by Liam Gumley, Paul Hubanks, and Ed Masuoka
April 1994

Table of Contents

1.0 Introduction

2.0 MAS Instrument Summary

2.1 MAS Specifications

2.2 MAS Channel Configuration Summary

2.3 MAS Experiment Acronyms

2.4 Glossary

3.0 MAS Data Calibration

3.1 Visible/Near-Infrared Band Calibration

3.2 Temperature Correction of Visible/Near-Infrared Bands

3.3 Infrared Band Calibration

4.0 MAS Data Geolocation

5.0 Obtaining MAS LEVEL-1B Data

5.1 Distribution Contact Points

5.2 SDST Distribution Tape Format

6.0 How to decode MAS LEVEL-1B Data

6.1 Introduction to HDF

6.2 Obtaining HDF Libraries

6.3 Decoding and Extracting MAS Level-1B Data

7.0 Structure and Contents of MAS LEVEL-1B Data

7.1 Data Structure Summary

7.2 Data Structure Term Definitions

8.0 Obtaining Summary and Browse Information

9.0 MAS Contact List

10.0 References

1.0 Introduction

The MODIS Airborne Simulator (MAS) is a modified Daedalus Wildfire scanning spectrometer which flies on a NASA ER-2 and provides spectral information similar to that which will be provided by the Moderate Resolution Imaging Spectroradiometer (MODIS), scheduled to be launched on the EOS-AM platform in 1998 (King et al. 1992). The principal investigators for the MAS are Dr. Michael King (NASA/GSFC, Greenbelt MD), and Dr. Paul Menzel (NOAA/NESDIS, Madison, Wisconsin).

The Wildfire Spectrometer was delivered to NASA Ames Research Center in April 1991. A single visible channel was added and several spectral channels in the infrared port were altered to configure the instrument for the FIRE Cirrus-II experiment. In January 1992 the modified Wildfire was then further modified to become MAS. Beginning in June of 1992, the MAS has been flown in a series of experiments that lasted on average 2 to 4 weeks and were held approximately every 6 months.

The MAS spectrometer acquires high spatial resolution imagery in the wavelength range 0.55 to 14.3 microns. A total of 50 spectral bands are available in this range, and currently the digitizer is configured before each mission to record any 12 of these bands during flight. For all pre-1994 MAS missions the 12-channel digitizer was configured with four 10-bit channels and seven 8-bit channels. The MAS spectrometer is mated to a scanner sub-assembly which collects image data with an IFOV of 2.5 mrad, giving a ground resolution of 50 meters from 20000 meters altitude, and a cross track scan width of 85.92 degrees. More details on the MAS sensor and scanner characteristics are shown in Section 2. A 50-channel digitizer which will record all 50 spectral bands at 12 bit resolution is currently under development and is expected to be completed in mid 1994.

In support of the MODIS program, a MAS LEVEL-1B processing system was designed and implemented by the MODIS Science Data Support Team (SDST) at NASA Goddard Space Flight Center in 1991. The purpose of the processing system is to ingest MAS LEVEL-0 aircraft sensor, engineering and navigation data, and produce calibrated, geolocated radiances (LEVEL-1B) in a portable format (HDF). The system was designed by Liam Gumley, and is currently maintained by Paul Hubanks of Research and Data Systems, Corporation.

The purpose of this document is to describe the characteristics of MAS LEVEL-1B data, the calibration and geolocation methods used in processing, the structure and format of the LEVEL-1B data files, and methods for accessing the data.

Questions about MAS LEVEL-1B data should be directed to:

Paul A. Hubanks

MAS Level-1B Processing Manager

Research and Data Systems Corporation

phone: (301) 982-3724

email: hubanks@ltpmail.gsfc.nasa.gov

2.0 MAS Instrument Summary

The MAS spectrometer is mated to a scanning system of the type described by Jedlovec et al. (1989). Table 1 summarizes that characteristics of the MAS sensor, platform, and scanning system.

In November/December 1991, the modified Wildfire instrument was flown in the FIRE Cirrus-II experiment onboard a NASA ER-2 in coordination with other aircraft and satellites over the Coffeyville KS field site as well as the TX/LA Gulf coast. In January 1992, the modified Wildfire was converted to MAS configuration. In June 1992 the MAS was flown over portions of the Atlantic Ocean in the region of the Azores during the ASTEX experiment. During early 1993 the MAS was flown in the southwestern Pacific Ocean during the TOGA/COARE and CEPEX experiments. In July 1993 the MAS was flown over the northeastern United States during the SCAR-A experiment. During 1994, two missions are planned: a Navy ship-tracks experiment (MAST) and a fire experiment over California (SCAR-C).

Although the MAS instrument is a 50 band spectrometer, the data system used until mid 1994 has had the capability to record only 12 channels (at 8 bit resolution). The tables that follow show the channel configuration of the MAS data system for missions flown through July 1993. The recording of particular bands in specified channels is arbitrary and is determined before each experiment by the principal investigators (scientists) for each mission.

By January 1995, modifications that allow the Modis Airborne Simulator to record information in 50 visible and infrared spectral band regions at 12-bit resolution had been completed. Offered below is the spectral band configuration for Alaska-April95, ARMCAS, and SCAR-B (missions flown in 1995). Later missions had unique spectral band configurations which are offered on their respective Field Experiment Data pages.

Band  Left  Peak  Right
      50%   100%   50%

 01  0.523  0.546  0.569 


 02  0.627  0.653  0.680 


 03  0.679  0.700  0.721 


 04  0.720  0.741  0.762 


 05  0.761  0.781  0.802 


 06  0.802  0.823  0.844 


 07  0.844  0.865  0.886 


 08  0.888  0.906  0.922 


 09  0.921  0.945  0.967 


 10  1.595  1.622  1.650 


 11  1.652  1.678  1.705 


 12  1.706  1.732  1.757 


 13  1.759  1.786  1.812 


 14  1.813  1.838  1.863 


 15  1.863  1.889  1.914 


 16  1.914  1.939  1.963 


 17  1.964  1.989  2.013 


 18  2.014  2.040  2.064 


 19  2.066  2.090  2.113 


 20  2.114  2.138  2.162 


 21  2.163  2.187  2.210 


 22  2.211  2.237  2.261 


 23  2.262  2.287  2.310 


 24  2.312  2.337  2.360 


 25  2.361  2.386  2.410 


 26  2.875  2.959  3.039 


 27  3.031  3.110  3.195 


 28  3.202  3.276  3.358 


 29  3.338  3.422  3.507 


 30  3.511  3.588  3.668 


 31  3.669  3.743  3.819 


 32  3.819  3.901  3.986 


 33  3.976  4.054  4.132 


 34  4.126  4.208  4.288 


 35  4.288  4.363  4.435 


 36  4.439  4.516  4.594 


 37  4.585  4.667  4.748 


 38  4.745  4.823  4.901 


 39  4.897  4.974  5.047 


 40  5.046  5.124  5.203 


 41  5.205  5.281  5.362 


 42  8.386  8.600  8.822 


 43  9.484  9.785 10.101 


 44  10.300 10.546 10.794 


 45  10.746 11.022 11.289 


 46  11.736 11.976 12.185 


 47  12.651 12.859 13.109 


 48  12.998 13.244 13.470 


 49  13.421 13.723 14.021 


 50  13.960 14.193 14.383 

Prior to 1995, the MAS data system was capable of recording only 12 spectral bands at 8-bit resolution. This required unique instrument configurations for each Field Experiment.

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01       (bit bucket for 10-bit data: Channels 7,8,9,10)
    02           N/A          0.680             0.010
    03           N/A          1.630             0.050
    04           N/A          1.930             0.050 
    05           N/A          2.080             0.050
    06           N/A          2.130             0.050
    07           N/A          3.750             0.150
    08           N/A          4.650             0.150
    09           N/A          4.500             0.150 
    10           N/A          8.800             0.400
    11           N/A         10.950             0.500
    12           N/A         11.950             0.500

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01     (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           02           0.664             0.055
    03           07           0.875             0.041
    04           09           0.945             0.043
    05           10           1.623             0.057
    06           20           2.142             0.047
    07           31           3.725             0.151
    08           49          13.952             0.517
    09           42           8.563             0.396
    10           45          11.002             0.448
    11           48          13.186             0.352
    12           46          12.032             0.447

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01     (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           02           0.664             0.055
    03           07           0.875             0.041
    04           10           1.623             0.057
    05           14           1.830             0.050
    06           20           2.142             0.047
    07           31           3.725             0.151
    08           49          13.952             0.517
    09           42           8.563             0.396 
    10           45          11.002             0.448 
    11           48          13.186             0.352
    12           46          12.032             0.447

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01     (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           02           0.664             0.055
    03           07           0.875             0.041
    04           10           1.623             0.057
    05           15           1.880             0.050
    06           20           2.142             0.047
    07           31           3.725             0.151
    08           49          13.952             0.517
    09           42           8.563             0.396
    10           45          11.002             0.448
    11           48          13.186             0.352
    12           46          12.032             0.447

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01       (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           01           0.547             0.043
    03           02           0.664             0.055
    04           07           0.875             0.041  
    05           09           0.945             0.043 
    06           15           1.880             0.050
    07           20           2.142             0.047
    08           31           3.725             0.151
    09           42           8.563             0.396
    10           45          11.002             0.448
    11           48          13.186             0.352
    12           46          12.032             0.447

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01     (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           01           0.547             0.043
    03           02           0.664             0.055 
    04           05           0.786             0.040
    05           06           0.834             0.042 
    06           07           0.875             0.041
    07           09           0.945             0.043 
    08           10           1.623             0.057
    09           15           1.830             0.050
    10           31           3.725             0.151 
    11           42           8.563             0.396 
    12           45          11.002             0.448 

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01       (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           02           0.664             0.055  
    03           07           0.875             0.041 
    04           10           1.623             0.057
    05           15           1.880             0.050 
    06           20           2.142             0.047
    07           23           2.280             0.050
    08           32           3.900             0.150 
    09           31           3.725             0.151
    10           42           8.563             0.396
    11           45          11.002             0.448
    12           46          12.032             0.447

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01           01           0.547             0.043
    02           02           0.664             0.055
    03           04           0.745             0.040 
    04           05           0.786             0.040
    05           06           0.834             0.042
    06           07           0.875             0.041
    07           09           0.945             0.043
    08           10           1.623             0.057
    09           20           2.142             0.047  
    10           32           3.900             0.150  
    11           45          11.002             0.448  
    12           46          12.032             0.447 

DATA CHANNEL   MAS BAND   CentralWavelength   50%Bandwidth     
    01       (bit bucket for 10-bit data: Channels 9,10,11,12)
    02           01           0.547             0.043  
    03           10           1.623             0.057        
    04           10           1.623             0.057     
    05           15           1.880             0.050   
    06           20           2.142             0.047 
    07           46          12.032             0.447   
    08           32           3.900             0.150  
    09           02           0.664             0.055         
    10           07           0.875             0.041          
    11           45          11.002             0.448         
    12           32           3.900             0.150       

FIRE = First ISCCP Radiation Experiment (10/91-12/91)
ASTEX = Atlantic Stratocumulus Transition Experiment (5/92-6/92)
TOGA/COARE = Tropical Ocean Global Atmosphere/Coupled Ocean-Atmosphere Response Experiment (1/93-3/93)
CEPEX = Central Equatorial Pacific Experiment (3/93-4/93)
SCAR-A = Sulfates, Clouds, and Radiation/America (7/93)

Ames = NASA Ames Research Center
Band = One detector/filter assembly
Blackbody = Radiation source, emittance=absorptance=1, reflectance=0
Browse = Uncalibrated reduced resolution image product
Calibration = Conversion from sensor digital counts to physical units
Channel = Area in LEVEL-0 format reserved for data from one band
Data system = MAS digital flight data digitizer/recorder onboard ER-2
ER-2 = NASA high altitude research version of U-2 spy plane
Exabyte = 8 mm tape drive for digital data storage
Exabyte format = Storage format for ER-2 Exabyte data system
Flight track (line) = Portion of aircraft track where heading is constant
FTP = File Transfer Protocol (Internet)
Geolocation = Computation of earth coordinates of a sensor measurement
GIF = Graphics Interchange Format
HDF format = Hierarchical Data Format
IFOV = Instantaneous Field Of View
INS = Inertial Navigation System onboard ER-2
Intermediate Format = Data format created by unpacking LEVEL-0 data
IR (Infrared) = Wavelengths of 3.7 microns and longer
ISCCP = International Satellite Cloud Climatology Project
LEVEL-0 = Unformatted, raw, packed, unprocessed sensor data
LEVEL-1A = Unpacked and reformatted sensor data
LEVEL-1B = Calibrated and geolocated sensor data
MAS = MODIS Airborne Simulator
Metadata = Descriptive information pertaining to a dataset
MODIS = Moderate Resolution Imaging Spectroradiometer
Navigation Data = Record of the position and orientation of a sensor platform
NIR (Near-Infrared) = Wavelengths from 0.7 to 3.7 microns
PCM = Tape drive used for old ER-2 data system
Scanner = MAS scan mirror and blackbody target assembly
SGI = Silicon Graphics Workstation
Spectrometer = MAS optics and detector assembly
VIS (Visible) = Wavelengths from 0.4 to 0.7 microns

3.0 MAS Data Calibration

Calibration coefficients for the MAS visible (~ < 0.7 um) and near-infrared (0.7 um < ~ < 3.7 um) bands are derived from integrating spheres used pre and post launch. The transformation from digital count to radiance is defined by

where,

Radiance = computed radiance
Count = recorded count
Gain = channel gain
CBBavg = running average of 30 previous cold black body counts
Slope = slope coefficient, determined for each band

The radiance units are Watts per square meter per steradian per micron (W m^-2 sr^-1 um^-1). Integrating sphere calibrations may be carried out several times during a MAS flight experiment. There may be several sets of calibration slopes, each applied to data from a discrete time period, during a single experiment. For more details, see Arnold et al. (1994a, 1994b).

MAS testing in a cold chamber at NASA Ames Research Center has shown that some MAS VIS/NIR bands are subject to sensitivity changes as a function of temperature. This is important, since at cruise altitude (20000 meters) the MAS instrument environmental temperature is typically around -35deg.C. The cold chamber data is used to develop equations which predict the sensitivity of the VIS/NIR channels at given temperatures. Temperature sensors mounted on the MAS in-flight are then used as a reference for computing the sensitivity correction. For example in ASTEX data it was found that only channels 5 and 6 were sensitive to this phenomenon. For these two channels Count=Ceff, for all other VIS/NIR channels Count=Crec.

The form of the equation for Ceff during ASTEX was:

where,

Ceff = temperature corrected effective digital count
Gain = channel gain
Ccbb = recorded in-flight digital count for cold black body
Crec = recorded original in-flight digital count
a1, b1 = coefficients determined for each band
Tmas = temperature of cold black body (degrees Celsius)

The coefficients a1 and b1 are determined for each temperature sensitive channel in each MAS experiment. This correction is applied before the linear transformation to radiance units. For more details on the corrections used in various field experiments, see Arnold et al. (1994a, 1994b).

Calibration data for the MAS infrared (IR) bands (~ > 3.7 um) are obtained during flight from two blackbody sources which are viewed during every mirror scan. The first (cool) blackbody is usually left to float at ambient temperature, although it can be temperature controlled. The second (warm) blackbody is maintained at a higher temperature than the cool blackbody, and the temperature level is set depending on the desired scene temperature range. Thermistors measure the temperature of the blackbodies during flight. The equivalent Planck radiances for the blackbodies (assuming unit emissivity) are computed using a sensor weighted integral of the Planck function over the wavelength range of the spectral band :

R(L,T) = sensor weighted equivalent Planck radiance
B(L,T) = Planck radiance at wavelength and blackbody temperature T
S(L) = sensor spectral response

The Planck radiance in Watts per square meter per steradian per micron (W m^-2 sr^-1 um^-1) is computed by the equation

where

C1 = 2 * h * c² = 1.1910439 * 10^-16 W m^-2
C2 = ( h * c ) / k = 1.4387686 * 10^-2 mK
L = wavelength in meters
T = temperature in degrees Kelvin

Once the equivalent Planck radiances for the blackbodies are known, the calibration slope and intercept are computed by the relationships

where,

R(L,T)BB1 and R(L,T)BB2 are the equivalent Planck radiances for the cool and warm blackbodies respectively
CountBB1 and CountBB2 are the digital radiance counts for the cool and warm blackbodies respectively.

The calibration from digital counts to radiance is then a linear transformation of the form

The radiance units are Watts per square meter per steradian per micron (W m^-2 sr^-1 um^-1). This procedure is performed for every IR channel on every scanline. No averaging of MAS blackbody data is done.

Conversion from IR radiance to Planck equivalent temperature or 'brightness temperature' may be done by inverting the Planck equation. The inverse equation is of the form

where,

T(L,B) = brightness temperature in degrees Kelvin,
C2 = ( h . c ) / k = 1.4387686 . 10^-2 m K
C1 = 2 . h . c² = 1.1910439 . 10^-16 W m^-2
l = wavelength in meters
B(L,T) = Planck radiance in W m^-2 sr^-1 um^-1

4.0 MAS Data Geolocation

Navigation data onboard the ER-2 is recorded by an Inertial Navigation System (INS). This system has a dedicated tape recorder, which is separate from the MAS tape recorder. Before February 1992, INS data was obtained from the INS recorder itself. This data was unpacked and formatted by staff at Ames Research Center, and distributed as a separate file along with the MAS data. Since February 1992, the new MAS Exabyte data recorder has had the capability to record data from the INS system at the same time as it records MAS data, thus giving one integrated dataset which contains both MAS and navigation information. In the old INS recording system, aircraft position and attitude were updated every 5 seconds. In the new recording system, the update interval is 1 second.

The geolocation algorithm is only applied to those portions of a flight where the aircraft flew straight, level, and on a constant heading. Data recorded during ascent, turns, and descent is not processed to LEVEL-1B. Ascent data is not processed since the scanner is not looking at nadir, and thus the imagery is distorted. During turns, the scanner roll correction system cannot compensate for high roll angles, and the imagery is distorted so that black bands appear on either side of the flight track. During descent, the scan mirror optics become covered with condensation, and the image information is lost.

The straight line flight track start and end times are identified by an automated program which analyzes altitude, heading, pitch, and roll. Once the start and end times for all flight tracks are determined, linear regressions for aircraft latitude, longitude, heading and altitude versus time are computed. This is the simplest way to represent the INS data set when it is referenced to the MAS data set, which is sampled at a higher rate (6.25 scanlines/second). To geolocate a given MAS flight track, the MAS start time and scanline number at the beginning of the flight track are determined. These are used as a reference for the rest of the flight track, since the time code recorded by the MAS is stored as integer seconds. The scanline number and scan rate are used to determine the time at subsequent scanlines in the flight track by

where,

t0 = time of the first scan line in a flight track
S = scan line number of a data line in the flight track
S0 = scan line number of the first data line in the flight track
r = scan rate (6.25 scans/second)

Once time for a given scanline is computed, the linear regression relationships are used to compute aircraft latitude, longitude, heading and altitude at that time. The scan angle is determined from the pixel number by the relationship

where,

DD = scan angle (displacement from aircraft nadir point, degrees)
K = angular distance between IFOVs ( 85.92 degrees / 715 steps )
E = pixel number
En = pixel number at nadir (358.5)

where,

Z = aircraft altitude

Once the distance and azimuth from nadir to the pixel are known (azimuth is computed from the heading), trigonometric transformations are used to compute the latitude and longitude for every 10th pixel on a scanline (pixels 1, 10, 20, 30, ...., 690, 700, 710, 716). The reason for geolocating every 10th pixel on a scan line is to save space in the output data set. Solar zenith and azimuth angles and sensor zenith and azimuth angles are also computed for every 10th pixel. More details on the MAS geolocation algorithm are given in Jedlovec et al. (1989).

5.0 Obtaining MAS LEVEL-1B Data

MAS LEVEL-1B data are currently distributed by the MODIS Science Data Support Team (SDST). However it is expected that by the end of 1994 the NASA Langley and Goddard DAAC's (Distributed Active Archive Center) will be the distribution points for MAS data. At that time users will need to contact the appropriate DAAC for specific media and format information. MAS LEVEL-1B data from the FIRE II Cirrus and ASTEX missions will be distributed by the Langley DAAC. MAS LEVEL-1B TOGA/COARE data will be distributed by the Goddard DAAC. Distribution points for other experiments have yet to be determined.

MODIS Science Data Support Team
Paul A. Hubanks
Research and Data Systems Corporation
7855 Walker Drive, Suite 460
Greenbelt MD 20770
Phone: (301) 982-3724
Internet: hubanks@ltpsun.gsfc.nasa.gov

Langley DAAC
User and Data Services
Mail Stop 157-B
NASA Langley Research Center
Hampton VA 23681
Phone: (804) 864-8656
Internet: userserv@eosdis.larc.nasa.gov

Goddard DAAC
DAAC User Support Office
Global Change Data Center
Code 902.2
NASA Goddard Space Flight Center
Greenbelt MD 20771
Phone: (301) 286-3209
Internet: daacuso@daac.gsfc.nasa.gov

MAS LEVEL-1B data distributed by the MODIS SDST are stored on Exabyte 8500 8mm tape. The tapes are formatted as simply as possible, so they may be read on any other system without the problems imposed by system dependent header files, block lengths, and formatting. Each tape contains all the flight lines for one MAS flight (one day). The number of flight lines per flight varies, but is generally between 10 and 20. The volume of data varies, but is generally 1 to 3 gigabytes per flight.

The archive tapes are created by writing each output data file (1 straight-line flight track) to tape in fixed-length blocks of 16384 bytes, in time ascending order. One end-of-file (EOF) mark is written at the end of the data blocks for each file, and an extra EOF is written at the end of the data on the tape. The last block of each file has good data at the start of the block and unused bytes (filled with null characters) at the end. Information on the length of the file is encoded in the header when the file is created. No file name, protection, or ownership information is written onto the archive tape. All information necessary to identify the file is stored in the file itself.

To read the data files from tape, system utilities should be used. For example on a UNIX system, the following command could be used to read the first file on the tape:

Note that "/dev/mt/tps0d3nrnsv.8500" is the tape device driver. This is system dependent. However, the user should ensure that the device driver is for an Exabyte 8500 (not 8200) drive, and handles fixed length blocks which are 16384 bytes long. The output file name is a sample only: the user could specify any output file name.

To read the first straight-line flight track file on a DEC VAX/VMS system, the following commands could be used (note that any VMS utility which reads tapes with fixed-length blocks is suitable):

6.0 How to decode MAS Level-1B Data

MAS LEVEL-1B data are stored in Hierarchical Data Format (HDF). The data are organized into individual straight-line flight tracks. Each MAS LEVEL-1B HDF file contains one straight-line flight track, and defines a granule of data. In general there are 10 to 20 flights (or flight days) in a given experiment and 10 to 20 flight tracks in a given flight.

HDF is developed and maintained by the National Center for Supercomputing Applications (NCSA), and has been selected as the standard data format for all Earth Observing System (EOS) data products. HDF is a multi-object file format for the transfer of graphical and numerical data between machines, and allows the user to create, access, and share scientific data in a form that is self-describing and network-transparent. "Self-describing" means that a file includes information defining the data it contains. "Network-transparent" means that a file is represented in a form that can be accessed by computers with different ways of storing integers, characters, and floating-point numbers. HDF files also allow direct access, so that a small subset of a large dataset may be accessed efficiently, without first reading through all the preceding data. It should be noted that HDF files should only be accessed through the HDF library of subroutine and function calls (from FORTRAN or C). HDF is not a data format which can be unpacked easily by knowing word locations, byte ordering and so on. With HDF, the user is insulated from these details, so that differences in system specific storage details are transparent.

MAS LEVEL-1B data are stored in a subset of HDF known as HDF/netCDF. Initial versions of the MAS LEVEL-1B data were written in netCDF (Network Common Data Form) format, developed by the University Corporation for Atmospheric Research (UCAR). In HDF Version 3.3, the netCDF interface was integrated with the HDF software library. All MAS LEVEL-1B data is now written in HDF/netCDF to comply with the EOS data format standard.

From the user's perspective, this means that while the data is stored in HDF/netCDF, the data may be extracted using netCDF-like software calls, from either FORTRAN or C. This is the preferred means of extracting data from MAS LEVEL-1B data files. While it is possible to use HDF software calls alone, this method is not described in this document.

Data Format Summary: MAS data is stored in HDF/netCDF, and may be extracted using netCDF-like calls to the HDF library.

In order to extract data from MAS LEVEL-1B data files, the HDF software library must be compiled and installed. The source code is available from NCSA through their HDF Information Server or by anonymous FTP on Internet.

The following commands may be used on a Unix system to retrieve the source code by ftp:

The instructions provided in the README files should then be followed to compile the source code and construct the library. If problems are encountered, please contact HDF user support by electronic mail at hdfhelp@ncsa.uiuc.edu.

To determine on which Platforms the latest release of the HDF software has been tested by NCSA go to the HDF Information Server.

To make it easier to get started using the HDF library, a compiled version for a Silicon Graphics (SGI) platform is available via anonymous FTP at ltpiris2.gsfc.nasa.gov. In the future, compiled libraries for several different platforms will be available. The user should log in with username anonymous and password guest. The files are stored in the directory

pub/MAS/hdf_netcdf/SGI_IRIX4.05

The following files are provided
(A = ASCII mode, B = binary mode in FTP)

libdf.a = compiled HDF software object library (B)
libnetcdf.a = compiled HDF/netCDF software object library (B)
ncdump = executable utility program to dump HDF files as formatted text (B)
netcdf.h = C header file needed to call the HDF/netCDF library (A)
netcdf.inc = FORTRAN include file needed to call the HDF/netCDF library (A)
simple = executable utility program to read MAS LEVEL-1B HDF files (B)
simple.f = FORTRAN program to read MAS LEVEL-1B HDF data (A)
simple.sh = script to compile the FORTRAN program simple.f (A)

These files provide the tools to begin extracting data from the MAS LEVEL-1B HDF data files using the FORTRAN or C netCDF interface. For further information on programming using the netCDF FORTRAN or C interface consult the NetCDF User's Guide

Note that if you downloaded the HDF3.3r3 source code as previously described, a copy of the netCDF User's Guide is included in the distribution. In addition, FORTRAN programs are available which can unpack MAS radiance and meta data from a LEVEL-1B HDF file.

7.0 Structure and Contents of MAS LEVEL-1B Data

Each individual MAS LEVEL-1B file contains calibrated, geolocated radiances for all MAS channels for one straight line flight track. Each file also contains various ancillary data such as instrument configuration, and time and date codes. When the MAS LEVEL-1B data structure was designed, a major concern was that all LEVEL-0 engineering and ancillary input data be maintained in the output data set. For this purpose, all LEVEL-0 parameters (except the raw digital image counts) are retained in the LEVEL-1B structure. In this way, the health and performance of the sensor may be monitored, and the data may be reversed to LEVEL-0 if necessary.

The MAS LEVEL-1B data structure was designed to be easy to understand and to reflect the physical characteristics of the data. All variables, dimensions, attributes, and data are named descriptively to aid interpretation. The following listing was created from an actual MAS LEVEL-1B HDF file using NCDUMP (see Section 7.2). Due to size constraints only the dimensions, variables, and global attributes are listed here. The data portion is not shown.

This section defines the dimensions and gives a brief description of each dimension contained in the data set. Dimensions are named integers which specify the shape of one or more multi-dimensional variables in the HDF file.

Time = UNLIMITED ; // (5367 currently)
The 'length' dimension of the data set. This dimension corresponds to elapsed time, and also corresponds with the incrementing of the scan line counter.

NumberOfChannels = 12 ;
The number of data channels

NumberOfPixels = 716 ;
The number of pixels (IFOVs) across a MAS scan.

HeaderLength = 6160 ;
The length of the ASCII data set header. The header is composed of blocks of 80 characters, which contain no carriage return, line feed or similar record separators. The header contains a copy of the MAS instrument configuration file that was used in processing.

AnchorIndexSize = 73 ;
The number of geolocation anchor points per scan line. Each MAS scan line has geolocation data for every 10th pixel (to save space). The geolocation data is defined for pixel numbers 1, 10, 20, 30, 40,......680, 690, 700, 710 and 716 which makes a total of 73 geolocation anchor points. It should be noted that pixels 1 to 358 are on the starboard (right) side of the aircraft, while pixels 359 to 716 are on the port (left) side of the aircraft.

Variables:
This section defines the type and size of the variables, along with a brief description of each, in the output data set. A variable represents a multi-dimensional array of values of the same type. The type definitions are from the C language, however the equivalent FORTRAN types are

byte: BYTE 8 bit data
char: CHARACTER synonymous with byte
short: INTEGER*2 16 bit integers
long: INTEGER*4 32 bit integers
float: REAL*4 32 bit IEEE floating point
double: DOUBLE PRECISION 64 bit IEEE floating point

Note that the listing of variable names which follows show the dimensions in C order. In FORTRAN, the order of dimensions is reversed. For example, the variable declared as CalibratedData(Time, NumberOfChannels, NumberOfPixels) in C would be declared as CalibratedData(NumberOfPixels, NumberOfChannels, Time) in FORTRAN.

float Left50%ResponseWavelength(NumberOfChannels) ;
Left50%ResponseWavelength:units = "microns" ;
The wavelength corresponding to the left hand 50% response point of the spectral response function for each channel.

float Central100%ResponseWavelength(NumberOfChannels) ;
Central100%ResponseWavelength:units = "microns" ;
The wavelength corresponding to the central 100% response point of the spectral response function for each channel.

float Right50%ResponseWavelength(NumberOfChannels) ;
Right50%ResponseWavelength:units = "microns" ;
The wavelength corresponding to the right hand 50% response point of the spectral response function for each channel.

float SolarSpectralIrradiance(NumberOfChannels) ;
SolarSpectralIrradiance:units = "watts/meter2/micron" ;
SolarSpectralIrradiance:title = "Sensor Weighted and Orbit Corrected Solar Spectral Irradiance" ;
The sensor weighted and orbit corrected solar spectral irradiance at the top of the atmosphere for each channel. This is computed using the LOWTRAN7 database of solar spectral irradiance and the spectral response function for each channel by

where,

Sw(L) = sensor weighted orbit corrected exoatmospheric solar spectral irradiance
S(L) = orbit corrected exoatmospheric solar spectral irradiance
R(L) = sensor spectral response

short AnchorPointIndex(AnchorIndexSize) ;
The pixel numbers for which geolocation information is defined. The geolocation data is defined for pixel numbers 1, 10, 20, 30, 40,......680, 690, 700, 710 and 716 which makes a total of 73 geolocation anchor points.

char DataSetHeader(HeaderLength) ;
The data set header text. The header is composed of blocks of 80 characters, which contain no carriage return, line feed or similar record separators. The header contains a copy of the MAS instrument configuration file that was used in processing. This indicates, among other things, which spectral bands were enabled, which channels had 8 or 10 bit data, which channels were calibrated using the MAS blackbodies (IR channels), and slopes/intercepts for the visible/near-IR channels.

short DataFrameStatus(Time) ;
(LEVEL-0 MAS engineering data) Zero indicates good data. Bits are set to 1 to indicate errors as follows:

Bit Position    Meaning
    1       Bad digit in scanline count
    2       Bad digit in time code
    4       Bad digit in thumbwheel data
    8       Bad digit in reference temperature
   16       Incorrect channel
   64       Bad sync word
  128       No match for end-of-frame code (last word)
 sign       Set to 1 if any other bit set

: FIREClockOffset (INS-MAS) Hours = -99 .f ;
Offset between INS and MAS clocks for pre-Exabyte recording system LEVEL-0 data only. Before the Exabyte recording system was brought into use in February 1992 the MAS and INS data streams were recorded separately, each using it's own internal clock. During the FIRE mission, the INS clock was found to be leading the MAS clock by anywhere from 30 to 90 seconds. This offset had to be corrected to ensure accurate geolocation, so the INS clock time was adjusted in post-processing to match the MAS clock time. This variable records the time subtracted from the INS clock time. After February 1992 the Exabyte LEVEL-0 recording system encoded the MAS and INS in one synchronous data stream, so this problem was eliminated.

8.0 Obtaining Summary and Browse Information

MAS flight summary information and browse data are available via the Browse Imagery Archive within the MAS World Wide Web site. Available are summary and browse information for each MAS experiment. The flight summary file is an ASCII text summary of the flight lines processed for a particular flight. The browse imagery files for each flight line are stored in Graphics Interchange Format (GIF). These images contain every 4th pixel on every 4th line from one short-wave visible and one long-wave infrared channel. They have been contrast stretched and as such do not contain quantitative radiance information. There are 2 GIF image files per flight line, named 92170??v.gif and 92170??i.gif, where ?? is the flight line number, v denotes visible (VIS) imagery, and i denotes infrared (IR) imagery. The date corresponds to year and day of year as mentioned previously. The VIS and IR channels are selected such that the VIS channel is at around 0.68 microns, and the IR channel is at around 12 microns.

Example browse images from the MAS flight on 17 June 1992 are shown in Figure 1. These were acquired over the Atlantic Ocean, off the coast of the Azores Islands. Stratiform cloud features are visible, with marine stratocumulus in the upper region of the image and continental stratocumulus in the lower region of the image. A cold cirrus cloud formation can be seen (as a black feature) in the infrared browse image in the upper portion. This feature is not distinct in the visible browse image.

Visible (0.68 um) browse image Infrared (11.95 um) browse image

 

F igure 1. MAS Browse Images (17 June 1992 - Flight Line #11)

 

Public domain or shareware software packages to view GIF imagery on various systems are available from the following anonymous FTP sites:

IBM/PC
oak.oakland.edu, pub/msdos/gif/cshw101a.zip

Macintosh
mac.archive.umich.edu, mac/graphics/graphicsutil/gifconverter2.37.cpt.hqx

Unix/X
ftp.x.org, contrib/xloadimage.4.1.tar.gz

9.0 MAS Contact List

Below is a list of persons involved with the processing and analysis ofMAS data.

Paul Hubanks, MAS LEVEL-1B Processing Manager
Research and Data Systems Corporation
7855 Walker Drive, Suite 460
Greenbelt MD 20770, USA
(301) 982-3724, phubanks@ltpsun.gsfc.nasa.gov

Liam Gumley, MAS LEVEL-1 Processing System Designer
Code 913, NASA Goddard Space Flight Center
Greenbelt MD 20771, USA
(301) 286-8789, gumley@climate.gsfc.nasa.gov

Tom Arnold, MAS VIS/NIR Calibration
Code 913, NASA Goddard Space Flight Center
Greenbelt MD 20771, USA
(301) 286-4805, arnold@climate.gsfc.nasa.gov

Michael King, MAS Principal Investigator
Code 900, NASA Goddard Space Flight Center
Greenbelt MD 20771, USA
(301) 286-8228, king@climate.gsfc.nasa.gov

Paul Menzel, MAS Principal Investigator
NOAA/NESDIS
1225 W. Dayton St.
Madison WI 53706 USA
(608) 263-4930, paulm@ssecmail.ssec.wisc.edu

Chris Moeller, MAS Investigator; calibration, geolocation, MCIDAS
Space Science and Engineering Center, UW-Madison
1225 W. Dayton St.
Madison WI 53706, USA
(608) 263-9597, chrism@ssecmail.ssec.wisc.edu

Jeff Myers, MAS LEVEL-0 Data Distribution
High Altitude Missions Branch, NASA Ames Research Center
Mail Stop 240-6
Moffett Field CA 94035, USA
(415) 604-6252, jmyers@msmail.arc.nasa.gov

Pat Grant, MAS Engineering
High Altitude Missions Branch, NASA Ames Research Center
Mail Stop 240-6
Moffett Field CA 94035, USA
(415) 604-6393, pgrant@msmail.arc.nasa.gov

10.0 References

Arnold, G.T., M. Fitzgerald, P.S. Grant, and M.D. King, 1994a: MODIS Airborne Simulator Visible and Near-Infrared Calibration - 1991 FIRE-Cirrus Field Experiment. NASA Goddard Space Flight Center, NASA Technical Memorandum 104600.

Arnold, G.T., M. Fitzgerald, P.S. Grant, and M.D. King, 1994b: MODIS Airborne Simulator Visible and Near-Infrared Calibration - 1992 ASTEX Field Experiment. NASA Goddard Space Flight Center, NASA Technical Memorandum 104599.

Jedlovec, G.J., K.B. Batson, R.J. Atkinson, C.C. Moeller, W.P. Menzel, and M.W. James, 1989: Improved Capabilities of the Multispectral Atmospheric Mapping Sensor (MAMS). NASA Marshall Space Flight Center, NASA Technical Memorandum 100352.

King, M. D., Y. J. Kaufman, W. P. Menzel and D. Tanré, 1992: Remote sensing of cloud, aerosol, and water vapor Properties from the Moderate Resolution Imaging Spectrometer (MODIS). IEEE Trans. Geosci. Remote Sens., 30, 2-27.

Back to: [Documents and Research]