Title: INDOEX C-130 AIMR Data Contact: Craig Walther, Brian Lewis, or Rich Neitzel National Center For Atmospheric Research/ATD P.O. Box 3000 Boulder, Colorado 80303-3000 USA Phone: 303-497-2054 Email: craigw@ucar.edu Or Julie Haggerty University of Colorado Boulder, Colorado USA Data Set Description: REFURBISHMENT OF THE AIRBORNE IMAGING MICROWAVE RADIOMETER AT THE NATIONAL CENTER FOR ATMOSPHERIC RESEARCH ABSTRACT The Airborne Imaging Microwave Radiometer (AIMR) was purchased by the Atmospheric Environment Service in Canada and was later loaned to the National Center for Atmospheric Research (NCAR). This is a scanning four-channel radiometer operating at 37 and 90 GHz. It was originally used to map the age of sea ice in the Arctic Ocean. Its base products are brightness temperature maps of the earth's surface at two different polarizations at the two frequencies. In 1997 NCAR decided to operate the AIMR as a supported instrument. An engineering study of the instrument determined that the computers and software used both in real time and in analyzing the data were fairly outdated and would be hard to support and maintain. Therefore, NCAR decided to replace the real time computer system and rewrite all of the software. In order to fully understand the instrument, the radio frequency (RF) portions were also totally researched and characterized. The newly refurbished instrument has now completed three successful research campaigns. Plans to make AIMR more reliable and more versatile have been made for the future 1.0 HISTORY The AIMR (for Airborne Imaging Microwave Radiometer) is a four channel scanning radiometer that was jointly developed by the Atmospheric Environment Service (AES) of Canada and MPB Technologies of Dorval, Quebec Canada. The radiometer was originally used by the AES Ice Branch to detect frozen or open water and to estimate the age of the ice patches in the Arctic Ocean. It was flown in a large fiberglass housing beneath a Lockheed Electra aircraft. The AIMR was first used in February of 1989. When the Electra was no longer available to AES, they were not able to use AIMR, so it was loaned to the National Center for Atmospheric Research (NCAR) in Boulder Colorado. NCAR developed a mount for AIMR inside their Lockheed C-130 "Hercules" aircraft and used the radiometer on several projects. In 1998 NCAR decided to include the AIMR as one of its supported instruments. In order to fully support the instrument, NCAR replaced much of the original computer equipment and recharacterized the receiver. 2.0 OVERVIEW OF THE SYSTEM AIMR has four radiometric channels, two orthogonally polarized channels at 37 GHz and two at 90 GHz. The channels are polarized in a manner such that the amount of vertically polarized radiation and the amount of horizontally polarized radiation can be uniquely calculated at most look angles of the instrument. These channels are detected and amplified by receivers whose outputs are digitized and written on tape by a set of computer systems. The base products from the radiometer are the brightness temperatures, of each of the four channels, at each sample location. 2.1 THE ANTENNA SYSTEM The AIMR scanning antenna consists of a 25 cm diameter (minor axis) elliptical mirror constructed from black anodized aluminum. The antenna is mounted at a 45 degree angle from its spin axis which is along the longitudinal axis of the aircraft. The bottom of the aircraft has a large "smile" cut in it for the AIMR to look out. This smile allows the AIMR to see approximately +/- 60 degrees from nadir. As the antenna spins, a swath of the earth is observed below the aircraft. The width and length of the swath depends on the aircraft's altitude. If the antenna is spun quickly enough, each successive swath overlaps the previous one slightly. This provides a complete brightness temperature map of a ribbon of the earth's surface along the flight track of the aircraft. The 25 cm diameter antenna leads to a beamwidth of about 1 degree in the 90 GHz channels and 2.5 degrees in the 37 GHz channels. The maximum rotational velocity of the antenna is 4.9 revolutions per second. At this speed, on the C-130, full overlap of the 90 GHz channel is obtained at altitudes of 1400 meters or greater and the 37 GHz channels overlap above 550 meters. The width of the observed ribbon of data on the earth's surface is 2400 meters at an aircraft altitude of 1400 meters and the 90 GHz beam's footprint is about 25 meters in diameter. The antenna reflects the energy incident on it into a Gaussian Optical Lens Assembly (GOLA). This assembly contains one diachronic reflector that separates the 37 GHz energy from the 90 GHz energy. Two polarization separators are then used to orthogonally polarize each of the channels such that they receive radiation polarized approximately +45 degrees and -45 degrees from the vertical. The polarized energy is then received with four RF horns and fed to the mixers. The lens itself is constructed to reflect all energy below 30 GHz. 2.2 THE RECEIVERS The RF energy is mixed to an intermediate frequency (IF) by mixers mounted directly on the feed horns. The local oscillator for these mixers is provided by four Gunn diodes that are also mounted on the end of the feed horns. The IF signal is then bandpass filtered to 0.5 to 2.0 GHz in the 37 GHz channels and 2.0 to 4.0 GHz in the 90 GHz channels. A tunnel diode square law video detector then detects the filtered signals. Each detected signal is then amplified by a temperature compensated video amplifier and fed to a software programmable analog signal conditioning board. The signal conditioning boards allow the computer to select an offset, gain (attenuation) and filter cutoff frequency for each channel. The filters for both 37 GHz channels and both 90 GHz channels are not individually programmable and they must be set the same. The gain and offset is variable for each of the four channels. The computer algorithm that controls the analog signal conditioning boards uses the gain and offset to keep each channel within the range of the 12 bit (0-10 volt) analog to digital (A/D) converter. The programmability of the filters is used to match the filter time constant with the time between each sample. In this way the maximum sensitivity is obtained while maintaining low correlation between samples. 2.3 THE CALIBRATION LOADS Two calibration loads are used to calibrate the data (and control the gain and offset of the signal conditioning boards) with each rotation of the antenna. These loads are constructed from the RF energy absorbing epoxy (in a traditional array of small pyramids) and are covered with a low loss foam to keep their temperature more stable. One of these loads, called the "hot load", is kept at a constant temperature of about 350 K. The other load (the "cold load") is allowed to float at the ambient temperature inside of the AIMR structure. NCAR's mounting configuration for AIMR (inside of the C-130 cabin) has led to cold load temperatures that are higher than the outside ambient temperature. This is unfortunate since AIMR is calibrated between these two loads and the response of the receivers is assumed to be linear outside of the temperatures spanned by the loads. However, much of the important data acquired by AIMR is well outside of this calibration region. This highlights the need for an ability to calibrate AIMR at temperatures higher and (especially) lower than the span of the calibration loads. These calibration points do not need to be checked regularly, during flight, as long as they can show that the linearity assumption for the receivers is correct. The temperature of each load is monitored with several precision calibrated temperature sensors. 2.4 THE DIGITIZER AND HANDLING AND DISPLAY OF THE DATA The rotational speed of the motor; the sampling rate of the A/D converters; and the gains, offsets, and filter time constants of the receivers are all controlled by a Motorola 68020 processor mounted in a VME bus chassis. These operational parameters are set to defaults each time the system starts, but the operator is free to change these parameters in flight. The sample rate of the A/D converter sets the sample spacing along the scan. This spacing is often set to 0.5 degrees so that each sample in the 90 GHz channels is completely overlapped by the two adjacent samples. With each rotation of the antenna, the temperature of the calibration loads and the observed brightness temperature of the loads is calculated and attached to the data stream. The actual counts out of the A/D converter for each of the four channels at each sample position is also recorded. The raw data are actually recorded across one calibration load, throughout the entire viewing aperture and then across the other calibration load. In this way the post acquisition analysis software can recalculate all parameters, if desired. In real time, the data for a single rotation are grouped together and sent to a Sun Microsystems computer for display and recording. The Sun computer uses the load brightness temperature as calculated by the VME bus computer to calculate the brightness temperatures for each of the channels. It also provides the ability to look at a display of the averaged brightness temperature of the 37 GHz or the 90 GHz channels. The real time display corrects the data for aircraft attitude and position, which provides a display with an aspect ratio of 1. 2.5 TIMING AND ACQUISITION OF THE ATTITUDE AND POSITION DATA On board the NCAR C-130 a universal time stamp is provided with a clock that is synchronized to the Global positioning System (GPS) time and then passed around the aircraft in IRIG-B time standard format. The AIMR system uses a clock card that is attached to this time standard. The clock is read and the current time is added to the data stream with each rotation of the antenna. In this way the AIMR data can easily be directly compared with any other instruments that might be in use on the aircraft. Also available on the NCAR C-130 is a stream of data in RS-232 format that contains the aircraft attitude, position, altitude AGL, and velocity data. These data are also ingested by the VME bus computer and appended to the data stream and recorded. It is these data that the system uses to control the scan speed of the antenna to maintain the desired along track overlap of the data when the aircraft changes altitude. 2.6 CONTROL OF THE SYSTEM An ASCII control file that is easily edited with a standard text editor provides many of the default operating parameters for the AIMR system. This file is read each time the VME bus computer is booted. The operator, through a Graphical User Interface (GUI) can also change many of these operating parameters in real time. Some of the many parameters the operator has control over include the limits for the built in test parameters. Many temperatures, the scan rate and a data quality parameter are checked on each rotation of the antenna for being out of bounds. If any of the parameters are out of bounds the operator is given a warning. Also, the operator can select the amount of overlap of the data along the track, with each scan of the instrument; whether to record actual aircraft parameters or a fake set of default parameters and which temperature sensors to use in the calculation of the actual calibration load temperatures. Other items under the operator's control include many things that do not effect the basic data that are written to tape. This includes complete control of the display. The display is capable of displaying any two of the 4 channels or their averages. The distance between the ticks on the display and the color scale for the brightness temperatures can also be set. There are parameters that can be set in the ASCII control file that fix the desired operating points (i.e. the actual voltage out of the receivers) while the antenna is pointed at the calibration loads. Moving the desired voltages from the hot load and the cold load, for a particular channel, closer together allows AIMR to cover a wider range of brightness temperatures (at lower resolution) and visa versa. Also moving the operating point up and down in the range of the A/D converter (while maintaining the difference in the desired load voltages) allows the AIMR to measure lower (or higher) temperatures, at the same resolution. Future plans for the AIMR GUI interface include allowing the user to select a high and low brightness temperature in real time, from which the operating points would be calculated and set. A thorough characterization of the AIMR receiver determined that the linear dynamic range of the receivers is large. Therefore, the ability to modify the AIMR operating point allows it to be used in many different applications. 3.0 REFURBISHMENT AT NCAR NCAR replaced much of the computer hardware used with the original AIMR and completely rewrote all of the real time and post acquisition analysis software. In the real time VME bus crate the sampling computer was upgraded to a faster processor and an IRIG-B interface was added. The control and display computer was upgraded to a Unix Sun Sparc-Station. Communications between the two nodes is now done over electronic Ethernet, not optically as done previously. The real time display was recreated in an X-window, and the controls and status information were recreated in a GUI interface. The post acquisition analysis software was totally rewritten to provide ground relative arrays of brightness temperature (including calculating the horizontal and vertical polarizations) from a single pass of the data tape. These arrays of data are written to disk files during the "straight and level" segments of a flight. Automatic detection of these segments is attempted by the software, but can be overridden by the user. Future plans call for additional programs to read these data files and allow the user to insert their own analysis code to produce additional parameters (or statistics) and to rewrite the data out in the same format. Also a converter from the NCAR described format of these files and a NetCDF file has been developed. 4.0 CALIBRATION OF THE INSTRUMENT As mentioned in Section 2.3 it is highly desirable to perform a calibration of the AIMR at temperatures that are on the same order as those observed with the instrument. Since the emissivity of unfrozen water (at these frequencies) is nearly 0.5, the brightness temperatures, observed by AIMR over water, are very low, often below 200K. Therefore, a calibration technique that will allow AIMR to look at an object of known brightness temperature on the order of 200K has been sought. An attempt was made to turn AIMR over and point it up at the nighttime sky. A radiosonde was launched to measure the humidity in the atmosphere. If a night is selected with very little wind and few clouds, the angular dependence in the resulting AIMR data can be assumed to be due to the varying amount of atmosphere it looks through at the different angles. The brightness temperature of the night sky is fairly easy to determine and it is very stable on a 1 hour time scale (about the time it takes the radiosonde to reach its maximum height). Unfortunately the brightness temperature of the nighttime sky in Colorado can be very cold, especially in winter when the humidity is low. The only serious attempt NCAR has made to perform a sky calibration of this sort occurred when the sky was too cold (less than 50 K) and was off the low end of the AIMR receiver. An attempt will be made this year in summer to find a sky that is not so cold. Another calibration load was constructed from large sheets of RF absorbing foam. This foam is specified to be able to absorb over 40 dB of RF energy and is often used to line anachronic chambers. Anything with this good of absorptive properties should have an emissivity of about 1.0. Two shallow pans were constructed, one the size of the foam and the other slightly larger. Precision wound platinum wire temperature sensors were inserted into the foam. The original results were very promising as the sensors and the AIMR read almost exactly the same foam temperature while everything was at thermal equilibrium in the laboratory. This shows that the original assumption that the emissivity of the foam is near 1.0 is a good assumption. Then liquid nitrogen was poured into the large pan and the smaller pan with the foam in it was floated on the liquid nitrogen. The temperature of the platinum sensors could be recorded as AIMR observed the foam. Very quickly AIMR was measuring temperatures much higher (up to 30 K) than those of the sensors. The reason for this discrepancy is still unexplained. The wires running to the sensors were in the liquid nitrogen, perhaps they pulled heat out of the sensors faster than it could be pulled out of the foam. The platinum sensors were not calibrated at temperatures as low as they experienced in this experiment, perhaps they need calibrated down to lower temperatures. This technique is still promising for several reasons. It is easy and inexpensive to do and could be performed while the AIMR is installed on the C-130. It also provides more than one data point, so if the AIMR receivers are slightly non-linear the method would show that. 5.0 USE OF THE NCAR MODIFIED AIMR The NCAR modified AIMR was mounted in the C-130 and used in the SHEBA program in May and July of 1998. In this program it was used to map the polar ice cap in a manner similar to how it was used at AES. Later in 1998, AIMR was used on the WIFE program to study wild fires. This is a totally new use for AIMR and much research is needed to develop post acquisition analysis software for this type of study. It is anticipated that AIMR can be used to map out an estimate of the fuel available around these fires as well as seeing through the smoke to map the fire directly. In February and March of 1999 AIMR was flown in the INDOEX program, south of the Indian sub-continent over the Indian Ocean. It is anticipated that AIMR can provide column integrated liquid and vaporous water content, below the aircraft, for this program. Work on retrieving these parameters from the AIMR data is proceeding. Formats: 6.0 DATA FORMAT DESCRIPTION FOR THE "GEO" FORMAT The AIMR post processed data are stored in a data format referred to as the "geo" format. This is a generic data format that can be used to hold data from any scanning radiometer (including MCR). However, here are a few AIMR specifics: 1) Each file begins with a "Global Header". This header consists of a Basic Header Information Block and twelve (12) Channel information blocks. There are no comment blocks. 2) The data for all channels, but the angle channels, are given in brightness temperature in degrees K, the angles are in radians. 3) The recorded channels are (in order): T37-1, T37-2, T90-1, T90-2, Avg T37, Avg T90, T37 Horiz, T37 Vert, T90 Horiz, T90 Vert, Angle 37, Angle 90. 4) All of the data are written in floating point. 5) The data pixels are always square (same length along track and cross track). Other Generic Data Format "Rules": 1. The Global Header Block (GHB) is written out once at the beginning of each file. Data Records (DRs) are written out for each row of pixels in each file. 2. The latitude and longitude values listed in the DR correspond to the center of the center data pixel. The data pixels "spread out" sequentially and are orthogonal to the aircraft track. 3. All instrument channels contain the same number and size of data pixels. 4. In most cases, MCR and AIMR data will be processed and output for straight and level flight segments only. Each straight and level segment will be output into an individual data file in accordance with the below data format. 5. The size of the DR (i.e., the number of pixels) cannot change within a given data file. 6. The Global Header Block consists of the Basic Header Information followed by as many Channel Information blocks as there are channels recorded. 7) The Data Record consists of the Basic Header Information followed by paired sets of Channel Headers and Data. There are as many paired sets in each Data Record I) THE GLOBAL HEADER BLOCK (GHB) A) The Basic Header Information (88 Bytes Long) Data Type Description Char (16 bytes) Instrument name (MCR or AIMR) Char (16 bytes) Project name (Ex: SHEBA) Char (16 bytes) Platform or site name (Ex: C 1 30) Char (16 bytes) Flight number or other mission number indicator Long Int Size of global header block (bytes) Long Int Size of each data record (bytes) Long Int Number of pixels (points) per data record Long Int Byte ordering flag (O=Little endin; I =Big endin) Long Int Missing data flag (Ex: -32767) Long Int Number of instrument channels (Ex: MCR=7; AIMR=4) B) The Channel Information (Repeated once for each channel) (48 Bytes Long) Data Type Description Char (8 bytes) Channel name Float Channel central wavelength or frequency Float Channel bandwidth Float Channel instantaneous field of view (IFOV) Float Channel swath width Long Int Channel data type flag (1 = byte data; 2 = 2 byte data; 3 = 4 byte long integer data; 4 = 4 byte floating data) Long Int Channel data descriptor (O = MCR spectral radiance values inmW/cm2-um-sr; 1 = AIMR brightness temperature values in K; other values as defined) Long Int Channel data conversion slope value (0 if not used) Long Int Channel data conversion intercept value (0 if not used) Long Int Byte offset to channel header, measured from start of data record (DR) Long Int Byte offset to channel data, measured from start of data record II) THE DATA RECORD A) The Basic header Information block (80 Bytes Long) Data Type Description Long Int Record number (O=first record) Short Int Year (YYYY) Short Int Month (MM) Short Int Day (DD) Short Int Hour (GMT hour value; 0-24) Short Int Minutes (0-59) Short Int Seconds (0-59) Short Int Milliseconds (0-999) Short Int Unused Float Instrument scan rate (revolutions/sec) Float Latitude (at start of data record, center of center data pixel) Float Longitude (at start of data record, center of center data pixel) Float Track angle (degrees); clockwise from N (N=O degrees) Float Aircraft heading (degrees); clockwise from N (N=O degrees) Float Altitude (AGL; meters) Float Altitude (MSL; meters) Float Wind speed (m/sec) Float Wind direction (degrees); clockwise from N (N=O degrees) Float True airspeed (m/sec) Float Ground speed (m/sec) Float Solar zenith angle (degrees); 0-90 degrees Float Solar azimuth angle (degrees); clockwise from N (N=O degrees) Float Pixel width (meters), cross-track Float Pixel length (meters), along-track B) The Channel Header (16 Bytes Long) Data Type Description Char (8 bytes) Channel name Float Channel central wavelength or frequency Long Int Channel data quality flag (O=good; non-zero values indicate bad data, with exact meanings of non-zero values defined in the "Comments" block in the GHB) C) The Data (Need to check the Data Type and multiply it times the Number Of Pixels to determine the length of this block. Note: this will not work in the case of type 3 data) Data Type Description Type determined by Channel data values value of channel data type flag in GHB Note: The Channel Header and Data pairs will repeat Number of Instrument Channels times in each Data Record 7.0 The netCDF format The following is a header dump from a typical AIMR netCDF file. netcdf aimrnetCDF1 { dimensions: record = UNLIMITED ; // (158 currently) Number_Pixels = 200 ; variables: short Month(record) ; short Day(record) ; short Hour(record) ; short Minute(record) ; short Second(record) ; short Millisecond(record) ; float Latitude(record) ; Latitude:units = "deg" ; Latitude:definition = "Aircraft Latitude" ; float Longitude(record) ; Longitude:units = "deg" ; Longitude:definition = "Aircraft Longitude" ; float Track(record) ; Track:units = "deg" ; Track:definition = "Direction Aircraft is Moving 0=North +=cw looking down" ; float Heading(record) ; Heading:units = "deg" ; Heading:definition = "Direction Aircraft is Pointed 0=North +=cw looking down" ; float Altitude_agl(record) ; Altitude_agl:units = "m" ; Altitude_agl:definition = "Aircraft Altitude Above Ground Level" ; float Altitude_msl(record) ; Altitude_msl:units = "m" ; Altitude_msl:definition = "Aircraft Altitude Above Mean Sea Level" ; float Wind_speed(record) ; Wind_speed:units = "m/s" ; Wind_speed:definition = "Wind Speed" ; float Wind_dir(record) ; Wind_dir:units = "deg" ; Wind_dir:definition = "Wind Direction 0=North +=cw looking down" ; float True_airspeed(record) ; True_airspeed:units = "m/s" ; True_airspeed:definition = "Aircraft True Air Speed" ; float Groundspeed(record) ; Groundspeed:units = "m/s" ; Groundspeed:definition = "Aircraft Ground Speed" ; float Solar_zenith(record) ; Solar_zenith:units = "deg" ; Solar_zenith:definition = "Angle Of Sun From Horizontal" ; float Solar_azimuth(record) ; Solar_azimuth:units = "deg" ; Solar_azimuth:definition = "Angle Of Sun From North +=cw looking down" ; float P37_1(record, Number_Pixels) ; P37_1:Frequency_or_Wavelength = 37.f ; P37_1:Bandwidth = 1500.f ; P37_1:Instantaneous_FOV = 2.5f ; P37_1:Swath_Width = 0.f ; P37_1:Data_Descriptor = 1 ; P37_1:scale_factor = 0 ; P37_1:add_offset = 0 ; int Quality_flag_P37_1(record) ; float P37_2(record, Number_Pixels) ; P37_2:Frequency_or_Wavelength = 37.f ; P37_2:Bandwidth = 1500.f ; P37_2:Instantaneous_FOV = 2.5f ; P37_2:Swath_Width = 0.f ; P37_2:Data_Descriptor = 1 ; P37_2:scale_factor = 0 ; P37_2:add_offset = 0 ; int Quality_flag_P37_2(record) ; float P90_1(record, Number_Pixels) ; P90_1:Frequency_or_Wavelength = 90.f ; P90_1:Bandwidth = 2000.f ; P90_1:Ins File Name Conventions: This data is an off-line dataset which is available on tape. Each tape contains several tar files which contain data files that are named sequentially (i.e., aimrOutputxx where xx is an integer number) with aimrOutput1 being the first file of a flight, aimrOutput2 is the second file of a flight, etc. All the files for a single flight are in one tar file. Only "straight and level" sections of each flight were processed, so there are a different number of files for each flight, depending on how many turns and altitude changes were made by the aircraft. The tar files also contain a file named "aimrOutput.files" which is an ASCII text file giving the start and stop times for each of the aimrOutputxx files. The flight naming convention was RFxx where xx is an integer. Note that these data are available in both the GEO and netCDF formats. References: None.