|
In the weeks following the August 1997 launch of SeaWiFS, several
pertinent questions were received by the Ocean Color Data Support Team (OCDST)
at the Goddard Distributed Active Archive Center (DAAC). Because the answers
to these questions were of potential interest to the expanding
community of ocean color data users, the following question-and-answer page
was composed. New and interesting questions sent to the OCDST will
occasionally be added here.
- What's ocean color?
- What's the difference between wave height, ocean color, and sea surface temperature images?
- What was CZCS?
- What is the difference between Level 1 GAC and Level 2 GAC?
- How can I get a GAC image showing a wider area?
- What software do I need to view images?
- What do I get when I order a file?
- Why is there an approximately one-week-long break in the data
coverage in October 1997?
- Why does the SeaWiFS Project data browser have files
that are not available from the DAAC?
- Can I get real-time data support from SeaWiFS?
- What does "reprocessing" mean?
- Can I get SeaWiFS data on a CD-ROM?
- What is a SeaWiFS data subscription?
- What information is necessary for a SeaWiFS data subscription?
- Why doesn't the DAAC have HRPT data that I can see on the
SeaWiFS Data Browser?
- Why is EPTOMS ozone data missing between December 12, 1998,
and January 3, 1999?
- Why do you have subsets of just the U.S. East Coast and the
Mediterannean Sea region?
- I'm writing a paper. How should I give credit for the
SeaWiFS data?
- How do I get data from two separate HRPT stations and put it
in the same order?
- I don't like the new data browser. Can I use the old one?
- What's the "Ocean Color Spectrum", and how can I get into it?
- Why isn't there any SeaWiFS data on November 17 and
18, 1999?"
- Why are there only two ancillary meteorological
data files per day in October and part of November, 1999?"
- How can the NDVI land data be complete over
South America in the summer if the tilt segments always take place over
the same region?
The most familiar images from the Coastal Zone Color Scanner (CZCS), and
now SeaWiFS, are false-color representations of phytoplankton pigment
concentration, calculated from the radiances (light intensities) measured by
the sensor. Sensors like CZCS and SeaWiFS measure radiance in specific bands
of the visible light spectrum, and the data is used to calculate such
variables as pigment concentration and light attenuation. False color is
simply used to emphasize the concentration ranges. For the CZCS, violet and
blue represent low concentrations (below 1.5 milligrams per cubic meter), green
and yellow are intermediate, and orange and red are high concentrations (up to
30 milligrams per cubic meter). The SeaWiFS color scale is similar, but the
improved data and better algorithms allow a better estimate of chlorophyll,
rather than pigment, concentration.
Phytoplankton pigment is predominantly chlorophyll, but also contains
degradation byproducts (phaeopigments). In addition, light-absorbing
substances or reflective particles in the water can "confuse" the algorithms
used to calculate the pigment concentration, particularly in coastal regions.
Newer sensors, such as SeaWiFS and MODIS (the Moderate Resolution Imaging
Spectroradiometer), have more spectral bands, i.e., they measure more "slices"
of the visible light spectrum, allowing more sophisticated algorithms and
better discrimination between chlorophyll, phytoplankton pigments, and other
substances and particles in the water.
One other note: the algorithms also involve correcting for light scattering
in the atmosphere. Atmospheric correction is quite important in the process,
as about 90% of the light received by a satellite- borne sensor is from the
atmosphere, and only about 10% from the ocean.
Ocean color is primarily used for research into phytoplankton, the base of
the oceanic food web and the main biological entity in the marine carbon cycle.
Phytoplankton are the plants of the ocean, producing organic carbon by
photosynthesis. Ocean color is therefore valuable in research into
biological systems and patterns in the ocean, but the data is quite dependent
on the algorithms used to derive pigment concentration and other geophysical
variables from the measured radiances. Considerable research effort with
SeaWiFS data will concentrate on algorithm development. For this reason, the
data is somewhat dependent on research conducted at sea to refine the
algorithms. As this research progresses, the reliability of the calculated
geophysical variables will improve. Use of the data also requires statistical
combination of several views of the same region to eliminate the pervasive
presence of clouds and haze.
Sea surface temperature (SST) measures the infrared radiation emitted by
the ocean surface. It is simply a measure of the intensity of this radiation
and relies less on in-situ verification. However, atmospheric correction and
comparison to in-situ sensors for calibration is still quite important.
Because oceanic currents and water masses can vary considerably in
temperature, SST data is particularly useful in observing currents and
circulation in the oceans. The data is quite sensitive to atmospheric effects,
and is also obscured by clouds. SST data is most commonly obtained from
polar-orbiting satellites operated by the National Oceanic and Atmospheric
Administration.
Wave height and sea surface height (SSH) are obtained from altimetry, which
measures a radar signal broadcast from the satellite and reflected back.
Precise timing of the signal indicates the distance between the satellite and
the surface of the ocean. Use of orbit and gravity models provides the actual
"height of the ocean", which indicates highs and lows in the ocean surface
that are somewhat analogous to meteorological highs and lows. The data can be
used to indicate large scale patterns of circulation and variability over
large scale. The current El Nino -- Southern Oscillation event is quite
visible in SSH data, as the high "dome" of heated water normally present in
the western Pacific essentially migrates eastward in the absence of the normal
wind pattern. Waves of warm water can be observed moving eastward. Wave
height, which can indicate both storm influence and regions of high variability
(such as the Agulhas retroflection) is derived from the spectral waveform of
the radar signal reflected from the ocean surface. The radar signal is
somewhat influenced by water vapor in the atmosphere, but this can be corrected
for with a separate instrument that measures specific water absorption bands in
the microwave region. The NASA and Centre National Etudes Spatiales (CNES)
satellite TOPEX/Poseidon has such
a microwave instrument, and the current accuracy of the SSH measurement is
better than 5 cm.
The Coastal Zone Color Scanner (CZCS) was a NASA sensor, flown on the
NIMBUS-7 satellite. It collected ocean color data from 1978 - 1986. For a good
overview of the sensor, go to the URL
/guides/GSFC/guide/CZCS_Sensor.gd.html
Though CZCS was a NASA instrument, calibration of the data relied on an
international team of scientists. Considerable work with the CZCS data was
also performed by research groups in France and Italy.
The GAC (Global Area Coverage) images that you are seeing are browse
images. The browse image for Level 1A data is made from the raw radiance
received at the satellite in band 8. This essentially allows land and clouds
to be distinguished from open ocean. Level 1 data can be processed to Level 2
using the free software package called SeaDAS, which can be accessed and
obtained from at the URL http://shark.gsfc.nasa.gov. Note
that SeaDAS only works on UNIX platforms at present and it requires an IDL
license.
NOTE: On October 16, 1997, the SeaWiFS Project changed the Level 1A
browse product to a true-color image that uses bands 1, 5, and 6. The browse
image should be much more pleasing to the eye, as well as more informative.
The SeaWiFS chlorophyll product is a Level 2 geophysical data product.
Remember that the current data uses an at-launch algorithm which will be
analyzed by the Calibration and Validation element of the SeaWiFS Project, and
improved algorithms will be used in the future. The entire database of
SeaWiFS data will be periodically reprocessed using new algorithms during the
course of the mission.
The only wide-area images are from the HRPT stations, which are LAC (Local
Area Coverage) data. LAC data is only available for level 1, and will require
SeaDAS (or software developed by the user) for processing. The GAC data is in
1500 km wide strips for a purpose - the pixel sizes are more uniform in the
center of the SeaWiFS scan. The width of a LAC scan is 2800 km, but the
pixels are elongated near the edge due to the curvature of the Earth.
If you prefer a very wide area, you can also look at the level 3 product,
which is a map of the whole Earth. The resolution is considerably reduced, of
course.
Yes, additional software is required. The image itself is a GIF file, so
you will need software capable of displaying GIF images. On a UNIX system, a
package like XV will work. There should be software packages capable of
displaying GIF images on a PC, and either Netscape or Internet Explorer
(properly configured) will also display the images.
If you download the browse image data, you are receiving a Hierarchical Data
Format (HDF) file that includes the image. 2008 update: The following site has a Windows HDF browser that allows viewing of HDF files: http://hdf.ncsa.uiuc.edu/plugins/UsersGuide/index.html
If you order data, you will receive the digital HDF files that contain all
the data obtained by SeaWiFS, and additional data that allows the data to be
processed (to add coastlines to an image, for example). The browse files are
smaller files with less information, so they cannot be used as effectively for
research.
The satellite went into "safe" mode during this period, because it received
erroneous navigation information. Virtually all satellites have this
contingency in their control software, which helps to prevent a catastrophic
failure. The instrument essentially goes to "sleep" until the ground
controllers can diagnose the problem and wake the instrument up again in the
proper fashion.
The SeaWiFS Project at Goddard Space Flight Center receives twice-daily
transmissions of data from the satellite. Once they have received the raw
data, they can immediately produce images and data files, and make them visible
on their data browser. In some cases a browse image will be produced, but it
will take longer to make the actual digital data file. The initial processed
data are then examined by the Calibration/Validation element of the SeaWiFS
Project to insure data integrity. Occasionally the data will be processed
again. The data files have to "pass" calibration and validation before they
are sent to the DAAC. It is common to have a 1-3 day delay from the time an
image can be seen on the SeaWiFS Project data browser to the time the data is
ingested at the DAAC and it is visible on the DAAC Web browser.
If the browse image is obtained very soon after it is first produced, it may
not have been processed with the best meteorological and ozone data available.
(These data types are called "ancillary" data, because they are not data from
the sensor but they are necessary for data processing.) Climatological data,
data which is the average value over year or longer timespan, may be used to
initially process the data. When the ancillary data corresponding to the time
the data was obtained is substituted, the appearance of the new browse image
may be slightly different.
Your research group can't get SeaWiFS data in real time from the DAAC, but
there is a group to contact for real-time data support. As part of the SIMBIOS
(Sensor Intercomparison and Merger for Biological and Interdisciplinary Ocean
Studies) Project, real-time images can be provided. It is necessary to be a
SeaWiFS Authorized Research User and to receive approval from SeaWiFS Project
Scientist Charles R. McClain. Researchers who receive this support are
expected to provide optical and/or pigment data to the SeaWiFS/SIMBIOS Projects
within six months of the cruise. For more detailed information and online data
request forms, go to the SIMBIOS/SeaWiFS
Support Services and Schedules page.
It is also possible to request recorded LAC data for a specific region
through the URL given above, because there is a limited amount of storage
(about 10 minutes of data daily) available onboard the satellite for
one-kilometer resolution LAC data. If a location and date are provided, LAC
recording can be scheduled for that area. This option might be important if a
study area is out of the receiving range of any HRPT station.
When the data comes from the satellite, it's just digital data -- strings
of numbers. The data has to be processed to derive meaningful remote sensing
data. The first step in SeaWiFS data processing is to produce "raw" radiance
data, the intensity of light for each band detected at the satellite. When
navigation information is added, this becomes Level 1A data. Algorithms are
then used to remove the effect of atmospheric light scattering and to account
for the angle of the sun, producing normalized water-leaving radiances. This
data can be combined in algorithms to produce other parameters, such as
chlorophyll concentration. This level of processing is called Level 2 data.
Level 3 data means that the data has been combined statistically into "bins"
which contain all the data from a certain region collected daily, weekly,
monthly, or annually. The SeaWiFS bins are square areas 9x9 kilometers in
dimension.
"Reprocessing" means that the SeaWiFS data has been processed again, using
new and improved algorithms, or by applying calibration data from the
instrument itself. When SeaWiFS was launched, the algorithms used to produce
the initial images were "at-launch" algorithms. In the early days of the
mission, the operating characteristics of the sensor were adjusted, and the
algorithms used to produce the data products were tested and refined by
comparing in situ data to data from the sensor. The definitions of
data flags (for conditions in which the data may be inaccurate) or masks (which
screen out clouds or land) were examined and improved as well.
The SeaWiFS Project began reprocessing the data in early January of 1998.
The reprocessed data is considered "science-quality" data, more accurate and
reliable than the data which was initially distributed during the last quarter
of 1997. The reprocessed data can now be obtained from the DAAC SeaWiFS data
browser. Of course, all of the data now being received is processed with the
new algorithms and calibration data. For a detailed description of the changes
involved, see SeaWiFS Data Reprocessing.
The current algorithms aren't the final word, however. The SeaWiFS Project
plans to reprocess the whole data set about once a year, as more research in
optical oceanography takes place. The goal is to make the current algorithms
even better.
CD-ROMs are not one of the current data distribution options for SeaWiFS
data. CD-ROMs hold a large volume of data, and although SeaWiFS is currently
producing a lot of data, it isn't enough (yet) to merit distribution on
CD-ROM. NASA customarily puts a comprehensive data set, such as a year of data
or a decade of data, on CD-ROMs. Very large amounts of data, such as from the
TOPEX/Poseidon altimetry mission, are also distributed this way. Because of
the costs of producing CD-ROMS, this option is usually used for data with a
great deal of potential interest and a large community of users. Because
SeaWiFS data is restricted to Authorized Research Users, it can't be
distributed as freely as other kinds of data -- at this time.
A data subscription means that an Authorized Research user has requested
that certain types of data be acquired for their use, and that the data should
be sent to them at regular intervals. The data subscription works this way:
when data is sent to the DAAC from the SeaWiFS Project, the subscription
software selects data corresponding to a subscription request. The data is
placed in a file corresponding to an individual researcher. When the time
interval is completed, the data is either written on magnetic tape and sent by
mail, or placed in an FTP directory. In the latter case, an email notification
is sent to the user that the data is ready to be transferred electronically.
First of all, data subscriptions are only available for Authorized Research
Users, so you must be registered with the SeaWiFS Project. Registration also
serves to put mailing address and email address information in our database.
Once that is accomplished, we need to know the following: the data products
desired in the subscription, how you would like to get the data (by FTP or on
magnetic tape), and what interval for data delivery is desired (daily, weekly,
or monthly). Many subscribers currently get data from individual HRPT
stations.
There are two possible reasons for this discrepancy. The most likely reason
is the two-week data embargo. The SeaWiFS Global Browse Utility, which is
intended to be used primarily by SeaWiFS Authorized Research Users for planning
purposes, displays GIF images from real-time HRPT stations, but no actual HDF
data files can be downloaded from there. The Goddard DAAC does not receive
any SeaWiFS data until the end of the two-week data embargo period. I.e., for
data received on May 1, the first day the DAAC could receive it would be May
15.
However, if the data embargo time period has expired, the other possible
reason is that the DAAC is not storing duplicate HRPT station data. If the
data collected by one HRPT station is entirely contained in the data from
another station, the DAAC will only receive the larger data file.
This situation might be a little difficult to visualize. As SeaWiFS
observes the Earth, it broadcasts to any station that can receive the data.
For example, as it passes over California, HRPT stations in Monterey, Los
Angeles, and even Texas can receive the satellite telemetry. (It's the same
data!) Because the station in Texas is farther from the satellite, and may
only receive a portion of the data that the Monterey station receives. And if
the antenna at the Monterey station is larger than the antenna at the station
in Los Angeles, the Monterey station may get all the data received at the Los
Angeles station, and more on either "end" of the data transmission. Thus, all
the data received at the Los Angeles station is included in the data received
by the Monterey station, and the DAAC will only get the data file from the
Monterey station.
The Earth Probe - Total Ozone Mapping Spectrometer (EPTOMS) satellite
suffered a serious navigational malfunction on December 12, 1998. Basically,
the navigational system indicated that the instrument was not properly
oriented. The satellite tried to re-orient itself before going into "safe"
mode, losing almost all of its maneuvering propellant. The ground controllers
were able to devise a new method of orienting the satellite (using magnetic
torque control) and the instrument was restored to normal science mode on
January 3, 1999. The TOMS Project cautions that the January 3rd data may
contain small errors. For more information about the TOMS Project, visit
their Web site,
http://toms.gsfc.nasa.gov.
The backup ozone data source is the TIROS Operational Vertical Sounder
(TOVS). Ancillary ozone data from TOVS is also available as SeaWiFS Ancillary
Data at the DAAC.
Our regional subsets were created to reduce both processing time and memory
storage requirements for areas (primarily coastal regions) where there is
considerable localized oceanographic and estuarine research activity. In
order to create them to be compatible with SeaDAS processing, we actually
create them using the subsetting function within SeaDAS. Therefore, it takes
considerable processing time to create them, and we chose areas where interest
in this type of research is particularly high. Remember that one of the
features of our regional subsets is that they can be directly downloaded by FTP
(they don't have to be ordered through the DAAC system). For more information
about the subsets, contact the OCDST.
You should use the following template as a guide:
"The authors would like to thank the SeaWiFS Project (Code 970.2) and the
Goddard Earth Sciences Data and Information Services Center/Distributed Active
Archive Center (Code 610.2) at the Goddard Space Flight Center, Greenbelt, MD
20771, for the production and distribution of these data, respectively. These
activities are sponsored by NASA's Earth Science Enterprise."
In addition, If you have used data from an HRPT station, please give an
attribution to that station.
The new browsing system offers several different ways to search for, and
order, data from a specific time period. The method to use primarily depends
on whether or not you wish to examine browse images of the data files.
Examining browse images and selecting specific data files reduces the volume
of data that will be sent to you, but it takes longer.
The first method is to select your specific time period. There are two
ways to do this -- one way is to choose data by year or by month by
"navigating" the calendars and checking the year(s) or month(s) you prefer.
If you want one specific time range in one year or one month, you can provide
the beginning and end dates below the calendar table. If the time range spans
the New Year (January 1), you'll have to order data from individual years
separately using this method. (You can even select data from the same range
of dates from different months using this method, perhaps for comparison to
data from a buoy that is collected at the same time each month.) The main
drawback to this method of ordering is that it orders the files in "blocks";
browse images can't be examined, and files can't be deleted from a "block".
The other way to select a specific time period is to use the "Temporal
Search" feature. Click on the words "Temporal Search" or the spinning globe.
Using temporal search allows the specification of ANY time range; just make
sure that the start date is earlier than the end date! The output is a list
of files from the HRPT station. The browse images can be examined so that
only the desired files are placed in the order, using the "Add to Order"
button. Skip the file without ordering by using "Next File", and "Continue"
when you're done examining the files. If you reach the last file on the list,
you will "Continue" automatically to the next step in the ordering process.
The information above covers how to search for data from one HRPT station.
To get data from two (or more) stations in the same order, you first have to
create a data order. Once you click on the "Create Order" button and you see
the phrase "The following has been added to your order:" and a descriptive
table below, the order has been created. "Continue Ordering" takes you back
to the first page for the HRPT station supplying the data you just ordered.
To get to a different station, we recommend clicking "Back to TOP", and
navigating back to the HRPT station selection page. Select the next station,
and begin ordering data by your favorite method.
Sorry, we can't put the old browser back "on-line" again. You might not
think so, but the new browsing system is better than the old one. One of the
main reasons is that now the Goddard DAAC has the same system for ordering data
from any dataset (ocean color data is only one of several Earth Science data
sets at the DAAC). Previously, there were several different kinds of data
browsers for the different data sets.
The new browser allows a lot more flexibility in the creation of data
orders, and it helps us save on resources and the "manual" steps needed to get
the data to you. Plus, it's modular, so we can make changes faster.
However, we understand that using the new browser might be confusing. If
you have questions about how to use it, you should email us (or call us) to
describe your specific requirements, and we'll suggest the best strategy to
use.
Just ask!
"Ocean Color Spectrum" is a collaborative effort between the Goddard DAAC
and the Alliance for Marine Remote
Sensing (AMRS), which publishes Backscatter . Besides being
published in the magazine, "Ocean Color Spectrum" also goes out to every email
address on the Goddard DAAC's distribution list, composed of marine scientists
associated with ocean color-related research.
You can submit any related information to James Acker at the
Goddard DAAC, acker@disc.gsfc.nasa.gov.
The Orbview-2 satellite was commanded to enter "safe" mode and aligned to
present the smallest possible impact profile, in order to safeguard against
possible damage from meteoritic particles associated with the Leonid meteor
shower. There is a similar data gap for the period November 17-20, 1998.
No. A fire damaged the main computer that NCEP used to
generate the ancillary data files. The two files per day were
generated by a slower backup system while the main computer
system was repaired.
The tilt segment (the period of time when the instrument is tilting, which
is done to minimize sun glint) is generally performed over the same region of
South America in the Northern Hemisphere summer so that critical sites for
in situ research in the Caribbean Sea will be imaged.
The NDVI images posted to the SeaWiFS Web site use max-NDVI compositing
(i.e., the highest NDVI value observed within each bin is selected). No
masking of any kind is done. Clouds are minimized by the fact that they have
low NDVI. The tilt is slow enough that we probably get some data in every 9-km
bin of the mapped data over a 1-week period.
|
 |
Atmos Composition
