GES DISC DAAC Data Guide:
SeaWiFS Level 1A and Level 2 HDF Dataset

Sea-viewing Wide Field-of-view Sensor (SeaWiFS)
Level 1A and Level 2 HDF
Dataset Guide Document
Version 2.0, June 2000

Coastal Regions of Chile and Argentina, Level 1A and Level 2 GAC Browse Imagery

Abstract:

The Sea-viewing Wide Field-of-view Sensor (SeaWiFS) is an eight-channel visible light radiometer dedicated to global ocean color measurements. The mission parameters of SeaWiFS allow coverage of more than 90% of the ocean surface every two days. SeaWiFS will map global ocean color at a resolution of 4.5 kilometers, and it also provides regional data at a resolution of 1 kilometer. SeaWiFS is the follow-on mission to the Coastal Zone Color Scanner (CZCS), and the predecessor to several ocean color satellite sensors scheduled for deployment in the years 1998-2002.

Table of Contents

Introduction
	Sample Images
	Abstract
Investigators
	Title of Investigation
	Contacts
Dataset Information
	Introduction
	Objectives/Purpose
	Summary of Level 1 Parameters
	Summary of Level 2 Geophysical Products
Theory of Measurements
Equipment
	Instrument Description
	Platform Description
	Key Variables
	Manufacturer of Instrument
	Calibration
Procedure
Observations
Data Granularity
Data Description
	Spatial Characteristics
	Temporal Characteristics
	Parameter/Variable Description/Definition
	HDF Overview
Data Analysis and Product Generation
	General Discussion
	Data Processing Sequence
Errors
Notes
Applications of the Dataset
Dataset Plans
References
Related Software
Data Access
Output Products and Availability
Glossary of Terms
List of Acronyms

2. Investigators:

Dr. Charles McClain, Project Scientist

Goddard Space Flight Center, Code 970.2

Greenbelt, MD 20771

(301)286-5377

Email: mcclain@calval.gsfc.nasa.gov

Dr. Wayne Esaias, MODIS Oceans Team Leader

Goddard Space Flight Center, Code 971

Greenbelt, MD 20771

(301) 614-5709

Email: wayne@puffin.gsfc.nasa.gov

Dr. Stanford Hooker, Field Program Manager

Goddard Space Flight Center, Code 971

Greenbelt, MD 20771

(301) 286-9503

Email: stan@ardbeg.gsfc.nasa.gov

Dr. Gene Feldman, Data System Manager

Goddard Space Flight Center, Code 610.2

Greenbelt, MD 20771

(301) 286-9428

Email: gene@seawifs.gsfc.nasa.gov

Title of Investigation:

Sea-viewing Wide Field-of-view Sensor

Contacts (Data Production)

Data:

Dr. Gene Feldman

Goddard Space Flight Center, Code 610.2

Greenbelt, MD 20771

(301)286-9428

email: gene@seawifs.gsfc.nasa.gov

Software:

Frederick Patt

SeaWiFS Project, Code 970.2

Goddard Space Flight Center

Greenbelt, MD 20771

(301) 286-2866

Email: Frederick.S.Patt.1@gsfc.nasa.gov

3. Dataset Information

Introduction:

This dataset consists of satellite measurements of global and regional ocean color data obtained by the Sea-viewing Wide Field-of-view Sensor (SeaWiFS), in orbit on the OrbView-2 (formerly "SeaStar") platform. The concentration and predominant identity of substances and particles in the euphotic (lighted) zone of the upper ocean influences the apparent color of the ocean, which can range from deep blue to varying shades of green and ruddy brown. Living phytoplankton (which contain chlorophyll and associated photosynthetic pigments), inorganic sediments, detritus (particulate organic matter), and dissolved organic matter all contribute to the color of the ocean.

OrbView-2.

Objectives/Purpose:

A quantitative determination of global ocean primary productivity is crucial to understanding the ocean's role in the global carbon cycle. The SeaWiFS mission parameters were designed with this goal as a fundamental consideration. In addition, SeaWiFS will refine remote-sensing measurements of phytoplankton chlorophyll and associated pigments, organic matter, and suspended particulate matter in the oceans. SeaWiFs will provide the first continuous observations of global ocean color, anticipating ocean color data from several future ocean remote sensing missions.

The Coastal Zone Color Scanner (CZCS) went far beyond its original status as a proof-of-concept mission. During its eight years of operation (November 1978-June 1986) the CZCS clearly demonstrated that measurements of ocean color from space were possible, and also proved that this technology could be used to characterize the distribution of biological productivity in the surface ocean. CZCS mission constraints, however, prevented quantitative determination of global ocean primary productivity.

NASA's ocean biogeochemistry research program, of which SeaWiFS is a critical element, has established a set of science goals. Data from SeaWiFS is expected to be used in the following ways:

• Goal I: Determine the spatial and temporal distributions of phytoplankton blooms, along with the magnitude and variability of primary production by marine phytoplankton on a global scale.
• Goal II: Quantify the ocean's role in the global carbon cycle and other biogeochemical cycles.
• Goal III: Identify and quantify relationships between ocean physics and large-scale patterns of biological productivity.
• Goal IV: Understand the fate of fluvial nutrients and their possible effect on marine carbon budgets.
• Goal V: Identify the large-scale spatial and temporal distribution of spring blooms in the global oceans.
• Goal VI: Acquire global data on marine optical properties, accompanied by an improved understanding of processes associated with mixing along the edges of eddies and boundary currents.
• Goal VII: Advance the scientific applications of ocean color data and the technical capabilities required for data processing, management, and analysis, in preparation for future missions.

One of the primary goals of the SeaWiFS Project is the following stringent objective:

"To achieve radiometric accuracy to within 5% absolute and 1% relative, water-leaving radiances to within 5% absolute, and chlorophyll a concentration to within 35% over the range 0.05 - 50.0 mg m^-3."

Summary of Level 1 Parameters:

Level 1 data consists of at-spacecraft raw radiance counts with calibration and navigation information available separately in the data file. The following table lists the center wavelength for each of the eight SeaWiFS bands, along with the primary use of each wavelength. SeaWiFS bands 1-6 are 20nm wide, and bands 7 and 8 are 40 nm wide.

 Band    Center Wavelength    	Primary Use
          nm    ("color")
          
  1       412   (violet)        Dissolved organic matter (incl. Gelbstoffe)
  2       443   (blue)        	Chlorophyll absorption
  3       490   (blue-green)  	Pigment absorption (Case 2), K(490)
  4       510   (blue-green)    Chlorophyll absorption
  5       555   (green)         Pigments, optical properties, sediments 		
  6       670   (red)           Atmospheric correction and sediments (CZCS heritage) 
  7       765   (near IR)     	Atmospheric correction, aerosol radiance
  8       865   (near IR)      	Atmospheric correction, aerosol radiance
 

[ See Glossary for definitions of terms. ]

Summary of Level 2 Geophysical Products

Level 2 data consists of five normalized water-leaving radiances (radiance data corrected for atmospheric light scattering and sun angles differing from nadir), and seven geophysical parameters derived from the radinace data. The following table lists the 11 SeaWiFS geophysical products.

	Normalized water-leaving radiances at:
	412 nm
	443 nm
	490 nm
	510 nm
	555 nm
	670 nm
	Chlorophyll a concentration
	K(490)
	Angstrom coefficient, 510-865 nm
	Epsilon of aerosol correction at 765 and 865 nm
	Aerosol optical thickness at 865 nm

[ See Glossary for definitions of terms. ]

4. Theory of Measurements

Ocean color remote sensing is based on the principle that particulate and dissolved substances suspended in water will interact with incident light. Where concentrations of particulate matter and dissolved substances are low, conditions typical for the open ocean, water molecules scatter light similar to the way that the atmosphere scatters light, producing a characteristic deep blue color. The scattering of light by particulates and the absorption of light by dissolved substances will alter this color. Chlorophyll, the photosynthetic pigment found in phytoplankton, absorbs strongly in the red and blue regions of the visible light spectrum and reflects in the green. As the concentration of phytoplankton increases, the color of the water will therefore appear increasingly green. The absorption of light by chlorophyll can be quantified to determine the concentration of chlorophyll in water, allowing estimation of phytoplankton abundance in a given area.

The relationship between light absorption and chlorophyll concentration may be complicated by the presence of light-scattering inorganic particulate matter in the water. Particulate matter concentrations generally increase in coastal regions, such that the water color near the coast trends from green to brown or reddish-brown. Even though chlorophyll may be present in higher concentrations near the coast, the presence of particulate matter makes it more difficult to extract the amount of light absorption due solley to chlorophyll. In addition, certain classes of phytoplankton form hard mineral shells that scatter light very effectively, such that the water color can appear shade of aquamarine or milky white.

SeaWiFS measures light intensity in several bands. The measurements allow quantification of light absorption and subsequent estimation of chlorophyll and suspended matter concentrations. SeaWiFS improves on the CZCS mission by having better bands for atmospheric correction (i.e., removing the effect of light scattering by the Earth's atmosphere), which will particularly aid the estimation of chlorophyll and suspended matter in coastal regions.

5. Equipment

Instrument Description

The primary optics of SeaWiFS consist of an off-axis folded telescope and a rotating half-angle mirror. Radiation backscattered by the Earth's surface and atmosphere is collected by the telescope and reflected onto the mirror, and the beam path is then directed through beam splitters (dichroics, which transmit some wavelengths and reflect the rest) to separate the radiation into four wavelength regions. Spectral bandpass filters are used to narrow these regions to the 20 nm requirements of the eight SeaWiFS spectral bands, and the radiation then falls on silicon detector elements. The electronics module amplifies the detector signal, performs analog-to-digital conversion and time delay and integration for data transmission. Instrument calibration utilizes an on-board solar radiation diffuser and lunar observation. The instrument may be tilted forward or backward 20 degrees along the spacecraft orbital trajectory to minimize the effects of sun glint.

SeaWiFS

Platform Description

The OrbView-2 satellite (formerly called "SeaStar") orbits in a sun-synchronous, descending node orbit at an altitude of 705 km. The orbital period is 98.9 minutes, with an inclination of 98.217 degrees. Local time of descending node is 12:05 PM + 15 minutes. The satellite was launched on August 1, 1997 into a 305 km orbit, and 32 orbit-raising burns performed over the next month raised the orbit to its final altitude.

The satellite has a three-axis stabilized system consisting of orthogonal magnetic torque rods for roll and yaw control and two momentum wheels for pitch stabilization. The satellite is equipped with sun sensors, horizon sensors, and magnetometers.

The propulsion system consists of two subsystems, a reaction control system and a hydrazine propulsion system. The reaction control system uses nitrogen and provides third stage stabilization during the launch. The hydrazine propulsion system is used for raising the orbit from the nominal 278 km parking orbit to the 705 km sun-synchronous operational orbit. In addition, it is used for orbit trim requirements over the life of the mission. The spacecraft employs four Hamilton Standard one pound thrusters.

Redundant global positioning system (GPS) receivers are used for orbit determination, an essential component of satellite and data navigation (Earth location). The orbit state derived from GPS is included in the spacecraft health telemetry.

Two telemetry streams are transmitted. The first is real-time LAC data merged with spacecraft health and instrument telemetry at 665.4 kbps. This is transmitted at L-band with a frequency of 1702.56 MHz. The other telemetry stream consists of stored GAC and selected LAC, along with spacecraft health and instrument telemetry, at 2.0 Mbps. This is transmitted at S-band with a frequency of 2272.5 MHz. The command system uses S-band with an uplink of 19.2 kbaud at 2092.59 MHz.

Key Variables

Nominal operating parameters for SeaWiFS:	

	Scan Width		58.3 deg (LAC); 45.0 deg (GAC)
	Scan Coverage		2,800 km (LAC); 1,500 km (GAC)
	Pixels along Scan	1,285 (LAC);  248 (GAC)
	Nadir Resolution 	1.13 km (LAC);  4.5 km (GAC)
	Scan Period		0.167 seconds
	Tilt			-20, 0, +20 deg
	Digitization		10 bits

Hughes Electronics Santa Barbara Research Center (Hughes SBRC)

Calibration

Pre-Launch Calibration

Due to the stringent radiometric objectives of the SeaWiFS Project, SeaWiFS underwent an extensive prelaunch calibration program. Calibration was performed at Hughes SBRC, and included an open air observation of the Sun for solar calibration purposes. (The preflight solar calibration is described in Chapter 3 of Volume 19 in the SeaWiFS Technical Report Series, NASA Technical Memorandum 104566.) The prelaunch characteristics of SeaWiFS were analyzed in detail to provide a comprehensive understanding of the sensor's radiometric response. SBRC employed a 100cm Spherical Integrating Source (SIS) which with a spectral shape equivalent to a 2,850 K blackbody for calibration purposes. Despite the approximate three-year hiatus between instrument completion and spacecraft integration, the calibration of the instrument was essentially unchanged over that time.

On-orbit calibration

SeaWiFS does not carry calibration lamps, but will rely on views of a solar radiation diffuser and the moon for radiometric calibration. The solar radiation diffuser is viewed once per orbit (near the southern terminator) to monitor sensor calibration over several orbits. The moon is viewed via a spacecraft maneuver to monitor calibration over months or years. Lunar views take place when the lunar phase is few days prior to or past full moon.

Data validation is accomplished by comparing data from the sensor to ocean optical data obtained during a series of calibration cruises, both prelaunch (commencing in 1992) and soon after the launch of the satellite. The data validation process also utilizes data from the Marine Optical Buoy (MOBY) moored off of the island of Lanai, Hawaii. MOBY is a moored in- and above-water optical radiometer that can transmit data to a receiving station on Lanai, allowing frequent comparison to data from SeaWiFS.

6. Procedure

Raw radiance data from the instrument is either transmitted directly to ground High Resolution Picture Transmission (HRPT) stations or recorded on the instrument for later transmission during downlink sessions to GSFC. Direct broadcast data is 1 km resolution LAC data. Recorded data is either 1 km resolution LAC data (primarily for calibration and validation) and 4.5 km resolution GAC data. The GAC data is used for the production of the global data set. Data is placed on disk and then processed to Level 1A, Level 2, and Level 3 products by the SeaWiFS Project. The data are then transmitted to the Goddard DAAC for archive and distribution. (HRPT data are processed to Level 1A by the receiving station, then sent to the SeaWiFS Project and subsequently to the DAAC.)

7. Observations

Missing Data

Although the SeaWiFS data set is essentially complete for the first 6 months of operation, there are a few noteworthy gaps in coverage. During the initial instrument check-out period from September 4 to September 18, 1997, the instrument only collected data intermittently. From October 14-18, 1997, the sensor was non-operational in "safe" mode due to transmission of incorrect navigation data. During the reprocessing of the data in January 1997, a small number of files were identified with large navigation errors that could not be corrected. These files were deleted from the data archive.

Several short safe-hold periods have occurred during the mission. During November 1999 and November 2000, the instrument was turned off for several days, and the satellite orbital configuration was adjusted to minimize the potential impacts of particles in the anticipated peak Leonid meteor shower.

Masks and Flags

The SeaWiFS land mask is a 1 km resolution dataset that overlies all land areas in the Level 2 product. Slight offsets may be found in the Level 1A LAC and HRPT data up to approximately 2 pixels (~2 km). The bathymetry flag marks all pixels from water depths less than 30 meters due to the possible influence of bottom reflectance. The cloud/ice mask covers all pixels above an albedo threshold. The cloud/ice mask may be misidentify shallow water areas with particularly high albedo as clouds.

The following is a list (also found in the Operational Archive Product Specifications document) of the algorithm name and corresponding flag condition for the l2_flags field in a SeaWiFS version 3 data file:

ATMFAIL		Atmospheric correction failure from invalid inputs
LAND		Land (present in pixel)
BADANC		Missing ancillary data
HIGLINT		Severe Sun glint
HILT		Total radiance above knee in any band
HISATZEN	Satellite zenith angle above limit
COASTZ		Shallow water
NEGLW		Negative water-leaving radiance in any band
STRAYLIGHT	Stray light contamination
CLDICE		Clouds and/or ice
COCCOLITH	Presence of coccolithophores
TURBIDW		Turbid, case-2 water
HISOLZEN	Solar zenith angle above limit
HITAU		High aerosol concentration
LOWLW		Low water-leaving radiance at 555 nm
CHLFAIL		Chlorophyll concentration not calculable
NAVWARN		Questionable navigation (e.g. tilt change)
ABSAER		Absorbing aerosol index above threshold
TRICHO		Presenc of Trichodesmium
MAXAERITER	Maximum iterations of NIR algorithm
MODGLINT	Moderate Sun glint
CHLWARN		Chlorophyll out-of-range
ATMWARN		Epsilon out-of-range
DARKPIXEL	Dark pixel (Lt - Lt < 0) for any band
SPARE		Spare flags

8. Data Granularity

Due to the variety of SeaWiFS data products, the definition of a data granule varies depending on the product. For Level 1A and Level 2 GAC data, the granule is a single swath of consecutive scans obtained as the satellite orbit moves from the north polar region to the south polar region. A single swath therefore contains approximately 40 minutes of data. A single Level 1A data file is ~21.9 MB, and the corresponding Level 2 file is ~19.1 MB.

For Level 1A HRPT data, the data granule contains the consecutive scans broadcast to the station while the satellite was within the station's receiving horizon. The 1 km resolution of HRPT LAC data makes the file sizes much larger, in the range of 50-110 MB.

9. Data Description

Spatial Characteristics

Coverage:

Spatial coverage is global, with full GAC coverage every two days. (Cloud cover limits actual imaging of the entire ocean surface to approximately every eight days, meaning that at least one view of any area on the ocean surface can be obtained in an eight-day period.) Coverage along the equator is slightly degraded due to instrument tilt to avoid sun glint effects.

Resolution:

L1A and Level 2 GAC has a spatial resolution at nadir of 4.5 km. Level 1A LAC data has a spatial resolution of 1.13 km, also at nadir.

Projection:

Level 1A HRPT, Level 1A GAC, and Level 2 GAC data have satellite swath projection.

Temporal Characteristics

Coverage:

The archive of SeaWiFS data products began on September 18, 1997. Partial orbit data was first obtained on September 4, 1997. Full-time operation of SeaWiFS obtains approximately 14.5 swaths of Level 1A and Level 2 GAC data per day, with each swath representing approximately 40 minutes of data (~20 MB).

Resolution:

The SeaWiFS rotating mirror rotates at a rate of 6 resolutions per second, or 0.167 seconds per revolution. GAC scans represent an Earth view of 21 milliseconds.

Parameter/Variable Description/Definition

SeaWiFS Level 1A data are raw radiance values for each of the eight SeaWiFS bands, with calibration and navigation information included in the HDF data file. Solar calibration and lunar calibration data are available as separate HDF files. As for the CZCS, a GAC pixel is generated by subsampling the LAC data every fourth scan line and every fourth pixel, to generate an effective GAC resolution of 4.5 km. Visible wavelength radiances are measured for each of the 20nm wide SeaWiFs bands. The bands are centered on the wavelengths given in the following table.

	
 Band   Center Wavelength    	Primary Use
         nm    ("color")

  1     412   (violet)        	Dissolved organic matter (incl. Gelbstoffe)
  2     443   (blue)        	Chlorophyll absorption
  3     490   (blue-green)  	Pigment absorption (Case 2), K(490)
  4     510   (blue-green)      Chlorophyll absorption
  5     555   (green)           Pigments, optical properties, sediments 		
  6     670   (red)           	Atmospheric correction and sediments (CZCS heritage)
  7     765   (near IR)     	Atmospheric correction, aerosol radiance
  8     865   (near IR)      	Atmospheric correction, aerosol radiance

Note: The SeaWiFS Level 1A browse product originally displayed the raw radiance counts from band 8. The new browse product displays a pseudo true-color image derived using data from bands 1, 5, and 6. The Level 1A GAC browseproduct is subsampled by a factor of 2 to produce a 9 km resolution image.

SeaWiFS Level 2 GAC data consist of the following 11 geophysical products:

Normalized water-leaving radiances at:

1. 412 nm
2. 443 nm
3. 490 nm
4. 510 nm
5. 555 nm
6. 670 nm

And the following derived products:

1. Chlorophyll a concentration
2. K(490)
3. Angstrom coefficient, 510-865 nm
4. Epsilon of aerosol correction at 765 and 865 nm
5. Aerosol optical thickness at 865 nm

Notes: The SeaWiFS Level 2 GAC browse product consists of the chlorophyll a concentration values along with appropriate masks, and is 9 km resolution (similar to the Level 1A browse product).

Normalized water-leaving radiances are corrected for the effects of light scattering in the atmosphere and for sun angles differing from nadir. Aerosol radiances are calculated as part of the atmospheric correction process. The atmospheric correction algorithm utilizes radiance data at 765 and 865 nm. The chlorophyll a algorithm builds on the experience of CZCS and ocean optical research during the intervening years, and will be subject to refinement in the course of calibration and validation research. For version 3 data files, the CZCS pigment geophysical product, which was simply the chlorophyll a value multiplied by a constant, was deleted from the standard archive product. (See Aiken et al., 1995.) K(490) also builds on the algorithms developed for CZCS data. The aerosol data products are derived from the atmospheric correction process. Version 3 data files add the Angstrom coefficient from 510-865 nm, which is sensitive to aerosols.

Units of Measurement

Radiance units are milliwatts per square centimeter per micron per steradian (mW cm^-2 µm^-2 sr^-2).

Concentration units (for pigment and chlorophyll) are in milligrams per cubic meter of seawater (mg m^-3). K(490) is expressed in units of reciprocal meters (m^-1).

The epsilon of the aerosol correction, aerosol optical thickness at 865 nm, and the Angstrom coefficient are dimensionless units.

Data Source

SeaWiFS orbits on the OrbView-2 satellite, which carried the name "SeaStar" during most of the SeaWiFS development period. OrbView-2 is a registered trademark of Orbimage Inc.

HDF Overview

All SeaWiFS data is available in Hierarchical Data Format (HDF), a data format developed by the National Center for Supercomputing Applications (NCSA). HDF is a "self-describing" data format, which means that all of the information necesssary to examine the data in an HDF file is contained within the file.

HDF has several different "data models" which are used to store data products. The data models that are used to store data are Scientific Data Sets (SDS), Raster Image Sets, Vgroups, and Vdatas. Global Attributes contain data that is applicable to the entire data file. An entire HDF file may be visualized schematically as a set of objects containing different data variables. Vgroups act as directories to data arrays, and they can contain SDS objects.

HDF Structure of SeaWiFS Level 1A Data

For Level 1A data, the main object is a Scientific Data Set (in this case, a three-dimensional array) containing the raw radiances measured at the spacecraft for each pixel in a scan line, for all 8 SeaWiFS bands. Thus, for a Level 1A LAC file, there are a variable number of scan lines, each containing 1,285 pixels, with 8 radiance values associated with each pixel. L1A GAC data has 248 pixels per scan line due to the GAC subsampling scheme. Each L1A GAC swath has approximately 3,600 scan lines, though this number is not exact.

The Vgroups for SeaWiFS Level 1A data are:

	Sensor Tilt 
	Converted Telemetry 
	Navigation
	Scan-Line Attributes 
	Raw SeaStar Data 
	Calibration

The Vgroup "Raw SeaStar Data" contains the data object lla_data, the three- dimensional array described above, as well as several variables that are specific to each scan line. (Refer to the PDF document SeaWiFS OPERATIONAL ARCHIVE PRODUCT SPECIFICATIONS for information on the variables of each of these Vgroups.)

SeaWiFS Level 1A Global Attributes:

Mission and Documentation

	Product Name
	Title
	Data Center
	Station Name
	Station Latitude
	Station Longitude
	Mission
	Mission Characteristics
	Sensor
	Sensor Characteristics
	Data Type
	Replacement Flag
	Software Name
	Software Version
	Processing Time
	Processing Control
	Input Parameters
	

Data Time

	Start Time
	End Time
	Scene Center Time
	Node Crossing Time
	Start Year
	Start Day
	Start Millisec
	End Year
	End Day
	End Millisec
	Start Node
	End Node
	Orbit Number

Data Quality

	Pixels per Scan Line
	Number of Scan Lines
	LAC Pixel Start Number
	LAc Pixel Subsampling
	Scene Center Scan Line
	Number of Scan Control Points
	Number of Pixel Control Points
	Mask Names
	Flag Percentages

Scene Coordinates

	Latitude Units
	Longitude Units
	Scene Center Latitude
	Scene Center Longitude
	Scene Center Solar Zenith
	Upper Left Latitude
	Upper Left Longitude
	Upper Right Latitude
	Upper Right Longitude
	Lower Left Latitude 
	Lower Left Longitude
	Lower Right Latitude
	Lower Right Longitude
	Northernmost Latitude
	Southernmost Latitude
	Westernmost Longitude
	Easternmost Longitude
	Start Center Latitude
	Start Center Longitude
	End Center Latitude
	End Center Longitude
	Orbit Node Longitude	

HDF Structure for SeaWiFS Level 2 GAC Data

The structure of SeaWiFS Level 2 data is very similar to the structure of Level 1A GAC data. There are only two new Global Attributes, both in the Data Quality category: Mask Names and Flag Percentages. The former indicates the different masks (such as clouds or land) that were used in the processing. The latter indicates how much of the data was flagged for not meeting specific quality criteria during processing from Level 1 to Level 2. (Note that Level 1A LAC data is not processed to Level 2.)

The primary new HDF object in the Level 2 GAC data is the Vgroup "Geophysical Data", which contains 11 data arrays corresponding to the pixel and scan line dimensions of the parent Level 1A GAC file. Each array contains individual pixel values for each of the 11 SeaWiFS Level 2 geophysical products.

Related Datasets

The primary precursor dataset to the SeaWiFS dataset is the eight-year archive collected by the CZCS. Other ocean color datasets are from the Ocean Color and Temperature Scanner (OCTS) and the Modular Optoelectronic Scanner (MOS). The Moderate Resolution Imaging Spectroradiometer (MODIS) is slated to begin operations in 1998, and several additional ocean color sensors are slated for launch in the period 1997-2002. Data from the Advanced Very High Resolution Radiometer (AVHRR), primarily used to observe sea surface temperature but also employed to observe turbid water masses, can be correlated with ocean color data. Sea surface wind datasets (derived either from remote sensing, meteorological instruments, or meteorological observations) can also be used in concert with ocean color data.

10. Data Analysis and Product Generation

General Discussion

The process of deriving accurate geophysical values from remote sensing radiance data is conceptually simple yet operationally complex. In principle, the instrument in space detects the intensity of light at various wavelengths of the electromagnetic spectrum. In the case of SeaWiFS, all of the wavelengths it detects are in the narrow segment of the spectrum that is visible to the human eye. The sole function of the instrument and its associated electronics is to quantify the light intensity, translate it into digital form, append data that allows the data to be navigated (i.e., determine the location on Earth from where the light originated), and send it to an Earth-based receiving station.

The remainder of the data analysis takes place on Earth. Algorithms developed on the basis of radiative transfer physics and both oceanographic and meteorological observation are employed to accurately extract the faint signal of backscattered light radiating from the ocean surface from the pervasive influence of scattered light in the atmosphere, an effect that is accentuated by the presence of atmospheric aerosol particles. The CZCS employed the assumption that no light radiated from the ocean surface at 670 nm, and thus all of the light detected was due to Rayleigh scattering from air molecules and aerosol scattering. SeaWiFS improves on this scheme by detecting light at 765 and 865 nm, as a small amount of light may actually radiate from the ocean at 670 nm. Furthermore, the atmospheric correction scheme used by SeaWiFS more accurately reproduces variable atmospheric conditions (Gordon and Wang 1994).

Once the radiance signal has been corrected for atmospheric light scattering, the signal is then corrected for the solar zenith angle to derive normalized water-leaving radiances. Normalized water-leaving radiances are subsequently used in algorithms to produce geophysical values. These algorithms were developed through oceanographic research into the optical characteristics of oceanic surface waters. As the most significant influences on the optical nature of oceanic waters are the presence of chlorophyll in phytoplankton and the presence of suspended particles, the algorithms use the water-leaving radiances to calculate the values of the related geophysical parameters. The geophysical parameters are calculated from the radiance values on a pixel-by-pixel basis, allowing the values to be mapped to Earth coordinates.

Several different methods have been employed to allow an accurate continuous assessment of instrument calibration. These methods were previously described in Section 5. The data analysis utilizes observations of the onboard solar diffuser and of the nearly-full moon for onboard instrument calibration. Data from the Marine Optical Buoy (MOBY) moored off of Lanai, Hawaii, is used to monitor the accuracy of "system calibration", which refers to the interaction of sensor data and scientific data processing to derive geophysical values that approximate reality.

Data Processing Sequence

SeaWiFS Level 0 data is digitized at 10 bits for transmission to ground stations. The primary data elements in Level 0 data are the raw radiance counts for all eight bands, accompanied by spacecraft and instrument telemetry. Processing to Level 1A appends calibration and navigation data to the file, as well as instrument and selected spacecraft telemetry. There are several different forms of Level 1A data: HRPT LAC, recorded LAC (which includes several types of calibration data), and GAC. A single GAC file consists of a swath data recorded from one north-to-south orbital pass, and constitutes one HDF file. A single HRPT file contains all of the scans received by the ground station while the satellite was above the station's receiving horizon. Recorded LAC scans, which are usually recorded for calibration and validation purposes as well as for regions of special research interest, contain the number of scan lines ordered by Mission Operations to cover the designated region.

Processing to Level 2 requires several additional steps. The data is navigated so that land masks may be correctly placed. Ancillary meteorological data and ozone data is used for atmospheric correction. The computational steps described earlier are employed to produce normalized water-leaving radiances and derived geophysical products. Each Level 2 data file is one HDF file, and corresponds exactly in temporal and spatial extent to the parent Level 1A file. Note that only Level 1A GAC data is processed to Level 2. Recorded LAC and HRPT LAC data is not processed to Level 2, but the SeaWiFS Data Analysis System (SeaDAS), discussed further in the "Related Software" section, will be capable of processing Level 1A data to Level 2 geophysical products.

11. Errors

SeaWiFS data includes masks and flags for anomalous conditions. The Level 1 to Level 2 data processing sequence tests for each of these conditions, and either flags the data or places the appropriate mask. Land, sun glint, and the presence of clouds and ice are masked.

The following is a list of the flag algorithms and masks with a brief description of the condition that invokes them:

Flags:

1. ATMFAIL
   Atmospheric correction failure due to invalid inputs.

2. HISOLZEN
   Pixels with solar zenith angle greater than 75 degrees cause uncertainty in atmospheric correction.

3. HISATZEN
   For pixels with pixel-to-spacecraft angles greater than 56 degrees, causing distorted pixel sizes.

4. STRAYLIGHT
   Stray light (instrument effect) occurring in proximity to very bright pixels (high L_t).

5. BADANC
   Missing ancillary data: indicates if interpolated rather than real values for ozone or surface meteorological data were used.

6. COASTZ
   Water depth shallower than 30 meters where bottom reflectance effects may occur.

7. HITAU
   High atmospheric turbidity indicator (aerosols), making atmospheric correction less reliable.

8. NEGLW
   Negative water-leaving radiance: The normalized water-leaving radiance will be set to zero for Level 3 binning. Can occur in cloud shadows, very high productivity Case 1 waters and turbid Case 2 waters.

9. LOWLW
   Normalized water-leaving radiance is below 0.15 mW cm^-2 µm^-2 sr^-2 at 555 nm. Indicates an anomalous condition as the water-leaving radiance is less than the clear water value.

10. COCCOLITH
    Indicates presence of coccolithophores (Brown and Yoder 1994).

11. TRICHO
    Indicates presence of Trichodesmium (Subramaniam et al., 2000)

12. TURBIDW
    Distinguishes Case 1 and Case 2 waters using an irradiance reflectance algorithm (Morel 1988).

13. CHLFAIL
    Failure of the semi-analytic chlorophyll algorithm. Chlorophyll concentration is set to zero for this case.

14. NAVWARN
    Questionable navigation; occurs most often in tilt segments.

15. ABSAER
    Absorbing aerosol index above a threshold of 0.5.

16. MAXAERITER
    The maximum number of iterations (10) for the NIR algorithm (Siegel et al., 2000) was exceeded.

17. CHLWARN
    Calculated chlorophyll values are out of range (0.001 - 64 mg m^-3).

18. DARKPIXEL
    Dark pixel: L_t - L_t < 0 for any band.

19. ATMWARN: Atmospheric Correction Failure
    Atmospheric correction algorithm fails to return epsilon values within a specified range.

Masks and Associated Flag Condition:

1. HILT: High L_t
   Radiances greater than the knee value in one or more bands, causing reduced precision.

2. LAND: Land
   The flagged pixel is over land.

3. CLDICE: Clouds and Ice
   Pixels with an albedo at 865 nm greater than 1.1%.

4. HIGLINT, MODGLINT: Sun Glint
   Pixels with a glint radiance greater than 0.005F₀(865) will be masked. F₀ is the extraterrestrial solar constant adjusted for Earth-Sun distance. MODGLINT indicates that a sun glint correction was applied, and is a flag; HIGLINT is a mask invoked by the condition defined above.

The conditions described above were designated for the purpose of identifying and marking conditions that can make the data unreliable for research purposes. After approximately four months of operation, the SeaWiFS Project evaluated the data that had been obtained by the instrument and concluded that that sensor had met the established criteria for mission success. As the SeaWiFS mission continues and the algorithms are refined, the reliability and integrity of the full dataset can be better assessed.

12. Notes

Several research projects managed by or in association with the Calibration/Validation element of the SeaWiFS Project are intended to analyze and refine SeaWiFS calibration and the algorithms used to produce Level 2 geophysical products. During January and February 1998, the entire dataset was reprocessed for the first time with improved algorithms and instrument calibration data. These algorithms provided the first science-quality SeaWiFS ocean color data, as the initial data produced with the at-launch algorithm set were considered preliminary results. The SeaWiFS Project plans to periodically reprocess the accumulated dataset using algorithms developed by ongoing research efforts.

The dataset has been reprocessed two subsequent times: from October 1998 to April 1999, and in May-June 2000.

As stated earlier, the goals of the SeaWiFS Project are stringent with regard to both radiometric accuracy and the accuracy of the primary geophysical products. The stated goal for chlorophyll concentration is an accuracy of 35% within the range 0.05 - 50.0 mg m^-3.

13. Applications of the Dataset

The main application of SeaWiFS data is research into the dynamics of phytoplankton populations in the surface waters of the ocean on a global scale. This research is important to investigations of carbon cycling in the ocean and the flux of carbon dioxide from the ocean to the atmosphere, which in turn is an important factor in the global climate system.

SeaWiFS data is also very useful into research on phytoplankton dynamics on the basin scale, mesocale, and even the fairly small scale of river outflow and phytoplankton distribution in large lakes. These types of investigations are very useful to marine ecosystem research, fisheries, and localized perturbations in the marine environment. Furthermore, SeaWiFS data on ocean clarity and turbidity can be used in investigations of sediment transport processes and various forms of anthropogenic input to the ocean.

On a fundamental level, SeaWiFS data allows a more detailed investigation of the optical physics of seawater on a global scale.

14. Future Dataset Plans

It is anticipated that the data from SeaWiFS will be augmented by data from the Moderate Resolution Imaging Spectroradiometer (MODIS), launched in December 1999 on the Terra platform. Several other countries have ocean color sensors in development that are slated for launch in the period 1998-2000. The Sensor Intercomparison and Merger for Biological and Interdisciplinary Oceanic Studies (SIMBIOS) project was formed for the purpose of intercalibration of ocean color data obtained by different ocean color instruments.

15. References

1. Aiken, J., G.F. Moore, C.C. Trees, S.B. Hooker, and D.K. Clark, 1995: The SeaWiFS CZCS-Type Pigment Algorithm. SeaWiFS Technical Report Series, Volume 29, NASA Technical Memorandum 104566, S.B. Hooker and E.R. Firestone Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 34 pages.
2. Biggar, S.F., P.N. Slater, K.J. Thome, A.W. Holmes, and R.A. Barnes, 1994: "Chapter 3: Preflight Solar-Based Calibration of SeaWiFS", IN: McClain, C.R., R.S. Fraser, J.T. McLean, M. Darzi, J.K. Firestone, F.S. Patt, B.D. Schieber, R.H. Woodward, E-n. Yeh, S. Mattoo, S.F. Biggar, P.N. Slater, K.J. Thome, A.W. Holmes, R.A. Barnes, and K.J. Voss, 1994: Case Studies for SeaWiFS Calibration and Validation, Part 2, SeaWiFS Technical Report Series, Volume 19, NASA Technical Memorandum 104566, S.B. Hooker, E.R. Firestone, and J.G. Acker, Eds., NASA Goddard Space Flight Center, Greenbelt, Maryland, 25-32.
3. Brown, C.W., and J.A. Yoder, 1994: Coccolithophorid blooms in the global ocean. J. Geophys. Res., 99, 7467-7482.
4. Gordon, H.R., and M. Wang, 1994: Retrieval of water-leaving radiance and aerosol optical thickness over the oceans with SeaWiFS: a preliminary algorithm. Appl. Opt., 33(3), 443-452.
5. Morel, A., 1988: Optical modeling of the upper ocean in relation to its biogenous matter content (Case I waters). J. Geophys. Res., 93, 10,749-10,768.
6. Siegel, D.A., M. Wang, S. Maritorena, and W.D. Robinson, 2000: Atmospheric correction of satellite color imagery: the black pixel assumption. Appl. Opt., (submitted).
7. Subramaniam, A., R.R. Hood, C.W. Brown, E.J. Carpenter, and D.G. Capone, 2000: A classification algorithm for mapping Trichodesmium blooms using SeaWiFS. Deep Sea Research, (submitted).

16. Related Software

SeaDAS was specifically developed for the processing and analysis of SeaWiFS HDF data. The following describes the what SeaDAS can do, as well as providing a path to obtain the software.

The SeaDAS Web site, http://seadas.gsfc.nasa.gov, is updated with current operating information for SeaDAS 4.0. System configuration and hardware requirements, and information on how to obtain SeaDAS 4.0, are given below and are also found on the SeaDAS Web site.

The SeaDAS software system was written for the specific purpose of analyzing and processing SeaWiFS HDF data. SeaDAS is a comprehensive image analysis package for all SeaWiFS data products and ancillary data (wind, surface pressure, humidity and ozone) from NMC (National Meteorological Center and TOVS (TIROS Operational Vertical Sounder). All SeaDAS source code is free and available for download via FTP.

The Interactive Data Language (IDL) from Research System Inc. (RSI) is used to build all the GUI and display related programs in SeaDAS. SeaDAS 4.0 is released with a blanket purchase of IDL-Runtime, so users do not have to acquire IDL or IDL-Runtime at their expense. SeaDAS includes the Hierarchical Data Format (HDF) libraries from National Center for Supercomputing Applications (NCSA) which are also required to build certain SeaDAS programs. IDL, C, FORTRAN77, and IMAKE from the vendors are required only if modifications to the source code and user-defined versions of the executables are desired.

Suggested Hardware Requirements:

• Platform: SGI 02, SUN UltraSparc workstations, or PC
• Memory:192 MB (regular users), 384 GB (HRPT users)
• Disk: 9 GB (actual SeaDAS installation requires ~330 MB without demo files and ~950 MB with demo files. Additional disk space is required for storing original data and processed data files.
• Tape Drive: 4MM(DAT) or 8mm Exabyte (for DAAC data)
• Display: 19" Console or X-terminal with 20 MB memory, 1280x1024 resolution, 8-bit, 256 colors

Software Requirements:

• Operating Systems: SGI: IRIX 6.3, or IRIX 6.5 SUN: Solaris 2.6 or Solaris 2.7
• Required Software: IDL-Runtime, IDL 5.1 or 5.2
• Languages: C (SGI V3.19, SUN V 3.0.1), FORTRAN(SGI V 4.0.2, SUN V 3.0.1), IDL 5.2 or 5.3 (IDL 5.1 may work but has not been re-tested)
• Software Libraries: HDF 4.1r1 (included in SeaDAS)

SeaDAS PC Linux Version

SeaDAS 4.0 for Linux/PC has been developed and tested under the following environment:

"Generic" PC with Pentium II 350 MHz CPU
Redhat Linux 6.0
IDL 5.1.2L for PC, IDL-Runtime

Compile-from-scratch support

• SGI: IRIX 6.3 and 6.5
• Sun: Solaris 2.6 and 2.7
• PC: RedHat Linux 6.0

Obtaining SeaDAS 4.0

SeaDAS is available for download via anonymous FTP from seadas.gsfc.nasa.gov. The /seadas directory contains the following compressed tar files:

• seadas_data1.tar.Z, seadas_data2.tar.Z, seadas_data3.tar.Z: SeaDAS required data files for L1, L2, and L3 processing
• seadas_demo.tar.Z: sample files for testing, and demonstrations
• seadas_irix6.3.tar.gz: SeaDAS for SGI IRIX 6.3 operating system
• seadas_irix6.5.tar.gz: SeaDAS for SGI IRIX 6.5 operating system
• seadas_solaris2.6.tar.gz: SeaDAS for SGI IRIX 6.3 operating system
• seadas_irix2.7.tar.gz: SeaDAS for SGI IRIX 6.3 operating system
• seadas_rhlinux6.0.tar.gz: SeaDAS for PC, RedHat Linux 6.0 operating system

Connect to the SeaDAS ftp site to download program files.

Note, some users, especially outside the U.S. have had trouble with the size of large SeaDAS files. Smaller "split" sections of the files have been created using the UNIX "split -20000" command. The split files can be found in the /seadas/split directory.

To put the split files back together (concatenate the files), use the following commands:
cat seadas_src.tar.Z?? > seadas_src.tar.Z (for source tar file), OR
cat seadas_data.tar.Z?? > seadas_data.tar.Z (for data tar file)

To put the pieces back together, and uncompress them, and unTAR the TAR file in one step:
cat seadas_src.tar.Z?? | zcat - | tar xvf - (for source files) OR
cat seadas_data.tar.Z?? | zcat - | tar xvf - (for data files)

SeaDAS can also be created on 4mm (DAT) or 8mm tape for those users who do not have Internet access or who have substantial difficulty with FTP of these large files. Please send your request to seadas@seadas.gsfc.nasa.gov.

SeaWiFS data has been successfully opened and examined using the Fortner Research prototype HDF Browser. The Research Systems Inc. software products Transform and Noesys (descriptions can also be found at the SciSpy Web site) have been used on SeaWiFS data files, and Noesys has been used to transfer SeaWiFS data to the EASI/PACE Geographical Information System (GIS) software package.

Two other software packages, HDF Explorer and Windows Image Manager, have also been used with SeaWiFS data files. Links to the sites where more information can be obtained are below. Windows Image Manager offers the capability of converting SeaWiFS data to many other image formats.

HDF Explorer

Windows Image Manager

17. Data Access

The SeaWiFS project is a "data buy" mission, for which NASA contracted with Orbital Sciences Corporation to build and launch the satellite. In return, Orbital Sciences Corporation was granted the opportunity to sell data from SeaWiFS for commercial applications. From September 18, 1997 to March 11, 1998, data from SeaWiFS was unrestricted. After March 11, 1998, the data is restricted to SeaWiFS Authorized Research Users solely for scientific research purposes. The data is subject to a two-week distribution embargo for normal research applications. Rea-time data access is granted to researchers and selected HRPT stations for specific research needs. After five years, all SeaWiFS data will be unrestricted.

In order to restrict data access to SeaWiFS Authorized Research Users, usernames and passwords are issued from the Goddard DAAC to each individual research user after they have provided the necessary documentation to the SeaWiFS Project.

Contacts for Archive/Data Access Information

Goddard DAAC Ocean Color Data and Resources Website

/OCDST/OB_main.html

Goddard DAAC Ocean Color Data Support Team
ocean@daac.gsfc.nasa.gov

James Acker, Team Lead
Code 610.2
NASA Goddard Space Flight Center
Greenbelt, MD 20771
USA
ocean@daac.gsfc.nasa.gov
301-614-5435
FAX: 301-614-5268

Goddard DAAC Helpdesk

Frances Bergmann
Code 610.2
NASA Goddard Space Flight Center
Greenbelt, MD 20771
USA
daacuso@daac.gsfc.nasa.gov
301-614-5224
FAX: 301-614-5268

Archive Identification:

NASA Goddard Space Flight Center DAAC

18. Output Products and Availability

Tape Media:

8mm EXABYTE tape (8200 and 8500 bpi)
4mm DAT tape (90m)

Other Products:

All SeaWiFS Products at the Goddard DAAC can be obtained by electronic transfer using the File Transfer Protocol (FTP.)

19. Glossary of Terms

20. List of Acronyms

	
AVHRR	Advanced Very High Resolution Radiometer
CZCS	Coastal Zone Color Scanner
DAAC	Distributed Active Archive Center
FTP	File Transfer Protocol
GAC	Global Area Coverage
HDF	Hierachical Data Format
HRPT	High Resolution Picture Transmission
LAC	Local Area Coverage
MOBY	Marine Optical Buoy
MODIS	Moderate Resolution Imaging Spectroradiometer
MOS	Modular Optoelectronic Scanner
NASA	National Aeronautics and Space Administration
NCSA	National Center for Supercomputing Applications
OCTS	Ocean Color and Temperature Scanner
SBRC	Santa Barbara Research Center
SDS	Scientific Data Sets

SIMBIOS	Sensor Intercomparison and Merger for Biological and 
		Interdisciplinary Oceanic Studies 

SIS	Spherical Integrating Source
SeaDAS	SeaWiFS Data Analysis System
SeaWiFS	Sea-viewing Wide Field-of-view Sensor