avcal [ parameter=value ... ] [ inputfile ... ]
Parameters are luts_only, radiance_only, save_cal_vars, temp_units, vis_coefs, line_interval, solar_corr, full_report.
avcal converts AVHRR sensor count values to values with engineering units. For channels 1 and 2 the units are percent albedo. For channels 3, 4, and 5 the units can be either percent albedo or brightness temperature. This process is often referred to as radiometrically calibrating the data. New calibration coefficients are computed for each 100 input lines.
The calibration algorithm is based on information contained in NOAA-NESS Publication #107, July 1979 (see NOTES). Briefly, the visible and near infrared data (channels 1 and 2) are converted from raw counts to percent albedo using a linear relationship determined prior to launch. The thermal infrared data (channels 3, 4, and 5) are converted from raw counts to radiances with a linear relationship that is based on the raw count value associated with cold space (roughly 3 degrees Kelvin) and the raw count value associated with the temperature of an onboard target (approximately 300 degrees Kelvin). A slight nonlinearity in channels 4 and 5 is corrected using a quadratic function of radiance. Lastly, the infrared radiances are converted to temperature using the inverse Planck function.
To calibrate visible channels, avcal uses the ASCII data file avcal.coef, the location of which is defined by the SATDATA environment variable. This is a time-dependent file, each record of which contains the calibration information on a per-satellite basis together with the date label, defining the starting date when this information became valid. See this file for more information.
The output report provides the intercept (intcpt) and slope (slope) for each channel's linear calibration equation. It also provides the space (space) and target (target) counts. Lastly, it provides the temperatures (average) of the four platinum resistance thermistors (prt), which are embedded in the onboard target.
The output data are stored as scaled 16 bit integers to reduce mass storage requirements. Subsequent processing functions convert these compressed values to their true values as needed.
One or more TeraScan datasets are required as inputs. There are no output TeraScan datasets; i.e. calibration is done in place with the results overwriting the input data.
OPTIONAL. This hidden parameter determines whether avcal writes lookup table variables to the input dataset, rather than modifying the actual AVHRR data. If yes, avcal outputs lookup table variables corresponding to each AVHRR channel in the input dataset.
Both temperature and radiance lookup table variables are output for the AVHRR IR channels. Temperature units are defined by the temp_units parameter.
The following are the names and units of the lookup table variables:
avhrr_ch1_lut (percent albedo)
avhrr_ch2_lut (percent albedo)
avhrr_ch3a_lut (percent albedo)
avhrr_ch3_lut (temperature)
avhrr_ch4_lut (temperature)
avhrr_ch5_lut (temperature)
avhrr_ch3_rad (radiance)
avhrr_ch4_rad (radiance)
avhrr_ch5_rad (radiance)Note that luts_only=yes does not allow for sensor recalibration. Only the results of the initial calibration are output as lookup table variables. The interval of this initial calibration can be set using the line_interval parameter.
Valid responses are yes and no. The default is no, which means do not output lookup table variables.
OPTIONAL. This parameter determines whether avcal converts raw AVHRR channel 3, 4, or 5 data to radiance or to temperature.
Valid responses are yes and no. The default is no, which means convert to temperature.
OPTIONAL. This parameter determines whether avcal creates additional calibration data variables in the given datasets. These variables are listed in NOTES. There are two classes of calibration variables: variables taken directly from the HRPT header, and variables computed during the calibration process.
avcal converts the raw AVHRR data to brightness temperatures with units of degrees Celsius, Kelvin, or Fahrenheit. Degrees kelvin or celsius should be used if the data will be used to estimate sea surface temperature using nitpix because nitpix does not currently support input data with units of fahrenheit.
Valid responses are [celsius, kelvin, fahrenheit]. The default is celsius.
OPTIONAL. avcal can overwrite gain and offset for visible channels with the user-supplied four real values of vis_coefs : slope (ch1), intercept (ch1), slope (ch2) and intercept (ch2).
Valid responses are any four real values. Accepting the default values (i.e. pressing RETURN instead of entering the actual values), no matter what they are, means that no overwriting will be done.
OPTIONAL. Defines the number of HRPT frames, after which the calibration coefficients will be updated.
Valid response is an integer value between 10 and 10240. The default value is 100 if luts_only=no. If luts_only=yes, the default value is 10240.
OPTIONAL. Determines whether or not to perform the solar zenith angle correction.
Valid responses are [yes, no]. The default is no.
If full_report=no, only the calibration report for the first 100 lines is printed. If full_report=yes, a calibration report is printed for every 100 lines of input data.
Valid responses are [yes, no]. The default is no.
Calibrate and report the contents of the TeraScan dataset named big. Notice that units, bad_value and real_scaling values all change after calibration. See bigdataset for information on the original contents of big.
[1] % avcal big
temp_units : char(10) ? [celsius]
full_report : char( 3) ? [no]
line 1 ch slope intcpt space target prt average
1 0.10634 -3.85 38.3 1 419.00
2 0.10749 -3.88 39.6 2 427.00
3 -0.00144 1.43 994.1 602.5 3 414.00
4 -0.17249 167.04 988.0 334.2 4 423.00
5 -0.19988 196.32 993.8 360.8
[2] % contents big
printout : char( 3) ? [no]
Dimensions:
1 1199 line
2 1410 sample
3 103
Global Attributes:
history byte
channels long 1 2 3 4 5
projection long std_projection sensor_scan
et_affine double 1 0 0 1 -67 -367
satellite byte noaa-9
sensor long std_sensor avhrr
pass_date long std_date 88/09/13
start_time double std_time 21:18:22.6
time_adjust double std_time 00:00:00
attitude double degrees 0 0 0
sensor_tilt double degrees 0
sensorient double degrees 0 45 0 55.3 -55.3
scan_samples long 2048
scan_rates double 6 40000 37.6991
orb_elem_date long std_date 88/09/13
orb_elements double 0.476962 0.828252 4e-05 0.001585
0.0363905 0.669068 0.275306
center_lat double std_latitude
center_lon double std_longitude
map_angle double degrees
Variable Attributes:
bad_value double
real_scaling double
Variables:
hrpt_header short
1 1199 line
2 103
avhrr_ch1 short albedo*100%
1 1199 line
2 1410 sample
bad_value : -32768
real_scaling : 0 100
avhrr_ch2 short albedo*100%
1 1199 line
2 1410 sample
bad_value : -32768
real_scaling : 0 100
avhrr_ch3 short temp_deg_c
1 1199 line
2 1410 sample
bad_value : -32768
real_scaling : 0 100
avhrr_ch4 short temp_deg_c
1 1199 line
2 1410 sample
bad_value : -32768
real_scaling : 0 100
avhrr_ch5 short temp_deg_c
1 1199 line
2 1410 sample
bad_value : -32768
real_scaling : 0 100
avhrr, nitpix, avin, avfix, fixline, avfilt, nitpix.
The calibration algorithm is based on information contained in Techniques for Data Extraction and Calibration of TIROS-N/NOAA Series Satellite Radiometers for Direct Readout Users by Levin Lauritson and Gary Nelson (NOAA-NESS Publication #107, July 1979).
The following variables are output as a result of save_cal_vars=yes.
Dimension Size Description line ??? number of AVHRR scan lines prt_samples 3 number of PRT samples in HRPT header cal_samples 10 number of space and target samples per channel in HRPT header cal_cases ??? number of AVHRR sensor re-calibrations prts 4 number of onboard PRTs HRPT Header Variables Units Dimensions ch1_ramp_cal line ch2_ramp_cal line ch3_ramp_cal line ch4_ramp_cal line ch5_ramp_cal line prt_array raw_counts line, prt_samples ch3_patch_temp line ch3_target raw_counts line, cal_samples ch4_target raw_counts line, cal_samples ch5_target raw_counts line, cal_samples ch1_space raw_counts line, cal_samples ch2_space raw_counts line, cal_samples ch3_space raw_counts line, cal_samples ch4_space raw_counts line, cal_samples ch5_space raw_counts line, cal_samples Computed Variables Units Dimensions cal_offset_line cal_cases cal_num_lines cal_cases cal_good_lines cal_cases cal_good_3a_lines cal_cases cal_good_3b_lines cal_cases cal_chN_offset percent_albedo cal_cases cal_chN_slope percent_albedo/raw_counts cal_cases cal_chN_offset2 percent_albedo cal_cases cal_chN_slope2 percent_albedo/raw_counts cal_cases cal_chN_space_median raw_counts cal_cases cal_chN_space_mean raw_counts cal_cases cal_chN_space_stdev raw_counts cal_cases cal_chN_space_nedn percent_albedo cal_cases cal_chN_target_median raw_counts cal_cases cal_chN_target_mean raw_counts cal_cases cal_chN_target_stdev raw_counts cal_cases cal_chN_target_nedt temp_deg_k cal_cases cal_prtN_median raw_counts cal_cases cal_prtN_mean raw_counts cal_cases cal_prtN_stdev raw_counts cal_cases cal_prt_median raw_counts cal_cases, prts cal_prt_mean raw_counts cal_cases, prts cal_prt_stdev raw_counts cal_cases, prts
The space and target statistics are computed after throwing out obviously bad data. For any given scan line (or HRPT minor frame), if any of the calibration data is zero, then all of the calibration information from that line is discarded.
The median is really the median line average, e.g., to compute cal_ch1_space_median, first compute the average channel 1 space counts for each of a set of scan lines (HRPT minor frames), and then take the median.
cal_chN_space_nedn is noise equivalent delta radiance, the product of the standard deviation of space counts data and the radiance/counts gain ratio. cal_chN_target_nedt is noise equivalent delta temperature, the product of the standard deviation of target counts data and the temperature/counts gain ratio.
In a calibration interval containing a transition between NOAA KLM channels 3A and 3B, if there are fewer than 5 good lines containing channel 3A data, then cal_good_3a_lines is forced to zero, and no attempt is made to calibrate channel 3A in the interval. Similarly, if there are fewer than 5 good lines containing 3B data, then cal_good_3b_lines is forced to zero, and no attempt is made to calibrate channel 3B in the interval.
Last Update: $Date: 2002/05/07 23:50:35 $