visible bands using the short wave infrared (SWIR) channels at 1240 nm and 2130 nm. The increased resolution spectral Rrs channels are input into bio-optical algorithms (Quasi-Analytical Algorithm (QAA), Water Mass Classification, OC2, etc.) that have traditionally used the 1 km reflectances resulting in finer resolution products. Finer resolution bio-optical properties are demonstrated in bays, estuaries, and coastal regions providing new capabilities for MODIS applications in coastal areas. The finer resolution products of total absorption (at), phytoplankton absorption (aph), Color-Dissolved Organic Matter (CDOM) absorption (ag) and backscattering (bb) are compared with the 1km products and in situ observations. We demonstrate that finer resolution is required for validation of coastal products in order to improve match ups of in situ data with the high spatial variability of satellite properties in coastal regions.

Satellite ocean color remote sensing is plagued by loss of coverage due to cloud obscuring, glint contamination, atmospheric correction failures, and other issues. We have developed a simple and efficient technique for estimating missing remote sensing data by taking advantage of the inter-pixel spatial and temporal coherency of individual ocean color products. The technique first employs a limited iterative triangular interpolation procedure. This procedure attempts to select three neighboring pixels forming the tightest triangle enclosing the data point we are attempting to recover; and then interpolating. On failure to find three suitable neighbors, a second procedure is employed which attempts to recover missing data points by using a time dependent "latest pixel" replacement. This procedure replaces the missing data point with the most recent data point collected at that grid point within the last seven days. This technique has been applied to MODIS (MODerate resolution Imaging Spectrometer) ocean color products of phytoplankton absorption, back-scattering coefficient, and chlorophyll concentration to produce cloud free bio-optical products on a daily basis and provide a new capability for monitoring coastal processes. We demonstrate a new method on MODIS products and show how bio-optical properties change over a daily and monthly time scale.

Numerical experiments conducted with an ocean general ocean circulation model reveal the potential influence of solar radiation absorbed by phytoplankton on the thermal structure and currents of the Tropical Atlantic Ocean. In the model, solar radiation penetration is parameterized explicitly as a function of chlorophyll-a concentration, the major variable affecting water turbidity in the open ocean. Two types of runs are performed, a clear water (control) run with a constant minimum chlorophyll-a concentration of 0.02 mgm^-3, and a turbid water (chlorophyll) run with space- and time-varying chlorophyll-a concentration from satellite data. The difference between results from the two runs yields the biological effects. In the chlorophyll run, nutrients and biology production are implicitly taken into account, even though biogeochemical processes are not explicitly included, since phytoplankton distribution, prescribed from observations, is the result of those processes. Due to phytoplankton-radiation forcing, the surface temperature is higher by 1-2 K on average annually in the region of the North Equatorial current, the Northern part of the South Equatorial current, and the Caribbean system, and by 3-4 K in the region of the Guinea current. In this region, upwelling is reduced, and heat trapped in the surface layers by phytoplankton is not easily removed. The surface temperature is lower by 1 K in the Northern region of the Benguela current, due to increased upwelling. At depth, the equatorial Atlantic is generally cooler, as well as the eastern part of the tropical basin (excluding the region of the sub-tropical gyres). The North and South equatorial currents, as well as the Equatorial undercurrent, are enhanced by as much as 3-4 cms^-1, and the circulation of the subtropical gyres is increased. Pole-ward heat transport is slightly reduced North of 35°N, suggesting that phytoplankton, by increasing the horizontal return flow in the subtropical region, may exert a cooling influence on higher latitude regions. The findings indicate that biology-induced buoyancy plays a significant role, in an indirect if not direct way, in the variability of the Tropical Atlantic Ocean, with consequences on atmospheric circulation and climate.

SeaWiFS ocean-colour products for the Southwestern Atlantic Ocean are evaluated in comparison with in situ bio-optical data collected during the March 10-12, 2002 cruise from Ushuaia to Uruguay. Radiometric measurements and surface water sampling were carried out at 14 stations. The in situ dataset included HPLC chlorophyll-a concentration (chl-a), a_ph( λ) spectral absorption coefficients of phytoplankton, suspended particles, and a_cdom( λ) dissolved organic matter, R_RS remote sensing reflectance, AOT and aerosol optical thickness. In general, the SeaWiFS-derived and SIMBAD measured AOT were low, but with a good agreement within SIMBAD uncertainty errors. The SeaWiFS-derived R_RS was systematically underestimated, but still with good fits. The a^*_ph (440) were high indicating the presence of small size cells with a low packaging effect. The HPLC pigment composition did not show strong variations amongst the sampled points, with communities most probably dominated by small cells. The phytoplankton community was more homogeneous in the southern stations, than in the northernmost stations influenced by the mixing of the Brazil and Malvinas Currents and the La Plata River discharge. The analyzes of the in situ a_cdom (440) characterized the sampling stations as CDOM rich waters. All SeaWiFS chl-a algorithms showed reasonable performances. The empirical algorithms overestimated lower chl-a while underestimated higher concentrations. The GSM01 semi-analytical algorithm underestimated most chl-a values, while CARDER underestimated only the lower concentrations. It is expected that the accuracy of chlorophyll retrievals in coastal areas of the BMC can be improved by a proper tuning of the semi-analytical models with regional inherent optical properties measurements.

This study employs SeaWiFS data over the waters off the southeastern China to evaluate a semi-analytical algorithm for euphotic zone depth (Z_e). This algorithm is based on water's inherent optical properties (IOPs), which can be near-analytically calculated from spectral remote-sensing reflectance, where remote-sensing reflectance can be derived from the normalized water-leaving radiance provided by SeaWiFS. In the Taiwan Strait, compared with in situ Z_e (±3 hour within SeaWiFS collection), average error (ε) is 15.0 % and root mean square error (RMSE) is 0.074, with Z_e in a range of 14-34 m from field measurements. In the South China Sea, compared with in situ Z_e (±48 hour within SeaWiFS collection),ε is 5.1 % in summer and 22.6 in winter, while RMSE is 0.032 in summer and 0.129 in winter, with Z_e in a range of 10-82 m from field measurements. For comparison, we also evaluate the performance of the empirical Z_e algorithm that is based on chlorophyll concentration. It is found that the IOP-centered approach has higher accuracy compared to the chlorophyll-a centered approach (e.g. in the South China Sea in winter, ε is 55.3 % and RMSE is 0.219). The new algorithm is thus found not only worked well with waters of the Gulf of Mexico, Monterey Bay and the Arabian Sea, but also worked well with waters of the China Sea.