[Next] [Prev] [Contents] [Top]

An Introduction to Atmospheric and Oceanographic Datasets


The first meteorological satellite, TIROS 1, was launched on 1 April 1960. Since that initial launch, the remote sensing of the atmosphere and earth's surface by satellites has made great progress and has contributed significantly to operational weather forecasts and our understanding of the atmosphere-ocean system. For example, satellites have been useful for measuring the areal extent of the polar ice packs and continental snow cover and their seasonal and interannual variability. Prior to satellite observations, these quantities which are important components of the climate system, had been difficult to quantify.

Measurements and Orbits

Satellite instruments detect radiances (i.e., electromagnetic energy over a finite range of wavelengths or, equivalently, frequencies). These radiances are the result of scattering, reflection or emission by the earth/atmosphere system. Although the instruments and computer algorithms are designed to minimize unwanted information (i.e., noise), the radiances received by satellite instruments contain both the desired signal and some noise. (Noise, in this discussion, means not only unwanted instrument/transmission signals but also unwanted signals from other geophysical variables.) The detected radiances are transmitted to centers which process and archive both the raw and processed data. The processing involves sophisticated `retrieval' algorithms (e.g., inversion techniques and/or statistical models) which can be used to derive estimates for a broad range of geophysical quantities. These quantities include, but are not limited to: sea surface and atmospheric temperatures, winds, water vapor, precipitation, cloud-cover, snow and/or ice cover, elements of the radiation budget, chemical components (e.g., ozone [o3], carbon dioxide [co2],...) and other physical quantities of interest at various space and time scales. These estimated quantities are used for visual weather displays, as input to operational weather forecast centers and for basic scientific research.

Most instruments on geophysical satellites are passive rather than active devices. A passive device detects radiances while an active device is the source of the energy. Active devices have been less frequently used for geophysical satellites because they require more power which usually limits instrument lifetime.

The sources of the radiances detected by passive satellite instruments are reflected solar energy and energy absorbed and reemitted by the earth/atmosphere system. The portions of the electromagnetic spectrum which are normally used for geophysical studies are: the ultraviolet (UV; ~ 0.0001-0.4 um) , the visible (VIS; ~ 0.4-0.74 um), the near infrared infrared (IR; sometimes called outgoing long wave radiation (OLR); ~ 0.74-100 um) and the microwave wavelengths (~ 100 um and longer). The wavelengths are inversely proportional to the temperature of the emitting body. Solar energy is contained within the shorter wavelengths (0.15-4 um) with about 9% in the UV range, 45% in the VIS range and the rest at the longer wavelengths. Various geophysical variables and chemical constituents of the atmosphere-ocean system emit radiation at the longer wavelengths due to much lower emitting temperatures. For example, the infrared and microwave wavelengths may be used to derive atmospheric and thermal information (e.g., the atmosphere's vertical temperature profile and cloud top temperatures). The visible wavelength range is mainly used to monitor cloud systems and infer precipitation.

Satellites may be categorized by their orbits. The two most common categories are polar-orbiting and geostationary (Figs. 5.1 and 5.2). Generally, polar-orbiting meteorological satellites maintain sun-synchronous orbits (an orbit whose plane is fixed relative to the sun). Sun-synchronous polar orbiters generally have low orbital altitudes (e.g., 850 km). At these altitudes it takes about 100 minutes to encircle the globe which yields 14-15 equatorial crossings per day. Sun-synchronous orbits are designed so that a satellite passes over a particular location at the equator twice per day at the same time every day. Being in a polar orbit does not necessarily mean that the satellite passes directly over the poles. Rather, their orbits are at an angle relative to the equator. This equator-crossing angle is called the inclination angle and it is a defined by a mission's objective. Operational meteorological satellites have large inclination angles (80 degree to 100 degree) while research satellites may have smaller inclination angles (Fig. 5.3). Operational polar orbiting satellites are principally used to obtain daily cloud cover, vertical temperature and water vapor distributions and global sea surface temperature. They are also used to receive/transmit data from moving platforms (e.g., drifting buoys or balloons) and determine their geographic position based upon the Doppler shift in the frequency received at the satellite.

Geostationary satellites have high altitudes (e.g., 36,000 km), maintain a fixed geographic location and orbit at the same speed that the earth rotates. They make observations at 20-30 minute intervals throughout each day. This allows continuous monitoring of a particular area of the earth. The size of the area monitored is a function of instrument design and satellite altitude. However, a single geostationary satellite can monitor about 25-30% of the earth's surface Figure 5.2 illustrates the areas covered by five geostationary satellites. Spatial resolution varies among the different satellites, ranging from 1 to 5 km (VIS) and from 5 to 8 km (IR). Geostationary satellites collect full images of VIS and IR each half hour and can do so more frequently if needed. Use of the IR allows images to be collected during the day or night. Uses of these data include locating and tracking tropical storms and deriving wind estimates which are used by operational weather forecasting models.

Both polar orbiting and geostationary satellites transmit data back to earth by radio. Polar orbiters record, and in some cases, preprocess the data prior to transmitting to earth. Geostationary satellites are always in contact with a ground station and need not record the data. Geostationary satellites provide two types of direct broadcast services: (i) a high-resolution transmission, most often used by researchers, and (ii) a low resolution transmission often called weather facsimile (WEFAX). Figures 5.4 and Chapters 3 and 4), which are site specific, irregularly spaced and of varying quality; satellite data represent spatial averages, cover wide areas and are of relatively consistent quality (if the calibration of the detecting instrument has remained stable). However, the absolute accuracy of the derived quantities is difficult to establish. Each meteorological satellite has different spatial sampling resolutions. For orbiting satellites, horizontal resolution is best along a satellite's track and is less over areas between tracks. As an example, the UARS, which is used to study the chemistry, dynamics, and energetics of the stratosphere, has a 500 km wide latitudinal swath.

One characteristic of raw satellite data is its very large volume. In general, NASA archives the experimental data and NOAA archives operational data. These data are available to researchers. However, the computing and storage requirements generally preclude individual researchers from doing their own processing. Thus, the research community uses the processed datasets which are of considerably smaller volume, yet may still be quite large.

Some sources of error for satellite measurements include: space-time sampling problems, instrumentation limitations, calibration drift, aerosol loading after volcanos, difficulty in removing noise, uncertainties in location, changes in equatorial crossing times and calibration of results with other observed variables. Sometimes (e.g., ERBE), the processed data are a combination of inversion methods and models which can make the data uncertainty estimates difficult to assess.

One difficult problem is how to assign various estimates to specific atmospheric altitudes. Simply stated, the problem is that the satellite instruments detect vertically integrated radiance values. Complex models exist that estimate how a particular vertical distribution of a geophysical variable emits radiation. However, there are a large number of complexities that make interpolation of a measured radiance value to specific atmospheric levels uncertain. Much work remains to be done but comparison studies with in situ data show that significant progress is being made.

Some Satellite Systems and Programs

TIROS and ESSA : The TIROS series (launch dates: 1960-63 for TIROS I-VIII; 1965 for TIROS IX and X) of satellites were the first to be launched specifically for atmospheric studies. There were essentially two sensors on board, an IR radiometer and a television camera. These satellites were spin-stabilized, meaning that the spin axis was fixed in space. Consequently, the instruments viewed the earth perpendicularly only once per orbit. This resulted in the use of elaborate spherical geometry to determine the latitude and longitude of the data. The later TIROS series, ESSA (1966-1969), was improved so that the spin axis was perpendicular to the orbital plane, enabling perpendicular measurements and easier rectification once per satellite rotation. The ESSA orbits were sun-synchronous.

NIMBUS: NIMBUS satellites (1964-78) wre used for cloud mapping and used wide field-of-view and fixed radiometers. However, the satellites were earth stabilized (as opposed to spin stabilized) so that the axis is always perpendicular to the surface of the Earth, and they occupied polar sun-synchronous orbits at approximately 1000 km altitude.

NOAA: NOAA (1970-present) operates a system of operational weather satellites. The first five were ITOS (Improved TIROS Operational System) type satellites, carrying three main instruments; two scanning radiometers (VIS and IR) and a Vertical Temperature Profile Radiometer (VTPR). These were polar orbiters that provided information on clouds, vertical temperature profiles, water vapor, outgoing long-wave radiation (OLR),etc.

TIROS-N: The third generation of operational meteorological polar-orbiting satellites was started by TIROS-N. The Stratospheric Sounding Unit (SSU) was provided by the United Kingdom, France provided the data collection system and it was launched by the U.S. in 1978. The first Microwave Sounding Unit (MSU) and a High Resolution Infra-Red Sounder (HIRS) were flown on this platform.

GOES: The first Geostationary Observational and Environmental Satellite (GOES-1) was launched in May 1974. The most recent launches occurred in March 1987 (GOES-7) and April 1994 (GOES-8). These NOAA satellites are usually located at 75 degree-west and 135 degree-west longitudes. The combination allows coverage of almost all of North America and South America and adjacent oceans (see Fig. 5.2). Originally, the GOES series provided only VIS and IR images. However, the instruments have improved in quality over the past 20-years allowing for more comprehensive examinations of the atmosphere. For example, the more recent GOES series provide the ability to infer the vertical profiles of temperature and moisture. This information is used by operational forecast centers.

Special Purpose Earth-Orbiting Satellites: Skylab demonstrated the ability to accurately measure the sea surface height (SSHT) from space. The SSHT is useful for geophysical studies because it is a proxy for the geoid, a surface of equal gravitational attraction. The anomalies in the gravitational field provide valuable information about changes in the density structure of the Earth. It is interesting to note that "sea level" is not the same over the globe. The SSHT near Sri-Lanka is more than 180 meters lower than near New Guinea, 6800 km away (granted, these are the extremes). The success of Skylab experiments spawned the use of radar altimeters on GEOS, Seasat, Geosat, TOPEX, and ERS-1.

Earth Radiation Budget Experiment: The ERBE satellite system consisted of an Earth Radiation Budget Satellite (ERBS), NOAA 9, and NOAA 10. The instrument package includes both scanning and nonscanning radiometers designed to provide high resolution, regional scale measurements of numerous radiative quantities. These provided a basis for long-term continent-scale monitoring of the radiation budget. More specifically, they measure the amount of solar radiation at the top of the atmosphere and the radiation emitted by the earth/atmosphere system. These are of fundamental importance since they determine the sources and sinks of energy that drive the climate system.

International Satellite Cloud Climatology Project : The purpose of ISCCP is to collect and analyze satellite observed radiances to infer the global distribution of radiative properties of clouds. The goal is to improve the modeling of cloud effects on climate. It uses data from the five geostationary satellites and polar orbiters of the NOAA/TIROS-N type.

European Remote Sensing: The ERS-1 satellite was launched in July 1991. It is a polar orbiting satellite operating at an altitude of about 785 km. It carries a number of instruments including an AMI scatterometer, which measures radar-backscatter from which wind estimates over the oceans may be derived. A second satellite (ERS-2) has recently been launched.

Table 5.1
Some Representative/Commonly Used NCAR Satellite Datasets
ds676.0TIROS N, NOAA1974-presdaily gridded VIS and IR brightnesses, OLR
ds684.1ECT-corrected1974-1999Global OLR Datasets
ds692.0NOAA series1972-1979vertical temperature profiles from 8 IR bands
ds700.0NOAA series1978-1992raw global radiances from MSU, SSU, TOVS (sounders)
ds701.0NOAA series1979-1993gridded mid-troposphere temperature from 53.74GHz
ds701.5NOAA series1979-presgridded precipitation from MSU channels (Spencer)
ds703.0NOAA series1989-1991Polar Orbiter Global (GAC) Data
ds710.0NIMBUS-71978-1986along-orbit ozone derived from backscattered UV
ds712.01974-1974reduced gridded dataset Atlantic brightnesses and IR
ds716.0INSAT1984-1989India: IR and VIS VHRR 2x, 8x daily
ds718.5NOAA series1974-1994monthly and half-monthly OLR
ds724.0Meteosat1993-presradiances: 2500x2500 pixels; start July 93; (G. Campbell)
ds725.0ISCCP1983-1987geostationary US cloud cover from radiances
ds727.1GEOSAT1986-1988along-orbit global ocean wind speed, wave height, SSHT
ds729.0SSMI1987-1991Chang's 5 degree precip estimates
ds733.0NIMBUS-71978-1987ERBE matrix of daily/monthly radiances
ds742.0ISCCP1983-1994global equal-area gridded cloud cover every 3 hours
ds744.0ERS-11991-1993along-track oceanic wind speeds and directions, SSHT

Measurements and Orbits
Satellite Data Characteristics
Some Satellite Systems and Programs

An Introduction to Atmospheric and Oceanographic Datasets
[Next] [Prev] [Contents] [Top]