Earth Radiation Budget Experiment Regional, Zonal, and Global Averages (S-4) Output Product Langley ASDC Data Set Document

Summary:

This document describes the Regional, Zonal, and Global Averages (S-4) Output Product and provides the user with the necessary information to use the Earth Radiation Budget Experiment (ERBE) data for scientific research studies.

The S-4 data set consists of scanner and nonscanner data processed for the months in which the scanners were operational. During this processing, the nonscanner data were processed using scene identification information (see Table 21) from the scanners.

Table of Contents:

1. Data Set Overview:

Data Set Identification:

ERBE_S4_NAT: Earth Radiation Budget Experiment (ERBE) S-4 (Scanner and Nonscanner) Regional, Global, and Zonal Averages of Radiant Flux and Albedo in Native (NAT) Format (ERBE_S4_NAT)

Data Set Introduction:

The S-4 contains averages of flux and albedo on regional, zonal, and global scales for both scanner and nonscanner data.

Objective/Purpose:

The objectives of ERBE are:

Summary of Parameters:

The S-4 contains averages of flux and albedo on regional, zonal, and global scales for both scanner and nonscanner data. It is available as a single or as a combination of all operational spacecraft (ERBS, NOAA-9, and NOAA-10) for the scanner data and the wide field-of-view (WFOV) data. It is available as a combination of the ERBS and NOAA-9 spacecraft for medium field-of-view (MFOV) data. Single satellite data will soon be available. The S-4 is a multifile product which has seven files of scanner data; scanner data are of 2.5 degree resolution and each region is nested to 5 and 10 degrees. The S-4 contains five files of MFOV numerical filter and five files of WFOV numerical filter. Numerical filter data are of 5 degree resolution and each region is nested to 10 degrees. Also on the S-4 are three files of MFOV shape factor and three files of WFOV shape factor data, which are of 10 degree resolution. Monthly (day), monthly (hour), daily, and monthly hourly averages are determined for each region. The data are represented in 8-, 16-, and 32-bit integers. The values contained are as follows:

Discussion:

The goal of the ERBE is to produce monthly averages of longwave and shortwave radiation parameters on the Earth at regional to global scales. Preflight mission analysis lead to a three-spacecraft system to provide the geographic and temporal sampling required to meet this goal. Three nearly identical sets of instruments were built and launched on three separate spacecraft. These instruments differ principally in the spacecraft interface electronics and in the field-of-view limiters for the nonscanner instruments because of differences in the spacecraft orbit altitudes.

The ERBS spacecraft was launched by Space Shuttle Challenger in October 1984 and was the first spacecraft to carry ERBE instruments into orbit. The ERBS was designed and built by Ball Aerospace Systems under contract to NASA Goddard Space Flight Center (GSFC), and ERBS was the first spacecraft dedicated to NASA science experiments to be launched by the Space Shuttle. The ERBS carries the Stratospheric Aerosol and Gas Experiment (SAGE II) instrument in addition to the ERBE instruments. The Payload Operation and Control Center (POCC) at GSFC directs operations of the ERBS spacecraft and the ERBE and SAGE II instruments, employing both ground stations and the Tracking and Data Relay Satellite System (TDRSS) network. Spacecraft and instrument telemetry data are received at GSFC where the data are processed by the Information Processing Division that provides ERBE and SAGE II experiment data to the NASA Langley Research Center (LaRC).

The second and third spacecraft launched with ERBE instruments are Television Infrared Radiometer Orbiting Satellite (TIROS) N-class spacecraft, which are part of the NOAA operational meteorological satellite series. The NOAA-9 and NOAA-10 spacecraft were launched in December 1984 and September 1986, respectively. The NOAA spacecraft include other instruments, such as the Advanced Very High Resolution Radiometer (AVHRR) and the High-Resolution Infrared Radiometer Sounder (HIRS), which provide NOAA with data for near-real-time weather forecasting. Both spacecraft are in nearly sun-synchronous orbits. The equator-crossing times (at launch) of the orbital nodes for the NOAA-9 and NOAA-10 orbits were 1420 UT (ascending) and 1930 UT (descending), respectively, where UT denotes universal time. The Satellite Operations and Control Center (SOCC) at the National Environmental Satellite and Data Information Service (NESDIS) operates the NOAA spacecraft. NOAA also provides decommutation processing of the telemetry data.

NASA tracks the ERBS spacecraft, and the North American Aerospace Defense Command (NORAD) tracks the NOAA spacecraft. The tracking data are provided to GSFC where orbit ephemeris data are calculated for all three spacecraft and provided to LaRC.

Related Data Sets:

2. Investigator(s):

Investigator(s) Name and Title:

Dr. Bruce R. Barkstrom
ERBE Principal Investigator
NASA Langley Research Center

Title of Investigation:

Earth Radiation Budget Experiment (ERBE)

Contact Information:

3. Theory of Measurements:

The theory behind the measurements made to collect the ERBE data is non-trivial and well beyond the scope of this document. However, interested readers are referred to the following publications: NASA Reference Publication 1184, Angular radiation models for Earth-atmosphere system, Volume 1: Shortwave radiation, and Volume 2: Longwave radiation; NASA Technical Paper 2670, Calculation and accuracy of ERBE scanner measurement locations; and Smith (Reference 15).

4. Equipment:

Sensor/Instrument Description:

Collection Environment:

All three sets of ERBE instruments were designed to collect data for one year but had a goal of two years. The nonscanner instruments continue to collect data for ERBS; however, the nonscanner instruments on-board NOAA-9 and NOAA-10 have been deactivated. Table 1 describes the nominal orbit parameters for each satellite at launch.

Source/Platform:

The ERBE instruments are on the ERBS, NOAA-9, and NOAA-10 satellites.

Source/Platform Mission Objectives:

ERBS was the first spacecraft dedicated to NASA science experiments to be launched by the Space Shuttle. ERBS carries SAGE II instruments in addition to the ERBE instruments. The NOAA spacecraft include other instruments, such as the Advanced Very High Resolution Radiometer (AVHRR) and the High-Resolution Infrared Radiometer Sounder (HIRS), which provide NOAA with data for near-real-time weather forecasting.

Key Variables:

A complete list of the measured parameters is found in Table 2.

Principles of Operation:

The ERBE is a multisatellite system designed to measure the Earth's radiation budget. The ERBE instruments fly on a mid-inclination NASA satellite, (ERBS), and two sun-synchronous NOAA satellites, (NOAA-9 and NOAA-10). Each satellite carries both a scanner and a nonscanner instrument package with characteristics listed in Table 2.

The scanner package contains three radiometric detectors each of which consists of an f/1.84 Cassegrain telescope. All are located within a single, rotating scan-head which, when operating in the cross track azimuth position, scans the Earth perpendicular to the satellite ground track from horizon to horizon. The scan-head can also be rotated in azimuth at a slow rate (0.9 degrees/second NOAA, 0.675 degrees/second ERBS). Each detector samples 74 measurements per scan. The total detector has no filter and so absorbs all wavelengths. The shortwave detector has a Suprasil-W1 filter which transmits only shortwave radiation. The longwave detector has a multilayer filter on a diamond substrate to reject shortwave and accept longwave radiation. To enhance the spectral flatness of the detectors, each thermistor chip is coated with a thin layer of black paint.

The nonscanner instrument package contains four Earth-viewing channels and a solar monitor. The Earth-viewing channels have two spatial resolutions: a horizon-to-horizon view of the Earth, and a field-of-view limited to about 1000 km in diameter. The former are called the wide field-of-view (WFOV) and the latter the medium field-of-view (MFOV) channels. For each of the two fields of view, there is a total spectral channel which is sensitive to all wavelengths and a shortwave channel which uses a high purity, fused silica filter dome to transmit only the shortwave radiation from 0.2 to 5 microns. The solar monitor is a direct descendant of the Solar Maximum Mission's Active Cavity Radiometer Irradiance Monitor detector. Because of the concern for spectral flatness and high accuracy, all five of the channels on the nonscanner package are active cavity radiometers.

Sensor/Instrument Measurement Geometry:

The nonscanner elevation beams can be rotated to any of three positions: launch/stow/internal calibration position (180 degrees), solar calibration position (78 degrees), and Earth-viewing (nadir) position (0 degrees). The WFOV detectors view the Earth from limb-to-limb (plus a small ring of space). The MFOV detectors are designed to include approximately an Earth view of 10 geocentric degrees within the unencumbered field of view (FOV).

The scanner can rotate in azimuth between 0 degrees and 180 degrees with an accuracy of 0.075 degrees. The normal scan mode is cross-track. The effective FOV of the scanner is 3 degrees.

Manufacturer of Sensor/Instrument:

The ERBE instruments were developed by TRW, Inc.

Calibration:

Specifications:

Not applicable.

Tolerance:

The tolerance is 1 percent for the total channel and 2 percent for the shortwave channel.

Frequency of Calibration:

For the scanner instruments, in-flight calibrations were accomplished every scan, as well as on a bi-weekly basis. In-flight calibrations of the nonscanners were normally performed on a bi-weekly basis.

Other Calibration Information:

The ERBE instruments were developed by TRW, Inc. Laboratory calibrations of the ERBE nonscanner and solar monitor instruments were completed in the TRW calibration facility at Redondo Beach, California in 1984. The fundamental standards used for the ERBE instruments were the International Pressure and Temperature Standard of 1968 (IPTS-68) and the World Radiation Reference (WRR). The TRW master reference blackbody (MRBB) was calibrated using these, and the MRBB was subsequently used to transfer the calibrations to the internal blackbody (IBB) and to the shortwave channels via an integrating sphere. The results of the calibrations were reported in detail in TRW calibration documents.

In-flight calibrations are performed in order to maintain the accuracy of radiometric measurements by accounting for internal instrument component parametric changes brought about by the spacecraft's environmental variables. In-flight calibrations of the nonscanners were normally performed on a bi-weekly basis. These included internal calibrations, space looks, and solar calibrations. Internal calibrations consist of cycling of IBB temperatures (total sensors) and shortwave internal calibration source (SWICS) voltages. Space looks consist of observations of "cold" space, both before and after solar calibrations. Solar calibrations consist of measurements made while the solar disc is within the instrument's field-of-view.

On days when internal calibrations are performed, shortwave offsets are independently determined in four ways:

In cases where the first option is not viable, the second option is used, along with a linearly-fitted delta based upon the historical differences between method 1 and method 2. The offsets determined using options 3 and 4 have never been used in production processing.

5. Data Acquisition Methods:

The ERBE nonscanner instrument consists of four Earth-viewing detectors and one solar monitor detector located on the head assembly. The four Earth-viewing detectors are unchopped active cavity radiometers (ACR), whereas the solar monitor is an unfiltered chopped ACR designed to measure direct solar radiation for calibrating the Earth-viewing detectors. Two of these detectors have wide field-of-view (WFOV) apertures allowing the detectors to view the entire disk of the Earth; the other two detectors have medium field-of-view (MFOV) apertures allowing the detectors to view an area about 1000 km in diameter. Two of the Earth-viewing detectors, one WFOV and one MFOV, and the solar monitor detector measure total radiation, whereas the other two Earth-viewing detectors measure shortwave radiation. The total radiation detectors are unfiltered, and the shortwave spectral bands are achieved by use of fused silica dome filters placed over the detectors.

The nonscanner instrument microprocessor processes and executes ground-commanded and stored commands to direct and control the instrument operations. The instrument can operate in several modes so that radiation measurements can be made over a wide range of operational conditions. The instrument can operate at azimuth angles between 0 and 180 degrees, and at fixed elevation beam positions of 0(nadir), 78 (solar ports), and 180 (stow or internal calibration position) degrees. Normal Earth-viewing operation is at the nadir elevation position and at an azimuth position of 180 degrees for NOAA-10, 170 degrees for NOAA-9, and 0 degrees for ERBS. The ERBE nonscanner instrument output consists of a complete cycle of radiometric and housekeeping measurements every 16 seconds. There are 20 radiometric measurements every 16 seconds, while the frequency of housekeeping measurements is either 1, 2, or 4 measurements per 16 seconds, depending on the type of measurement.

Telemetry data from the ERBE instruments on the NOAA-9 and NOAA-10 spacecraft are transmitted to Control and Data Acquisition (CDA) ground stations at Gilmore Creek, Alaska, and Wallops Island, Virginia that relay the data through a geostationary communications satellite to the SOCC at NESDIS in Suitland, Maryland. NOAA provides decommutation processing of the telemetry data and provides the data to LaRC. During portions of the ERBE mission, telemetry data from the NOAA spacecraft were transmitted to GSFC for decommutation processing and delivery to LaRC. Telemetry and tracking data from the ERBE instrument on ERBS are transmitted to the NASA ground terminal at White Sands, New Mexico through the Tracking and Data Relay Satellite System (TDRSS). The data are transmitted from the ground terminal to the NASA communications center at GSFC, where the data are processed by the Information Processing Division (IPD) that provides ERBE telemetry data to LaRC.

6. Observations:

Data Notes:

Not applicable.

Field Notes:

Not applicable.

7. Data Description:

Spatial Characteristics:

Spatial Coverage:

The spatial coverage differs with the channel and the spacecraft, as described below.

Spatial Coverage Map:

Though a map is not available, the limits of coverage are discussed in the Spatial Coverage Section.

Spatial Resolution:

The spatial resolution differs with the four types of instruments and the two types of spacecraft (ERBS and NOAA). The WFOV instruments have 136 degree FOV on ERBS and 126 degree FOV on the NOAA satellites. The MFOV instruments have footprints of approximately 5 geocentric degree radius or 1000 km at the TOA. The scanner instruments have an instantaneous hexagonal FOV with an angular size of 3 X 4.5 degree, which is equivalent to a 31 X 47 km footprint at nadir for ERBS and 44 X 65 km for NOAA. The solar instrument has an unencumbered FOV which observes the entire solar disk.

Gridded products of the scanner data are available in 2.5 X 2.5 degree resolutions. S-4 and S-4G scanner data are also available as 5 X 5 degree and 10 X 10 degree nested grids. 5 X 5 degree resolution and 10 X 10 degree nested grids are available for numerical filter nonscanner data, and 10 X 10 degree resolution is available for the shape factor nonscanner data on the S-4 output product.

Projection:

Gridding is an equal-angle projection of 2.5 X 2.5 degree (NFOV, 10368 bins), 5.0 X 5.0 degree (MFOV, 2592 bins), and 10.0 X 10.0 degree (WFOV, 648 bins).

Grid Description:

Binning of the data is based on an equal-angle grid of 2.5 X 2.5 degree (NFOV, 10368 bins), 5.0 X 5.0 degree (MFOV, 2592 bins), and 10.0 X 10.0 degree (WFOV, 648 bins). In each resolution, the bin number 1 is found at 90 degree N, 0 degree W with the bin number increasing in an easterly direction.

Temporal Characteristics:

Temporal Coverage:

Instruments on the three satellites (ERBS, NOAA-9, and NOAA-10) began acquiring Earth viewing data in November 1984, February 1985, and October 1986, respectively. All of the scanner instruments outlived their life expectancy of one year. The NOAA-9 scanner ceased operations on January 20, 1987 and the NOAA-10 scanner on May 22, 1989. The ERBS scanner ceased operations on February 28, 1990. All of the Earth-viewing nonscanner instruments collect measurements continuously except during calibrations. The solar instrument collects about 20 minutes of usable data during regularly scheduled bi-weekly solar calibration periods. Additional solar measurement data are sometimes obtained for special projects.

Temporal Coverage Map:

Table 3 shows the archival status of the S-4 product. The various combinations of the satellites reflect the actual duration of the scanners as indicated in Table 3. Note that the MFOV data were not processed.

Please consult the Langley ASDC IMS for available data granules for both single and combination satellites.

Temporal Resolution:

Data records for the Level 2 products are instantaneous measurements and estimates. Gridded data (the S-9, S-10, S-4, and S-4G products) are daily, monthly hour (hourly averages for a month), monthly day (daily averages for a month), and hourly.

Data Characteristics:

Parameter/Variable:

Although the data are stored in two different record formats (scanner and nonscanner), the data item definitions are the same for both. The main difference between the two formats is the presence of clear-sky data in the scanner records. Also, the nonscanner region numbers are calculated on a 5.0 or 10.0 grid versus a 2.5 grid for the scanner region numbers.

The definitions below apply directly to the input data (from Monthly Time/Space Averaging) found in files 4, 11, 16, 21, and 24 as listed in Table 11. These same data items in the remainder of the files are averaged values which have passed through the nesting process (files 5, 6, 12, and 17), the zonal averaging process (files 7, 8, 9, 13, 14, 18, 19, 22, and 25) or the global averaging process (files 10, 15, 20, 23, and 26).

The data items are divided into two groups: descriptive data items and scientific data items. The scientific data item are defined by data type (Monthly (Day), Monthly (Hour), Daily, or Monthly Hourly) rather than according to the order found in Table 16 and Table 17, to preserve the original order among the items. (See the Variable Description/Definition Section directly below).

Variable Description/Definition:

Descriptive Data Items

Scientific Data Items

In the following definitions, numbers in parentheses refer to equations in Reference 2.

The default or fill values for all missing data items are listed below in Table 5. Note that such values are not scaled by S-4 prior to being packed and written to the output tape.

Monthly (Day) Quantities

Monthly (Day) Quantities: These are monthly means based on daily calculations of flux. For longwave (LW) quantities, the daily means are obtained from the extrapolation, interpolation, and diurnal modeling algorithms that operate on the existing longwave measurements. The extrapolation and interpolation algorithms will, in general, cross daily boundaries, but the longwave diurnal model applied to land scenes operates on a specific day.

The shortwave (SW) quantities are based on calculations for specific days. The days are defined to be symmetric about solar noon.

Monthly (day) values for clear-sky: The previous quantities are calculated from all data including clear, partly cloudy, mostly cloudy, and overcast conditions. Clear-sky information (identified with a CS subscript) for longwave (means and statistics defined in Reference 2) is calculated in the Inversion Subsystem and passed through Monthly Time/Space Averaging without modification. The shortwave clear-sky values are calculated by Monthly Time/Space Averaging according to the distribution of cloud conditions as indicated by the scene fraction vector.

Monthly (Hour) Values

Monthly (Hour) Quantities: These items are monthly means based on values averaged over the month at each local hour. In general, they result in different values for the same quantity, compared to the monthly (day) means.

Monthly (hour) values for clear-sky: Clear-sky information for longwave (means and statistics defined in Reference 2) is based on values calculated in the Inversion Subsystem. The shortwave clear-sky values are calculated by Monthly Time/Space Averaging according to the distribution of cloud conditions as indicated by the scene fraction vector.

Daily Values

Daily values: These quantities are calculated for each day in the month.

Daily values for clear-sky: The above values are repeated for clear-sky conditions (except for the SOLARD). The LW clear-sky values are passed from the Inversion Subsystem, but the SW values are calculated within Monthly Time/Space Averaging.

Monthly Hourly Values

Monthly hourly values: These values are calculated for the month at each local hour.

Monthly hourly values for clear-sky: The above values are repeated for clear-sky conditions as defined by the Inversion Subsystem.

Geotype

Scanner:
An integer from 1-5 denoting the surface type for the region. The types are:

1 = ocean
2 = land
3 = snow
4 = desert
5 = land/ocean mix (coastal regions)

For the land/ocean mix, the corresponding directional models (clear, partly, or mostly cloudy over this scene) are linear composites of land and ocean models and not independent models.

Nonscanner:
The fraction of cloud-free (as determined by the Inversion Subsystem) land and desert geotype. If greater than 0.5, the half-sine model is applied in the calculation of LWF (Reference 2).

Unit of Measurement:

Units of measurement for the calculated and measured science variables for the S-4 data product can be found in the Variable Description/Definition Section of this document.

Data Source:

The purpose of the S-4 Output Product is to provide averages of radiant flux values and albedos using data from the Monthly Time/Space Averaging Subsystem (Reference 4) on a regional, zonal, and global basis. The basic structure of the ERBE grid system lends itself to calculating several types of averages. Reference 1 provides a detailed description of the ERBE grid system and gives the necessary background information for the design of this subsystem.

The S-4 product contains data which have been averaged to 2.5, 5.0, and 10.0 degree grid scales. The layout of a 2.5 degree system is given; the 5.0 and 10.0 degree systems are designed similarly. In this grid system, L = longitude and = latitude is replaced with colatitude, where = 90 - , so that .

The following list shows the number of regions for each resolution:

To facilitate comparison with nonscanner data, the scanner data are nested with area weighting to 5.0 and 10.0 degree regions, with four scanner 2.5 degree regions producing a 5.0 degree region and sixteen 2.5 degree regions producing a 10.0 degree region. Similarly, four nonscanner numerical filter 5.0 degree regions are nested to produce a 10.0 degree region. This nesting is pictured in Figure 2, and the weighting is described further on in this section.

The S-4 product also contains averages over the latitudinal bands (zones). The following list shows the number of latitudinal bands for each resolution:

The equation below gives the formula for calculating averages:
where,
= nested average flux value
N = number of regions included in nested average
W_i = area weighting factor
M_i = individual values

The final type of average is on a global level. Each parameter is averaged over the entire globe with area weighting.

For each of the three ERBE spacecraft, (ERBS, NOAA-9, and NOAA-10), there are three sets of measurements (scanner, nonscanner MFOV, and nonscanner WFOV). For the nonscanner measurements, there are two data reduction techniques (shape factor and numerical filter). For each satellite, the on-line processing proceeds as follows:

Table 6 provides a summary of the type of data available in the S-4 output product.

Processing is controlled by the lower resolution region numbers. A simple set of calculations can be used to derive the four higher resolution region numbers which will be nested into the lower resolution region (see Figure 2). The formulas for finding the four 2.5 degree region box numbers which are nested into a 5.0 degree region box are:

The formulas for finding the four 5.0 degree region box numbers which are nested into a 10.0 degree region box are:

where,
B₁₀(N) = 10.0° region box number

The colatitude index is used to obtain the correct area weighting value and to distinguish between polar latitudinal bands and nonpolar latitudinal bands. It can be derived from the region number as follows:

COLAT = INT[B₆ - 1)/N] + 1

where,
B₆ = higher resolution region number
N = number of regions in a latitudinal band

The 2.5 degree product is used to nest the 5.0 degree product which is then used to nest to the 10.0 degree product. This process is done for each satellite and each combination of satellites. For each of these processes, separate zonal and global products are produced for each resolution,

where for any data product, designates the source and final product resolution. Analogously, there will also be similar paths for numerical filter data:

and

and for shape factor data:

and

So, for each satellite and combination of satellites, there are potentially nine different global (monthly) averages of each calculated quantity. Why is there a difference between, for example, G(2.5-2.5) and G(2.5-5)? If there were no missing data on the high resolution grid, these numbers should be the same. However, missing data forces nesting procedures which can produce some discrepancies between the two products.

Because the ERBE grid system divides the globe into regions which are defined by equal increments of latitude and longitude rather than equal areas, the nested averages must be normalized by weighting each region based upon its area. Since regions across a latitudinal band have the same area, only one weighting factor is needed for each latitudinal band. It also follows that since the area weighting factors across a latitudinal band are the same, the zonal averages do not have to be normalized. The following equation gives the formula for calculating the area weighting factors.

where,
W_i = area of by region in km²
= resolution in degrees (2.5, 5.0, 10.0)
R = distance from the center of the Earth to the top-of-the atmosphere (km)
= colatitude in degrees of the center of the latitudinal band region

Polar day-night indicators are used to identify those regions within approximately 23.5 degrees of the poles that experience continuous darkness or continuous daylight at certain times of the year. These regions are treated differently from those which always experience day-night cycles.

In general, radiant flux values for regions not observed by the satellite are not accumulated as part of the averaging process. However, since it is known that the shortwave radiant flux is zero when there is no daylight, those unobserved regions which are in complete darkness for the entire month (i.e., near the polar regions at certain times of the year), will have the shortwave set to zero. The reason for doing this is to reduce the error in the zonal and global averaging process.

There is a direct relationship between the daily solar declination and the colatitude of a region. This can be used to determine whether or not that region is experiencing total darkness or has some daylight and which part of the month is affected. Of course, none of this has any effect on regions in latitudinal bands which are not near the north or south poles.

The criteria for setting the polar day-night indicators for latitudes in the northern hemisphere are:

The criteria for setting the polar day-night indicators for latitudes in the southern hemisphere are:

In order to clarify this concept, some tables have been provided to illustrate the results of applying the logic described above. Table 7 shows which colatitudes are considered to be the northern and southern polar latitudes for each resolution.

Table 8 shows the solar declinations for 1985, which was chosen as the example year.

Table 9 shows the sunlit days for the northern and southern polar regions for the 2.5 degree resolution.

Table 10 gives a sample of the polar day-night indicator values for some latitudes at the 2.5 degree resolution.

Data Range:

Please refer to the Temporal Coverage Map Section of this document for the archival status of the ERBE S-4 product.

Sample Data Record:

ERBE data records are stored as binary integers in the ERBE native format. When ordering this data set, the user has the option of receiving the sample read software for this data set. The "README" file for executing this software provides the necessary information to obtain data records from this data set.

8. Data Organization:

Data Granularity:

A general description of data granularity as it applies to the IMS appears in the EOSDIS Glossary.

The S-4 Output Product contains two types of data files: scanner files (files 4 through 10 in Table 11) which consist of the scanner input data file, and the resulting nested, zonal, and global data files; and nonscanner files (files 11 through 26 in Table 11) which consist of the nonscanner input data files and the resulting nested (if any), zonal, and global file(s).

The maximum number of records possible for each file is shown in Table 11. The actual number of records will depend upon which regions were represented in the input files. For example, if regions 1 through 144 were not present on the scanner input tape, then the actual number of records present on the S-4 Output Product will be 10224. This in turn, may affect the nesting process and may affect the total number of records produced. The scanner files normally include clear-sky data and have a record format which is different than that for the nonscanner files. These formats are discussed below.

*yy represents the year (e.g., 89 - 1989); mm represents the number value of a month (e.g., 01 = January, 12 = December)

s represents the satellite code:
1 = NOAA-9
2 = ERBS
3 = NOAA-10
4 = NOAA-9/NOAA-10
5 = ERBS/NOAA-9
6 = ERBS/NOAA-10
7 = NOAA-9/ERBS/NOAA-10

Scanner Data Records

The cumulative total bits for a single scanner record (or the scanner scale factor record) is 10320 bits. These are grouped into 32-, 16-, and 8-bit words as shown in Table 12.

The actual data items are listed in Table 16 in the same order as they appear on tape and are defined in the Variable Description/Definition Section. This format is used in files 4 through 10 as listed in Table 11.

Nonscanner Data Records

The cumulative total bits for a single nonscanner record (or the nonscanner scale factor record) is 5760 bits. These are grouped in 32-, 16-, and 8-bit words as shown below in Table 13.

The actual data items are listed in Table 17 in the same order as they appear on tape and are defined in the Variable Description/Definition Section. This format is used in files 11 through 26 as listed in Table 11.

Data Format:

The S-4 Output Product contains 26 files as shown in Table 11. The first file contains the standard ERBE header record, the second file contains the scanner scale factors, and the third file contains the nonscanner scale factors. Files 4 through 26 contain the data records.

The archival status of this product at the Langely ASDC is found in the Temporal Coverage Map Section. The S-4 files can be obtained on 9-track, 8mm, or 4mm tape media or as electronic disk files via FTP. When the user is connected to the on-line Langley ASDC system, he will be able to select a particular data set pertaining to the ERBE S-4 data in which he is interested. The name of the S-4 set, which the user will see as he orders his data is ERBE_S4_NAT. The names of the S-4 files which the user will receive from the ASDC are listed in column 2 of Table 11. Column 3 of the same table gives a description of each file.

The first file on the S-4 Output Product contains one header record which identifies the data on the tape. It is a 30-byte record formatted as 8-bit bytes and defined in Table 14.

Table 15 shows an example of the S-4 header.

The second file contains the integer scanner scale factors. This file contains the same number of scale factor values as there are scanner data values. Table 16 lists a scanner output record and includes a column which lists the corresponding scanner scale factors.

Table 16: Packing Scheme for the Scanner Output Record

Item Name	Data Type	No. of Values per Region	Starting Item No.	Ending Item No.	Scale Factor	Cumulative No. Bits	Minimum Value	Maximum Value
32-bit words:
FILE ID	-	1	1	1	1	32	1	7
TSOLRD	MONTHLY (DAY)	1	2	2	100	64	0	500000
	MONTHLY (DAY)	1	3	3	100	96	-200	200
CS	MONTHLY (DAY)	1	4	4	100	128	-200	200
TSOLRDCS	MONTHLY (DAY)	1	5	5	100	160	0	500000
	MONTHLY (HOUR)	1	6	6	100	192	-200	200
TSOLRH	MONTHLY (HOUR)	1	7	7	100	224	0	500000
CS	MONTHLY (HOUR)	1	8	8	100	256	-200	200
TSOLRHCS	MONTHLY (HOUR)	1	9	9	100	288	0	500000
SOLARD	DAILY	31	10	40	100	1280	0	500000
SOLARH	HOURLY	24	41	64	100	2048	0	500000
SOLARHCS	HOURLY	24	65	88	100	2816	0	500000
SPARE	-	2	89	90	1	2880	127	127
Total number of bits = 32 x 90 = 2880
16-bit words:
Region #	-	1	91	91	1	2896	1	10368
Date: YYMM	-	1	92	92	1	2912	8411	to present
Spacecraft ID	-	1	93	93	1	2928	1	7
	MONTHLY (DAY)	1	94	94	10	2944	0	400
	MONTHLY (DAY)	1	95	95	10	2960	0	800
ALBEDO	MONTHLY (DAY)	1	96	96	1000	2976	0	1
CS	MONTHLY (DAY)	1	97	97	10	2992	0	400
CS	MONTHLY (DAY)	1	98	98	10	3008	0	800
ALBEDOCS	MONTHLY (DAY)	1	99	99	1000	3024	0	1
	MONTHLY (HOUR)	1	100	100	10	3040	0	400
	MONTHLY (HOUR)	1	101	101	10	3056	0	800
ALBEDO	MONTHLY (HOUR)	1	102	102	1000	3072	0	1
CS	MONTHLY (HOUR)	1	103	103	10	3088	0	400
CS	MONTHLY (HOUR)	1	104	104	10	3104	0	800
ALBEDOCS	MONTHLY (HOUR)	1	105	105	1000	3120	0	1
	DAILY	31	106	136	10	3616	0	400
	DAILY	31	137	167	10	4112	0	800
ALBEDO	DAILY	31	168	198	1000	4608	0	1
CS	DAILY	31	199	229	10	5104	0	400
CS	DAILY	31	230	260	10	5600	0	800
ALBEDOCS	DAILY	31	261	291	1000	6096	0	1
	HOURLY	24	292	315	10	6480	0	400
	HOURLY	24	316	339	10	6864	0	1000
ALBEDO	HOURLY	24	340	363	1000	7248	0	1
CS	HOURLY	24	364	387	10	7632	0	400
CS	HOURLY	24	388	411	10	8016	0	1000
ALBEDOCS	HOURLY	24	412	435	1000	8400	0	1
Total number of bits = 16 x (435 - 90) = 5520
8-bit words:
NDLW	DAILY	31	436	466	1	8648	0	24
NDSW	DAILY	31	467	497	1	8896	0	24
NDLWCS	DAILY	31	498	528	1	9144	0	24
NDSWCS	DAILY	31	529	559	1	9392	0	24
NHLW	HOURLY	24	560	583	1	9584	0	31
NHSW	HOURLY	24	584	607	1	9776	0	31
NHLWCS	HOURLY	24	608	631	1	9968	0	31
NHSWCS	HOURLY	24	632	655	1	10160	0	31
SPARES	-	20	656	675	1	10320	127	127
Total number of bits = 8 x (675 - 435) = 1920

NOTE: CS at the end of an item name indicates that the item is for clear-sky conditions only.

The scanner scale factors file contains 10320 bits which are divided into 32-, 16-, and 8-bit words giving a total of 675 scale factors as shown in the Data Granularity Section. These quantities are used to scale the integer data quantities in each of the scanner data files (files 4 through 10 in Table 11) as follows:

ith Real Scanner Quantity = (ith Integer Scaled Quantity Rrom File) / (ith Scale Factor)

The third file contains the integer nonscanner scale factors written in the same order as their corresponding nonscanner data items as shown in Table 17.

Table 17: Packing Scheme for the Nonscanner Output Record

Item Name	Data Type	No. of Values per Region	Starting Item No.	Ending Item No.	Scale Factor	Cumulative No. Bits	Minimum Value	Maximum Value
32-bit words:
FILE ID	-	1	1	1	1	32	8	23
TSOLRD	MONTHLY (DAY)	1	2	2	100	64	0	500000
	MONTHLY (DAY)	1	3	3	100	96	-200	200
	MONTHLY (HOUR)	1	4	4	100	128	-200	200
TSOLRH	MONTHLY (HOUR)	1	5	5	100	160	0	500000
SOLARD	DAILY	31	6	36	100	1152	0	500000
SOLARH	HOURLY	24	37	60	100	1920	0	500000
Total number of bits = 32 x 60 = 1920
16-bit words:
Region #	-	1	61	61	1	1936	1	2592
Date: YYMM	-	1	62	62	1	1952	8411	to present
Spacecraft ID	-	1	63	63	1	1968	1	7
	MONTHLY (DAY)	1	64	64	10	1984	0	400
	MONTHLY (DAY)	1	65	65	10	2000	0	800
ALBEDO	MONTHLY (DAY)	1	66	66	1000	2016	0	1
	MONTHLY (HOUR)	1	67	67	10	2032	0	400
	MONTHLY (HOUR)	1	68	68	10	2048	0	800
ALBEDO	MONTHLY (HOUR)	1	69	69	1000	2064	0	1
	DAILY	31	70	100	10	2560	0	400
	DAILY	31	101	131	10	3056	0	800
ALBEDO	DAILY	31	132	162	1000	3552	0	1
	HOURLY	24	163	186	10	3936	0	400
	HOURLY	24	187	210	10	4320	0	1000
ALBEDO	HOURLY	24	211	234	1000	4704	0	1
GEOTYPE	-	1	235	235	1000	4720	0	1
SPARES	-	5	236	240	1	4800	127	127
Total number of bits = 16 x (240 - 60) = 2880
8-bit words:
NDLW	DAILY	31	241	271	1	5048	0	24
NDSW	DAILY	31	272	302	1	5296	0	24
NHLW	HOURLY	24	303	326	1	5488	0	31
NHSW	HOURLY	24	327	350	1	5680	0	31
DEAD-SCANNER FLAGS	-	3	351	353	1	5704	0	32767
SPARES	-	7	354	360	1	5760	127	127
Total number of bits = 8 x (360 - 240) = 960

The nonscanner scale factors file contains 5760 bits which are divided into 32-, 16-, and 8-bit words giving a total of 360 scale factors as shown in the Data Granularity Section. These quantities are used to scale the corresponding integer data quantities in each of the nonscanner data files (files 11 through 26 in Table 17) as follows:

ith Real Nonscanner Quantity = (ith Integer Scaled Quantity From File) / (ith Scale Factor)

An output summary report is created when the S-4 product is produced. Because the output is created sequentially, it is impossible to store the header information such as the number of records in each file. If one of the five input files is missing, the corresponding output files will each contain a record of fill values (also known as default values) according to the byte size of the data within that one record, or they will contain a record of ones. Table 18 lists the fill values.

To illustrate the file structure when one or more input files are missing, a printed summary for the output tape is shown in Figure 3 and Figure 4; "file empty" indicates a record of ones or fill values. When reading the S-4 output product, the processing date must be checked to see if the run date was before or after February 1992. Data processed before February 1992 will have these records filled with ones. Data processed after this date will have one record of fill values.

Also if some regions are missing from the input files, the output files may contain less than the maximum possible number of records.

9. Data Manipulations:

Formulae:

Derivation Techniques and Algorithms:

There are a number of steps in the processing of the ERBE data (see Flowchart Figure). The mathematics involved in each of these steps is beyond the scope of this document. However, interested readers are referred to the following: NASA Reference Publication 1184, Angular radiation models for Earth-atmosphere system, Volume 1: Shortwave radiation, and Volume 2: Longwave radiation; NASA Technical Paper 2670, Calculation and accuracy of ERBE scanner measurement locations; and Smith (Reference 15).

Data Processing Sequence:

Processing Steps:

The input to the standard S-4 product is provided by the Monthly Time/Space Averaging Subsystem, which accumulates data for each region into a 32x25 matrix (see Data Matrix Figure). There is a 31x24 submatrix within the larger matrix with each row representing one day of the month and each column representing a local hour for the region. The 25th column contains the daily averages for the month, referred to as the daily averages. The 32nd row contains the hourly averages for the month, referred to as the monthly hourly averages. The lower right-hand box contains the average of the daily averages, referred to as the monthly (day) average, and the average of the monthly hourly averages, referred to as the monthly (hour) average. The daily, monthly hourly, monthly (day), and monthly (hour) averages are calculated and stored in a regional average data base in the Monthly Time/Space Averaging Subsystem and passed to the S-4 Output Products Subsystem (Reference 5). These values are used to provide the nested, zonal, and global averages mentioned earlier.

Up to five separate types of data can be received from Monthly Time/ Space Averaging. One or more (up to four) of these types of data may be missing without upsetting the processing. The five different types of data are:

The Langley Research Center (LaRC) has the responsibility of processing and validating all science data from the ERBE mission and of distributing the resulting data products to the science community. The ERBE data processing system at LaRC uses a modular software subsystems approach to process the ERBE data, starting with the input telemetry and ephemeris data from Goddard Space Flight Center (GSFC) and NOAA and ending with the production of the required science data products.

The diagram in the Flowchart Figure shows the major steps in the science data processing, together with the primary input and output data products. The first step in this processing procedure is to ingest 24 hours of telemetry data from the ERBS, NOAA-9, or NOAA-10 spacecraft into the front-end processing subsystem of the Data Processing System. The processing at this step accounts for spacecraft differences and for differences in the data acquisition and handling systems of the ERBS and TIROS N satellites. The data are organized into a format that is common to data from GSFC and NOAA. Extensive data quality editing and evaluation are performed, including the checking of quality flags appended by the tracking networks and processing systems at GSFC and NOAA. The operational status of the instruments is determined, and all instrument housekeeping data and selected spacecraft housekeeping measurements are converted to engineering units and edited. Point vectors for the optical axes of the detectors are calculated in a local horizon coordinate system at the spacecraft.

The 8-day ephemeris data sets are processed and validated separately before combining them with the corresponding telemetry data. Orbital data are tested for consistency with data from the previous week, and tests are performed to verify the consistency of the orbit calculations between 1-minute data points. The tests include checks for in-plane and out-of-plane consistency and precision. The routine verification processing and other analyses performed to verify the accuracy of the ephemeris data have generally demonstrated accurate orbit determination for both the ERBS and NOAA spacecraft.

The next major processing stage begins with the merging of the output data from telemetry processing with data output from the ephemeris processing. The FOV locations on a surface at the TOAare determined for every radiometric measurement. The FOV locations are more critical for the scanner measurements than those of the nonscanner because of the small FOV of the scanner instrument. A FOV accuracy analysis has shown that the calculated locations of the scanner measurements are well within the FOV footprint of the instrument on the Earth.

At this processing stage, the raw measurements for each radiometric detector are also converted to incident radiances at the spacecraft. The conversion algorithms employ calibration coefficients that are based primarily on ground-based calibration data, but which are updated with results from in-flight calibrations.

In the inversion processing stage, data from the scanner detectors are used to identify the type of scene or source at the TOA that produced the raw radiometric measurements. Based on the scene type and geographical location, the scanner measurements are adjusted to account for changes in the spectral response in each detector. Using the selected scene-type, one of several angular directional models is selected for inverting or reducing the measurements from satellite altitude to radiant fluxes at the TOA. The nonscanner measurements are inverted using scene information determined during scanner data processing and two different inversion algorithms. One algorithm employs geometric shape factors and the other employs numerical filtering.

The time-ordered estimates of TOA fluxes are sorted into spatial sequences for both the scanner and nonscanner measurements, grouping all estimates for a month together on a regional basis. A full calendar month of estimates is then retrieved for each region of the Earth. Hourly, daily, and monthly estimates of several different parameters are derived by interpolation using directional models that describe the temporal variation of the radiation budget components. Archival products of monthly averages of radiation components for both the scanner and nonscanner are produced at this point.

Several archival products are produced at the final stage of data processing. The nested averages product gives values of the scanner and nonscanner fluxes from each instrument averaged over various spatial scales. The processing at this stage also combines data from all available spacecraft to produce a combined-satellite product of TOA fluxes averaged over the same spatial scales. An archival product for solar monitor measurements is also produced to provide time histories of solar calibration data. Finally, a scene validation product is produced that combines ERBE data with measurements from the AVHRR and the HIRS instruments. Data from these two NOAA instruments are used to validate the scene identification algorithm. Currently all archival data products are distributed first to the ERBE Science Team for review and validation and then to LaRC ASDC for archival.

Processing Changes:

Processing changes are described in the Special Corrections/Adjustments Section below.

Calculations:

Special Corrections/Adjustments:

Several modifications have been made to Monthly Time/Space Averaging Algorithms which affected the S-4 product in the following areas:

Calculated Variables:

Before the data are packed and written to the output file, albedos are calculated for monthly (day), monthly (hour), daily, and monthly hourly average quantities. The albedos are calculated on a regional, zonal, and global basis using the following equations:

For monthly (day):

For monthly (hour):

For daily values (for each day):

For monthly hourly (for each hour of a given month):

Graphs and Plots:

10. Errors:

Sources of Error:

A discussion of various factors that may lead to errors are discussed in the Confidence Level/Accuracy Judgement Section of this document.

Quality Assessment:

Data Validation by Source:

The measurement of radiation budget requires a massive data processing system. ERBE's system uses about 250,000 lines of FORTRAN code. This system also uses an additional 150,000 lines of off-line diagnostic work. The stringent requirements for accuracy in the budget dictate an acute attention to detail.

The ERBE data processing system uses about 25,000 coefficients. These coefficients are conveniently arranged in three groups. The first group is the set of "calibration coefficients" that appear in the algorithms converting telemetry counts to instrument irradiation. Ground- and in-flight-calibration sources provided these coefficients. The second group includes the angular distribution models (ADMs) and spectral unfiltering coefficients needed for inversion. A categorization of the Nimbus-7 ERB measurements forms the base for the ADM's. Missing bins were filled using the reciprocity principle. A combination of radiative transfer results and measurements of the instrument spectral responses provides the spectral correction coefficients. The third and final group of parameters consists of the coefficients needed for time averaging, mainly the directional models. These models describe the dependence of each scene type's albedo upon solar zenith angle. These directional models also came from the Nimbus-7 ERB, but have been suitably supplemented by Geostationary Operational Environmental Satellite (GOES) observations where needed. The majority of the coefficients are used in the inversion process.

The earth's radiation budget is not easy to measure, even indirectly. The ERBE Science Team has relied on consistency and measurement intercomparisons for validation. Fortunately, ERBE data provides a number of these checks. These criteria provide a way of judging the consistency of the various parameters in the data processing system.

Confidence Level/Accuracy Judgement:

The ERBE data products are complex assemblages of data and models. Thus, their uncertainties are difficult to compute. The following numbers represent estimates of the standard deviations about a given data point within which the true measurement might lie. They are not definitive confidence intervals, but are intuitively based on the observed discrepancies in the intercomparisons. It is also important to remember that different measurements have different uncertainties. First, for instantaneous radiances, we expect uncertainties of about 10f longwave observations of filtered radiance and 2 0.000000or shortwave. Radiative transfer comparison and spectral consistency provide the basis for this uncertainty estimate. Second, on an instantaneous observation of flux from 2.5 X 2.5 degree geographic regions, the ERBS/NOAA-9 intercomparisons offer reasonable estimates of uncertainty. These are 5 Wm^-2 in the longwave and 15 Wm^-2 in the shortwave. Third, on a monthly average, regional basis, the uncertainties in the scanner data are about 5 Wm^-2 for shortwave and 5 Wm^-2 for longwave. These come from simulations with GOES data. This uncertainty represents no change from the preflight estimate. The nonscanner averages may be somewhat more uncertain because of sampling and diurnal averaging process. Fourth, the uncertainty in global, annual average net radiation is probably about 5 Wm^-2. This estimate is based on the imbalance obtained using scanner data from the four validation months (April, July, and October 1985; January 1986).

Measurement Error for Parameters:

Measurement error is mentioned in the Confidence Level/Accuracy Judgement Section of this document.

Additional Quality Assessments:

None.

Data Verification by Data Center:

The data were received on 12 inch worm media. Before the data were archived, the ASDC checked all granules to ensure that the size of the granules matched that what was delivered on the media. The version number of the granules were also checked so that the most current version of the data are available to the user community. Granule level metadata were extracted from the granules such as the product ID, satellite(s) ID, and data date.

11. Notes:

Limitations of the Data:

There are no known limitations or unreliable aspects in the algorithms implemented to generate the ERBE science data.

Known Problems with the Data:

There are no known problems or inconsistencies in the ERBE data.

Usage Guidance:

The monthly hourly averaged results are a combination of measurements and models. The mean of these results represents the best estimate of the monthly hourly results. Also note that one should not just average the measurements alone to determine the monthly hourly means, because it will give a misleading diurnal cycle. The combination of measurements and models gives a more reasonable estimate when compared to full-time sampling of the GOES.

Any Other Relevant Information about the Study:

None.

12. Application of the Data Set:

Measurements of the radiation budget provide one of the important tools for the validation of numerical models of the atmosphere. They also provide possibilities for "climate experiments" by allowing the sensitivity of the radiation budget to various forcings to be studied empirically.

The use of cloud discrimination has provided a significant new source of information on the influence of clouds and the characteristics of clear-sky fluxes. This information is particularly important in understanding cloud forcing. It is also important in describing the response of clouds to climate change: the climate cloud sensitivity.

13. Future Modifications and Plans:

The ERBE project plans to complete the reprocessing, which is currently in progress, of the nonscanner data using inversion and time/space averaging processes which do not use scanner scene identification information.

Current plans are to reprocess the ERBE scanner data beginning in 1996 using the CERES algorithms.

To continue the measurements of the radiation budget, a second project,the Clouds and the Earth's Radiant Energy System (CERES), is currently being developed. CERES is a key component of the Earth Observing System (EOS). The CERES instruments are improved models of the Earth Radiation Budget Experiment (ERBE) scanner instruments. The strategy of flying instruments on Sun-synchronous, polar orbiting satellites, such as NOAA-9 and NOAA-10, simultaneously with instruments on satellites that have precessing orbits in lower inclinations, such as ERBS, was successfully developed in ERBE to reduce time sampling errors. CERES will continue that strategy by flying instruments on the polar orbiting EOS platforms simultaneously with an instrument on the Tropical Rainfall Measuring Mission (TRMM) spacecraft, which has an orbital inclination of 35 degrees. In addition, to reduce the uncertainty in data interpretation and to improve the consistency between the cloud parameters and the radiation fields, CERES will include cloud imager data and other atmospheric parameters. The first CERES instrument is scheduled to be launched on the TRMM spacecraft in 1997. Additional CERES instruments will fly on the EOS-AM platforms, the first of which is scheduled for launch in 1998, and on the EOS-PM platforms, the first of which is scheduled for launch in 2000.

14. Software:

Software Description:

Read software is available for this data set.

Software Access:

The read software can be obtained from the NASA Langley ASDC.

15. Data Access:

Contact Information:

Data Center Identification:

Procedures for Obtaining Data:

Data, programs for reading the data, and user's guides can be obtained through the EOSDISLangley ASDC on-line system which will allow users to search through the data inventory and place orders on-line. Users may also order data by letter, telephone, electronic mail (INTERNET), or personal visit.

The Langley ASDC User and Data Services staff provides technical and operational support for users ordering data.

Data Center Status/Plans:

On a regular basis, individual ERBE data granules are reviewed by local members of the ERBE Science Team. Upon Science Team approval, the ERBE Data Management Team releases the data granule to the LaRC ASDC for archive.

16. Output Products and Availability:

Browse images are available for the ERBE S-4 data set as well as an ERBE movie (video tape).

17. References:

18. Glossary of Terms:

19. List of Acronyms:

20. Document Information: