![]() |
Quality of MISR Radiometric Calibration |
MISR has challenging radiometric specifications of 3% absolute, and 1% band and camera-relative calibration. The radiometric calibration is facilitated by use of an on-orbit calibrator (OBC) experiment, which is utilized at bi-monthly intervals. During these experiments a diffuse panel reflects solar light into the cameras. The intensity of this light is measured with on-board photodiode "detector standards." With the measured incident radiance and output DN, radiometric calibration coefficients are computed. Following each experiment the coefficients are packaged into a file called the Ancillary Radiometric Product (ARP). This new ARP file is used for the next two months to produce the MISR Level 1B1 radiance product, and in turn the Level 1B2 geo-located product.
Because of the challenging radiometric requirements, the calibration team at JPL has been researching various procedures for reducing the calibration data and producing coefficients. Calibration-algorithm changes have occurred over time, and thus the initial ARP time-series files have not been produced using a common algorithm. The full history of this activity is reported in Table 2 below.
It is desirable to generate all reprocessed data at a uniform quality level which reflects the most current and accurate approach available to calibrating the instrument. To this end, the ARP time-series files have been regenerated twice during the mission with fixed algorithms. The first regeneration effort, conducted in January of 2004, and described in Table 3 below, was used to produce reprocessing Collection 4 data. The second and final ARP regeneration effort, described in Table 1 below, was conducted in November of 2004. The user must pay attention to the version number portion of the ARP_INFLTCAL Filename in order to determine the uncertainty of the resulting data. Users are advised to obtain data products generated with the ARP file version specified in table 1.
ARP ID | File Version* | Radiometric uncertainty (for uniform targets) |
---|---|---|
T002-T056 | F02_0010 | Same as T029, described in Table 2. |
* Filename conventions: Txxx = Time window, defined in Table 2. Fxx = Format version xxxx = revision number of this Time window file |
ARP ID | Valid For Data Acquired On/After: | Time | Orbit | Filename* | Calibration algorithm updates | Radiometric uncertainty (for uniform targets) |
---|---|---|---|---|---|---|
T056 | 04 Feb 2009 | 23:47:13 | 48590 | T056_F02_0010 | Same algorithm as T055. | Same as T018 |
T055 | 21 Jan 2009 | 23:34:51 | 48386 | T055_F02_0010 | CALFACT, the precursor to ARPGEN which calculates the calibration factors contained in the calfactor file, was modified to remove any dependencies on goniometer movement. Calibration sequences T054 and prior computed Pin3_cal and Pin4_cal based on the goniometer movement. Due to the MISR anomaly during the T054 calibration sequence on September 30, 2008, the MISR calibration sequences were modified to exclude any goniometer movement. Therefore, the algorithm which calculates the Pin3_cal and Pin4_cal calibration factors was changed to use predictions derived from historical goniometer calibration data through T054. For Pin4_cal a south_Pin2 / north_Pin2 ratio was also used to compensate for panel variations. Calibration sequences from T055 on will use this new algorithm. | Same as T018 |
T054 | 16 Oct 2008 | 17:54:17 | 46970 | T054_F02_0010 | Same algorithm as T031. | Same as T018 |
T053 | 06 Aug 2008 | 22:44:45 | 45939 | T053_F02_0010 | Same algorithm as T031. | Same as T018 |
T052 | 11 Jun 2008 | 23:34:21 | 45124 | T052_F02_0010 | Same algorithm as T031. | Same as T018 |
T051 | 04 May 2008 | 22:32:21 | 44570 | T051_F02_0010 | Same algorithm as T031. | Same as T018 |
T050 | 19 Feb 2008 | 22:50:44 | 43478 | T050_F02_0010 | Same algorithm as T031. | Same as T018 |
T049 | 16 Dec 2007 | 23:46:09 | 42532 | T049_F02_0010 | Same algorithm as T031. | Same as T018 |
T048 | 30 Oct 2007 | 22:50:45 | 41847 | T048_F02_0010 | Same algorithm as T031. | Same as T018 |
T047 | 29 Aug 2007 | 22:38:38 | 40944 | T047_F02_0010 | Same algorithm as T031. | Same as T018 |
T046 | 28 Jun 2007 | 01:00:34 | 40028 | T046_F02_0010 | Same algorithm as T031. | Same as T018 |
T045 | 01 May 2007 | 00:23:28 | 39183 | T045_F02_0010 | Same algorithm as T031. | Same as T018 |
T044 | 28 Feb 2007 | 00:11:21 | 38280 | T044_F02_0010 | Same algorithm as T031. | Same as T018 |
T043 | 15 Jan 2007 | 01:25:13 | 37640 | T043_F02_0010 | Same algorithm as T031. | Same as T018 |
T042 | 15 Nov 2006 | 00:16:38 | 36751 | T042_F02_0010 | Same algorithm as T031. | Same as T018 |
T041 | 06 Sep 2006 | 00:54:00 | 35732 | T041_F02_0010 | Same algorithm as T031. | Same as T018 |
T040 | 21 Jul 2006 | 01:37:18 | 35048 | T040_F02_0010 | Same algorithm as T031. | Same as T018 |
T039 | 18 May 2006 | 01:36:55 | 33819 | T039_F02_0010 | Same algorithm as T031. | Same as T018 |
T038 | 21 Mar 2006 | 00:59:46 | 33271 | T038_F02_0010 | Same algorithm as T031. | Same as T018 |
T037 | 20 Jan 2006 | 00:34:26 | 32397 | T037_F02_0010 | Same algorithm as T031. | Same as T018 |
T036 | 10 Nov 2005 | 00:25:18 | 31363 | T036_F02_0010 | Same algorithm as T031. | Same as T018 |
T035 | 26 Sep 2005 | 00:56:43 | 30707 | T035_F02_0010 | Same algorithm as T031. | Same as T018 |
T034 | 21 Jul 2005 | 00:26:35 | 29732 | T034_F02_0010 | Same algorithm as T031. | Same as T018 |
T033 | 01 Jun 2005 | 00:39:31 | 29004 | T033_F02_0010 | Same algorithm as T031. | Same as T018 |
T032 | 21 Mar 2005 | 01:28:20 | 27956 | T032_F02_0010 | Same algorithm as T031. | Same as T018 |
T031 | 31 Jan 2005 | 00:47:05 | 27242 | T031_F02_0010 | Subsequent to the overall band-to-band calibration adjustments made in ARP T024, camera-by-camera radiometric calibration refinements were made based on the combined analysis of multiple calibration techniques (Diner et al., 2004; Kahn et al., 2005). They amounted to 1% or less in all cases except the near-infrared channel of the Bf camera, which was reduced by 2.5%. | Same as T018 |
T030 | 19 Nov 2004 | 00:27:53 | 26237 | T030_F02_0001 | Same algorithm as T024. | Same as T018 |
T029 | 28 Sep 2004 | 00:22:20 | 25436 | T029_F02_0001 | Same algorithm as T024. | Same as T018 |
T028 | 29 Jul 2004 | 00:10:15 | 24533 | T028_F02_0001 | Same algorithm as T024. | Same as T018 |
T027 | 16 Jun 2004 | 23:46:38 | 23921 | T027_F02_0001 | Same algorithm as T024. | Same as T018 |
T026 | 30 Mar 2004 | 00:17:39 | 22771 | T026_F02_0001 | Same algorithm as T024. | Same as T018 |
T025 | 20 Jan 2004 | 10:48:50 | 21758 | T025_F02_0001 | Same algorithm as T024. | Same as T018 |
T024 | 5 Dec 2003 | 01:08:28 | 21023 | T024_F02_0002 | Band-to-Band Adjustment: The red band calibration was reduced by 3% relative to the calibration in use since the year 2000. The NIR band was reduced by 1.5% (Bruegge et al., 2004). Blue and Green band relative calibrations were not changed. Additional bug fixes amounting to roughly a 1% change include: calibrating the G1 coefficients over the slowly-changing clear atmosphere portion of the cal acquisition and fixing some minor time and angle problems in the ARP generation software. | Same as T018 |
T023 | 21 Oct 2003 | 00:19:32 | 20441 | T023_F02_0001 | Same algorithm as T022. | Same as T018 |
T022 | 20 Aug 2003 | 01:03:31 | 19524 | T022_F02_0001 | Minor Bug Fix to T018 algorithm. Solar Angles on the diffuse panel are computed at the correct time, removing a 1.5 degree error; and the aft-camera BRF scale factors have been adjusted appropriately. The resulting change in the calibration is less than 1%. | Same as T018 |
T021 | 09 Jun 2003 | 00:14:54 | 18475 | T021_F02_0001 | Same algorithm as T018. | Same as T018 |
T020 | 22 Apr 2003 | 00:15:25 | 17776 | T020_F02_0001 | Same algorithm as T018. | Same as T018 |
T019 | 18 Feb 2003 | 01:08:38 | 16859 | T019_F02_0001 | Same algorithm as T018. | Same as T018 |
T018 | 12 Jan 2003 | 00:41:19 | 16320 | T018_F02_0001 | Same off-axis corrected algorithm as used to generate T017, but an additional bug fix regarding sun angles on the calibration panel was made to the code which calculates BRF. | Absolute: 4% Camera and band relative: 2% Pixel relative: 0.5% |
T017 | 24 Oct 2002 | 00:42:14 | 15155 | T017_F02_0001 | Off-axis corrected algorithm. An error in the BRF database indexing has been corrected in the code used to process MISR calibration coefficients. | Absolute: 4% Camera and band relative: 2% Pixel relative: 0.5% |
T016 | 6 Sep 2002 | 00:42:47 | 14456 | T016_F02_0001 | Linear. A linear calibration equation is again implemented. The DN versus radiance curve is constrained to go through zero. This algorithm gives more consistent calibrations among data collected over the North and South poles. | Exceptions to the 4% absolute uncertainty are: AN camera
western edge is 10% too bright; AN camera eastern edge is 5% too dim.
Error varies between these values for other pixels, with best radiometry
at the swath center.
Relative uncertainties: Same as most recent ARP. |
T015 | 26 Jun 2002 | 01:33:03 | 13408 | T015_F02_0001 | No algorithm change. | |
T014 | 25 Apr 2002 | 01:21:24 | 12505 | T014_F02_0001 | No algorithm change. | |
T013 | 25 Feb 2002 | 00:05:41 | 11645 | T013_F02_0001 | No algorithm change. | |
T012 | 22 Dec 2001 | 01:02:27 | 10699 | T012_F02_0001 | BRF correction algorithm. Data from the on-board goniometer have been used to update the North calibration panel reflectance data base. The D-pin photodiodes are used to calibrate all but the nadir camera; the -yn-pn photodiode was used for the An camera only. We believe this to correct an aft-to-fore camera bias; camera-relative uncertainties using this algorithm are believed to be less than 2%. Level 1 data produced using this ARP are Provisional Data Quality. | Absolute: Larger than current algorithm. See T016. Camera relative: 4% Band and pixel: Same as most recent ARP (2% and 0.5% respectively). |
T011 | 19 Nov 2001 | 00:19:10 | 10218 | T011_F02_0001 | D-PIN photodiodes used to calibrate C and D cameras; HQE for all other cameras. | |
T010 | 26 Sep 2001 | 00:56:52 | 9432 | T010_F02_0001 | In an effort to study camera-to-camera relative calibrations, the D-PIN photodiodes were used to calibrate the C and D cameras (HQE photodiodes were used for other cameras). An evaluation of the camera-relative calibrations followed. | |
T009 | 11 Jul 2001 | 01:27:11 | 8311 | T009_F02_0001 | A 5% aft-to-fore camera-relative bias has been discovered. This bias is present for all data products produced through T011 (T012 has applied the final correction algorithm). For ARP T009, separate calibration coefficients have been developed for the photodiodes as they view the North panel (used for aft and AN-red, nir channels) and the South panel (used for the fore- and AN-blue, green channels). This change was expected to improve fore-aft camera biases. In practice, several additional updates were made in T010 though T012 to correct this problem. Photodiodes used for T009: D-PIN for D camera, HQE for others. | Absolute: Larger than current algorithm. See T016. Camera relative: 5% Band and pixel: Same as most recent ARP (2% and 0.5% respectively). |
T008 | 17 May 2001 | 01:19:09 | 7510 | T008_F02_0001 | Beginning with this delivery, a quadratic calibration equation is being used to convert the sensor data from DN to radiances values. This algorithm may change the radiances reported over dark, or ocean targets, by a few percent. Radiances reported over bright scenes are believed to be invariant with algorithm. Photodiodes used: HQE only. | |
T007 | 07 Mar 2001 | 01:17:44 | 6476 | T007_F02_0001 | Use was made of the June 11th Vicarious Calibration experiment to validate the Blue HQE photodiode. As a result of this study all radiances computed for MISR have been adjusted upwards by a factor of 1.10, irrespective of data acquisition time. The Linear Algorithm is used both to convert DN to Radiance and to generate ARP coefficients. | Absolute: Radiances are increased to remove a 10% bias from previous data products. Errors are still larger than the current algorithm. See T016. Camera relative: 5% Band and pixel: Same as most recent ARP (2% and 0.5% respectively) |
T006 | 19 Dec 2000 | 19:29:18 | 5351 | T006_F02_0001 | ||
T005 | 01 Nov 2000 | 21:06:26 | 4653 | T005_F02_0001 | ||
T004 | 29 Aug 2000 | 14:25:43 | 3717 | T004_F02_0001 | ||
T003 | 12 Jun 2000 | 4:16:43 | 2575 | T003_F02_0001 | ||
T002 | 24 Feb 2000 | 16:41:00 | 995 | T002_F02_0005 | ||
* Filename convention: F = Format version T = Time window xxxx = revision number of this Time window file |
Table 3 below describes ARPs that were used to create MISR reprocessing collection 4. These products were widely distributed. Collection 4 products will eventually be replaced by another reprocessing collection. This set of ARPs was generated with uniform quality in January, 2004, before the camera-by-camera adjustment was available.
ARP ID | File Version* | Radiometric uncertainty (for uniform targets) |
---|---|---|
T025-T026 | F02_0001 | Same as T024, described in Table 2. |
T024 | F02_0002 | Same as T024, described in Table 2. |
T019, T020, T022, T023 | F02_0003 | Same as T024, described in Table 2. |
T021, T018 | F02_0004 | Same as T024, described in Table 2. |
T017 | F02_0005 | Same as T024, described in Table 2. |
T015 | F02_0004 | Same as T024, described in Table 2. |
T002-T014, T016 | F02_0003 | Same as T024, described in Table 2. |
* Filename conventions: Txxx = Time window, defined in Table 2. Fxx = Format version xxxx = revision number of this Time window file |
Warning: Do not use MISR Level 1 radiance products with the version number F03_0021. Products with this version number were generated with ARP files containing an error in the band-to-band calibration. This error resulted in Red band radiances that were 6% too high and Near-Infrared band radiances that were 2% high. All data products containing this error have been removed from the Langley DAAC archive. The first ARP to include the erroneous correction was T024_F02_0001, but T017-T023 were redelivered with the error at the same time. The Langley DAAC used these bad ARPs between December 5, 2003 and January 23, 2004.
ARP ID | File Version* | Radiometric uncertainty (for uniform targets) |
---|---|---|
T024 | F02_0001 | N/A |
T018 | F02_0003 | N/A |
T017, T019-T023 | F02_0002 | N/A |
* Filename conventions: Txxx = Time window, defined in Table 2. Fxx = Format version xxxx = revision number of this Time window file |
Further information may be found in the following papers:
The MISR Access Data Table contains a link to the MISR Validation website for those who desire more details.