Initial analysis of ocean color data from the ocean color and temperature scanner. I. Imagery analysis.
TL;DR: The ocean color and temperature scanner (OCTS) collected global ocean color data from November 1996 to June 1997 andalyses of OCTS imagery indicate three features that impair scientific research uses: band misalignments, image striping, and image noise.
Abstract: The ocean color and temperature scanner (OCTS) collected global ocean color data from November 1996 to June 1997. Analyses of OCTS imagery indicate three features that impair scientific research uses: (1) band misalignments, (2) image striping, and (3) image noise. These are due to (1) band offsets in the sensor design, (2) detector radiometric response variability, and (3) primarily cloud contamination, respectively. Methods are analyzed to ameliorate the effects of each that facilitate use of OCTS ocean color data for quantitative scientific analyses.
Citations
More filters
TL;DR: Results show that there are no obvious bias differences between the OCTS- and POLDER-derived ocean-color products, whereas the differences due to noise, which stem from variations in sensor characteristics, are difficult to correct at the pixel level.
Abstract: We describe our efforts to study and compare the ocean-color data derived from the Japanese Ocean Color and Temperature Scanner (OCTS) and the French Polarization and Directionality of the Earth's Reflectances (POLDER). OCTS and POLDER were both on board Japan's Sun-synchronous Advanced Earth Observing Satellite from August 1996 to June 1997, collecting approximately 10 months of global ocean-color data. This operation provided a unique opportunity for the development of methods and strategies for the merging of ocean-color data from multiple ocean-color sensors. We describe our approach to the development of consistent data-processing algorithms for both OCTS and POLDER and the use of a common in situ data set to calibrate vicariously the two sensors. Therefore the OCTS- and POLDER-measured radiances are bridged effectively through common in situ measurements. With this approach to the processing of data from two different sensors, the only differences in the derived products from OCTS and POLDER are the differences that are inherited from the instrument characteristics. Results show that there are no obvious bias differences between the OCTS- and POLDER-derived ocean-color products, whereas the differences due to noise, which stem from variations in sensor characteristics, are difficult to correct at the pixel level. The ocean-color data from OCTS and POLDER therefore can be compared and merged in the sense that there is no significant bias between two.
69 citations
TL;DR: In this paper, a general modeling formalism is used to extend near-field point spread function (PSF) measurements over a wide-field off-axis angular range, covering a dynamic range of 5 to 6 orders of magnitude, using bidirectional reflectance distribution function (BRDF) measurements of the scan mirror and other key optical elements.
Abstract: A general modelling formalism is used to extend near-field point spread function (PSF) measurements over a wide-field off-axis angular range, covering a dynamic range of 5 to 6 orders of magnitude, using bidirectional reflectance distribution function (BRDF) measurements of the scan mirror and other key optical elements. Line spread function (LSF) model results were compared with measured near-field LSF measurements for the Moderate Resolution Imaging Spectroradiometer (MODIS) of the Earth Observing System (EOS), demonstrating excellent agreement between model and measurements. When realistic effects of sensor scattered light are taken into account, significant radiometric bias errors are produced near high-contrast structured scenes (e.g. bright clouds over dark oceans and land; broken snow and ice scenes). Image restoration using an asymptotically exact PSF is shown to produce results significantly different from those produced by traditional near-field 5 × 5, 7 × 7, ..., PSF kernel matrix inversion techniques. The results reported have implications for future remote-sensor specifications and testing, in-flight and surface-based calibration comparisons, and the assessment of radiometric bias errors in the presence of moderate- to high-contrast Earth scenes.
24 citations
TL;DR: Comparison of OCTS-computed water-leaving radiances with colocated in situ measurements showed that the prelaunch calibration required adjustment from -2% to +13%.
Abstract: We assessed the geometric and radiometric performance of the ocean
color and temperature scanner (OCTS) using data acquired over the
United States. Initial results indicated a geometric offset in the
along-track direction of 4–5 pixels that was attributed to a tilt
bias. OCTS radiometric data appeared to suffer from near-field and
possibly far-field scatter effects. Analysis of radiometric
stability was inconclusive because of daily variability and the absence
of a full seasonal cycle. Comparison of OCTS-computed water-leaving
radiances with colocated in situ measurements showed that
the prelaunch calibration required adjustment from -2% to
+13%. Minor modification of OCTS data processing based on these
results and avoidance of near-field scatter effects can enable improved
and more-reliable OCTS data for quantitative scientific
analyses.
7 citations
TL;DR: Analysis of fast-ion D-alpha (FIDA) data on National Spherical Torus Experiment-Upgrade (NSTX-U) shows that the cold Dα line contaminates the FIDA baseline and both the Fida spectra and spatial profile are in better agreement with theoretical predictions.
Abstract: Analysis of fast-ion D-alpha (FIDA) data on National Spherical Torus Experiment-Upgrade (NSTX-U) shows that the cold Dα line contaminates the FIDA baseline. The scattered light is comparable to the FIDA emission. A scattering correction is required to extract the FIDA signal. Two methods that relate the scattered light contamination to the intensity of the cold Dα line are employed. One method uses laboratory measurements with a calibration lamp; the other method uses data acquired during plasma operation and singular value decomposition analysis. After correction, both the FIDA spectra and spatial profile are in better agreement with theoretical predictions.
4 citations
1 citations
References
More filters
01 Jan 1983
TL;DR: In this paper, the processing algorithms used for relating the apparent color of the ocean observed with the Coastal Zone Color Scanner on Nimbus-7 to the concentration of phytoplankton pigments (principally the pigment responsible for photosynthesis, chlorophyll a) are developed and discussed in detail.
Abstract: The processing algorithms used for relating the apparent color of the ocean observed with the Coastal-Zone Color Scanner on Nimbus-7 to the concentration of phytoplankton pigments (principally the pigment responsible for photosynthesis, chlorophyll a) are developed and discussed in detail. These algorithms are applied to the shelf and slope waters of the Middle Atlantic Bight and also to Sargasso Sea waters. In all, four images are examined, and the resulting pigment concentrations are compared to continuous measurements made along ship tracks. The results suggest that over the 0.08-1.5-mg/m3 range the error in the retrieved pigment concentration is of the order of 30-40% for a variety of atmospheric turbidities. In three direct comparisons between ship-measured and satellite-retrieved values of the water-leaving radiance the atmospheric correction algorithm retrieved the water-leaving radiance with an average error of approximately 10%. This atmospheric correction algorithm does not require any surface measurements for its application.
790 citations
TL;DR: The processing algorithms used for relating the apparent color of the ocean observed with the Coastal-Zone Color Scanner on Nimbus-7 to the concentration of phytoplankton pigments are developed and discussed in detail and suggest the error in the retrieved pigment concentration is of the order of 30-40% for a variety of atmospheric turbidities.
Abstract: The processing algorithms used for relating the apparent color of the ocean observed with the Coastal-Zone Color Scanner on Nimbus-7 to the concentration of phytoplankton pigments (principally the pigment responsible for photosynthesis, chlorophyll-a) are developed and discussed in detail. These algorithms are applied to the shelf and slope waters of the Middle Atlantic Bight and also to Sargasso Sea waters. In all, four images are examined, and the resulting pigment concentrations are compared to continuous measurements made along ship tracks. The results suggest that over the 0.08-1.5 mg/cu m range, the error in the retrieved pigment concentration is of the order of 30-40% for a variety of atmospheric turbidities. In three direct comparisons between ship-measured and satellite-retrieved values of the water-leaving radiance, the atmospheric correction algorithm retrieved the water-leaving radiance with an average error of about 10%. This atmospheric correction algorithm does not require any surface measurements for its application.
764 citations
TL;DR: The results suggest that for wind speeds ≲ 10-12 m/s, models that relate whitecap reflectance to wind speed are sufficiently accurate to meet the SeaWiFS accuracy goal for retrieval of the water-leaving radiance in the blue when the aerosol scattering is weakly dependent on wavelength.
Abstract: The effects of oceanic whitecaps on ocean-color imagery are simulated and inserted into the proposed Sea-Viewing Wide-Field-of-View Sensor (SeaWiFS) atmospheric-correction algorithm to understand its tolerance to error in the estimated whitecap contribution. The results suggest that for wind speeds ≲ 10–12 m/s, present models that relate whitecap reflectance to wind speed are sufficiently accurate to meet the SeaWiFS accuracy goal for retrieval of the water-leaving radiance in the blue, when the aerosol scattering is weakly dependent on wavelength. In contrast, when the aerosol scattering has a strong spectral signature, the retrievals will meet the goal only when the whitecap reflectance is underestimated.
207 citations
TL;DR: In this article, a closed-form solution for geolocation of satellite-based Earth sensor data has been developed, depending only upon the assumption of an ellipsoidal Earth model, and rely on vector and matrix algebra to improve computational efficiency.
Abstract: A closed-form solution for geolocation of satellite-based Earth sensor data has been developed. The algorithms are exact, depending only upon the assumption of an ellipsoidal Earth model, and rely on vector and matrix algebra to improve computational efficiency. The values computed are the geodetic coordinates of observed locations (latitude and longitude) and the sensor and solar angles required by radiative transfer models (sensor and solar zenith and azimuth angles). An improvement in computation speed of a factor of two is achieved over a spherical trigonometry solution, along with a much greater accuracy on an ellipsoidal Earth. As much as a factor of six increase in speed is obtained over the more accurate iterative solution for just geodetic latitude and longitude.
47 citations